Biotechnology Applications for Banana and Plantain Improvement

Biotechnology Applications for Banana and Plantain Improvement INIBAP’s Mandate The International Network for the Improvement of Banana and Plantain (INIBAP) was established in 1984 and has its headquarters in Montpellier, France. INIBAP is an autonomous and nonproﬁt intergovernmental organization whose aim is to increase the production of banana and plantain on smallholdings by: – initiating, encouraging, supporting, conducting and coordinating research aimed at improving the production of banana and plantain, with particular reference to the people of developing countries; – strengthening regional and national programs concerned with improved and disease- free banana and plantain genetic material, and facilitating the interchange of such material by assisting in the establishment and analysis of regional and global trials of new and improved cultivars; – coordinating and supporting training programs for developing-country technicians and scientists. Planning for the creation of INIBAP began in 1981 in Ibadan with a resolution passed at a conference of the International Association for Research on Plantain and Bananas. INIBAP is a member of the Consultative Group on International Agricultural Research (CGIAR).

held in San José, Costa Rica 27-31 January 1992

INTERNATIONAL NETWORK FOR THE IMPROVEMENT OF BANANA AND PLANTAIN The opinions in this publication are those of the authors and not necessarily those of INIBAP.

Acknowledgments

INIBAP is grateful to all participants, who came from all over the world, for their contributions to the achievement of the Workshop’s objectives. Likewise, INIBAP is grateful to: • The Centro Agronómico de Investigación y Enseñanza (CATIE), especially to JV Escalant and the staff of the Biotechnology Unit, for hosting and handling the ﬁrst day of Workshop sessions. • The Corporación Bananera Nacional de Costa Rica (CORBANA) for its cooperation in the local arrangements and for its gracious hospitality. • The Dutch Ministry of Foreign Affairs, Special Programme on Biotechnology and Development Cooperation, for sponsoring the production of these proceedings. •B Wills and RD Huggan who enthusiastically undertook the task of editing the proceedings, and to F Malafosse for her dedication in preparing the necessary documents for the editors. •R Jaramillo, Coordinator of the INIBAP Network for Latin America and the Caribbean, who included even his family in the organization and administration of this event. • USAID for its participation in the organization of the meeting.

Editorial note

In some papers references have been submitted without complete publishing data. They may therefore lack article titles, the full names of journals, and/or the place of publication and the publisher. Unavoidable delays in publication have also led to the omission of full publishing data about some items listed as “in press.” Should readers have difﬁculty in identifying particular references, staff at INIBAP will be glad to assist. Contents Page

Introduction E De Langhe (Director, INIBAP): Opening Speech 7 R Tarté (Director General, CATIE): Introductory Remarks 9 I Buddenhagen: Whence and Whither Banana Research and Development? 12

Part 1: Nonconventional Strategies for Producing Banana and Plantain Resistant to Pathogens and Pests Genetic Maps D González de León, S Fauré Genetic Mapping of the Banana Diploid Genome 29 Transformation F Bakry, R Haïcour, JP Horry, R Megia, L Rossignol Applications of Biotechnologies to Banana Breeding 52 Identifying and Generating Resistance C Rivera, P Ramírez, R Pereira Preliminary Characterization of Viruses Infecting Banana in Costa Rica 63 CM Fauquet, RN Beachy Status of Coat Protein-Mediated Resistance and its Potential Application for Banana Viruses 69 J Dale, T Burns, S Oehlschlager, M Karan, R Harding Banana Bunchy Top Virus 85 GA Strobel, AA Stierle, R Upadhyay, J Hershenhorn, G Molina The Phytotoxins of Mycosphaerella ﬁjiensis 93 RC Ploetz Molecular Approaches to Identifying Fusarium Wilt Resistance 104 D De Waele Potential of Gene Transfer for Engineering Resistance against Nematode Attack 116 N von Mende, P Burrows, J Bridge Molecular Aspects of Resistance to Nematodes 125 Cross-Pollination Breeding PR Rowe, FE Rosales Breeding Cooking Bananas for areas with Marginal Growing Conditions 128 Part 2: In-Vitro Strategies for Musa Germplasm Handling O Arias Commercial Micropropagation of Banana 139 FJ Novak, H Brunner, R Afza, R Morpurgo, RK Upadhyay, M Van Duren, M Sacchi, J Sitti Hawa, A Khatri; G Kahl, D Kaemmer, J Ramser, K Weising Improvement of Musa through Biotechnology and Mutation Breeding 143 M Perea-Dallos Contribution to the study of Banana Anther Culture 159 MR Söndahl, C Noriega Bioreactor Micropropagation Technology for Rapid Commercialization Opportunities in Plantation Crops 163 Genetic Improvement D Vuylsteke, R Swennen Genetic Improvemement of Plantains 169 JV Escalant, C Teisson Somatic Embryogenesis and Cell Suspensions in Musa 177 H Leblanc, JV Escalant Induced Parthenogenesis to Obtain Haploid Plants in Musa 181 W Parrott Cell-Culture Techniques 183 Somaclonal Variation FX Côte, X Perrier, C Teisson Somaclonal Variation in Musa sp. 192 LA Withers Early Detection of Somaclonal Variation 200 Part 3: Technology Transfer JA Chambers, JI Cohen Donor Assistance in Plant Biotechnology for the Developing World 211 Part 4: Recommendations (Working Group 1) Nonconventional Strategies for Producing Banana and Plantain Resistant to Pathogens and Pests 225 (Working Group 2) In-Vitro Strategies for Musa 235 (Working Group 3) Technology Transfer 236 Annexes 1. Acronyms and Abbreviations 247 2. Participants 249 Introduction

E De Langhe 7

Opening Speech

E De Langhe

Ladies and Gentlemen, this Musa Biotechnology Workshop has been made possible with the collaboration and input of several persons and institutes. It is a real pleasure for me to explain their roles. Dr Rodrigo Tarté, Director of CATIE, has been continually and consistently encouraging his Center in developing biotechnology research, and was most supportive of the initiative to have the workshop take place in Costa Rica. CATIE is the Regional Base of INIBAP in Latin America and the Caribbean, and INIBAP also considers CATIE as an excellent center for the transfer of new technology into the region. The United States Agency for International Development (USAID) provided substantial financial support, allowing the participation of many specialists from the American continent. Moreover, it allowed Dr Judy Chambers to collaborate actively in the development of the workshop’s concept and technical organization. The Technical Centre for Agriculture and Rural Cooperation (CTA) of the EEC supports the participation of our colleagues from Africa and from Asia. It will also ﬁnance the publication of the proceedings. Dr Ramiro Jaramillo, our Regional Coordinator for Latin American countries, has, during the last 3 weeks, devoted most of his time to the numerous details regarding the practical organization of the workshop, and the fax-telex line with headquarters, especially with Ms Susan Fauré of our secretariat, was “red-hot” at times. His base of operations, during the last few days, was at San José airport, rather than at CATIE. Bob Huggan, our Editor, feels quite comfortable, having already received the text of about half of the presentations, and will no doubt be consulting you in the further preparation of the proceedings. His long-standing professional experience is a guarantee of the quality of that publication. Dr Ivan Buddenhagen has agreed to make the general introduction to the scientiﬁc contributions, and I have no doubt that he will do this in his typically provocative and neuron-stimulating style. Our Germplasm Research Coordinator, Hugues Tezenas du Montcel, and our Crop Protection Research Coordinator, David Jones, are here and are of course readily available to provide any technical information you may wish to ask for. INIBAP attaches particular importance to the recommendations of this workshop and on their follow-up within the Network. The background paper, which you have already received, explains what we expect from this event, and I will, by the end of the presentations, elaborate on how INIBAP expects to be helpful as a Network.

Director, INIBAP 8 Opening Speech

I am conﬁdent that the workshop will be productive in several senses. A good many scientiﬁc contacts will be made—which is part of INIBAP’s task, to encourage informal networking. If the workshop formulates guidelines regarding INIBAP’s future role in Musa biotechnology, then its output will be maximal. R Tarté 9

Introductory Remarks

R Tarté

On behalf of the Tropical Agricultural Research and Training Center (CATIE), I want to extend the warmest welcome to all participants in this extremely important workshop. As you all know, CATIE has been the host for INIBAP’s regional coordination for Latin America and the Caribbean. We have, together, undertaken several important activities with the purpose of improving the production of bananas and plantains in the region. Those activities are oriented towards coordinating the research efforts conducted by many institutions committed to this goal at a global level. The search for new technologies, the exchange of information, and the performance of joint collaborative activities require an institutional complementarity that must avoid the unnecessary duplication of efforts. Such complementarity and collaboration are regarded by CATIE as the most important elements of its strategic development. They are also, I believe, the essence of a networking mechanism like that promoted by INIBAP. And complementarity and collaboration are the only ways of facing today’s challenges of agricultural research and development: the challenge of increasing productivity while maintaining the natural resource base — the challenge of making agriculture sustainable, not only ecologically, but also economically, socially, and culturally. Such challenges perhaps comprise a greater challenge than the one faced by those whose research led to the Green Revolution some 30 years ago. Commodity and disciplinary research has evolved over many years into multidisciplinary farm research, and today into what we call multidimensional research. Multidimensional because the search for agricultural development that is sustainable must take into account the different dimensions involved in an integrated way: the ecological, the economic, the social, the cultural, the policy, and the institutional dimensions. And those require not only multidisciplinary research, but also integrated multi-institutional action. Multidimensional research, as we understand it in CATIE, can be regarded as research focused with an ecoregional approach. In other words, one that takes into account the integrated management of natural resources at a regional level along with the development of sustainable production systems. It is along these lines that the new agricultural research for the 1990s must be designed. It is along these lines that our efforts will give rise to the new era of

Director General, CATIE 10 Introductory Remarks

modernization of agriculture that is demanded as we approach the beginning of the 21st century: the era of modernization with sustainability, equity, and quality of life. Now, as the need for a new way of conducting agricultural research is outlined, where do we stand and where do we want to go in research on bananas and plantains? In this part of the world, banana production is regarded as a very modern operation because of the high technological level required to assure the highest productivity and the best quality fruit for our export markets. But it is also the most contaminating and ecologically damaging of all agricultural operations. All of the technology used in banana production to control major pests and diseases and to protect the fruit that is exported, is based on the heavy use of chemicals. In the last 20 or 30 years the major technological changes that have occurred in the protection of bananas for export against the threatening diseases, are changes in the kind of pesticide or fungicide used. Banana technology has depended on applications of chemicals, with new brands of products appearing with relative frequency, and pesticide trials have become major applied research activities. This also accounts for the very high cost of banana production. While this is true for bananas, plantain growers, who are usually small farmers operating on a small scale, are demanding low-cost technologies to improve their production and to control the severe incidence of diseases affecting their crops, particularly black Sigatoka. There is obvious need for a reorientation of banana and plantain research, a need to introduce the new concept of plantain research, and a need to introduce the new concept of modernization with sustainability that must guide all agricultural research today. This is now urgent, particularly in the light of the extensive expansion of the banana-growing area in Costa Rica and other Central American countries. For all those reasons, we must maximize our creativity and imagination by planning our research activities through a major, unselfish cooperative effort among our scientists, policymakers, and farmers. There are many opportunities for transforming research on bananas and plantains. Such opportunities must be found in the more fundamental areas related to the biological and behavioral aspects of the major pests and diseases, such as black Sigatoka and nematodes, in order to help identify integrated pest management methods with emphasis on biological, genetic, and cultural methods of control rather than on chemicals. But perhaps the most challenging opportunities are in the field of biotechnology. CATIE has taken up such a challenge by designating biotechnology as the main research activity of the plantain working group that was set up 3 years ago. With the support of INIBAP, this research involves micropropagation, somatic embryogenesis and cell suspensions, haplomethods, somaclonal variation, and medium- and long-term germplasm conservation. The use of biotechnology for the improvement of bananas and plantains will, no doubt, contribute to the modernization of their production on a sustainable basis, provided that efforts are directed towards the reduction of environmental hazards and production costs and to the development of new production systems for small farmers. I sincerely hope that the creativity needed, enhanced through worldwide collaboration in R Tarté 11

the ﬁeld of biotechnology, will emerge as a very important outcome of this workshop. But we must also bear in mind that a multidimensional research approach must include interaction with other disciplines in the biological, economic, and social sciences as well as with those related to policy, services, and marketing environments. I am sure CATIE and INIBAP will remain committed to this goal, and will ﬁnd new and effective ways of coordinating this type of research. Let me give you, once again, our warmest welcome and best wishes for a very successful meeting. 12 Whence and Whither Banana Research and Development?

Whence and Whither Banana Research and Development?

IW Buddenhagen

Preface This paper is dedicated to RH Stover and NW Simmonds — the two men who have contributed the most to banana research both directly, and indirectly by influencing others. Through thick and thin of political, bureaucratic, and funding problems they have remained steadfast to scientific enquiry and useful accomplishment on this wonderful crop over their long careers.

Introduction I view my task at this workshop as providing a historical setting of banana research and banana problems to give perspective to the modern biotechnologist who wishes to apply wonderful new techniques to this ancient crop. I use “banana” generically, to cover also plantains and various cooking bananas: in short, to mean Musa. Banana research has had a fascinating but chequered history. Much has been learned; but the inability to develop a good long-term organizational structure and support system for multidisciplinary teamwork research has slowed progress. But it is important to realize how different the tropics were even only 50 years ago. French, British, Dutch, and Belgian empires dominated much of the tropics, except for Latin America which had thrown off the feudal Iberian yoke before anyone had evolved the concept of applied agricultural science. The Spanish empire fragmented into many small and a few larger tropical countries whose elites often lived in the highlands and had no interest in lowland (banana) agriculture. (And little enough interest in agriculture per se except as a taxable resource.) Much later, as science became applied to agriculture, the Latins and the Filipinos became influenced — and educated — directly or indirectly by the United States. But to back up a bit, “modern science” evolved in Europe, mostly in northern Europe, in spite of an abortive try long before in Greece and elsewhere. Science that reveals an understanding of biological phenomena and of plants, which could be applied to agriculture, is very recent — say 150 years or so — and it was related to the unique

Department of Agronomy and Range Science, University of California, Davis, CA 95616, USA IW Buddenhagen 13

north-European creation of the “industrial revolution”. Both activities were rapidly adopted and amplified in the USA, with its early strong north-European heritage and connections. Also, where new societies were formed of British people—in Canada, New Zealand, and Australia. They were also adopted soon by the Japanese. But in the tropics, the ancient cultures, rich in many ways but poor in others, remained largely outside the advances of biological science and the industrial revolution. Neither science nor agricultural science evolved in any tropical country. What did ﬁnally develop was the application of evolving agricultural science through the various European colonial services to the crops of interest to the home country — oil palm, cocoa, coffee, cotton, various spices, tobacco, sugarcane, and rubber.

The Early Days: 1900 to the Second World War (WW2) The Americans (USA) started the banana trade without any science, just buying, shipping, and selling bananas to fruit-hungry Americans. The British became interested through the potential of bananas for their Caribbean Islands, and the British market, and they went at it with typical British-botany bias—doing the first good work on the crop in taxonomy and cytogenetics. They even initiated a breeding program. This British initiative, beginning in the 1920s, was the first real flowering of banana research. The American banana companies (especially United Fruit Company) were soon confronted with the dying of their one and only “shipping” banana, Gros Michel, by what became known as Panama disease (Fusarium wilt) and they, too, started an abortive breeding program — in Panama in the late 1920s — and hired the great German Fusarium expert O Reinking to look at the problem. He went to Asia to collect resistant bananas and collected the eventual replacement for Gros Michel, in a yard in Saigon in 1926. The great depression caused the abandonment of the company breeding program, but Sigatoka — the great leaf blight of bananas —finally reached the Americas in the 1930s from its Asian homeland, causing a major shakedown of the Latin American banana trades. The French developed the agronomy of the crop on their hilly islands of Martinique and Guadeloupe and, much later, serendipitously discovered the effectiveness of oil spray for Sigatoka control. Local country agricultural departments and agricultural universities developed slowly in the poor Latin American countries and banana research hardly developed. The idea, if considered at all, was that the banana companies should do banana research. Thus, the companies and the Latin American peasant farmer, growing plantains or bananas, were left to their own devices. Since the USA had no colonial service and no tropical empire and little public interest in tropical agriculture, their potential agricultural research skills had little public-sector outlet. Their three tropical territories — Puerto Rico, Hawaii, the Philippines — were not seen as potential banana producers and, although the USDA 14 Whence and Whither Banana Research and Development?

established agricultural research positions in these territories and agricultural universities were also established, no long-term commitment was made to conduct research on bananas. There was no visionary in the USDA on tropical agriculture, let alone on bananas. There was no-one to link private and public sector research interests and, indeed, the US attitude was and still largely is—let the companies do their own research. Thus, the poor tropical peasant farmer remained short-changed, as did tropical agriculture in general, including the large economic sectors dependent upon the banana trade. Of the three American tropical territories, Hawaii became the most westernized in agricultural tradition through American company developments of sugarcane and pineapple (not bananas), and very successful research organizations were set up on these crops, across company interests, serving all companies contributing to the Hawaiian Sugar Planters’ Association, and the Pineapple Research Institute. Later, both USDA and university input to these research organizations developed. Such good models were never copied for bananas by the American companies operating in Latin America (or later in the Philippines), even though several were the same companies that proﬁted from the pineapple and sugarcane research groups. Thus, no collective large research group was ever formed for bananas, and no link with USDA or universities was established. (As an aside, the C4 photosynthetic pathway was first discovered by scientists at the Hawaiian Sugar Planters’ Association.) Back in the homeland of Musa, even less was done. The Dutch established much good tropical agriculture work in their “Indies” (now Indonesia) but they hardly touched bananas nor educated the local people in agricultural science. The heartland of Musa evolution was held by the French (Indochina) and the British (Malaya/Borneo/Burma), but in these regions essentially nothing was done on bananas. Although the Americans established a big collection in the Philippines by 1915, little else was done. Apparently, the colonial powers saw bananas in these regions as something nice to eat, and appreciated their diversity in local dishes, but basically accepted bananas as part of the background of the native cultures, which were exotic, complex, and left alone. New outside crops were seen as exploitable — oil palm, rubber, tobacco, etc., and these received all the nascent tropical research. Back in Africa, the Germans, French, and Belgians did just enough to try out the crop and cover its local agronomy. It was in the Caribbean and Central American countries where the industry was pioneered before the 1st World War where initial successes led to large investments, and practical problem-solving and development activities on a big scale. Making a commercial success of bananas meant exploiting the north American and European markets, and it required trustable suppliers of quality fruit, new shipping and handling practices, good and fast communications, diversified geographic sources, and fertile, well-drained soils. The “banana business” required organization and efficient engineering, shipping, and marketing. Engineers and shippers ran the business, not agriculturalists. Everyone did the necessary local agronomy, but the key stimulus to banana research came with the early appearance of Panama disease, both in IW Buddenhagen 15

northeastern South America, Panama, and Central America and then in the Caribbean and far-away Australia. This devastating and intractable Fusarium wilt was not understood, was spread everywhere in planting stock, and was fought in various unsuccessful ways.

The Post-WW2 Roller-Coaster to the Early 1970s Since the American banana companies (especially the United Fruit Company) were the primary victims of Fusarium wilt, they invested in various attempts at finding a solution. It was their “company” problem. They hired plant pathologists but their key approach was escape—move to new lands. As escape to new lands became more and more costly, both economically and politically, “the Company” invested more heavily in research. A new breeding program was initiated in Honduras in about 1960 and this, combined with a last-ditch switch to Fusarium-resistant Cavendish, stimulated the second great flowering of banana research. Multidisciplinary research on postharvest diseases and physiology was established, as well as research on resistance, breeding, and cytogenetics, epidemiology, etc. New non-banana men at the top of the research management ladder opened up the narrowness developed under the old banana hands. Consultants from different fields were hired and many new things happened. This flowering began in about 1958, grew more expansive (and more resented) in the mid-1960s and gradually withered with political unrest, antitrust action, and the buying out of the United Fruit Company by first one group and then another. Banana overproduction reduced prices and profits; new management at the very top viewed bananas as a commodity like wheat, and saw little commitment to long-term tropical agriculture, to tropical-country development, or to investing in the future through research. The other big banana companies also were “acquired” by conglomerates, and even Del Monte is now only a subsidiary of a big, almost bankrupt, British conglomerate. Early on, the British considered the banana wilt problem as a colonial problem affecting their prosperous Jamaican colony, not as a company problem. Thus their approach was to have the problem shouldered by the Imperial College of Tropical Agriculture in Trinidad and by the Jamaican Government Department of Agriculture. The prewar breeding and research programs were resurrected after WW2, with new collection expeditions and further research in botany and cytogenetics. But by 1960 all work was shifted to Jamaica under the Banana Board Research Department and the main “spring” of inspiration from the Imperial College dried up. But the important legacy from this source came to us through two superb books by NW Simmonds, published in 1959 and 1962. Following Jamaican independence, support to the Banana Board organization dwindled and little research is supported there today. But the key point is that the early banana breeding effort sparked the basic cytogenetic studies of bananas and much of the basic work on bananas. Much credit was due to the liberal policies of the Imperial College of Tropical Agriculture, which encouraged the development of long-term basic research programs. But that is past, and nowhere else did such an attitude prevail, nor does it fully exist today anywhere. 16 Whence and Whither Banana Research and Development?

The French only gradually recovered an effective involvement with their tropical colonies after the war and the agitation of independence movements. But in the Caribbean, their islands remained a part of France and, early-on, they converted to wilt- resistant Cavendish. This stimulated a continued intensification of research through their federal (public) overseas fruit research organization (IRFA), headquartered in Guadeloupe and Montpellier, France. (Like the British, the French viewed bananas as a national commitment, not as a company problem.) Work centered around plant protection, nutrition, general agronomy, and shipping problems. Their early growing of Cavendish and their good research on “efficiency” enabled them to avoid being concerned with a breeding program until quite recently.

Historical Forces Driving Banana Breeding and Other Banana Research Although Fusarium wilt was the driving force that led to the first (Trinidad) and second (United Fruit/Honduras) flowering of banana research and breeding programs, it lost its punch with economic success of the conversion of company bananas from Gros Michel to resistant Cavendish clones. Banana research had not flourished anywhere else, even in the presence of Fusarium wilt, in regions not participating in export markets or where local people had a diversity of clones only some of which were susceptible. There was little stimulus to start new, or continue old, breeding programs once Cavendish removed the Fusarium threat. Those areas converting first had the least need to do banana research except for adaptive agronomy. However, once the United Fruit Company switched to Cavendish, the push was for high yields and high quality and, with the new variety, new problems, as always, appeared. Thus, the second banana research flowering in Honduras (1958-75) grew and continued, but with different objectives — more efficient control of Sigatoka, nematode, and postharvest fruit disease and insects. Also, more efficient nutrition and agronomy and fruit quality control. Others had to play catch- up, be subsidized (as for the French and British markets), or wither away. A temporary spark for research and breeding was kindled by the appearance of black Sigatoka in Honduras in 1972 and its subsequent spread as far as Colombia. Soon, however, a combination of good epidemiology research and new systemic fungicides reduced control costs so that, once again, the spark for supporting resistance breeding dwindled. The pathogen has not yet got to Jamaica, and this absence combined with political changes destabilized research support and the breeding program there collapsed. By 1975, interest in a new bred-variety also decreased in the United Fruit Company. Partly to blame was the lack of success from decades of investment in breeding. But the problem had other roots. The Company breeding program had developed as the British/Caribbean program was waning, but there was hardly any connection with this experienced program. Discussions were held and attempts made to obtain germplasm, but no agreement could be reached. The Company went at it alone, hiring new IW Buddenhagen 17

inexperienced people, and never utilizing the British talents directly. But they hired good people and mounted three very successful collection expeditions to Asia and took every lead from Simmonds’ excellent books. Enough money and talent led to the best multidisciplinary effort ever mounted on bananas, to many new research findings, to the largest well-classified germplasm collection, to the development of the best male hybrids and to the best tetraploids for a dessert banana. But there were several key problems: the breeding target kept changing as practical Company agronomists switched from Gros Michel to Valery and then to Grande Naine. This shift to ever-shorter and higher-yielding cultivars made for a shifting and much more difficult breeding target. Pathology research and new chemicals leading to cheaper control of Sigatoka, and then of black Sigatoka led to less interest by growers and managers in any new untested (and possibly untrustworthy) cultivar. Moreover, Company management became convinced that any new cultivar would be stolen and used by others and the logical question — why subsidize everybody else?—was asked. (The logical answer—get together on funding with the other companies and enlist public funds from governments with international interests for the common good—was, unfortunately, never approached.) Furthermore, in the 1970s, various Latin American exporting countries formed their own private or public organizations dealing with export, shipping, policy, applied research, and with self-financing through box taxes. These activities, plus strikes and labor movements, led to a decrease of transnational involvement in land ownership and banana growing and less interest in “Company” investment in research and in breeding.

Other Initiatives Since the Mid-1970s

Musa and the CGIAR The CGIAR is the “overseeing body” of the international agricultural research centers (IARCs) that grew out of original initiatives by the Rockefeller Foundation. The first centers were concerned with cereals but, with the formation of CIAT in the 1960s, the clonal crop cassava became a research target. I was working on bananas in Central America at the time and was interviewed by the organizers of CIAT for a position on rice. My suggestion that Musa should be an important research target and that they were a staple food was met with skepticism and lack of interest. The formation of IITA in Nigeria as the CGIAR center for the lowland humid tropics added more clonal crops to the CGIAR system in about 1972 (Dioscorea yams, cocoyams, sweet potatoes). In addition, plantains were mentioned as a possible crop, but no advocate came forward to develop a research program on them. Bananas, of course, were not mentioned. (Bananas, for the CGIAR in general, and for donors who live far from the tropics, have a negative connotation because they can be a commercial crop from which proﬁt can be derived by big companies. Their true nature and pervasive importance in local economies are little appreciated.) However, by 1976 a meeting was held at IITA to form an International Association for Research on Plantain and Other Cooking Bananas. 18 Whence and Whither Banana Research and Development?

IITA’s charter was revised and “plantains” as a crop was made more clear. IITA’s initial work was minor and was all on plantain agronomy in poor delta soils. Management, board, and donors saw no reason to ﬁnance scientists to work on plantains full-time, nor on a breeding program, nor on diseases; and dessert bananas were never included as a crop. Thus, there was a CGIAR center in the country of greatest plantain production in the world (Colombia), one in the Philippines, at a humid lowland site ideal for Musa research and in a center of origin of Musa (IRRI), and one dedicated to the lowland humid tropics but located in a non-Musa seasonally-dry site (IITA, Ibadan). Musa was ignored by the ﬁrst two and was given only cursory treatment at IITA until about 1983. It was not until 1985 that IITA hired a full-time scientist for plantains; and breeding work was begun only in 1987 and a breeder hired only in December 1991. No-one had had the vision in the IARCs nor in the CGIAR or its Technical Advisory Committee (TAC) to mount a study on the importance, problems, and research needs of Musa worldwide, and to go after solutions in a logical manner. Meanwhile the key banana research center and only viable breeding program (United Fruit Company in Honduras) was dying by default.

The Brazilian initiative About this time (1981), EMBRAPA of Brazil decided to go after solving their long- standing problem of Fusarium wilt and to preclude any problem from a future invasion of black Sigatoka for their very different AAB Brazilian dessert bananas — through breeding. They very sensibly dealt with the senior breeder (K Shepherd) in the waning Jamaican breeding program. Shepherd, who inherited all the old British breeding/genetics knowledge and germplasm, and who knew what to do, set up an operation in Bahia and went after a Fusarium-wilt resistant dessert AAB type. This breeding target was both “new” and much easier than trying to beat the best Cavendish, since the existing dessert AABs have low yields and poor agronomy. By dealing with Costa Rica for ﬁeld screening of good progeny against black Sigatoka, they were able to obtain resistance to this pathogen also. Brazil has recently released a new dessert tetraploid (AAB characters) from this program. Anyone familiar with the variability of Fusarium may wonder how good the resistance will be, and it is not yet clear to an outsider how well the work of the program is continuing and how much multidisciplinary research is involved. Nevertheless, this EMBRAPA initiative was useful to help get banana breeding off dead-center.

The United Fruit Co./FAO/USAID/FHIA Meanwhile, back in Central America, the 1970s were fraught with problems of antitrust, new Latin exporting groups, tariffs, overproduction, hurricanes, buy-outs, and takeovers. Not a good climate for research, and only a few scientists remained in the United Fruit Company’s research group at La Lima. The advent and spread of black Sigatoka in 1972- 80 gave a new brief spark to breeding but, by 1983, the Company even gave away its 24- year-old breeding program. IW Buddenhagen 19

Actually, the idea was good—to create an “Agricultural Research Foundation” with the Company facilities, which could work on many crops to benefit Honduras at large. USAID and FAO were alerted to this opportunity, and they responded in different ways. A man with vision at FAO Rome (A Bozzini) decided the Musa germplasm and breeding program were too precious to let die. FAO is not a research-granting agency, but special funds were made available for 6 months (later renewed for another 6 months), to maintain the germplasm and PR Rowe’s breeding program at survival level. Also—and this is a key point—FAO sent a lawyer to help draft the FHIA charter, who insisted that the Musa part of the “Foundation” would be internationalized and not be just a Honduran activity. This created two great opportunities. 1. The breeding program could be internationalized and broadened in scope and in clientele. It could approach world needs without the past company narrowness. Moreover, while under FAO, the germplasm could become available to the world community. 2. The breeding program and Musa research overall could become the group for receipt of international monies from international donors of the CGIAR and others for solving any Musa problem, anywhere, that would be amenable to a breeding solution. Unfortunately, neither great opportunity was realized, although some progress was made. The FAO connection was a short bridging one before a USAID commitment of 10 years to FHIA was finalized. But this AID connection came out of the Latin American desk and had no connection with the part of AID that partially funds the CGIAR system. Moreover, they were not interested in supporting research on the banana export crop and they pushed for work on other crops. Musa was not really separated administratively within FHIA or in the minds of the US donors or their program reviewers. The connection with the international agricultural research system was never really made. Musa research was hardly expanded, nor did it become multidisciplinary and fully rounded. Other forces were at work to undermine the great opportunities. 1. The new Latin American marketing and exporting unions and other banana or research-related organizations could not grasp that a really good noncompany breeding program would be beneficial for themselves, as well as others. They could not separate FHIA La Lima from the “Company” La Lima, and the “Company” carried too much negative stigma. 2. IDRC in Canada, in looking around for a new “networking” project, seemed willing to put money into Musa unconnected to the CGIAR system, and listened to their Latin contacts who considered that the Honduras program should not be backed. (But IDRC did put money directly into FHIA in support of the small breeding program.) 3. The French and Belgians wanted to establish their own “international” Musa organization, mainly based on their African experience and the connection with the growing internationalization of IITA’s small plantain program, and IDRC Canada was also interested. 4. IITA leadership could not visualize the world needs or opportunities on Musa, a crop considered to be just one more of its too-many-crops roster. Thus, the stage for the formation of INIBAP was set. 20 Whence and Whither Banana Research and Development?

The Taiwan situation Meanwhile, at about this time (1983), Fusarium wilt had become so common on the “resistant” Cavendish clones growing in Taiwan that a solution was needed. This “breakdown” of Cavendish resistance, starting in the late 1960s was also a red ﬂag of fear to the tropical banana trades. It helped fuel the bare maintenance of the Honduran breeding program. It stimulated research because it was gradually realized that “new” Fusarium races or strains could attack resistant Cavendish also in subtropical Australia, South Africa, and the Canaries. Even Cavendish in tropical Philippines was attacked, but at a low level. The solution in Taiwan was to work out an economical replanting rotational system with paddy rice. The key was cheap, efficient raising of clean planting stock through meristem tissue culture. This provided the possibility for detecting somaclonal variants from large populations —and just maybe, resistant ones. The dividend from a straight commercialized laboratory propagation system, of a Cavendish resistant to these resistance-breaking strains, would be great indeed. Most variants have poor agronomic characters, even if they appear to have short-term resistance. However, some good variants have been identified which also appear to have some resistance (Hwang, Ko 1987; and personal information). This program appears not to have been broadened by attempting mutation treatments or in-vitro selection. These possible approaches were tried for a short time in the nearby Philippines by Del Monte Corporation, an effort abandoned in 1983 (Epp 1987).

IAEA, Vienna The International Atomic Energy Agency in Vienna has a breeding and genetics section which conducts and sponsors mutation breeding activities in many developing countries. This long-standing joint FAO/IAEA division is funded by the United Nations, and among its activities is a banana mutation/somaclonal laboratory at Seibersdorf near Vienna. Somaclonal material of potentially mutated variants is sent to any cooperator interested in testing it. Somaclonal projects on Musa have been backed in several countries, including Costa Rica. Although mutation treatments may be given, in-vitro selection with fungal-product pressures have not yet been well worked out. However, interest has been sufficiently strong to generate research on somatic embryogenesis (a natural for mutation breeding and now for biotechnology initiatives). Success in somatic embryogenesis has been reported (also at Louvain) and we should be brought up to date at this workshop. It seems that IAEA is a logical place to add modern biotechnology efforts to its existing older biotechnology efforts of mutation, applied to crop improvement. They wish to do so.

The unique Australian experience (old and new) Over the years the pragmatic Australians gradually expanded their banana production (mostly in Queensland) to satisfy Australian demand. Long ago bunchy top virus was accidentally brought from Fiji and this prompted research and quarantine activities. IW Buddenhagen 21

Banana research developed much as it does on crops in the USA — with farmers organizing growers’ groups and pushing for public-sector help through extension and research — centered at both the State Agricultural Department (QDPI) at local agricultural universities and at the Department of Agriculture in New South Wales. I think this unique situation for bananas was due to the very nature of society and agricultural development there. It is the only place where banana farmers and workers were not part of a peasantry separate from a ruling elite, but were part and parcel of a locally developing economy which became self-supporting largely through a marriage of independent enterprise and local government concerned with local development of education, industry, and farming, and a tradition of self-help through research, extension, and education. The farmers and farm workers were just as good as anyone else, and anyone might end up as an elected representative or with a PhD in agriculture. Research was done as was locally needed, mostly on agronomy, but also some good physiology and pathology have been done over the years. Banana researchers have long-term careers, though on a small scale, because funding and horizons largely reﬂect the problems of bananas in Queensland. A major push to broaden the Queensland scope was provided by ACIAR (Australian Centre for International Agricultural Research) in developing an international banana project in 1985, focused on international concerns of black Sigatoka, bunchy top virus, and breakdown of Fusarium-wilt resistance in Cavendish. The project was ostensibly focused on the “Pacific Islands” for the international dimension, with the first two problems. But the latter two problems were also Queensland problems, so a base for research was already ongoing. Adding ACIAR monies also gave a “germplasm diversity” dimension and the project became a focus for the IBPGR Musa collection from New Guinea, and a base for evaluation. Thus, research became much broadened and deepened so that, now, monoclonals and a DNA probe for bunchy top virus exist, and work on tissue-culture variation and mutation, as well as basic work on Fusarium variability, is carried out. The ﬁnal pay-off would be for direct involvement in a breeding program, but this most important and interesting aspect did not quite get off the ground.

FAO and IBPGR FAO continued its interest in Musa following its brief banana breeding rescue initiative, but mostly in the pure plant protection area. Reviews of black Sigatoka and of banana research in South America were made (1984), but the lack of mechanisms for follow- through action by this nonresearch organization more or less invalidated this effort. In 1990, FAO pushed an initiative to hold a meeting (jointly with INIBAP) to help resolve the conﬂicting reports of what causes banana bunchy top. Also, they have teamed up with IBPGR to write a series of guidelines on the safe international movement of germplasm, one of which covers Musa. It is, of course, important to have “safe” movement of germplasm but, like many quarantine initiatives, one can always use them for delaying or inhibiting germplasm movement to one’s own advantage. The Musa guidelines are, in my view, overly restrictive, overly centralized, and yet still unsafe, and they have tended to 22 Whence and Whither Banana Research and Development?

inhibit or delay exchange of material among researchers — the more-apparent ready- availability and safety of which would have accelerated cooperative research. FAO has been pressured by some of its developing-country members to help them not to be left out in the race to capitalize on modern biotechnology. FAO has recently hired a plant biotechnology officer to manage FAO’s incipient activities in this area. Also, a strategy paper has been developed that plots a course of action which attempts to balance any new biotechnology initiatives with the old crop improvement projects and extensive country knowledge of FAO staff. It remains to be seen how efficient an FAO role in brokering useful activities in this area can be. Without substantial expansion, FAO may find it difficult to become a viable player in this complex area. At FAO, but completely unrelated to the Crop Production and Protection Division, is housed an International Commodity Group which concerns itself with helping the big agricultural commodities of international trade. Their concern is stabilization of prices through various mechanisms. Banana is one of their commodities and, over the last 5 years, they have decided to try to apply the reserve funds from banana trade levies to lowering production costs by backing research projects. Much discussion over division and use of these funds is under way. Theoretically, some could go to useful biotechnology. IBPGR, earlier part of FAO, is now essentially an autonomous (and newly renamed) center within the CGIAR. Musa is just one of over 200 crops they are concerned with, and how active a role they will take in being involved or catalyzing funding for collection, analysis, and evaluation of Musa remains to be seen. They were involved in a small way in helping to catalyze the recent collecting expeditions to New Guinea, with ACIAR and AIDAB funding. What role they should play, since they are not directly involved with banana breeding programs which would use and should study the germplasm, needs careful consideration.

Recent Events The original meetings which later catalyzed the formation of INIBAP were held at IITA in Nigeria in July 1981 and later at IDRC in Ottawa in July 1982. At the first meeting Jamaica was heavily represented, two IITA staff were present, and PR Rowe of SIATSA (now FHIA), as well as E De Langhe, now Director of INIBAP; the only potential donor present was IDRC. Since breeders were present, one suggestion was to perform induced mutation as well as expand Dr Rowe’s small project in breeding AAB plantains. A strong suggestion was made to reactivate the Jamaican program for breeding plantain and cooking bananas. The reason given for all this was only the threat of black Sigatoka— but no pathologist attended the meeting! By this time black Sigatoka had arrived in Africa. Since black Sigatoka attacks plantains, Musa became more acceptable to receive international public funds. There was no mention of a “network”, nor of “coordination”. At the second meeting in Ottawa the pitch was that (a) “... urgent action is needed to establish an informal international network linking national plantain and banana research organizations ... so that disease-resistant planting material can be made available as soon as possible.” And (b) “An important aspect of this network would be IW Buddenhagen 23

the strengthening of the two existing banana breeding programs, in Honduras and in Jamaica.” In retrospect, which gives the wisdom of hindsight, these conclusions and suggestions seem terribly naive. Now, 10 years later, one may ask how many farmers are growing the “disease-resistant planting material [to] be made available as soon as possible”. Also, one may question if the subsequently created INIBAP network has indirectly resulted in the diversion of potential funds away from the key need of expanded, well-rounded breeding programs. Also, 10 years later, the Jamaican breeding program is defunct; but, unrelated to the INIBAP network, three new breeding programs have been established, in disparate locations (Brazil, Guadeloupe, and Nigeria). Moreover, two of these (Brazil and IITA) seem well on the way to solving, with minimal funding, the major problems of the AAB dessert and plantain groups. The United Fruit breeding program is now that of FHIA, still under PR Rowe, still underfunded, but it is coming up with potential solutions for the eastern African cooking bananas and for ABBs in general, through a new breeding thrust for these objectives. Separate from the network, the French have pushed hard to expand Musa research under CIRAD-IRFA and have started a breeding program and considerable basic research. It is hoped we will be updated on all these new activities at this workshop. INIBAP has done various useful things, in organizing meetings such as this one, in contributing to meetings of others, in communications of various sorts including the abstracting service, in sponsoring tours and consultancies. And on very limited research seed-money disbursement arrangements. Very recently INIBAP has organized global and regional testing of a few new potential cultivars bred by FHIA and IITA. By assuming this as an INIBAP activity, the active breeding programs tend to be less involved directly with potential clientele and with analytical appraisal of environment/genotype interactions. Direct involvement by an active multidisciplinary crop improvement team of scientists in both of these activities is very useful for any breeding program, to provide feedback insights. Where testing is separated from these scientists, the testing tends to become an end in itself, as has been seen in the massive international testing of rice and maize. Local research suffers. INIBAP concentrates on networking and coordination of others, and this activity is fraught with pitfalls. Research is competitive. Funding is short and struggled for on an individual basis. Researchers don’t like to be “coordinated”. There is a long legacy of country, regional, institute, and commercial rivalries. The carrot held out, of small handouts of funds, is accepted by some, but often resented. So there are many problems. But the fault lies less with INIBAP than with the very conceptualization of the organization by the original organizers and donors. Most of the original CGIAR centers perform networking and coordination functions of some sort, and even they find these activities difficult, not very cost-efficient, and often politically sensitive. But at least they are operating from a research base. One may question the wisdom of the formation of an organization to conduct “networking and coordination” when it has no research base, especially on a tropical crop direly needing both better breeding and a better research base. To headquarter the organization in a nontropical country, far from banana 24 Whence and Whither Banana Research and Development?

problems and from research programs (except for the nearby French laboratories) has also posed great difficulties. TAC has finally realized the importance of bananas and plantains and in their 1990 priority study lists them among the top 10 world commodities in value of production. This may have influenced the move whereby INIBAP was recently allowed to enter the CGIAR system, as one of several new centers under that mantle. Unique among the crop centers, it still is constituted not to have a research base. And, like IITA, it does not really address dessert bananas. Thus, the CGIAR continues to side-step the banana industry that is so vital to the economic well-being of quite a few countries. While all this has been going on many new initiatives have occurred. Some of the new associations in Latin America and the Caribbean, such as CORBANA, UPEB, and others, as well as the older WINBAN, carry out applied research and technical services of various kinds. In addition to those mentioned above in Brazil, Australia, Taiwan, IAEA, through FAO, IBPGR and at IITA, scientists scattered here and there, with a dedicated interest in banana research, have plugged away at modern research on embryogenesis, on Fusarium, on Mycosphaerella, on moko and blood diseases, on the banana weevil and banana nematodes and on biotechnology applied to taxonomy and evolution. Many of them are here at this workshop, very valuably arranged by INIBAP. Many of these scientists are distant from banana lands, and they struggle for funds to keep a part-time project going. Few are really connected to a banana breeding program.

The Future Much new knowledge needs to be discovered that is useful for knowledge’s sake. But where there are real problems and opportunities in the tropical banana world that can affect the lives and livelihoods of millions, research should be applied to direct solutions. The solutions will come through breeding. And through developing the best possible compatible marriage of breeding and biotechnology. Both need to go forward together. And industry, international donors, foundations, and private donors should join together to construct rational, productive, useful and used research programs centered on several fully developed breeding programs. More scientists at universities should be involved and more PhD theses developed. There is much to accomplish. More and better research and cooperation is needed and less jockeying for political power and control. There are many new players interested in the biotechnology of tropical crops: companies; national “overseas” organizations such as CIRAD in France, DGIS in the Netherlands, NRI in England; individuals and groups at universities; FAO/IAEA; “brokering” groups of various types; national agricultural systems in quite a few countries; and even the World Bank. The World Bank has appointed a biotechnology specialist to review and establish useful biotechnology projects connected to or appropriate for World Bank projects and interests. So there are enormous interest and possibilities. The art and wisdom will be to add biotechnology where appropriate and in such a way as not to deplete resources for both applied and basic research that is needed that is not classifiable as biotechnology. For bananas, this is especially difficult since IW Buddenhagen 25

fully-rounded multidisciplinary breeding programs are even yet not in existence nor well supported! Biotechnology’s main opportunity will be a marriage with breeding and genetics of bananas, to focus on the most difficult breeding target: a resistant Grande Naine or Williams. These need transformation and appropriate genes for transformation. All the other banana classes should be bred conventionally since they are now proving so easy to breed. Biotechnology may help as a breeding tool. There are other opportunities to contribute to banana needs by a focus on enhancing potential biocontrol mechanisms of animal parasites. Whether the needs of postharvest physiology offer opportunities is uncertain. Genes are available to alter enzymes of the ripening process. Certainly the main investment should relate to solving the major biological pest and pathogen problems. Pathogen variability and host/pathogen coevolution have recently received new emphasis, but much more needs to be done, combining fieldwork and detailed collecting with biotech-assisted analysis. No fungal disease of any crop has yet fallen to the powerful techniques of biotechnology, so the challenge is very great. I see no reason why a banana pathogen should not be the first—it is up to you, and others who wished to come to this workshop but could not, to make it happen. It is also up to good leadership to catalyze good communication between real problems, researchers, and those who make decisions on the allocation of research funds. The model used for the biotechnology thrust of the Rockefeller Foundation, which has had great success, is well worth considering. Let this workshop be a turning point in banana research by catalyzing proper funding of multidisciplinary breeding programs and enabling their flowering, with the aid of biotechnology, actually to solve the important problems. Sufficient appreciation of our mistakes and the chequered history of banana research should enable a better course to be charted for the future.

References and Bibliography

BUDDENHAGEN IW. 1977. Resistance and vulnerability of tropical crops in relation to their evolution and breeding. Annals New York Acad. Sci. 287:309-326. BUDDENHAGEN IW. 1987. Disease susceptibility and genetics in relation to breeding of bananas and plantains. Pages 95-109 in Banana and Plantain Breeding Strategies (Persley GJ, De Langhe EA, eds). ACIAR Proceedings no.21. Canberra, Australia: ACIAR. BUDDENHAGEN IW. 1990. Banana breeding and Fusarium wilt. Pages 107-113 in Fusarium Wilt of Banana (Ploetz RC, ed.). St Paul, MN, USA: APS Press. DHED’A D, DUMORTIER F, BARRIS B, VUYLSTEKE D, DE LANGHE E. 1991. Plant regeneration in cell suspension cultures of the cooking banana cv. `Bluggoe’ (Musa spp. ABB group). Fruits 46:125-135. EPP MD. 1987. Somaclonal variation in bananas: a case study with Fusarium wilt. Pages 140-150 in Banana and Plantain Breeding Strategies (Persley GJ, De Langhe EA, eds). ACIAR Proceedings no.21. Canberra, Australia: ACIAR. FAWCETT W. 1913. The Banana. London. HWANG SC, KO WH. 1987. Somaclonal variation of bananas and screening for resistance to Fusarium wilt. Pages 151-156 in Banana and Plantain Breeding Strategies (Persley GJ, De Langhe EA, eds). ACIAR Proceedings 21, Canberra. 26 Whence and Whither Banana Research and Development?

KERVÉGANT D. 1935. Le Bananier et son Exploitation. Paris. KRIKORIAN AD, CRONAUER SS. 1984. Aseptic culture techniques for plantain improvement. Economic Botany 38:322-331. KRIKORIAN AD, SCOTT ME, CRONAUER-MITRA SS, SMITH DL. 1990. Musa callus and cell culture: strategies, achievements and directions. Pages 9-23 in: In-Vitro Mutation Breeding of Bananas and Plantains I. Report of the First Research Coordination Meeting for the FAO/IAEA, May 29 - June 2, 1989. Technical Document IAEA 312.D2.RC.411. Vienna, Austria: International Atomic Energy Agency. PERSLEY GJ (ed.). 1990. Agricultural Biotechnology: Opportunities for International Development. Wallingford, Oxon, UK: CAB International. PERSLEY GJ. 1990. Beyond Mendel’s Garden: Biotechnology in the Service of World Agriculture. Wallingford, Oxon, UK: CAB International. PERSLEY GJ, De Langhe EA (eds). 1987. Banana and Plantain Breeding Strategies. ACIAR Proceedings no.21. Canberra, Australia: ACIAR. PLOETZ RC (ed.). 1990. Fusarium Wilt of Banana. St Paul, MN, USA: APS Press. QUISUMBING E. 1919. Studies of Philippine bananas. Agric. Rev. 12:1-73. REYNOLDS PK. 1927. The Banana. Boston and New York. RIOS PG. 1930. Cultivo del banana en Puerto Rico. Ins. Agr. Exp. Sta. Rio Piedras Bull. 36. 58 pp. ROWE PR. 1981. Breeding an “intractable” crop: bananas. Pages 66-83 in Genetic Engineering for Crop Improvement (Rachie KO, Lyman JM, eds). New York, USA: The Rockefeller Foundation. SIMMONDS NW. 1959. Bananas. London, UK: Longman. SIMMONDS NW. 1962. The Evolution of the Bananas. London, UK: Longman. SIMMONDS NW. 1983. Plant breeding: the state of the art. Pages 5-25 in: Genetic Engineering of Plants: An Agricultural Perspective (Kosuge T, Meredith CP, Hollaender A, eds). New York and London: Plenum Press. STOVER RH. 1972. Banana, Plantain and Abaca Diseases. Kew, Surrey, UK: Commonwealth Mycological Institute. STOVER RH, BUDDENHAGEN IW. 1986. Banana breeding: polyploidy, disease resistance and productivity. Fruits 41:175-191. STOVER RH, SIMMONDS NW. 1987. Bananas. London, UK: Longman. VALMAYOR RV. 1976. Plantains and bananas in Philippine agriculture. Fruits 31:661-663. WARD FS. 1930. Banana growing in Malaya and the presence of diseases. Malayan Agric. J. 18:63-70. Part 1

Nonconventional Strategies for Producing Banana and Plantain Resistant to Pathogens and Pests D González de León, S Fauré 29

Genetic Mapping of the Banana Diploid Genome: toward an integrated approach to the study of the Musa genome and the use of molecular marker technologies in Musa breeding

D González de León, S Fauré

“So many genes. So little time.” (Ed Coe, maize geneticist)

Introduction Like so many crops of relative importance in human diets, bananas and plantains have not been studied as much as they deserve. In particular, much of what is known about their natural history, genetics, and evolution is limited to the contributions of comparatively few researchers since the beginning of the century. Thus, while many of the cytotaxonomic riddles of the genus Musa have been elegantly unravelled (Simmonds 1962, and references thereof), there is almost no knowledge of the genetics of any of its taxa, let alone of their population genetics and the minutiae of their reproductive biology. Modern banana breeding has suffered substantially from this situation, and progress has been rather slow in breeding clones that fulﬁll current breeding objectives, such as resistance to major diseases (Buddenhagen 1993; Rowe, Rosales 1993). Even when some of the objectives have been attained, it is typical, if not unfortunate, that the agronomic characters of the new clones are not as good as the most widely grown triploid varieties. It seems therefore imperative that more knowledge be gained about the banana genome in general, in order to facilitate its manipulation and thus contribute to giving new impetus to current breeding practices and helping develop novel improvement strategies, especially at the diploid level.

Genome Manipulation and Genetic Mapping Over the last 6-10 years, with the ﬂourishing of plant molecular genetics, new tools have been developed and reﬁned that hold enormous promise in what they can reveal about

CIRAD-BIOTROP, BP 5035, 34032 Montpellier Cedex 1, France 30 Genetic Mapping of the Banana Diploid Genome

the genome of almost any organism. These tools provide new kinds of genetic markers, that is characters that show inheritable variation at the DNA level, rather than at the morphological or biochemical levels. These characters are called markers because we use them to obtain, albeit indirectly, information about the genetics of other traits of interest in the organism under study. Restriction fragment length polymorphisms (RFLPs), random amplified polymorphic DNAs (RAPDs or AP-PCR markers) and other derived or analogous methodologies that generate molecular markers have been expertly reviewed by numerous authors (see for example Beckmann and Soller 1986; Burr et al. 1983; Rafalski et al. 1991; Stuber 1992; Tanksley et al. 1989; Williams et al. 1990). With these genome manipulation techniques it is relatively easy to identify specific chromosome regions that are important for a particular phenotype. Thus, genetic mapping of the genome of hitherto poorly studied species has become a feasible, almost trivial endeavor, as long as some conditions are fulfilled such as the availability of suitable segregating populations. Once particular genomic regions have been identified, they can be transferred from germplasm to germplasm by conventional crossing. In these procedures, it is the markers themselves that may serve as selection criteria. This sort of “marker-assisted selection” has been around since the early days of modern genetics (see for example Stuber 1992 for a brief review). The difference with today’s approaches is simply in the kind of markers then available. Only a handful of crops have benefited from such early markers. Take, for example, the case of one of the best genetically characterized plants: after more than 50 years of intensive genetic investigations, the current genetic database for maize (Zea mays) consists of over 600 individual loci identified by their visible phenotype (“naked eye polymorphisms” or NEPs) or by the analysis of their biochemical bases (Coe et al. 1988). Of these, some 400 have been placed onto specific segments of the corn genetic map; using conventional approaches, an enormous effort would be required to gather all of these points into a unique, comprehensive map, simply because it is quite impossible to accumulate more than a few of these mutations into individual breeding stocks. By contrast, in less than 10 years, research in only three different public laboratories has resolved more than 1000 new loci defined by DNA polymorphisms. These have been added to the various existing versions of the corn map (Burr et al. 1988; Helentjaris 1987; Hoisington, Coe 1990). Simple mapping strategies based on DNA markers such as “interval mapping” (Lander, Botstein 1987), are now being used to construct the comprehensive corn map (Hoisington, Coe 1990). Anyway, extensive genetic maps constructed on the basis of morphological, isozyme, and some other phenotypic traits (e.g., physiological mutants, disease resistances) exist for very few crop plant species (e.g., tomato, pea, maize). In the bananas, what little is known of the formal genetics of NEPs is limited to one or perhaps two, not too recent papers (Simmonds 1953; Vakili 1968). However, the maize example given above serves at least one purpose: it gives us the reassurance that rapid progress in dissecting the genetics of the bananas is within our grasp. In effect, molecular tools provide a potentially indefinite number of markers that can serve as selection criteria in the manipulation of the banana genome. D González de León, S Fauré 31

The use of molecular markers to locate major genes and loci contributing to the expression of a continuously varying character (i.e., quantitative trait loci or QTLs) affecting agriculturally important traits is rapidly becoming the primary approach for understanding the genetic basis of these traits and, more importantly in the case of crops such as the bananas, to help develop novel breeding strategies. High-density molecular marker linkage maps have now been constructed for more than 15 different plant species including Arabidopsis thaliana (Chang et al. 1988; Nam et al. 1989), barley (Graner et al. 1991; Heun et al. 1991), Brassica spp. (Landry et al. 1991; Slocum et al. 1990; Song et al. 1991), chili peppers (Tanksley et al. 1988), lentils (Havey, Muehlbauer 1989), lettuce (Landry et al. 1987), maize (Burr et al. 1988; Hoisington, Coe 1990), potato (Gebhardt et al. 1989), rice (McCouch et al. 1988), rye (Wang et al. 1991), soybeans (Keim et al. 1990), tomato (Bernatsky, Tanksley 1986), and wheat (Gale et al. 1990). Some examples of the localization of agriculturally important traits relative to genetic markers are briefly described elsewhere in these proceedings (Ploetz 1993). The ultimate use of molecular markers is the detection and potential manipulation of loci affecting quantitative characters; at least two convincing demonstrations of the reduction of QTLs to Mendelian factors have very recently been published for tomato and maize. These should indicate what strategies to adopt for the genetic dissection of QTLs in other crops that are far more difficult to manipulate (Paterson et al. 1991; Paterson et al. 1988; Stuber et al. 1992). In the next section we briefly review some aspects of Musa biology that should not be overlooked if the use of diploid germplasm is to be increased in current and future breeding strategies.

The Banana Genome: Genetic Constraints and their Consequences The typical seedless edible banana appears to be the product of two essential evolutionary processes accompanying domestication (Simmonds 1962): parthenocarpy, that is the autonomous stimulus to pulp growth; and sterility, both genic female (and male?) sterility and that arising from structural hybridity. A third, additional and subsequent process was the formation of triploid, fully sterile cultivated forms arising from the fusion of reduced and nonreduced gametes following nuclear restitution. Both at the diploid and triploid levels, domestication has therefore resulted in an obligatory form of vegetative propagation of sexually sterile clones that has ensured the long-term preservation of relatively unchanged phenotypes. It is not difficult, therefore, to understand that breeding new triploid cultivars of bananas or plantains has not been an easy task and that success has been rare (Rowe 1984; Stover, Simmonds 1987). The classical schemes involve, for example, crossing a diploid, usually a fertile wild one showing some desirable trait, onto a good triploid recipient cultivar having good female restitution. Not much had been drawn from this 32 Genetic Mapping of the Banana Diploid Genome

method (Rowe 1984), but some recent successes are reported in these proceedings (Vuylsteke 1993). Schemes of this sort are nonetheless hindered by intrinsic genetic constraints. Insufficient knowledge about the genomic origins and composition of the best triploid cultivars complicates even further the choice of the best diploid materials for desirable introgressions. Other strategies, based on breeding better diploids using both cultivated clones and wild accessions, are becoming increasingly important, not to say inescapable (Bakry et al. 1990; Rowe 1987). The use of M. acuminata diploids is far from trivial. Take the cultivated types, for example: most of the accessions investigated have been found to be heterozygous for one or several translocations and/or inversions (reviewed by Simmonds 1962; pers. com. Bakry et al., CIRAD), and thus tend to be at least fairly male-sterile; genic female sterility may further complicate the use of these materials also as female parents. In the case of the wild acuminata taxa, the situation seems superficially better since these are fertile, seed-bearing types having normal meiosis, that should be easy subjects in hybridization schemes. But data have been accumulating since the first half of this century that there exists considerable chromosome structural differentiation within the species (reviewed by Simmonds 1962). Thus, most currently recognized subspecies such as banksii and burmannicoides, to name but two, produce hybrids that are partially sterile as a result of chromosome structural differences between the parents (e.g., Dodds, Simmonds 1948; pers. com. Shepherd 1987: ms. on translocations in Musa acuminata). Bananas are one among many examples of plants and animals whose cytogenetic structure is made more complex through extensive chromosomal repatterning (Burnham 1956; Jones 1970; Stebbins 1971; White 1978). These include taxa from genera such as Triticum spp. (Fominaya, Jouve 1985), Lens (Tadmor et al. 1987), Gossypium (Menzel et al. 1978), Secale (Rees, Sun 1965), Capsicum (González de León 1986), Helianthus (Chandler et al. 1986), Caenorhabditis (McKim et al. 1988), and Drosophila (Sharp, Hilliker 1989). This phenomenon establishes natural reproductive barriers between acuminata wild diploids, and is an obvious component of subspecies divergence and increased genetic diversity of the species as a whole. It is important for breeders, because it can seriously hinder their efforts at recombining certain characteristics of wild materials before creating suitable cultivars. In effect, there are at least three important genetic consequences of structural hybridity. First, there is interchromosomal linkage resulting from the fact that the nonhomologous chromosomes involved in a reciprocal translocation, for example, will associate during the first meiotic prophase to form multivalents (as opposed to normal bivalents) which behave as a single recombination unit or linkage group. Thus, genes normally situated on different chromosomes in one of the parents, that is genes that usually segregate independently, will show varying degrees of genetic linkage in the progeny of structural hybrids. Secondly, there is reduced recombination of genes located near the break-points at the origin of the translocations: see Figure 1 (Burnham 1962; Carlson 1988; Rees 1961; Rees, Jones 1977; Rickards 1983); this is reflected by D González de León, S Fauré 33

Figure. 1. Diagrammatic representation of meiotic pairing in a hypothetical heterozygote for two reciprocal translocations involving three pairs of chromosomes. reductions in the genetic distances between those loci. And thirdly, given that the disjunction of multiple chromosome associations at meiotic metaphase usually results in an unbalanced distribution of interchanged chromosomal segments, some gametes will be inviable and their gene combinations will not be represented in the progeny, i.e., there is differential gametic viability. This is simply reﬂected in the progenies as a more or less severe distortion of the expected segregation of genes carried on rearranged chromosomes (e.g., Rees 1961; Sharp, Hilliker 1989). 34 Genetic Mapping of the Banana Diploid Genome

Thus, both the probability of finding unwanted linkages and the strength of these linkages may increase; concomitantly, the frequency of certain genotypic classes may decrease; the overall effects could well interfere with the breeder’s efforts at recombining and transferring desirable traits from wild and cultivated diploids, sometimes forcing them, even unwittingly, to attempt bridging crosses to circumvent these reproductive barriers or to avoid apparent linkage of undesirable traits. Conversely, translocations can also be used to the breeder’s advantage for the transfer or the localization of particular gene loci (Beckett 1991; Jones, Rawls 1988; Weber 1989). Either way, knowledge of the extent and distribution of structural rearrangements would be an invaluable asset to expedite breeding strategies involving diploid bananas. This could be achieved through the combined investigations of meiotic behavior of hybrids of interest to improvement programs (e.g., Bakry et al., CIRAD) and the search for specific genetic markers (RFLP or other) closely associated to the rearranged segments. In other plant families, different RFLP marker orders have been shown to exist within and between species or higher taxonomic orders. For example, the pepper genetic map is highly rearranged with respect to those of the tomato and the potato, while in the latter two, marker order is conserved (Bonierbale et al. 1990; Bonierbale et al. 1988; Tanksley et al. 1988). Within pepper (Capsicum) alone, there are large numbers of structural differences between and within the ﬁve species that have given rise to today’s cultivated types (González de León 1986; Pickersgill 1977). Mapping structural differences will certainly help manage and perhaps take advantage of their genetic effects for breeding purposes. Such information will guide breeders as to the choice of better progenitors while providing them with tools that can help them detect desirable structural homozygotes that can more easily be used for further recombination and introgression.

First Elements of a Genetic Map of Musa acuminata: a glimpse at work in progress at CIRAD The ﬁrst elements of an RFLP linkage map of the Musa acuminata genome have been derived by screening two genomic libraries isolated from young banana leaves. Each library was developed through digestion of nuclear-enriched genomic DNA with one of two restriction enzymes (EcoRI and PstI). Fragments from 500 to 2000 bp were selected, cloned into a suitable plasmid, and transformed into a bacterial host. The libraries were screened for repeated sequence inserts (Amson 1989; Fauré et al. 1991; Fauré et al. 1992). Some of the results of the screening of a number of banana accessions with inserts from these libraries are given in Table 1. Separate samples of DNA of each accession were digested with one of two restriction enzymes (DraI and EcoRV). Apart from the ease of isolation of some 250 usable probes that reveal unique sequences in diploid bananas, the essential fact illustrated by these results is the high degree of polymorphism revealed by the probes; thus, most markers will segregate in the progenies D González de León, S Fauré 35

Table 1. Summary of the characteristics of the probe libraries (unpublished data available up to December 1991). The results correspond to the overall polymorphism revealed after restriction digestion of the DNA of each accession with two different enzymes (see also Fig.3).

Figure 2. Four of the mapping populations under current development and/or evaluation in CIRAD’s banana improvement program. of most crosses. Furthermore, by using two more enzymes, between 60% and 98% polymorphism is revealed between any pair of accessions tried so far. This is comparable to some of the most polymorphic plant species reported in the literature. At present, the mapping work is being conducted strictly within the framework of CIRAD’s banana improvement program. Mapping populations are being developed in Guadeloupe (CIRAD-IRFA) in collaboration with F Bakry and JP Horry. The nature of four of these populations is outlined in Figure 2. Field evaluations of important 36 Genetic Mapping of the Banana Diploid Genome

characters such as parthenocarpy and disease resistance will be made on all of these populations. When possible, all other characters that show variation in the progeny will be measured and their inheritance analyzed.

The ﬁrst population available for study was the F2 of SF265 x banksii. This was easy to create since SF265 (AAcv) appears to be a derivative of ssp. banksii; for the same reason, however, it is less polymorphic than the others. In addition, the female parent was found to be heterozygous at many loci thus reducing even further the number of usable probes in the F2 progeny. Two examples of segregation proﬁles revealed by such probes are shown in Figure 3.

Up to December 1991 the linkage map developed using 82 F2 individuals of this population comprises 27 RFLP loci and 5 isozyme loci. About a third of these markers showed a signiﬁcantly distorted segregation (data not shown). Linkage analysis of the segregation data was performed using multipoint techniques as implemented in the computer package Mapmaker V1.0 (Macintosh) (pers. com. Tingey et al., Dupont Co. 1990; Lander et al. 1987). The resulting provisional map is shown in Figure 4; 6 linkage groups with 2 to 5 loci each have been found so far. Twelve other markers segregated independently. Our preliminary cytological investigations indicate that the wild banksii parent has normal male meiosis, that SF265 is heterozygous for a reciprocal interchange, and that the F1 hybrid plant that gave rise to the F2 is heterozygous for two such rearrangements involving three pairs of nonhomologous chromosomes (a situation similar to that depicted in Figure 1). As was outlined in an earlier section, there is a strong probability

Figure 3. Typical RFLP segregation patterns observed among the F2 progeny of SF265 x banksii. A. Hybridization signal of probe pMaCIR1006 of the EcoRI library onto DNA digested with DraI; B. Hybridization signal of probe pMaCIR27 of the PstI library onto DNA digested with EcoRV. D González de León, S Fauré 37

Figure 4. First elements of a linkage map of the banana genome (Musa acuminata) constructed from unpublished data available up to December 1991.

that the distorted segregation ratios were due to differential gamete viability in the F1 parent. This connection is being investigated at present. Additional genetic consequences have already been discussed, notably some form of interchromosomal linkage between the rearranged chromosomes. In this cross, we could then expect to detect fewer than the 11 linkage groups corresponding to the basic number of chromosomes in these diploids. Besides, not only should one of these groups be relatively larger, but it should also contain all or some of the markers that showed distorted segregation. The possibility of stronger linkage among these, owing to the likely reduction in recombination near the break-points of the interchange meiotic conﬁguration, will be explored. Our plan is to put between 60 and 100 markers on this map in 1992. These will include RFLP, RAPD, isozyme and perhaps a few morphological markers. With the publication of the map toward the end of 1992, we shall endeavor to make freely available all of the probes that will have been used in this or other maps based on other populations (Fig.2).

The Suitability of Musa Plant Materials for Molecular Marker Analyses The feasibility of constructing a genetic map of the M. acuminata genome was, we think, well established in the previous section. In order to support this claim even further, and perhaps motivate potential or actual banana breeders and geneticists to take a deeper 38 Genetic Mapping of the Banana Diploid Genome

interest in this kind of research, we rapidly review below what, in our experience, makes diploid bananas suitable for molecular marker analyses. Indeed, the obvious shortcomings of bananas as model genetic organisms, resulting from their relatively long life cycle and enormous plant habit as compared with typical annual crops, are outweighed by numerous advantages for genetic and molecular marker work. Pollinations between numerous diploid accessions are fairly easy to perform at least in one direction, even when the flowers are perfect, as in the case of M. acuminata ssp. banksii. Besides, the relatively slow sequential development of the female inflorescence, the long-term availability of large amounts of pollen from single individuals, and the clonal reproduction of each plant, all allow for self- and cross-pollinations to be performed given enough planning. There are difficulties, of course, when different degrees of female and/or male sterility complicate the task, or when there is a need for rescuing embryos through tissue culture to ensure their survival; but none of these obstacles is unique to or specially acute in the bananas. In favorable cases, very large amounts of seed can be produced from a single cross. Once the progeny of a given cross is established in the field, it essentially represents an eternal population, indefinitely available for further study and evaluation, given enough maintenance. Not only that, it can also be cloned and thus replicated in one or more different environments. The equivalent operation can be more difficult and costly in the case of more widely studied annuals. Leaf material is available in great, almost unlimited amounts from any young to adult individual plant, at any time of the year, and is very easily harvested and stored. Extraction of fairly pure DNA in large quantities and for large numbers of samples is easily and cheaply achieved. In our laboratory for instance, in 2 weeks, a single person can extract, quantify, and check for quality enough DNA for a fair-sized project involving 200 individuals to be analyzed for 40 probe-enzyme combinations (protocols, from harvest to final results, are available from the authors). The high polymorphism that has been revealed at the DNA level means that an essentially unlimited number of markers can be developed even for crosses between fairly related accessions (e.g., SF265 x banksii, etc., Table 1). Finally, owing to the increased representation of wild diploids in the larger collections (CIRAD, Guadeloupe; Vakili 1968), it should be possible to establish a wide variety of segregating populations arising from crosses between them. Prior screening for polymorphic loci will contribute to the genetic description of these materials well before much else is known about them on other grounds.

Towards an Integrated Approach to the Study of the Musa Genome It has been suggested above that at least some of the genetic characteristics and constraints of the diploid bananas can be dissected using molecular markers. Later, it was argued that one of the means toward such a target is the construction of a genetic D González de León, S Fauré 39

Figure 5. A scheme for an integrated approach to the study of the Musa genome using molecular marker technologies: the approximate chronology of CIRAD’s current program is outlined on the left. map, and that this, in itself, is now a tangible reality. This reality was made possible by the suitability of the bananas themselves as materials for genetic and molecular marker manipulations. It is therefore only natural to propose that molecular markers should be used to complete our picture of the banana genome in terms of its wealth of variation throughout Musa and even beyond it. The program developed at CIRAD is an attempt at integrating mapping and formal genetic studies with the investigation, description, and analysis of molecular diversity in the two putative progenitor species of the edible bananas and plantains. A synoptic diagram of this program, giving an approximate chronology for the start of various subprograms and ﬁnal objectives, is given in Figure 5. The central concept in this scheme is that of genome “reconstruction”: it may well be possible to reconstruct improved polyploid assemblages (AAA, AAB, ABB, AABB, etc.) from chosen recombined diploids; the building blocks would be provided by data on gene 40 Genetic Mapping of the Banana Diploid Genome

orders found in the genus and their correlation with the presence or absence, throughout the genome, of particular alleles in specific combinations. Inherent to this hypothesis is the idea that the genome architecture that will be revealed in today’s bananas still reflects the ancestral contributions to existing domesticates. Current diversity data, whether at the morphological or genetic levels, give sufficient support to our working hypothesis. The classic classificatory scheme, whose main elements were the morphological recognition of the bispecific origins of bananas and plantains coupled with knowledge of chromosome numbers (Simmonds, Shepherd 1955), has been, in essence, supported by recent studies using various kinds of genetic or biochemical markers. Thus, isozyme markers (Horry 1989; Jarret, Litz 1986a,b), flavonoids (Horry, Jay 1988), and chloroplast DNA markers (Gawel, Jarret 1991a,b), all support the concept of bispecific origins. Indeed, diagnostic isozymes and chloroplast DNA bands have been described. But perhaps the more pertinent data in the present context is that the isozyme, flavonoid and, more recently, the data obtained from the analysis of length variation in the intergenic gene spacer of banana ribosomal genes (rDNA) (Lanaud et al. 1992), all confirm earlier beliefs that the diversity of the cultivated types represents a fraction of that encountered at the wild, diploid level. The use of mapped loci as markers for diversity studies will help strengthen the current biosystematic treatment of the genus, by providing data on numerous chromosome segments well distributed over the entire genome. Moreover, it will be possible to verify the existence of linkage disequilibria fixed in present sterile, clonal edible types. In order to build this integrated approach into the study of the banana genome on solid ground, it is essential that the mapping effort itself rests on a structured, coordinated plan that it is hoped will involve many participants. The following paragraphs suggest some ideas as to what ingredients such a plan should include.

An integrative core map of Musa and beyond: some suggestions and propositions 1. Development of reference mapping populations A concerted effort, possibly coordinated through INIBAP, for the development and establishment of multiple reference mapping populations (RMPs) should be made within the framework of collaborative efforts between existing field and laboratory facilities; some guidelines for prospective developers of such populations are given in the appendix to this paper. Moreover, some or all of these populations should be multiplied for multisite evaluations after quarantine clearing. Tests of the behavior of these populations under various biotic and abiotic stresses would then be simplified and each breeding program would be able to obtain data pertaining to their specific improvement objectives during all or some specific growing cycles. 2. Construction and use of a “core” map It is proposed that initial efforts should concentrate on the construction of a basic reference or “core” map to be used for coordinating all mapping efforts and applications. D González de León, S Fauré 41

Recently developed for maize (Gardiner et al.: pers. com. EHJ Coe), the idea of a core map consists in identifying a standardized set of RFLP markers that would be more or less evenly spaced in the genetic map(s) developed from the RMPs. The probes for revealing such loci should give strong, simple-to-read signals, and their relative distances should be such that their relative order be at least 1000 times more probable than any other order for the same probes. Other probes revealing loci within shorter distances of the core markers will simply be put into the intervals or “bins” defined by the latter. In other words, depending on the size of the starting mapping population, the core map will have a defined density (see the appendix). As more information is obtained from other crosses and experiments, some of the bins defined by core markers can be further subdivided into smaller segments. Also, as more information about structural rearrangements becomes available, different bins will “contain” break-points marking specific translocations or inversions. The core map will slowly evolve into an integrative one (see below). The main advantage of a core map is that it simplifies data-sharing between research groups and should expedite the integration of segregation data obtained from RMPs and other, more specific, populations. Eventually, any genetic trait would be localized to a small region of the genome by using a limited number of agreed-upon markers. Furthermore, the bins themselves should serve as the indexing basis of a mapping database system (Gardiner et al.).

3. From a core map to an “integrative” map and beyond Eventually the map should be “integrative” in at least two ways. First, it should aim at representing the plethora of structural rearrangements that have accompanied genome evolution and domestication in Musa (see above). This will not only clarify phylogenetic relationships between taxa (especially in conjunction with molecular systematic data), it will also help guide crossing strategies so as to avoid unwanted linkages and favor specific genic recombinations. Secondly, the integrative map should combine all kinds of markers at the morphological, biochemical, and DNA levels. Each marker type has its own intrinsic advantages and limitations, and they all should be used to answer specific questions in specific ways. Molecular genetic markers should include those based on anonymous RFLP probes such as those described in the previous section, new marker types such as RAPDs, and cloned genes, of which there are more than 100 now available that have been isolated from other species. These can reveal RFLPs corresponding to putatively functional genes when used as heterologous probes. In the case of PCR-based methods such as RAPDs that have a few intrinsic disadvantages, such as the detection of different loci using the same oligonucleotide (acting as a “probe” in this case) on different populations (Williams et al. 1991), researchers would be encouraged to isolate the polymorphic DNA bands and use them as RFLP probes on some or all of the RMPs (Williams et al. 1990). 42 Genetic Mapping of the Banana Diploid Genome

Conversely, sequencing of the ends of existing probes should help develop PCR-based detection systems for the fine genetic analysis of specific loci or for increasing the probability of finding a polymorphism when this has not been possible through other means. In effect, the sequenced ends can be used as primers for amplification of the sequence corresponding to the original probe, and the product can be digested by one or more restriction enzymes to reveal differences between otherwise monomorphic accessions at that locus (Williams et al. 1991). Such methods could conceivably evolve into +/- detection systems that rely on visible products that can be measured directly in the test-tube used for DNA amplification, (e.g., Higuchi et al. 1992). Progressively, a map should emerge that combines all types of available markers and therefore becomes neutral with respect to technical approaches. While the minute chromosomes found in Musa have resisted attempts at establishing adequate karyological descriptions, their low number (2n=2x=22) and the rather small size of the genome as a whole (Arumuganathan, Earle 1991, our own unpublished data, Novak 1993) could help motivate the construction of a library of very large banana nuclear DNA fragments cloned into yeast artificial chromosomes (YACs: see for example Ganal et al. 1989; Garza et al. 1989; Guzmán, Ecker 1988). The sequence-tagged sites generated upon partial sequencing of RFLP probes could serve as a starting point for rapidly locating these “long-distance” probes (i.e., several hundred kilobases) on the core map. The eventual construction of physical maps of the banana genome, i.e., the ordering of YAC clones that cover most, if not all, the genome would then be achieved and would open up new options such as map-based cloning of banana genes (e.g., Tanksley et al. 1989). 4. Coordination of mapping activities: a mapping information database and clone bank The scant present research on banana genetics certainly needs a substantial boost; meanwhile, it is all the more important to optimize what little is being done around the world. As more laboratories start applying molecular technologies, it seems desirable to start coordinating their activities as soon as possible. One way of achieving this would be to set up a mapping information database and a clone (probe) bank that would promote sharing of tools and information, and help define genetic nomenclatural rules right from the start. Initially, INIBAP could help set up a probe transfer and exchange system to facilitate and promote the use of the best probes that are already available.

Conclusions Despite their major importance as a staple starch source and their signiﬁcant place in the export market under their dessert forms, bananas and plantains still remain on the margin of modern genetics research and its applications to breeding practices. On account of their highly polymorphic nature and their ease of handling for molecular marker studies, the bananas are an easy target of such studies. D González de León, S Fauré 43

Moreover, with the increasingly important role that wild diploid clones play in the improvement of the crop, and the complex biological problems that these diverse materials pose to breeders and geneticists alike, it seems logical, if not imperative, to increase our investigations of their genetics and reproductive biology. We have argued that genetic mapping provides powerful tools to achieve these goals, and have demonstrated that it is feasible. Further progress in the genetic dissection of Musa spp. should beneﬁt from an approach that would integrate cytogenetic, mapping, and genetic diversity data into a coherent description of the Musa genome. Thus, far from being an end in itself, mapping is a means toward a better understanding of genome evolution, organization, and function; it provides a set of powerful handlers—or genetic markers—that allow extensive genetic manipulations. Genetic markers should quickly enhance and strengthen our knowledge of genome diversity and organization of the diploid bananas and their ancestral contributions to triploid cultivars; in gaining such knowledge, breeding practices will have more rational bases and, eventually, we shall be in a better position to assess how and for what speciﬁc purposes marker technologies and other molecular tools can have a direct and cost-effective impact on banana and plantain improvement. Acknowledgments JL Noyer’s expert technical support is greatly appreciated. Our special thanks to F Bakry and JP Horry and their team in Guadeloupe, to J Ganry and H Tezenas du Montcel, without whose dedication and support little would have been possible. This research is being supported in part by a grant from the Commission of European Communities’ STD2 program (TS 2A-0094-F(SD)), and a doctoral studentship (S Fauré) from the French Ministry of Research and Technology.

References

AMSON C. 1989. Recherche de marqueurs génétiques chez le bananier (Musa spp.) par la méthode du polymorphisme de longueur des fragments de restriction. Rennes, France: DEA, ENSA. ARUMUGANATHAN K, ED EARLE. 1991. Nuclear DNA content of some important plant species. Plant Mol. Biol. Reporter 9:208-218. BAKRY F, HORRY JP, TEISSON C, TEZENAS DU MONTCEL H, GANRY J. 1990. L’amélioration génétique des bananiers à l’IRFA/CIRAD. Fruits numéro spécial:25-40. BECKETT JB. 1991. Cytogenetic, genetic and plant breeding applications of B-A translocations in maize. Pages 493-529 in Chromosome Engineering in Plants: Genetics, Breeding, Evolution (Tsuchiya PKGaT, ed.). Amsterdam, The Netherlands: Elsevier. BECKMANN JS, SOLLER M. 1986. Restriction fragment length polymorphisms and genetic improvement of agricultural species. Euphytica 35:111-124. BERNATSKY R, TANKSLEY SD. 1986. Toward a saturated linkage map in tomato based on isozymes and random cDNA sequences. Genetics 112:887-898. BONIERBALE MW, GANAL MW, TANKSLEY SD. 1990. Applications of restriction fragment length polymorphisms and genetic mapping in potato breeding and molecular genetics. Pages 13-24 in The Molecular and Cellular Biology of the Potato (Vayda ME, Park WD, eds). Wallingford, Oxon, UK: CAB International. BONIERBALE MW, PLAISTED RL, TANKSLEY SD. 1988. RFLP maps based on a common set of clones reveal modes of chromosomal evolution in potato and tomato. Genetics 120:1095-1103. BUDDENHAGEN I. 1993. Whence and whither banana research and development? (This volume.) 44 Genetic Mapping of the Banana Diploid Genome

BURNHAM CR. 1956. Chromosomal interchanges in plants. Bot. Rev. 22:419-552. BURNHAM CR. 1962. Discussions in Cytogenetics. Minneapolis, USA: Burgess. BURR B, BURR FA, THOMPSON KH, ALBERTSON MC, STUBER CW. 1988. Gene mapping with recombinant inbreds in maize. Genetics 118:519-526. BURR B, EVOLA SV, BURR FA, BECKMANN JS. 1983. The application of restriction fragment length polymorphism to plant breeding. Pages 49-59 in Genetic Engineering Principles and Methods, vol.5 (Setlow SK, Hollaender A, eds). New York and London: Plenum Press. CARLSON WR. 1988. The cytogenetics of corn. Agronomy Monograph 18:250-343. CHANDLER JM, JAN CC, BEARD BH. 1986. Chromosomal differentiation among the annual Helianthus species. Syst. Bot. 11: 354-371. CHANG C, BOWMAN JL, DEJOHN AW, LANDER ES, MEYEROWITZ EM. 1988. Restriction fragment length polymorphism linkage map for Arabidopsis thaliana. Proceedings of the National Academy of Sciences (USA) 85:6856- 6860. COE EHJ, NEUFFER MG, HOISINGTON DA. 1988. The genetics of corn. Pages 81-257 in Corn and Corn Improve- ment (Sprague GF, Dudley JW, eds). Madison, Wisconsin, USA: American Society of Agronomy. DODDS KS, SIMMONDS NW. 1948. Sterility and parthenocarpy in diploid hybrids of Musa. Heredity 2:101-117. FAURÉ S, CARREEL F, AMSON C, AUBERT G, HORRY JP, GONZÁLEZ DE LEÓN D, LANAUD C. 1991. RFLP analysis in the banana (Musa spp.). Poster presented at the EUCARPIA Symposium on Genetic Manipulation in Plant Breeding: Molecular Biology/Breeding Interface, Tarragona, Spain. FAURÉ S, GONZÁLEZ DE LEÓN D, HORRY JP, LANAUD C. 1992. Construction of a genetic map of the genome of the bananas (Musa spp.). Poster presented at the International Colloquium in Honour of Jean Pernes: Gene pools, gene flows, and plant genetic resources, 8-10 January 1992, Paris, France, FOMINAYA A, JOUVE N. 1985. Metaphase I centromere coorientation in interchange heterozygotes of Triticum aestivum L. J. Hered. 76:191-193. GALE MD, CHAO S, SHARP PJ. 1990. RFLP mapping in wheat: progress and problems. Pages 353-363 in Gene Manipulation in Plant Improvement II (Gustafson JP, ed.). New York, USA: Plenum Press. GANAL MW, YOUNG ND, TANKSLEY SD. 1989. Pulsed field electrophoresis and physical mapping of large DNA fragments in the Tm-2a region of chromosome 9 in tomato. Mol. Gen. Genet. 215:395-400. GARZA D, AJIOKA JW, CARULLI JP, JONES RW, JOHNSON DH, HARTL DL. 1989. Physical mapping of complex genomes. Nature 340:577-578. GAWEL N, JARRET RL. 1991a. Chloroplast DNA restriction fragment length polymorphisms (RFLPs) in Musa species. Theor. Appl. Genet. 81:783-786. GAWEL N, JARRET RL. 1991b. Cytoplasmic genetic diversity in bananas and plantains. Euphytica 59:19-23. GEBHARDT C, RITTER E, DEBENER I, SCHACHTSCHABEL U, WALKEMEIER B, UHRIG H, SALAMI F. 1989. RFLP analysis and linkage mapping in Solanum tuberosum. Theor. Appl. Genet. 78:65-75. GONZÁLEZ DE LEÓN D. 1986. Interspecific hybridisation and the cytogenetic architecture of two species of chili pepper (Capsicum—Solanaceae). Ph.D. thesis, University of Reading, UK. GRANER A, JAHOOR A, SCHONDELMAIER J, SIEDLER H, PILLEN K, FISCHBECK G, WENZEL G, HERRMANN RG. 1991. Construction of an RFLP map of barley. Theor. Appl. Genet. 83:250-256. GUZMÁN P, ECKER JR. 1988. Development of large DNA methods for plants: molecular cloning of large segments of Arabidopsis and carrot DNA into yeast. Nucleic Acids Res. 16:11091-11105. HAVEY MJ, MUEHLBAUER FJ. 1989. Linkages between restriction fragment length, isozyme, and morphological markers in lentil. Theor. Appl. Genet. 77:395-401. HELENTJARIS T. 1987. A genetic linkage map for maize based on RFLPs. Trends Genet. 3:217-221. HEUN M, KENNEDY AE, ANDERSON JA. 1991. Construction of a restriction fragment length polymorphism map for barley (Hordeum vulgare). Genome 34:437-447. HIGUCHI R, DOLLINGER G, WALSH PS, GRIFFITH R. 1992. Simultaneous amplification and detection of specific DNA sequences. Bio/Technology 10:413-417. HOISINGTON DA, COE EH. 1990. Mapping in maize using RFLPs. Pages 331-352 in Gene Manipulation in Plant Improvement II (Gustafson JP, ed.). New York, USA: Plenum Press. HORRY JP. 1989. Chimiotaxonomie et organisation génétique dans le genre Musa. Parts 1,2,3. Fruits 44:455- 474, 509-520, 573-578. HORRY JP, JAY M. 1988. Distribution of anthocyanins in wild and cultivated banana varieties. Phytochemistry 27:2667-2672. D González de León, S Fauré 45

JARRET RL, LITZ RE. 1986a. Isozymes as genetic markers in bananas and plantains. Euphytica 35:539-549. JARRET RL, LITZ RE. 1986b. Enzyme polymorphism in Musa acuminata Colla. J. Hered. 77:183-186. JONES K. 1970. Chromosome changes in plant evolution. Taxon 19:172-179. JONES WK, RAWLS JMJ. 1988. Genetic and molecular mapping of chromosome region 85A in Drosophila melanogaster. Genetics 120:733-742. KEIM P, DIERS BW, OLSON TC, SHOEMAKER RC. 1990. RFLP mapping in soybean: association between marker loci and variation in quantitative traits. Genetics 126:735-742. LANAUD C, TEZENAS DU MONTCEL H, JOLIVOT MP, GLASZMANN JC, GONZÁLEZ DE LEÓN D. 1992. Variation of ribosomal gene spacer length among wild and cultivated banana. Heredity 68:147-156. LANDER ES, BOTSTEIN D. 1987. Homozygosity mapping: a way to map human recessive traits with the DNA of inbred children. Science 236:1567-1570. LANDER ES, GREEN P, ABRAHAMSON J, BARLOW A, DALY MJ, LINCOLN SE, NEWBURG L. 1987. Mapmaker: an interactive computer package for constructing primary genetic linkage maps of experimental and natural populations. Genomics 1:174-181. LANDRY BS, HUBERT N, ETOH T, HARADA JJ, LINCOLN SE. 1991. A genetic map for Brassica napus based on restriction fragment length polymorphisms detected with expressed DNA sequences. Genome 34:543-552. LANDRY BS, KESSELI RV, FARRARA B, MICHELMORE RW. 1987. A genetic map of lettuce (Lactuca sativa L.) with restriction fragment length polymorphism, isozyme, disease resistance and morphological markers. Genetics 116:331-337. MCCOUCH SR, KOCHERT G, YU ZH, WANG ZY, KHUSH GS, COFFMAN WR, TANKSLEY SD. 1988. Molecular mapping of rice chromosomes. Theor. Appl. Genet. 76:815-829. MCKIM KS, HOWELL AM, ROSE AM. 1988. The effects of translocations on recombination frequency in Caenorhabditis elegans. Genetics 120:987-1001. MENZEL MY, BROWN MS, NAQI S. 1978. Incipient genome differentiation in Gossypium. I. Chromosomes 14, 15, 16, 19 and 20 assessed in G. hirsutum, G. raimondii and G. lobatum by means of seven A-D translocations. Genetics 90:133-149. NAM HG, GIRAUDAT J, BOER BD., MOONAN F, LOOS WDB, HAUGE BM, GOODMAN HM. 1989. Restriction fragment length polymorphism linkage map of Arabidopsis thaliana. Plant Cell 1:699-705. NOVAK FJ, BRUNNER H, AFZA R, MORPURGO R, UPADHYAY RK, VAN DUREN M, SACCHI M, SITTI HAWA J, KHATRI A. 1993. Improvement of Musa through biotechnology and mutation breeding. (This volume.) PATERSON AH, DAMON S, HEWITT JD, ZAMIR D, RABINOWITCH HD, LINCOLN SE, LANDER ES, TANKSLEY SD. 1991. Mendelian factors underlying quantitative traits in tomato: comparison across species, generations, and environments. Genetics 127:181-197. PATERSON AH, LANDER ES, HEWITT JD, PETERSON S, LINCOLN SE, TANKSLEY SD. 1988. Resolution of quantitative traits into Mendelian factors by using a complete linkage map of restriction fragment length polymorphisms. Nature 335:721-726. PICKERSGILL B. 1977. Chromosomes and evolution in Capsicum. Pages 27-37 in ‘CAPSICUM 77’, Comptes Rendus du 3ème Congrès EUCARPIA Piment. France: INRA. PLOETZ RC. 1993. Molecular approaches to identify fusarium wilt resistance. (This volume.) RAFALSKI JA, TINGEY SV, WILLIAMS GK. 1991. RAPD markers - a new technology for genetic mapping and plant breeding. AgBiotech News Info. 3:645-648. REES H. 1961. The consequences of interchanges. Evolution 15:145-152. REES H, JONES RN. 1977. Chromosome Genetics. London, UK: Edward Arnold. REES H, SUN S. 1965. Chiasma frequency and the disjunction of interchange associations in rye. Chromosoma (Berl.) 16:500-570. RICKARDS GK. 1983. Orientation behaviour of chromosome multiples of interchange (reciprocal translocation) heterozygotes. Ann. Rev. Genet. 17:443-498. ROWE P. 1984. Breeding bananas and plantains. Plant Breed. Rev. 2:135-155. ROWE P. 1987. Banana breeding in Honduras. Pages 74-77 in Banana and Plantain Breeding Strategies (Persley GJ, De Langhe EA, eds). ACIAR Proceedings no.21. Canberra, Australia: ACIAR. ROWE PR, ROSALES FE. 1993. Breeding cooking bananas for areas with marginal growing conditions using Cardaba (ABB) in cross-pollinations. (This volume.) SHARP CB, HILLIKER AJ. 1989. Linkage and segregation distortion in Drosophila melanogaster. Genome 32:840- 846. 46 Genetic Mapping of the Banana Diploid Genome

SIMMONDS NW. 1953. Segregations in some diploid bananas. J. Genet. 51:458-469. SIMMONDS NW. 1962. The Evolution of the Bananas. London, UK: Longmans. SIMMONDS NW, SHEPHERD K. 1955. The taxonomy and origins of the cultivated bananas. Journ. Linn. Soc. Bot. 50:302-312. SLOCUM MK, FIGDORE SS, KENNARD WC, SUZUKI JY, OSBORN TC. 1990. Linkage arrangement of restriction fragment length polymorphism loci in Brassica oleracea. Theor. Appl. Genet. 80:57-64. SONG KM, SUZUKI JY, SLOCUM MK, WILLIAMS PH, OSBORN TC. 1991. A linkage map of Brassica rapa (Syn Campestris) based on restriction fragment length polymorphism loci. Theor. Appl. Genet. 82:296-304. STEBBINS GL. 1971. Chromosomal Evolution in Higher Plants. London, UK: Arnold. STOVER RH, SIMMONDS NW. 1987. Bananas, 3rd edn. New York. STUBER CW. 1992. Biochemical and molecular markers in plant breeding. Plant Breed. Rev. 9:37-61. STUBER CW, LINCOLN SE, WOLFF DW, HELENTJARIS T, LANDER ES. 1992. Identification of genetic factors contributing to heterosis in a hybrid from two elite maize inbred lines using molecular markers. Genetics 132:823-839. TADMOR Y, ZAMIR D, LADIZINSKY G. 1987. Genetic mapping of an ancient translocation in the genus Lens. Theor. Appl. Genet. 73:883-892. TANKSLEY SD, BERTNATZKY R, LAPITAN NL, PRINCE JP. 1988. Conservation of gene repertoire but not gene order in pepper and tomato. Proc. Natl Acad. Sci. (USA) 85:6419-6423. TANKSLEY SD, YOUNG ND, PATERSON AH, BONIERBALE MW. 1989. RFLP mapping in plant breeding: new tools for an old science. BioTechnology 7:257-264. VAKILI NG. 1968. Responses of Musa acuminata species and edible cultivars to infection by Mycosphaerella musicola. Trop. Agric. (Trinidad) 45:13-22. VUYLSTEKE D, SWENNEN R. 1993. Genetic improvement of plantains: the potential of conventional approaches and the interface with in-vitro culture and biotechnology. (This volume.) WANG ML, ATKINSON MD, CHINOY CN, DEVOS KM, HARCOURT RL, LIU CJ, ROGERS WJ, GALE MD. 1991. RFLP-based genetic map of rye (Secale cereale L.) chromosome-1R. Theor. Appl. Genet. 82:174-178. WEBER DA TH. 1989. Mapping RFLP loci in maize using B-A translocations. Genetics 121:583-590. WHITE MJD. 1978. Chain processes in chromosomal speciation. Syst. Zool. 27:285-298. WILLIAMS JGK, KUBELIK AR, LIVAK KJ, RAFALSKI JA, TINGEY SV. 1990. DNA polymorphisms ampliﬁed by arbitrary primers are useful as genetic markers. Nucleic Acids Res. 18:6531-6535. WILLIAMS MNV, PANDE N, NAIR S, MOHAN M, BENNETT J. 1991. Restriction fragment length polymorphism analysis of polymerase chain reaction products ampliﬁed from mapped loci of rice (Oryza sativa L.) genomic DNA. Theor. Appl. Genet. 82:489-498.

Appendix: Mapping Populations — a few guidelines for prospective developers (by M Lorieux, D González de León, CIRAD, France) The purpose of these brief guidelines is to orient prospective developers of populations for genetic studies and mapping. A few theoretical considerations are discussed and a selected bibliography is suggested for further reading.

Construction of a genetic map: marker density, population sizes, and precision of estimated genetic distances Population types of practical interest to banana genetics Ideally, the type and size of a population chosen for mapping characters of interest should allow an optimal estimation of the recombination frequencies between segregating D González de León, S Fauré 47

markers and the traits themselves. Practical considerations may suggest special population types, such as recombinant inbreds. In the case of the bananas, populations are potentially eternal and therefore one should aim at generating F2 progeny populations given that these are the most informative when using markers such as RFLPs (see below). However, in many cases, the production of F2 populations may not be possible because it is impractical to obtain entirely homozygous diploid lines, unless these already exist (e.g., banksii accessions). Figure A(1) shows the normal crossing scheme for obtaining various types of segregating populations from two homozygous parents. This was illustrated in Figure 2 of the main article by the cross between the wild diploid clones Calcutta 4 and banksii.

Other population types, such as those shown on Figure A, may be easier to obtain in the short term. Although these populations may not be optimal for mapping, the relative loss of information at a given population size may not be significant for certain applications (Gebhardt et al. 1989; Ritter et al. 1990). For instance, in order to speed up the mapping of structural differences within and between Musa taxa, a number of progeny populations could be developed from crosses or selﬁngs within structural groups; maps could then be constructed using markers already localized in a reference map: only two markers per chromosome arm (about 44 in all) should help detect major changes in gene order. Then, more intensive mapping in and around that linkage region would be pursued with additional probes to get a better estimate of break-point location.

The precision of estimated genetic distances

The precision of estimations of recombination frequencies or genetic distances, depends on three interrelated factors: the size of the population, the types of markers to be mapped, and the density of these markers (and therefore the genetic size of the genome). The more individuals are genotyped, i.e., the more information is available, the more precise is the estimation. This precision can be expressed as the standard error sr of the recombination frequency r. Allard (1956) provided formulae and tables to calculate sr in the case of pair-wise segregation data (two-point analysis), for different population and marker types. The latter fall into two categories according to their mode of segregation: codominant markers (e.g., isozymes or RFLPs) and dominant markers (e.g., RAPDs and many morphological single gene mutations).

Codominant markers (isozymes, RFLPs). When markers are codominant, both homozygous and heterozygous individuals are detected. Thus, each individual is fully informative, and the estimation of the recombination frequency attains its maximal precision. Marker order is inferred from recombination fractions between two, three, or more loci taken simultaneously. It is therefore desirable that the standard errors be very small in comparison with the value of r itself. This is illustrated in Figure B for markers having a recombination frequency of about 20%: for example, an F2 of 50 individuals should suffice to keep sr < 5% when using codominant markers. By contrast, about 28 more individuals from a backcross population are needed to achieve the same precision. 48 Genetic Mapping of the Banana Diploid Genome

Figure A. Types of populations that can be developed for genetic studies in the bananas and plantains. 1. Generation of F2 and backcross progenies from homozygous parents (some wild diploids). BC progenies can be helpful in the analysis of dominant/recessive characters and of some nucleocytoplasmic interactions. 2. A “pseudo-F2” is obtained by self-fertilization of a heterozygous individual. The difﬁculty here is that the parental genotypes are unknown and neither is the phase of the markers (see Fig.B). 3. “Pseudo-backcross” obtained from parents having common alleles or not. 4. When mostly heterozygous parents are used, having common alleles or not, certain loci will be informative in an “F1” but uninformative in another one; a strategy would consist in developing several complementary “F1”s and F2s.

Dominant markers (RAPDs). In the case of dominant markers, the estimation of the recombination frequencies is less precise than with codominant markers unless the population is a backcross, because in an F2 homozygous and heterozygous individuals cannot be distinguished. Moreover, the precision of this estimation depends on the phase of the alleles at each of the two loci in the F1 parent (see Figure B for explanations). Analysis of dominant and codominant markers simultaneously. When both codominant (e.g., RFLPs) and dominant (e.g., RAPDs) markers segregate in the same population, the precision of the estimation of their recombination frequency is intermediate between the two preceding cases. D González de León, S Fauré 49

Figure B. Standard error (sr) of the recombination frequency (r) as a function of the population size (n), when the recombination frequency between two loci is 20%. sr decreases exponentially as the population size increases, with different rates according to the type of population. BC: backcross population, codominant or dominant markers; F2 2 codo: F2 population, two codominant loci. F2 2 domC: F2 population, two loci dominant and in coupling, i.e., the two dominant alleles are on the same chromosome of F1 parent. F2 2 domR: F2 population, two loci dominant and in repulsion, i.e., the two dominant alleles are each on one of two homologous chromosomes of F1 parent. F2 1codo 1dom: F2 population, one locus codominant , the other dominant.

Marker density and order, and the “core” map The concept of the “core” map has been deﬁned in the main paper. Here we shall touch on some statistical aspects. The relative order of core markers should be established with a high probability at a given density and for a specific population size. There are several methods for determining marker order. The most powerful takes into account all the recombination events that have occurred in a linkage group in all the population (“multipoint linkage analysis”: Lander et al. 1987). Here we shall illustrate the relationship between the density of the markers given by their mean distance (Dm), the population size, and the precision of estimated r values, using simple two-point analysis. Assuming a normal distribution of the values of r, there is a 99% probability that the estimation of r is within the interval [r-2.58sr, r+2.58sr]. Such conﬁdence intervals around the recombination frequencies should not overlap, so as to have a high probability that 50 Genetic Mapping of the Banana Diploid Genome

Figure C. Standard error (sr) of the recombination frequency (r) as a function of the population size (n), when the recombination frequency between two loci takes a range of different values (inset). These curves were constructed for an F2 population mapped with codominant markers. the relative sequence of the markers is true. This means that the mean distance between ≤ markers Dm should not have a conﬁdence interval larger than Dm/2, i.e., 2.58sr Dm/2 for ≤ a 99% probability, or 1.96sr Dm/2 for a 95% probability. For example, given an F2 population of n=80 individuals mapped with codominant markers, we have to find, in ≤ Figure C, a value of r (= Dm) for which sr r/(2*2.58). We then ﬁnd that for Dm= 0.2, sr = ≤ 0.036, so we have sr Dm/(2*2.58) =0.0388. Thus, a core map with an average distance between markers of 20 cM can be constructed with an F2 population of 80 individuals, for codominant markers, and with 99% confidence intervals for the estimations of the recombination values. If the genome has a total length of, say, 2000 cM, then about 100 regularly distributed markers would be needed to approximate a usable core map.

Suggested references ALLARD RW. 1956. Formulas and tables to facilitate the calculation of recombination values in heredity. Hilgardia 24: 235-278. BURR B, BURR FA. 1991. Recombinant inbreds for molecular mapping in maize: theoretical and practical considerations. TIG 7:55-60. LANDER ES, GREEN P, ABRAHAMSON J, BARLOW A, DALY MJ, LINCOLN SE, NEWBURG L. 1987. Mapmaker: an interactive computer package for constructing primary genetic linkage maps of experimental and natural populations. Genomics 1:174-181. D González de León, S Fauré 51

RITTER E, GEBHARDT C, SALAMINI F. 1990. Estimation of recombination frequencies and construction of RFLP linkage maps in plants from crosses between heterozygous parents. Genetics 125:645-654. SNAPE JW. 1988. The detection and estimation of linkage using doubled haploid or single seed descent populations. Theor. Appl. Genet. 76:125-128. TANKSLEY SD, YOUNG ND, PATERSON AH, BONIERBALE MW. 1989. RFLP mapping in plant breeding: new tools for an old science. BioTechonology 7:257-264. TANKSLEY SD, MILLER J, PATERSON A, BERNATSKY R. 1988. Molecular mapping of plant chromosomes. Pages 157- 173 in Chromosome Structure and Function: impact of new concepts (Gustafson JP, Appels R, eds). New York, USA: Plenum Press. WILLIAMS JGK, KUBELIK AR, LIVAK KJ, RAFALSKI JA, TINGEY SV. 1990. DNA polymorphisms ampliﬁed by arbitrary primers are useful as genetic markers. Nucleic Acids Res. 18:6531-6535.

Mapping QTLs When a quantitative trait is observed in a mapping population, it is possible to establish correlations between the “phenotypic values” of the trait for each individual and the corresponding genotypes for all segregating markers in the population. Obviously, the more genotypic classes there are, the more precise will be the analysis of the trait. Codominant markers such as RFLPs will therefore be more powerful than dominant markers in F2 progeny populations. Large populations may be needed for dissecting characters under the control of many genes. The power of different methods for detecting genomic segments underlying quantitative trait loci (QTLs) has been discussed by various authors: ASINS MJ, CARBONELL EA. 1988. Detection of linkage between restriction fragment length polymorphism markers and quantitative traits. Theor. Appl. Genet. 76:623-626. KNAPP SJ. 1991. Using molecular markers to map multiple quantitative trait loci: models for backcross, recombinant inbred, and doubled haploid progeny. Theor. Appl. Genet 81:333-338. LANDER ES, BOTSTEIN D. 1989. Mapping mendelian factors underlying quantitative traits using RFLP linkage maps. Genetics 121:185-199. PATERSON AH, DAMON S, HEWITT JD, ZAMIR D, RABINOWITCH HD, LINCOLN SE, LANDER ES, TANKSLEY SD. 1991. Mendelian factors underlying quantitative traits in tomato: comparison across species, generations, and environments. Genetics 127:181-197. STUBER CW, LINCOLN SE, WOLFF DW, HELENTJARIS T, LANDER ES. 1992. Identification of genetic factors contributing to heterosis in a hybrid from two elite maize inbred lines using molecular markers. Genetics 132:823-839. 52 Applications of Biotechnologies to Banana Breeding

Applications of Biotechnologies to Banana Breeding: haplogenesis, plant regeneration from protoplasts, and transformation

F Bakry1, R Haïcour2, JP Horry1, R Megia2, L Rossignol2

Introduction Creation of resistant varieties is necessary for banana cultivation In comparison with other food crops such as cereals, little is known about banana genetics, especially as regards the determinism of main characters such as parthenocarpy or female sterility, which are a requirement for the production of edible seedless bananas (Dessaw 1988). Despite 60 years’ work in the field of hybridization, all cultivated varieties are still natural and issued from clonal selection (Rowe 1984). Due to disease extension in bananas the creation of new cultivars or the improvement of already-existing bananas becomes necessary. The main diseases are black leaf streak due to Mycosphaerella ﬁjiensis Morelet, Panama disease due to Fusarium oxysporum Schlecht f. cubense (E.F. Smith) Snyd. & Hams race 1, with some new pathogenic strains, moko disease due to Pseudomonas solanacearum E.F. Smith race 2, banana bunchy top virus (BBTV), cucumber mosaic virus (CMV), weevils, nematodes, etc. The restricted genetic base of cultivated clones makes banana growing a particulary fragile activity because of the rapid extension of new strains of diseases. In addition, the threat of pest and disease attack is increased by monoculture and the lack of soil rotation (Delvaux et al. 1990). For many diseases— Panama, moko, BBTV, CMV — no chemical treatment exists, and destruction of the affected plants is the only method of control. For the others, the cost of chemical treatments is increasing as more virulent strains appear, or even prohibitive for small farmers in developing countries, or else damaging to the environment. In particular, this is obvious for black leaf streak which has expanded through South America and Africa (De Langhe 1987; Bakry et al. 1989). 1Station de Neufchâteau, CIRAD-IRFA, 97130 Sainte Marie, Guadeloupe, FWI; 2Laboratoire de Morphogenèse végétale expérimentale, Université Paris-Sud XI, Bât.360, 91405 Orsay Cedex, France F Bakry, R Haïcour, JP Horry, R Megia, L Rossignol 53

Banana breeding programs supported by the EEC Since 1982 CIRAD-IRFA has developed a program of banana breeding supported by the EEC. The results obtained during programs STD1 and STD2 improved our knowledge of the structure of the Musa genus and gave a better understanding of the mechanisms of sterility in Musa (CIRAD-IRFA, Neufchateau Station, Guadeloupe, FWI) (Dessauw 1988), as well as the resistance to Cercospora spp. (CIRAD-IRFA Phytopathology Laboratory, Montpellier, France; CRBP Cameroon) (Carlier, Mourichon 1990; Fouré et al. 1990). Progress has also been achieved in the genetic mapping of diploid bananas, which will provide better genetic control of agronomic traits (CIRAD-BIOTROP, Montpellier, France) (Fauré et al. 1991). In addition, progress has been obtained in in-vitro culture, especially regarding vegetative propagation and regeneration (in-vitro culture laboratories at KU Leuven, Belgium; CIRAD-IFRA/CATIE, Costa Rica; CIRAD-BIOTROP, Montpellier, France; Paris XI University, Orsay, France) (Banerjee et al. 1987; Sannasgala 1989; Escalant, Teisson 1988, 1989; Harran 1991; Megia et al. 1992). The aim of the STD3 program, which has just been agreed by the EEC, is to identify new banana cultivars suitable for export and new plantains for local use resistant to or tolerant of black leaf streak, which is the main threat to banana production. In addition, good productivity, structure, and organoleptic quality of fruits have to be maintained, as well as parthenocarpy and sterility. For that purpose six different projects will be developed by the groups involved in the previous STD1 and STD2 programs, as follows. 1. Genetic resources: collection and conservation of clones selected on the basis of morphotaxonomic and genetic evaluation, to permit the creation of databases that can be made available for breeders. 2. Genetic studies and diploid breeding: to control sterility, to understand the genetic coding for agronomic traits and resistance characters, and to create new lines for triploids. 3. Agronomic evaluation of diploids and newly created varieties. 4. Varietal improvement through 3x cv x 2x → 4x cv crosses. 5. Nonconventional varietal improvement by using protoplasts and genetic engineering. 6. Studies on the local distribution and genetic variability of Mycosphaerella ﬁjiensis in relation with the host. This paper describes original data obtained during the STD2 program by three groups that are now involved in the STD3 program in the ﬁeld of biotechnology: - from Paris XI University, Orsay, France; CIRAD-IRFA, Neufchâteau Station, Guadeloupe, FWI; and CATIE, Costa Rica.

The Strategy Chosen by CIRAD-IRFA and its Collaborators This strategy, as agreed for the STD3 program, is focused on two main topics. 1. The correction of defects in well-established cultivars, such as Cavendish (AAA) or Plantain (AAB), in order to acquire a speciﬁc resistance, for example. 54 Applications of Biotechnologies to Banana Breeding

2. The creation of new synthetic hybrids, which will be either sweet bananas or cooking bananas, according to the selection criteria chosen. A combination of conventional breeding methods with new techniques in genetic engineering is necessary to achieve these objectives.

Results Most of the natural edible parthenocarpic varieties of bananas are heterozygous, unlike their wild relatives. This implies a complicated inheritance pattern because the observed characters rely both on simple or polygenic traits (additive values) and on improved performance gained from poly-allelism (heterosis) at multiple loci (epistasis). Beside the genetic aspects, structural heterozygoty (exchanges, inversions) can also partly explain the sterility of edible diploid bananas. Thus the most rational use of diploid varieties is the obtention of more fertile pure lines in which the gene potential will be highly inheritable and can produce F1 hybrids usable in banana breeding. It is therefore necessary to obtain homozygotic bananas in order to create controlled F1 hybrids suitable for our breeding program. Haplo-diploidization techniques, already applied in some monocotyledon programs, are a valuable model for the creation of new pure lines in bananas. This is why we started a haplomethod program with banana using immature pollen, i.e., in-vitro androgenesis on the one hand, and as unfecunded oospheres (i.e., induced parthenogenesis) on the other.

Haplomethods

Androgenesis Recent successful results were obtained in Guadeloupe with anther culture of wild Musa acuminata ssp. burmannica cv Long Tavoy (Fig.1). A creamy white callus (1 mm diameter) emerging through the dehiscence split was isolated and subcultured. After transfer to a shoot-initiating medium, vegetative buds developed on the callus. Plantlets rooted when isolated, and were transferred to a basal medium (BM). About 20 weak plants were weaned, but 10 only survived in the greenhouse, showing more or less variegated leaves. Chromosome counts carried out on root tips showed four plants to be diploid (2n=2x=22) and the others are probably diploid too, as suggested by their identical form. The microsporial origin of these four plants was proved by isoenzymes studies: the mother plant presents a three-band pattern heterozygous locus with dimeric enzymes (Got. B and Est. A), whereas there is a one-band pattern homozygous locus with the regenerated plants.

Induced parthenogenesis The method currently developed at CATIE consists of the use of irradiated pollen (by gamma ray) to stimulate, after pollination, the embryo development of the oosphere F Bakry, R Haïcour, JP Horry, R Megia, L Rossignol 55

Figure 1. The different stages of anther culture, material, and culture conditions. without fertilization, to obtain, in the seedlings, haploid or spontaneously doubled haploid plants. Pollen from M. acuminata ssp. burmannicoides and M. balbisiana cv Tani was irradiated with different gamma radiation doses (3-100 krad, 60cobalt source), and then used to pollinate female ﬂowers of M. acuminata ssp. burmannicoides. Observations with a ﬂuorescent microscope revealed that the germinating tube grows to reach the ovule at any tested dose of pollen irradiation, but only the treatments between 0-10 krad permitted embryo development. Moreover, endosperm-vacant seeds were found in all doses used, but more markedly at those above 8 krad. 56 Applications of Biotechnologies to Banana Breeding

Embryo rescue by in-vitro transfer permitted the development of whole plants originating from hybridization with pollen irradiated with 3, 5, 7, 8, and 10-krad treatments. Karyotypes carried out on root tips showed the plants evaluated to be mainly diploid, although two weak plants were found to be haploid (one with 5-krad and the other with 7-krad treatments). Discussion These preliminary results obtained by our group for the first time show that both techniques described above (androgenesis and induced parthenogenesis) are applicable to banana. The techniques make possible a reduction in the cost and duration of experiments (at least eight generations are necessary to reach homozygosity by self- fertilization (Fig.2). Additionally, in-vitro androgenesis is the only technique available for obtaining homozygotic plants in some clones from 2x parthenocarpic clones as 2n, which are usually female-sterile, with weak male fertility. Finally, these techniques enable us to obtain new kinds of plant material such as haploid callus and haploid cell cultures, which create new perspectives in the field of somatic cell fusions or transformations.

Protoplast regeneration and transformation For this purpose the group of nonconventional banana breeders in Orsay has developed special protoplast techniques. These are of considerable interest in the production of nonconventional hybrids, and for electroporation transformation. At this time the main difficulty with banana protoplasts is the low potential of protoplasts to divide and to differentiate into a whole plant. To overcome this problem the Orsay group has focused its work on the culture and differentiation of protoplasts. Bakry (1984), followed by Da Silva Conceicâo (1989), observed low reactivity in banana protoplasts, and Da Silva Conceicâo reported that few divisions could be obtained, and only with embryogenic material. Working with the CIRAD-BIOTROP CIV laboratory (CIRAD-IRFA, Montpellier, France), which kindly provided appropriate plant material (embryogenic suspension of diploid Musa acuminata ssp. burmannica cv Long Tavoy obtained by Escalant [Escalant, Teisson 1989]), the Orsay team, using a feeder cell technique, succeeded in obtaining continuous divisions from protoplasts to microcolonies, up to the callus stage. This kind of experimentation, however, is limited by the long- term embryogenic culture used as cell suspension, which does not allow us to obtain embryos and whole plants. So we have used other plant material with better potential. Through suitable collaboration with the Laboratory of Tropical Crop Husbandry (at KU Leuven, Belgium), which has proven experience in somatic embryogenesis (Dhed’a et al. 1991), we have been able to improve our results signiﬁcantly. We cultivated thin layers (called scalps) from proliferating buds from shoot meristems of a Bluggoe clone (ABB). They provide highly morphogenic material (with highly embryogenic cell suspension) from which we have been able to obtain protoplasts. When cultured, these protoplasts differentiated into embryos and later in whole plants. F Bakry, R Haïcour, JP Horry, R Megia, L Rossignol 57

Figure 2. Comparison between the conventional pedigree method and in-vitro androgenesis. These last results are of considerable interest regarding the use of biotechnologies in CIRAD’s genetic program. In addition, unpublished original data have been obtained by the Orsay group. Using a particle gun, expression of a foreign inserted gene (GUS) in the diploid cell culture of bananas has been obtained for the ﬁrst time. This result opens up promising potentials in the transformation of this plant. 58 Applications of Biotechnologies to Banana Breeding

Conclusions and Prospects The techniques described above are useful for CIRAD-IRFA’s work on banana improvement (Bakry et al. 1990), in which the hybridization strategy is focused on diploid plant synthesis. Special attention is also being given to the improvement of wild clones and edible cultivars (Fig.3).

Figure 3. The diploid-triploid approach to banana breeding at CIRAD-IRFA. F Bakry, R Haïcour, JP Horry, R Megia, L Rossignol 59

First, however, we have to improve some diploid populations, targeting disease resistance and agronomic and organoleptic traits. For this purpose bulk selection followed by genetic mixing will be done in recurrent selection. From these improved populations, pure lines will be chosen and intercrossed to obtain heterozygous, interpopulation, diploid hybrids F1. To obtain homozygotic plants, the classical scheme consists in self-repeated pollinations. At present, ﬁrst-generation plants are observed and CIRAD-IRFA is working on in-vitro haplogenesis in order to obtain homozygotic plants more rapidly. Obtention of triploid varieties, which is the ultimate objective, may be achieved in three different ways.

1. By crossing a diploid F1 hybrid (restitution of the diploid female F1 hybrid genome in 2x female gametes) with another homozygotic male diploid (x male gametes). However this procedure is uncertain and uncontrolled.

2. By crossing a tetraploid plant (from an F1 diploid clone, coupled with colchicine treatment) with another homozygotic improved diploid. The success of this scheme depends on the meiotic behavior of the tetraploid. The best gametes, containing genetic potential of the F1 diploid parent, will be obtained from the allotetraploid AABB structure.

3. By using biotechnologies: fusion of 2x F1 protoplasts with a haploid protoplast produced from anther culture via haploid embryogenic callus and haploid cell suspension, like those obtained with other crops (Sun et al. 1989). These 3x somatic hybrids will be screened using flow-cytometry apparatus, for example. The advantage of this scheme is that the F1 diploid genetic structure of the 2x parent is always preserved. Our new results in the field of banana cell transformation open up new perspectives in the breeding of such edible diploid or triploid bananas as Cavendish. It is necessary to obtain transferable DNA, whether it comes from banana or not, and to choose with precision the best plant target, tissue, or cell type. A particle gun would be used to transform all plant material with morphogenic potential. Electroporation would also be used on suitable protoplasts. Many techniques are already used daily with success in our program. In the field of hybridization, electrophoresis reveals the similarity between clones or the degree of heterozygous differentiation; and cytogenetics permits the control of interspecificity of hybrids. In-vitro culture will strengthen our program with vegetative propagation, zygotic immature embryo culture, callogenesis, somatic embryogenesis, and polyploidization to tetraploid level with meristem diploid clones treated with colchicine. And finally, haplogenesis, all protoplast techniques, and genetic engineering enable us to anticipate what challenges lie ahead for us in our work.

Acknowledgments This research has been undertaken with financial support from CIRAD-IRFA via EEC funding of DG 12 (Contract no.TS ZA-0094-F). We also thank C Teisson and research workers in the Laboratory of Tropical Crop Husbandry for their invaluable assistance. 60 Applications of Biotechnologies to Banana Breeding

References

BAKRY F. 1984. Choix du matériel à utiliser pour l’isolement de protoplastes de bananiers (Musa spp.), Musacées. Fruits 39:449-452. BAKRY F, HORRY JP, TEZENAS DU MONTCEL H, GANRY J. 1989. L’amélioration génétique des bananiers à l’IRFA/CIRAD. Fruits (numéro spécial): 25-40. BANERJEE N, SCHOOFS J, HOLLEVOET S, DUMORTIER F, DE LANGHE E. 1987. Aspects and prospects of somatic enbrogenesis in Musa AAB, cv. Bluggoe. Acta Hort. 212:727-729. CARLIER J, MOURICHON X. 1990. Polymorphisme de fragments de restriction (RFLP) chez les Mycosphaerella ssp., pathogène des bananiers et plantains. IIème Congrès Société Française de Phythopathologie, Montpellier, November 1990. DA SILVA CONCEICÂO A. 1989. Isolement et culture de protoplastes de bananiers (Musa sp.): étude de divers facteurs. Orsay, Paris: DEA Génétique et Sélection Animale et Végétale, option Végétale UPS. 14pp. DESSAW D. 1988. Etude des facteurs de stérilité du bananier (Musa ssp.) et des relations cytotaxonomiques entre M. acuminata Colla et M. balbisiana Colla. Parts 1,2,3. Fruits 43:539-558;615-638; 685-700. DE LANGHE E. 1987. Necesidas de una estrategia internacional para al mejoramiento genetica del Banana y del Plantano. Pages 181-199 in Memoria de la Reunion Regional de INIBAP para America Latina y el Caribe, San José. DELVAUX B, PERRIER X, GUYOT P. 1990. Diagnostic de la fertilité de systèmes culturaux intensifs en bananeraies à la Martinique. Fruits 45:223-236. DHED’A D, DUMORTIER F, PANIS B, VUYLSTEKE D, DE LANGHE E. 1991. Plant regeneration in cell suspension cultures of the cooking banana cv “Bluggoe” (Musa ssp. AAB Group). Fruits 46:125-135. ESCALANT JV,TEISSON C. 1988. Embryogenèse somatique chez Musa sp. C.R. Acad. Sci. Paris, série III 306:277-281. ESCALANT JV, TEISSON C. 1989. Somatic embryogenesis and plants from immature zygotic embryos of species Musa acuminata and Musa balbisiana. Plant Cell Reports 7:665-668. FAURÉ S, CARREEL F, AMSON C, AUBERT G, HORRY JP, GONZÁLEZ DE LEÓN D, LANAUD C. 1991. RFLP analysis in the bananas (Musa ssp.). Eucarpia Symposium on Genetic Manipulation in Plant Breeding: Molecular Biology/Breeding Interface, Tarragona, Spain. FOURÉ E, MOULIOM PEFOURA A, MOURICHON X. 1990. Etude de la sensibilité variétale des bananiers et des plantains à Mycosphaerella ﬁjiensis Morelet au Cameroun. Caractérisation de la résistance au champ de bananiers appartenant à divers groupes génétiques. Fruits 45:339-345. HARRAN S. 1991. Etude de cals inﬂorescentiels de deux cultivars de bananiers (901 et GN) et des plantes qui en sont issues. Ph.D. thesis, Uuniversité de Paris-Sud XI, Orsay, France. 110 pp. MEGIA R, HAÏCOUR R, ROSSIGNOL L. 1992. Callus formation from banana (Musa sp.) protoplasts. Plant Breeding Reviews 7:135-155. ROWE PR. 1984. Breeding bananas and plantains. Plant Breeding Reviews 7:135-155. SANNASGALA K. 1989. In vitro somatic embrogenesis in Musa. Ph.D. thesis. KU Leuven, Heverlee, Belgium. SUN CS, PRIOLI LM, SÖNDAHL MR. 1989. Regeneration of haploid and dihaploid plants from protoplasts of supersweet (sh2sh2) corn. Plant Cell Reports 8:313-316.

Further Information on Protoplast Regeneration and Transformation in Musa (by R Haïcour, L Rossignol)

Introduction A combination of conventional breeding programs with new methods in genetic engineering is necessary for banana improvement. These new strategies must be interpreted as complements to and not replacements for conventional breeding. F Bakry, R Haïcour, JP Horry, R Megia, L Rossignol 61

This is why at the Laboratoire de Morphogenèse végétale expérimentale, Université de Paris-Sud XI, we decided to develop protoplast techniques for use in the banana breeding program. Protoplasts, as single-cell systems, are particularly appropriate for applications of plant genetic engineering such as somatic hybridization and transformation by protoplast fusion and electroporation, respectively. Protoplasts appear to be tools of special interest in banana and plantain improvement.

Protoplasts: obtention, division, and regeneration The first banana protoplasts were isolated in our group in 1983 by Bakry. Then Cronauer and Krikorian (1984-1986), Chen (1985), Chaput in our laboratory (1986), and Da Silva Conceicâo (1989) also obtained protoplasts from different explants of banana. But, up to now (1993), the main difficulty with banana protoplasts has been their low potential to divide. The very low frequency of first and second divisions of banana protoplasts cultured in sea plaque agarose has been reported, but only when embryogenic cells were used (Da Silva Conceicâo, 1989). However, sustained protoplast division and differentiation in plants are a prerequire- ment for using protoplasts in a breeding program. To overcome this problem, we decided to work with the most reactive material, i.e., with embryogenic material. So we used embryogenic cells of an aged diploid suspension of Musa acuminata ssp. burmannica cv Long Tavoy (AA), kindly provided in 1987 by JV Escalant (CIRAD, Montpellier, France) and maintained in our laboratory by weekly subcultures. To stimulate protoplast division, we tried to use feeder cells. Cell suspensions from three species, Lolium multiflorum, Triticum monococcum, and Zea mays were tested. The best results were obtained with Lolium feeder cells, with a high density of protoplasts plated above the feeder cell layer and separated from the nurse cells by a physical barrier (a double nylon mesh). The nurse cells in the feeder layers proved to be critical in supporting the sustained protoplast divisions, the formation of microcolonies, and subsequent growth up to the callus stage. After 3 months of culture, protoplast-derived calli of banana died, even though the nurse cells were still growing. Conversely, calli continued to grow when subcultured on fresh nurse medium. Cytohistological sections showed that the cells of protoplast-derived calli retained the embryogenic characters previously observed in the cell suspensions used as a source of protoplasts: particularly a large nucleus, important reserves, and a thick cell wall. Moreover, morphological and cytological study revealed that cells of banana were different from those of Lolium and the protoplast-derived calli obtained had all the characteristics of banana cells. These latter were smaller in size than Lolium cells and had a denser cytoplasm. Chromosomes were larger in Lolium than in dividing cells of banana. In order to confirm that calli developed above the nylon filters derived from banana protoplasts and not from the contamination of nurse cells, such calli were examined for isoenzymes. Phosphoglucomutase (PGM) and alcohol deshydrogenase (ADH) were found 62 Applications of Biotechnologies to Banana Breeding

to distinguish banana from Lolium cells. Since the zymogram of the calli located above the nylon ﬁlters showed an identical pattern to that of banana cells, it was concluded that they really derived from banana protoplasts. But this technique was still limited by a lack of regeneration due probably to the long- term embryogenic culture used as cell suspension. At present (1993), this suspension does not allow us to obtain morphogenesis with calli derived from protoplasts (Megia et al. 1992). So we have used other plant material with better potential. Collaboration with the Laboratory of Tropical Crop Husbandry (at KU Leuven, Belgium), which is experienced in somatic embryogenesis, permitted us signiﬁcantly to improve our results. Referring to data from Dhed’a et al. (1991), we cultivated thin layers (scalps) from proliferating shoot-tip cultures of a cooking banana (ABB), subgroup Bluggoe cv Matavia. They provided a good embryogenic cell suspension from which we have been able to obtain protoplasts. When cultured on medium with feeder cells, protoplasts divided and differentiated into globular embryos and later in whole plants (pers. com. Megia et al.). These original results obtained by the Orsay group are of particular interest in relation to the use of biotechnologies in the CIRAD genetic program.

First results with cell transformation An unpublished original result has just been obtained (1993) by the Orsay team using a powder gun, as described by Zunbrunn et al. (1989). Fresh diploid cell suspension culture of M. acuminata ssp. burmannica cv Long Tavoy were shot with tungsten particles of 0.5-1.0 µm diameter coated with a DNA solution. Plasmid DNA derived from PI BI 221, containing the Ca MV 35S promoter, Gus coding region, and Nos terminator was used as a chimeric gene introduced in banana cells. Histochemical Gus assay was performed with X glu substrate, and then we obtained for the ﬁrst time the expression of a foreign-inserted gene in banana. This preliminary result shows the feasibility of transformation in this important crop.

Reference ZUNBRUNN G, SCHNEIDER M, ROCHAIX JD. 1989. A simple particle gun for DNA mediated cell transformation. Technique A. Journal of Methods in Cell and Molecular Biology 1: 204-216. C Rivera, P Ramírez, R Pereira 63

Preliminary Characterization of Viruses Infecting Banana in Costa Rica

C Rivera, P Ramírez, R Pereira

Introduction Costa Rica, the second largest worldwide exporter of bananas, has retained this position by increasing yields through better methods of production and the recuperation of cultivated areas (Informe UPEB 1990). Some viral diseases are a real threat for the production of bananas (Dale et al. 1986). Different symptoms have been observed in commercial banana fields in Costa Rica that could be associated with viral infection. These symptoms are “distorted candle leaf”, rugose leaves with thickened and chlorotic veins, pockets of necrosis in the sheaths, mild or severe chlorosis, chlorotic streaking or flecking, and bunched leaves at the apex of the plant. CORBANA (Corporación Bananera Nacional) and the University of Costa Rica have started a cooperative research project to characterize the viral diseases that could be associated with these different symptoms. Although only one virus disease has been reported in Costa Rica’s plantations —cucumber mosaic virus (CMV), the type member of the Cucumovirus group (Brioso et al. 1986)—the possibility of ﬁnding other virus diseases is present. The Cucumovirus is a small group of serologically interrelated viruses each of which includes numerous biological variants (Francki 1985). A large number of strains of CMV have been recognized by host range and symptomatology (Kaper, Waterworth 1981), serological relationships (Devergne, Cardin 1973), peptide mapping of the viral coat protein (Edward, Gonsalves 1983), physical and chemical properties (Lot, Kaper 1976), and nucleic acid hybridization (Gonda, Symons 1978; Piazolla et al. 1979; Owen, Palukaitis 1988). Devergne and Cardin (1973) and Devergne et al. (1981) have grouped a large number of isolates of CMV on the basis of biological properties as well as precipitin and Elisa testing into two main groups: ToRS and DTL. They have reported that there is a third serotype—CMVCo—that appears to be serologically distinct from both of these groups. This paper describes the preliminary characterization of two CMV-related viruses isolated from banana by symptomatology, host range, serology, and electron microscopy, and reports the presence of two other virus-like particles in commercial banana plantations in Costa Rica.

Centro de Investigación en Biología Celular y Molecular, Universidad de Costa Rica, San José, Costa Rica 64 Preliminary Characterization of Viruses Infecting Banana in Costa Rica

Materials and Methods

Preliminary analysis by Elisa of the different symptomatologies In 1990, a survey was made in several banana fields, in which plants were selected that presented different symptomatologies frequently observed. Leaf samples were collected from 39 different plants. Sheath samples were collected only from those plants showing pockets of necrosis. All samples were analyzed by DAS-Elisa using the CMVer and CMVvi diagnosis kit from AGDIA and the Sanofi CMV of banana diagnosis kit. The Sanofi kit is made of polyclonal antibodies produced against the three main serotypes of CMV, known as DTL, ToRS, and Co. The AGDIA CMVvi kit reacts with some strains of serotype DTL and the AGDIA CMVcr kit reacts with some strains of serotype ToRS. The Elisa test was done following the protocols recommended by the manufacturer. One plant from each of two different symptom prototypes (Types 1 and 2) commonly observed in the field (Fig.1), pre- senting also very different results by Elisa testing (Table 1), were selected for further analysis. Both selected plants were of cv Grande Naine. Type 1 was reproduced by rhizome, while Type 2 was reproduced by tissue culture. Type 1 showed bunched leaves at the apex of the plant, leaf distortion (Fig.1,A), pockets of necrosis in the sheaths (Fig.1,B), and distorted fruit (Fig.1,C). Type 2 showed “distorted candle”, rugose leaves with thickened and chlorotic veins (Fig.1,D). Type 2 Figure 1. Symptoms associated with banana collected in Costa Rica. A: Bunched leaves. B: Necrotic pockets in the sheaths. did not show pockets C: Distorted fruits. D: Distorted candle—rugose leaves with of necrosis in the sheaths. thickened and chlorotic veins. C Rivera, P Ramírez, R Pereira 65

Virus purification Virus was partially purified from banana Types 1 and 2 by a modification of the method described by Palukaitis (pers. com.). Infected leaf tissue, minus midrib, was finely chopped and ground in liquid nitrogen and homogenized in 1.5 vol (w/v) of 0.5 M sodium-citrate buffer, pH 6.5, containing 5mM EDTA and thioglycolic acid 0.5% and 1.5 vol (v/v) of chloroform. After a cycle of differential centrifugation (15 000 g for 10 min, 186 500 g for 90 min) the pellet was resuspended overnight at 4°C in 5 mM sodium- borate buffer, pH 9, containing 5mM EDTA and 2% Triton X-100. The suspension was clarified by low-speed centrifugation (7000 g for 10 min) and the virus was precipitated from clarified extract by ultracentrifugation (218 000 g for 60 min) through a 10% (w/v) sucrose cushion prepared on 5 mM sodium-borate buffer, pH 9, containing 2% (v/v) Triton X-100, 5 mM EDTA. The virus preparation was resuspended in 1 mL of the same buffer.

Electron microscopy Partially puriﬁed virus preparations of Types 1 and 2 were negative-strained with 1% uranyl acetate, pH 5, on Formvar-Carbon coated grids (400 mesh), and observed in an H-7000 Hitachi electron microscope.

Host range symptomatology Partially puriﬁed virus preparations from banana Type 1 and crude sap of banana Type 2 were mechanically inoculated into healthy plants of Curcurbita pepo, Cucumis sativus, and Nicotiana tabacco cv Samsun. Inoculations were made by rubbing carborundum-dusted cotyledons and expanding ﬁrst leaves. The inocula were prepared in cold 5-mM sodium-borate buffer, pH 9, containing 5 mM EDTA, 2% Triton X-100. The leaves were washed with distilled water after inoculation and were observed for symptoms.

Results

Table 1. Elisa test results of the two selected types. Absorbance 405 nm Elisa Type 1 X + 3s Type 2 X + 3s healthy healthy CMVvi 0.265 0.047 0.049 0.014 CMVcr 0.203 0.054 0.036 0.022 CMV Sanoﬁ 0.134 0.038 0.003 0.005 To establish limits for the negative and positive values the upper negative limit was taken as the absorbance mean of the healthy values (X) plus 3s (standard deviation) (P =0.003). 66 Preliminary Characterization of Viruses Infecting Banana in Costa Rica

Electron microscopy Isometrical particles of about 29 nm with hollow centers were observed in the partial purified preparation from banana Types 1 and 2 (Fig.2, A-B). Flexuous-filamentous and rigid-filamentous virus-like particles were also observed in banana Type 2 (Fig.2, C-D).

Figure 2. Virus particles associated with infected field bananas. A: Virus particles purified from banana Type 1. B,C,D: Virus particles purified from banana Type 2. The bar represents 100 nm.

Host range symptomatology Symptoms induced by Types 1 and 2 were compared on zucchini, cucumber, and tobacco (Table 2). Figure 3 shows some of the most important symptoms observed. Table 2. Host range symptomatology. Hosts Inoculumsource Cucumber Zucchini Tobacco Type 1 Mild, mosaic, Mild mosaic, Symptomless rugose leaf rugose leaf Type 2 Chlorotic local Mild mosaic, Symptomless lesion, later rugose and systemic rugose curling leaf and curling leaf, enations C Rivera, P Ramírez, R Pereira 67

Figure 3. Symptoms observed in indicator plants inoculated separately with banana Type 1 and Type 2. A: Local lesions induced in cucumber by inoculation with Type 2. B: Enation observed in cucumber after inoculation with Type 2. C: Chlorotic leaf veins and severe mosaic observed in cucumber by inoculation with Type 2. D: Mild mosaic and leaf curling after the inoculation of zucchini separately with Type 1 and Type 2.

Conclusions The results obtained in this work suggest the presence of CMV associated with the Type 1 plant. The symptoms observed in this type are similar to some of those reported for some CMV strains of banana (Frison, Putter 1989). The electron micrographs of partially purified virus preparation showed spherical particles of about 29 nm with hollow centers similar to those reported for CMV (Tolin 1977). The DAS-Elisa results confirmed the CMV presence in Type 1. However, these results did not permit differentiation to be made between serological subgroups. The electron microscopy of partially purified samples of plant Type 2 showed very few isometrical viral particles similar to those found in the Type 1 plant. However, the DAS- Elisa results of these samples did not permit confirmation of the presence of CMV. By the standards commonly used in virus-plant Elisa testing (X + 3s of the healthy control samples), our results would be rated positive. However, these values are very low positives or unusually high negatives. There exist very few test results with banana sap analysis by DAS-Elisa commercial kits. It is necessary to analyze a bigger number of healthy samples and samples showing Type 2 symptoms in order to establish positive- negative thresholds. The symptoms observed in zucchini and tobacco when they were inoculated with both types of banana plants, were similar. Some differences were observed in cucumber plants inoculated with Types 1 and 2. Only Type 2 developed chlorotic local lesions in the inoculated leaves and enations on systemic infected leaves. 68 Preliminary Characterization of Viruses Infecting Banana in Costa Rica

The two other virus-like particles associated also with Type 2 have not yet been analyzed by serology. It is necessary to complete the characterization of the banana CMV on the following basis: (a) biological properties, using a large variety of indicator plants; (b) serological properties, determining the serological relationship with known CMV strains and other members of the Cucumovirus group, such as TAV and PSV; and (c) molecular properties, analyzing the ability to hybridize with eDNA from the members of the Cucumovirus group. It is also important to start the characterization of the two other virus-like particles associated with the Type 2 plant.

References

BRIOSO PST, RIVERA C, ARROYO T, RODRIGUEZ CM. 1986. Identificao e distribucao du virus do mosaico do pepino em areas de cultivo de bananeira e platano na Costa Rica (America Central). Fitopatología Brasileira 11:392. DALE JL, PHILLIPS DA, PARRY JN. 1986. Double-stranded RNA in banana plants with bunchy top disease. J. Gen. Virol. 67:371-375. DEVERGNE JC, CARDIN L. 1973, Contribution a l’étude du virus de la mosaïque du concombre (CMV). IV. Essai de classification de plusieurs isolats sur la base de leur structure antigénique. Ann. Phytopathol. 5:409-430. DEVERGNE JC, CARDIN L, BURCKARD J, VAN REGENMORTEL MHV. 1981. Comparison of direct and indirect ELISA for detecting antigenically related cucumovirus. J. Virol. Methods 3:193. EDWARDS MC, GONSALVES D. 1983. Grouping of seven biologically defined isolates of cucumber mosaic virus by peptide mapping. Phytopatology 73:1177-1120. FRANCKI RIB. 1985. The viruses and their taxonomy. Pages 1-15 in The Plant Viruses, vol.1 (Francki RIB, ed.). New York, USA: Plenum Press. FRISON EA, PUTTER CAJ (eds). 1989 FAO/IBPGR Technical guidelines for the safe movement of Musa germplasm. Rome, Italy: Food and Agriculture Organization of the United Nations/International Board for Plant Genetic Resources. GONDA TJ, SYMONS RH. 1978. The use of hybridization analysis with complementary DNA to determine the RNA sequence homology between strains of plant viruses: its application to several strains of cucumoviruses. Virology 88:361-370. Informe UPEB. 1990. Panamá 14:7-20. KAPER JM, WATERWORTH HE. 1981. Cucumoviruses. Pages 257-332 in Plant Virus Infections and Comparative Diagnosis (Kurstak E, ed.). New York, USA: Elsevier/Holland Biomedical. LOT H, KAPER JM. 1976. Physical and chemical differentiation of three strains of cucumber mosaic virus and peanut stunt virus, Virology 74:209-222. OWEN J, PALUKAITIS P. 1988. Characterization of cucumber mosaic virus. I. Molecular heterogeneity mapping of RNA3 in eight CMV strains. Virology 166:495-502. PIAZZOLA P, DIAZ-RUIZ JR, KAPER JM. 1979. Nucleic acid homologies of eighteen cucumber mosaic virus isolates determined by competition hybridization, J. Gen. Virol. 45:361-369. TOLIN SA. 1977. Cucumovirus group. Pages 303-309 in The Atlas of Insects and Plant Viruses, including mycoplasma viruses and viroids. New York, USA: Academic Press. CM Fauquet, RN Beachy 69

Status of Coat Protein-Mediated Resistance and its Potential Application for Banana Viruses

CM Fauquet1, RN Beachy2

Introduction The recent development of gene transfer technologies to plants has made it possible to transfer useful traits to a number of crop plants. Among the different possibilities, virus resistance is probably one of the most successful applications of plant genetic engineering. When used to complement current breeding programs, these technologies have the potential to help control plant viruses and consequently to decrease the impact of plant viruses on crop productivity. During the past 6 years different molecular approaches have been developed to attempt to control viruses, some of which are still under investigation in laboratories, while others have reached the field-testing level. This paper briefly reviews these different approaches, discussing the state of development, the efficacy, and the stability of each concept for controlling viruses by genetic engineering. Particular emphasis is placed on coat protein-mediated resistance (CPMR) because of the success of this strategy and because of the large resource of data and examples now available derived from the use of this technique. The type of resistance obtained and the spectrum of protection produced is exposed, and the examples of viruses belonging to virus groups the members of which are infecting bananas are particularly detailed in order to evaluate the potential of application of the coat protein strategy for this crop.

Multiple Strategies for Controlling Viruses With the increase of the knowledge of virus replication and virus genome organization, scientists are testing different approaches for controlling virus infection. Different viral genes or sequences, inserted into the plant genome, interfere with the virus replication in a beneﬁcial way for the infected plant.

1ORSTOM, Division of Plant Biology, MRC7, 10666 Torrey Pines Road, La Jolla, CA 92037, USA; 2TSRI Plant Division, MRC7, at the same address 70 Status of Coat Protein-Mediated Resistance and its Potential Application for Banana Viruses

Ribozyme strategy A novel approach to achieve virus resistance is the use of autocatalytic RNA-cleaving molecules, also called “ribozymes” (Cech 1986). Viroid RNAs such as avocado sunblotch viroid and satellite RNAs such as the satellite of tobacco ring spot virus (TobRSV) have the possibility of self-cleavage during replication (Buzayan et al. 1986; Forster, Symons 1987; Hutchins et al. 1986; Prody et al. 1986). Cleavage is effective both on the positive and negative strand of the RNA and is highly specific and associated with conserved sequence domains. Several studies have been conducted to determine the optimal in-vitro conditions of cleavage (Gerlach 1989; Haseloff, Gerlach 1988). Replicase genes of tobacco mosaic virus (TMV) and barley yellow dwarf virus (BYDV) encoding sequences bearing specific virus cleavage sites have been integrated in transgenic plants and should generate sequence-speciﬁc endonuclease activities (Gerlach 1989), but no in-vivo results have been published yet.

Translation strategy By translation strategy we refer to the strategies involving integration in the plant genome of sequences generating complementary sequences to viral RNA that interfere with the translation of viral genes by hybridization of the coding sequence. It has been reported that synthesis of complementary RNA (antisense RNA) can decrease the accumulation of gene products in both procaryotes and eukaryotes (Ecker, Davis 1986; Green et al. 1986). It is likely that antisense RNAs anneal with sense RNAs to form a double-strand complex which is rapidly degradated or which inhibits translation of the RNA. Several viral coat protein antisense constructs including TMV, cucumber mosaic virus (CMV), and potato virus X (PVX) have been integrated into plants and tested for resistance to infection (Cuozzo et al. 1988; Hemenway et al. 1988; Powell et al. 1989). In all these cases resistance has been reported against infection by the homologous virus but only at low virus inoculum concentrations. Recently, a similar study done with the potato leafroll virus (PLRV) (Kawchuk et al. 1991) proved to induce resistance to the virus inoculation of the same level as the CPMR.

Replication strategy One of the ﬁrst steps of the virus cycle is replication of the viral genome and it seems logical that any approach to block this phase should be an efﬁcient way to protect plants. Two approaches have been considered to block a virus infection, the antisense and the sense approaches. Antisense approach of the replication strategy This strategy attempts to block the replication of virus by hybridization of complementary sequences to the replicase viral gene or to sequences recognized by the replicase during replication. This strategy is at a preliminary stage of investigation but is presented because of promising in-vitro results and the potential that it may be applicable when other approaches fail. The complete antisense sequence of the replicase gene of the tomato golden mosaic virus (TGMV) was integrated into tobacco genome and several lines CM Fauquet, RN Beachy 71

were reported to exhibit a high level of resistance when challenged with different concentrations of TGMV (Lichtenstein, Buck 1990). This approach may be an interesting alternative but must be further tested before being considered as a useful strategy for practical purposes. The last example concerns turnip yellow mosaic virus (TYMV) where antisense sequences corresponding to the tRNA-like structure of the 3’ extremity of the TYMV RNA can strongly inhibit replicase activity (Cellier et al. 1990). Sense approach of the replication strategy Competitor RNA. A study using sense sequences comprising the tRNA-like structure of TYMV (see above) in order to compete with the similar viral sequences and thereby decrease virus replication activity has been conducted. In-vitro studies have shown such competition (Morch et al. 1987) and in-vivo experiments show some level of resistance (Cellier et al. 1991). In contrast a similar approach used with the TMV seemed not to induce any resistance (Powell et al. 1990). Subgenomic DNA. Some viruses produce subgenomic molecules during virus infection; for example several geminiviruses produce subgenomic DNA molecules of the B component. Insertion of one copy of such DNA of the African cassava mosaic virus (ACMV) into the tobacco genome reduced disease symptoms when the plants were challenged with ACMV (Stanley et al. 1990). Symptom amelioration is associated with a reduction in the level of viral DNA, including B DNA which is responsible for the symptomatology, and the resistance is specific to ACMV. Satellite RNA. Another approach taken to confer protection against viruses is to cause the expression of virus satellite (Sat) RNAs. Sat RNAs are associated with several viruses and are dependent upon a helper virus for their replication and spread in the infected plant. It has been reported that the presence of Sat RNAs in cucumber mosaic virus (CMV)-infected tobacco reduces disease symptoms (Mossop, Francki 1979). Similarly, tobacco plants infected with a mixture of tobacco ring spot virus (TobRSV) and ToBRV Sat caused amelioration of symptoms (Gerlach et al. 1986). Transgenic plants that express these satellite sequences were shown to decrease symptoms and virus replication (Gerlach et al. 1987; Harrison et al. 1987; Jacquemond et al. 1988). However, when the plants expressing CMV Sat RNA are infected with a related but different cucumovirus, there is symptom amelioration but no decrease in virus replication. Recently, this strategy has been applied to tomato (Tien et al. 1990; Tousch et al. 1990), and proven to be efficient for the reduction of symptoms both in greenhouse and field experiments (Tien et al. 1990). Not all satellite sequences provide symptom attenuation but sometimes they can also cause necrosis; the sequences responsible for severe symptoms are reduced to a few nucleotides (Devic et al. 1990; Jaegle et al., 1993). There is a risk that amplifying a satellite in transgenic plants may result in some of the molecule reverting to a necrotic form, causing dramatic symptoms when naturally infected by the helper virus. This possibility will greatly limit the utilization of the Sat RNA strategy unless further studies can demonstrate a great stability of the system. TMV replicase strategy. Lately, a new source of genetically engineered resistance has been identified involving the transformation of plants with nonstructural viral genes. 72 Status of Coat Protein-Mediated Resistance and its Potential Application for Banana Viruses

Tobacco plants transformed with the TMV 54 kDa gene, which is derived from a portion of the replicase complex, are immune to extremely high concentrations of TMV virions or µ -1 RNA (up to 500 g mL ) of the strain U1 (Golemboski et al. 1990). This immunity is highly speciﬁc to the strain U1, or a mutant YS1/1 of TMV and susceptible to the other strains, including the related U2 strain of TMV. This approach is undertaken with several viruses and preliminary results with the expression of the entire replicase gene of the potato virus X are producing a high level of resistance into transgenic tobacco plants (Braun, pers. com.) and portions of the replicase gene of CMV as well (Anderson, pers. com.).

Status of the Coat Protein Strategy

Among the different strategies for controlling viruses by genetic engineering, the coat protein strategy is currently the most promising. Many examples have been published and the efficiency in terms of protection, stability, and specificity has been evaluated for several viruses. The type of resistance and the mechanisms of action of the CPMR have been investigated and much information is available. Finally, both laboratory experiments and ﬁeld tests with different crops have been conducted and the ﬁrst commercial use of this type of resistance is scheduled for 1995.

Definition, concept, and production of coat protein-mediated resistance (CPMR) CPMR refers to the resistance to virus infection caused by expression of a coat protein (CP) gene in transgenic plants. The expression of a CP gene confers resistance to the virus from which the CP gene was derived. Resistance is associated to the expression of the CP gene, and is stably inherited to subsequent generations. The CP strategy is the expression by the plant of the viral CP gene integrated into the plant genome. Construction of the chimaeric gene should include the selection of an appropriate transcriptional promoter to cause the expression of the CP gene at sufficient levels to produce disease resistance. Several different transcriptional promoters have been used and the promoter that has proven most effective is the CaMV 35S. This promoter leads to high levels of mRNA and protein in most of the plants in which it has been tested. Furthermore, an enhanced 35S promoter, e35S, which is produced by duplication of an upstream regulatory sequence (Kay et al. 1987), causes higher levels of gene expression and is now widely used. The coding region used for the gene is obtained by deriving double-stranded DNA from the virus genome. When necessary, specific mutations of the gene may be needed to increase the translation of any mRNA following the consensus sequence rules described by others (Kozak 1988). The third part of the chimaeric genes is a sequence to confer transcript termination and polyadenylation. Little evidence has been published to date to indicate that a specific 3’ end is preferred in transgenic plants. The 3’ ends used for most chimaeric genes expressed in plants have CM Fauquet, RN Beachy 73

been taken from the T-DNA region of the Ti plasmid (Nelson et al. 1988; Powell et al. 1986; Powell et al. 1990). Assessing disease resistance involves inoculation with virus plants that express the CP gene (CP+) and those that do not (CP-). There is also increasing evidence that nontranslated messengers might interfere with virus replication and, consequently, transgenic plants expressing the CP messengers without CP production should be used as control. A precise estimation of the number of inserted copies of the chimaeric gene, the level of messenger expression, and the amount of CP produced are required for a valuable evaluation of the CPMR. A comparison in the numbers of sites of infection, the disease incidence, the development of disease symptoms, and the accumulation of virus are generally used to evaluate resistance. In order to use populations of plants that are identical in age, growth rate, and size, R1 or successive generations of plants are used. Prior to the inoculation of seedlings with virus, the segregation of the introduced gene is determined generally by an immunological reaction to detect the CP or by following the expression of a gene that is co-introduced with the CP gene.

Efﬁcacy of coat protein-mediated resistance The efﬁcacy of CPMR is demonstrated by the number of examples where resistance has been achieved, by the spectrum of specificity of protection, and also by the type of resistance achieved.

Speciﬁcity of coat protein-mediated resistance Since 1986, the date of the ﬁrst publication describing the coat-protein strategy (Powell et al. 1986), there have been a number of reports of CPMR involving a variety of different viruses and host plants (Beachy et al. 1990). A list of published and unpublished reports is presented in Table 1. It includes viruses belonging to 13 different virus groups and hosts that belong to dicotyledons and monocotyledons. Most of the examples are from nonenveloped ssRNA(+) viruses, but it seems that the morphology of the virus, the fact that the viral genome is divided or not, and the type of viral genome organization does not matter for the obtention of resistance. There are positive examples of CPMR for two groups of enveloped ssRNA(-) viruses: Tenuivirus and Tospovirus, but there is some question about the mode of action of the strategy because the CP mRNA could also act as antisense of the complementary strand of the CP. As for the DNA viruses, we have information only about geminiviruses where the CPMR has been achieved: the tomato yellow leaf curl virus from Thailand (TYLCV-Th) and the African cassava mosaic virus (ACMV); but in these cases the CP levels were extremely low and the protection limited (Rochester, pers. com.). There is only one example of the use of CPMR in monocotyledons where the CP expression is very high and the resistance to the virus evaluated by insect transmission is also very high (Hayakawa et al. 1991). It is anticipated that more examples of CPMR for different virus groups will be achieved in the very near future, which will provide substantial information about the strategy. 74 Status of Coat Protein-Mediated Resistance and its Potential Application for Banana Viruses

Spectrum of coat protein-mediated resistance The best TMV CP(+) tobacco lines were inoculated with members of different virus groups including CMV, AlMV, PVX, and PVY (see Table 1), but there was no noticeable resistance against infection by any of the tested viruses (Anderson et al. 1989). Though there is no protection for viruses belonging to other virus groups, there is growing evidence that the CPMR has a large spectrum of speciﬁcity within the group and that a particular CP can provide resistance to more than the homologous virus. A complete study on the spectrum of resistance of transgenic plants expressing a CP gene was carried out on tobamoviruses (Nejidat, Beachy 1990); CP(+) tobacco plant lines that expressed the U1 TMV CP gene were inoculated with other tobamoviruses. Based upon comparisons of amino acid sequences of CPs (Gibbs 1986), tobamoviruses have degrees of relatedness to TMV ranging from 85 to 39%. Infection by TMV, ToMV, pepper mild mosaic virus (PMMV), and tomato mild green mosaic virus (TMGMV) were inhibited by 95-98%, Odotonglossum ring spot virus (ORSV) by 80-95%, ribgrass mosaic virus (RMV), and sunn hemp mosaic virus (SHMV) by 40-60% (Nejidat, Beachy 1990). Similarly, transgenic tomato plants expressing the CP of TMGMV were found resistant to TMV and ToMV (Sturtevant et al. 1991). On the basis of these studies it was concluded that viruses that are related to the CP of TMV by more than 60% are less able to infect the resistant lines than are less related tobamoviruses. Results of experiments of CPMR in the Potyvirus group, particularly important since many economically important plant viruses belong to this virus group, demonstrated that expression of a particular CP gene can induce protection for several other potyviruses. The soybean mosaic virus (SMV) CP protected tobacco plants from infection by tobacco etch virus (TEV) and potato virus Y (PVY) (Stark, Beachy 1989). Transgenic tobacco plants expressing the CP of papaya ringspot virus (PRSV) were found resistant to PVY, PeMV and TEV (Ling et al. 1991). The CP gene sequences of these potyviruses are homologous to the level of 65%. In the case of tobraviruses the heterologous protection is also effective for viruses having about 60% homology (van Dun, Bol 1988). For cucumoviruses it has been proven for several cases that the CPMR is extended to several strains among and across the two subgroups of the Cucumovirus group (see below). Multiple manifestations of coat protein-mediated resistance Resistance to inoculation. In each of the examples of CPMR described to date, resistance is manifested by several features. First, there is a reduction in the numbers of sites where infection occurs on inoculated leaves. Fewer starch lesions were produced after inoculation with PVX on CP(+) tobacco plants than on CP(-) plants (Hemenway et al. 1988), and there are fewer chlorotic lesions caused by TMV infection on tobacco plants that express the TMV CP gene than on those that did not (Powell et al. 1986). Likewise, the numbers of necrotic local lesions caused by TMV infection on CP(+) Xanthi nc tobacco local lesion were 95-98% lower than on CP(-) plants (Nelson et al. 1987). These experiments indicate that the expression of a CP gene causes a reduction in the number of sites where infection is established upon inoculation. CM Fauquet, RN Beachy 75

Table 1. Examples of coat protein-mediated resistance in transgenic plants. Virus group CP gene Transgenic plant Virus resistance Reference Tobamovirus TMV tobacco TMV (Powell et al. 1986) "" ToMV (Nelson et al. 1988) "" PMMV (Nejidat, Beachy 1990). "" TMGMV " "" HRSV " "" ORSV " " tomato TMV (Nelson et al. 1988) "" ToMV " ToMV tomato ToMV (Sanders et al. 1990) TMGMV tobacco TMGMV (Sturtevant et al. 1991) " tomato TMV " "" ToMV " Tobravirus TRV tobacco TRV (van Dun, Bol 1988) "" PEBV " Carlavirus PVM potato PVM (Wefels et al. 1990) PVS tobacco PVS (MacKenzie, Tremaine 1990) potato PVS (MacKenzie et al. 1991) Potexvirus PVX tobacco PVX (Hemenway et al. 1988) " potato PVX (Lawson et al. 1990) CCMV tobacco CCMV (Fauquet et al. 1991) Potyvirus PVY potato PVY (Lawson et al. 1990) SMV tobacco PVY (Stark, Beachy 1989) "" TEV " ZYMV tobacco ZYMV (Namba et al. 1990) WMV II tobacco WMV II " PRSV tobacco PVY (Ling et al. 1991) "" TEV " "" PeMV " Furovirus BNYVV beet (protoplast) BNYVV (Kallerhoff et al. 1990) AlMV group AlMV tobacco AlMV (Loesch-Fries et al. 1987) "" "(Tumer et al. 1987) "" "(van Dun et al. 1987) " tomato " (Tumer et al. 1987)

Cucumovirus CMV-D tobacco CMV-D (Cuozzo et al. 1988) "" "(Nakayama et al. 1990) " tomato CMV-D (Cuozzo et al. 1988) CMV-C tobacco CMV-C (Quemada et al. 1991) CMV-Chi " CMV-WL " CMV-WL tobacco CMV-WL (Namba et al. 1991) CMV-C " CMV-Chi " Ilarvirus TSV tobacco TSV (van Dun et al. 1988) Luteovirus PLRV potato PLRV (Tumer et al. 1990) "" "(Kawchuk et al. 1990) Tenuivirus RSV rice RSV (Hayakawa et al. 1991) Tospovirus TSWV tobacco TSWV (MacKenzie, Ellis 1992) "" "(Gielen et al. 1991) Abbreviations: AlMV, alfalfa mosaic virus; BNYVV, beet necrotic yellow vein virus; CCMV, cassava common mosaic virus; CMV, cucumber mosaic virus; ORSV, Odotonglossum ringspot virus; PEBV, pea early browning virus; PLRV, potato leafroll virus; PMMV, pepper mild mosaic virus; PRSV, papaya ringspot virus; PVM, potato virus M; PVS, potato virus S;.PVX, potato virus X; PVY, potato virus Y; RSV, rice stripe virus; SMV, soybean mosaic virus; TEV, tobacco etch virus; TMGMV, tobacco mild green mosaic virus; TMV, tobacco mosaic virus; ToMV, tomato mosaic virus; TRV, tobacco rattle virus; TSV, tobacco streak virus; TSWV, tomato spotted wilt virus; WMV II, water melon mosaic virus II; ZYMV, zucchini yellow mosaic virus. 76 Status of Coat Protein-Mediated Resistance and its Potential Application for Banana Viruses

Resistance to virus spread within the plant. The second manifestation of resistance of CP-engineered plants is a reduced rate of systemic disease development throughout the CP(+) plants. Thus, if inoculation results in infection on the inoculated leaves, the likelihood that the infection will become systemic is considerably lower in CP(+) plants than in CP(-) plants. Grafting studies in which a stem segment of a transgenic TMV (CP+) tobacco plant was inserted between the rootstock and apex of a wild-type tobacco, demonstrated that the (CP+) segment prevented the virus from moving to the upper part of the grafted plant. This experiment shows that the CP may play a role in the long- distance movement of the virus and, consequently, resistance has a component that affects systemic spread of the infection, at least in the TMV-tobacco system (Wisniewski et al. 1990). Resistance to symptom expression. A third manifestation of resistance is a reduced rate of disease development on systemic hosts that are CP(+). In most of the examples of CPMR, CP(+) plant lines were less likely to develop systemic disease symptoms than those that were CP(-). Several plant lines that expressed the PVX CP gene did not become severely infected when inoculated with high levels of virus (Hemenway et al. 1988). Similar results were reported for CPMR against CMV (Cuozzo et al. 1988), TMV (Powell et al. 1986), PVY and TEV (Stark, Beachy 1989), and other viruses. On the contrary, in other cases the transgenic plants that are becoming infected showed similar symptoms to those of the control plants as, for example, the CMV in transgenic tobacco plants (Quemada et al. 1991) and the cassava common mosaic virus (CCMV) in transgenic Nicotiana benthamiana (Fauquet et al. 1991). In one transgenic line of tobacco with CMV the plants were asymptomatic but showed a normal level of virus replication (Quemada et al. 1991), suggesting that the CP might play a direct role in the suppression of symptoms in addition or not to suppression of virus replication. Resistance to virus multiplication. Another manifestation of resistance is lower accumulation of virus in CP(+) compared with CP(-) plant lines. Elisa and semi- quantitative western blots have been used to quantify virus accumulation in inoculated leaves and other plant parts in most examples of CPMR (Cuozzo et al. 1988; Hemenway et al. 1988; Lawson et al. 1990; Nelson et al. 1987; Powell et al. 1986). In certain examples of resistance, plants accumulate no virus after inoculation (Hemenway et al. 1988; Lawson et al. 1990), and can therefore be considered to be immune to infection under the conditions of the tests. In other cases, the virus accumulation in the infected transgenic plants is normal and comparable to control plants (Fauquet et al. 1991; Quemada et al. 1991). All resistance manifestations of CPMR can usually, but not always, be overcome by inoculating with relatively high concentrations of virus. A virus concentration of 10 µg mL-1 of TMV nearly breaks the CPMR to TMV in a system where 0.01 µg mL-1 causes disease in CP(-) plants (Powell et al. 1986). Fifty µg mL-1 are needed to overcome CP resistance to AlMV, PVX, PVY and TEV (see Table 1) (Hemenway et al. 1988; Lawson et al. 1990; Stark, Beachy 1989; Tumer et al. 1987) and 100 µg mL-1 of CCMV are required to break the CPMR in tobacco plants (Fauquet et al. 1991). Resistance is largely overcome by inoculation with RNA rather than virions in many cases except the PVX CP(+) lines of CM Fauquet, RN Beachy 77

tobacco and potato (Hemenway et al. 1988; Lawson et al. 1990), and the CCMV lines of tobacco (Fauquet et al. 1991); this feature, shared by two members of the potexvirus group may reflect a property specific to viruses of this group. The majority of evaluation of CPMR has been assessed by mechanical inoculation of the virus and, of course, the most important criterion for virus resistance is its evaluation in natural modes of contamination, i.e., in vegetative propagation and with the natural vectors. Information related to these points is limited but significant. In the case of the dually engineered resistance against PVX and PVY in potatoes (Lawson et al. 1990), it has not been possible to recover in the potato tubers any of these two viruses. At least one line of potato showed a good level of resistance through aphid inoculation of PVY. PLRV, a member of the Luteovirus group, is nonmechanically transmissible and the CP- engineered potato plants have all been challenged by using aphid inoculation (Kawchuk et al. 1990; Tumer et al., 1991), and demonstrated some degree of resistance. CMV CP(+) transgenic tobacco plants have been challenged with viruliferous aphids and proven to be resistant as for mechanical inoculations (Quemada et al. 1991). Stability of coat protein-mediated resistance CPMR is, in most cases, monogenic and is inherited by the following generations, like any other genetic trait. Consequently the genetic stability of this resistance gene is expected to be the same as any other gene. The biological stability of such resistance can be questioned, but no answer can be provided until it is used in natural conditions. One may argue that a single-point mutation can change a vital amino acid in the expressed coat protein and consequently alter the resistance, but the probability will not be greater than for any other monogenic resistance gene. Furthermore, we know that CPMR is effective for viruses differing in up to 40% in their CP sequences. In order to test the stability of the system, TMV CP(+) lines of tomato have been inoculated successively 15 times with the same isolate of virus, with the idea of selecting TMV molecules able to overcome the protection; however, no resistance breaking strain has yet been identified (White and Beachy, pers. com.). Field experiments with coat protein-mediated resistant plants There have been several field tests of virus-engineered resistant plants. Tomato plants expressing the TMV and ToMV CP genes have been tested in the field for several years, and potato plants expressing the PVX and PVY CP genes have also been tested in the field. Tobacco plants that expressed the CP gene of AlMV were field-tested in Wisconsin in 1988 (Krahn, pers. com.). CP(+) plants developed disease more slowly or not at all compared with CP(-) plants; symptom development was correlated with virus accumulation. At 85 days after inoculation only 9% of the CP(+) plants had developed a systemic infection, while 93% of the CP(-) plants had systemic infections. The first field test with TMV-resistant tomato plants, from cultivar VF36, was conducted in 1987 (Nelson et al. 1988). CP(+) plants, mechanically inoculated with TMV, exhibited a delay in the development of disease symptoms or did not develop symptoms compared with the CP(-) plants. No more than 5% of the CP(+) plants developed disease 78 Status of Coat Protein-Mediated Resistance and its Potential Application for Banana Viruses

symptoms by fruit harvest, while 99% of the VF36 plants developed symptoms. Lack of visual symptoms was associated with a lack of virus accumulation. Fruit yields of the infected VF36 plants decreased 26-35% compared with healthy plants, whereas yields from the CP(+) line were equal to those of uninoculated VF36 plants. To determine if the TMV CP gene conferred protection against infection by field isolates of ToMV, tests were conducted in 1988 in Florida and Illinois. Progeny that were homozygous for the TMV CP gene and VF36 CP(-) plants were challenged with a Florida isolate of ToMV, Naples C, in Florida. The field test in Illinois was conducted to determine if expression of the TMV CP gene in tomato would protect it against a number of different strains of TMV and ToMV. The TMV CP gene conferred resistance against ToMV-Naples C infection under Florida field conditions and against two strains of TMV under Illinois field conditions. Only weak protection was conferred against infection by the ToMV strains under Illinois field conditions (Sanders et al. 1990). To enhance protection against ToMV, plants were produced that expressed a CP gene derived from ToMV-Naples C. These lines were evaluated under field conditions in Illinois along with the control tomato line UC82B, with lines expressing the TMV CP gene, and with lines expressing both TMV and ToMV CP genes. The TMV CP(+) lines were resistant to TMV infection, as shown in the earlier field test; however, they were less resistant to infection by ToMV-Naples C. The tomato lines expressing ToMV CP gene were highly resistant to infection by ToMV-Naples C. Plants that expressed both TMV and ToMV CP genes were equally well protected against TMV and ToMV. Field tests were conducted with Russet Burbank potato plants expressing the CP genes of PVX and PVY (Lawson et al. 1990). PVY causes significant yield depression in potato and, in combination with PVX, PVY produces a severe disease called “rugose mosaic”. To determine if expression of PVX and PVY CP genes would protect potato plants from the synergistic effects of PVX and PVY infection in the field, plants propagated from CP(-) Russet Burbank and from plant lines expressing both CP genes were inoculated with both PVX and PVY and transplanted into the field (Kaniewski et al. 1990). Plants from four CP(+) lines were significantly protected from infection by PVX. However, three of the lines were not protected from infection by PVY when simultaneously inoculated by both viruses. Plants of one line, however, were highly resistant to PVX and PVY, as predicted from growth-chamber tests. Tuber yields at maturity in uninoculated plots were the same for all the lines. In contrast, tuber yields of all inoculated lines were markedly reduced, except the line resistant to both viruses, which was unaffected by virus inoculation. For the last 3 years, field trials of cucurbits transformed with the CMV CP gene have been conducted with success (Gonzalves et al. 1991). In 1991, four transgenic lines of the cv Poinsett 76 were compared with resistant cv Marketmore 76; infection was allowed to occur naturally by aphids using a low percentage of infected plants as initial virus sources. After 4 months of growth, the transgenic plants performed much better than the control plants; 8% of transgenic plants showed symptoms versus 98% for the control and 8% for the resistant line. CM Fauquet, RN Beachy 79

Potential Use of Coat Protein-Mediated Resistance for Banana Viruses The choice of a particular strategy for controlling plant viruses by genetic engineering depends greatly on the types of virus chosen as a target; but, because of the apparent universality of the efﬁcacy of CPMR and because of the strength of the resistance, coat protein strategy should be ﬁrst attempted.

Viruses infecting bananas Viruses infecting banana are limited in number and still insufficiently described; however, the following have been identified : Banana bunchy top DNA-virus (Dietzgen, Thomas 1991; Harding et al. 1991; Thomas, Dietzgen 1991; Wu, Su 1990). This virus has been isolated only in Australia, and it is not known whether it is the causal agent of the banana bunchy top disease or a helper of the banana bunchy top luteovirus. This virus with a genome of multipartite single-stranded DNA most probably belongs to a new group of viruses. The genome comprises a large number (seven or more) of monogenic components. It is somewhat similar to the geminiviruses for which CPMR so far has not been extremely effective and, if the CP strategy was poorly efficient with this virus, another strategy, like the sense or antisense strategy for the replicase gene, should be considered. Banana bunchy top luteovirus (Brunt et al. 1990). Luteoviruses are single-stranded RNA+ viruses with a monopartite genome. With the example of the PLRV, we know that CPMR is an efficient strategy to create resistance by biotechnology, provided that the CP is engineered and the gene expression increased. Resistance to the virus through insect- vector inoculation has been demonstrated in potato and tobacco, and should therefore be effective in banana with appropriate chimaeric gene constructs. If banana bunchy top disease results from the effect of two viruses, the usefulness of CPMR would have to be investigated using one or the other CP or a combination of the two CPs. Cucumber mosaic cucumovirus (Brunt et al. 1990). This virus is one of the most commonly used for CPMR and we know that the CP strategy is extremely effective with a wide spectrum of protection (see below). In any project of banana transformation to control viruses through biotechnology, the CMV CP gene should be used because it is an important viral disease of banana and because it can be an internal marker for other viral genes. We know that if CPMR cannot be achieved with CMV, the unique banana environment should be taken into account to improve gene expression and efficacy of CPMR. Banana bract mosaic potyvirus (Dale, pers. com.). This virus has been recently discovered in the Philippines where it is causing a lot of damage. There are many examples of CPMR for such viruses and the strategy could be applied directly to banana with the appropriate cloning of the CP gene for this potyvirus. There are also indications that different potyviruses in other places in the world infect banana: proper identification is needed, but, with the wide spectrum of protection effected by CPMR, an overall protection against banana potyviruses using one CP could be envisaged. 80 Status of Coat Protein-Mediated Resistance and its Potential Application for Banana Viruses

Banana streak badnavirus (Lockhart 1990). Badnaviruses are double-stranded DNA viruses for which we have not achieved any examples of CPMR. Studies with a virus of the same group are under way with the rice tungro bacilliform virus, and should provide valuable information to extrapolate to the banana streak virus. This virus does not seem to be very widespread in the world and, therefore, is not an important economic target.

Summary of coat protein-mediated resistance for PLRV The CPMR for PLRV has been achieved by two different groups which obtained similar results (Kawchuk et al. 1991; Tumer et al. 1991). The CP of the PLRV is expressed at a very low level, at the limit of the detection (0.01%), but resistance to the virus replication was nevertheless effective. The transgenic plants were resistant to inoculation by the natural aphid vector, even in the case where large numbers of insects were used. There was no relation between the amount of CP detected in the transgenic lines and the level of resistance estimated in the same lines. Field trials have been successfully conducted, but the spectrum of speciﬁcity remains to be established. The CP expression has been improved by sequence protein engineering, removing the internal ORF, resynthesizing the CP coding sequence, and by expressing two chimaeric genes in the same transgenic plant. In these conditions up to 9% of the regenerated lines were highly resistant or immune (Cuozzo et al. 1988; Tumer et al. 1991). Furthermore, it has been proven that the CP antisense strategy is also effective against PLRV, providing an alternative to the CPMR (Kawchuk et al. 1991).

Summary of coat protein-mediated resistance for CMV CPMR for CMV has been achieved by three different groups which obtained similar results in different plants (Ye et al. 1991; Cuozzo et al. 1988; Beachy, pers. com.). The CP of CMV was expressed at a very high level (0.6%), and resistance to the virus multiplication was effective. The transgenic plants were resistant to mechanical inoculation and/or inoculation with the natural aphid vector. There was no direct relation between the amount of CP detected in the transgenic lines and the level of resistance estimated in the same lines, though the highest expressors were among the most resistant lines, and no line was highly resistant without any CMV CP. Field trials of transgenic cucumbers have been conducted successfully for 3 years, where the best transgenic lines had less than 10% of contaminated plants after 4 months of survey. Studies have been conducted to evaluate the spectrum of speciﬁcity of the resistance, and it was shown that the plants were resistant to strains within one subgroup of CMV strains and across the two subgroups of the cucumovirus group.

Conclusion The last few years have provided a profusion of techniques to control viruses by genetic engineering. Most have demonstrated potential for conferring resistance to plants and constitute future hopes for crop improvement. Several techniques have already been CM Fauquet, RN Beachy 81

applied in field experiments and have proven to be real sources of resistance to viruses, e.g., the satellite and the coat protein strategies. The ribozyme strategy is interesting because of its potential applicability to all viruses, and because of its high specificity, but it needs to be demonstrated under in-vivo conditions. The satellite strategy, though very effective, is strictly limited to viruses having satellites, which are very few in the plant virus world, and its use is limited by the risk that mutations would accompany widespread use. CPMR for controlling viruses is the safest, most efficient, and widely documented type of engineered resistance. Several field experiments have been conducted with different crops and CPMR was shown effective in all cases. The spectrum of specificity of CPMR is broader than in any other type of virus resistance (including natural resistance), because plants can be protected against viruses of the same group sharing up to 60% in their coat protein sequence. Coat protein transgenic plants are resistant to high concentrations of virus inoculum, and even in some cases to RNA inoculum. Resistance against vector inoculations and vegetative propagation have also been demonstrated. The resistance generated by the coat protein strategy is of a multiple type, reducing the number of infection sites, the symptoms, the virus multiplication and the long-distance spread through plants. It seems that the banana is an excellent candidate for using biotechnology to control viruses. And because there is no natural source of resistance for any banana virus, because it is a vegetatively propagated crop, and because CPMR has been used to describe most of the viruses of banana described to date, CPMR is applicable and should be useful. All the manifestations of CPMR can be the result of a single mechanism or of multiple effects of the coat protein on different targets. Nevertheless, despite the number of studies conducted to understand the mechanisms of the action of the coat protein strategy, the question of how the coat protein confers resistance to engineered plants remains unanswered. Acknowledgments This publication has been made possible with the participation of ORSTOM and TSRI.

References

ANDERSON EJ, STARK DM, NELSON RS, POWELL PA, TUMER NE, BEACHY RN. 1989. Transgenic plants that express the coat protein genes of tobacco mosaic virus or alfalfa mosaic virus interfere with disease development of some non-related viruses. Phytopathology 79:1284-1290. BEACHY RN, LOESCH-FRIES S, TUMER NE. 1990. Coat protein-mediated resistance against virus infection. Annual Review of Phytopathology 28:451-474. BRUNT A, CRABTREE K, GIBBS A. 1990. Viruses of Tropical Plants. Wallingford, Oxon, UK: CAB International. 389 pp. BUZAYAN JM, GERLACH WL, BREUNING G. 1986. Satellite tobacco ringspot virus RNA: a subset of the RNA sequence is sufﬁcient for autolytic processing. Proceedings of the National Academy of Sciences (USA) 83:8859-8862. CECH TM. 1986. The chemistry of self-splicing RNA and RNA enzymes. Science 236:1532-1539. CELLIER F, ZACCOMER B, BOYER JC, MORCH MD, HAENNI AL, TEPFER M. 1991. The “sense” strategy confers protection against turnip yellow mosaic virus (TYMV) in transgenic Brassica plants. Third International Congress of Plant Molecular Biology, 6-11 October 1991, Tucson, Arizona, USA. 82 Status of Coat Protein-Mediated Resistance and its Potential Application for Banana Viruses

CELLIER F, ZACCOMER B, MORCH MD, HAENNI AL, TEPFER M. 1990. Strategies for interfering in vivo with replication of turnip yellow mosaic virus. Proceedings of the VIIIth International Congress of Virology, 26-30 August 1990, Berlin, Germany. CUOZZO M, O’CONNELL KM, KANIEWSKI W, FANG RX, CHUA NH, TUMER NE. 1988. Viral protection in transgenic plants expressing the cucumber mosaic virus coat protein or its antisense RNA. Bio/Technology 6: 549-557. DEVIC M, JAEGLE M, BAULCOMBE D. 1990. Cucumber mosaic virus satellite RNA (Y strain): analysis of sequences which affect systemic necrosis on tomato. Journal of General Virology 90: 1443-1449. DIETZGEN RG, THOMAS JE. 1991. Properties of virus-like particles associated with banana bunchy top disease in Hawaii, Indonesia and Tonga. Australian Plant Pathology 20:161-165. ECKER JR, DAVIS RW. 1986. Inhibition of gene expression in plant cells by expression of antisense RNA. Proceedings of the National Academy of Sciences (USA) 83:5372-5376. FAUQUET CM, BOGUSZ D, FRANCHE C, SCHOPKE C, CHAVARRIAGA P, CALVERT L, BEACHY, R. 1991. Evaluation of the coat protein mediated resistance to cassava common mosaic virus (CCMV) in Nicotiana benthamiana. Third International Congress of Plant Molecular Biology, 6-11 October 1991, Tucson, Arizona, USA. FORSTER AC, SYMONS RH. 1987. Self-cleavage of plus and minus RNAs of a virusoid and a structural model for the active sites. Cell 49:211-220. GERLACH WL. 1989. Use of plant virus satellite RNA sequences to control gene expression. Viral Genes and Plant Pathogenesis, 15-17 Oct 1989, Lexington, Kentucky, USA. GERLACH WL, BUZAYAN JM, SCHEIDER IR, BREUNING G. 1986. Satellite tobacco ringspot virus RNA: biological activity of DNA clones and their in vitro transcripts. Virology 151:172-185. GERLACH WL, LLEWELLYN D, HASELOFF, J. 1987. Construction of a plant disease resistance gene from the satellite RNA of tobacco ringspot virus. Nature 328:802-805. GIBBS A. 1986. Tobamovirus classification. Pages 168-180 in The Plant Viruses, vol.2: The Rod-shaped Plant Viruses. New York, Plenum Press. GIELEN JJL, DE HAAN P, KOOL AJ, PETERS D, VAN GRINSVEN MQJM, GOLDBACH RW. 1991. Engineered resistance to tomato spotted wilt virus, a negative-strand RNA virus. Bio/Technology 9:1363-1367. GOLEMBOSKI DB, LOMONOSSOFF GP, ZAITLIN M. 1990. Plants transformed with a tobacco mosaic virus nonstructural gene sequence are resistant to the virus. Proceedings of the National Academy of Sciences (USA) 87:6311- 6315. GONZALVES CV, CHEE P, SLIGHTOM JL, PROVVIDENTI R. 1991. Field evaluation of transgenic cucumber plants expressing the coat protein gene of cucumber mosaic virus. Abstracts of presentations, APS annual meeting, 17-21 August 1991, St Louis, Missouri, USA. GREEN PJ, PINES O, INOUYE M. 1986. The role of antisense RNA in gene regulation. Annual Review of Biochemistry 55:569-597. HARDING RM, BURNS TM, DALE JL. 1991. Virus like particles associated with banana bunchy top disease contain single-stranded DNA. Journal of General Virology 72:225-230. HARRISON BD, MAYO MA, BAULCOMBE DC. 1987. Virus resistance in transgenic plants that express cucumber mosaic virus satellite RNA. Nature 334:799-802. HASELOFF J, GERLACH WL. 1988. Simple RNA enzymes with new and highly speciﬁc endoribonuclease activities. Nature 334:585-591. HAYAKAWA T, ZHU Y, ITOH K, KIMURA Y, IZAWA T, SHIMAMOTO K, TORIYAMA S. 1991. Genetically engineered rice resistant to rice stripe virus, an insect transmitted virus. Third International Congress of Plant Molecular Biology, 6-11 October 1991, Tucson, Arizona, USA. HEMENWAY C, FANG RX, KANIEWSKI WK, CHUA NH, TUMER NE. 1988. Analysis of the mechanism of protection in transgenic plants expressing the potato virus X coat protein or its antisense RNA. The EMBO Journal 7:1273- 1280. HUTCHINS CJ, RATHJEN PD, FORSTER AC, SYMONS RH. 1986. Self cleavage of plus and minus RNA transcripts of avocado sunblotch viroid. Nucleic Acids Research 14:3627-3640. JACQUEMOND M, AMSELEM A, TEPFER M. 1988. A gene coding for a monomeric form of cucumber mosaic virus satellite RNA confers tolerance to CMV. Molecular Plant Microbe Interactions 1:311-316. JAEGLE M, DEVIC M, LONGSTAFF M, BAULCOMBE D. 1993. Cucumber mosaic virus satellite RNA (Y strain): analysis of sequences which affect mosaic symptoms on tobacco. Journal of General Virology (in press). CM Fauquet, RN Beachy 83

KALLERHOFF J, PEREZ P, GÉRENTES D, PONCETTA C, BEN TAHAR S, PERRET J. 1990. Sugarbeet transformation for resistance to rhizomania. Proceedings of the VIIIth International Congress of Virology, 26-30 August 1990, Berlin, Germany. KANIEWSKI WK, LAWSON C, SAMMONS B, HALEY L, HART J, DELANNAY X, TUMER NE. 1990. Field resistance of transgenic Russet Burbank potato to effects of infection by potato virus X and potato virus Y. Bio/Technology 8:750-754. KAWCHUK LM, MARTIN RR, MCPHERSON J. 1990. Resistance in trangenic potato expressing the potato leafroll virus coat protein gene. Molecular Plant Microbe Interactions 3:301-307. KAWCHUK LM, MARTIN RR, MCPHERSON J. 1991. Sense and antisense RNA-mediated resistance to potato leafroll virus in Russet Burbank potato plants. Molecular Plant Microbe Interactions 4:247-253. KAY R, CHAN A, DALY M, MCPHERSON J. 1987. Duplication of CaMV 35S promoter sequences creates a strong enhancer for plant genes. Science 236:1299-1302. KOZAK M. 1988. Leader length and secondary structure modulate mRNA function under condition of stress. Molecular Cell Biology 6:2737-2744. LAWSON C, KANIEWSKI WK, HALEY L, ROZMAN R, NEWELL C, SANDERS PR, TUMER NE. 1990. Engineering resistance to mixed virus infection in a commercial potato cultivar: resistance to potato virus X and potato virus Y in transgenic Russet Burbank. Bio/Technology 8:127-134. LICHTENSTEIN CL, BUCK KW. 1990. Expression of antisense RNA in transgenic tobacco plants confers resistance to geminivirus infection. Molecular Strategies for Crop Improvement, 17-24 April 1990, Keystone, Colorado, USA. LING K, NAMBA S, GONZALVES C, SLIGHTOM JL, GONZALVES D. 1991. Protection against detrimental effects of potyvirus infection in transgenic tobacco plants expressing the papaya ringspot virus coat protein gene. Bio/Technology 9:752-758. LOCKHART BEL. 1990. Evidence for a double stranded circular DNA genome in a second group of plant viruses. Phytopathology 80:127-131. LOESCH-FRIES LS, MERLO D, ZINNEN T, BURHOP L, HILL K, KRAHN K, JARVIS N, NELSON S, HALK E. 1987. Expression of alfalfa mosaic virus RNA4 in transgenic plants confers virus resistance. The EMBO Journal 6:1845-1851. MACKENZIE DJ, ELLIS PJ. 1992. Resistance to tomato spotted wilt virus infection in transgenic tobacco expressing the viral nucleocapsid gene. Molecular Plant Microbe Interactions 5:34-40. MACKENZIE DJ, TREMAINE JH. 1990. Transgenic Nicotiana debneyii coat protein are resistant to potato virus S infection. Journal of General Virology 71:2167-2170. MACKENZIE DJ, TREMAINE JH, MCPHERSON J. 1991. Genetically engineered resistance to potato virus S in potato cultivar Russet Burbank. Molecular Plant Microbe Interactions 4:95-102. MORCH MD, JOSHI RL, DENIALT TM, HAENNI AL. 1987. A new sense RNA approach to block viral RNA replication in vitro. Nucleic Acids Research 15:4123-4129. MOSSOP DW, FRANCKI RIB. 1979. Comparative studies on two satellite RNAs of cucumber mosaic virus. Virology 95:395-404. NAKAYAMA M, YOSHIDA T, OKUNO T, FURUSAWA I. 1990. Protection against cucumber mosaic virus and its RNA infection in transgenic tobacco plants expressing coat protein and antisense RNA of the virus. Proceedings of the VIIIth International Congress of Virology, 26-30 August 1990, Berlin, Germany. NAMBA S, LING K, GONZALVES C, SLIGHTOM JL. 1991. Expression of the gene encoding the coat protein of cucumber mosaic virus (CMV) strain-WL, appears to provide protection to tobacco plants against infection by several different CMV strains. Gene (in press). NAMBA S, LING K, GONZALVES C, SLIGHTOM JL, GONZALVES D. 1990. Comparative expression of coat protein genes of PRV, ZYMV and WMW II in transgenic tobacco plants. Proceedings of the VIIIth International Congress of Virology, 26-30 August 1990, Berlin, Germany. NEJIDAT A, BEACHY RN. 1990. Transgenic tobacco plants expressing a tobacco mosaic coat protein gene are resistant to some tobamoviruses. Molecular Plant Microbe Interactions 3:247-251. NELSON RS, MCCORMICK SM, DELANNAY X, DUBÉ P, LAYTON J, ANDERSON EJ, KANIEWSKA M, PROKSCH RK, HORSCH RB, ROGERS SG, FRALEY RT, BEACHY RN. 1988. Virus tolerance, plant growth, and ﬁeld performance of transgenic tomato plants expressing coat protein from tobacco mosaic virus. Bio/Technology 6:403-409. NELSON RS, POWELL AP, BEACHY RN. 1987. Lesions and virus accumulation in inoculated transgenic tobacco plants expressing the coat protein gene of tobacco mosaic virus. Virology 158:126-132. POWELL AP, NELSON RS, DE B, HOFFMANN N, ROGERS SG, FRALEY RT, BEACHY RN. 1986. Delay of disease development in transgenic plants that express the tobacco mosaic virus coat protein gene. Science 232:738-743. 84 Status of Coat Protein-Mediated Resistance and its Potential Application for Banana Viruses

POWELL PA, SANDERS PR, TUMER NE, BEACHY RN. 1990. Protection against tobacco mosaic virus infection in transgenic plants requires accumulation of capsid protein rather than coat protein RNA sequences. Virology 175:124-130. POWELL PA, STARK DM, SANDERS PR, BEACHY RN. 1989. Protection against tobacco mosaic virus in transgenic plants that express TMV antisense RNA. Proceedings of the National Academy of Sciences (USA) 86:6949-6952. PRODY GA, BAKOS JT, BUZAYAN JM, SCHNEIDER IR, BREUNING G. 1986. Autolytic processing of dimeric plant virus satellite RNA. Science 231:1577-1580. QUEMADA H, GONSALVES D, SLIGHTOM JL. 1991. Expression of coat protein gene cucumber mosaic virus strain-C in tobacco: protection against infections by CMV strains transmitted mechanically or by aphids. Phytopathology 81:794-802. SANDERS PR, SAMMONS B, KANIEWSKI W, HALEY L, LAYTON J, LAVALLÉE BJ, DELANNAY X, TUMER NE. 1990. Field resistance of transgenic tomatoes expressing the tobacco mosaic virus or tomato mosaic virus coat protein genes. Phytopathology 82:683-690. STANLEY J, FRISCHMUTH T, ELLWOOD S. 1990. Defective viral DNA ameliorates symptoms of geminivirus infection in transgenic plants. Proceedings of the National Academy of Sciences (USA) 87:6291-6295. STARK DM, BEACHY RN. 1989. Protection against potyvirus infection in transgenic plants: evidence for broad spectrum resistance. Bio/Technology 7:1257-1262. STURTEVANT AP, NEJIDAT A, BEACHY RN. 1991. Transgenic plants expressing the coat protein of tobacco mild green mosaic virus are resistant to viral infection. Third International Congress of Plant Molecular Biology, 6-11 October 1991, Tucson, Arizona, USA. THOMAS JE, DIETZGEN RG. 1991. Purification, characterisation and serological detection of virus-like particles associated with banana bunchy top disease in Australia. Journal of General Virology 72:217-224. TIEN P, ZHAO SZ, WANG X, WANG GJ, ZHANG CX, WU SX. 1990. Virus resistance in transgenic plants that express the monomeric gene of cucumber mosaic virus (CMV) satellite RNA in greenhouse and ﬁeld. Proceedings of the VIIIth International Congress of Virology, 26-30 August 1990, Berlin, Germany. TOUSCH D, JACQUEMOND M, TEPFER M. 1990. Behaviour towards cucumber mosaic virus of transgenic tomato plants expressing CMV-satellite RNA genes. Proceedings of the VIIIth International Congress of Virology, 26-30 August 1990, Berlin, Germany. TUMER NE, KANIESWKI WK, LAWSON C, HALEY L, THOMAS PE. 1990. Engineering resistance to potato leaf roll virus in transgenic Russet Burbank potato. Proceedings of the VIIIth International Congress of Virology, 26-30 August 1990, Berlin, Germany. TUMER NE, LAWSON C, HEMENWAY C, WEISS J, KANIESWKI WK, NIDA D, ANDERSON J, SAMMONS B. 1991. Engineering resistance to potato leafroll virus in Russet Burbank potato. Third International Congress of Plant Molecular Biology, 6-11 October 1991, Tucson, Arizona, USA. TUMER NE, O’CONNELL KM, NELSON RS, SANDERS PR, BEACHY RN. 1987. Expression of alfalfa mosaic virus coat protein gene confers cross-protection in transgenic tobacco and tomato plants. The EMBO Journal 6:1181- 1188. VAN DUN CMP, BOL JF. 1988. Transgenic tobacco plants accumulating tobacco rattle virus coat protein resist infection with tobacco rattle virus and pea early browning virus. Virology 167:649-652. VAN DUN CMP, BOL JF, VAN VLOTEN-DOTING L. 1987. Expression of alfalfa mosaic virus and tobacco rattle virus coat protein genes in transgenic tobacco plants. Virology 159:299-305. VAN DUN CMP, OVERDUIN B, VAN VLOTEN-DOTING L, BOL JF. 1988. Transgenic tobacco expressing tobacco streak virus or mutated alfalfa mosaic virus coat protein does not cross-protect against alfalfa mosaic virus infection. Virology 164:383-389. WEFELS E, COURTPOZANIS A, SALAMINI F, ROHDE W. 1990. Analysis of transgenic potato lines expressing the coat proteins of potato virus Y (PVY) or potato virus M (PVM). Proceedings of the VIIIth International Congress of Virology, 26-30 August 1990, Berlin, Germany. WISNIEWSKI LA, POWELL PA, NELSON RS, BEACHY RN. 1990. Local and systemic spread of tobacco mosaic virus in transgenic plants. Plant Cell 2:559-567. WU RY, SU HJ. 1990. Purification and characterization of banana bunchy top virus. Journal of Phytopatholy 128:153-160. YE Y, ZHAO P, ZHAO SZ, LIU YZ, TIEN P. 1991. Transgenic tobacco plants with high resistance to cucumber mosaic virus, by expressing satellite RNA and coat protein. Third International Congress of Plant Molecular Biology, 6-11 October 1991, Tucson, Arizona, USA. J Dale, T Burns, S Oehlschlager, M Karan, R Harding 85

Banana Bunchy Top Virus: prospects for control through biotechnology

J Dale, T Burns, S Oehlschlager, M Karan, R Harding

Introduction The banana, as with most major food crops, is plagued by a number of virus diseases that can considerably reduce both the yield and economics of growing bananas. The most important of these viruses is banana bunchy top virus (BBTV). This virus was first recorded in Fiji in 1879 and, since that time, has spread to many banana-producing countries. Cucumber mosaic virus (also known as banana mosaic virus and infectious chlorosis) infects bananas to a greater or lesser extent in most, if not all, banana- producing countries; banana streak virus, a mealybug-transmitted badnavirus, was first recorded in Morocco and appears to be confined to Africa; banana bract mosaic virus is a probable potyvirus reported only from the Philippines. There are also two virus diseases of abaca (Musa textalis) reported in the Philippines: abaca bunchy top, which appears to be a strain of banana bunchy top virus and abaca mosaic, which is probably caused by sugarcane mosaic virus. Considering the importance of both bananas and BBTV, there has been relatively little research effort devoted to this virus. Consequently, there has been little information published about BBTV until recently. There are probably a number of reasons for this neglect including: (a) bananas are grown primarily in the tropics and subtropics whereas most virologists and research funds are concentrated in the temperate regions; (b) the banana is a difficult plant to work with from a virological aspect; (c) BBTV does not occur in Central or South America, the regions which supply nearly 90% of the world’s banana exports and where most international banana funding has been directed; and (d) until relatively recently, there was no international organization to coordinate banana research. There have, however, been considerable advances in our knowledge of BBTV in the last 4-5 years and, with the integration of biotechnology and BBTV research, there is now a realistic prospect of developing resistant cultivars in the foreseeable future.

Centre for Molecular Biotechnology, Queensland University of Technology (QUT), 2434 Brisbane, 4001 Queensland, Australia 86 Banana Bunchy Top Virus: prospects for control through biotechnology

Banana Bunchy Top: the disease

The importance of the disease BBTV is considered the most economically damaging of the viruses that infect bananas. It virtually destroyed the banana industry in the southern Queensland and northern New South Wales regions of Australia in the 1920s in probably the best documented outbreak of BBTV. Currently, the virus is a major limiting factor in the production of bananas in many countries including the Philippines, China, and India. There is, of course, the ever- present danger of the inadvertent introduction of BBTV into Central and South America.

Symptoms of the disease The symptoms of advanced infections in Musa spp. are quite distinctive. The leaves at the apex of the plant become narrower, dwarfed, and upright giving the top of the plant a bunched appearance. The leaves also show marginal chlorosis. Infected plants may not produce a bunch or the bunch does not emerge from the pseudostem depending on the time the plant has been infected. The ﬁrst symptoms of infection are less distinct but very characteristic. Dark green streaks develop on the petioles, midribs, and leaf veins. These streaks are of varying length and the resulting dot-dash pattern is often referred to as a “morse code” pattern. Variation in the severity of symptoms has been reported and it is possible that this is due to strain variation (H Su, pers. com.).

Geographical distribution The disease was ﬁrst recorded in Fiji in 1879 and, since this initial report, the disease has been recorded in most banana-producing regions of the world including Australia, Asia (China, Indonesia, the Philippines, Vietnam, India, Pakistan, Taiwan), Africa (Egypt, Gabon, Burundi), Sri Lanka, and a number of the Paciﬁc islands (Hawaii, Tonga, Western Samoa, Guam). There are a number of notable regions or countries where BBTV has not been recorded, particularly South and Central America, the Caribbean, Papua New Guinea, and Thailand as well as many African countries. BBTV is almost certainly continuing to be moved internationally; for instance, it seems that the virus has only recently reached Hawaii. The first report of the virus in Africa was from Egypt in 1901. There was then a considerable time before reports of the virus in central Africa (Burundi and Gabon); it appears that BBTV is now spreading in central Africa. It is somewhat puzzling as to how countries such as Thailand have avoided the virus considering the close proximity to “infected” countries and the ease of movement of germplasm in the region.

Transmission of the disease There are two reported mechanisms for the transmission of BBTV. The virus is persistently transmitted by the black banana aphid, Pentalonia nigronervosa. There is J Dale, T Burns, S Oehlschlager, M Karan, R Harding 87

no information available as to whether the virus replicates in its aphid vector. This aphid has been recorded in virtually all banana-growing countries. The virus can also be transmitted through the vegetative parts of the banana plant including suckers and the corm. It has been demonstrated that the virus can readily pass through the tissue-culture process. The virus cannot be transmitted by mechanical (sap) inoculation.

Banana Bunchy Top: the virus

The search for the virus Banana bunchy top disease was always assumed to be caused by a virus and was classified as a possible luteovirus by Matthews (1982). The rationale for such a classification was the biological properties of the disease: it is transmitted by aphids in a persistent manner but not by mechanical inoculation; the symptoms of the disease included chlorosis and stunting; and the phloem of infected plants was damaged (Dale 1987). These are all primary biological characteristics of luteoviruses and luteovirus infections. Numerous attempts were made to purify virions from BBTV-infected plants using procedures successfully adopted for luteovirus purification but without significant success. However, further evidence for the association of a luteovirus with the disease came from the reported occurrence of double-stranded RNAs in bananas infected with BBTV but not in healthy bananas (Dale et al. 1986). These dsRNAs were of similar size to those reported in luteovirus-infected plants. These observations have been confirmed by a number of other groups (B Kummert, pers. com.; H Su, pers. com.; M Iskra, pers. com.).

Small isometric virus-like particles In 1989, two groups reported the purification of virus-like particles (VLPs) from BBTV- infected plants. Iskra et al. (1989) obtained isometric particles about 28 nm in diameter, within the size range expected for a luteovirus; Su and Wu (1989) also obtained isometric particles but these were 20-22 nm in diameter, which apparently contained single- stranded RNA approximately 2.0 x 106 in molecular weight and a single protein of 21 000 molecular weight. Wu and Su (1990) classified these particles as belonging to a “small” luteovirus. Harding et al. (1991) and Thomas and Dietzgen (1991) also reported the purification of small isometric VLPs, 18-20 nm in diameter, from BBTV-infected plants. However, Harding et al. determined that these particles contained single-stranded DNA of about 1 kb. Thomas and Dietzgen confirmed this result. Both groups reported a single coat protein of about 20 000 molecular weight. Subsequently, Harding et al. cloned a portion of this ssDNA and, by using this clone as a DNA probe, demonstrated that this 1 kb DNA was present only in BBTV-infected bananas, not healthy bananas, and was transmitted from infected bananas to healthy bananas with the disease and the 18-20 nm VLPs. The DNA probe hybridized with all BBTV-infected samples including samples obtained from 88 Banana Bunchy Top Virus: prospects for control through biotechnology

Australia, Burundi, Egypt, Gabon, Indonesia, the Philippines, and Taiwan, indicating that at least one virus is common to all BBTV infections and that this virus contains ssDNA. Antisera, both polyclonal and monoclonal, has been produced by Su and Wu (1989) and Thomas and Dietzgen (1991). These antisera react with isolates of BBTV from many different regions conﬁrming that a small isometric virus is intimately associated with the disease. The properties of the particles of BBTV reported by Harding et al. and Thomas and Dietzgen preclude these particles as being virions of a luteovirus because luteoviruses have virions of about 25 nm containing ssRNA and a coat protein subunit with a molecular weight of about 24 000. The properties are, however, most similar to the properties of the VLPs associated with subterranean clover stunt virus (SCSV) (Chu, Helms 1988). These particles are isometric, 17-19 nm in diameter, and contain ssDNA (0.85-0.88 kb) and a single protein about 19 000 in molecular weight. SCSV is also transmitted by aphids in a persistent manner and is not transmissible by sap.

Current status of research at QUT Since the publication of the association of ssDNA with the virions of BBTV (Harding et al. 1991), our research has centered on the cloning sequencing and analysis of the BBTV genome. The original clone, pBT338, contained a 977 bp insert. The nucleotide sequence of this insert was determined. From this sequence, two adjacent oligonucleotide primers were synthesized. In a polymerase chain reaction (PCR), these primers would be extended away from each other; thus, amplification would occur only if the molecule were circular. When these primers were used in a PCR reaction with BBTV ssDNA extracted from purified virions, a 1.1 kb product was amplified. This demonstrated that the ssDNA from BBTV virions is circular. The PCR product, which should represent the complete sequence of this circle, was cloned into a “T-tailed” vector and the nucleotide sequence determined. Analysis of this sequence revealed the following. •A sequence of 1111 bases. •A strong possible stem/loop structure which could represent the origin of replication. The stem consisted of 11 base pairs. The loop, 10 nucleotides, contained the consensus sequence common to gemini viruses and coconut foliar decay virus. • One large open reading frame of 858 nucleotides with an associated polyadenylation signal. This ORF, when translated, contained an NTP binding site usually associated with viral replicases. • An untranslated region. The PCR primers derived from the sequence of pBT338 were used to amplify this BBTV circle from a Philippine isolate of BBTV. The product was similar in size to that obtained from the Australian isolate and was subsequently cloned and sequenced. The Philippine sequence had the same features as the Australian sequence and the nucleotide sequence of the large ORF has very high sequence similarity with the Australian isolate. J Dale, T Burns, S Oehlschlager, M Karan, R Harding 89

DNA from at least three other unique circles from BBTV virions has been cloned and partially sequenced. Attempts are now being made to clone and sequence the complete circles and determine their function. Analysis of SCSV has revealed that the genome is composed of at least seven unique circles of ssDNA each containing a single large ORF. It would appear at this time that BBTV will have a very similar genome organization. Coconut foliar decay virus shares many properties with BBTV and SCSV (small isometric particles containing ssDNA). One circle of this virus has been cloned and sequenced. This sequence contains one large ORF (possible replicase) and five other smaller ORFs. No other different DNA circles have been reported from this virus. Attempts have also been made to determine the N terminal amino acid sequence of the coat protein; these attempts have been unsuccessful, suggesting that the N terminus is blocked.

A possible second virus There is some evidence that a second virus may be associated with banana bunchy top disease. Initially, Dale et al. (1986) reported the presence of “luteovirus-like” dsRNA in infected but not healthy bananas; this has since been conﬁrmed by three other groups. Secondly, Iskra et al. (1989) reported the puriﬁcation of 28 nm VLPs from BBTV infected bananas. Thus there is a real possibility of a second virus, probably an RNA virus, involved in the disease. This situation can be resolved only by further experimentation. There are a number of instances where it has been demonstrated that the association of two different viruses together cause a particular disease. In fact, there is already one instance of a DNA and a RNA virus combining to cause a disease: rice tungro bacilliform and rice tungro spherical viruses.

Diagnosis and Control of BBTV The most likely early beneﬁts of biotechnology as applied to bananas will be in the area of disease diagnostics and control. This is already happening in relation to the diagnosis of BBTV, and research on control of BBTV using biotechnology is well established.

Diagnosis of BBTV Until recently, diagnosis of BBTV infection was based on the detection and identification of symptoms. The symptoms of BBTV are quite distinctive once fully developed and are easily recognized by experienced personnel. However, there are two main dangers: first, in many instances, personnel charged with the responsibility of detecting BBTV-infected bananas have no practical experience of the symptoms. This is often the case of quarantine officers in countries where BBTV does not occur. Secondly, there are situations where the symptoms of infection are indistinct, masked, or absent; this is particularly true of infected plantlets in tissue culture, the method of choice for disseminating banana germplasm internationally. 90 Banana Bunchy Top Virus: prospects for control through biotechnology

More advanced and accurate techniques of diagnosis have become available since methods for the puriﬁcation of BBTV particles were developed. Both Su and Wu (1989) and Thomas and Dietzgen (1991) have produced monoclonal and polyclonal antisera and utilized these in Elisa-based detection assays. Harding et al. (1991) have developed DNA probes for the diagnosis of BBTV and these appear to be of at least equal sensitivity to the Elisa-based systems. More recently, we have investigated the polymerase chain reaction (PCR) for “ultra-sensitive” and rapid detection of BBTV in infected tissue. Initial experiments indicate that it is possible to detect the virus in less than 5 µg of tissue. The development of these advanced diagnostic techniques greatly decreases the risk of inadvertent dissemination of BBTV to “uninfected” regions, and should improve the effectiveness of other current and future control strategies.

Control options for BBTV The primary objective of the international research effort on BBTV is to control the virus and therefore the disease. There are essentially two strategies for plant virus control: avoidance and resistance. Avoidance includes minimization and exclusion. Resistance can be subdivided into selection, conventional breeding, and genetically engineered resistance. When the options for controlling BBTV are evaluated, there are a number of important points that must be considered: (a) Musa spp. appear to be the only important natural hosts of BBTV; (b) bananas are a vegetatively propagated perennial crop; (c) BBTV does not occur in all banana-producing countries or continents; and (d) bananas are produced both as a large-scale plantation crop and as a domestic subsistence crop. Almost certainly, therefore, different BBTV control strategies must be used for different situations. Avoidance Virus minimization depends very heavily on the ability to be able to diagnose an infection rapidly and sensitively. With the availability of sensitive and specific serological and molecular tests for BBTV, virus avoidance becomes a more attractive option. Exclusion. This operates at two levels. The first is the regional, international, or intercontinental level where the approach is to exclude the virus from an area where it does not occur. This is usually implemented through quarantine. Quarantine may involve the total exclusion of all plant species known to be hosts of the virus, exclusion of those plant species if the particular plants had been transported from or through an “infected” country or region and, ﬁnally, quarantine can involve the testing of plants to determine whether they are infected with a particular virus prior to release. Quarantine is particularly relevant to the control of BBTV as the virus does not occur in South and Central America. There is, however, pressure to import new Musa germplasm into the Americas as the major banana breeding programs are in the Americas whereas the center of diversity of Musa is in Southeast Asia where BBTV occurs. Further, there are a number of countries in “infected” continents that have not recorded BBTV and wish to continue to J Dale, T Burns, S Oehlschlager, M Karan, R Harding 91

exclude the virus, usually by means of quarantine. Also, BBTV has been successfully excluded from regions of Australia using a regional quarantine policy (Dale 1987). The second level at which the exclusion strategy can be implemented is within regions or countries where a particular virus occurs. This normally involves the establishment of a virus-tested scheme where planting material is made available that is known to be virus- free or in which the percentage infection is very low. This has the effect of reducing the amount of virus inoculum within the crop and has been used very successfully with a number of vegetatively propagated crops including potatoes and citrus. A number of countries have used such a policy with BBTV and, at least in Australia, this strategy has contributed to the high level of control of BBTV in that country. Eradication. This strategy involves the destruction of plants that are infected with or are possible hosts of the virus of interest. This may involve the removal of infected crop plants and/or the removal of noncrop host plants. These noncrop hosts may be a species other than the crop or be the same species but growing wild. Again, eradication played an important role in the control of BBTV in Australia. Almost certainly, eradication of BBTV- infected bananas in other countries would involve not only the removal of infected bananas in plantations and gardens but also the removal of wild or noncultivated bananas. Resistance to BBTV Selection. For most crop plants, there are numerous cultivars differing in various agronomic characters. It is often possible to select cultivars resistant to a particular virus by screening. This normally involves assembling a cultivar collection, inoculating each cultivar and finally examining each cultivar for evidence of infection. This usually involves both a visual inspection and screening with a suitably sensitive and specific diagnostic system. Such a selection procedure will identify cultivars that are truly resistant to the virus and also those that are tolerant. Unfortunately, there are no confirmed reports of any banana or plantain cultivars being resistant to BBTV. It would appear that a greater effort should be made to identify resistance or tolerance in banana cultivars. Conventional resistance breeding. Where no cultivars have been identified with resistance or tolerance or where the resistant cultivars are unsuitable, it is possible with many crops to breed resistance into a crop by conventional breeding, that is, crossing a resistant cultivar or species with a suitable other parent. Again, this strategy is unlikely to be successful for the control of BBTV, at least in the short term. There are three major reasons for this: (a) no confirmed BBTV resistance has been recorded in any Musa spp.; (b) there are no current commercial cultivars of bananas that are the product of a banana breeding program; and (c) none of the current banana breeding programs includes BBTV resistance as a target, nor do they screen for BBTV resistance. Genetically engineered resistance. Recently, a number of techniques have been developed that can induce artificial resistance in a plant to a particular virus. The most successful of these techniques is coat protein gene-mediated resistance. Since it was first reported (Powell et al. 1986), this technique has been shown to be applicable to the development of resistance against a wide range of RNA viruses in a number of different hosts and is now accepted as a standard strategy in the “molecular breeding” of virus- 92 Banana Bunchy Top Virus: prospects for control through biotechnology

resistant plants. However, it is unknown at this time as to whether such a strategy will be effective against DNA plant viruses such as BBTV. Despite this lack of prior knowledge, our research is concentrated towards investigating coat protein gene-mediated resistance against BBTV in bananas. We are, however, investigating a number of alternative approaches that may be applicable to development of resistance to BBTV. Our research, in 1992, is centered on three major objectives. 1. The cloning, sequencing, and analysis of the BBTV genome initially to identify the coat protein gene. 2. The development of a convenient and efﬁcient banana transformation system. Our efforts have concentrated on the use of microprojectiles using reporter genes. At this time, this appears to be the most promising technique. 3. The identiﬁcation of suitable plant promoters that will express appropriate genes in bananas. To date, our efforts have concentrated on constitutive monocotyledon promoters.

Conclusions There appear to be excellent prospects for the application of biotechnology in the control of BBTV. Advances have already been made in the diagnosis of the disease using both serological and DNA-based techniques, and these new diagnostic systems are being adopted rapidly, particularly where international movement of germplasm is concerned. The greatest beneﬁts are probably still to come. The most important objective will be the development and distribution of genetically engineered bananas resistant to BBTV. There may also be beneﬁts derived from the study and utilization of the virus itself.

References

CHU PWG, HELMS K. 1988. Novel virus-like particles containing circular single-stranded DNA associated with subterranean clover stunt disease. Virology 167:38-49. DALE JL. 1987. Banana bunchy top, an economically important tropical plant virus. Advances in Virus Research 33:301-325. DALE JL, PHILLIPS DA, PARRY JN. 1986. Double-stranded RNA in banana plants with bunchy top disease. Journal of General Virology 67:371-375. HARDING RM, BURNS TM, DALE JL. 1991. Small single-stranded DNA associated with banana bunchy top disease. Journal of General Virology 72:225-230. ISKRA ML, GARNIER M, BOVE JM. 1989. Purification of banana bunchy top virus. Fruits 44:63-66. MATTHEWS REF. 1982. Classification and nomenclature of viruses. Intervirology 17:1-200. POWELL AP, NELSON RS, DE B, HOFFMAN N, ROGERS SG, FRALEY RT, BEACHY RN. 1986. Delay of disease development in transgenic plants that express the tobacco mosaic virus coat protein gene. Science 232:738-743. SU HJ, WU RY. 1989. Characterisation and monoclonal antibodies of the virus causing banana bunchy top. Technical Bulletin no. 115. Tapai: ASPAC Food and Fertiliser Technology Center. 10 pp. THOMAS JE, DIETZGEN RG. 1991. Purification, characterisation and serological detection of virus-like particles associated with banana bunchy top disease in Australia. Journal of General Virology 72:217-224. WU RY, SU HJ. 1990. Purification and characterisation of banana bunchy top virus. Journal of Phytopathology 128:153-160. GA Strobel, AA Stierle, R Upadhyay, J Hershenhorn, G Molina 93

The Phytotoxins of Mycosphaerella ﬁjiensis, the Causative Agent of Black Sigatoka Disease, and Their Potential Use in Screening for Disease Resistance

GA Strobel1, AA Stierle1, R Upadhyay1,3, J Hershenhorn1, G Molina2

Introduction Bananas and plantains are the primary food source for millions of people in many areas of the world, including Central Africa, Southeast Asia, Central and South America, and the Caribbean. People in these regions are generally faced with the problems of enormous population growth and recurring food shortages, conditions that augment the importance of high-yield, low-cost crops such as bananas and plantains (Stover, Simmonds 1987, pp.397-438). They yield a sweet, nutritious fruit and produce a starch that can be used to prepare a variety of staple foods. But bananas and plantains provide more than just complex carbohydrates. They also yield a diverse array of useful secondary products such as fibers, wrappers, confectioneries, vegetables, catsup, beer, wine, and vinegar (Stover, Simmonds 1987; Valmayor 1986). Commercial cultivation of bananas in tropical America, the Philippines, and Caribbean and Pacific islands has emerged as a significant economic resource. In these regions they not only maintain their importance as a food staple, but also contribute to the GNP, provide employment and fiscal earnings, and generate foreign currency. Of all the plants grown as food, bananas and plantains produce the highest yield for the lowest cost (Stover, Simmonds 1987, pp.402-422). Unfortunately, this vital resource is currently being threatened by the devastating leaf disease known as black Sigatoka. Black Sigatoka is the most severe form of the Sigatoka disease complex involving three closely related fungal pathogens: Mycosphaerella musicola Leach ex Mulder, causative agent of yellow Sigatoka, first identified in the Sigatoka district of Java in 1902;

1Department of Plant Pathology, Montana State University, Bozeman, Montana 59717-0314, USA; 2FHIA, Apdo 2067, San Pedro Sula, Honduras; 3Present address: Directorate of Plant Protection, Quarantine & Storage, NHIV, Faridabad, Haryana, India 94 The Phytotoxins of Mycosphaerella ﬁjiensis

M. fijiensis Morelet, causative agent of black leaf streak, described in Fiji in 1964; and M. fijiensis var difformis Mulder and Stover, causative agent of black Sigatoka discovered in Honduras in 1972 (Fouré 1986). As the Sigatoka diseases have progressed in time their virulence has increased. Their host range has broadened to include plantains, which are resistant to yellow Sigatoka disease. Black Sigatoka has supplanted yellow Sigatoka as the major threat to the cultivation and economic importance of bananas and plantains (Buddenhagen 1986). The black Sigatoka epidemic in Honduras and several other countries in 1972-73 resulted in crop losses approaching 47% (Jaramillo 1986, p.46). Crop yields for Horn-type plantains have been reduced by 50% (Stover, Simmonds 1987, p.293). Control costs for this disease can be astronomical, totalling more than US$17 million per year in Costa Rica alone; no other disease or pest control program in that country can compare (Jaramillo 1986). The cost of controlling black Sigatoka disease in Central America, Colombia, and Mexico surpassed US$350 million in less than 8 years. Large fruit- producing companies must spray fungicide mixtures up to 30 times annually to control black Sigatoka, expending up to 30% of their gross incomes. Annual fungicide costs amount to 10’s of millions of dollars (Stover, Simmonds 1987). Small landowners cannot control the disease in their fields, lacking both the equipment and the funds required for fungicide application. The impact of this plant disease on the poor people of the tropics has been enormous. An alternative solution is the isolation and dissemination of banana and plantain cultivars resistant to the Sigatoka disease complex. Such a program could involve either breeding for resistance or the screening of existing cultivars for inherent resistance. Breeding bananas and plantains is a slow, tedious process, due to serious problems with polyploidy and poor seed yields and pollen infertility (De Langhe 1986). Furthermore, movement of the germplasm of bananas and plantains and different pathogenic biotypes in the screening process could inadvertently introduce new forms of Sigatoka or other diseases into disease-free areas. The isolation and identification of wild-type resistant cultivars or the selection of somaclonal mutants may circumvent these difficulties. Unfortunately, finding such resistant cultivars can be an expensive, time-consuming process. Using traditional methodology, mature plants are challenged by fungal spores produced on naturally infected leaves according to fastidious inoculation schemes to artificially induce the disease. This process may require over 12 months to establish unequivocally the susceptibility or resistance of a particular cultivar to the disease (Stover 1986, pp.114-118). Fortunately, a preliminary report in 1989 indicated that disease symptoms could be induced by the crude extract of the pathogen, suggesting the presence of one or more phytotoxins (Upadhyay et al. 1990). Because of the extreme importance of this disease we began a comprehensive investigation of the phytotoxins of M. fijiensis. It was hoped that phytotoxins isolated from the fungus could provide a rapid indication of cultivar susceptibility or resistance using a simple leaf assay. This paper describes the isolation, characterization, and biological significance of the phytotoxins isolated to date. GA Strobel, AA Stierle, R Upadhyay, J Hershenhorn, G Molina 95

Materials and Methods

Fungal culture The culture of M. fijiensis (IMI 105378) used in this study was originally isolated from Musa sapientum by R Leach in 1964 in Fiji, and was supplied by the Commonwealth Mycological Institute, London. M. musicola (357) was provided by RA Fullerton, DSIR, New Zealand; M. fijiensis cv difformis was isolated by GC Molina, FHIA, Honduras. All cultures were maintained at 26°C on modified M-1-D medium to which 12 mL L-1 coconut milk was added (Pinkerton, Strobel 1976).

Mass culture and extraction The fungi were grown in 2-L Erlenmeyer flasks containing 800 mL modified M-1-D medium (autoclaved) (Pinkerton, Strobel 1976). The medium was inoculated with two mycelial disks (1 cm2) cut from the edge of a vigorously growing fungal colony. Each ﬂask was shaken at 140 rev min-1 for 28 days at 26°C with 12 hours of light (50 EM-2 S-1). After incubation, an equal volume of methanol was added to the fungal culture, which was left for 24 hours at 4°C. The culture was ﬁltered through eight layers of cheesecloth and the volume of the filtrate was reduced by 2/3 by rotary evaporation at 40°C. The concentrated filtrate was exhaustively extracted (3x) with an equal volume of ethyl acetate. The three extracts were combined and reduced to a brown oil (ca 90 mg L-1). For the investigation of fungal metabolites at timed intervals, cultures were grown for 10, 15, and 20 days following the protocol outlined above. The crude extract yields were ca 35, 75, and 86 mg L-1, respectively.

Bioassay A simple leaf puncture bioassay was used to guide us in the isolation of phytotoxins (Sugawara et al. 1985). Five test varieties of bananas and plantains including Grande Naine (Musa acuminata Colla, AAB), Horn plantain (M. acuminata x M. balbisiana), Saba (M. balbisiana, ABB), IV-9 (new line of FHIA), and Bocadillo (AA) were propagated through meristem culture to produce plantlets in test tubes. Plantlets were ready for testing at the 3-5 leaf stage when plants were approximately 10 cm tall. Test fractions were dissolved in 10% methanol and applied to the nicked surface of the fully opened heart leaves in 5 µL droplets. The punctures were encircled with Vaseline petroleum jelly to maintain droplet integrity. Leaves were placed in sterile plastic chambers with moistened ﬁlter paper to provide humidity. Crude extracts were tested at 25, 50, and 100 µg per puncture wound; pure compounds were tested at 0.01, 0.1, 1.0, 5.0, and 10 µg per puncture. Necrotic tissue around the wound was measured at 24 and 48 hours after treatment. Activity indexing utilized a 0-5 scale (+), ranging from no necrosis (0) to lesions 12 mm in diameter (5+). All tests were repeated three times. Methanol (10%) and water were used as controls. Organic extracts of sterile (noninoculated) medium did not exhibit phytotoxicity. 96 The Phytotoxins of Mycosphaerella ﬁjiensis

Results and Discussion

Isolation of phytotoxins The ethyl acetate soluble extracts of the culture filtrate of M. fijiensis were chromato- graphed on Sephadex LH-20 [CHCl3-MeOH, 1:1 (v/v)]. Fractions 3, 4, and 5 were phytotoxic. Fraction 4, the most active fraction, was further resolved by reverse-phase high-performance liquid chromatography on an Altex C-18 column using a gradient elution of acetonitrile-water. This step concentrated the primary phytotoxins in the sixth and eighth of eight fractions. Repeating the HPLC procedure on the sixth fraction yielded 2,4,8-trihydroxytetralone 1. Flash silica gel chromatography of the eighth fraction yielded juglone 2 as the first eluant. Fraction 3 was resolved into nine fractions by reverse-phase HPLC using a MeOH- water gradient. The third fraction was the known phthalide isoochracinic acid 5. Fraction 5 yielded 2-carboxy-3-hydroxycinnamic acid 3 after chromatography on the LH-20 column, as described above. Fraction 6 was 3,4,6,8-tetrahydroxytetralone 6 (4-hydroxyscytalone). Table 1 summarizes the data for the isolation of compounds 1-6. Table 1. Comparison of molecular formulae and separation procedures of phytotoxins isolated from Mycosphaerella fijiensis. a b Compound Molecular Final Rf-TLC % Yield formula isolation c 1 C10H10O4 RP-HPLC 0.39 0.63

2 C10H6O3 Flash silica 0.95 0.02

3 C10H8O5 Sephadex LH-20 0.0 0.03 d 5 C10H8O5 RP-Flash 0.57 0.03

6 C10H10O4 Sephadex LH-20 – 0.50 a: Silica gel developed in chloroform-methanol, 10-1. b: % yield = (mass of compound isolated/mass of total organic extract) x 100. c: 25%-75% acetonitrile in water over 25 min. d: 40%-70% MeOH in water over 20 min.

Structural elucidation of phytotoxins (each shown as a numbered structure in Fig.1)

The molecular formula of compound 1 was determined to be C10H10O4 by high-resolution mass spectrometry, indicating a compound with six sites of unsaturation. Phenolic and ketone functionalities were indicated by infrared absorptions at 3590 and 1640 cm-1. Both 2D COSY and 1D decoupling experiments indicated a trisubstituted benzene ring with three adjacent protons and a cis-1,3-diol system. Detailed analysis of the mass spectral and 1H NMR data resulted in the proposal of structure 1, 2,4,8-trihydroxytetralone. Compound 1 was proposed as a melanin shunt-pathway product by Stipanovic and Bell (1977), although it was isolated only when an appropriate precursor was supplied to GA Strobel, AA Stierle, R Upadhyay, J Hershenhorn, G Molina 97

Figure 1. The seven phytotoxins of Mycosphaerella ﬁjiensis. Each is numbered according to the order shown in the text. Phytotoxin 7 was previously described by Upadhyay et al. 98 The Phytotoxins of Mycosphaerella ﬁjiensis

the fungus Verticillium dahliae (Fig.1). The compound had also been isolated by Fujimoto et al. (1986) from Penicillium diversum var aureum. However, no biological activity has ever been ascribed to the compound. Compound 2, juglone, has been previously isolated from several plant sources including walnuts (Juglans nigra) (Binder et al. 1989) and species of Penicillium (Fujimoto 1986) and Verticillium (Greenblatt, Wheeler 1986). It was readily identiﬁed by spectral data analysis: both spectroscopic and physical characteristics were identical to the authentic compound purchased from the Aldrich Chemical Company.

Compound 3 has a molecular formula of C10H8O5, as determined by chemical ionization mass-spectrometry, indicating seven sites of unsaturation. 1H NMR analysis indicated three adjacent protons on a tri-substituted benzene ring, with the highest ﬁeld aromatic proton, δ6.95 ppm, ortho to a hydroxyl group. Two mutually coupled (J=16Hz) olefinic absorptions at δ8.32 and δ6.18 ppm suggested a trans-α,ß-unsaturated carboxylic acid derivative. This was a very polar compound which streaked on silica TLC plates unless 0.5% triﬂuoroacetic acid was added to the solvent system. This behavior suggested the presence of carboxylic acid functionality; infrared analysis supported this idea with broad OH stretch absorption from 3150 to 3550 cm-1 and carbonyl absorption at 1720 cm-1. The compound was methylated by dropwise addition of freshly prepared diazomethane (Aldrich, no date). Examination of the 1H NMR spectrum of the single product 4 indicated the addition of three methyl groups at δ3.94, δ3.83 and δ3.77 ppm, which indicated the presence of either three carboxylic acid moieties or two acids and a phenolic moiety in the original compound. Difference nOe experiments with the methylated product led to the proposed structure of the novel compound 2-carboxy-3- hydroxycinnamic acid.

Compound 5, isoochracinic acid, C10H8O5, is also a bicyclic compound with the same aromatic substitution pattern found in 1 and 2. It contains both a fused-ring aromatic lactone and a carboxylic acid, as indicated by the infrared absorptions at 1723 and 1718 cm-1. Isoochracinic acid has been isolated from Hypoxylon coccineum and Alternaria kikuchiana, the causative agent of black leaf spot of pear (Kamida, Namiki 1974; Trost et al. 1980). The biogenesis of isoochracinic acid has been of particular interest. It is one of the few naturally occurring phthalides. During this study we were able to establish 3 as a possible precursor of 5. If left in contact with CHCl3-MeOH (1:1) for 24 hours at 25°C, cyclization would occur spontaneously, resulting in a racemic mixture of 5.

Compound 6 has a molecular formula of C10H10O4, with six sites of unsaturation. It has the same aromatic moiety as 1 and 2 but differs in the aliphatic portion. Examination of the spectral data indicates the structure shown, 4-hydroxyscytalone, which is also a known metabolite in the melanin shunt pathway (Bell et al. 1976). 2,4,8-trihydroxytetralone 1. MS: m/z(%) 194 (13.4), 176 (5.6); HRMS: 194.0550 1 (observed), 194.0552 (actual); H NMR (MeOH-D4): 7.54 (t, J=8.1Hz), 7.21 (d, J=8.1), 6.86 (d, J=8.1), 4.89 (dd, J=11.4, 5.1), 4.41 (dd, J=13.3, 4.8), 2.63 (dt, J=4.8, 11.4) and 1.98 (dd, J=13.3, 5.1) 1 juglone 2. MS: m/z 174; H NMR (MeOH-D4): 7.68 (t, J=7.5Hz), 7.59 (d, J=7.5), 7.31 (d, J=7.5), 7.00 (s), 7.01 (s) GA Strobel, AA Stierle, R Upadhyay, J Hershenhorn, G Molina 99

1 2-carboxy-3-hydroxycinnamic acid 3. MS: m/z 208; H NMR (MeOH-D4): 8.32 (d, J=16Hz), 7.39 (t, J=7.1 Hz), 7.09 (d, J=7.1), 6.95 (J=7.1) 2-carboxy-3-methoxycinnamic acid dimethyl ester 4. HRMS: 250.0821 (observed), 1 250.0817 (actual); H NMR (CDCl3): 7.63 (d, J=16Hz), 7.38 (t, J=8.0), 7.19 (d, J=8.0), 6.94 (d, J=8.0), 6.37 (d, J=16) isoochracinic acid 5. HRMS: 208.0364 (observed), 208.0368 (actual); 1H NMR (MeOH- D4): 7.52 (t, J=7.5Hz), 7.03 (d, J=7.5), 6.86 (d, J=7.5), 5.81 (dd, J=6.0, 5.8), 2.85 (dd, J=15.4, 5.8), 2.71 (dd, J=15.4, 6.0) 4-hydroxyscytalone 6 MS: m/z(%) 210 (22.1), 191.9 (9.2), 137 (39.1); 1H NMR (MeOH- D4): 6.73 (d, J=2.5Hz), 6.21 (d, J=2.5), 4.30 (d, J=3.0), 3.97 (ddd, J=3.0, 4.8, 8.1), 2.96 (dd, J=4.8, 17.5), 2.66 (dd, J=8.1, 17.5).

Biological activity Each of the compounds isolated displayed phytotoxicity at one or more test concentrations on one or more cultivars of banana or plantain. Compound 1 exhibited the greatest host selectivity, analogous to the fungal pathogen itself, particularly at the 5 µg 5 µL-1 level. That is, the variety IV-9 is resistant to black Sigatoka disease and was insensitive to 1 up to the 10 µg 5 µL-1 application rate in the leaf bioassay test. Saba, a tolerant banana cultivar, was only slightly reactive to 1, unlike the extremely disease-susceptible varieties Boca and Horn plantain which developed large necrotic lesions after application of toxin 1. Compound 1 shows definite potential as a screening tool for toxin sensitivity in tissue culture systems, especially with the recent development of such systems (Novak et al. 1989). The most biologically active phytotoxin was juglone 2, which induced the formation of necrotic lesions on all cultivars of bananas and plantains at 0.1 µg (Table 2). Unfortunately, it was not host-selective at any test concentration, making it unsuitable for cultivar screening. Juglone was isolated at extremely low concentrations (Table 1) and therefore may not play an important role in symptom induction. The novel phytotoxin 2-carboxy-3-hydroxycinnamic acid 3 demonstrated some host selectivity at the 5 µg 5 µL-1 level, which again paralleled that of the fungus. Toxins 5 and 6 were considerably less active than the other phytotoxins isolated and displayed no host selectivity. Fijiensin 7, previously reported as a phytotoxin from the fungus, exhibits a low level of bioactivity and no host selectivity (Upadhyay et al. 1990). In this study we found no evidence of phytotoxins in the aqueous phase of the fungal culture extracts and no evidence of additional phytotoxins in the ethyl acetate extract. It appears on the basis of yield, level of bioactivity, and host selectivity, that 2,4,8-trihydroxytetralone 1 is the most important phytotoxin produced by M. fijiensis. It is interesting to note that all of the phytotoxins isolated might possess a much greater level of bioactivity than appears in Table 2. The leaf puncture method is a reliable but relatively crude technique which has recently been measurably improved by a fluorescence labeling method employing flow-cytometry (Bergland et al. 1988). Compounds with low solubility can have more ready access to intact protoplasts, and 100 The Phytotoxins of Mycosphaerella fijiensis

Table 2. Comparison of the reactions of various banana/plantain cultivars to the phytotoxins produced by Mycosphaerella fijiensisa. µg/puncture Grande Naine Horn plantain Saba Boca IV-9 Compound 1 0.10 µg- - --- 1.00 - + - - - 5.00 + ++ - + - 10.00 ++ +++ + +++ - Compound 2 0.01 µg - - - - - 0.10 + + + + + 1.00 +++ +++ ++ +++ +++ 5.00 +++ ++++ ++ ++++ ++++ Compound 3 1.00 µg - - - - - 5.00 + + - + - 10.00 +++ ++++ + ++++ ++++ 20.00 ++++ ++++ ++ ++++ ++++ Compound 5 1.00 µg - - - - - 5.00 - + + + + 10.00 + ++ ++ ++ + 20.00 ++ +++ +++ +++ ++ Compound 6 1.00 µg - - - - - 5.00 - - - - - 10.00 - + + + + 20.00 + + + + + a Toxins were applied at the rate indicated in 5 mL of solution (2% EtOH in water) as described in Materials and methods. Readings recorded above were made 48 hours after treatment. may demonstrate toxicity levels that are orders of magnitude greater than those determined by the leaf puncture method. Our study also involved analysis of toxin production at different times during the fermentation process. We were particularly interested in determining when 1 was produced in culture. It has been noted in field studies that a 3-4 week lag time occurs between inoculation of a host with the pathogen and onset of disease symptoms (Stover 1986). It seems unlikely that a toxin that is capable of inducing disease symptoms in the laboratory in a matter of hours would require 3 weeks for symptom induction in the field. To this end, we grew the fungus for intervals of 10, 15, 20, and 28 days. The ethyl acetate soluble extract of each culture was analyzed in precisely the same manner as before. No GA Strobel, AA Stierle, R Upadhyay, J Hershenhorn, G Molina 101

Figure 2. Proposed pentaketide pathway of melanin biosynthesis and shunt pathways leading to the synthesis of 1, 2, and 6 (Greenblatt, Wheeler 1986). evidence of either 1 or 2 was found until the cultures had reached 28 days (or past 20 days), although isoochracinic acid 5 was found in all of the cultures. This interesting discovery correlates the observed lag time between inoculation and onset of disease symptoms in the field with the production of 1 in culture. Further investigation of the phytotoxins of the 10-, 15-, and 20-day cultures is in progress. One extremely important aspect of this study concerns the determination of any morphological or chemical differences among the three fungi of the Sigatoka complex, M. musicola, M. fijiensis, and M. fijiensis var difformis. Preliminary data show some interesting chemical trends. All three fungi were grown and processed as described above. M. musicola causes the least virulent form of the Sigatoka disease complex, and it 102 The Phytotoxins of Mycosphaerella fijiensis

does not appear to produce 1 or 2. M. fijiensis var difformis, causal agent of the most severe form of black Sigatoka, produces toxin 1 at 6-10 times the level of its less virulent counterpart M. fijiensis. Detailed comparisons of phytotoxin production by these three fungi will be presented at a later date. It is interesting to note the biosynthetic origin of the major toxin, 1, and related metabolites in the fungal extract. Examination of the literature (Stipanovic, Bell 1977; Bell et al. 1976a,b; Wheeler, Stipanovic 1985) indicated that 2,4,8 trihydroxytetralone 1, juglone 2, and 4-hydroxyscytalone 6 are likely melanin shunt-pathway metabolites (Fig.2). Wheeler studied the biosynthesis of melanin in 20 species of ascomycetous and imperfect fungi, and found that, in all species tested except Aspergillus niger, the melanin precursor is 1,8-dihydroxynaphthalene, not indole, as was previously suspected (Wheeler 1983). In a series of elegant experiments using melanin-deficient mutants, these investigators carefully mapped not only the melanin pathway but also the major shunt-pathway metabolites (Tokonsbalides, Sisler 1979). Many of the compounds isolated from M. fijiensis may arise via these pathways. Melanin occurs in the cell walls of many fungi, apparently protecting the cells from desiccation and ultraviolet radiation. It may also protect fungal cells from other microorganisms (Greenblatt, Wheeler 1986). The production of melanin can also be an important determinant of pathogenicity: melanization of the fungal appressorium is a necessary prerequisite to invasion of rice by Pyricularia oryzae (Woloshuk et al. 1980, 1983), of cucumbers by Colletotrichum lagenarium (Kubo et al. 1985), and of beans by C. lindemuthianum (Wolkow et al. 1983). Application of the fungicide tricyclazole, a systemic inhibitor of melanin biosynthesis, can control these fungal infections by interfering with melanization of the fungal cells (Greenblatt, Wheeler 1986; Tokousbalides, Sisler 1979; Woloshuk et al. 1983). Examination of the biosynthetic melanin pathway, however, indicates an interesting departure from the norm. Tricyclazole inhibits pathogenicity by blocking melanization, which leads to accumulation of shunt metabolites in the fungal cells (Tokousbalides, Sisler 1979). In M. fijiensis, it is the shunt-pathway metabolites that are the probable agents causing cell death. We could predict that application of tricyclazole could actually increase the virulence of the fungi by increasing the production of these phytotoxins. We ground and extracted, with EtOAC, 500 g of infected banana leaf tissue in order to find evidence of any or all of the phytotoxins within the plant. Chromatography on an LH- 20 column, as previously described, did not yield any fraction that possessed biological activity. Likewise, no evidence for the presence of any of these compounds could be found by TLC (Table 1), or NMR. This is not surprising, considering the strong possibility that in the process of fungal infection, these phytotoxins could conceivably be oxidized, reduced, or derivitized into other compounds not having biological activity. Comparable studies using plant material in the very early stages of infection may prove more fruitful in discovering these substances in the tissue. Such information would be helpful in more firmly establishing their role in the infection process. As a test for its potential usefulness, a purified preparation of tetralone from M. fijiensis, was applied to a series of meristem plantlets at varying concentrations. The GA Strobel, AA Stierle, R Upadhyay, J Hershenhorn, G Molina 103

banana and plantain cultivars showed differential activity in concentrations as low as 5 µg µL-1. This activity paralleled the known resistance or susceptibility of these cultivars to M. ﬁjiensis. Therefore, we feel that preparations of the toxin have potential usefulness in plant breeding/selection programs in all parts of the world where bananas and plantains are grown. Acknowledgments The authors wish to acknowledge the ﬁnancial assistance of NSF grant DMB-8607347, an AID grant in cooperation with G Molina and the Montana Agricultural Experiment Station.

References

ALDRICH CHEMICAL COMPANY. No date. Diazomethane Generators. Technical Information Bulletin AL-103. Milwaukee, WI, USA: the Company. BELL AA, STIPANOVIC RD, PUHALLA JE. 1976a. Tetrahedron 32:1353-1356. BELL AA, PUHALLA JE, TOLMSOFF WJ, STIPANOVIC RD. 1976a. Can. J. Microbiol. 22:787-799. BERGLUND D, STROBEL S, SUGAWARA F, STROBEL GA. 1988. Pl. Sci. 56:183-188. BINDER RG, BENSON ME, FLATH RA. 1989. Phytochemistry 28:2799-2801. BUDDENHAGEN IW. 1986. Pages 95-109 in Banana and Plantain Breeding Strategies (Persley GJ, De Langhe EA, eds). ACIAR Proceedings no.21. Canberra, Australia: ACIAR. DE LANGHE EA. 1986. Pages 19-23 in Banana and Plantain Breeding Strategies (Persley GJ, De Langhe EA, eds). ACIAR Proceedings no.21. Canberra, Australia: ACIAR. FOURÉ E. 1986. Pages 110-113 in Banana and Plantain Breeding Strategies (Persley GJ, De Langhe EA, eds). ACIAR Proceedings no.21. Canberra, Australia: ACIAR. FUJIMOTO Y, YOKOYAMA E, TAKAHASHI T, UZAWA J, MOROOKA N, TSUNODA H, TATSUNO T. 1986. Chem. Pharm. Bull. 34:1497-1500. GREENBLATT GA, WHEELER, MH. 1986. J. LIQ. Chrom. 9:971-981. JARAMILLO RC. 1986. Page 42 in Banana and Plantain Breeding Strategies (Persley GJ, De Langhe EA, eds). ACIAR Proceedings no.21. Canberra, Australia: ACIAR. KAMEDA K, NAMIKI M. 1974. Chem. Lett. (1974 vol.):1491-1492. KUBO Y, SUZUKI K, FURUSAWA I, YAMAMOTO M. 1985. Pest. Biochem. Physiol. 23:47-55. NOVAK F, AFZA R, VAN DUREN M, PEREA-DALLOS M, CONGER BV, XIALONG T. 1989. Biotech. 7:154-159. PINKERTON F, STROBEL GA. 1976. Proc. Nat. Acad. Sci.(USA) 73:4007-4011. STIPANOVIC RD, BELL AA. 1977. Mycologia 69:165-171. STOVER RH. 1986. Pages 114-118 in Banana and Plantain Breeding Strategies (Persley GJ, De Langhe EA, eds). ACIAR Proceedings no.21. Canberra, Australia: ACIAR. STOVER RH, SIMMONDS NW. 1987. Bananas. New York, USA: Wiley. SUGAWARA F, STROBEL GA, FISHER LE, VAN DUYNE GD, CLARDY J. 1985. Proc. Natl. Acad. Sci. (USA) 82:8291-8294. TOKOUSBALIDES MC, SISLER HD. 1979. Pest. Biochem. and Physiol. 11:64-73. TROST BM, RIVERS GT, GOLD JM. 1980. J. Org. Chem. 45:1835-1838. UPADHYAY RK, STROBEL GA, COVAL SJ, CLARDY J. 1990. Experientia 46:983-984. WHEELER MH. 1983. Trans. Br. Mycol. Soc. 81:29-63. WHEELER MH, STIPANOVIC RD. 1985. Arch. Microbiol. 142:234-241. WOLKOW M, SISLER HD, VIRGIL EL. 1983. Phys. Plant Path. 23:55-71. WOLOSHUK CP, SISLER HD, TOKOUSBALIDES MC, DUTKY SR. 1980. Pest. Biochem. Physiol. 14:256-264. WOLOSHUK CP, SISLER HD, VIRGIL EL. 1983. Phys. Plant Path. 22:245-259. VALMAYOR RV. 1986. Pages 50-55 in Banana and Plantain Breeding Strategies (Persley GJ, De Langhe EA, eds). ACIAR Proceedings no.21. Canberra, Australia: ACIAR. 104 Molecular Approaches to Identifying Fusarium Wilt Resistance

Molecular Approaches to Identifying Fusarium Wilt Resistance

RC Ploetz

Introduction Fusarium wilt has a long and destructive history in the world’s banana-producing regions (Ploetz et al. 1990; Stover 1962). Also known as Panama disease, it is caused by a variable soilborne fungus, Fusarium oxysporum f.sp. cubense. Fusarium wilt almost destroyed the export trade during the ﬁrst half of the 20th century. Although continued damage on Gros Michel (AAA), the cultivar used by exporters at that time, was avoided by the conversion to cultivars of the Cavendish subgroup (AAA), the later clones are now succumbing to a new variant of the pathogen, race 4, in subtropical production areas in the eastern hemisphere. Although signiﬁcant damage has yet to occur on the Cavendish clones in the tropics, there is concern that race 4, or something like it, will develop in the important export plantations found in these regions. Less publicized, but potentially more destructive than the epidemics on Gros Michel, are outbreaks that affect nonexported, locally consumed cultivars. Susceptible clones are used as staple foods or important dietary supplements in eastern Africa, Brazil, Southeast Asia, and other production regions (Alves et al. 1987; Ploetz et al. 1990; Valmayor 1987). Since preferred cultivars are heavily damaged or have been eliminated by fusarium wilt in many of these areas, there is an urgent need to develop or identify resistant replacements that would be accepted by local consumers.

Resistance to Fusarium Wilt Banana breeding began in 1922 at the Imperial College of Tropical Agriculture in Trinidad and in 1924 in Jamaica (Stover, Buddenhagen 1986). The impetus for these programs, and one that subsequently began and still continues in Honduras [Fundación Hondureña de Investigación Agrícola (FHIA)], was fusarium wilt. Others at this workshop will address past and current breeding efforts (see Rowe 1993; Vulsteke, Swennen 1993), but it is instructive here to note the record of the early programs and how resistance to fusarium wilt among products of these programs was assessed.

University of Florida, Tropical Research and Education Center, 18905 SW 280th Street, Homestead, FL 33031, USA RC Ploetz 105

Despite their long history, acceptable commercial replacements for the AAA dessert cultivars have not been developed by the breeding programs (Buddenhagen 1990; Stover, Buddenhagen 1986). Genetic constraints, such as triploidy and sterility of some desirable parents, impede straightforward progress in this area. In addition, the exceptionally good agronomic attributes of Gros Michel and the Cavendish cultivars have been difficult to match in the resistant products that have been produced to date (in the final analysis, producing a fusarium-wilt resistant AAA dessert banana with horticultural traits that equal those of the current Cavendish ideotype may be an unrealistic goal). Although success in breeding resistant AAB dessert or ABB cooking bananas may be more easily achieved, breeding for these targets has started only recently (Buddenhagen 1990; Rowe 1990; Shepherd et al. 1987; Vulsteke 1993). In general, the resistance of AAB and ABB replacements produced in Brazil, Honduras, and Nigeria is either not well characterized or not known. New genotypes produced by the breeding programs are evaluated in field trials. Unfortunately, field screening has provided erroneous information about the resistance of products of these programs (for examples of “resistant” clones succumbing to fusarium wilt after deployment to different areas, see Stover, Buddenhagen 1986). Until recently, screening sites were limited to the few areas in which active breeding programs existed. Unfortunately, these sites provided imperfect screening environments. Either narrow pathogenic variability in F. oxysporum f.sp. cubense at the sites compromised their effectiveness (e.g., FHIA in Honduras), or edaphic factors at a location restricted disease development (i.e., disease-suppressive soils at the EMBRAPA site in Brazil). Even now that the value of geographically diverse screening sites is recognized, it is not clear that field trials in different locations will provide the necessary information. Since F. oxysporum f.sp. cubense is a diverse pathogen, and because the factors that determine whether or not fusarium wilt develops on a given clone in a given area are complex and not completely understood, determining whether new clones possess global or durable resistance to fusarium wilt will be difficult (Ploetz 1990). Banana breeding is time-consuming and expensive. Thus, methods either to expedite the selection process or to generate new, resistant genotypes would be very useful. Below, I briefly review and discuss means by which fusarium wilt-resistant bananas could be produced in the future.

Resistance Mechanisms Although resistance to diseases has been identiﬁed in other crops when pathogenesis was not well understood, knowledge of this process is often useful during these endeavors. For fusarium wilt on banana, the interaction between the pathogen and host is complex. Work to define ways in which F. oxysporum f.sp. cubense and banana interact was begun by Wardlaw (1930) who noted the formation of tyloses in banana in response to vascular colonization by this fungus. Although much remains to be learned about these processes, a good basic understanding of the steps leading from root colonization to death of a susceptible cultivar, or defense of a resistant plant, is now 106 Molecular Approaches to Identifying Fusarium Wilt Resistance

available (those interested in in-depth treatments of this subject should refer to Beckman 1987, 1989, 1990). I am not aware of specific studies on infection of banana by F. oxysporum f.sp. cubense, but results from work with other plants and F. oxysporum suggests that resistance in other hosts to different formae speciales of this fungus is expressed after infection occurs; the roots of resistant and susceptible plants are both infected (references in Beckman 1987). Resistance and susceptibility are ultimately determined by a series of chemical and physical events that occur in the xylem (Beckman 1987; Pegg 1985). Resistant hosts effectively wall off the invading parasite by forming gels, gums, tyloses and other xylem-occluding products (e.g., Vander Molen et al. 1977a,b). In turn, a successful pathogen breaches host defenses by virtue of, among other things, an array of hydrolytic enzymes that degrade host cell walls and reaction products (Pegg 1985).

Phenolic compounds Phenolic compounds that impregnate host reaction products are thought to play an important role in the resistance process, either making physical barriers stronger or chemically impervious to the previously mentioned hydrolytic enzymes of the pathogen (Beckman 1987; MacHardy, Beckman 1981; Pegg 1985). Enzymes that are instrumental in the formation of these compounds have been previously studied in many different hosts, including banana (Mace, Wilson 1964; Mace et al. 1972; Mueller, Beckman 1978). For example, phenol-oxidizing enzymes such as peroxidase, tyrosinase and laccase, are associated with many different vascular diseases (Pegg 1985). In general, these enzymes are inducible. Their role in resistance may be a timing phenomenon; that is, hosts which quickly produce sufﬁcient levels of the important products in response to the pathogen resist disease development. Peroxidases and polyphenoloxidases are stored, preformed, in various, localized sites in banana (Mace et al. 1972; Mueller, Beckman 1978). Different isoforms of the enzymes (isozymes) and the quantities produced are known to differ among different banana genotypes (Jarret, Litz 1986a,b) and, since the levels and number of peroxidase isozymes produced are greatest in roots of banana and other hosts, it has been postulated that they may help protect plants against infection by root pathogens (Bonner et al. 1974; Lagrimini, Rothstein 1987). The involvement of these enzymes in the resistance process and the considerable enzymatic variation that exists among different banana genotypes indicate that enzyme assays could help characterize and identify fusarium wilt resistance. If speciﬁc forms or relatively high quantities of these enzymes are correlated with resistance in the host, it is conceivable that products of breeding programs or pre- existing clones could be quickly and effectively evaluated via enzyme analyses, without resort to conventional disease screening trials. The potential for this approach is borne out by recent, unpublished work by Morpurgo (as cited in Novak 1992). He indicated that constitutive levels of peroxidase in SH-3362, a race-4 resistant, synthetic AA hybrid produced at FHIA, were greater by 10- fold than in Pisang Mas, a susceptible AA cultivar. Additional work in this area is warranted. RC Ploetz 107

Somaclonal Resistance Somaclonal variation is a common phenomenon that has drawn the attention of those interested in all aspects of crop improvement (Larkins, Scowcroft 1983). Somaclonal variants with enhanced resistance to various diseases are known for many plants, including banana (Hwang 1990). Resistant somaclones have been recovered from cell or tissue cultures with and without applying selection pressure prior to field screening. Hwang (1990) generated about 20,000 plantlets of Giant Cavendish via conventional meristem culture and screened these, without prior selection, in a field infested with race 4. Six somaclones with tolerance to race 4 were identified. Obviously, field selection of disease-resistant somaclones is a time-consuming and labor-intensive procedure. Although preselection of “resistant” somaclones in the laboratory prior to field testing could speed this process, such an approach requires reliable selection criteria. Perhaps the simplest procedure by which cell or tissue cultures have been screened is by direct exposure to pathogens (Heath-Pagliuso et al. 1988; McComb et al. 1987). Cell or tissue cultures, however, are often very sensitive to this sort of treatment and it may not be possible to distinguish resistant from susceptible selections in vitro. At least with the different fusarium wilt systems that I am aware of, both resistant and susceptible materials die after challenge (Heath- Pagliuso et al. 1988; pers. com. Kistler; Ploetz 1990). Several pathogen-free methods have been used that are generally less destructive, more reproducible, and amenable to manipulation than the direct exposure method. Since they are pathogenic determinants (i.e., responsible for pathogenesis), host- specific toxins are ideal for this purpose. For example, T-toxin, produced by the fungus Bipolaris (Helminthosporium) maydis, has been used to select somaclones of maize (Zea mays) with resistance to southern corn leaf blight (Gegenbach et al. 1978). However, from a selection standpoint, host-specific toxins are, unfortunately, not always involved in pathogenesis. Nonspecific toxins produced by pathogens may be useful, even when the relationship between tolerance of a toxin and resistance to a disease is not clear. Fusaric acid is a good example of such a host-nonspecific toxin. It is produced in vitro by many formae speciales of F. oxysporum and has been detected in infected hosts, including banana (see Beckman 1987 and references therein). Although its role as a wilt toxin is controversial, fusaric acid has been used to select fusarium wilt-resistant variants of tomato (Lycopersicon esculentum) (Shahin, Spivey 1986), and suggested as a selective agent for fusarium wilt-resistant musk melon (Cucumis melo) (Mégnégneau, Branchard 1988). Results with strawberry (Fragaria X ananassa) (Toyoda et al. 1991) and banana (Epp 1987; Krikorian 1990), however, have been less encouraging. Fusaric acid is highly toxic. During work on banana, there appeared to be an either/or phenomenon: either meristems and cell cultures died at a given concentration of the toxin or, at lower concentrations, no growth reduction or mortality was observed (Epp 1987; Krikorian 1990). Since sublethal concentrations of fusaric acid that could be used to select tolerant somaclones were never identified, fusaric 108 Molecular Approaches to Identifying Fusarium Wilt Resistance

acid may not be a useful tool for selecting fusarium wilt-resistant somaclones of banana. When culture filtrates of pathogens have been used as screening agents in plant tissue cultures, the results have also been mixed. Arcioni et al. (1987) and Gray et al. (1986) reported using culture filtrates to identify, respectively, fusarium wilt-resistant alfalfa (Medicago sativa) (F. oxysporum f.sp. medicaginis) and brown stem rot- resistant soybean (Glycine max) (Phialophora gregata). However, Mégnégneau and Branchard (1991) and Vardi et al. (1986) indicated that culture filtrates were not useful for identifying, respectively, fusarium wilt-resistant musk melon (F. oxysporum f.sp. melonis) and different cultivars of Citrus spp. with resistance to Phytophthora citrophthora. A wide range of compounds and stimuli induce resistance in plants to disease agents (Kuc 1985), some of which have been used to select disease-resistant somaclones in plant cell or tissue cultures. For example, Buiatti and colleagues (Buiatti et al. 1985, 1987a,b) used cell wall components of F. oxysporum f.spp. dianthi and lycopersici as elicitors during work on, respectively, carnation (Dianthus caryophyllus) and tomato. Clones with altered phytoalexin production, compared with the parental lines, were recovered in both systems, and they suggested that such a screening technique could provide a new method for selecting pathogen-resistant genotypes. Krikorian (1990) used potassium vanadate as an elicitor in cell cultures of Cardaba, an ABB cooking banana. He indicated that small percentages of cells survived treatment and that organized propagules were recovered. At this time, I am not aware of results from pathogenicity trials with material produced during this work. Although Hwang’s results (1990) make it clear that somaclonal variation could provide new sources of resistance in banana to fusarium wilt, I am not aware of results other than his in this area. Certainly, as cell or tissue culture techniques for banana are refined or new ones are developed, investigation of somaclonal variation for resistance purposes should be expanded, especially when conventional breeding strategies are unsuccessful.

Markers for Resistance The identification of markers for disease resistance has received significant attention as a means to increase the efficiency of screening programs (Behare et al. 1991; Medina, Stevens 1980; Michelmore et al. 1991; Rick, Fobes, 1974; Robinson et al. 1970; Sarfatti et al. 1989). The objective of these studies is to identify phenotypic or genetic traits that are linked to resistance genes. Marker identification relies on populations of host progeny that segregate for resistance and susceptibility and nonambiguous means by which the presence or absence of the marker and disease response of the progeny can be scored. For example, Robinson et al. (1970) identified an anthocyaninless gene in tomato (ah) that was linked to a gene for resistance to tobacco mosaic virus (Tm2). By selecting progeny with green stems, breeders were also able to select for resistance to this virus. Rick and Fobes (1974) observed linkage between resistance in tomato to RC Ploetz 109

Meloidogyne incognita (rootknot nematode) and a particular acid phosphatase allele (Aps-1). Thus, progeny that possessed Aps-1 could be selected as nematode-resistant without recourse to nematode challenge. Recently, two additional types of genetic markers have been used to expedite resistance breeding. In fundamentally different ways, restriction fragment length polymorphisms (RFLPs) and randomly amplified polymorphic DNAs (RAPDs) identify variations in host DNA sequences that can be exploited as markers for resistance genes (for detailed discussions of these techniques refer to Beckman and Soller 1986; Michelmore et al. 1991; and Tanksley et al. 1989). For example, RFLPs have been used as markers for genes for resistance to Stemphylium spp. (Sm) and race 2 of F. oxysporum f.sp. lycopersici (I2) (Behare et al. 1991; Sarfatti et al. 1989), whereas Michelmore et al. (1991) described screening bulked, segregating host populations for RFLP or RAPD markers and provided data on the use of RAPD markers for identifying resistance genes in lettuce (Lactuca sativa) to downy mildew (Bremia lactucae). Although resistance markers can be powerful tools for plant breeders, their identification can be quite time-consuming and expensive. In addition, they may rely upon considerable background information that has been generated for a given host genome. Compared with lettuce or tomato, two of the above hosts for which impressive linkage maps exist, almost nothing is known about banana (see González de León 1993). Also lacking in banana is a progeny population that segregates for resistance to fusarium wilt. Presumably, segregating populations could be generated by breeding programs at FHIA, or elsewhere, which have diverse collections of resistant and susceptible parents. Information on the inheritance of fusarium wilt resistance among different Musa acuminata subspecies and M. balbisiana was generated by Vakili (1965). Genetic control of resistance was not complex [for example, he determined that resistance to race 1 in Lidi (AA) was controlled by a single dominant gene]. Unfortunately, additional research in this area has apparently not been conducted after the publication of Vakili’s (1965) work. However, if we assume that resistance to different races of fusarium wilt in other banana genotypes is controlled by single genes, it should be possible to identify markers for resistance to this disease in the future (see Gonzáles de León 1993 for information on genetic mapping of the M. acuminata genomes). His, and other research on this topic, could begin to provide the foundation on which marker work could proceed.

Genetic Transformation Genetic transformation of plants to enhance either disease or insect resistance is discussed by others at this workshop (Dale et al. 1993; Fauquet, Beachy 1993; De Waele 1993). Transformation with coat protein genes of viral pathogens is now a widely used disease-control strategy (see Dale et al. 1993; Fauquet, Beachy 1993), but transformation to enhance resistance to fungal-incited diseases has received little attention. Although experience in the latter area is scant, production of transgenic banana with resistance to fusarium wilt is conceivable (Murfett, Clarke 1987). 110 Molecular Approaches to Identifying Fusarium Wilt Resistance

Several substantial hurdles must be cleared before such an objective can be realized. First, it will be necessary to develop transformation protocols for banana (see Dale et al. 1993; Bakry et al. 1993). Many different methods have been developed for transforming plants (for a comprehensive list and discussion of the available methods, see Potrykus 1991); however, I know of no published reports on banana transformation. Since a reliable transformation technique could be used for diverse and important applications, work in this area is of great interest (see Novak 1992). Unfortunately, the widely used and successful method, involving genetically engineered strains of Agrobacterium tumifaciens as DNA vectors, is not effective with all plant species. In particular, monocotyledons are not usually amenable to A. tumifaciens-mediated gene transfer, but there are notable exceptions (Potrykus 1991). Some “transitional” monocotyledons, such as asparagus (Asparagus officinalis) (Potrykus 1991) and black pepper (Piper nigrum) (pers. com. RE Litz) are easily transformed. Other monocotyledons, such as maize, have been transformed via agroinfection, but this appears to be more difficult than for the previously mentioned monocotyledons (Grimsley et al. 1987). Still, it should not be assumed that attempts to transform banana by this route would be futile (Murfett, Clarke 1987). Other methods are used less frequently and, to date, have been less successful than the Agrobacterium-based systems. This may change as new transfer techniques are developed and refined. Microprojectile or biolistic transformation is one of the more unusual approaches to the problem of introducing foreign DNA into plant cells (see Klein et al. 1987; Sanford et al. 1987; Dale et al. 1993; Bakry et al. 1993). Transformation with microprojectiles has several advantages. Two of several listed by Potrykus (1991) are that it may be used on many different types of tissue (i.e., specific tissues are usually not needed), and that it is host-range nonspecific. Thus, the biolistic approach may be especially useful for species, such as bananas, that might not lend themselves to Agrobacterium-mediated gene transfer (Dale et al. 1993). DNA has been introduced into plant protoplasts by a variety of techniques (see Potrykus 1991 for details). Like the microprojectile strategies, protoplast transformation avoids the host-range constraints of A. tumifaciens-based systems. Physical methods for transforming protoplasts include electroporation and microinjection, whereas chemical methods rely on polyethylene glycol or other osmotica for DNA transfer. The genetic transformation of banana with these techniques would rely on procedures by which protoplasts could ultimately be used to regenerate plants. Since totipotency for banana protoplasts has only recently been reported (referenced in Novak 1992), it may be some time before these systems can be used to produce transgenic banana plants. Results for other crops indicate that these protocols can be used rarely for more than a few cultivars or types of given crops. Although gene transfer via banana protoplasts may not be feasible in the near future, continued attention should be given to this possibility. Obviously, for any transformation technique to be useful, a gene or genes of interest must have been previously identified and cloned for use. In addition, marker gene(s) must be incorporated with the chosen gene(s) to facilitate selection of transformed tissues or cells. An effective promoter for banana must also be found and used. I am aware of no work on banana in these areas. RC Ploetz 111

Recently, plant transformation for enhanced chitinase production has been considered as a means to increase resistance to diseases caused by fungi (Jones et al. 1986; Nitzsche 1983). Chitin is a cell-wall component of fungi, but not plants (Bartnicki- Garcia 1968). Plants normally produce constitutive, low levels of chitinases, enzymes that have been shown to inhibit fungal growth in vitro (Schlumbaum et al. 1986). Since infection by fungal pathogens induces an increase in the levels of these enzymes in plants, and because there are no known substrates for chitin in higher plants, chitinases are thought to play a role in the inducible defense systems of plants. Chitin-specific strategies are aimed at raising the constitutive levels of chitinase production in hosts. It is hoped that plants with altered chitinase production will, as a consequence, possess enhanced protection against infection and colonization by fungal pathogens that have chitinous cell walls. In general, these strategies involve transforming plants with genetically engineered constructs of constitutive promoters and foreign chitinase genes from either bacteria or higher plants (Broglie et al. 1991; Chet et al. 1991; Jones et al. 1986; Linthorst et al. 1990). Recently, Broglie et al. (1991) reported enhanced resistance in tobacco to the soilborne fungus Rhizoctonia solani after transformation with a bean chitinase gene under constitutive control of the 35S promoter from cauliflower mosaic virus. When transgenic seedlings were planted in soil infested with this pathogen, seedling survival was increased and disease development was delayed relative to nontransformed plants. Since transformation for chitinase production is a very new strategy, it will probably be some time before its usefulness against diseases other than those caused by necrotrophic fungi is known. It is hoped that vascular wilt diseases, such as that caused by F. oxysporum f.sp. cubense, will also be amenable to these schemes. Manshardt (1992) discussed two additional classes of compounds that target chitin. Lectins that agglutinate and precipitate chitin are known and one from Utrica dioica, UDA, has been shown to have activity against several plant-pathogenic fungi. Partial knowledge of its amino acid composition and its small size (8500 MW) make the eventual identification and isolation of a gene that encodes the production of UDA hopeful. The other class of chitin-specific compounds, nikkomycins, are nucleoside peptide antibiotics produced by the actinomycete Streptomyces tendae. Nikkomycins are inhibitors of chitin synthases, the enzymes responsible for chitin production, and they are capable of inhibiting fungal growth. Although their production is encoded by more than one gene (probably several), work on the isolation of these genes and their use to engineer plants capable of producing nikkomycins is under way. Receiving less attention than the chitin-specific compounds are low molecular weight lytic peptides that disrupt cell membrane integrity (discussed in Manshardt 1992). They have been isolated from several different animals, and genes for their production have been isolated. Tomato and potato (Solanum tuberosum) have been transformed with modified genes which produce one class of these compounds, cecropins, and enhanced resistance to Pseudomonas solanacearum, a bacterial pathogen of both hosts, was observed. Currently, little is known of the potential for controlling fungal-incited diseases with such strategies. 112 Molecular Approaches to Identifying Fusarium Wilt Resistance

Utsumi et al. (1988) recently cloned a gene for detoxifying fusaric acid from the fungus Cladosporium werneckii in Escherichia coli. When they treated either tomato callus or cuttings with the transformed bacterium prior to treatment with fusaric acid, callus viability was dramatically increased and wilting in cuttings was reduced. As previously mentioned, fusaric acid has been used to screen various plants for resistance to fusarium wilt. Although it has not been used successfully to identify fusarium wilt- resistant banana somaclones (Epp 1987; Krikorian 1990), transformation with this, or other fusaric acid-detoxifying genes, should be attempted if gene-transfer protocols for banana become available in the future. Although conceptually intriguing, transformation of banana for fusarium wilt resistance will require much more work before it becomes possible. Certainly, much could be learned from successes with other crops, and it is hoped that some of these applications will be used in the future to control this important disease.

References

ALVES EJ, SHEPHERD K, DANTAS JLL. 1987. Cultivation of bananas and plantains in Brazil and needs for improvement. Pages 44-49 in Banana and Plantain Breeding Strategies (Persley G, De Langhe E, eds). ACIAR Proceedings no.21. Canberra, Australia: ACIAR. ARCIONI S, PEZZOTTI M, DALIANI F. 1987. In vitro selection of alfalfa plants resistant to Fusarium oxysporum f.sp. medicaginis. Theoretical and Applied Physics 74:700-705. BAKRY F, HAÏCOUR R, HORRY JP, MEGIA R, ROSSIGNOL L. 1993. Applications of biotechnologies to banana breeding: haplogenesis, plant regeneration, and transformation. (This volume.) BARTNICKI-GARCIA S. 1968. Cell wall chemistry, morphogenesis, and taxonomy of fungi. Annu. Rev. Microbiol. 22:87-108. BECKMAN CH. 1987. The Nature of Wilt Diseases of Plants. St Paul, MN, USA: APS Press. 175 pp. BECKMAN CH. 1989. Colonization of the vascular system of plants by fungal wilt pathogens: a basis for modeling the interactions between host and parasite in time and space. Pages 19-32 in Vascular Wilt Diseases of Plants: Basic Studies and Control (Tjamos EC, Beckman CH, eds). Heidelberg, Germany: Springer-Verlag. BECKMAN CH. 1990. Host responses to the pathogen. Pages 93-105 in Fusarium Wilt of Banana (Ploetz RC, ed.). St Paul, MN, USA: APS Press/Amer. Phytopathol. Soc. BECKMAN JS, SOLLER M. 1986. Restriction fragment length polymorphisms in plant genetic improvement. Oxford Survey. Plant Mol. Cell Biol. 3:197-250. BEHARE J, LATERROT H, SARFATTI M, ZAMIR D. 1991. Restriction fragment length polymorphism mapping of the Stemphylium resistance gene in tomato. Mol. Plant-Microb. Interact. 4:489-492. BONNER JE, WARNER RM, BREWBAKER JL. 1974. A chemosystematic study of Musa cultivars. HortScience 9:325- 328. BROGLIE K, CHET I, HOLLIDAY M, CRESSMAN R, BIDDLE P, KNOWLTON S, MAUVIS CJ, BROGLIE R. 1991. Transgenic plants with enhanced resistance to the fungal pathogen Rhizoctonia solani. Science 254:1194-1197. BUDDENHAGEN IW. 1990. Banana breeding and fusarium wilt. Pages 107-113 in Fusarium Wilt of Banana (Ploetz RC, ed.). St Paul, MN, USA : APS Press/Amer. Phytopathol. Soc. BUIATTI M, SCALA A, BETTINI P, NASCARI G, MORPURGO R, BOGANI P, PELLIGRINI MG, GIMELLEI F, VENTURO R. 1985. Correlations between in vivo resistance to Fusarium and in vitro response to fungal elicitors and toxic substances in carnation. Theor. Appl. Genet. 70:42-47. BUIATTI M, MARCHESCHI G, VENTURO R, BETTINI P, BOGANI P, MORPURGO R, NACMIAS B, PELLIGRINI MG. 1987a. In vitro response to Fusarium elicitor and toxic substances in crosses between resistant and susceptible carnation cultivars. Plant Breed. 98:346-348. BUIATTI M, SIMETI C, VANNINI G, MARCHESCHI G, SCALA A, BETTINI P, BOGANI P, PELLIGRINI MG. 1987b. Isolation of tomato cell lines with altered response to Fusarium cell wall components. Theor. Appl. Genet. 75:37-40. RC Ploetz 113

CHET I, BROGLIE K, BROGLIE R. 1991. The role of lytic enzymes in controlling soilborne plant pathogens. (Abstr.) Symposium NBR 32. XIIth Int. Plant Prot. Cong. Rio de Janeiro. DALE J, BURNS T, OEHLSCHLAGER S, KARAN M, HARDING R. 1993. Banana bunchy top virus: prospects for control through biotechnology. (This volume.) DE WAELE D. 1993. Potential of gene transfer for engineering resistance against nematode attack. (This volume.) EPP MD. 1987. Somaclonal variation in bananas: a case study with Fusarium wilt. Pages 140-150 in Banana and Plantain Breeding Strategies (Persley G, DeLanghe E, eds). ACIAR Proceedings no. 21. Canberra, Australia: ACIAR. FAUQUET CM, BEACHY RN. 1993. Status of the coat protein-mediated resistance and its potential application for banana viruses. (This volume.) FITCH MMM, MANSHARDT RM, GONSALVES D, SLIGHTOM JL, SANFORD JC. 1990. Stable transformation of papaya via microprojectile bombardment. Plant Cell Rep. 9:189-194. GEGENBACH BG, GREEN CE, DONOVAN CN. 1978. Inheritance of selected pathotoxin resistance in maize plants regenerated from cell cultures. Proceedings of the National Academy of Science (USA) 74:5113-5117. GONZÁLEZ DE LEÓN D, FOURÉ S. 1993. Genetic mapping of the banana diploid genome. (This volume.) GRAY LE, QUAN YQ, WIDHOLM JM. 1986. Reaction of soybean callus to culture ﬁltrates of Phialophora gregata. Plant Science 47:45-55. GRIMSLEY N, HOHN T, DAVIES JW, HOHN B. 1987. Agrobacterium-mediated delivery of infectious maize streak virus into maize plants. Nature 325:177-179. HAGIO T, BLOWERS AD, EARLE ED. 1991. Stable transformation of sorghum cell cultures after bombardment with DNA-coated microprojectiles. Plant Cell Rep. 10:265-268. HEATH-PAGLIUSO S, PULLMAN J, RAPPAPORT L. 1988. Somaclonal variation in celery: screening for resistance to Fusarium oxysporum f. sp. apii. Theor. Appl. Genet. 75:446-451. JARRET RL, LITZ RE. 1986a. Determining the interspeciﬁc origins of clones within the ‘Saba’ cooking banana complex. HortScience 21:1433-1435. JARRET RL, LITZ RE. 1986b. Enzyme polymorphism in Musa acuminata Colla. J. Heredity 77:183-188. JONES J, TAYLOR J, GRADY K, MUELER G, SUSLOW T. 1986. Engineering bacterial chitinase genes for crop protection. J. Cell. Biochem. Suppl. 10C, J30, 15. KLEIN TM, WOLF ED, WU R, SANFORD JC. 1987. High velocity microprojectiles for delivering nucleic acids into living cells. Nature 327:70-73. KRIKORIAN AD. 1990. Baseline studies and cell culture studies for use in banana improvement schemes. Pages 127-133 in Fusarium Wilt of Banana (Ploetz RC, ed.). St Paul, MN, USA: APS Press/Amer. Phytopathol. Soc. KUC J. 1985. Expression of latent genetic information for disease resistance in plants. Pages 302-412 in Cellular and Molecular Biology of Plant Stress (Key J, Kosuge T, eds). New York: Alan R. Liss. LAGRIMINI ML, ROTHSTEIN S. 1987. Tissue specificity of tobacco peroxidase isozymes and their induction by wounding and tobacco mosaic virus infection. Plant Physiol. 84:438-442. LARKINS PJ, SCOWCROFT WR. 1983. Somaclonal variation and crop improvement. Pages 289-314 in Genetic Engineering of Plants: An Agricultural Perspective (Kosuge T, Meredith CP, Hollaender A, eds). New York, USA, London, UK: Plenum Press. LINTHORST JM, VAN LOON LC, VAN ROSSUM CMA, MAYER A, BOL JF, VAN ROEKEL JSC, MEULENHOFF EJS, CORNELISSEN BJC. 1990. Analysis of acidic and basic chitinases from tobacco and petunia and their constitutive expression in transgenic tobacco. Mol. Plant-Microbe Interact. 3:252-258. MACE ME, WILSON EM. 1964. Phenol oxidases and their relation to vascular browning in Fusarium-infected banana roots. Phytopathology 54:840-842. MACE ME, VEECH JA, BECKMAN CH. 1972. Fusarium wilt of susceptible and resistant tomato isolines: histochemistry of vascular browning. Phytopathology 62: 651-654. MACHARDY WE, BECKMAN CH. 1981. Vascular wilt fusaria: Infection and pathogenesis. Pages 365-390 in Fusarium: Diseases, Biology and Taxonomy (Nelson PE, Tousson TA, Cook RJ, eds). Pennsylvania State University Press. MANSHARDT RM. 1992. Papaya. Pages 489-511 in Biotechnology of Perennial Fruit Crops (Hammerschlag FA, Litz RE, eds). Wallingford, Oxon, UK: CAB International. 114 Molecular Approaches to Identifying Fusarium Wilt Resistance

MACCOMB JA, HINCH JM, CLARKE AE. 1987. Expression of field resistance in callus tissue inoculated with Phytophthora cinnamomi. Phytopathology 77:346-351. MEDINA FHP, STEVENS MA. 1980. Tomato breeding for nematode resistance: survey of resistant varieties for horticultural characteristics and genotype of acid phosphotase. Acta Hort. 100:383-393. MÉGNÉGNEAU B, BRANCHARD M. 1988. Toxicity of fusaric acid observed on callus cultures of various Cucumis melo genotypes. Plant Physiol. Biochem. 26:585-588. MÉGNÉGNEAU B, BRANCHARD M. 1991. Effects of fungal culture filtrates on tissue from susceptible and resistant genotypes of muskmelon to Fusarium oxysporum f.sp. melonis. Plant Sci. 79:105-110. MICHELMORE RW, PARAN I, KESSELL RV. 1991. Identification of markers linked to disease resistance genes by bulked segregant analysis: a rapid method to detect markers in specific genomic regions using segregating populations. Proc. Natl. Acad. Sci. (USA) 88:9828-9832. MUELLER WC, BECKMAN CH. 1978. Ultrastructural localisation of polyphenoloxidase and peroxidase in roots and hypocotyls of cotton seedlings. Can. J. Bot. 56:1579-1587. MURFETT J, CLARKE A. 1987. Producing disease-resistant Musa cultivars by genetic engineering. Pages 87-94 in Banana and Plantain Breeding Strategies (Persley G, De Langhe E, eds). ACIAR Proceedings no.21. Canberra, Australia: ACIAR. NITZSCHE W. 1983. (Title not known.) Theor. Appl. Genet. 65:171-172. NOVAK FJ. 1992. Musa biotechnology. Pages 449-488 in Biotechnology of Perennial Fruit Crops (Hammerschlag FA, Litz RE, eds). Wallingford, Oxon, UK: CAB International. NOVAK FJ, AFZA R, VAN DUREN M, PEREA-DALLOS M, CONGER BV, TANG X. 1989. Somatic embryogenesis and plant regeneration in suspension cultures of dessert (AA and AAA) and cooking (ABB) bananas (Musa spp.). Biotechnology 7:154-159. PEGG GF. 1985. Life in a black hole: The micro-environment of the vascular pathogen. Trans. Brit. Mycol. Soc. 85:1-20. PLOETZ RC. 1990. Roundtable discussions. Pages 135-138 in Fusarium Wilt of Banana (Ploetz RC, ed.). St Paul, MN, USA: APS Press/Amer. Phytopathol. Soc. PLOETZ RC, HERBERT J, SEBASIGARI K, HERNANDEZ JH, PEGG KG, VENTURA JA, MAYATO LS. 1990. Importance of fusarium wilt in different banana-growing regions. Pages 9-26 in Fusarium Wilt of Banana (Ploetz RC, ed.). St Paul, MN, USA: APS Press/Amer. Phytopathol. Soc. POTRYKUS I. 1991. Gene transfer to plants: assessment of published approaches and results. Annu. Rev. Plant Physiol. Plant Mol. Biol. 42:205-225. RICK CM, FOBES JF. 1974. Association of an allozyme with nematode resistance. Rep. Tomato Genet. Coop. 24:25. ROBINSON RW, PROVVIDENTI R, SCHROEDER WT. 1970. A marker gene for tobacco mosaic resistance. Rep. Tomato Genet. Coop. 20:55-56. ROWE PR. 1990. New genetic combinations in breeding bananas and plantains resistant to disease. Pages 114- 123 in Identification of Genetic Diversity in the Genus Musa: Proceedings of an international workshop held at Los Baños, Philippines, 5-10 September 1988 (Jarret RL, ed.). Montpellier, France: INIBAP. ROWE PR, ROSALES FE. 1993. Breeding cooking bananas for areas with marginal growing conditions by using Cardaba (ABB) in cross-pollinations. (This volume.) SANFORD JC, KLEIN TM, WOLF ED, ALLEN N. 1987. Delivery of substances into cells and tissues using a particle bombardment process. J. Part. Sci. and Tech. 5:27-37. SARFATTI M, KATAN J, FLUHR R, ZAMIR D. 1989. An RFLP marker in tomato linked to the Fusarium oxysporum resistance gene I2. Theor. Appl. Genet. 78:755-759. SCHLUMBAUM A, HIGHTOWER RC, MEAGHER RB. 1986. Plant chitinases are potent inhibitors of fungal growth. Nature 324:365-367. SHAHIN EA, SPIVEY R. 1986. A single dominant gene for Fusarium wilt resistance in protoplast-derived tomato plants. Theor. Appl. Genet. 73:164-169. SHEPHERD K, DANTAS JLL, ALVES EJ. 1987. Banana breeding in Brazil. Pages 78-83 in Banana and Plantain Breeding Strategies (Persley G, De Langhe E, eds). ACIAR Proceedings no.21. Canberra, Australia: ACIAR. STOVER RH. 1962. Fusarial Wilt (Panama Disease) of Bananas and other Musa species. Kew, Surrey, UK: Commonwealth Mycological Institute. 117 pp. RC Ploetz 115

STOVER RH, BUDDENHAGEN IW. 1986. Banana breeding: polyploidy, disease resistance and productivity. Fruits 41:175-191. TANKSLEY SD, YOUNG ND, PATERSON AH, BONIERBALE MW. 1989. RFLP mapping in plant breeding: new tools for an old science. Bio/Technol. 7:257-264. TOYODA H, HORIKOSHI K, YAMANO Y, OUCHI S. 1991. Selection for fusarium wilt disease resistance from regenerants derived from leaf callus of strawberry. Plant Cell Rep. 10:167-170. UTSUMI R, HADAMA T, NODA M, TOYODA H, HASHIMOTO H, OHUCHI S. 1988. Cloning of fusaric acid-detoxifying gene from Cladosporium werneckii: a new strategy for the prevention of plant diseases. J. Biotechnol. 8:311- 316. VAKILI, N.G. 1965. Fusarium wilt resistance in seedlings and mature plants of Musa species. Phytopathology 55:135-140. VALMAYOR RV. 1987. Banana improvement imperatives: The case for Asia. Pages 44-49 in Banana and Plantain Breeding Strategies (Persley G, De Langhe E, eds). ACIAR Proceedings no.21. Canberra, Australia: ACIAR. VANDER MOLEN GF, BECKMAN CH, RODEHORST E. 1977a. Vascular gelation: a general response phenomenon following infection. Physiol. Plant Pathol. 11:95-100. VANDER MOLEN GF, LABAVITCH JM, STRAND LL, DEVAY JE. 1977b. Pathogen-induced vascular gels: ethylene as a host intermediate. Physiol. Plant. 59:573-580. VARDI A, EPSTEIN E, BRIEMAN A. 1986. Is the Phytophthora citrophthora culture ﬁltrate a reliable tool for the in vitro selection of resistant Citrus variants? Theoretical and Applied Genetics 72:569-574. VUYLSTEKE D, SWENNEN R. 1993. Genetic improvement of plantains: the potential of conventional approaches and the interface with in-vitro culture and biotechnology. (This volume.) WARDLAW CW. 1930. The biology of banana wilt (Panama disease). III. An examination of sucker infection through root bases. Ann. Bot. 45:381-399. 116 Potential of Gene Transfer for Engineering Resistance against Nematode Attack

Potential of Gene Transfer for Engineering Resistance against Nematode Attack

D De Waele

Introduction Genetically engineered plants expressing completely new traits are no longer a dream of the future. Recent advances in recombinant DNA and tissue-culture technology have made it possible successfully to transfer genes from any organism, be it a microorganism, plant, or animal, to higher plants. The performances of these “new” so-called transgenic plants are already being evaluated. The first results, covering a variety of applications, are promising. Gene transfer has also a great potential for engineering resistance against nematode attack in agricultural crops. Nematologists, however, have been surprisingly slow in recognizing the potential of this new technology. As a result of this lack of interest, the development of a genetically engineered nematode-resistant plant is far behind the engineering in plants of such traits as virus and insect resistance. Fortunately, this situation is changing. During the last 2-3 years, several research groups in Europe and the USA, in the public as well as in the private sector, have initiated albeit modest R&D programs aimed at engineering nematode resistance in plants. The present paper consists of three major parts. First, the present status of engineering new traits in plants is summarized. This information should demonstrate to the reader that this new technology is working. Secondly, it is shown that plant genetic engineering offers an interesting alternative for managing nematodes on bananas (= bananas and plantain throughout) in view of the existing limitations of the traditional management methods used. It is also shown that this technology is complementary to conventional plant breeding. Thirdly, the major strategies that are currently being developed to engineer nematode resistance in plants are discussed. This discussion permits the identiﬁcation of the different steps in the development of a program aimed at the genetic engineering of nematode resistance in bananas. It should then become clear that a lack of vital, mostly basic knowledge is constraining the development of such a program. It is hoped that the presented information will offer the reader a good idea of the type of research that can and should be initiated in order to obtain this knowledge.

Plant Genetic Systems NV, Jozef Plateaustraat 22, 9000 Gent, Belgium D De Waele 117

Status of Engineering New Traits in Plants Since 1983, when the first successful transformation of a plant cell with a foreign gene and the regeneration of a fertile plant from this cell was reported, about 30 different plant species have been transformed (Fraley 1992). The list of transgenic plants includes such important horticultural and field crops as cucumber, eggplant, lettuce, melon, pea, strawberry, tomato, alfalfa, cotton, maize, oilseed rape, potato, rice, soybean, sugar beet, sunflower, and tobacco. For most of these plants, Agrobacterium tumefaciens, a plant pathogen causing tumerous crown galls upon infection, was used as the vector to introduce the genes in the plant cells. For plants resistant to Agrobacterium-mediated transformation, including most monocotyledonous plants, alternative transformation techniques, such as electroporation, polyethylene-mediated transfer, microinjection, and particle bombardment, were developed (Uchimiya et al. 1989). Today, transgenic plants resistant to viruses (Beachy et al. 1990), bacteria (Anzai et al. 1989), fungi (Broglie et al. 1991), insects (Brunke, Meeusen 1991), and herbicides (Botterman, Leemans 1988) exist. Also available are transgenic plants that produce higher amounts of nutritious proteins (Altenbach, Simpson 1990) or new biologically active peptides used in the pharmaceutical industry (Krebbers, Vandekerckhove 1990). Hybrid seeds can now be obtained through genetic engineering of nuclear male sterility (Mariani et al. 1990) while the ripening of fruit can be controlled (Oeller et al. 1991) and the color of flowers can be changed (Meyer et al. 1987) through the genetic engineering of antisense genes. The agronomic performances of many of these transgenic plants and the expression of some of the new traits are being evaluated in field trials throughout the world. The cumulative total of field trials with genetically engineered organisms (also including other transgenic organisms) was estimated at about 500 by the end of 1991. Commercialization of the first transgenic plants is expected in the second half of the 1990s. However, in addition to a number of remaining technical limitations, regulatory approval and public perception will also be important factors determining the future use of transgenic plants.

Genetic Engineering and Conventional Plant Breeding In contrast with what is sometimes postulated, plant genetic engineering will not replace conventional plant breeding. On the contrary, it provides the breeder with a novel and powerful tool to overcome some limitations inherent in conventional plant breeding. In traditional plant breeding, genes are transferred among sexually compatible individuals from the same or closely related species. The source of resistance is limited; the genetic basis is rather narrow. Through genetic engineering it is possible to bring genes into a plant from any organism. The resistance sources are unlimited; the genetic basis is wide. In traditional plant breeding, traits from wild plant varieties are transferred into a commercial plant variety without knowing the genetic or molecular basis responsible for this trait. In genetic engineering, characterized genes, coding for well characterized gene 118 Potential of Gene Transfer for Engineering Resistance against Nematode Attack

products, are transferred. Since the entire genomes of both parents are combined, conventional plant breeding is time-consuming. Since in genetic engineering the gene is transferred in “one step”, while other characters of the recipient plant are in principle not affected, the amount of backcrossing needed is considerably reduced. A good illustration of the unlimited and well characterized transfer of genes inherent in genetic engineering is the successful development of tobacco plants resistant to the disease wildfire (Anzai et al. 1989). Wildfire is caused by a toxin produced by Pseudomonas syringae pathovar tabaci upon infection. The resistance was obtained by engineering in the tobacco plants a detoxifying enzyme, derived from the pathogen itself. Through genetic engineering it will be possible to add new traits to superior plant varieties obtained from a conventional plant breeding program.

Nematode Pests of Bananas In most regions of the world, nematodes are recognized as important pests of bananas (Gowen, Quénéhervé 1990). Crop losses caused by nematodes to bananas are very high. Average annual yield losses due to damage by nematodes on bananas are estimated at 19.7% worldwide, representing an estimated annual monetary loss of about US$178 million based on 1984 commodity prices (Sasser, Freckman 1987). In Africa, estimated crop losses vary from 20 to 80% (Sarah 1989). In Central America, reductions of 50% in fruit yields and an increase of 60% in lodging have been observed (Roman 1986). In general, three types of plant-parasitic nematodes can be distinguished. The ectoparasitic nematodes live completely outside the plant and pierce the outermost plant cell layers with their stylet in order to feed. The migratory endoparasitic nematodes live completely inside the plant tissues but are able to move freely between the root and the soil. They feed on normal plant cells inside the plant. Eggs are laid either outside or inside the plant. Migratory endoparasitic nematodes have no complex interaction with their host. The sedentary endoparasitic nematodes also live inside the plant but the adult female becomes sedentary (immobile). Eggs are laid outside the plant. Sedentary endoparasites have a complex interaction with their host. They induce normal plant cells to form specialized feeding structures. These structures serve as food transfer cells and are vital for the survival of the nematodes. The most damaging and widespread nematodes attacking bananas are migratory endoparasites: the burrowing nematode Radopholus similis (Cobb) Thorne, the root-lesion nematodes Pratylenchus coffeae (Zimmerman) Filipjev & Schuurmans Stekhoven and P. goodeyi Sher & Allen, and the spiral nematode Helicotylenchus multicinctus (Cobb) Golden (Gowen, Quénéhervé 1990). The migratory behavior of the migratory endoparasites in the cortex of bananas, feeding upon and destroying the parenchymous cells, causes the formation of cavities and, consequently, the disruption of the root system. Since, in bananas, the root system is not only vital for absorption and transportation of water and solutes but also for providing a ﬁrm anchorage in the soil, the presence of these nematodes often results in toppling of the bananas. The symptoms caused by these nematodes are very similar. The roots show dark, D De Waele 119

brown or red, lesions. The above-ground symptoms are characterized by dwarfing, fewer and smaller leaves, leaf chlorosis, premature defoliation, small bunches with reduced weight, and lodging of plants. The sedentary endoparasitic root-knot nematodes Meloidogyne incognita (Kofoid & White) Chitwood, M. javanica (Treub) Chitwood, M. arenaria (Neal) Chitwood, and M. hapla Chitwood also occur in banana corms and roots (Gowen, Quénéhervé 1990). Juveniles (J2) emerge from the egg, invade a root and induce speciﬁc feeding structures, the so-called giant cells, which progress within the vascular cylinder, and also induce the cortical cells to multiply and form galls. The J2 swell and molt three times to form the females which enlarge rapidly and become immobile and globose. The adult female produces several hundred eggs which are laid via a canal outside the host tissue. Root-knot nematodes cause root injury while restriction of the total water ﬂow through the infected roots by the giant cells results in stress and a reduction in the uptake of water and nutrients. Infected plants show stunted root growth. The most obvious symptoms are galling on primary and secondary roots. In addition to the damage directly caused by nematodes, nematode-induced root lesions create a food base for weak, unspecialized fungal parasites of banana (Fusarium spp., Cylindrocarpon spp., Rhizoctonia spp.), enabling the fungi to invade the stele and to increase the amount of root necrosis. Aggravation of fusarium wilt or Panama disease, one of the most devastating diseases affecting bananas, caused by Fusarium oxysporum f.sp. cubensis in the presence of nematodes, has been reported (Newhall 1958; Loos 1959).

Limitations of Traditional Nematode Management in Bananas Where nematode parasitism of bananas cannot be prevented by using clean planting material in nematode-free soil and growing the plants under strict quarantine conditions (which is a difficult task since nematodes are present in every agricultural field and disperse easily), nematode management on bananas is mainly based on crop rotation and chemical control (Gowen, Quénéhervé 1990). However, the successful application of both methods is severely restricted. In those areas where bananas are grown continuously, crop rotation is useless while chemical control is usually also not applied when traditional mixed cropping systems are being used. Not only is the high price of nematicides often prohibitive for the small farmer, but most nematicides are also extremely toxic for nontarget organisms, including the user, and may pollute the environment. In the USA and Europe the detection of DBCP, EDB, D-D, carbofuran, and aldicarb in groundwater has resulted in the suspension or restriction of the use of these nematicides (Thomason 1987). This example is being followed by more and more governments in developing countries. Recently, even a ﬁrm that produces a nematicide (Rhône-Poulenc) has initiated a worldwide suspension on the sale and use of one of its products (aldicarb) on bananas, following the detection of higher-than-expected residue levels in one of five trials being conducted in Central and South America (Agrow, 14 June 1991). 120 Potential of Gene Transfer for Engineering Resistance against Nematode Attack

No banana variety speciﬁcally resistant to nematodes has been released so far. Although sources of resistance, be it only against Radopholus similis, are known, the development of a nematode-resistant banana variety is extremely difficult because of the genetic complexity of this crop, its low fertility, and the long period required for the nematological evaluation of crossings (Pinochet 1988). In view of the limitations of the traditional management methods used, genetic engineering offers an attractive alternative for managing nematodes on bananas.

Strategies for Engineering Nematode Resistance in Plants Since the existent recombinant DNA technology is not suited for the engineering of complex pathways in plants, strategies for engineering nematode resistance in plants are being focused on proteins which can be engineered by transferring a single gene, coding directly for the protein. The strategies for engineering nematode resistance in plants that are currently being developed can be classiﬁed as: (1) strategies aimed at the direct killing of the nematode; (2) strategies aimed at using or reinforcing existing plant defense mechanisms; and (3) strategies aimed at interfering in the compatible nematode-plant interaction. Most of these strategies are being developed. No transgenic plants resistant to nematodes have been obtained so far. In the ﬁrst strategy, external or internal nematode structures are targeted with either lytic enzymes or oral toxic proteins. Examples of lytic enzymes are the chitinases and collagenases. Since chitin is an important component of the nematode egg shell (Bird 1976) and collagen of the cuticle of juvenile and adult nematodes (Reddigari et al. 1986), it is hoped that the expression of these lytic enzymes in plants will lyse the external nematode structures and kill the nematodes (Havstad et al. 1991). The expression of chitinases in plants will work only against the eggs of migratory endoparasitic nematodes and not against the eggs of ectoparasitic nematodes, which do not enter the roots, or the eggs of the sedentary endoparasitic nematodes which are deposited outside the roots. The expression of collagenases will of course not work against ectoparasitic nematodes. The development of this strategy is seriously hampered by the incomplete knowledge of the precise nature of both the egg shell and the cuticle of plant-parasitic nematodes. This lack of basic knowledge contrasts sharply with the data currently available on the same structures in animal-parasitic nematodes. A good example of the search for oral toxic proteins is the screening for toxins produced by Bacillus thuringiensis. This gram-positive soil bacterium is better known worldwide for its insecticidal crystal proteins, which have been used successfully as bio-insecticides and to engineer insect resistance in plants (Lambert, Peferoen 1992). Bacillus thuringiensis isolates with (ovicidal and larvicidal) activity against nematodes have been found, but this activity was almost exclusively effective against free-living and animal-parasitic nematodes D De Waele 121

(Bottjer et al. 1985; Meadows et al. 1990). Only some nematocidal activity against Pratylenchus scribneri Steiner, a migratory endoparasite, has been reported so far (Bradfisch et al. 1991). The molecular basis of the observed ovicidal and larvicidal activity is unknown. Screening of B. thuringiensis isolates for oral toxic proteins active against plant-parasitic nematodes is seriously hampered by the obligate biotrophic way of feeding of most plant-parasitic nematodes. This implies that plant-parasitic nematodes feed only upon living plant cells. The cloning of genes conferring resistance to nematodes forms the subject of the second strategy. Since the recombinant DNA technology currently available permits only the transfer of traits encoded by single genes, only single, dominant resistance genes are being cloned. Examples of such genes are the Mi gene from Lycopersicon peruvianum which is effective against all Meloidogyne species attacking tomato except M. hapla (Williamson et al. 1991). The molecular basis of the resistance is unknown. Several routes to clone these resistance genes can be followed (North 1990). Mapping of the genes relative to polymorphic markers followed by chromosome walking is favored by most researchers. Cloning of resistance genes is a time-consuming activity. Because only resistance genes conferring resistance to sedentary endoparasitic nematodes have been identified so far, this strategy is not applicable to ectoparasitic and migratory endoparasitic nematodes species. The third strategy aims at interfering in the nematode-plant interaction. Although this strategy includes such activities as the search of repellants and of proteins blocking the nematode-plant recognition, most research efforts are aimed at the inhibition or destruction of the feeding structures induced by the sedentary endoparasitic nematodes. This limits eventual application to these nematodes. The expression in plants of antibodies, the so-called plantibodies, directed against proteins secreted by the nematodes, and which are essential for the induction and/or maintenance of the feeding structures, is one example of a route followed to disrupt the feeding structures (Schots et al. 1991). This method is based on the assumption that expression of these antibodies will inhibit the biological activity of the proteins. Another route followed is based on the idea that the sedentary endoparasitic nematodes specifically induce genes that are switched on only in the feeding structures. Induction of genes in nematode-infected plant roots has been reported (Gurr et al. 1991). The regulatory sequences of such genes could be used to direct cytotoxic proteins to the feeding structures. The development of both methods is hampered by a lack of basic knowledge on the precise function of the secreted proteins on the one hand and the specificity of the induced genes on the other.

Steps in the Development of a Program aimed at the Genetic Engineering of Nematode Resistance in Bananas The first step in the development of a program aimed at the genetic engineering of resistance against nematode attack in any crop, and thus also in bananas, is the identiﬁ- 122 Potential of Gene Transfer for Engineering Resistance against Nematode Attack

cation of the most important nematode species. The nematode life cycle, feeding behavior, and interaction with other pathogens and the host plant will all have a decisive influence on the choice of the strategy to be followed. Also, the ease with which these nematodes can be cultured in order to obtain living specimens for bioassays will be of importance. There is sufficient information that at least three distinct pathotypes of R. similis occur in Central America and the Caribbean (Pinochet 1979; Tarté et al. 1981). Information on the existence of other pathotypes outside this region or of pathotypes of the other migratory endoparasitic nematodes is completely lacking but necessary for the correct identification of the target nematodes. Bananas, as well as most other subtropical and tropical crops, are often parasitized simultaneously by several nematode species (Gowen, Quénéhervé 1990). This implies that the genetic engineering of resistance to only one of these species may result in the increased occurrence of one or more of the other nematode species against which the resistant genes offer no protection. The strategy should thus be aimed at engineering global resistance. The second step in the development of a program aimed at the genetic engineering of nematode resistance in plants is the identification and purification of the protein on which the chosen strategy will be based and the cloning of the gene that encodes this protein. There are many biological substances that show activity against nematodes, but most of these products are secondary metabolites. Secondary metabolites are usually the products of complex biochemical pathways involving many different proteins. Engineering nematode resistance in a plant with a secondary metabolite would involve the cloning of a complete metabolic pathway. For this task, the existent recombinant DNA technology is not suited. Therefore, strategies for engineering nematode resistance in plants must focus on proteins that can be engineered by transferring a single gene, coding directly for the protein. Very few of these proteins have been identified so far. In their quest for biological control of nematodes, nematologists have identified a relatively large number of bacteria and, especially, fungi showing activity against plant- parasitic nematode species (Kerry 1990). Unfortunately, in most cases, the molecular basis for this activity has not been investigated. Once a protein has been identified, the encoding gene has to be cloned. A clone consists of multiple copies of identical DNA sequences that are produced when they are inserted and replicated in cloning vehicles using recombinant DNA technology. The gene must then be transferred to cells of bananas, using one of the transformation techniques, and the transformed cells regenerated. Finally, the transgenic bananas must be thoroughly evaluated on resistance to the target nematodes and on agronomic performance.

Constraints for the Engineering of Nematode Resistance in Bananas Major constraints obstructing the development of virtually every program aimed at the engineering of nematode resistance in bananas are: (1) insufficient knowledge on the D De Waele 123

occurrence of pathotypes or races of the migratory, especially R. similis, and sedentary nematodes parasitizing bananas; (2) the obligate biotrophic way of feeding of all banana nematodes; (3) the primitive level of nematode-plant interaction of most banana nematodes, except the Meloidogyne species; (4) the rather high difﬁculty of establishing monoxenic cultures of banana nematodes; and (5) the occurrence of multispecies infections on the same banana. In addition to these constraints, there is a general lack of basic knowledge of many external and internal structures and processes of plant-parasitic nematodes and virtually no information on the molecular and genetic basis of most nematode-plant interactions.

Conclusion All indications are that recombinant DNA technology, as a complementary tool to conventional plant breeding, will have a major and beneﬁcial impact on the agriculture of tomorrow. Although genetic engineering offers certainly no cure-all for obtaining nematode-resistant crops, opportunities to use this technology also for obtaining nematode- resistant bananas, be it against one or more nematode species simultaneously, should not be neglected.

References

ALTENBACH SB, SIMPSON RB. 1990. Manipulation of methionine-rich protein genes in plant seeds. Trends in Biotechnology 8:156-160. ANZAI H, YONEYAMA K, YAMAGUCHI I. 1989. Transgenic tobacco resistant to a bacterial disease by the detoxiﬁcation of a pathogenic toxin. Molecular and General Genetics 219:492-494. BEACHY RN, LOESCH-FRIES S, TUMER NE. 1990. Coat protein-mediated resistance against virus infection. Annual Review of Phytopathology 28:451-474. BIRD AF. 1976. The development and organization of skeletal structures in nematodes. Pages 107-137 in The Organization of Nematodes (Croll NA, ed.). New York, USA: Academic Press. BOTTERMAN J, LEEMANS J. 1988. Engineering herbicide resistance in plants. Trends in Genetics 4:219-222. BOTTJER KP, BONE LW, GILL SS. 1985. Nematoda: susceptibility of the egg to Bacillus thuringiensis toxins. Experimental Parasitology 60:239-244. BRADFISCH GA, HICKLE LA, FLORES R, SCHWAB G. 1991. Nematocidal Bacillus thuringiensis toxins: opportunities in animal health and plant protection. Abstract presented at the First International Conference on Bacillus thuringiensis held in Oxford, UK. BROGLIE K, CHET I, HOLLIDAY M, CRESSMAN R, BIDDLE P, KNOWLTON S, JEFFREY MAUVAIS C, BROGLIE R. 1991. Transgenic plants with enhanced resistance to the fungal pathogen Rhizoctonia solani. Science 254:1194-1197. BRUNKE KJ, MEEUSEN RL. 1991. Insect control with genetically engineered crops. Trends in Biotechnology 9:197- 200. FRALEY R. 1992. Sustaining the food supply. Bio/Technology 10:40-43. HAVSTAD P, SUTTON D, THOMAS S, SENGUPTA-GOPALAN C, KEMP J. 1991. Collagenase expression in transgenic plants: an alternative to nematicides. Abstract 345 presented at the Third International Congress of Plant Molecular Biology held in Tucson, USA. GOWEN S, QUENHERVE P. 1990. Nematode parasites of bananas, plantains and abaca. Pages 431-460 in Plant Parasitic Nematodes in Subtropical and Tropical Agriculture (Luc M, Sikora RA, Bridge J, eds). Wallingford, Oxon, UK: CAB International. GURR SJ, MCPHERSON MJ, SCOLLAN C, ATKINSON HJ, BOWLES DJ. 1991. Gene expression in nematode-infected plant roots. Molecular and General Genetics 226:361-366. 124 Potential of Gene Transfer for Engineering Resistance against Nematode Attack

KERRY B. 1990. An assessment of progress towards microbial control of plant-parasitic nematodes. Supplement to Journal of Nematology 22(4S):621-631. KREBBERS E, VANDEKERCKHOVE J. 1990. Production of peptides in plant seeds. Trends in Biotechnology 8:1-3. LAMBERT B, PEFEROEN M. 1992. Insecticidal promise of Bacillus thuringiensis. BioScience 42:112-122. LOOS CA. 1959. Symptom expression of Fusarium wilt disease of the Gros Michel banana in the presence of Radopholus similis (Cobb, 1893) Thorne, 1949 and Meloidogyne incognita acrita Chitwood, 1949. Proceedings of the Helminthological Society of Washington 26:103-111. MARIANI C, DE BEUCKELEER M, TRUETTNER J, LEEMANS J, GOLDBERG RB. 1990. Induction of male sterility in plants by a chimaeric ribonuclease gene. Nature 347:737-741. MEADOWS J, GILL SS, BONE LW. 1990. Bacillus thuringiensis strains affect population growth of the free-living nematode Turbatrix aceti. Invertebrate Reproduction and Development 17:73-76. MEYER P, HEIDMANN I, FORKMANN G, SAEDLER H. 1987. A new Petunia ﬂower colour generated by transformation of a mutant with a maize gene. Nature 330:677-678. NEWHALL AG. 1958. The incidence of Panama disease of banana in the presence of the root knot and burrowing nematodes (Meloidogyne and Radopholus). Plant Disease Reporter 42:853-856. NORT G. 1990. The race for resistance genes. Nature 347:517. OELLER PW, MIN-WONG L, TAYLOR LP, PIKE DA, THEOLOGIS A. 1991. Reversible inhibition of tomato fruit senescence by antisense RNA. Science 254:437-439. PINOCHET J. 1979. Comparison of four isolates of Radopholus similis from Central America on Valery Bananas. Nematropica 9:40-43. PINOCHET J. 1988. Comments on the difﬁculty in breeding bananas and plantains for resistance to nematodes. Revue de Nématologie 11:3-5. REDDIGARI SR, JANSMA PL, PREMACHANDRAN D, HUSSEY RS. 1986. Cuticular collagenous proteins of second-stage juvenile and adult females of Meloidogyne incognita; isolation and partial characterization. Journal of Nematology 18:294-302. ROMAN J. 1986. Plant-parasitic nematodes that attack bananas and plantain. Pages 6-19 in Plant-Parasitic Nematodes of Bananas, Citrus, Coffee, Grapes and Tobacco. Union Carbide Agricultural Products Company. SARAH JL. 1989. Banana nematodes and their control in Africa. Nematropica 19:199-216. SASSER JN, FRECKMAN DW. 1987. A world perspective on nematology: the role of the Society. Pages 7-14 in Vistas on Nematology (Veech JA, Dickson DW, eds). Society of Nematologists, Inc. SCHOTS A, ROOSIEN J, DE BOER JM, SCHOUTEN A, OVERMARS HA, ZILVERENTANT JF, POMP H, BOUWMAN-SMITS L, VISSER B, STIEKMA WJ, GOMMERS F, BAKKER J. 1991. Plantibodies: a flexible approach to design resistance against pathogens. International Symposium on Advances in Potato Crop Protection, Wageningen, The Netherlands: 39 (abstract). TARTÉ R, PINOCHET J, GABRIELLI C, VENTURA O. 1981. Differences in population increase, host preferences and frequency of morphological variants among isolates of the banana race of Radopholus similis. Nematropica 11 : 43-52. THOMASON IJ. 1987. Challenges facing nematology: environmental risks with nematicides and the need for new approaches. Pages 469-476 in Vistas on Nematology (Veech JA, Dickson DW, eds). Society of Nematologists, Inc. UCHIMIYA H, HANDA T, BRAR DS. 1989. Transgenic plants. Journal of Biotechnology 12:1-20. WILLIAMSON VM, HO JY, MA HM, WEIDE R, LANKHORST KLEIN R, LIHARSKA T, ZABEL P. 1991. A high resolution linkage map of the nematode resistance gene region in tomato. Abstract 1710 presented at the Third International Congress of Plant Molecular Biology held in Tucson, USA. N von Mende, P Burrows, J Bridge 125

Molecular Aspects of Resistance to Nematodes

N von Mende, P Burrows, J Bridge

Plant parasitic nematodes are important pests in tropical and subtropical agriculture. These microscopic plant parasites are frequently divided into three major groups according to their migratory and feeding behavior. The first group, the migratory, ectoparasitic nematodes, remain outside the root and insert their stylet into epidermal or meristematic cells to feed. In general they cause little damage to plant tissue unlike the second group, the migratory, endoparasitic nematodes. Endoparasitic nematodes not only kill the cells they feed upon but, by burrowing through the root tissues, they cause extensive destruction leading to cavitation and secondary infection (Dropkin 1980). The third group includes the sedentary, endoparasitic nematodes which have evolved highly sophisticated relationships with their hosts. These nematodes enter roots as vermiform juveniles and induce certain root cells to transform into nurse or transfer cells. The function of the nurse cells is to act as a metabolic sink that provides the sedentary nematodes with a continuous supply of nutrients. This feeding strategy allows the female to swell and produce a large number of eggs (Jones et al. 1981). The nematodes causing the most damage to bananas and plantains are migratory endoparasites, and include Radopholus similis, Pratylenchus goodeyi, P. coffeae, and Helicotylenchus multicinctus. These cause destruction of roots, initially stunting the plant but ultimately causing disruption of the anchorage and toppling of the plant. It is also common to find such sedentary endoparasites as Meloidogyne spp. and Rotylenchulus reniformis parasitizing the root system. In addition to members of these ﬁve genera, over 146 other nematode species are associated with Musa spp. throughout the world (Gowen, Quénéhervé 1990). A variety of measures is used to control the damage by these parasitic nematodes in banana plantations. To prevent the introduction of these pests into new plantations, carefully selected pared suckers (free of nematode damage) are heat-treated or dosed with nematicides before planting. Currently, a more favored approach is to grow nematode-free plants from meristematic tissue. Nematicides are used widely by growers producing fruit for the international export market. Less specialized production serving local markets may not justify the high cost of chemical treatment. There is considerable interest in nonchemical, environmentally friendly, alternative methods of nematode control. Attention has focused on producing nematode-resistant plants through

Research Group on Nematodes in Tropical and Subtropical Agriculture, IACR, Rothamsted Experimental Station, Harpenden, Herts, AL5 2LQ, and CAB International Institute of Parasitology, St Albans, UK 126 Molecular Aspects of Resistance to Nematodes

conventional breeding and modern biotechnological means, and on the use of naturally occurring biological control agents (De Leij, Kerry 1991). Resistance to pests and diseases in crops can be obtained either using available experimental lines/germplasm accessions in conventional breeding programs or through molecular genetic and plant transformation technologies. Molecular techniques can overcome some limiting factors inherent in plant breeding, e.g., selected genes can be inserted into a new host thus bypassing infertility and other breeding barriers. This approach has been used successfully for the production of transgenic plants resistant to viruses (Grumet 1990) and insects (Brunke, Meeusen 1991). Programs have also been initiated recently to produce nematode-resistant transgenic plants. The most important question, however, is which gene to select for insertion into the plant to induce resistance. Two strategies have been considered. Firstly, to exploit the plant’s natural defense mechanisms, and secondly to insert genes which, directly or indirectly, disrupt nematode development. Both of these strategies are discussed below. Some progress has been made in the selection of nematode-resistant banana clones but less headway has been made in identifying genes for resistance. Some clones appear to be resistant to some species of nematode and not others. For example, Pinochet and Rowe (1978, 1979) found two clones of the diploid Pisang Jari Buaya cultivar were resistant to R. similis but not to P. coffeae. In contrast with the lack of information concerning resistance to migratory nematodes, several genes resistant to sedentary endoparasitic nematodes have been identified through conventional plant breeding programs (Cook 1991). There is growing evidence, through extensive studies of plant- nematode interactions, that in many cases a gene-for-gene system exists in which a virulence gene in the nematode has been identified which is matched by a resistance gene in the plant (e.g., Jones et al. 1981). One resistance mechanism is the hypersensitive response involving rapid cell necrosis which stops the nematode from feeding and becoming established on the plant. A few resistance genes have been identiﬁed, such as the Mi gene which confers resistance to Meloidogyne incognita and the H1 gene conferring resistance to the potato cyst nematode Globodera rostochiensis. Current efforts are under way to isolate and characterize these genes (Barone et al. 1990; Klein-Lankhorst et al. 1991). In addition to the hypersensitive response, plants have many other defense mechanisms such as enzyme inhibitors, lytic enzymes, and ligniﬁcation (Lamb et al. 1989; Bowles 1990), but their role in resistance to nematodes has not been elucidated. The aim of the second strategy for producing nematode-resistant plants is to disrupt nematode development. One approach involves the use of antisense constructs which inhibit the translation of specific transcripts essential for host-plant recognition and parasitism. The success of this approach is dependent upon a thorough understanding of host-parasite interactions at the cell and molecular level. Research at Rothamsted Experimental Station is currently under way to identify signal molecules and gene regulation involved in host-recognition and induction of nurse cells. For this purpose endoparasitic nematodes and Arabidopsis thaliana are being used as a model system (Sijmons et al. 1991). A. thaliana is ideally suited for molecular studies because of its N von Mende, P Burrows, J Bridge 127

small genome size, short life cycle, and well developed classical genetics. Furthermore, a wide range of well characterized ecotypes and mutants are available. A further approach is the insertion of functional genes that interfere with the feeding or development of nematodes. So far this strategy has only been developed for insect pests using genes encoding, for example, for the cowpea trypsin inhibitor or toxins produced by Bacillus thuringiensis (Brunke, Meeusen 1991). Also envisaged is the insertion of genes for collagenases or chitinases, which would digest collagen (a major component of nematode cuticle), or chitin (a component of the egg shell), respectively. Some workers are considering incorporating genes encoding for mammalian antibodies with afﬁnity for important receptors or essential structural proteins that are involved in successful parasitism, e.g., salivary proteins or feeding tubes (Hussey 1989). In conclusion, several diverse strategies involving molecular techniques are available to develop nematode-resistant plants. It is highly likely that some of these strategies could be extended to Musa species. Before this can happen in a rational and informed way leading to durable resistance much more needs to be known about communication between nematodes and their respective hosts at the cell and molecular level.

References

BARONE A, RITTER E, SCHACHTSCHABEL U, DEBENER T, SALAMINI F, GEBHARDT C. 1990. Localization by restriction fragment length polymorphism mapping in potato of a major dominant gene conferring resistance to the potato cyst nematode Globodera rostochiensis. Mol. Gen. Genet. 224:177-182. BOWLES DJ. 1990. Defense-related proteins in higher plants.Annu. Rev. Biochem. 59:873-907. BRUNKE KJ, MEEUSEN RL. 1991. Insect control with genetically engineered crops. TIBTECH 9:197-200. COOK R. 1991. Resistance in plants to cyst and root-knot nematodes. Pages 213-240 in Agricultural Zoology Reviews vol.4. DROPKIN VH. 1980. Introduction to Plant Nematology. New York, USA: John Wiley. DE LEIJ FAAM, KERRY BR. 1991. The nematophagous fungus Verticillium chlamydosporium as a potential biological control agent for Meloidogyne arenaria. Revue Nematol. 14:157-164. GOWEN S, QUÉNÉHERVÉ P. 1990. Nematode parasites of bananas, plantains and abaca. Pages 431-460 in Plant Parasitic Nematodes in Subtropical and Tropical Agriculture (Luc M, Sikora A, Bridge J, eds). Wallingford, Oxon, UK: CAB International. GRUMET R. 1990. Genetically engineered plant virus resistance. HortScience 25:508-513. HUSSEY RS. 1989. Disease-inducing secretions of plant-parasitic nematodes. Annu. Rev. Phytopathol. 27:123-141. JONES FGW, PARROTT DM, PERRY JN. 1981. The gene-for-gene relationship and its signiﬁcance for potato cyst nematodes and their Solanaceous hosts. In Plant Parasitic Nematodes vol.III (Zuckerman BM, Rohde RA, eds). London, UK: Academic Press. KLEIN-LANKHORST R, RIETVELD P, MACHIELS B, VERKERK R, WEIDE R, GEBHARDT C, KOORNNEEF M, ZABEL P. 1991. RFLP markers linked to the root knot nematode resistance gene Mi in tomato. Theor. Appl. Genet. 81:661-667. LAMB CJ, LAWTON MA, DRON M, DIXON RA. 1989. Signals and transduction mechanisms for activation of plant defenses against microbial attack. Cell 56:215-224. PINOCHET J, ROWE P. 1978. Reaction of two banana cultivars to three different nematodes. Plant Disease Reporter 62:727-729. PINOCHET J, ROWE P. 1979. Progress in breeding for resistance to Radopholus similis from Central America on Valery Bananas. Nematropica 9:40-43. SIJMONS PC, GRUNDLER FMW, VON MENDE N, BURROWS PR, WYSS U. 1991. Arabidopsis thaliana as a new model host for plant-parasitic nematodes. The Plant Journal 1:245-254. 128 Breeding Cooking Bananas for Areas with Marginal Growing Conditions

Breeding Cooking Bananas for areas with Marginal Growing Conditions by using Cardaba (ABB) in Cross-Pollinations

PR Rowe, FE Rosales

Introduction The ABB cooking bananas (Bluggoe, Saba, Cardaba, and Pelipita) have a high level of resistance to black Sigatoka (Mycosphaerella ﬁjiensis), are resistant to drought, and are tolerant of soil nutritional deficiencies. Of these cultivars, Saba is produced in the greatest amount for a particular area. It is the most important cooking banana in the Philippines where its estimated annual production is over 1.5 million t (Valmayor 1987). Saba and Cardaba were earlier thought to have a BBB genomic composition, but results from an isozyme analysis (Jarret, Litz 1986) indicate that both should be classiﬁed as ABB. All these robust ABB cultivars can be grown in areas with soil and climatic conditions unfavorable for cultivation of the AAB plantains and AAA dessert bananas. Plantains predominate as the preferred starchy banana in Latin America and the Caribbean, and most of the nearly 5 million t of these AAB cooking types produced annually in this region are consumed domestically (Jaramillo 1987). Plantains are even more important in western and central Africa where they account for more than 25% of the carbohydrates in the diets of some 60 million people. Over 60% of the world’s plantains are produced and consumed in this area of Africa (INIBAP 1987). The other major nondessert types are the highland AAA cultivars of eastern Africa which are divided into cooking bananas and beer bananas. For about 20 million people in Burundi, Rwanda, Uganda, northern Tanzania, and eastern Zaire, these cooking bananas (boiled green) are the primary staple food. The annual production of about 9 million t of these two types of bananas (all for domestic consumption) in this area constitutes close to one-half of the African production of bananas and plantains (INIBAP 1986). Before the advent of black Sigatoka in Latin America and Africa, no serious attention was given to breeding new disease-resistant hybrids which could substitute for the AAB and AAA cooking cultivars. Stover and Dickson (1976) described the 1973 black Sigatoka

Fundación Hondureña de Investigación Agrícola (FHIA), Apto Postal 2067, San Pedro Sula, Honduras PR Rowe, FE Rosales 129

epidemic in Honduras. This disease is now present in most of the plantain-producing countries of Latin America (Fullerton, Stover 1990). It is already widespread in western and central Africa (Wilson, Buddenhagen 1986) and was identiﬁed in eastern Africa in 1987 (Sebasigari, Stover 1988). All the different plantain clones that have been tested are susceptible to black Sigatoka (Swennen, Vulysteke 1990). The AAA highland cooking and beer bananas of eastern Africa are also susceptible to this leaf spot disease (Sebasigari 1990). With the outbreak of the black Sigatoka epidemic in Honduras, and the drastic effects of this new disease for Latin America on plantain yield and quality, investigations of possible approaches to genetic improvement of plantains were begun in the Honduran banana breeding program. Now, breeding for black Sigatoka-resistant hybrids which could replace the AAA clones of eastern Africa has also become critically important. The four ABB cooking banana cultivars mentioned earlier are seed-fertile when crossed with diploids, and all were used in 3n x 2n cross-pollinations made in the initial attempts to breed black Sigatoka-resistant hybrids which could substitute for plantains. It was learned later that the French Plantain and Maqueño AAB clones could be used as female parental lines in breeding plantains (Rowe, Rosales 1990), but now it appears that crosses made onto one of the ABB clones (Cardaba) have provided some exceptional hybrids in breeding for new cooking bananas. The development and apparent value of three of these hybrids (one triploid and two tetraploids) with Cardaba parentage are discussed in this paper.

Development of a Dwarf Triploid with Cardaba Ancestry Genetic behavior of the ABB clones during meiosis is variable. Cheesman and Dodds (1942) found that when Bluggoe was crossed with diploids, the resultant hybrids included diploids, triploids, and tetraploids as well as apparent aneuploids with little vigor and abnormal leaf characters. Rowe (1976) reported preliminary attempts to use different ABB clones in breeding. Bred diploids of M. acuminata (AA) and accessions of M. balbisiana (BB) were crossed onto Chato (synonymous with Bluggoe), Pelipita and Cardaba (listed as Saba in the article, but later correctly identiﬁed as Cardaba). The weak and abnormal hybrids were discarded as seedlings, while normal progenies were transplanted to the field for subsequent verification of ploidy in individual plants selected for further observation. The hybrids with the best plant and bunch features from this series of crosses were tetraploids from Cardaba x M. balbisiana. Subsequently, a triploid hybrid derived from crossing a M. acuminata diploid onto one of these ABBB selected tetraploids had a much larger bunch (43 kg) than either Cardaba or its tetraploid parental line (Rowe 1983). This triploid was tall and was later discarded because of its slow growth, but it provided indications that hybrids with Cardaba in their pedigree are potentially useful in breeding for genetic improvement of cooking bananas. 130 Breeding Cooking Bananas for Areas with Marginal Growing Conditions

Since Cardaba is a tall plant (Fig.1), one of the primary objectives was to introduce the dwarfing gene of bred dwarf diploids into hybrids with Cardaba parentage. Rowe (1987) reported development of the SH-3386 dwarf triploid by crossing a dwarf diploid (AA) onto a tetraploid derived from Cardaba x M. balbisiana. SH- 3386 very nearly approaches commercial acceptability, but thin fruit at maturity has precluded its evaluation as a prospective cooking banana. While SH-3386 has no pollen, it does produce a few seeds when pollinated. Most hybrid seedlings obtained from crossing diploids onto SH-3386 are useless thick- leaved dwarfs characteristic of heptaploids, but a few grow normally. One of the tetraploid hybrids that has been selected from among pro- Figure 1. Plant with bunch of the Cardaba (ABB) cultivar genies of SH-3386 has many which has a tall plant stature. desirable qualities, and merits evaluation as a possible new commercial type with characters superior to those of the traditional cooking bananas in certain areas.

A Tetraploid Cooking Banana Hybrid with Commercial Potential This tetraploid with apparent potential as a new cooking banana is SH-3565 which was derived from SH-3386 x SH-3320, and was brieﬂy described in an earlier report (Rowe 1990). SH-3320 is a bred diploid (AA) which has a high level of resistance to black Sigatoka. A code name of FHIA-03 was assigned to SH-3565 when this hybrid was chosen as one of the entries for evaluation in the International Musa Testing Program (IMTP) which is supervised by INIBAP. PR Rowe, FE Rosales 131

Figure 2. Plants showing relative reactions to severe black Sigatoka pressure of the Nyamwihogora (AAA) cooking banana of eastern Africa (left) and FHIA-03 when bunches are ready to harvest. All the leaves of the former are necrotic while the latter has ﬁve functional leaves.

One traditional cultivar for which FHIA- 03 could possibly be a superior replacement is Bluggoe. It is widely grown and is a very important source of food in parts of eastern Africa (e.g., the southwestern slopes of Kilimanjaro and the eastern side of Zanzibar) and the West Indies (Simmonds 1966). This ABB clone is susceptible to bacterial wilt or moko disease (Pseu- domonas solanacearum) and race 2 of Panama disease (Fusarium oxysporum f.sp. cubense) (Stover 1972). The reactions of FHIA-03 to moko and race 2 of Panama disease are not yet known. However, one of the characters of FHIA-03 is that the male flowers are persistent, which should reduce contamination by the insect- transmitted strain of the moko disease bacteria. Further tests for resistance to Panama disease are pending, but FHIA-03 has been observed to remain healthy in an area where surrounding plants were infected with race 1 of this pathogen. At the time cross-pollinations were begun with Cardaba to evaluate the usefulness of this clone in breeding, black Sigatoka was not yet present in eastern Africa. Before the arrival of this disease, the most serious restriction on production of the highland bananas was soil nutritional and mineral deficiency problems which exist in all countries of eastern Africa (Sebasigari, Stover 1988). Now it appears that possibly the greatest value of FHIA-03 could be as a vigorous black Sigatoka-resistant substitute for the susceptible AAA cooking and beer varieties of that region. The susceptibility to black Sigatoka of the Nyamwihogora (AAA) cooking banana of eastern Africa as compared with the resistance of FHIA-03 in Honduras is shown in Figure 2. Under severe disease pressure FHIA-03 has some spotting, but still has about five functional leaves as compared with none for the eastern African clones when the bunches are ready to harvest. FHIA-03 is an extremely vigorous plant and has produced well under prolonged drought (4 months) and marginal soil (clay loam) conditions. This hybrid appears to 132 Breeding Cooking Bananas for Areas with Marginal Growing Conditions

Figure 3. From left: bunch features of the Nyamwihogora and Bluggoe cultivars of eastern Africa compared with the FHIA-03 hybrid, which merits evaluation as a possible disease- resistant replacement for these two traditional cooking bananas. have equal or better vigor than its Cardaba progenitor, even after the introduction of two M. acuminata diploids in its pedigree. If FHIA-03 is otherwise adapted to the altitudes of eastern Africa, its apparent tolerance to soil nutritional and mineral deficiencies should also be a valuable trait that would promote its cultivation as a new cooking banana. No precise data have been taken on flowering and ratooning speeds, but observations have been that FHIA-03 is relatively fast in both respects. By counting the number of pseudostem stumps for harvested bunches in individual stools from the same plot, FHIA- 03 was judged to be equal to the AAA clones of eastern Africa in ratooning speed. An additional desirable character of this hybrid is that it has a dwarf plant stature (which can be compared with the Grande Naine or Valery Cavendish clones depending upon whether it is grown under poor or optimum conditions). The pseudostems are very strong, and no doubled plant has been observed even when not propped and with bunches weighing more than 45 kg. Another attractive feature of FHIA-03 is its eating qualities. When green fruit is fried as thin slices, the flavor and texture are very similar to those of Bluggoe. When boiled whole, the green fruit has a flavor and texture almost identical to those of the AAA clones of eastern Africa. While the eastern African cooking bananas are eaten only as boiled green fruit, Bluggoe is cooked both green and ripe. FHIA-03 has a soft texture and sweet-acid flavor when ripe, as compared with a firmer texture and somewhat starchy flavor for ripe PR Rowe, FE Rosales 133

Bluggoe. It remains to be determined if FHIA-03 compares favorably with Bluggoe when cooked ripe. Perhaps any deficiencies in ripe FHIA-03 cooking qualities would be compensated for by its pleasant flavor and texture when eaten raw as a dessert banana. It could provide fresh fruit with an apple-like flavor in areas where dessert bananas have traditionally not been readily available, since ripe Bluggoe and AAA fruit of eastern Africa do not have pleasing fresh flavors. Another possible use for FHIA-03 would be as a beer banana. The Pisang Awak ABB cultivar is a popular beer banana in eastern Africa, but is susceptible to Panama disease (Sebasigari, Stover 1988). Since FHIA-03 has an ABB clone in its pedigree, this genetic background could contribute to its suitability for this purpose. FHIA-03 appears to have potential as a new hybrid for domestic consumption, as discussed above. Yields of FHIA-03 would be expected to be considerably greater than those of the traditional clones for which this hybrid could be an attractive alternative. Bunch features of a typical eastern African cooking banana and Bluggoe, as compared with those of FHIA-03, are shown in Figure 3. FHIA-03 should not be considered for export since, like its Cardaba parental line, it tends to ripen rather quickly after harvest. This relatively short green life would present risks that ethylene released from ripening FHIA-03 could cause premature ripening of other export bananas if shipped together. Pollen production is sparse in FHIA-03, and only one seed has been produced from pollination of about 30 bunches of this hybrid with a diploid that produces abundant pollen. Since pollen from diploids is much more efficient in fertilization than pollen from tetraploids, and FHIA-03 produces scant amounts of pollen, it is expected that this tetraploid would remain seedless if grown commercially. This effective seedlessness is a valuable trait for commercial production, but it essentially eliminates FHIA-03 as a parental line in further cross-pollinations. However, another tetraploid which has been selected from the segregating progenies of SH-3386 produces abundant pollen and is seed-fertile.

A Pollen- and Seed-Fertile Tetraploid Breeding Line This pollen- and seed-fertile tetraploid is SH-3648 (Fig.4) which was derived from SH- 3386 x SH-3362. SH-3362 is a bred diploid (AA) which has shown resistance to race 4 of Panama disease in Australia (Langdon, Pegg 1988). Triploids cannot be crossed with triploids. However, seed-fertile tetraploids can be crossed with pollen-fertile tetraploids since the even number of chromosome sets permits normal divisions and recombinations during reproduction. Thus, tetraploids derived from crossing a diploid onto one particular triploid clone can be crossed with tetraploids produced by crossing a diploid onto a different triploid genotype. This 4n x 4n breeding scheme has provided possibilities for genetic recombinations that heretofore were not practical. One factor that has made 4n x 4n crosses an attractive approach to breeding new types of bananas has been the identification of an array of seed-fertile triploid clones. Also, the development of diploids with desirable agronomic and disease- 134 Breeding Cooking Bananas for Areas with Marginal Growing Conditions

Figure 4. Plant and bunch features of the dwarf SH-3648 tetraploid which has Cardaba in its pedigree. SH-3648 is being used as both a pollen and seed parent in further crosses for breeding new cooking banana hybrids.

resistance features have now resulted in superior genetically diverse tetraploids from 3n x 2n crosses (Rowe, Rosales 1993). The three main groups of cooking bananas are plantains (AAB), eastern African highland bananas (AAA), and the ABB clones. Rowe and Rosales (1990) reported the development of tetraploids from crosses onto a French Plantain and the Maqueño (AAB) cooking banana. Now, SH-3648 has the Cardaba ABB clone in its pedigree and provides an additional tetraploid for making 4n x 4n crosses in attempts to produce hybrids that have combinations of the best features of different cooking bananas. For development of plantain types with potential for expanding the export market, the most promising 4n x 4n cross is between tetraploids derived from the French Plantain and tetraploids with Maqueño parentage. However, since this paper is about hybrids with Cardaba as a progenitor, discussion will be limited to crosses involving SH-3648. All the tetraploids produced so far from crosses onto French Plantain and Maqueño are tall. The first seeds obtained from crossing the dwarf SH-3648 onto these AAAB tetraploids have been produced and are being germinated by embryo culture. These 4n x 4n crosses were made with the anticipation that some of the resultant hybrids will have the eating qualities of plantains plus the dwarfness and vigor of SH-3648. Such hybrids would be expected to be productive in areas where plantains are preferred but presently have limited productivity due to infertile soils. Another use of SH-3648 is as a pollen parent in crosses onto the triploid Bluggoe, Pelipita, and Cardaba clones. These 3n x 4n crosses are rationalized by taking into PR Rowe, FE Rosales 135

consideration that meiosis is variable in these ABB clones, and the expectation that some of the progenies will be triploids and tetraploids. Such crosses could result in hybrids with the dwarf plant habit of SH-3648 and the preferred cooking qualities of the traditional tall ABB clones. The ﬁrst seedlings produced from this series of crosses have already germinated in embryo culture. Crosses between SH-3648 and different advanced bred diploids are expected to give genetic diverseness in triploid hybrids that would be seedless due to the uneven number of chromosome sets. The first hybrids obtained from crossing the black Sigatoka- resistant SH-3437 diploid onto SH-3648 have been transplanted to the field. Selected hybrids from this cross are anticipated to provide additional prospective commercial types that could have even higher levels of resistance to black Sigatoka than that of FHIA-03.

Conclusions More than 85% of the total world production of about 60 million t of the different types of bananas are consumed domestically. A large portion of this production consists of plantains, cooking bananas, and beer bananas. With the exception of the Saba and Cardaba cooking bananas, which are cultivated in the Philippines (Valmayor et al. 1981; Valmayor 1987), all the major clones used for cooking or making beer are susceptible to diseases that either are already causing increased production costs and reduced yields or are constant threats to the few remaining relatively disease-free areas. Attempts to use Saba in breeding have not yet resulted in any useful hybrids. However, as has been discussed in this paper, some hybrids with exceptional plant and bunch features have been developed by using Cardaba in cross-pollinations. One hybrid with Cardaba parentage has qualities that indicate it merits evaluation to determine further its potential for commercial cultivation, and another provides a superior tetraploid breeding line in subsequent crosses for development of additional new cooking bananas. Acknowledgment Grateful thanks are given to the International Development Research Centre (IDRC) of Canada which has provided the major ﬁnancial support that made continued breeding for genetic improvement of cooking bananas possible.

References

CHEESMAN EE, DODDS KS. 1942. Genetical and cytological studies of Musa IV. Certain triploid clones. Journal of Genetics 43:337-357. FULLERTON RA, STOVER RH (eds). 1990. Sigatoka Leaf Spot Diseases of Bananas: proceedings of an international workshop held at San José, Costa Rica, March 28 - April 1, 1989. Montpellier, France: INIBAP. 374 pp. INIBAP. 1986. Banana Research in Eastern Africa. Proposal for a Regional Research and Development Network. Montpellier, France: INIBAP. 136 Breeding Cooking Bananas for Areas with Marginal Growing Conditions

INIBAP. 1987. Plantain in West and Central Africa. Proposal for a Regional Research and Development Network. Montpellier, France: INIBAP. JARAMILLO CR. 1987. Banana and plantain production in Latin America and the Caribbean. Pages 39-43 in Banana and Plantain Breeding Strategies (Persley GJ, De Langhe EA, eds). ACIAR Proceedings no.21. Canberra, Australia: ACIAR. JARRET RL, LITZ RE. 1986. Determining the interspecific origins of clones within the `Saba’ cooking banana complex. HortScience 21: 1433-1435. LANGDON PW, PEGG KG. 1988. Field screening for resistance to Fusarium wilt. In Research Report: ACIAR Banana Improvement Project for the Pacific and Asia — Fusarium Wilt (Panama Disease) Studies. Canberra, Australia: ACIAR. 12 pp. ROWE P. 1976. Possibilités d’amélioration génétique des rendements de plantain. Fruits 31:531-536. ROWE P. 1983. L’hybridation pour l’amélioration des plantains et autres bananas à cuire. Fruits 38:256-260. ROWE PR. 1987. Breeding plantains and cooking bananas. Pages 21-23 in International Cooperation for Effective Plantain and Banana Research: proceedings of the third meeting of IARPB held in Abidjan, Côte d’Ivoire, 27-31 May 1985. Montpellier, France: INIBAP. ROWE PR. 1990. New genetic combinations in breeding bananas and plantains resistant to diseases. Pages 114- 123 in Identification of Genetic Diversity in the Genus Musa: proceedings of an international workshop held at Los Baños, Philippines, 5-10 September 1988 (Jarret RL, ed.). Montpellier, France: INIBAP. ROWE PR, ROSALES FE. 1990. Breeding bananas and plantains with resistance to black Sigatoka. Pages 243-251 in Sigatoka Leaf Spot Diseases of Bananas: proceedings of an international workshop held at San José, Costa Rica, March 28 - Abril 1, 1989 (Fullerton RA, Stover RH, eds). Montpellier, France: INIBAP. ROWE PR, ROSALES FE. 1993. Bananas and plantains. In Advances in Fruit Breeding, 2nd edn (Janick J, Moore JN, eds). Timber Press (in press). SEBASIGARI K. 1990. Effect of black Sigatoka (Mycosphaerella fijiensis Morelet) on bananas and plantains in the Imbo plain in Rwanda and Burundi. Pages 61-63 in Sigatoka Leaf Spot Diseases of Bananas: proceedings of an international workshop held at San José, Costa Rica, March 28 - April 1, 1989 (Fullerton RA, Stover RH, eds). Montpellier, France: INIBAP. SEBASIGARI K, STOVER RH. 1988. Banana Diseases and Pests in East Africa: report of a survey made in November 1987. Montpellier, France: INIBAP. SIMMONDS NW. 1966. Bananas, 2nd edn. London, UK: Longmans. 512 pp. STOVER RH. 1972. Banana, Plantain and Abaca Diseases. Kew, Surrey, UK: Commonwealth Mycological Institute. 316 pp. STOVER RH, DICKSON JD. 1976. Banana leaf spot caused by Mycosphaerella musicola and M. fijiensis var. difformis: a comparison of the first Central American epidemics. FAO Plant Protection Bulletin 24:36-42. SWENNEN R, VULYSTEKE D. 1990. Aspects of plantain breeding at IITA. Pages 252-266 in Sigatoka Leaf Spot Diseases of Bananas: proceedings of an international workshop held at San José, Costa Rica, March 28 - April 1, 1989 (Fullerton RA, Stover RH, eds). Montpellier, France: INIBAP. VALMAYOR RV. 1987. Banana improvement imperatives - the case for Asia. Pages 50-56 in Banana and Plantain Breeding Strategies (Persley GJ, De Langhe EA, eds). ACIAR Proceedings no.21. Canberra, Australia: ACIAR. VALMAYOR RV, RIVERA FN, LOMULJO FM. 1981. Philippine Banana Cultivar Names and Synonyms. IPB Bulletin no.3. Los Baños: University of the Philippines. 16 pp. WILSON GF, BUDDENHAGEN I. 1986. The black Sigatoka threat to plantain and banana in West Africa. IITA Research Briefs 7(3):3. Part 2

In-Vitro Strategies for Musa O Arias 139

Commercial Micropropagation of Banana

O Arias

During this century, the production of bananas has been increased significantly. The increase was interrupted temporarily only by the two wars and the depression of the 1930s. World production of export bananas over the period increased 18 times, from less than half a million tonnes exported in 1900, to 2.3 million t in 1950 and 9 million t in 1990. Despite the fact that the total value of bananas in the international market is not high, the impact of this activity in the economies of banana-exporting countries in Latin America has been considerable. Not only has there been a monetary contribution, but also work and income have been created for a large sector of the population. The export trade additionally contributes to national economies through taxes. Banana growing has led to an important increase in jobs. It has been estimated that in Latin America 250 000 are employed and 450 000 more are involved in complementary activities. This means that 700 000 families or a total of 35 million people depend on this activity. Even though during 1981-90 the standard of living has decreased because of setbacks in the economies of banana-exporting countries in Latin America, this has not affected the people involved in banana production because, in most cases, the people involved earned more than others who work in different areas of agriculture (UPEB 1991). Opportunities for new markets, especially in Europe, has led to an expansion of banana plantations. Between 1985 and 1991 the number of hectares cultivated in Latin America increased tremendously: by approximately 118% in Ecuador, 61% in Costa Rica, 65% in Colombia, and 67% in Mexico. In other countries the increase was not very signiﬁcant, as shown in Table 1. The establishment of new banana-growing areas has increased the demand for planting materials. Growers have mainly used three types: large suckers with an upper diameter of 10-15 cm; small suckers with a diameter of 5-10 cm; and in-vitro plants. The first two are inconvenient because the vegetative part of the plant which is used frequently carries pathogens and pests, especially nematodes such as Radopholus similis, the borer Cosmopolitus sordidus, race 4 of Fusarium oxysporum f.sp. cubense, and viruses such as cucumber mosaic and bunchy top. In-vitro plantlets provide excellent alternative planting material free of these pests and diseases for use in the new areas of cultivation. This advantage has been intensively exploited over the last

Agribiotecnología de Costa Rica S.A., PO Box 25-4001, Alajuela, Costa Rica 140 Commercial Micropropagation of Banana

Table 1. Estimates of the areas planted with export bananas in Latin America (ha). Country 1985 1991 Increase Costa Rica 20 285 33 000 12 415 Colombia 24 191 40 000 15 809 Guatemala 7 688 8 500 812 Honduras 17 590 19 500 1 910 Nicaragua 2 761 3 000 239 Panama 13 541 15 000 1 459 Ecuador 48 000 105 000 57 000 Mexico 12 000 20 000 8 000 Total 134 356 244 000 97 644

5 years by such Latin American countries as Colombia and Costa Rica. The technology has been transferred in part from countries with a longer tradition in the use of banana tissue-culture plantlets, such as Israel, Australia, and Taiwan (Reuveni 1986). The main advantages of the in-vitro propagation technique for bananas are the following. 1. Plants can be rapidly multiplied from a mother plant of known, desirable characters. 2. Selected and screened plants can be maintained free of serious diseases and pests. 3. The use of in-vitro plantlets in areas not infected with nematodes avoids the use of nematocides, at least for a period of 5 years; it also avoids environmental contamination and residues, and gives significant savings to the grower of approximately US$400 ha-1 a-1. 4. The use of single-cycle high-density banana plantations could be adopted more widely, especially for the window market. The expected advantages of this practice are the following: very high yield in a short time; efficient control of flowering and harvesting time; saving of hand labor; the possible use of poor land marginal for permanent cultivation; saving expenditure on infrastructure (Israeli, Nameri 1987). 5. 98% survival under field conditions. 6. Plants from in-vitro plantlets grow faster in the early growing stages than those from suckers. 7. Uniformity of flowering. 8. Short harvesting period. 9. In comparison with the suckers, plants are cheaper and easier to propagate and transport. 10. There are important advantages regarding germplasm conservation and the possibility of international transfer. 11. The material produced should be true-to-type and conform to the characters of the mother plant. (Such conformity is currently causing some concern in the industry.) With the increased use of in-vitro plant multiplication, the widespread presence of somaclonal variation has been detected, from negligible levels up to 40% of a field. This O Arias 141

has been found in all countries that have used in-vitro planting material, e.g., Australia, Israel, South Africa, Taiwan, and Latin America (Arias, Valverde 1987; Reuveni 1986). Experience gained in Costa Rica in the commercial use of in-vitro plantlets suggests the incidence of off-type plants produced in vitro varies between 1 and 10%, when big populations of carefully selected plants are planted. Almost all the off-types have morphological characters in stage V (nursery), making elimination possible at this stage. Table 2 shows results from the evaluation of 66 299 plants grown in three different locations. The percentage of mutants was between 6.5 and 8% for cv Grande Naine and 8.5% for cv Williams. Table 2. Off-types observed in three ﬁelds in Costa Rica. Plantation Cultivar No. of plants evaluated % of off-types 1Grande Naine 50179 8.0 2Grande Naine 10260 6.5 3Williams 5860 8.5

When the plants were analyzed individually, we were able to establish the frequency of occurrence of some of the mutants. It was then apparent that changes in plant stature and bunch characters were the most common off-types, as shown in Table 3. Table 3. Approximate % distribution of mutants in different clonal classes (pers. com. A Azofeita, O Arias 1991). Type % Dwarf ...... 40 Leaves upright ...... 25 Valery-like ...... 12 Plantain-like ...... 5 Variation in the amount of brown to purple pigmentation ...... 7 Abnormal foliage ...... 5 Leaf variegation ...... 4 “Masada” ...... 2

Dwarﬁsm, as indicated by the mutation of the Grande Naine height class, accounted for 40% of the mutants. A second class of mutants comprises those that have: leaves upright; closer internodes and abnormal phyllotaxy, giving the impression of a travelers palm (Ravenala madagascariensis); and fruit of no commercial value. The incidence of these plants was 25%. Twelve per cent of the plants were found to revert to Valery height with normal bunches. We also found plants with bunches similar in appearance to plantain. At the nursery stage we distinguished these plants by the large size of the leaf blade and dropping leaves. They accounted for 5% of the mutants. These plants may have a height similar to cv Grande Naine, or even a little smaller, but with abnormal phyllotaxy distribution and persistent bracts on the bunch. 142 Commercial Micropropagation of Banana

The absence of or abnormal pigmentation was also related to somaclonal variation. The plants concerned had darker and thickened leaves that also showed purple pigmentation on the lower side of the leaf. They represented 7% of the total of mutants. This pigmentation is different from the known yellow to green variegation (that accounted for 4%). Occasionally plants had abnormal foliage characters: irregular shape of the lamina, sometimes with a portion missing or a wavy edge of the leaf that accounted for 5% of the mutants. We found that 2% of the plants looked normal but had greasy spots on the lower side of the leaf, which the Israeli investigators called “Masada”. These plants produced small fruits of no commercial value. If careful screening is carried out, almost all the somaclonal variations can be detected at the nursery stage, as already noted. In this case we found in the field an approximate error of 1.5%, which corresponded to dwarf plants most of the time because they were the most difficult to observe at the nursery stage. In conclusion, the experience obtained in Latin America over the last 5 years in which thousands of plants were produced and planted in the field, has produced enough technology for the large-scale use of in-vitro plants, especially for the establishment of new plantations and the substitution of traditional varieties in old fields. Using somaclonal variants, the Taiwanese researchers have also demonstrated the possibility of using these plants for screening for disease resistance (Wang, Ko 1987). The tissue-culture production of bananas in Latin America is probably the most important example of the application of biotechnology in agriculture. In spite of the progress achieved, we still have an enormous lack of information on the horticultural management of the plants at nursery level (stage V). The development of more practical methods for the early detection of mutants in the laboratory would assist the commercial laboratories considerably, and thus help in the coming years in the development of genetically engineered plants with resistance to and tolerance of the most important pests and diseases, within the overall objectives of biotechnology research for improved banana cultivation.

References

ARIAS O, VALVERDE M. 1987. Producción y variación somaclonal de plantas de banano, variedad grande naine, producidas por cultivos de tejidos. Asbana 28:6-11. ISRAELI Y, NAMERI N. 1987. A single-cycle high-density banana plantation planted with in vitro propagated plants. Israel: Jordan Valley Banana Experimental Station. 14 pp. REUVENI O. 1986. Performance and genetic variability in banana plants propagated by in vitro techniques. Bet Dagan, Israel: The Volcani Center, Department of Subtropical Horticulture, Agricultural Research Organization. 26 pp. UPEB. 1991. La actividad bananera mundial en 1990. No.91-92:9-20. WANG HSC, KO WH. 1987. Mutants of Cavendish banana resistant to race 4 of Fusarium oxysporum f.sp. cubense. Pages 42-46 in Taiwan Research Institute Annual Report 1987. FJ Novak et al. 143

Improvement of Musa through Biotechnology and Mutation Breeding

FJ Novak, H Brunner, R Afza, R Morpurgo, RK Upadhyay1, M Van Duren, M Sacchi, J Sitti Hawa2, A Khatri3 (IAEA); and G Kahl, D Kaemmer, J Ramser, K Weising (University of Frankfurt)

Introduction The genetic system of Musa is extremely complicated to study because of the serious problems of sterility, interspecific hybridity, heterozygosity, and polyploidy that are common in most of the clones. The asexual behavior is often an inseparable barrier in using cross-breeding as a tool for genetic improvement. Consequently, to date, the breeding of bananas and plantains having desirable characters and acceptability at the farmers’ level has been without much success. Therefore, the complexity of Musa genetics needs a more sophisticated system to support conventional breeding programs, and prospects of using potentials of biotechnology in this crop are very high (Persley, De Langhe 1987; Dale 1990; Novak 1992). This report describes the results of development of the following systems of Musa biotechnology: (a) in-vitro mutation induction; (b) cell (protoplast) culture and somatic embryogenesis; (c) markers of genetic diversity; (d) in- vitro screening and selection for resistance against Fusarium oxysporum f.sp. cubense and Mycosphaerella ﬁjiensis.

Materials and Methods Materials Clones of the following Musa species and cultivars have been used as experimental material: M. acuminata subsp. burmannica (clone IV-9), M. balbisiana, Pisang Mas

Plant Breeding Unit, Joint FAO/IAEA Programme, IAEA Laboratories, A-2444 Sebersdorf, Austria; and Plant Molecular Biology, Department of Biology, University of Frankfurt/Main, D-6000 Frankfurt/Main, Germany. Permanent addresses: 1PPO (Plant Pathology), Directorate of Plant Protection, Quarantine, and Storage, Department of Agriculture and Co-operation, N.H.-IV, Faridabad, Haryana, India; 2Fruit Research Division, Malaysian Agricultural Research and Development Institute (MARDI), PO Box 12301 General Post Ofﬁce, 50774 Kuala Lumpur, Malaysia; 3Atomic Energy Agricultural Research Centre, Tandojam, Pakistan 144 Improvement of Musa through Biotechnology and Mutation Breeding

(AA), SH-3362 (bred diploid clone AA), Grande Naine (AAA), Saba-Cardaba (ABB), Pisang Rastali (Silk banana, AAB).

Shoot-tip culture Shoot tips (2-5 mm in size) consisting of the meristematic dome with 2-4 leaf primordia were cultured on MS medium with 10 µM 6-benzylaminopurine (BAP) and 5 µM indoleacetic acid (IAA). The medium was solidiﬁed with gelrite (1.75 g L-1). Multiple shoot formation was achieved by subculturing shoot tips on MS medium supplemented with 20 µM BAP. Shoots were elongated and rooted on MS medium with 5 µM 2ip, 1 µM IBA. Detailed methods of in-vitro production of banana plantlets were described by Vuylsteke (1989).

Induced mutagenesis Meristem tips, 1-2 mm in size, were excised from in-vitro growing shoots and irradiated in a gamma cell with a 60Co source with 15-60 Gy at a dose rate of 8 Gy min-1 (Novak et al. 1990). Uniform longitudinally dissected meristem tips were dipped into an aqueous solution of 50 µM ﬁlter-sterilized ethylmethanesulphonate (EMS) with 2% v/v dimethylsulphoxide (DMSO) for 3 h at a constant temperature of 28°C (Omar et al. 1989). The explants were then thoroughly washed with ample amounts of sterile water prior to culture. The nutrient media were the same as described for shoot-tip culture.

Embryogenic callus and cell suspension, plant regeneration The extreme basal parts of the leaf sheaths were cut from young leaves of axenic meristem-derived plantlets. Corm tissues deprived of the apical meristem were cut into segments (5 x 5 mm). Both types of explants were plated on petri dishes on Schenk and Hildebrandt (1972) medium supplemented with a modiﬁed mixture of Staba’s vitamins (Novak et al. 1989), 40 g L-1 sucrose, 40 mg L-1 cystein HCl, 30 µM Dicamba. The cultures were maintained in the dark at 28°C for 4-6 weeks. Embryogenic calli were subcultured into liquid, hormone-free medium MS (half strength) and incubated on shakers (120 rev min-1) for 1 week. Cell suspensions were transferred into half-strength SH medium with 20 µM Dicamba for long-term culture of cell suspension (more than 1 year). The cell suspension maintained in the Dicamba medium lost its regeneration capacity within a 2- 3 month period. To regenerate plants, the fresh suspensions of cells and proembryogenic structures were transferred into half-strength, hormone-free MS and then plated onto solid medium composed of half-strength MS + vitamins + 40 g L-1 cystein + 20 g L-1 sucrose, 1 g charcoal, 5 µM zeatin, 4 g L-1 agar and 1.75 g L-1 gelrite. In this double-layer system, somatic embryos converted into plants within 2 weeks. The detailed protocol is described elsewhere (Novak et al. 1989).

Encapsulation of embryos Carefully selected, complete bipolar somatic embryos were transferred to a sterile 3% (w/v) solution of sodium alginate (Sigma, medium viscosity), prepared in MS basal FJ Novak et al. 145

medium (half concentration) without calcium chloride, supplemented with Staba vitamins, 20 g L-1 sucrose, 100 mg L-1 inositol, 40 mg L-1 cystein-HCl and 5 µM zeatin. Embryos, together with the alginate solution, were then taken up in a sterile pipette with an opening of 4 mm, and at close distance carefully dropped into a 50 µM solution of calcium chloride dihydrate on a gyratory shaker (60 rev min-1) at room temperature. The beads were left in the solution for 30 min to allow proper hardening. Selected, totally encapsulated beads were then transferred to the double-layer medium described above.

Protoplast isolation and culture The enzymatic mixture of Macerozyme R-10 (1%), cellulase Onozuka R-10 (3%) and pectinase (1%) released a large number of palisade mesophyll cells from leaves of aseptic plants and parenchyma cells of rhizome tissue during 12 h of cultivation at 28°C. The suspension of intact protoplasts, cell debris, starch grains, and raphides was washed in CPW13M, transferred to conical tubes and after 10 min of centrifugation at 100x g the protoplasts were collected from the surface of the media. The washing procedures were repeated twice to obtain pure fractions of protoplasts. The protoplasts were cultured in the half-strength MS with 20 µM Dicamba in thin layers of liquid media in petri dishes.

Protein extraction and electrophoresis Soluble proteins were extracted from young folded leaves by using SDS buffer. Samples were homogenized and centrifuged at 13 000 rev min-1 at 4°C for 15 min. A 12% polyacrylamide gel with a 4% stacking gel was used for electrophoretic separation of proteins at 1500 V/h. The gel was stained with comassie blue.

DNA fingerprinting Extraction of DNA. High molecular weight DNA was isolated from lyophilized leaves or other organs according to a modified cetyl-trimethylammonium bromide (CTAB) method (Weising et al. 1991). The lyophilized material (0.5 g) was ground to a fine powder and subsequently transferred to 15 mL of hot (60°C) 2 x CTAB extraction buffer. After gently swirling the resulting cell lysate for 30 min, nucleic acids were isolated by extraction with an equal volume of chloroform/isoamylalcohol (24:1) followed by precipitation with 0.6% volume of isopropanol. After centrifugation pellets were resuspended in TE (10 mM Tris-HCl, 1 mM EDTA, pH 8), and DNA was further purified by ultracentrifugation in CsCl/ethidium bromide followed by extraction with TE-saturated 1-butanol and ethanol precipitation. Restriction of DNA, agarose gel electrophoresis, and gel-drying. During initial experiments the purified DNA was digested with Hinf I, Alu I, and Taq I (according to supplier’s instructions; 6 units of enzyme per mg DNA). Since Hinf I gave the best results it was used in all subsequent experiments. After digestion the fragmented DNA was electrophoresed in 1.0% agarose gels and TAE buffer (40 mM Tris-acetate, 20 mM sodium acetate, 1 mM EDTA; pH 8.3). The gels were dried on a vacuum gel dryer. 146 Improvement of Musa through Biotechnology and Mutation Breeding

Probe labelling, radioactive hybridization, and autoradiography. The gels were denatured, neutralized and hybridized to 32P endlabeled oligonucleotides essentially as described (Ali et al. 1986). Temperatures of hybridization and stringent washing steps ° ° ° were 35 C for (GATA)4, 43 C for (CA)8, and 45 C for (GTG)5. One and the same gel was used for different hybridization probes. Before reprobing, probes were stripped off the gel by washing in 5 mM EDTA at 60°C (2 x 15 min). DNA amplification fingerprinting. The polymerase chain reaction was performed with 100 ng of high molecular weight genomic DNA in a volume of 50 µL. The tubes contained 150 µmol L-1 of each dNTP, 20 pmol of the primer(s) and 25 units of DNA polymerase from Thermus aquaticus (Ampli-Taq polymerase, Perkin-Elmer/Cetus). The reaction mix was overlaid with 30 µL mineral oil, incubated for 5 min at 94°C, and amplified in a waterbath thermocycler for 30 cycles consisting of 1 min at 94°C, 1 min at annealing temperature, and 2 min at 72°C. The last elongation step was extended to 10 min. One and the same DNA preparation served as a source for both oligonucleotide and amplification fingerprinting.

Fusarium culture and bioassays of plant-fungus interactions Fungal culture and filtrate production. Fusarium oxysporum f.sp. cubense (FOC) was cultured on Czapek-Dox medium and transferred to potato dextrose agar for sporulation. One mL of conidial suspension containing 450 000 conidia mL-1 was transferred to a 500-mL Erlenmeyer flask containing 250 mL of liquid Czapek-Dox for crude filtrate production. Cultural conditions were: 28°C, continuous light for sporulation on PDA, 28°C dark condition for production of mycelium and crude filtrate. After removal of mycelium, the liquid phase was filtered through three layers of cheese cloth, refiltered on Watmann paper no. 1 and finally filtered through a 0.22 µm Millipore membrane. Production of crude filtrate activity was assayed with a spectrophotometer following the absorbance of fusaric acid at 270-272 nm. Plant-fungus (FOC) coculture. The fungus (FOC) was inoculated on a slant potato dextrose agar (PDA) medium in tubes and allowed to grow for about 3 weeks in incubators at 28°C and continuous light. The plantlets were rooted in vitro on solidified MS medium with 20 µM BAP and 2 g L-1 charcoal, for about 3-4 weeks under 16 h day-1 light. For the treatment, uniform growing fungus and plantlets were selected. The plantlets were washed thoroughly with sterile water and the roots trimmed. Meanwhile liquid MS medium with 5 µM 2ip and 1 µM IBA was added on top of the cultured fungus. The prepared plantlets were then introduced into the above tubes and placed in incubators at 18°C, 25°C, and 32°C with 16 h day-1 light. For the control, the plantlets were prepared in the same manner but grown in fungus- free PDA slant medium topped with similar liquid MS medium, as described above. Observations for the wilting symptoms were carried out, such as the first appearance of the yellowing of leaves and final death of the plants. FJ Novak et al. 147

Bioassay of crude filtrate. Banana shoot tips cultured as already mentioned were transferred to a medium containing 3, 6, 9, and 12% of crude filtrate obtained after 21 days and placed in the same condition as before for 3 weeks. After that time, the plant material was analyzed for different parameters such as plant height, total fresh weight, shoot production, and root fresh weight. Plant inoculation. Rooted plantlets were obtained and after removal of gelrite the roots were cut and plants dipped into a conidial suspension (500 000 conidia mL-1) for 10 min and then transplanted in Erlenmeyer flasks containing sterile perlite. Total soluble peroxidase activity. At regular intervals infected plants were removed from the flask and after removal of roots and leaves the remaining corm tissues with some pseudostem tissue were crushed in a mortar using phosphate buffer, pH 6.8, as extraction buffer. The homogenized tissues were collected and placed in Eppendorf tubes and centrifuged for 15 min at 14 000 rev min-1. The supernatant was transferred to new Eppendorf tubes and immediately assayed for total peroxidase activity. Peroxidases were assayed with guaiacol as the hydrogen donor. The reaction mixture consisted of 0.3% v/v guaiacol and 15 µL of hydrogen peroxide in phosphate buffer, pH 6.8. Five µL of supernatant were added to 3 mL of reaction mixture and changes in absorbance were followed at 470 nm. The reaction was stopped after 2 min. Isoelectrofocusing (IEF) of total soluble peroxidases. Isoelectrofocusing of peroxidases was performed in a horizontal apparatus (Pharmacia Multiphor II) at 4°C. Ampholines (Pharmacia LKB) with a pH range of 3-10 were used as electrolyte carriers and prefocused for 20 min. Fifteen µL of sample per lane was added. Running conditions were fixed at 8 W, the mA from 33 to 10, and the voltage up to 2500 V. After focusing, the gels were soaked in 100 mL 0.1 M phosphate buffer at pH 6.8 containing 0.3% v/v guaiacol and 15 µL of hydrogen peroxide. After 15 min of incubation the gels were analyzed for the stained bands.

Isolation of toxins of Mycosphaerella fijiensis and bioassay The pathogen (MF) was grown in a modified M-1-D plus coconut water liquid medium in an Erlenmeyer flask at 25±1°C and 12 h day-1 light and dark cycles on a rotary shaker at 120 rev min-1 for 28 days. After the incubation period was over, an equal volume of methanol was added in the culture and placed overnight at 4°C. The culture was then filtered and the filtrate evaporated to reduce the volume on a rotary evaporator under vacuum at 36°C. The filtrate so obtained was extracted three times with an equal amount of ethylacetate which was evaporated under reduced pressure and mild temperature to obtain the crude extract (CE). CE containing bioactive compounds was fractionated after passing through a column loaded with Sephadex LH-20, using chloroform and methanol (1:1) as solvent system. The tetralone, isoochracinic acid and 2-carboxy-3-hydroxycinnamic acid of 90% purity were isolated using the protocol of Stierle et al. (1991). Juglone was obtained from Aldrich Chemical Company. Fijiensin was isolated using the method described elsewhere (Upadhyay et al. 1990). Bioassays were conducted on leaves of plantlets from different cultivars comprising highly 148 Improvement of Musa through Biotechnology and Mutation Breeding

susceptible to resistant cultivars using the leaf puncture bioassay test method (Upadhyay et al. 1990).

Results and Discussion

Mutation breeding In-vitro mutagenesis in shoot-tip culture of banana is schematically represented in Figure 1. This system was performed in the following steps: (1) vegetative shoot apices were micropropagated before mutagenic treatment; (2) tiny meristems permitted

Figure 1. Schematic representation of the in-vitro mutation breeding system in Musa (Novak et al. 1990). FJ Novak et al. 149

Figure 2. Schematic representation of biotechnological methods supporting the conventional breeding system of Musa (Vakili 1967). The techniques in dotted rectangles are not fully established for banana. 150 Improvement of Musa through Biotechnology and Mutation Breeding

treatments on a large number of propagulas and even the use of mutagenic chemicals which otherwise may have difﬁculty in penetrating target primordial cells; (3) replated subculture on a proliferation medium forced formation of adventitious buds which reduced chimerism in vegetative offsprings; (4) an in-vitro system was used for mass screening of the plantlets’ response to pathogenic fungi or their toxic products; (5) plants were regenerated and evaluated in ﬁeld conditions.

The variants in M1V4 vegetative progenies of irradiated and/or chemical mutagen (EMS)-treated shoot tips comprised phenotypic changes in morphological (plant stature, leaf shape), physiological (sucker growth and multiplication, flowering time, fruit ripening), and agronomic characters (bunch quality). The outstanding plant of the cultivar Grande Naine (putative mutant GN 60GyA) was identified and multiplied for field-testing. The mutant plant showed differences in the zymograms of soluble proteins and esterase isozymes (Novak et al. 1990). Induction of somatic mutations may substantially contribute to the broadening of the genetic variation among Musa clones with obligate vegetative reproduction (e.g., Cavendish). The system of in-vitro mutagenesis, however, may also be important for the breeding of fertile diploids before their use for crossing. Figure 2 schematically outlines the biotechnology supporting the Musa conventional breeding system proposed by Vakili (1967). Two steps of mutagenesis are proposed: (1) mutation-breeding of female diploids before crossing, or (2) mutation- induction and selection among F1 recombinant diploids before polyploidy induction. Somatic embryogenesis Realization of somatic embryogenesis (Fig. 3) in economically important Musa cultivars (AAA and ABB) opened new approaches to somatic cell manipulation and breeding (Novak et al. 1989). Somatic embryos may be encapsulated and used for propagation and in-vitro storage; however, the frequency of “germinating” artificial seeds was very low (less than 0.1%). Embryogenic suspension represents a superior source of genetic variation (“true somaclonal variation”) as well as a unique unicellular system for mutation induction and selection. Plants of Grande Naine regenerated via somatic embryogenesis were sent to FHIA, Honduras, and Maroochy Experimental Station, Queensland, for field observation of phenotypical variation. The material grown in a greenhouse at Seibersdorf will be analysed for DNA fingerprinting patterns to determine the genetic (in)stability of the somatic embryogenesis process. Our experience in somatic embryogenesis of different Musa clones clearly showed that a modified protocol must be empirically developed for each particular genotype. Embryogenic suspension is the preferred source for protoplast isolation and culture (Fig. 4). However, the attempts to induce cell wall formation and callus-colony proliferation were unsuccessful.

Genetic markers Differences among four different clones in the soluble protein electrophoresis were recognized as reliable markers for discrimination among Musa clones of genomic group A. However, consistency in the number of polymorphic bands on electrophoreto-graphs was FJ Novak et al. 151

Figure 3. Somatic embryogenesis in Musa. A: embryonic callus. B: cell suspension. C: somatic embryos. D: encapsulated somatic embryos (“artiﬁcial seeds”). E: germination of somatic embryos. F: regenerated plants. 152 Improvement of Musa through Biotechnology and Mutation Breeding

Figure 4. Protoplast suspension of Grande Naine (AAA). strictly dependent on the precise determination of developmental stages of leaves before protein extraction. The young, fully folded leaves were found as standard material for sample preparation (Fig. 5). A band of 30 000 d has been identified only in SH-3362, while a band of 32 000 d characterized Grande Naine (Fig. 6). Quantitative differences were identified among clones (Sacchi, Novak, 1992). DNA oligonucleotide and DNA amplification fingerprinting were successfully used to detect genetic diversity between 13 out of a total of 15 representative Musa species and cultivars. DNA fragments specific for the A genome as well as the B genome can be found with both techniques, which also permit discrimination between triploids (AAA) differing in their susceptibility or resistance towards Fusarium (e.g., Gros Michel, Cavendish). Both the oligonucleotide and the amplification fingerprints (RAPiD) are somatically stable, i.e., do not exhibit any differences between tissues of one and the same plant. The probes and restriction endonucleases used for DNA fingerprinting (see Material and Methods) did not distinguish among individuals of the same clone, Figure 5. SDS electrophoresis though a unique fingerprint was found for each clone of soluble proteins extracted from different leaves of (Fig. 7). However, certain amplimer combinations Grande Naine (AAA). FJ Novak et al. 153

Figure 6. SDS electrophoresis of soluble proteins extracted from folded leaves of four different cultivars of Musa: DC = Dwarf Cavendish; HG = Highgate; GN = Grande Naine; 3362 = SH-3362. Figure 7. DNA oligonucleotide fingerprinting of different clones of Musa. a = Pisang Mas; b = SH-3362: a bred diploid; c = Musa acuminata ssp. burmannica; d = M. balbisiana; e = Highgate; f = Dwarf Parfitt; g = Williams; h = Grande Naine; i = GN 60A: an induced mutant of Grande Naine; k = Tafetan verde; l = SH-3436: a bred tetraploid; m = AVP-67; n = Horn; o = Pelipita; p = Cardaba. The group of AA clones is genetically unrelated (a,b,c). Note the differences between Gros Michel (e) and Cavendish (f,g,h,i) groups. There are B genome aspecific fragments in d,m,n,o,p. discriminate between an original clone and a mutant in amplification fingerprinting (Fig. 8; Kaemmer et al. 1992). DNA fingerprinting produces a dendrogram closely related to the evolutionary history of the banana clones under study (Fig. 9). Our future cooperative research aims at detecting individual differences in breeding populations after crossing or mutagenesis. Bioassays of plant-fungus (FOC) interactions Crude filtrates obtained from the FOC culture expressed toxic activities which depended on the age of Fusarium culture and final concentration in the banana tissue culture medium. Shoot tips of bananas were affected in proportion to the concentration of crude filtrate present in the culture medium. No conclusive differential reaction has been 154 Improvement of Musa through Biotechnology and Mutation Breeding

Figure 8. DNA amplification fingerprinting (RAPiD) of various Musa clones (see Fig. 7 for gel position). At the top is simplex RAPiD and at the bottom is duplex RAPiD. Molecular weight standards are indicated in kb. RAPiD fingerprinting detects genetic diversity among all 15 clones. Both amplimer combinations discriminate clones within the Cavendish group (f,g,h,i). RAPiD fingerprinting indicates (among others) differences between the original clone Grande Naine (h) and the mutant GN 60A (i). obtained with respect to the known response of the pathogen on banana cultivars in vivo, suggesting that toxic compounds excreted by the pathogen in culture might not be responsible for the different levels of resistance expressed by the clones. On the other hand, the very real fact that crude filtrates showed a toxic activity supported the idea that toxin(s) play some role in pathogenesis; but this could be due to a postinfectional process. Since the establishment of pathogenesis is a multiple step process it is likely that the different levels of resistance shown by Pisang Mas and SH-3362 in vivo were due to some mechanism of early plant-pathogen interaction including wall-to-wall recognition eliciting defense processes. FJ Novak et al. 155

Figure 9. Dendogram of the phylogenetic relationships among various Musa cultivars.

Wilting symptoms on plants cocultured with FOC began to appear at about 2-3 weeks after treatment for all three temperatures under study, although the degree of infection varied, the most severely infected being the treatment under 32°C, whereby the older leaves turned brown and there was no or a very scarce new root growth. By weeks 5-6, all the treated plants succumbed to the infection for 32°C, 70% for 25°C, and about 50% for 18°C. In contrast, the control for all temperatures showed good shoot and root growth throughout. The results indicated that the major temperature (32°C) seemed to have 156 Improvement of Musa through Biotechnology and Mutation Breeding

enhanced the infection and thus the appearance of the wilting symptoms. With the encouraging results obtained so far, this technique could be used as a potential tool for early screening for FOC resistance in vitro. Levels of peroxidase (PRX) activity differ in the two clones. SH-3362 showed a 10-fold higher activity in uninoculated plants than Pisang Mas. SH-3362 exhibited a sharp increase of peroxidase activity after inoculation with both race 1 and 4. Differences have been recorded on time of induction and activity in the two fungal races. In fact, PRX activity of plants treated with race 1 increased 5 days after inoculation while the increase was recorded 8 days after inoculation in plants treated with race 4. When SH- 3362 was treated with race 4 the PRX activity continued to increase up to 11 days, followed by a decrease while the plants inoculated with race 1 PRX activity started to decrease immediately after inoculation and rose once again after 13 days. These results were confirmed by IEF of infected and uninfected banana plants grown in growth chambers, showing that the increase of activity was mainly due to the basic class of peroxidases. In Pisang Mas we were not able to find a similar pattern both in total activity and in IEF. Our preliminary results on PRX seem to indicate that this enzyme is involved in an active defense mechanism against FOC, but that this mechanism is not the only one that counteracts Fusarium infection in banana, as demonstrated by the lack of response in Pisang Mas plants infected with FOC race 1. This is not surprising; in fact, the role of peroxidases has been studied in many plant-pathogen interactions and their role has not been unequivocally established in any plant pathogen system. More investigations are needed in banana to establish the actual role of peroxidases in the defense response to Fusarium.

Bioassays of Mycosphaerella toxins Each toxin isolated from MF exhibited phytotoxic activity and induced necrotic lesions only within clones of bananas and plantains except for juglone, which damaged every plant tissue tested. However, juglone has shown a certain degree of differentiation at very low concentrations (0.01 mg 5 µL-1) with a wide array of plants tested from dicotyledon and monocotyledon groups. Since juglone is produced by the pathogen in traces (Stierle et al. 1991), its differential activity at lower concentrations has special importance and needs further investigations at biochemical and molecular levels. In a preliminary test, juglone was found to stimulate shoot formation in poor shoot-forming banana clones at lower concentrations (<20 µM-1). This observation indicates its potential use as a growth regulatory substance. Fijiensin, isoochracinic acid and cinnamic acid have been found active at 5-10 µg 5 µL-1 levels of concentration, but they did not show a signiﬁcant differential activity within clones of different cultivars of bananas and plantains. Tetralone expressed host selectivity and a clearcut differential activity among different clones of banana and plantain cultivars even at the higher test concentrations (10 µg 5 µL-1). The varieties IV-9 (M. acuminata ssp. burmannica) and Saba, known as cultivars resistant to and tolerant of the Sigatoka disease complex in the ﬁeld, exhibited no reaction or a very mild reaction in vitro, whereas susceptible cultivars such as Boca, Horn plantain, and Grande FJ Novak et al. 157

Naine developed large necrotic areas, thereby showing conformity with the field observations. These ﬁndings suggest that tetralone could be an important host-speciﬁc toxin with a fair chance of its use in inducing new clones and screening varieties for resistance in tissue culture against the Sigatoka disease complex. However, such experiments must also be supplemented with the pathogen itself to establish a selection pressure in tissue culture using new techniques.

Conclusion The current status of Musa biotechnology is very close to making a significant contribution to the breeding of new cultivars. Tissue-culture techniques are being developed for the induction of heritable variation with a high potential for Musa breeding. In-vitro mutagenesis may contribute to cross-breeding programs to extend the genetic pool for recombination. Somatic embryogenesis has been realized in economically important cultivars. This technique is especially important in the study of “true” somaclonal variation in Musa and for the use of this unicellular system in mutation induction. Disease resistance is a main target for biotechnological research in Musa. Reliable selection systems of resistant genotypes against FOC may be soon available in juvenile plant stages, while in-vitro cell selection seems to be a long-term target for basic research before its application in practical breeding programs. Tetralone and juglone produced by MF offer great potential as a screening tool in the future, with reference to black Sigatoka disease.

Acknowledgments The Musa clones and strains of FOC were kindly provided by PR Rowe, FHIA, La Lima, Honduras, and M Smith, QDPI, Maroochy Experimental Station, Australia. Thanks are also due to GA Strobel, Department of Plant Pathology, Montana State University, Bozeman, MT, USA, for extending scientific advice regarding toxins produced by Mycosphaerella. This research was supported by the IAEA Fellowship Programme granted to S Hawa and A Khatri, as well as by a Special Service Agreement granted to RK Upadhyay by the Joint FAO/IAEA Division.

References

ALI S, MÜLLER CR, EPPLEN JT. 1986. DNA fingerprinting by oligonucleotide probes specific for simple repeats. Hum. Genet. 74:239-243. DALE JL. 1990. Banana and plantain. Pages 225-240 in Agriculture Biotechnology: Opportunity for International Development (Persley GJ, ed.). Wallingford, Oxon, UK: CAB International. KAEMMER D, AFZA R, WEISING K, KAHL G, NOVAK FJ. 1992. DNA oligonucleotide and DNA amplification fingerprinting of wild species and cultivars of banana (Musa spp.). Bio/Technology 10:1030-1035. NOVAK FJ. 1992. Musa (bananas and plantains). Pages 449-488 in Biotechnology of Perennial Fruit Crops (Hammerschlag FA, Litz R, eds.). Wallingford, Oxon, UK: CAB International. 158 Improvement of Musa through Biotechnology and Mutation Breeding

NOVAK FJ, AFZA R, VAN DUREN M, PEREA-DALLOS M, CONGER BV, XIAOLANG T. 1989. Somatic embryogenesis and plant regeneration in suspension cultures of dessert (AA and AAA) and cooking (ABB) bananas (Musa spp.). Bio/Technology 7:154-159. NOVAK FJ, AFZA R, VAN DUREN M, OMAR MS. 1990. Mutation induction by gamma irradiation of in vitro cultured shoot-tips of banana and plantain (Musa cvs.). Tropical Agriculture (Trinidad) 67:21-28. OMAR MS, NOVAK FJ, BRUNNER H. 1989. In vitro action of ethylmethanesulfonate on banana shoot tips. Scientia Hort. 40:283-295. PERSLEY GJ, DE LANGHE EA (eds). 1987. Banana and Plantain Breeding Strategies. ACIAR Proceedings no. 21. Canberra, Australia: ACIAR. SACCHI M, NOVAK FJ. 1992. Use of protein electrophoresis for characterization of genetic diversity in Musa. Personal communication. SCHENK RU, HILDEBRANDT AC. 1972. Medium and techniques for induction and growth of monocotyledonous and dicotyledonous plant cell cultures. Can. J. Bot. 50:199-204. STIERLE A, UPADHYAY R, HERSHENHORN J, STROBEL GA, MOLINA G. 1991. The phototoxins of Mycosphaerella fijiensis, the causative agent of Black Sigatoka disease of bananas and plantains. Experientia 47:853-859. UPADHYAY RK, STROBEL GA, COVAL SJ, CLARDY J. 1990. Fijiensin, the first phytotoxin from Mycosphaerella fijiensis, the causative agent of Black Sigatoka disease. Experientia 46:982-984. UPADHYAY RK, STROBEL GA, COVAL SJ. 1990. Some toxins of Mycosphaerella fijiensis. Pages 231-236 in Sigatoka Leaf Spot Diseases of Bananas (Fullerton RA, Stover RH, eds). Montpellier, France: INIBAP. VAKILI NG. 1967. The experimental formation of polyploidy and its effect in the genus Musa. Amer. J. Bot. 54:24-36. VUYLSTEKE DR. 1989. Shoot-tip culture for the propagation, conservation and exchange of Musa germplasm. Rome, Italy: IBPGR. WEISING K, BEYERMANN B, RAMSER J, KAHL G. 1991. Plant DNA fingerprinting with radioactive and digoxigenated oligonucleotide probes complementary to simple repetitive DNA sequences. Electrophoresis 12:159-169. M Perea-Dallos 159

Contribution to the study of Banana Anther Culture

M Perea-Dallos

Introduction Banana and plantain are two of the most important staple crops for several millions of people in the world (Cronauer, Krikorian 1984). The food value of Musa cultivars as a source of such minerals as calcium, phosphorus, potassium, magnesium, and iron, is already well known. They also have high contents of vitamins A and C, which have an important value in human nutrition, and additionally high amounts of carbohydrate (Novak 1992). Since the beginning of banana production, Colombia has been one of the most important countries for the export of bananas and plantains. The main banana-producing area, Uraba, covers an area of 42 000 ha, with a yield of 246 000 t a-1. Most plantain farms are located on the mountains, where coffee is the main crop. Under these conditions, plantain is grown to provide needed supplies for the growers as a “survival crop.” It gives “security and cheap food,” not only for peasants in the growing areas, but also for different Colombian social groups. The genetic improvement of banana and plantain through conventional breeding methods is hampered because the pollen grains and ovule do not produce fertile seeds (Novak 1992). Sterility in edible bananas is the result of a complex of factors, the most important of which is triploidy and attendant meiotic anomalies, and parthenocarpy and sterility have arisen through gene mutation in fertile diploids (Krikorian, Cronauer 1984; Rowe 1984; Shepherd 1987; Simmonds 1966). The combination of traditional cross-breeding with biotechnology could provide a powerful tool for the development of new clones with desirable features, such as, for example, resistance to or tolerance of major diseases. Diploid clones of both wild and edible bananas are essential starting material for banana breeding as such sources of genetic resistance or tolerance, except for bunchy top disease which has been identiﬁed among the extensive collections of diploid accessions of Musa acuminata (Novak et al. 1989).

Anther Culture The importance of crop improvement through anther culture has been widely discussed, and signiﬁcant advances have been achieved in the last two decades.

Departamento de Biologia, Facultad de Ciencias, Universidad Nacional de Colombia, AA 23227, Bogotá, Colombia 160 Contribution to the study of Banana Anther Culture

Anther culture of monocotyledons has been extensively researched for many years, because of the importance of the crops that belong to this group. However, the low frequency of success, as a result of the very strong genotypic effects and, in some cases, a high proportion of albino regenerants, have seriously hindered progress in this area (Kasha et al. 1991). To date (1993), a great majority of plant species have been successfully regenerated from isolated anthers or grains of pollen at different stages of development. However, in Musa clones, especially the cultivated clones, most of the genotypes do not give any pollen grains, and eggs of diploid clones are haploids. This phenomenom excludes many valuable clones from hybridization. In view of the importance of haploids in plant breeding, more genotypes need to be tried in order to determine the role of genotypes in regard to media composition and environmental conditions, with speciﬁc reference to cell division and differentiation. Anther culture is not frequently used in Musa breeeding due to genotypically dependent differentiation of the genome, but there are efforts to resolve this problem.

Materials and Methods Plant material Experiments were carried out with three selected diploid clones of banana: Bocadillo, Pisang Lilin, and Malascensis. This material was collected from nursery fields of the Instituto Colombiano Agropecuario (ICA) at Carepa-Uraba. In order to study the appropiate stage of inﬂorescence, they were chosen at different stages of development, as for example, when the bunch had emerged (15 cm), and when the bunch had grown to about 30-35 cm (i.e., when it began to bend down). To allow for the asynchronic development of anthers, several hands were selected for research. The most appropiate stage of development was when inﬂorescences had 1-12 hands. The pollen grains were thus cultured when they reached the uninucleate stage and when the exine and intine layers were observable.

Method of anther culture After collection, the inflorescences were stored at 4°C for 4-7 days. This material was disinfected in 5% sodium hypochlorite (Clorox) for 45 min, followed by several rinses in sterile water. The isolation of anthers was carried out under aseptic conditions and then placed onto Murashige and Skoog (1962) basal medium, supplemented with various concentrations of Dicamba (3,6-dichloro-o-anisic acid at 5, 10, 15, 20, 25 µM), 2 g L-1 casein hydrolysate, and 60 g L-1 sucrose; the pH was adjusted to 5.8. Four anthers were placed in each vessel, and then incubated under dark conditions at 28°C. To avoid browning due to the phenolic compounds, tests were carried out at several concentrations of activated charcoal (0.5, 1.0, and 2.0%), and, after the appearance of the calli, subcultures were done every week. M Perea-Dallos 161

Results The first evidence of callus appeared after 26 days of culture. The uniformity of cell development was observed in all anthers of the inﬂorescence. The reduction of browning was noneffective through the activated charcoal treatment and the calli became dark after several days (1 week). However, cell proliferation continued and, after 8 weeks of anther culture, globular-like structures were formed. Table 1 and Figure 1 show the effect of callus formation and growth from the anther culture of diploid cultivars of banana in different concentrations of an auxin (Dicamba). Table 1. Base data from which Figure 1 is derived. Concentration % Callus formation and growth of Dicamba (µM) Bocadillo Pisang Lilin Malascensis 5887550 10 75 58 37 15 49 35 25 20 32 28 12 25 15 10 8

Figure 1. The effect of callus growth from anther cultures on three diploid bananas. 162 Contribution to the study of Banana Anther Culture

Discussion As callus induction manifests rapid and intensive blackening, it is necessary to make subcultures every week and incubate them in dark conditions. Efforts are in progress to develop structure-like embryos. Acknowledgments The author is indebted to the IAEA/FAO Joint Programme for its support for part of this work. Also to R Bayona, Director of the Laboratorio de CENIBANANO and M Mayorga, Director of Regional ICA at Carepa-Uraba for the provision of plant material. Also to AUGURA and the Universidad Nacional for the support of this research. Gratitude is additionally expressed to M Espitia for his help in typing the manuscript.

References

CRONAUER SS, KRIKORIAN AD. 1984. Multiplication of Musa from excised stem tips. Ann. Bot. 53. KASHA KJ, ZIAUDDIN A, REINBERGS E, FALK E. 1991. Use of haploids in induced mutation in barley and wheat. In Proceedings of the Second FAO/IAEA Meeting: Use of Induced Mutation in connection with Haploids and Heterosis in Cereals, Katowice, Poland (Maluszynski M, Barabas Z, eds). KRIKORIAN AD, CRONAUER SS. 1984. Banana. In Handbook of Plant and Cell Culture, vol 2 (Sharp W, Evans DA, Ammirato PV, Yamada Y, eds). New York, USA: Macmillan. MURASHIGE T, SKOOG F. 1962. A revised medium for rapid growth and bioassays with tobacco tissue cultures. Physiol. Plant. 15:473-497. NOVAK FJ. 1992. Musa (bananas and plantains). In Biotechnology of Perennial Fruit Crops (Hammerschlag FA, Litz RE, eds). Wallingford, Oxon, UK: CAB International. NOVAK FJ, AFZA R, VAN DUREN M, PEREA-DALLOS M, CONGER BV, XIOLANG T. 1989. Somatic embryogenesis and plant regeneration in suspension culture of dessert (AA and AAA) and cooking bananas (ABB), Musa spp. Biotechnology 7:154-159. ROWE P. 1984. Breeding bananas and plantains. Plant Breeding Reviews 2:135-155. SIMMONDS NW. 1966. Bananas, 2nd edn. London, UK: Longman. SHEPHERD K. 1987. Banana breeding, past and present. Acta Horticulturae 196:37-43. MR Söndahl, C Noriega 163

Bioreactor Micropropagation Technology for Rapid Commercialization Opportunities in Plantation Crops

MR Söndahl, C Noriega

Importance of Micropropagation Vegetative propagation includes budding, cuttings, and grafting, and micropropagation relates to all in-vitro techniques of cloning. Most current commercial plant micropropagation methods involve exploring the presence of axillary or apical meristems of donor genotypes (Kurtz et al. 1991). The cultures are maintained on semisolid medium, and the process requires three steps: explant inoculation and shoot development, shoot multiplication, and rooting/hardening-off. Each step requires a great amount of labor and laboratory space for incubation of the semisolid cultures. An enormous potential for short-term return can be realized through the utilization of vegetative propagation in perennial crops. This technique seems especially attractive for many tropical species that possess a high degree of heterozygosity and suffer from systemic infections. Examples of economically important crops where vegetative propagation is being used include: palms (oil palm, date palm, peach palm), banana, agave, bamboo, orchids, coffee, cacao, plantain, colocasia, cassava, potato, pineapple, and many others (Söndahl et al. 1983). Most of the time micropropagation provides an economic advantage for the propagation of a particular crop. In other cases, there is a technical advantage in cloning a selected genotype, as, for example, when seed propagation procedures will not sufﬁce because there is heterozygosity and segregation in the next generation. A case-by-case analysis is necessary to determine the pros and cons in applying micropropagation to a crop. Usually in-vitro propagation can provide very large numbers of cloned materials rapidly, using small, confined facilities. If these advantages are applied properly to a particular vegetatively propagated species, micropropagation techniques will be commercially superior to other conventional methods of propagation. Micropropagation methods have also been applied extensively for disease elimination. Most of the systemic parasites (virus, bacteria, fungi) are present in the

DNA Plant Technology Corporation (DNAP), 2611 Branch Pike, Cinnaminson, NF 08077, USA 164 Bioreactor Micropropagation Technology for Rapid Commercialization Opportunities

vascular system of plants. The culture of small meristems or cells can eliminate the great majority of such diseases because the vascular system is not yet differentiated or is completely absent. Sometimes, heat treatment is associated with meristem culture for more efficient disease elimination. There are cases, however, when cuttings or graftings cannot be applied and micropropagation techniques would be the only commercial solution for crop propagation (oil palm being one example). Many tropical plants fit into this category mainly due to their high levels of phenols and tannins which may inhibit rooting and grafting. Many other reasons for using vegetative propagation by conventional methods would also apply to micropropagation methods; for example, maintenance of heterozygosity, cloning superior individual plants, presence of sexual sterility, or incompatibility problems. Production of sufficient numbers of plants of a unique genotype for field evaluation followed by further developmental studies may be another objective of micropropagation, especially in the case where seed production is not possible. Finally, it also may be useful to apply micropropagation to establish seed orchards for bulk seed production or hybrid seed synthesis (cloning parental lines). Great advantages can be gained by applying micropropagation methods to perennial crops, due to the long time required for breeding. Here, one can easily evaluate the commercial advantages of cloning selected individual trees, thus bypassing all the time involved in out-crossing and selecting homozygous seed donors. The advantages would not only relate to the time to establish high-value commercial plantations, but also to the preservation of heterozygosity and plasticity in such cloned populations. Fruit trees, oil and coconut palms, ornamental trees, and woody species for pulp and lumber are the main candidates for micropropagation methods. Certainly the most appropriate technique (meristem culture, somatic embryo, or shoot production) will be determined, in part, by the kind of tree species, the economic value of each individual tree, and the size of the operation. Vegetative propagation by culturing apical and axillary meristems has been explored quite effectively and proved to be a reliable as well as a commercial method for many plant species. However, in many perennial trees, the number of available meristems is not sufficient to support a large-scale vegetative propagation, when use of somatic embryos would be the next best choice (Gupta et al. 1991). High-frequency production of somatic embryos has been described for more than 60 plant species. However, the challenge is now to produce such embryos from cultured tissues of the most important crop species. Moreover, there is a culture medium/genotype interaction and the production of large numbers of somatic embryos will require optimization for the elite genotypes. With time, field trials will provide information to ascertain the degree of fidelity of cloned plants derived from somatic embryos, as well as organogenic shoots. In this particular aspect of clonal fidelity, the experience gained with one plant species should not be extrapolated in totum to other species. The micropropagation method adopted, the source of initial explant, the genetic background of the stock plants among others, may determine the value and fidelity of a particular cloned species. Quick and reliable methods for evaluating clonal fidelity as early as possible in the cloning process are highly desirable. MR Söndahl, C Noriega 165

Somatic embryogenesis has been established as a means of regenerating plants from single cells. Somatic embryos can be produced via “direct” and “indirect” embryogenesis. Somatic embryos obtained by direct embryogenesis are more limited in numbers than embryos derived from indirect embryogenesis. In the latter case, the cells undergo extensive proliferation (formation of friable embryogenic tissue) before differentiation into normal embryos occurs (Söndahl et al. 1985). This self-cloning process may or may not produce a certain degree of variation among plants derived from either direct or indirect embryogenesis.

New Advances Widespread use of micropropagation for major crops in agriculture and forestry is still restricted as a result of relatively high production costs. These high costs in micropropagation are mainly due to high labor costs, limited rates of growth and development of plantlets in culture vessels, and low percentage survival of plantlets when transferred to greenhouse conditions (Kozai et al. 1991). New micropropagation processes are being developed whereby the cloned plants will be produced inside bioreactor vessels taking advantage of regeneration systems via somatic embryogenesis, or organogenesis, or even axillary shoot multiplication (Ammirato, Styer 1986; Cazzulino et al. 1991). The introduction of robotics in several steps of the process is also under development. The advantage of these new micropropagation technologies is the possibility of producing a much larger number of cloned plants in a smaller laboratory space, with reduced labor input, with, probably, a potential reduction in the total process time. A theoretical example with an embryogenic cell suspension system can illustrate the potential of the bioreactor micropropagation technique: A 1-L bioreactor vessel charged with embryogenic suspension cells at a density of 1 x 106 cells mL-1 and capable of 0.1% embryo regeneration would yield the following numbers of cloned material: - somatic embryos = 1 000 000 - germinated embryos = 500 000 -plantlets = 375 000 - young plants in greenhouse = 280 000 Note: These ﬁgures assume efﬁciencies of germination at 50%, plantlet development at 75%, and hardening at 75%. The use of a bioreactor will increase several-fold the multiplication capacity of one particular cell line. Also, several units can be operated simultaneously for multiple lines. Bioreactors will permit a high degree of culture control and the process is amenable to robotics. Low-cost plant bioreactor units can be devised for commercial micropropagation use. The risks of this process will be the incidence of contamination and the danger of some degree of variability among cloned individuals due to excessive multiplication cycles (long-term cultures). To overcome these problems, maximum aseptic techniques and multiple bioreactor units must be implemented. In addition, it 166 Bioreactor Micropropagation Technology for Rapid Commercialization Opportunities

may be possible to monitor clonal ﬁdelity of the propagated plants by rapid molecular screening methods, e.g., RAPD (random ampliﬁed polymorphic DNA). The micropropagation process via bioreactor vessels (Fig.1) will achieve its maximum efficiency if the regeneration process is based on somatic embryogenesis. Nevertheless, it will also provide advantages to propagation systems based on organogenesis, or axillary shoots. As the regeneration process changes from embryogenesis to organogenesis to axillary shoots, the multiplication rate will decrease. A general scheme for exploring the micropropagation process via bioreactor cultures is outlined in Figure 1.

Figure 1. Bioreactor micropagation ﬂow. MR Söndahl, C Noriega 167

Cloning in a liquid phase via bioreactor vessels has the potential to (a) reduce both cost and total time; and (b) substantially increase the availability of high-quality cloned materials. For example, a coffee micropropagation process based on liquid cultures will produce somatic embryos in large quantities. A recent communication from a tissue- culture laboratory in France demonstrated production rates of 460 000 robusta coffee embryos per liter after 7 weeks in Erlenmeyer flasks, or 600 000 embryos per 3-L bioreactor vessel after 8 weeks (Zamarripa et al. 1991). Our preliminary data with arabica coffee tissues (Yellow Catuai) revealed that the production of 1 million coffee embryos would require an initial inoculum of 80 g f.w. of embryogenic cells cultured in four 5-L bioreactor vessels. Efforts are being made to develop similar embryogenic systems for oil palm and roses. We are also interested in exploring the bioreactor culture technique for micropropagation of bananas and pineapple. These two crops are very amenable to propagaton via axillary shoot multiplication, and so the development of similar protocols for bioreactor vessels seems to be highly feasible. Several other research groups in the USA, Europe, and Japan are working to develop bioreactor micropropagation systems for celery, alfalfa, pinus, poinsettia, gladiolus, and lilies. Over several years we have researched the application of bioreactor vessels and/or liquid cultures for large-scale micropropagation. Initial efforts were made in a joint R&D program with Arthur D. Little Inc., which focused on adapting a spin filter bioreactor design to embryogenic cell systems such as carrot and celery (Ammirato, Styer 1986; Styer 1986). This line of research has continued to further develop this technique for coffee, oil palm, and roses. If successful with these species, other ornamental plants, industrial plant species, or forest trees should follow. The bioreactor micropropagation technology will offer advantages to this field, but it will also bring new technical challenges. This propagation process will produce approximately 0.5 million units per bioreactor unit every 5-10 weeks, and so new methods need to be devised to handle such large numbers to go through the next phases of the micropropagation process, i.e., germination, plantlet development, and hardening. Also, there is a great deal of potential to introduce robotics and automation in this process as well as the use of imaging systems to screen embryos at the proper morphological stage. Recent developments of robotic micropropagation systems have been reviewed by Kozai et al. (1991). Toshiba Corporation (Japan) developed a robot that mimics a series of manipulations by an operator at the laminar flow hood, and it is testing the system with such plants as potato, carnation, eucalyptus, and redwood. Silsoe Research Institute (England) is working on robotic micropropagation with a vision system that identifies Y- shaped nodes. Commonwealth Industrial Gases (Australia) is also developing robotic micropropagation with a vision system for daisies (Chrysanthemum cinerariaefolium), which it claims is able to produce more than 10 million plantlets per year. Kirin Brewery (Japan) and Twyford (USA) are jointly developing a robotic multiplication system without vision analysis for Ficus benjamin and ferns. Miwa and Mitsubishi (Japan) are developing a system for explanting bulb scales as a rate of 4800 scales per day. Komatsu Ltd (Japan) has a system for plants with upright shoots, such as potato. Mitsui and 168 Bioreactor Micropropagation Technology for Rapid Commercialization Opportunities

Ohtake Techno Union (Japan) jointly developed a robotic multiplication for sectioning (10-20 cm each) and transplanting segments of miniature roses. Automation of micropropagation systems will be a key development step for labor saving and better environmental control, thus meeting the increasing demands for low- cost and high-quality transplants in agriculture, forestry, and horticulture. Commercial mass production of transplants in the nursery industry is expected to increase in the order of 10 times, 100 times, and 1000 times if the production cost is decreased by 50%, 75%, and 90%, respectively (Kozai et al. 1991). An increasing number of farmers and growers will be inclined to buy transplants rather than raise them by themselves if the transplants are available at a higher quality and a lower price. Such a drastic reduction of production costs is most likely to be achieved by combining bioreactor micropropagation with automation. Banana and plantation crops would be natural candidates to receive the beneﬁts of future micropropagation technology, which should stimulate the clonal propagation market and facilitate production management.

References

AMMIRATO PV, STYER DJ. J. 1986. Strategies for large-scale manipulation of somatic embryos in suspension culture. Pages 161-178 in Biotechnology in Plant Science: Relevance to Agriculture in the Eighties (Zaitlin M et al., eds). New York, USA: Academic Press. CASSULINE D, PEDERSON H, CHIN CK. 1991. Bioreactors and image analysis for scale-up and plant propagation. Pages 147-177 in Scale-up and Automation in Plant Propagation (Vasil IK, ed.). New York, USA: Academic Press. GUPTA PK, TIMMIS R, PULLMAN G, YANCEY M, KREITINGER M, CARLSON W, CARPENTER C. 1991. Development of an embryogenic system for automated propagation of forest trees. Pages 75-93 in Scale-up and Automation in Plant Propagation (Vasil IK, ed.). New York, USA: Academic Press. KOZAI T, TINE KC, AITKEN-CHRISTIE J. 1991. Considerations for automation of micropropagation systems. Pages 503-517 in Automated Agriculture for the 21st Century. Chicago, USA: American Society of Agricultural Engineers. KURTZ SL, HARTMAN RD, CHU IYE. 1991. Current methods of commercial micropropagation. Pages 7-34 in Scale- up and Automation in Plant Propagation (Vasil IK, ed.). New York, USA: Academic Press. SÖNDAHL MR, MAKAMURA T, SHARP WR. 1985. Propagation of coffee. Pages 215-232 in Tissue Culture in Forestry and Agriculture (Henke RR et al., eds). New York, USA: Plenum Press. SÖNDAHL MR, SHARP WR, EVANS DE. 1983. Biotechnology of cultivated crops. Pages 98-114 in Biotechnology: the Challenges Ahead. Singapore: Science Council. STYER DJ. 1986. Applications and bioreactor technology to seed production and plant propagation. Pages 112- 127 in Proc. Agriculture and Life Sciences in China. Seed Certiﬁcation and Seed Technology. Beltsville, USA: Inst. Intl. Dev. Edu. Agric. Life Sci. ZAMARRIPA A, DUCOS JP, TESSEREAU H, BOLLON H, ESKES AB, PETIER V. 1991. Développement d’un procédé de multiplication en masse du caféier par embryogenèse somatique en milieu liquide. Proc. 14th Intl. Conf. Coffee Science, San Francisco, USA. pp.44. D Vuylsteke, R Swennen 169

Genetic Improvement of Plantains: the potential of conventional approaches and the interface with in-vitro culture and biotechnology

D Vuylsteke, R Swennen1

Introduction Genetic improvement of banana and plantain has gained renewed impetus over the past few years (Persley, De Langhe 1987). One of the reasons for this interest may be found in the increased awareness among both the donor and the international and national research communities of the importance of banana and plantain as a local food crop and as a source of revenue for the many smallholders that grow the crop, particularly in sub-Saharan Africa. Furthermore, the interest in banana and plantain breeding also emanates from the increased disease and pest pressure that is threatening the crop’s productivity and sustainability. Apart from the potential of cultural and/or biological interventions to control some of the diseases and pests affecting Musa, host-plant resistance is often perceived as the most appropriate technology to check yield losses due to biotic (and abiotic) stresses. Yet banana breeding is generally considered to be notoriously difﬁcult due to the speciﬁc biology of the preferred and most widely cultivated varieties. Triploidy and low levels of fertility have resulted in the perception that the crop is intractable in terms of conventional breeding approaches. The fact that 70 years of classical breeding endeavors in the Caribbean and Central America have achieved limited success—no man-bred cultivar has yet been grown successfully (Rowe 1984)— is certainly not strange to this view. But one must bear in mind that the primary objective of those earlier programs, i.e., the production of a disease-resistant banana for the export trade (only 10% of world production), may be of limited relevance to the breeding targets and methodologies for the bulk of the locally consumed bananas and plantains. Because of the crop’s recalcitrance to genetic improvement, novel approaches have been explored. Biotechnology has been proposed as a possible avenue for providing an

International Institute of Tropical Agriculture, Oyo Road, PMB 5320, Ibadan, Nigeria. 1Present address: Laboratory of Tropical Crop Husbandry, Katholieke Universiteit Leuven, Kardinaal Mercierlaan 92, 3001 Heverlee, Belgium 170 Genetic Improvement of Plantains

array of techniques to overcome the limitations of conventional techniques (Dale 1990). In some cases, biotechnology has even been advocated as the sole option to banana and plantain improvement, for example in producing materials resistant to banana bunchy-top virus (BBTV). In this paper, we would like to demonstrate the potential of conventional, yet modified, breeding approaches by providing an example of plantain improvement work in which tissue-culture techniques were fully integrated. In-vitro culture techniques and biotechnology have been and will continue to be valuable tools in the improvement of the crop, but we would like to emphasize, as many others have done, that this should be pursued only when integrated in breeding programs in order to maximize benefits from the limited human and financial resources devoted to genetic improvement.

Plantain Improvement at IITA In 1987, the International Institute of Tropical Agriculture (IITA) initiated a plantain and banana improvement program in response to the rapid spread of the virulent black Sigatoka disease (Mycosphaerella fijiensis Morelet), which is generally considered the major constraint to plantain and banana production worldwide. The program targets the incorporation of durable host-plant resistance to Sigatoka leaf spot in the plantains (Musa spp., AAB group) of western and central Africa, the highland cooking and beer bananas (Musa spp., AAA and ABB group) of eastern Africa, and ABB cooking bananas. Black Sigatoka disease causes yield losses of 30-50% in plantains (Stover 1983; IITA 1992) and further compounds the pest complex attacking the crop in sub-Saharan Africa. This warrants investigations into appropriate control strategies and provides the context for IITA’s work in this area. Because of earlier reports on the extreme difficulties encountered in plantain breeding (De Langhe 1976; Rowe 1984), the program started by screening the plantain collection (113 cultivars) for female fertility. Several different plantain cultivars were identified as seed-fertile as they produced viable seed upon hand-pollination. Twenty- nine French plantain cultivars set seed at averages ranging from less than 1 to over 20 seeds per bunch. Also, 8 False Horn plantains were found to be able to produce seed, yet generally at average rates below 1 seed per bunch. In addition to genotype, season was also found to have a marked influence on seed-set (Swennen, Vuylsteke 1990). The early identification of these 37 seed-fertile plantain cvs, and their relatively high seed- set, challenged the concept of female sterility in plantains and set the pace for IITA’s increased efforts in pursuit of conventional breeding. The breeding program was strongly supported by complementary plant tissue- culture work. Simple in-vitro culture techniques, such as shoot-tip culture and embryo culture, greatly facilitated the breeding activities by surmounting some of the obstacles impeding breeding progress. The most fertile plantains, among which the French plantain cvs Bobby Tannap, Obino l’Ewai, Mbi Egome 1, and Bungaoisan were rapidly propagated in vitro to a total D Vuylsteke, R Swennen 171

of 6000 plants, to set up a massive pollination scheme. Annual seed production from crosses of plantains with different diploid male parents averaged around 10 000. Only about 40% of these were viable seeds, containing a zygotic embryo embedded in endosperm. In-vitro embryo germination varied from 5 to 30%, depending on the cross and the season (Vuylsteke et al. 1990a). Of the many seedlings produced, about half were rogued at the nursery stage because of leaf abnormalities typical of aneuploidy and euploidy in excess of tetraploidy. About 250 plantain progenies — diploids, tetraploids and triploids — were field- established for evaluation. Noteworthy was the occurrence of a larger number of diploid (78%) than tetraploid (22%) progenies from the triploid (AAB plantain) by diploid (AA) crosses. Triploid progenies were very few. The frequency of tetraploids also differed according to the plantain cultivar used in crosses with the same male diploid. For example, cv Obino l’Ewai gave 42% tetraploid progenies, while cv Bobby Tannap produced only 12% tetraploids. Among the ﬁeld-established progenies, 20 plantain-derived tetraploid hybrids were selected for their reduced susceptibility to black Sigatoka disease, their high yields and good fruit parthenocarpy (Table 1). They also had generally shorter plant stature and better ratooning, but required about 1 month longer for the bunch to reach maturity. Most tetraploid hybrids demonstrated a capacity for seededness due to increased pollen fertility. The wild diploid banana clone Calcutta 4 was the male parent of 17 of the 20 selected hybrids. Nevertheless, fruit parthenocarpy and bunch size of the tetraploid hybrids was not inferior to that of their French plantain parents (Fig.1). This indicates that the poor bunch characters of this wild, nonparthenocarpic banana were generally not transmitted to the tetraploid progenies. Conversely, the high level of black Sigatoka resistance of Calcutta 4 was readily inherited in its offspring. Variation in growth and yield parameters (Table 1), qualitative morphological traits, and black Sigatoka reaction (Table 2) were observed within the same tetraploid family obtained from crosses of plantain with the male diploid parent Calcutta 4, a true- breeding line. This observation was surprising because one would expect an entirely uniform progeny from the combination of unreduced female gametes and the male gametes of a true-breeding species. Indeed, Calcutta 4 is a true-breeding line because its selfed progeny did not show segregation for black Sigatoka reaction or morphological traits (Simmonds 1952; IITA, unpublished results). Therefore, we infer the occurrence of segregation and recombination in the triploid plantain genome during the modiﬁed megasporogenesis, even when there is restitution of all three sets of maternal chromosomes. This inference challenges the “dead end” target of the primary triploid by diploid cross (Stover, Buddenhagen 1986; Bakry et al. 1990). The concept of banana/plantain breeding being essentially breeding-improved diploid pollinator lines (Rowe 1984; Shepherd 1987; Simmonds 1987; Stover, Buddenhagen 1986) may thus need to be reviewed, as variability is also increased by segregation on the female triploid side. 172 Genetic Improvement of Plantains

Table 1. Preliminary evaluation of growth and yield parameters and of black Sigatoka resistance in the plant crop of some selected tetraploid plantain hybrids (nonfungicide treated) as compared with their respective triploid female plantain parents (fungicide-treated). Cultivar and Male Plant YLS2 at HTS3 at Fruit Bunch derived hybrids parent1 height ﬂowering harvest ﬁlling weight (cm) (cm) time (days) (kg) Obino l’Ewai - 370 7.0 188 92 12.4 TMPx 548-9 C4 290 11.0 310 131 16.5 TMPx 2637-49 C4 360 11.6 225 127 16.6 TMPx 4698-1 C4 345 10.5 275 130 20.0 TMPx 5511-2 C4 345 10.0 165 118 17.8 TMPx 5706-1 C4 340 7.5 250 133 13.6 TMPx 6930-1 C4 330 10.3 240 134 18.5 TMPx 1658-4 Pl 320 9.3 185 134 21.5 Bobby Tannap - 340 7.0 171 92 14.0 TMPx 582-4 C4 300 11.0 270 135 14.3 TMPx 4479-1 C4 295 9.0 290 114 13.2 TMPx 2796-5 Pl 335 10.0 205 123 21.3 1 Male parents: C4 = Calcutta 4; Pl = Pisang Lilin. 2 Youngest leaf spotted (Stover, Dickson 1970). 3 Height of tallest sucker.

Table 2. Segregation of black Sigatoka reaction and other qualitative traits among selected tetraploid progenies obtained from crosses of the triploid French plantain (Musa spp., AAB group) cultivars Obino l’Ewai and Bobby Tannap with the true- breeding wild diploid M. acuminata ssp. burmannicoides clone Calcutta 4. Cultivar and Black Sigatoka Bunch Neutral Male bud Ratooning derived hybrids reaction orientation ﬂowers imbrication Obino l’Ewai susceptible pendulous persistent yes inhibited TMPx 548-9 mod. resistant1 pendulous deciduous yes regulated2 TMPx 2637-49 mod. resistant pendulous deciduous no regulated TMPx 4698-1 mod. resistant pendulous semipersistent3 yes regulated TMPx 5511-2 mod. resistant pendulous deciduous yes inhibited TMPx 5706-1 less susceptible pendulous deciduous no regulated TMPx 6930-1 mod. resistant pendulous deciduous yes regulated Bobby Tannap susceptible subhorizontal persistent yes inhibited TMPx 582-4 mod. resistant subhorizontal deciduous yes regulated TMPx 4479-1 less susceptible pendulous deciduous yes regulated 1Moderately resistant. 2Regulated: 1 to 3 suckers develop freely while others remain inhibited. 3Semipersistent: some neutral ﬂowers persist but the majority are deciduous. D Vuylsteke, R Swennen 173

Figure 1. The ﬁrst hand of the black Sigatoka-resistant tetraploid TMPx 548-9 (left) and of its female plantain parent Obino l’Ewai (right).

In-Vitro Culture and Biotechnology in support of Plantain/Banana Breeding

Shoot-tip culture Increased impetus for genetic improvement programs has focused attention on the collection, movement, and conservation of Musa germplasm. However, efforts to handle germplasm of this vegetatively propagated crop are typically fraught with a series of obstacles, such as slow multiplication, bulkiness, and quarantine-related problems of conventional propagules. Hence, a wide array of plant tissue-culture techniques is increasingly being used as an enabling and enhancing technology for the handling of Musa germplasm (Vuylsteke 1989). Aseptic shoot-tip culture, in combination with third-country quarantine, has been used as a vehicle for the safe exchange of banana/plantain germplasm (Vuylsteke et al. 1990b). Shoot-tip cultures confer considerable advantages for the international transfer of germplasm because the mass of plant material involved in the movement is greatly reduced and contained, and they overcome nearly all of the problems associated with nonobscure pests and pathogens. From 1985 to 1990, 320 new accessions were introduced by IITA as in-vitro cultures, thereby quadrupling the number of accessions held in the collection. These genetic resources, which provide the basis for the breeding program, were introduced in a joint effort with the Nigerian Plant Quarantine Service 174 Genetic Improvement of Plantains

and the INIBAP Transit Center at KU Leuven (Belgium). IITA has also deposited 17 hybrids of banana and plantain at the INIBAP Transit Center for virus indexing in view of their possible future international dissemination. Shoot-tip culture is a well-established, adequate, and relatively simple in-vitro method for the rapid propagation of selected Musa materials and the production of clean planting material. Multiplication rates are several orders of magnitude higher than those obtained with conventional methods, which is of great value for the multiplication of newly bred genotypes or to speed up the testing of selections by plant breeders. Micropropagation has been pivotal in the rapid deployment of IITA’s breeding program by supplying large numbers of plants of female and male parents for the crossing blocks and of black Sigatoka-resistant hybrids for the evaluation trials.

Embryo culture/rescue Hybrid plant production in the most common triploid Musa clones is hampered by low seed-set and by low seed germination rates. Seeds of plantain crosses germinate in soil at a rate of only 1%. Aseptic embryo culture techniques, already in use in banana breeding for 3 decades, are routinely applied to increase seed germination rates by a factor of 3 to 10. An average of about 900 plantain hybrid seeds are handled in vitro on a monthly basis, in support of the breeding program. Investigations into rescuing immature embryos are under way further to enhance germination rates. Preliminary results showed that immature embryos of 50-55 days after pollination had 20.7% germination, as compared with the 1.0% germination of mature embryos (80-85 days after pollination).

Cell-suspension cultures and somatic embryogenesis Though we have attempted to demonstrate above that conventional breeding of plantains has great potential, certain improvement targets remain virtually impossible without the input of molecular genetic techniques. Such procedures, however, largely depend on the successful regeneration of plants from cells or protoplasts. Prospective beneﬁts that might accrue from the integration of biotechnologies into banana and plantain breeding programs therefore require access to reliable cell culture protocols (Krikorian 1987). Yet, reports on cell suspension cultures of Musa are few. Moreover, most procedures for cell suspensions are based on callus (Novak et al. 1989; Escalant, Teisson 1989), which has obvious disadvantages in applications that require clonal uniformity, such as the recovery of transgenic plants. In addition, some procedures use zygotic embryos to produce suspension cultures (Escalant, Teisson 1989), therefore excluding the edible bananas. Recently, however, an elegant protocol for direct embryogenesis in somatic cell suspensions initiated from meristem explants was developed at KU Leuven, Belgium (Dhed’a et al. 1991). These embryogenically competent cell cultures have already proven to be the material of choice for the cryopreservation of germplasm (Panis et al. 1990; Panis et al. 1991), and protoplast culture (IITA unpublished results), and are currently evaluated for plant transformation studies. D Vuylsteke, R Swennen 175

Somaclonal variation The increased genetic variation among plants regenerated from in- vitro culture has been termed as somaclonal variation, an ubiquitous phenomenon. The frequent use of in- vitro culture techniques for the handling of Musa germplasm warrants investigations into the occurrence of somaclonal variation in this genus. Somaclonal variation in banana and plantain plants derived from shoot-tip culture can be a significant problem in micropropagation and has been described by many authors (see Israeli et al. 1991; Vuylsteke et al. 1991, and references cited therein). Somaclonal variation had earlier been proposed as a source of new useful variability for banana and plantain improvement, but it is now generally agreed that it should not be overestimated as a beneﬁcial adjunct to breeding. On the contrary, somaclonal variation will be of particular concern when new genotypes are rapidly propagated in vitro, either produced by conventional means or by recombinant DNA technology. In this framework, the usefulness of mutagenesis should also be questioned. The authors’ experience with plants recovered from gamma-irradiated meristems was that it did not create any valuable variability, but rather resulted in many grossly abnormal plants.

Conclusion Results from plantain breeding at IITA emphasize that promising black Sigatoka- resistant tetraploid hybrids, which combine high yields, adequate plant height, and improved ratooning, can be produced when using a wild diploid banana as a male source of disease resistance. In addition, seed production in a range of highland cooking and beer bananas (AAA group), and the recent planting of their first progenies, suggests that this major eastern African crop is also amenable to conventional genetic improvement. Our results challenge the commonly accepted concept of the intractability of plantain for genetic improvement by conventional breeding methods. The utilization of interspecific hybridization, embryo culture, rapid in-vitro multiplication, field-testing, and selection, has resulted in the identification of improved Musa (plantain) germplasm at IITA. Established tissue-culture techniques have thus contributed significantly to banana/plantain improvement. Also, molecular genetic methods are becoming available for investigating genetic variability (Gawel, Jarret 1991). RFLP and RAPD mapping will assist breeding programs by enabling early identification of promising new genotypes. Notwithstanding the recent success in conventional breeding at IITA, recombinant DNA technology may be of great benefit in the improvement of Musa cultivars that are difficult to breed (e.g., Cavendish) and to incorporate genes coding for characters that are not available in the Musa gene pool, such as BBTV resistance. We, however, advocate that biotechnology projects be fully integrated into conventional breeding programs in order to take full advantage of biotechnological developments and products. 176 Genetic Improvement of Plantains

References

BAKRY F, HORRY JP, TEISSON C, TEZENAS DU MONTCEL H, GANRY J. 1990. L’amélioration génétique des bananiers à l’IRFA/CIRAD. Fruits numéro spéciale: 25-40. DALE JL. 1990. Banana and plantain. Pages 225-240 in Agricultural Biotechnology: opportunities for international development (Persley GJ, ed.). Wallingford, Oxon, UK: CAB International. DE LANGHE E. 1976. Pourquoi l’amélioration génétique du plantain n’est-elle pas actuellement réalisable? Fruits 31:537-539. DHED’A D, DUMORTIER F, PANIS B, VUYLSTEKE D, DE LANGHE E. 1991. Plant regeneration in cell suspension cultures of the cooking banana cv. ‘Bluggoe’ (Musa spp., ABB group). Fruits 46:125-135. ESCALANT J, TEISSON C. 1989. Somatic embryogenesis and plants from immature zygotic embryos of species Musa acuminata and Musa balbisiana. Plant Cell Reports 7:665-668. GAWEL N, JARRET RL. 1991. Cytoplasmic genetic diversity in bananas and plantains. Euphytica 52:19-23. IITA. 1992. Crop Improvement Division, Plantain and Banana Improvement Program: 1991 Annual Report and 1992 Workplan. Ibadan, Nigeria: IITA. ISRAELI Y, REUVENI O, LAHAV E. 1991. Qualitative aspects of somaclonal variation in banana propagated by in vitro techniques. Scientia Hortic. 48:71-88. KRIKORIAN AD. 1987. Callus and cell culture, somatic embryogenesis, androgenesis and related techniques for Musa improvement. Pages 128-135 in Banana and Plantain Breeding Strategies (Persley GJ, De Langhe EA, eds). ACIAR Proceedings no.21. Canberra, Australia: ACIAR. NOVAK FJ, AFZA R, VAN DUREN M, PEREA-DALLOS M, CONGER BV, XIAOLANG T. 1989. Somatic embryogenesis and plant regeneration in suspension cultures of dessert (AA and AAA) and cooking (ABB) bananas (Musa spp.). Bio/Technology 7:154-159. PANIS B, WITHERS L, DE LANGHE E. 1990. Cryopreservation of Musa suspension cultures and subsequent regeneration of plants. Cryo-letters 11:337-350. PANIS B, DHED’A D, SWENNEN R. 1991. Freeze-preservation of embryonic Musa suspension cultures. Pages 183- 195 in Conservation of Plant Genes: DNA banking and in vitro biotechnology (Adams RP, Adams JE, eds). New York, USA: Academic Press. PERSLEY GJ, DE LANGHE EA (EDS). 1987. Banana and Plantain Breeding Strategies. Proceedings of an international workshop held at Cairns, Australia, 13-17 October 1986. ACIAR Proceedings no.21. Canberra, Australia: ACIAR. ROWE P. 1984. Breeding bananas and plantains. Plant Breeding Reviews. 2:135-155. SHEPHERD K. 1987. Banana breeding - past and present. Acta Horticulturae 196:37-43. SIMMONDS N W. 1952. Segregations in some diploid bananas. J. Genet. 51:458-469. SIMMONDS NW. 1987. Classiﬁcation and breeding of bananas. Pages 69-73 in Banana and Plantain Breeding Strategies (Persley GJ, De Langhe EA, eds). ACIAR Proceedings no.21. Canberra, Australia: ACIAR. STOVER RH. 1983. Effet du Cercospora noir sur les plantains en Amérique Centrale. Fruits 38:326-329. STOVER RH, DICKSON JD. 1970. Leaf spot of bananas caused by Mycosphaerella musicola: methods of measuring spotting prevalence and severity. Trop. Agr. (Trinidad) 47:289-302. STOVER RH, BUDDENHAGEN IW. 1986. Banana breeding: polyploidy, disease resistance and productivity. Fruits 41:175-191. SWENNEN R, VUYLSTEKE D. 1990. Aspects of plantain breeding at IITA. Pages 252-266 in Sigatoka Leaf Spot Diseases of Bananas. Proceedings of an international workshop held at San José, Costa Rica, March 28 - April 1 1989 (Fullerton RA, Stover RH, eds). Montpellier, France: INIBAP. VUYLSTEKE D. 1989. Shoot-tip culture for the propagation, conservation and exchange of Musa germplasm. Practical manual for handling crop germplasm in vitro no.2. Rome, Italy: IBPGR. 56 pp. VUYLSTEKE D, SWENNEN R, DE LANGHE E. 1990a. Tissue culture technology for the improvement of African plantains. Pages 316-337 in Sigatoka Leaf Spot Diseases of Bananas. Proceedings of an international workshop held at San José, Costa Rica, March 28 - April 1 1989 (Fullerton RA, Stover RJ, eds). Montpellier, France: INIBAP. VUYLSTEKE D, SCHOOFS J, SWENNEN R, ADEJARE G, AYODELE M, DE LANGHE E. 1990b. Shoot-tip culture and third- country quarantine to facilitate the introduction of new Musa germplasm into West Africa. FAO/IBPGR Plant Genetic Resources Newsletter 81/82: 5-11. VUYLSTEKE D, SWENNEN R, DE LANGHE E. 1991. Somaclonal variation in plantains (Musa spp., AAB group) derived from shoot-tip culture. Fruits 46:429-439. JV Escalant, C Teisson 177

Somatic Embryogenesis and Cell Suspensions in Musa

JV Escalant1, C Teisson2

Introduction Cultivated bananas and plantains generally originate from interspecific hybridization between two fertile diploid species, Musa acuminata (AA) and Musa balbisiana (BB) (Stover, Simmonds 1987). The most widespread edible genomic groups are: AAA (e.g., Cavendish), AAB (e.g., plantains), and ABB (e.g., Bluggoe). Over the past few years, diseases such as black Sigatoka (Mycosphaerella fijiensis), fusarial wilt (Fusarium oxysporum), and the banana bunchy top virus (BBTV) have resulted in increasing efforts to improve Musa disease resistance genetically. Because of the high sterility levels and polyploidy of banana and plantain, conventional breeding using traditional hybridization techniques remains difficult (Rowe 1984). Tissue-culture and molecular biology techniques have great potential for overcoming the problems yet unsolved by traditional improvement methods (Murfett, Clarke 1987). Such techniques rely on successful regeneration of plants from cell suspensions or protoplasts. The ﬁrst report on cell suspensions in Musa described the obtention of slow-growing cell suspensions from immature fruit-derived callus (Moham Ram, Steward 1964). Novak et al. (1989) reported recovery of plants from somatic embryos obtained in cell suspensions from rhizome tissue in diploid and triploid banana cultivars. More recently, cell suspensions and subsequent plant regeneration have been obtained applying somatic embryogenesis to meristematic ‘scalps’ of cooking banana cv Bluggoe (Sannasgala 1989; Dhed’a et al. 1991). In this paper, we report on the progress achieved on wild diploid species and triploid cultivars of Musa in the collaborative work between CATIE and CIRAD biotechnology laboratories.

Methods Depending on diploid or triploid material, different methods can be used to obtain cell suspensions and subsequent plant-regeneration techniques.

1CATIE/CIRAD-IRFA, PO Box 104, 7170 Turrialba, Costa Rica; 2CIRAD-BIOTROP, Laboratoire de Culture in vitro, BP 5035, 34032 Montpellier Cedex 1, France 178 Somatic Embryogenesis and Cell Suspensions in Musa

Cell suspension in diploid wild species Embryogenic calli were obtained using the method described by Escalant and Teisson (1989). When callus embryogenic characteristics were evident, showing first somatic embryos in its surface, calli were transferred into 25 mL of liquid medium. Cultures were kept in 125-mL Erlenmeyer flasks on a rotary shaker and refreshed every 2 weeks using the same medium. Once a high cell density was obtained (visual estimate), calli were sieved (500 µm) and cell suspensions (filtrated) maintained in the same medium, being renewed every week by decanting the old medium. Cell suspensions were initiated with somatic embryogenic callus at different culture periods. Inoculated calli released lots of cells into the liquid medium. After 3-5 weeks of culture, when cell density was adequate (visual estimate), callus was removed by sieving it (500 µm). At this stage, microscopic control of cell suspensions was done displaying many single embryogenic cells mixed with very small clusters (2-5 cells) and a few elongated cells. After 5 weeks a rapid cell growth was observed, reaching a weekly average increase of 10 µL mL-1. During the growing period, many cell divisions took place and, approximately 8-12 weeks later, somatic embryos appeared in large amounts. This growth was recorded measuring the total cell mass as packed cell volume using the PCV method (Reinert, Yeoman 1982, pp.12-14). After this, the most developed embryos were isolated (100 µm) and transferred to a germination medium, resulting in plant recovery frequencies from 20 to 36%. Separation of these embryos was carried out every month. After each filtration, it was possible to isolate more than 300 somatic embryos. This procedure started in mid-1991 and, since then, it has been maintained in the embryo-producing suspensions. With the second procedure, aliquots (0.5 mL) of filtered (100 µm sieved) cell suspensions, formed numerous somatic embryos after 3-4 weeks of inoculation in a hormone-free semisolid MI medium. Germination started after transferring embryos to the MG medium, with a 30% germination rate. Using this method, it was possible to have up to 70 somatic embryos from 1.5 mL of cell suspension. An advantage of this technique is that it facilitates the selection of the best-developed embryos, allowing immature ones to remain in the culture medium until reaching maturity. Also the fact that embryos are formed in a hormone-free medium will favor formation of more normal embryos, thus increasing germination percentage and obtention of complete plants. Histological sections of differentiated somatic embryos from both methods, showed a bipolar structure with the shoot and root apices connected to a storage parenchyma with vascularization, and entirely surrounded by an epidermis. After 1 month of in-vitro growth, plantlets from somatic embryos germination were transferred to a greenhouse for acclimatization.

Somatic embryogenesis and cell suspension in triploid Musa cultivars Somatic embryogenesis Explants used consisted of very young male ﬂowers of different cultivars: Musa AAA cv Grande Naine and 901; Musa AAB cv Currare French type; Musa ABB cv Pelipita. JV Escalant, C Teisson 179

After 3-4 months on an MS-derived semisolid medium, full calli and somatic embryos were observed: some of them yellow and compact and others more friable and white. Histological analysis showed that calli are originated from reactivation of perivascular cell divisions and that they are composed by meristematic cells associated with starch and proteic reserves. Transversal sections of somatic embryos revealed shoot and root apices connected to a storage parenchyma with trifid vascularization. Germination of somatic embryos occurred by development of the plumule and formation of a whole plant. After 1 month of in-vitro growth and 2 months in the greenhouse, plants were transferred to field conditions for further evaluation. Cell suspensions with triploid cultivars There are two ways of establishing cell suspensions from triploid cultures: (a) using male flower callus; and (b) using somatic embryos from male flowers. a) This technique is based on Professor Ma’s findings, and it consists in transfering callus with explant into a liquid medium. Even though inoculated calli released numerous cells in the liquid medium, during the initial steps of culture, most of them were meristematic and only a few showed some embryogenic characteristics. During the growth period, more embryogenic cells appeared and some somatic embryos were formed. However, these results are only preliminary, and continuing research is needed. b) In this technique, somatic embryos from male flowers are used as immature zygotic embryos for the diploid cell suspensions method. Grande Naine and Currare somatic embryos responded after approximately 3 weeks, forming small embryogenic callus. When this material became friable, it was transferred to a liquid medium releasing many embryogenic cells and small clusters. The first results from this research are considered very promising because of their similarity to those obtained from wild diploid species.

Conclusion Embryogenic diploid cell suspensions were obtained by culturing embryogenic callus. Calli from zygotic immature embryos and triploid somatic embryos were an excellent material for this purpose. The simplicity of this technique permits its application in plant transformation studies and its unquestionable future integration into banana and plantain improvement programs. Moreover, this method could be extended to other cultivars (2n, 3n, 4n) using embryos as raw material to obtain embryogenic callus and plant regenerations from cell suspensions.

References

DHED’A D, DUMORTIER F, PANIS B. 1991. Plant regeneration in cell suspension cultures of the cooking banana cv. “Bluggoe” (Musa spp., ABB group). Fruits 46:135. ESCALANT JV, TEISSON C. 1989. Somatic embryogenesis and plants from immature zygotic embryos of species Musa acuminata and Musa balbisiana. Plant Cell. Rep. 7:665-668. 180 Somatic Embryogenesis and Cell Suspensions in Musa

MOHAM RAM HY, STEWARD FC. 1948. The induction of growth in explanted tissue of banana fruit. Can. J. Bot. 42:1559-1579. MURFETT J, CLARKE A. 1987. Producing disease-resistant Musa cultivars by genetic engineering. Pages 87-94 in Banana and Plantain Breeding Strategies (Persley GJ, De Langhe E, eds). ACIAR Proceedings no.21. Canberra, Australia: ACIAR. NOVAK FJ, AFZA R, VAN DUREN M. 1989. Somatic embryogenesis and plant regeneration in suspension cultures of dessert (AA and AAA) and cooking (ABB) bananas (Musa spp.). BioTechnol. 7:147-158. REINERT J, YEOMAN MM. 1982. A laboratory manual: plant cell and tissue culture. Berlin, Germany: Springer- Verlag. ROWE P. 1984. Breeding bananas and plantains. Pages 135-155 in Plant Breeding Review (Janick J, ed.). Westport, CT, USA: AVI Publishing. SANNASGALA K. 1989. In vitro somatic embryogenesis in Musa. Dissertationes de Agricultura no.180. Ph.D. thesis. Heverlee, Belgium: KU Leuven. 172 pp. STOVER RH, SIMMONDS NW. 1987. Bananas. London, UK: Longman. H Leblanc, JV Escalant 181

Induced Parthenogenesis to Obtain Haploid Plants in Musa

H Leblanc, JV Escalant

Introduction This research began in CATIE in 1991, with the following main objectives: *To determine if pollen of Musa wild diploid species was able to germinate after being treated with different doses of gamma radiation. *To determine if, after such treatment, pollen was still able to stimulate embryo development without fertilization of the oosphere. *To determine if this technique helped to induce the formation of Musa spp. haploid plants.

Materials and Methods Pollen from M. acuminata ssp. burmannicoides and M. balbisiana type Tani was irradiated with different gamma radiation doses (5-100 krad) (60cobalt source), and then used to pollinate female flowers of M. acuminata ssp. burmannicoides. Pollen germination and evolution after pollinization were monitored using a fluo- rescent microscope with Z stainer and BV-2A filters. Observation of histological sections permitted the embryo evolution to be observed. Embryo rescue was carried out using tissue-culture techniques to obtain whole plants. The ploidy level of plants was determined using two methods: (a) evaluation of chromosome number with karyotypic analysis, and (b)evaluation of the ADN nucleus volume using flow-cytometry analysis.

Results and Discussion Pollen was able to germinate after irradiation with any of the doses used. Observation with a ﬂuorescent microscope revealed that the germinating tube grows to the ovule and makes contact with it whatever radiation dose is used. Fruit and seed development were obtained with all treatments. However, only treatments between 0 and 10 krad permitted embryo development. Moreover, endosperm-

CATIE/CIRAD-IRFA, PO Box 104, 7170 Turrialba, Costa Rica 182 Induced Parthenogenesis to Obtain Haploid Plants in Musa

vacant seeds were found in all the doses used, but at a higher percentage at doses in excess of 8 krad. The lack of endosperm is an important parameter because it can be related to fertilization absence, conﬁrming the presence of haploid embryos. Embryo rescue by in-vitro transfer permitted the culture of whole plants from seeds irradiated with treatments of 3, 5, 7, 8, and 10 krad. Karyotype analysis revealed that many of the plants evaluated were diploid, although two haploid plants were found among them (one with a 5-krad treatment and the other with 7-krad). It is also possible that there were more haploid plants not detected because they have the ability to double their chromosome number spontaneously. For this reason, it is important to follow up this research in order to ﬁnd a simple and accurate technique for detecting homocigous plants. W Parrott 183

Cell-Culture Techniques

W Parrott

Introduction The discovery that plant cells and tissues could be grown in vitro created a new discipline in the plant sciences. Cell-culture techniques quickly progressed beyond academic curiosities as their potential for plant propagation and plant improvement became evident. Cell-culture techniques have been extended to ever greater numbers of plant species, the techniques have become increasingly more reﬁned, and the biology of the systems better understood. Presently, cell- and tissue-culture technology for some species is much further advanced than for other species. Nevertheless, several universal patterns and similarities in procedures are now clearly evident across the most efficient cell culture systems. This information will be invaluable to optimize cell-culture techniques for those species still difﬁcult to handle in vitro.

Deﬁnition of Terms In its broadest sense, cell and tissue culture encompasses virtually any technique whereby aseptic plant organs, tissues, or cells are grown on a nutrient medium inside sealed containers. Thus, this broad sense has come to include techniques as diverse as meristem micropropagation, embryo, anther, tissue, and protoplast culture. In the strictest sense, and in the sense used for this review, cell culture refers exclusively to the culture of nonmeristematic cells, a distinctly different process from micropropagation, in which the original meristematic state of the explant tissue is maintained in vitro and used to proliferate numerous plantlets. Micropropagation remains an efﬁcient method for mass propagation of, or virus elimination from, desired genotypes. Micropropagation can be a source of novel variation, but thus far has not been readily amenable to gene transfer technology. In contrast to micropropagation, cell-culture techniques are not limited to explants from meristematic tissue. Virtually any part of a plant can serve as the explant. During the culture process, cells within the explant tissue divide to form a dedifferentiated tissue called a callus. From the perspective of plant improvement, cell-culture techniques are most powerful when the callus can be induced to redifferentiate into shoots (organogenesis) or embryos (somatic embryogenesis) from which whole plants can then

Department of Crop and Soil Sciences, University of Georgia, Athens, GA 30602-7272, USA 184 Cell-Culture Techniques

be recovered in a process called regeneration. Because a callus phase intervenes prior to regeneration, such systems are said to be indirect. Various selection, mutagenesis, or genetic transformation schemes may be superimposed on the system prior to indirect regeneration. However, at times it is possible for cells within the explant tissue to redifferentiate directly into shoots or embryos. Such direct regeneration systems are more challenging to use for selection or transformation, but are less subject to somaclonal variation than indirect systems.

Banana Cell Culture The overwhelming amount of in-vitro research on banana has centered on micropropagation techniques and will not be covered here. While organogenesis from nonmeristematic explants of bananas or plantains has not been reported, de novo regeneration of Musa species has been achieved via somatic embryogenesis (Escalant, Teisson 1988; Cronauer-Mitra, Krikorian 1988; Novak et al. 1989). Somatic embryogenesis (reviewed in Merkle et al. 1990) offers advantages over regeneration via organogenesis. First among these advantages is that complete propagules are formed. As somatic embryos have both shoot and root meristems, separate shoot-induction and root- induction steps are not required. Another advantage is that somatic embryogenesis tends to be less subject to somaclonal variation than organogenesis, presumably because somatic embryos are not very tolerant of somaclonal variation which can disrupt their ontogeny (Hanna et al. 1984; Ozias-Akins, Vasil 1988). This can be particularly advantageous during genetic transformation of a given genotype, in which case changes other than the engineered trait are undesirable. As technology continues to develop, somatic embryos will be amenable to large-scale production and artificial seed technology. Given the amenability of Musa species to somatic embryogenesis, optimization of efficient embryogenic protocols and their use in genetic transformation is a logical goal.

Factors Affecting Somatic Embryogenesis Somatic embryogenesis has several distinct stages, beginning with induction of the embryogenic state, followed by differentiation into an embryo, and subsequent maturation, germination, and conversion into plants. Induction of the embryogenic state usually requires the presence of an auxin. The best type and concentration of auxin for induction of somatic embryogenesis is determined empirically for each system. For Musa ornata, 2,4-D [2,4-dichlorophenoxyacetic acid] (Cronauer-Mitra, Krikorian 1988) has been effective, and Dicamba [3,6-dichloro-2- methoxybenzoic acid] has been effective for both bananas and plantains (Novak et al. 1989). Induction on 2,4,5-T [2,4,5-trichloro- phenoxyacetic acid] (Cronauer, Krikorian 1983) or induction on picloram [4-amino-3,5,6- trichloropicolinic acid] or Dicamba, followed by transfer to NAA [Ó-naphthaleneacetic W Parrott 185

acid] (Escalant, Teisson 1988), has permitted the recovery of somatic embryos of plantain. These results suggest that there is flexibility in the type of auxin used for induction of somatic embryogenesis in Musa species. A series of experiments critically to compare all the various growth regulators under controlled conditions should be useful for the optimization of the current regeneration protocols. Several plant tissues can serve as explants, with the caveat that stage of development or maturity of the tissue can be very important. The one tissue that is most frequently embryogenic is the immature zygotic embryo and, in several species, somatic embryogenesis occurs exclusively from immature zygotic embryos. However, zygotic embryo explants are of limited use in vegetatively propagated cultivars, as they are a different genotype from the parent cultivar. For seedless species such as cultivated bananas and plantains, the use of zygotic embryo explants is impossible. Meristematic tissue in which cells are actively undergoing mitosis, sometimes can be embryogenic. Such tissues include immature leaves of dicotyledons, leaf bases and internodes of monocotyledons, and apices of shoots, corms, or rhizomes. For those species in which an extensive callus phase precedes the formation of somatic embryos, virtually any tissue can serve as an explant. While immature zygotic embryos from seeded bananas (Cronauer-Mitra, Krikorian 1988) or from artiﬁcial hybrids (Escalant, Teisson 1988) can be induced to undergo somatic embryogenesis, the meristematic regions from shoot tips (Cronauer, Krikorian 1983) and rhizomes (Novak et al. 1989) are also embryogenic. The induction of somatic embryogenesis theoretically involves the induction of the same genetic pathways that lead to zygotic embryogenesis, and as such should be a universal phenomenon in all angiosperms. Nevertheless, individual genotypes within a species can differ widely in their ability to undergo somatic embryogenesis (Parrott et al. 1991). Hence such genotypic differences in embryogenic capacity probably reflect differences in the ability to activate the embryogenic pathway. When the molecular mechanisms of embryogenesis become understood, it should become possible to obtain somatic embryos from virtually any genotype. Fortunately, the cultivated bananas and plantains both appear to be genotypes with embryogenic capacity under the current protocols. The method by which auxins induce tissues to become embryogenic remains unknown, but it probably includes a series of steps. First, the presence of an auxin has been associated with DNA methylation, which putatively results in a decrease or termination of gene expression programs existing within the cell. In addition, auxins can isolate cells or clumps of cells from the remaining tissue, either by severing plasmodesmata and increasing friability of tissues, or through necrosis of surrounding tissues. Isolation may be reinforced by the formation of callose (Dubois et al. 1990; Dubois et al. 1991). The role of isolation has been reviewed by Williams and Maheswaran (1986) and by Smith and Krikorian (1989). The third step appears to be the imposition of polarity, possibly as a consequence of polar auxin transport (e.g., Brawley et al. 1984; Chée, Cantliffe 1989). Accordingly, the imposition of polarity by the application of an external electrical ﬁeld has increased the formation of somatic embryos in alfalfa (Dijak et al. 1986). 186 Cell-Culture Techniques

Once an embryogenic state has been induced, the continued presence of an auxin at a sufficiently high level in the medium can arrest the development of somatic embryos. Depending on the species, this arrest can occur at any time during ontogeny, ranging from the proembryo stage all the way to an early cotyledonary stage. A few species can even form mature somatic embryos in the presence of an auxin, but these have absent or poorly developed meristems (Halperin, Wetherell 1964; Parrott et al. 1988), which later makes it very difﬁcult to convert such embryos into plants. Hence, it is always preferable to remove the auxin from the culture medium as early as possible and, in some cases, add activated charcoal to the medium to assist in the auxin removal process (Zaghmout, Torello 1988; Ebert, Taylor 1990). When enough auxin is present in the medium to arrest embryogenesis at a proembryo to globular stage of development, the induced embryogenic state can be maintained for very long periods of time (Terzi, LoSchiavo 1990). This can give rise to cycles of repetitive embryogenesis, in which the original somatic embryos, instead of completing their ontogeny, give rise to other somatic embryos in a cyclical process. Such a system of repetitive or recurrent embryogenesis can effectively replace a callus phase during in- vitro selection or genetic transformation by Agrobacterium (McGranahan et al. 1988; McGranahan et al. 1990) or by microprojectile bombardment (Finer, McMullen 1991). The cycle of repetitive embryogenesis is broken by the removal of auxin from the medium, which allows somatic embryos to complete their development. Once somatic embryos have reached a cotyledonary stage of development, they require a maturation stage. The length of the maturation period is approximately as long as that required by zygotic embryos maturing in planta. The maturation process takes place in the absence of exogenous growth regulators, a phenomenon that may reﬂect the events that occur in planta, whereby cytokinin levels (Carman 1989) and auxin levels (Carnes, Wright 1988) in the ovule decrease once embryos are formed. As discussed previously, auxins inhibit the normal development of meristems, and cytokinins cause swollen hypocotyls in the resulting somatic embryos. Nevertheless, there is an apparent requirement for abscisic acid (ABA) for normal maturation of embryos. The continued presence of ABA beyond a few hours is probably not necessary, and a short pulse will initiate the synthesis of proteins associated with the late stages of embryogenesis and desiccation tolerance (Galau et el. 1990). While somatic embryos from most species probably have sufficient endogenous ABA that they do not require exogenous applications, some species exhibit improved conversion following exposure to ABA (Senaratna et al. 1989), and others, notably gymnosperms, have an absolute requirement for exogenous ABA to achieve normal maturation (Roberts et al. 1990). There are a few reports that maltose can be superior to sucrose (Button 1978; Strickland et al. 1987), but thus far, very little attention has been placed on the carbohydrate source used during maturation. However, recent reports (Finer, McMullen 1991), as well as current results from our laboratory, indicate that the use of 6-12% maltose promotes better embryo development than the traditional sucrose at any concentration. The use of maltose in a wider range of species merits further investigation. W Parrott 187

Somatic embryos of some species germinate readily and convert into plants, while those of other species germinate slowly and only sporadically convert into plants. Subjecting such mature somatic embryos to a desiccation step has been associated with enhanced germinability and conversion into plants in a range of monocotyledonous and dicotyledonous species (reviewed in Parrott et al. 1991). Embryos developing in planta undergo desiccation at the end of their maturation period, and desiccation has been associated with the synthesis of germination-speciﬁc proteins (Rosenberg, Rinne 1986, 1988). Protocols that have used high sucrose levels (e.g., Janick 1986) or paper bridges to suspend embryos above the medium (e.g., Cronauer-Mitra, Krikorian 1988) to achieve conversion may, in fact, have been achieving a degree of desiccation through osmotic stress. In addition, the use of ABA during the maturation process can enhance desiccation tolerance and plant conversion (Kumar et al. 1988; Senaratna et al. 1991) in some species. Germination of somatic embryos follows the return of embryos to a medium devoid of all growth regulators. Once seedlings have well developed roots and shoots, they are transferred to soil and hardened off. If a particular species is sensitive to daylength, attention should be provided to the length of the photoperiod to prevent the premature induction of ﬂowering.

Cell Culture, In-Vitro Selection, and Somaclonal Variation One of the characteristics frequently associated with plants regenerated from cell culture is that regenerated plants can differ by one or more traits from the plant that served as an explant donor. Such changes tend to occur in higher frequencies and represent more types of mutations than those resulting from conventional mutagenesis programs (Larkin et al. 1989). Depending on the research objectives, such variation may either be desirable or undesirable. Consequently, cell-culture protocols may be manipulated either to increase or decrease the amount of somaclonal variation that is recovered. In a case where genotypic ﬁdelity is desired, such as during the genetic transformation of an elite cultivar, somaclonal variation is most probably undesirable. As mentioned earlier, the use of somatic embryogenesis instead of organogenesis will limit the amount of somaclonal variation recovered. In addition, as the incidence of somaclonal variation increases with time in culture, somaclonal variation can be decreased by avoiding the use of long-term cultures and reducing the amount of time in culture to a minimum. Finally, callus tissue is notoriously susceptible to somaclonal variation. Hence it becomes especially important to minimize the length of any callus phase that may be present. If possible, repetitive embryogenesis should be used as a substitute for a callus phase. Alternatively, when maximizing somaclonal variation is the objective, such variation can be increased by increasing the amount of time in culture, especially the callus phase. The micropropagation of adventitious buds can also exhibit large amounts of variability (De Klerk 1990). Of particular interest to banana and plantain improvement, somaclonal 188 Cell-Culture Techniques

variation for disease resistance has been reported previously, in a range of species from poplars resistant to Septoria musiva (Ostrey, Skilling 1988) to tomato resistant to Fusarium (Evans 1989) and Pelargonium resistant to Xanthomonas (Dunbar, Stephens 1989). Perhaps the largest drawback to somaclonal variation is that it remains a random process (Evans 1989). When a specific type of somaclonal variant is desired, such as resistance to a pathotoxin, the process can be made less random through the use of in- vitro selection, a process which can select for dominant mutations when the mechanism of resistance at the cell level is the same as at the whole-plant level. The incorporation of pathotoxins into the culture medium has permitted the recovery of tobacco resistant to Pseudomonas and Alternaria (Thanutong et al. 1983), oilseed rape to Alternaria (MacDonald, Ingram 1986), and celery to Fusarium (Heath-Pagliuso et al. 1988). However, in at least the case of Leptosphaeria resistance in oilseed rape (Newsholme et al. 1989) and Verticillium resistance in alfalfa (MacDonald, Ingram 1986), the resistances proved to be either transient or nonheritable, underscoring a potential limitation to in-vitro selection. Given that apparently some callus formation precedes the formation of somatic embryos, Musa species should be amenable to in-vitro selection by incorporating culture ﬁltrates of Sigatoka or Panama disease in the culture medium. In addition, the ability to grow banana cells in suspension (Novak et al. 1989) prior to the onset of embryogenesis can facilitate the efﬁciency of in-vitro selection.

Protoplast Culture The ability to regenerate whole plants from isolated, single cells of plants represents one of the greatest achievements of academic importance for plant cell culture. However, the agricultural importance of protoplasts is much less clear at the present time. Techniques still in their infancy, such as asymmetrical hybridization or cytoplasm transfer, will continue to require the use of protoplasts. Other uses for protoplasts, such as interspecific hybridization and genetic transformation, have received considerable attention. Regeneration of plants from protoplast culture is extremely mutagenic and limited to very few species. It requires tremendous inputs of expertise and resources, and simpler alternatives are available. With the advent of microprojectile bombardment and repetitive embryogenesis technology, it is now clear that protoplasts are not necessary for plant transformation. While various reports of interspecific hybridization through protoplast fusion are available, simpler approaches were not tried prior to protoplast fusion. Approaches such as embryo rescue, in-vitro pollination, ovule culture, etc., should be exhausted before the use of protoplasts can be readily justified. These techniques lack the glamor of protoplast culture, but are much simpler and more economical. In the case of Musa species, the greatest return on investment can probably be best achieved by first devoting available resources to improving the efficiency of embryogenic and cell selection techniques. W Parrott 189

Conclusions The potential of cell-culture techniques to help solve a variety of agronomic problems has long been recognized, but results have usually fallen short of expectations. The use of cell-culture techniques offers the greatest potential in a genus such as Musa, which is characterized by the large-scale planting of elite, sterile genotypes. To the degree to which it is applicable, conventional breeding of bananas and plantains results in the loss of critical agronomic or horticultural characters, making it of limited value. The optimization of efficient regeneration techniques for this genus is necessary before cell-culture technology can be exploited. Fortunately, enough information is available to determine that efficient regeneration protocols are within reach for both bananas and plantains. Once such systems are in place, genetic transformation and in- vitro selection systems can be incorporated into the cell-culture protocols.

References

BRAWLEY SH, WETHERALL DF, ROBINSON KR. 1984. Electrical polarity in embryos of wild carrot precedes cotyledon differentiation. Proc. Natl Acad. Sci. (USA) 81:6064-6067. BUTTON J. 1978. The effects of some carbohydrates on the growth and organization of Citrus ovular callus. Z. Pﬂanzenphysiol. 88:61-68. CARMAN JG. 1989. The in ovulo environment and its relevance to cloning wheat via somatic embryogenesis. In Vitro Cell. Dev. Biol. 25:1155-1162. CARNES MG, WRIGHT MS. 1988. Endogenous hormone levels of immature corn kernels of A188, Missouri-17, and Dekalb XL-12. Plant Sci. 57:195-203. CHÉE RP, CANTLIFFE DJ. 1989. Embryo development from discrete cell aggregates in Ipomoea batatas (L.) Lam. in response to structural polarity. In Vitro Cell. Dev. Biol. 25:757-760. CRONAUER S, KRIKORIAN AD. 1983. Somatic embryos from cultured tissues of triploid plantains (Musa ‘ABB’). Plant Cell Rep. 2:289-291. CRONAUER-MITRA SS, KRIKORIAN AD. 1988. Plant regeneration via somatic embryogenesis in the seeded diploid banana Musa ornata Roxb. Plant Cell Rep. 7:23-25. DE KLERK GJ. 1990. How to measure somaclonal variation. Acta Bot. Neerl. 32:129-144. DIJAK M, SMITH DL, WILSON TJ, BROWN DCW. 1986. Stimulation of direct embryogenesis from mesophyll protoplasts of Medicago sativa. Plant Cell Rep. 5:468-470. DUBOIS T, GUEDIRA M, DUBOIS J, VASSEUR J. 1990. Direct somatic embryogenesis in roots of Cichorium: Is callose an early marker? Ann. Bot. 65:539-545. DUBOIS T, GUEDIRA M, VASSEUR J. 1991. Direct somatic embryogenesis in leaves of Cichorium. Protoplasma 162:120-127. DUNBAR KB, STEPHENS CT. 1989. An in vitro screen for detecting resistance in Pelargonium somaclones to bacterial blight of geranium. Plant Dis. 73:910-912. EBERT A, TAYLOR HF. 1990. Assessment of the changes of 2,4-dichlorophenoxyacetic acid concentrations in plant tissue culture media in the presence of activated charcoal. Plant Cell Tissue Organ Cult. 20:165-172. ESCALANT JV, TEISSON C. 1988. Embryogenèse somatique chez Musa sp. C. R. Acad. Sci. Paris 306:227-281. EVANS DA. 1989. Somaclonal variation—genetic basis and breeding applications. TIG 5:46-50. FINER JJ, MCMULLEN MD. 1991. Transformation of soybean via particle bombardment of embryogenic suspension culture tissue. In Vitro Cell. Dev. Biol. 27P:175-182. GALAU GA, JAKOBSEN KS, HUGHES DW. 1990. The controls of later dicot embryogenesis. Physiol. Plant. 81:280- 288. HALPERIN W, WETHERELL DF. 1964. Adventive embryony in tissue cultures of the wild carrot, Daucus carota. Am. J. Bot. 51:274-283. 190 Cell-Culture Techniques

HANNA WW, LU C, VASIL IK. 1984. Uniformity of plants regenerated from somatic embryos of Panicum maximum Jacq. (Guinea grass). Theor. Appl. Genet. 67:155-159. HEATH-PAGLIUSO S, PULLMAN J, RAPPAPORT L. 1988. Somaclonal variation in celery: screening for resistance to Fusarium oxysporum f.sp. apii. Theor. Appl. Genet. 75:446-451. JANICK J. 1986. Embryogenesis: the technology of obtaining useful products from the culture of asexual embryos. Pages 97-117 in Biotechnology of Plants and Microorganisms (Crocomo OJ, Sharp WR, et al., eds). Columbus, Ohio, USA: Ohio State University Press. KUMAR AS, GAMBORG OL, NABORS MW. 1988. Plant regeneration from cell suspension cultures of Vigna aconitifolia. Plant Cell Rep. 7:138-141. LARKIN PJ, BANKS PM, BHATI R, BRETTELL IS, DAVIES PA, RYAN SA, et al. 1989. From somatic variation to variant plants: mechanisms and applications. Genome 31:705-711. MACDONALD MV, INGRAM DS. 1986. Towards the selection in vitro for resistance to Alternaria brassicicola (Schw.) Wilts., in Brassica napus ssp. oleifera (Metzg.) Sinsk., winter oilseed rape. New Phytol. 104:621- 629. MCGRANAHAN GH, LESLIE CA, URATSU S, MARTIN LA, DANDEKAR AM. 1988. Agrobacterium-mediated transformation of walnut somatic embryos and regeneration of transgenic plants. Bio/Tech. 6:800-804. MCGRANAHAN GH, LESLIE CA, URATSU SL, DANDEKAR AM. 1990. Improved efﬁciency of the walnut somatic embryo gene transfer system. Plant Cell Rep. 8:512-516. MERKLE SA, PARROTT WA, WILLIAMS EG. 1990. Applications of somatic embryogenesis and embryo cloning. Pages 67-102 in Plant Tissue Culture: applications and limitations (Bhojwani SS, ed.). Amsterdam, the Netherlands: Elsevier. NEWSHOLME DM, MACDONALD MV, INGRAM DS. 1989. Studies of selection in vitro for novel resistance to phytotoxic products of Leptosphaeria maculans (Desm.) Ces. & De Not. in secondary embryogenic lines of Brassica napus ssp. oleifera (Metzg.) Sinsk., winter oilseed rape. New Phytol. 113:117-126. NOVAK FJ, AFZA R, VAN DUREN M, PEREA-DALLOS M, CONGER BV, XIAOLANG T. 1989. Somatic embryogenesis and plant regeneration in suspension cultures of dessert (AA and AAA) and cooking (ABB) bananas (Musa spp.). Bio/Tech. 7:154-159. OSTREY ME, SKILLING DD. 1988. Somatic variation in resistance of Populus to Septoria musiva. Plant Dis. 72:724-727. OZIAS-AKINS P, VASIL IK. 1988. In vitro regeneration and genetic manipulation of grasses. Physiol. Plant. 73:565-569. PARROTT WA, DRYDEN G, VOGT S, HILDEBRAND DF, COLLINS GB, WILLIAMS EG. 1988. Optimization of somatic embryogenesis and embryo germination in soybean. In Vitro Cell. Dev. Biol. 24:817-820. PARROTT WA, MERKLE SA, WILLIAMS EG. 1991. Somatic embryogenesis: potential for use in propagation and gene transfer systems. Pages 158-200 in Advanced Methods in Plant Breeding and Biotechnology (Murray DR, ed). Wallingford, Oxon, UK: CAB International. ROBERTS DR, FLINN BS, WEBB DT, WEBSTER FB, SUTTON BCS. 1990. Abscisic acid and indole-3-butyric acid regulation of maturation and accumulation of storage proteins in somatic embryos of interior spruce. Physiol. Plant. 78:355-360. ROSENBERG LA, RINNE RW. 1986. Moisture loss as a prerequisite for seedling growth in soybean seeds (Glycine max L. Merr.). J. Exp. Bot. 37:1663-1674. ROSENBERG LA, RINNE RW. 1988. Protein synthesis during natural and precocious soybean seed (Glycine max [L.] Merr.) maturation. Plant Phys. 87:474-478. SENARATNA T, MCKERSIE BD, BOWLEY SR. 1989. Desiccation tolerance of alfalfa (Medicago sativa L.) somatic embryos. Inﬂuence of abscisic acid, stress pretreatments and drying rates. Plant Sci. 65:253-259. SENARATNA T, KOTT L, BEVERSDORF WD, MCKERSIE BD. 1991. Desiccation of microspore derived embryos of oilseed rape (Brassica napus L.). Plant Cell Rep. 10:342-344. Smith DL, Krikorian AD. 1989. Release of somatic embryogenic potential from excised zygotic embryos of carrot and maintenance of proembryonic cultures in hormone-free medium. Am. J. Bot. 76:1832-1843. STRICKLAND SG, NICHOL JW, MCCALL CM, STUART DA. 1987. Effect of carbohydrate source on alfalfa somatic embryogenesis. Plant Sci. 48:113-121. TERZI M, LOSCHIAVO F. 1990. Somatic embryogenesis. Pages 54-66 in Plant Tissue Culture: applications and limitations (Bhojwani SS, ed). Amsterdam, the Netherlands: Elsevier. W Parrott 191

THANUTONG P, FURUSAWA I, YAMAMOTO M. 1983. Resistant tobacco plants from protoplast-derived calluses selected for their resistance to Pseudomonas and Alternaria toxins. Theor. Appl. Genet. 66:209-215. WILLIAMS EG, MAHESWARAN G. 1986. Somatic embryogenesis: factors inﬂuencing coordinated behaviour of cells as an embryogenic group. Ann. Bot. 57:443-462. ZAGHMOUT OMF, TORELLO WA. 1988. Enhanced regeneration from long-term callus cultures of red fescue by pretreatment with activated charcoal. HortSci. 23:615-616. 192 Somaclonal Variation in Musa sp.: theorical risks, risk management, future research prospects

Somaclonal Variation in Musa sp.: theoretical risks, risk management, future research prospects

FX Côte1, X Perrier2, C Teisson1

Introduction Over recent years in-vitro micropropagated Musaceae have been more and more widely used both for the international exchange of germplasm and the industrial production of healthy planting material. As a result of the increase in the latter use, the banana is probably the most intensely micropropagated crop. The appearance of somaclonal variations has introduced certain limits (Reuveni et al. 1986; Arias, Valverde 1987; Hwang, Ko 1987; Stover 1987; Daniells 1988; Vuylsteke et al. 1988; Krikorian 1989; Smith, Drew 1990; Israeli et al. 1991; Sandoval et al. 1991). However, considerable experience has been acquired, from which the following main conclusions may be drawn. 1. The main somaclonal variations that affect Musa are stable in the field after several crop cycles (Vuylsteke et al. 1988; Israeli et al. 1991). 2. The percentage of variants varies considerably, including in apparently identical micropropagation conditions. 3. The main characters concerned by the variations are those most frequently subject to natural mutation (Stover 1988; Vuylsteke et al. 1988). The two first above characteristics led us to approach the problem from the theoretical point of view: we looked for a mathematical model able to describe the variations in the percentage of variants over several successive cycles of in-vitro multiplication. The conclusions we drew from this study allowed us to deﬁne a number of pragmatic production strategies to limit the problem by combining risk management with early identiﬁcation of variants. The outlook for future research is described.

1CIRAD-BIOTROP, Laboratoire de Culture in vitro, BP 5035, 34032 Montpellier Cedex 1, France; 2CIRAD-IRFA, Laboratoire de Biométrie, BP 5035, 34032 Montpellier Cedex 1, France FX Côte, X Perrier, C Teisson 193

Theoretical Variations in the Percentage of Variants over Successive Multiplication Cycles Hypothesis and method of calculation The following hypotheses were put forward: • Only stable in-vitro variations are considered. Once a variation has occurred, all plants derived from this off-type plant are also variants. • The multiplication rate of variants is identical to that of true-to-type plants. • The probability of variation is constant during successive multiplication cycles and is independent of the rate of proliferation. •Variation can only affect a meristem during the initial stages of differentiation; as soon as this has taken place, risk of variation is zero in the plantlet concerned. For simpliﬁcation, we are not concerned with proliferation by direct multiplication of terminal meristems after splitting or fasciation. In these conditions, if p is the probability that a meristemic mass undergoes variation, t, the rate of multiplication and n the number of cultivation cycles, application of the recurrence relation shows that the percentage of variants will be equal to : 1 - [ 1 - ( p {t -1 } / t)] n x 100 called: function 1 (cf Appendix 1 for demonstration). Change in the percentage of variants is thus an exponential function which tends towards 100 when n tends to inﬁnity. This function depends on the rate of proliferation. Variation in the percentage of variants as a function of the number of cultivation cycles is shown in Figure 1 for the values p = 0.01, n = 1 to 20, and t = 2. The value of p was selected with reference to studies by Reuveni and Israeli (1990), Vuylsteke et al.

% variants 12

0 0 2 4 6 8 10 12 14 16 18 20 Subcultures Figure 1. Theoretical variation in the percentage of off-type plants over successive multiplication cycles. Hypothesis of calculation: y = 1 - [ 1 - ( p {t -1 } / t)]n (cf Appendix 1). n = the number of multiplication cycles, p = 0.01 is the probability that a meristemic mass undergoes variation, t = 2 is the proliferation rate. 194 Somaclonal Variation in Musa sp.: theorical risks, risk management, future research prospects

(1991), and Sandoval et al. (1991), who, in bananas of the Cavendish group and in plantains, observed rates of variants of a few percent after a small number of multiplication cycles. Taking into account the parameters selected, the curve of the function (1) may be considered as a linear one over the ﬁrst 20 multiplication cycles. The whole progeny of variants is off-type, and this is responsible for a large proportion of the total percentage of variants: after several cycles this percentage is far superior to the probability of appearance of variations.

Calculation of the variance of function 1-[1-(p{t-1}/t)]n Function (1) gives the expectation of the proportion of variants after n multiplication cycles. This value corresponds to the mean of the rates obtained with an inﬁnite number of lines derived from suckers established in vitro. However, very different percentages of off-types may be obtained, depending on how early variations occur. Let us consider the theoretical case where t = 2: if only one variation occurs during the ﬁrst multiplication cycle, 50% of the population will be off-type. With the same probability, if only one variation occurs during the second multiplication cycle, 25% of the descendants will be off-type, during the third multiplication cycle, 12.5%, and so on. The variance of function (1) cannot be analytically expressed; numerical examples of percentage of variants were determined using simulations. The simulations were carried out, for a given probability, p, by introducing variations in a random manner using a computer program. With t = 2, and p = 0.01, the population would include an average of 7.2% of variants after 15 multiplication cycles. Variance of this average determined after 500 simulations is shown in Figure 2. The distribution obtained is not normal; values of 3 to 55% of variants per line are observed with a mode of 5%. The peaks around 29 and 53%

% lines 50

0 171319 25 31 37 43 49 55 % variants

Figure 2. Estimation of the variance of function: 1 - [ 1 - (p {t - 1} / t)]n. Hypothesis of calculation: n = 15, p = 0.01, t = 2 (see caption of Fig.1). The estimation is determined after 500 simulations of function (1). Results are expressed in percentage of lines exhibiting a given percentage of variants. The appearance of variation is introduced in a random manner. The line is a population of plantlets derived from one sucker. Expectation of the mean percentage of variants = 7.2%. FX Côte, X Perrier, C Teisson 195

correspond to the rare cases when variations appeared in the first and second multiplication cycle respectively. The main information obtained from the above calculations is: - there is an increase in the percentage of variants during successive cultivation cycles; and - quite different and unpredictable rates of variants can be generated irrespective of the source of the material or the probability of the appearance of a variation.

Limiting the Risks of Variation linked to the use of Plantlets Production management The mathematical approach suggested two basic ways to reduce the risk of somaclonal variations: • Reduce the number of proliferation cycles to limit the multiplication of variant plantlets. •Use batches of plants from the greatest possible number of lines to reduce the possibility of rare ‘accidents’ linked to the early appearance of variants. These two strategies are already used, though doubtless not consistently enough, by industrial production units. They are more difﬁcult to apply in the case of a germplasm bank.

Removal of variants before planting The management plan described above should be completed by the earliest possible removal of the largest possible number of variant individuals before planting. In the case of a stable variation, in-vitro reintroduction and acclimatization of these off-types means a model will be available to facilitate early visual identification of morphological characters of variants. This method has been used to characterize dwarf variants in cv Grande Naine. During hardening, material obtained from off-types may be distinguished by morphological characters resembling those of dwarfism in adult plants, mainly reduced pseudostem height, a short petiole and leaf length (Côte et al. 1991). In the case of a large batch of plants hardened in natural conditions, identification is difficult. In these circumstances, identifying plants with morphological characters that fall between those of dwarf and true-to-type plants will be problematic. In order to facilitate identification in this particular case, we tested the use of gibberellin, as did Reuveni (1988) in vitro. Application of GA3 resulted in lengthening of the petiole in dwarf and true-to-type plants. Nevertheless their reaction to the growth regulator was different. In variant plants, only growth of the first leaf to emerge after treatment was stimulated, whereas, in true-to-type plants, stimulation was already observed in leaves developing at the time of treatment (Côte et al. 1991). Consequently there was a temporary accentuation in the morphological difference between the two types of banana. 196 Somaclonal Variation in Musa sp.: theorical risks, risk management, future research prospects

Conclusion and Outlook for Further Research In order to achieve better control of variation during micropropagation, more studies are needed on risk management and on phenonema that are responsible for the variations.

Risk management The theoretical model proposed is based on a number of hypotheses which certainly do not accurately represent the complexity of the phenomena. The model will undergo modification as our knowledge of somaclonal variation increases. However, in the meantime we hope it will suggest new fields of investigation. Our lack of knowledge about the biological signiﬁcance of risk p is one of its main limitations, but neverthless it is a model that may have broader applications. Compiling percentages of variation observed in large-scale trials or in industrial production would enable us to determine the value p more precisely, and there is a need for information from the scientists involved in these topics. Compilation should be carried out using batches of plants produced with identical micropropagation techniques. The value of p calculated in this way would enable us to determine the maximum number of plantlets that should be produced from one sucker for a given risk of variation. Already the conclusion to be drawn from the range of percentages of variants using the same micropropagation technique demonstrated with this model is that experiments to identify the origin of variations must be carried out on very large batches of plants.

Budding and its effect on variation The exact way of budding and proliferation of Musaceae may be of primary importance for the occurrence of somaclonal variation, as is the case in other species (Swartz 1991). In vitro, the great complexity rapidly attained in the proliferating shoots does not facilitate precise description. The problem nevertheless deserves attention, particularly the possible difference between neoformed meristem and simple multiplication of the terminal meristem, either after physical mutilation or due to a phenomenon resembling fasciation.

Use of variation markers The use of variation markers has several advantages: it can facilitate production management, the experimental approach by limiting the need for ﬁeld trials, and further knowledge of the origin of the variations. The dwarf variation in bananas of the Cavendish group could be used as a model in the search for early markers. This variation is stable and has similarities with natural differentiation in cultivars in this group. It is also the most common variation in the most frequently used bananas for in-vitro multiplication. The study of variation markers in other cultivars, and particularly those used in germplasm banks, is probably much more difﬁcult, given the diversity of genotype, the range of possible variations, and the limited number of individuals of a given genotype. FX Côte, X Perrier, C Teisson 197

Production management. An extremely limited number of culture cycles may be the simplest way to reduce the number of variants. This necessitates the frequent reintroduction of plants in vitro, which in turn increases the cost of plantlet production. The use of early markers may provide a way round this problem. Regular screening would mean that only plantlets identified as true-to-type are used for a limited number of multiplication cycles. The number of variants could be reduced to a very low value by frequent screening. This method presupposes a preliminary experimentation to ensure that risks of variation are identical for a population obtained from a true-to-type plantlet or from a sucker taken from the ﬁeld.

Types of early markers Morphological markers. Identiﬁcation of variants on a morphological basis at the test-tube level would be very useful. However, Israeli et al. (1991) have observed that this method is possible only for a few types of variants. Biochemical markers. Modifications in the endogenous hormone balance are frequently involved in somaclonal variation. The study of gibberellin metabolism is particularly interesting in the case of dwarﬁsm. The use of protein markers can also be envisaged, and has already been tested by Reuveni et al. (1986). Molecular markers. This should be one of the preferred methods for studying the origin of the variations. There are several research programs under way using molecular biology tools to classify closely related cultivars. Considering the similarity between natural differentiation and the main somaclonal variations, it is tempting to envisage characterizing variants using the same methods. A project of this type is already under consideration involving international collaboration between different research teams. Better knowledge of the origin of somaclonal variations in the banana has implications that go far beyond the already vast question of micropropagation by budding for mass multiplication and the conservation of germplasm. Efforts are currently under way to develop somatic embryogenesis as a back-up for genetic improvement programs. It is likely that similar problems of variation will occur in bananas produced with these techniques.

Appendix 1

Theoretical variations in the percentage of variants over successive multiplication cycles: methods of calculation (see hypothesis of calculation in the text) If p is the probability that a meristem undergoes variation during the early differentiation stage, t, the rate of multiplication, mn, the number of true-to-type plantlets in a population after n cycles of multiplication: 198 Somaclonal Variation in Musa sp.: theorical risks, risk management, future research prospects

• After n + 1 cycles of multiplication, t x mn plantlets are issued from the mn plantlets among which (t - 1)mn are neoformed plantlets and mn were present at the previous cycle. • The probability for a neoformed plantlet to be a true-to-type one is (1 - p); thus the total number of true-to-type plantlets after n + 1 cycle is

mn+1 = mn + (1 - p) (t - 1)mn mn+1 = mn [1 + (1-p) (t - 1)] or

mn+1 = mn [(t - p(t - 1)].

2 Thus, as mo = 1; m1 = [t - p(t - 1)]; m2 = [t - p(t - 1)] ... n and mn = [t - p(t - 1)] . • After n cycles of cultivation, tn plantlets are produced and the number of off-type plantlets is t n - [t - p(t - 1)]n. The proportion of variants is t n - [t - p(t - 1)]n tn and the percentage of variants is 1 - [1 - p(t-1)]n x 100. t

References

ARIAS O, VALVERDE M. 1987. Producción y variación de plantas de banano variedad Grande Naine producida por cultivo de tejidos. ASBANA 28:6-11. CÔTE FX, FOLLIOT M, KERBELLEC F, TEISSON C. 1991. Morphological characterization of banana vitroplantlets (cv Grande Naine) propagated from dwarf off-type and true-to-type plants during hardening. Effects of gibberellin treatment. ACORBAT Symposium 1991, Tabasco, Mexico (in press). DANIELLS JW. 1988. Comparison of growth and yield of bananas derived from tissue culture and conventional planting material. Banana Newsletter 11. HWANG SC, KO WH. 1987. Somaclonal variation in vitro of banana and its application for screening for resistance to fusarium wilt Pages 151-156 in Banana and Plantain Breeding Strategies (Persley GJ, de Langhe EA, eds). ACIAR Proceedings no.21. Canberra, Australia: ACIAR. ISRAELI Y, REUVENI O, LAHAV E. 1991. Qualitative aspects of somaclonal variations in banana propagated by in vitro techniques. Scientia Hortic. 48:71-78. KRIKORIAN AD. 1989. In vitro culture of bananas and plantains: background, update and call for information. Tropical Agriculture (Trinidad) 66:194-200. REUVENI O. 1988. Methods of detecting somaclonal variants in “Williams” bananas. Pages 108-113 in The Identiﬁcation of Genetic Diversity in the genus Musa, (Jarret RL, ed.). Montpellier, France: INIBAP. REUVENI O, ISRAELI Y. 1990. Measures to reduce somaclonal variation in in vitro propagated bananas. Acta Horticulturae 175:307-313. REUVENI O, ISRAELI Y, ESHDAT Y, DEGANI H. 1986. Genetic variability of banana plants multiplied via in vitro techniques. Final report submitted to IBPGR PR3/11. Israel: The Volcani Center. SANDOVAL JA, TAPIA SACM, VILLALOBOS AO. 1991. Observaciones sobre la variedad encontrada en plantas micropropagadas de Musa cv “Falso corno” AAB. Fruits 46:533-539. SMITH MK, DREW RA. 1990. Growth and yield characteristics of dwarf off-types recovered from tissue-cultured bananas Aust. J. Exp. Botany 30:575-578. FX Côte, X Perrier, C Teisson 199

STOVER RH. 1987. Somaclonal variation in Grande Naine and Saba bananas in the nursery and field. In Bananas and Plantain Breeding Strategies (Persley GJ, De Langhe EA, eds). ACIAR Proceedings no.21. Canberra, Australia: ACIAR. STOVER RH. 1988. Variation and cultivar nomenclature in Musa, AAA, group, Cavendish subgroup. Fruits 43:353-357. SWARTZ HH. 1990. Post culture behavior: genetic and epigenetic effects and related problems. Pages 95-121 in Micropropagation: Technology and Application (Debergh PC, Zimmerman RH, eds). Dordrecht, the Netherlands: Kluwer Academic Publishers. VUYLSTEKE D, SWENNEN R, WILSON GF, DE LANGHE E. 1988. Phenotypic variation among in vitro propagated plantain (Musa sp. cultivar “AAB”). Scientia Hortic. 36:79-88. VUYLSTEKE D, SWENNEN R, DE LANGHE E. 1991. Somaclonal variation in plantains (Musa spp., AAB group) derived from shoot-tip culture. Fruits 46:429-439. 200 Early Detection of Somaclonal Variation

Early Detection of Somaclonal Variation

LA Withers

In-Vitro Technology in Agriculture In-vitro (tissue culture) techniques have revolutionized many aspects of agriculture over recent years. Mass propagation in vitro is routine for large numbers of agricultural, horticultural, and forest species. The impact of in-vitro propagation is probably greatest for clonally propagated species but it is not limited to them (Beversdorf 1990). Propagation can be linked, with great benefit, to disease eradication and disease indexing, sometimes carried out in vitro, with the effect of increasing the scope and safety of germplasm exchange (IBPGR 1988). Moreover, these processes can be carried out in locations distant and climatically different from the natural growing region of the species in question, thereby remote from sources of infection with pathogens and risks of weather damage. Mass propagation can make elite material available to growers at a much earlier stage, thus enabling the efforts of the plant breeder to reach the farmer more rapidly. Biotechnological methods based on in-vitro culture have already made an impact on crop improvement and will become more important in the future as transformation techniques become reﬁned and more widely applicable (Tomes 1990). In the context of research, mass propagation can make uniform genotypes available to scientists, thereby eliminating one variable in experimental design. Genetic resources conservation is a crucial requirement for crop improvement and production. On the one hand, biotechnology and in-vitro culture amplify some of the problems of genetic conservation, e.g., by increasing the risks of narrowing of the gene pool through facilitating monoculture. However, on the other hand, they present new and improved ways of studying, understanding, and utilizing genetic diversity, and of conserving genetic resources. In-vitro conservation was a matter of speculation and curiosity little more than 15 years ago. Yet it now occupies a firm place in the portfolio of techniques used to safeguard genetic resources (Henshaw 1975; Withers 1980, 1989). Slow-growth storage for short- to medium-term conservation is now routine for many crops including banana and plantain (Dodds, 1991; Schoofs, 1990; Withers, 1987; 1990). Cryopreservation

International Board for Plant Genetic Resources (IBPGR), Via delle Sette Chiese 142, 00145 Rome, Italy LA Withers 201

(storage in liquid nitrogen) is virtually routine for cell-suspension cultures such as are used in physiological experiments, secondary-product synthesis, and some mass- propagation systems (Withers 1991a). Techniques are less well developed for organized cultures, although recent years have seen dramatic progress and the range of successfully cryopreserved species is now extensive. Particularly promising are the new techniques of encapsulation-dehydration and of vitriﬁcation, which appear to be able to preserve structural integrity in complex explants such as shoot-tips and embryos (Dereuddre et al. 1990; Towill 1990; Withers 1992). However, in propagation, in breeding and in conservation, there is one particular spectre that haunts the stage, namely somaclonal variation. The following sections explore the nature and potential impact of this phenomenon in general and in the case of Musa in particular.

Somaclonal Variation and Musa In-vitro culture practitioners have been aware for many years of the fact that what comes out of a culture procedure may well not be the same as that which goes into it. The culture process can lead to both genotypic and phenotypic changes (Larkin, Scowcroft 1981; Lee, Phillips 1988; Phillips et al. 1990; Scowcroft 1984). Cultures can exhibit changes in ploidy, chromosome number and structure, individual DNA base pairs, morphogenic potential, secondary product yields, disease resistance, and isozyme complement. Change is not an “unnatural” characteristic (Hohn, Dennis 1985; Walbot, Cullis 1985); indeed, somatic mutations have played an important part in the improvement of crops including Musa (Stover, Simmonds 1987). However, in vitro, it appears that the rate of change can be significantly accelerated. The concept of the absolute clone in vitro appears to be wrong and, if used incautiously as the basis for procedures with impact upon the genotype and phenotype of the in-vitro progeny, carries serious consequences. It is safer to consider the “clone” to be a tight population, the make-up of which can vary with time in terms of the range of genotypes present and the frequency of the predominant genotype. In the history of in-vitro linked variation, scientists moved from a position of ignorance of its existence, to awareness and fear of its consequences, but with little open discussion of the phenomenon. This was until the potential beneﬁts in crop improvement became apparent, coincident with the coining of the appealing and judgmentally neutral name “somaclonal variation” (Larkin, Scowcroft 1981; Scowcroft 1984). A period of excitement (and much investment) was followed by one of disappointment at the relatively few useful agronomic variants generated in vitro, and on to a period of realism, the situation today. Somaclonal variation exists and is one of the factors to key into the equation when planning and executing in-vitro procedures. Why is there special concern over somaclonal variation in the case of Musa? The answer to this question is related to the valuable niche that in-vitro technology occupies 202 Early Detection of Somaclonal Variation

in banana and plantain production, improvement, and conservation. A general appreciation of the importance of in-vitro technology in these aspects of agriculture is given above. Some specifics for Musa can be added to reinforce the argument. There is a huge demand for planting material in regions of the world where banana and plantain are important staple and export crops (Israeli et al. 1991; Vuylsteke, Swennen 1990). In-vitro propagation is the obvious and proven method to meet needs. Spread of disease in planting material is a serious problem that can be tackled by the exchange of cultures rather than in-vivo plant parts (Frison, Putter 1989; Schoofs 1990). Breeding for badly needed disease resistance in Musa could be accomplished by biotechnological approaches (Hwang 1990; Krikorian, Cronauer 1984; Novak et al. 1987; Stover, Buddenhagen 1986). The conservation of Musa genetic resources is problematic. Seeds are produced only by wild species. Sterile genotypes are, conventionally, conserved in field gene banks where costs, risks of disease, weather damage, and vandalism all pose serious threats to security (Withers, Engels 1990). In-vitro conservation, mainly by slow growth but with encouraging progress in cryopreservation, is providing a welcome alternative to the field gene bank. The most widely used propagation system of meristem proliferation, also used for slow-growth storage, is adventitious/semiadventitious (Banerjee et al. 1986). Such a system is considered risky from the point of view of somaclonal variation, although, in Musa, variants have been found even in directly regenerating, nonadventitious shoots (Vuylsteke, Swennen 1990). By its very nature, slow growth imposes stresses that could be selective. Successful cryopreservation work to date has involved the use of the apparently even more risky culture system of the embryogenic cell suspension (Panis et al. 1990). Promising as this technique is, cell and cell aggregate survival are well below 100% and, in principle, subject to risks of selection. Recovery appears to favor unorganized cell aggregates rather than organized, early-stage embryos. In-vitro improvement strategies involve similar systems or even potentially more unstable protoplasts. Whilst no data are yet available to suggest the existence of stability problems attributable to storage or genetic manipulation per se, these procedures themselves depend upon the development of satisfactory propagation systems. Evidence to date of somaclonal variation in Musa, gathered mainly, but not exclusively, from the standard mass-propagation system of in-vitro meristem proliferation, bears out the anticipated risks. Many studies have documented cases in Cavendish bananas (Musa spp., AAA group) with a smaller number describing variation in plantains (Musa spp., AAB group) (Arias, Valverde 1987; Hwang 1986; Israeli et al. 1991; Pool, Irizarry 1987; Ramacharan et al. 1987; Reuveni 1990; Reuveni et al. 1986; Smith 1988; Stover 1987; Vuylsteke, Swennen 1990; Vuylsteke et al. 1988). Qualitative and quantitative details were given in other presentations at this workshop. For the purposes of the present topic, some observations on the nature of somaclonal variation in Musa, taken largely from the work of Vuylsteke and colleagues (Vuylsteke, Swennen 1990; Vuylsteke et al. 1988) can help us characterize the phenomenon in Musa and put it into perspective. LA Withers 203

The variants generated in vitro are limited in number and most are not unique to in- vitro propagated material; they have been observed in the field but at lower frequencies. The type and frequency of variation is genotype-dependent. A spectrum of variants can arise from one explant and, whilst the phenomenon is clearly related to culture and probably to stress factors present in culture, for most variants it does not generally appear to be linked in magnitude to length of time in culture. (It must, however, be noted that the apparent linkage or lack of linkage between time in culture and variation could depend upon the multiplication regime and the contribution that one subculturing cycle makes to the subsequent one.) There is evidence suggesting the variation may be genetic in nature. However, reversion is observed in culture, thus showing that the unstable nature of some genotypes is conserved in vitro. This provides the chance of recovering the original genotype, but the problem of control remains. A unifying characteristic of the different variants is that they are generally manifested as changes in the plant’s habit, foliage, or inflorescence structure. Therefore, they are usually detectable only at the nursery, field, or even the fruit-production stage. Earlier visual detection in the laboratory or at the hardening stage before transfer to the nursery is possible only in a very limited number of cases such as extreme dwarfism (Israeli et al. 1991). Thus, for most variants, detection is on an extended time-scale, being months to over a year removed from the culture processes apparently generating the variation. The consequences of a significant time-lag in detecting somaclonal variation are several (Krikorian 1989). Material that reaches growers may not perform as expected. In extreme cases this could involve severe economic losses through poor yields, compounded by wasted investment of time, field space and other resources in cultivation of the propagated plants. In a collection of germplasm maintained in vitro, somaclonal variation could result in the loss of important genotypes. The damage to the reputation of in-vitro techniques, already felt in the case of some other crops where stability problems have been encountered, could be serious, resulting in a reaction against their development and implementation out of proportion to the real problem, and delaying or preventing a solution to the problem. From the conservation point of view in particular, the genotypic component in susceptibility to variation is particularly consequential. Genetic conservation should be applied to and driven by the needs of the gene pool, not driven by the available technologies. We should not just be conserving the “easy” genotypes. Instability in vitro is part of the biology of a genotype that must be handled in the best ways currently at our disposal. In any case, net risks resulting from several factors are experienced in the gene bank. Even if the genotype and associated risk factor is fixed, the way in which it is handled is not. That is under our control, to improve as our knowledge increases. Thus, if today that means risking a certain level of instability in vitro to counterbalance potentially greater risks of loss of entire accessions in the field, then there is no question that the in-vitro approach should be part of the conservation strategy. This is the essence of the integrated approach to genetic conservation, based on complementarity of different techniques that is now favored by IBPGR and others (Engels 1991; van Sloten 1990; Withers 1991b). 204 Early Detection of Somaclonal Variation

The transition of scientists through the spectrum of reactions to somaclonal variation described earlier, from ignorance and fear through to realism, needs to be matched by a strategy for confronting somaclonal variation as one of the realities of in- vitro culture of Musa. This could take the form of acknowledging the existence of somaclonal variation, learning how to recognize it, quantifying it, and then learning how to control it. Control is the key to handling any potentially beneficial but also potentially detrimental event. In the case of Musa, we are clearly at the second stage of this process in the sense of recognizing variants in the nursery and the ﬁeld, with some small inroads being made into the third and fourth stages of quantification and control. What is needed for signiﬁcant progress, however, is a way of overcoming the time-lag between generation and detection of variation to permit a real understanding of the linkages between variation and inducing factors.

Early-Stage Markers To bring the time scales of generation and detection more closely into line, there is a need for a system of early-stage markers that could be applied in vitro to identify variation as soon as it occurs. This is perhaps ambitious but should be viewed as a realistic expectation, particularly in view of the small number of important variants in Musa set against progress in the areas of biology that could provide the Potential early-stage marker systems foundation for such markers. One could speculate about the chromosome counting types of marker that could be devel- chromosome structure (karyotyping) oped (see box). This speculation is not chromosome banding based on extensive pilot work, nor is it without foundation. There are DNA measurements indications that at least some of the RFLP variants have a genetic basis, thus RAPD indicating the relevance of tests based on gross observations of chromosomes isozyme analysis and more detailed analysis of DNA by protein analysis molecular techniques. There are chlorophyll content speciﬁc indications that in Musa, as in other plants, some variants are the secondary-product analysis result of gross ploidy changes or endogenous levels of growth regulators aneuploidy (Hwang, Ko 1987; Israeli response to exogenous growth regulators et al. 1991; Lee, Phillips 1988; Phillips et al. 1990). Some variants behave in in-vitro growth rate a way that suggests the action of respiration rate transposable elements. Research has in-vitro morphology already been carried out on the LA Withers 205

application of isozyme analysis and RFLP (restriction fragment length polymorphism) to the identification of Musa varieties and somaclonal variants (Espino, Pimentel 1990; Jarret 1990). Very recently, work has started on the application of RAPD (random amplified polymorphic DNA) to varietal identification and the detection of variants. Preliminary results are very promising (pers. com. Ford-Lloyd and Newbury, Birmingham University, 1991, 1992). Slower field growth in some variants such as those with a drooping leaf habit might be reflected in the growth rate in vitro, expressed either as rate of accumulation of dry weight or rate of leaf primordium formation. Growth-rate abnormalities may be linked to changes in levels of endogenous growth regulators or a changed response to exogenously applied growth regulators such as gibberellins, as demonstrated by Reuveni (1990). Some morphological variants might simply have a different morphology in vitro. For example as already observed, increase in ploidy is marked by an increase in leaf thickness, the size of epidermal cells, and stomatal density (see Reuveni 1990). Protein analysis has been attempted in combination with isozyme analysis, but with little success (Reuveni 1990). However, there is scope for more extensive testing of this approach. Flavonoid analysis has been used in taxonomic studies (Horry, Jay 1990). Assays for these and other secondary products have potential as the basis for screening, particularly in view of the occurrence of variants with enhanced pigmentation (Israeli et al. 1991). Similarly, variegation involving loss of chlorophyll is a phenomenon with a quantifiable symptom. The ideas included in the box are not a definitive list; they are designed to stimulate the minds of scientists working with Musa in the laboratory and the field who are familiar with somaclonal variants, and of scientists whose work in anatomy, cytology, physiology, biochemistry, or molecular biology could be lent to the development of early markers. It is hoped that, as a result of this workshop, further ideas will be generated and potential tests discussed. What is important in designing such tests is that they should provide strong correlations, be rapid, be relatively inexpensive, and be reliable. Where it proves necessary to use more than one test, combinations should be limited in complexity so as to remain practical. An invaluable application of the early-stage markers would be in helping us design better experiments to try to identify and understand the linkages between somaclonal variation and factors in the culture procedure that could be manipulated to minimize variation. If the development and application of early-stage markers throws light onto the nature of somaclonal variants, that would be a bonus. However, it is neither an objective nor a requirement.

Developing a Strategy The development of the early-stage markers themselves is only part of the story. They need to be fully tested for their ﬁdelity and a strategy developed for incorporating them into propagation and conservation protocols, to provide the right tests at the right time to avoid the worst consequences of somaclonal variation. Components of the strategy will be 206 Early Detection of Somaclonal Variation

the nature of the tests themselves, their number, the procedure for sampling cultures, the timing and frequency of tests, streamlining and whenever possible mechanizing tests, and the incorporation of test results into procedural modiﬁcations. The scale of any testing operation will be crucial to its effective use. Testing too many cultures too frequently will deter progress because of costs and workloads. Testing too infrequently and on too few cultures will fail to pick up certain variants. The development of early-stage markers will require concerted efforts by many players and a considerable investment of effort and resources. However, the beneﬁts would add a dimension of security to in-vitro propagation and conservation of Musa, the lack of which at present is probably one of the major impediments to the biotechnological exploitation of the genus.

References

ARIAS O, VALVERDE M. 1987. Produccion y variacion somaclonal de plantas de banano variedad Grande Naine producidad por cultivo de tejidos. Rev. Asoc. Bana. Nac. (ASBANA) Costa Rica 28:6-11. BANERJEE N, VUYLSTEKE D, DE LANGHE E. 1986. Meristem tip culture of Musa: histomorphological studies of shoot bud proliferation. Pages 139-147 in Plant Tissue Culture and its Agricultural Applications (Withers LA, Alderson PG, eds). London, UK: Butterworth. BEVERSDORF WD. 1990. Micropropagation in crop species. Pages 3-12 in Progress in Plant Cellular and Molecular Biology (Nijkamp HJJ, van der Plas LHW, van Aartrijk J, eds). Dordrecht, the Netherlands: Kluwer. DEREUDDRE J, SCOTTEZ C, ARNAUD Y, DURON M. 1990. Résistance d’apex caulinaires de vitroplants de poirier (Pyrus communis L.ev. Beurre Hardy), enrobés dans l’alginate, à une deshydration puis à une congelation dans l’azote liquide: effect d’un endurcissement préalable au froid. Comptes Rendus de l’Academie des Sciences 310 Série III:371-323. DODDS JH. 1991. In Vitro Methods for Conservation of Plant Genetic Resources. London, UK: Chapman and Hall. ENGELS JMM. 1991. A holistic approach to germplasm conservation. IBPGR Newsletter for Asia and the Pacific 7:1-2. ESPINO RRC, PIMENTEL RB. 1990. Molecular methods for detecting genetic diversity in Musa. Pages 36-40 in The Identification of Genetic Diversity in the Genus Musa (Jarret RL, ed.). Montpellier, France: INIBAP. FRISON EA, PUTTER CAJ. 1989. FAO/IBPGR Technical Guidelines for the Safe Movement of Musa Germplasm. Rome, Italy: FAO/IBPGR/INIBAP. HENSHAW GG. 1975. Technical aspects of tissue culture storage for genetic conservation. Pages 349-358 in Crop Genetic Resources for Today and Tomorrow (Frankel OH, Hawkes JG, eds). Cambridge, UK: Cambridge University Press. HOHN B, DENNIS ES. 1985. Genetic Flux in Plants. Vienna, Austria: Springer-Verlag. HORRY JP, JAY M. 1990. Molecular methods for detecting genetic diversity in Musa. Pages 41-55 in The Identification of Genetic Diversity in the Genus Musa (Jarret RL, ed.). Montpellier, France: INIBAP. HWANG SC. 1986. Variation in banana plants propagated through tissue culture. J. Chin. Soc. Hortic. Sci. 32:117- 125 (in Chinese, with English summary). HWANG SC. 1990. Somaclonal resistance in Cavendish banana to fusarium wilt. Pages 121-125 in Fusarium Wilt of Banana (Ploetz RC, ed.). St Paul, MN, USA: APS Press/American Phytopathological Society. HWANG SC, KO WH. 1987. Somaclonal variation of bananas and screening for resistance to fusarium wilt. Pages 151-156 in Banana and Plantain Breeding Strategies (Persley GJ, De Langhe EA, eds). ACIAR Proceedings no.21. Canberra, Australia: ACIAR. IBPGR. 1988. Conservation and Movement of Vegetatively Propagated Germplasm: in vitro culture and disease aspects. Rome, Italy: IBPGR. ISRAELI Y, REUVENI O, LAHAV E. 1991. Qualitative aspects of somaclonal variations in banana propagated by in vitro techniques. B.V. Sci. Hort. 49:71-88. LA Withers 207

JARRET RL. 1990. Molecular methods for detecting genetic diversity in Musa. Pages 56-66 in The Identification of Genetic Diversity in the Genus Musa (Jarret RL, ed.). Montpellier, France: INIBAP. KRIKORIAN AD. 1989. In vitro culture of bananas and plantains: background, update and call for information, Trop. Agric. (Trinidad) 66(3). KRIKORIAN AD, CRONAUER SS. 1984. Aseptic culture techniques for banana and plantain improvement. Econ. Bot. 38:322-331. LARKIN PJ, SCOWCROFT WR. 1981. Somaclonal variation— a novel source of variability from cell cultures for plant improvement. Theoretical and Applied Genetics 60:197-214. LEE M, PHILLIPS RL. 1988. The chromosomal basis of somaclonal variation. Annual Review of Plant Physiology and Plant Molecular Biology 39:413-437. NOVAK FJ, DONINI B, HERMELIN T, MICKE A. 1987. Potential for banana and plantain improvement through in vitro mutation breeding. Pages 67-70 in Proceedings of the 7th ACORBAT Meeting, San José, Costa Rica, 23-27 Sep 1985 (Galindo JJ, Jaramillo R, eds). Technical Bulletin no.121. Turrialba, Costa Rica: CATIE. PANIS BJ, WITHERS LA, DE LANGHE EA. 1990. Cryopreservation of Musa suspension cultures and subsequent regeneration of plants. Cryo-Letters 11:357-363. PHILLIPS RL, KAEPPLER SM, PESCHKE VM. 1990. Do we understand somaclonal variation? Pages 131-141 in Progress in Plant Cellular and Molecular Biology (Nijkamp HJJ, van der Plas LHW, van Aartrijk J, eds). Dordrecht, the Netherlands: Kluwer. POOL DJ, IRIZARRY H. 1987. Off-type banana plants observed in a commercial planting of ‘Grand Nain’ propagated using the in vitro culture technique. Pages 99-102 in Proceedings of the 7th ACORBAT Meeting, San José, Costa Rica, 23-27 Sep 1985 (Galindo JJ, Jaramillo R, eds). Technical Bulletin no.121. Turrialba, Costa Rica: CATIE. RAMACHARAN C, GONZALEZ A, KNAUSENBERGER WI. 1987. Performance of plantains produced from tissue culture plantlets in St. Croix, U.S. Virgin Islands. Pages 36-39 in Proc. 3rd Meet. Int. Assoc. Research on Plantain and Banana, 27-31 May 1985, Abidjan, Côte d’Ivoire. REUVENI O. 1990. Methods for detecting somaclonal variants in ‘Williams’ bananas. Pages 108-113 in The Identification of Genetic Diversity in the Genus Musa (Jarret RL, ed.). Montpellier, France: INIBAP. REUVENI O, ISRAELI Y, ESHDAT Y, DEGANI H. 1986. Genetic variability of banana plants multiplied via in vitro techniques. Final report submitted to IBPGR (no.PR 3/11). Bet Dagan, Israel: Agricultural Research Organization, The Volcani Center. SCHOOFS J. 1990. The INIBAP Musa germplasm transit center. Pages 25-30 in Musa: Conservation and Documentation. Proceedings of INIBAP/IBPGR Workshop, Leuven, Belgium, 11-14 Dec 1989. Montpellier, France: INIBAP. SCOWCROFT WR. 1984. Genetic Variability in Tissue Culture: impact on germplasm conservation and utilization. Rome, Italy: IBPGR. SMITH MK. 1988. A review of factors influencing the genetic stability of micropropagated bananas. Fruits 43:219-223. STOVER RH. 1987. Somaclonal variation in Grand Naine and Saba bananas in the nursery and field. Pages 136- 139 in Banana and Plantain Breeding Strategies (Persley GJ, De Langhe EA, eds). ACIAR Proceedings no.21. Canberra, Australia: ACIAR. STOVER RH, BUDDENHAGEN IW. 1986. Banana breeding: polyploidy, disease resistance and productivity. Fruits 41:175-191. STOVER RH, SIMMONDS NW. 1987. Bananas. London, UK: Longman. 468 pp. TOMES D. 1990. Current research in biotechnology with application in plant breeding. Pages 23-32 in Progress in Plant Cellular and Molecular Biology (Nijkamp HJJ, van der Plas LHW, van Aartrijk J, eds). Dordrecht, the Netherlands: Kluwer. TOWILL LE. 1990. Cryopreservation of isolated mint shoot tips by vitrification. Plant Cell Reports 9:178-180. VAN SLOTEN DH. 1990. IBPGR and the challenges of the 1990s: a personal point of view. Diversity 6(2). VUYLSTEKE D, SWENNEN R. 1990. Somaclonal variation in African plantains. IITA Research 1:4-10. VUYLSTEKE D, SWENNEN R, WILSON GF, DE LANGHE E. 1988. Phenotypic variation among in vitro propagated plantain (Musa spp. cultivars AAB). Scientia Horticulturae 36:79-88. WALBOT V, CULLIS CA. 1985. Rapid genomic change in higher plants. Annual Review of Plant Physiology 36:367- 396. WITHERS LA. 1980. Tissue Culture Storage for Genetic Conservation. Rome, Italy: IBPGR. 208 Early Detection of Somaclonal Variation

WITHERS LA. 1987. Long-term preservation of plant cells, tissues and organs. Oxford Surveys of Plant Molecular and Cell Biology 4:221-272. WITHERS LA. 1989. In vitro conservation and germplasm utilization. Pages 309-334 in The Use of Plant Genetic Resources (Brown AHD, Marshall DR, Frankel OH, Williams JT, eds). Cambridge, UK: Cambridge University Press. WITHERS LA. 1990. Prospects and problems of in vitro genebanks. Pages 21-24 in Musa: Conservation and Documentation. Proceedings of INIBAP/IBPGR Workshop, Leuven, Belgium, 11-14 Dec 1989. Montpellier, France: INIBAP. WITHERS LA. 1991a. Maintenance of plant tissue cultures. Pages 243-267 in Maintenance of Microorganisms (Kirsop BE, Doyle A, eds). London, UK: Academic Press. WITHERS LA. 1991b. In vitro conservation. Biological Journal of the Linnean Society 43:31-42. WITHERS LA. 1992. In vitro conservation. In Biotechnology of Perennial Fruit Crops (Hammerschlag F, Litz RE, eds). Wallingford, Oxon, UK: CAB International (in press). WITHERS LA, ENGELS JMM. 1990. The test tube genebank — a safe alternative to field conservation. IBPGR Newsletter for Asia and the Paciﬁc 3:1-2. Part 3

Technology Transfer JA Chambers, JI Cohen 211

Donor Assistance in Plant Biotechnology for the Developing World

JA Chambers1, JI Cohen2

Introduction Technological innovation in agriculture is poised on the threshold of signiﬁcant achievement in the 20th century. Rapid advancements in the application of biotechnology to crop improvement programs promise to extend and expand the technological base which, during this century, marked the critical transition in agriculture to a system reliant on science and technology to increase agricultural yields. This trend towards knowledge-intensive agriculture, while well-developed in industrialized countries, leaves developing countries at a distinct disadvantage. Developing countries generally lack the well-established public and private technological infrastructure present in industrialized countries which enables the latter fully to realize and apply the beneﬁts of technological achievement in agriculture (Parton 1990). This disparity in technological benefit is becoming especially apparent with the increasing application and integration of biotechnology in agricultural improvement programs in industrialized countries. In recent years, development efforts have attempted to address this technological disparity in agriculture through multilateral and bilateral assistance efforts. Multilateral support to the international agricultural research centers (IARCs), such as those in the Consultative Group on International Agricultural Research (CGIAR) system, which has recently included INIBAP among its members, has been rewarded as evidenced by the significant increases in food production achieved during the green revolution of the 1960s. Increases in yield have resulted in increased income-generation for many farmers in developing countries. Yet despite these successes, food self-reliance continues to be an elusive goal for many developing nations, and worldwide agricultural production continues to suffer substantial losses of 20-40% from continuing pest, weed, and disease pressure (Walgate

1Biotechnology Specialist, USDA/RSSA, Office of Agriculture, Bureau for Research and Development, Agency for International Development, Washington, DC 20523-1809, USA; 2Biotechnology and Genetic Resource Specialist, Office of Agriculture, Bureau for Research and Development, Agency for International Development, Washington, DC 20523-1809, USA. (Note: Opinions presented in this paper are those of the authors and do not necessarily reflect those of the Agency for International Development or the United States Department of Agriculture.) 212 Donor Assistance in Plant Biotechnology for the Developing World

1990). Moreover, worldwide concern for a fragile natural resource base continues to mount in the face of increasing population pressures, decreased availability of prime lands for cultivation, and obvious environmental demise. Thus, concurrent with needs for greater production efficiency are new challenges, requiring continued scientific innovation to enhance environmental acceptability of agriculture. In an attempt to meet these challenges, biotechnology is becoming an integral part of the scientific foundation which comprises modern-day agriculture. The US Agency for International Development (AID) is committed to the application of biotechnology to address agricultural production constraints in developing countries. AID’s investment in this technology is apparent through many forms of support, including its continued direct support to the IARCs as they continue to integrate the tools of biotechnology with their conventional research programs. In addition, AID supports the IARCs through a program entitled “Collaborative Research on Special Constraints at the International Agricultural Research Centers”, which is designed to overcome specific obstacles to technological breakthroughs at the IARCs by funding collaborative research with US scientists. But perhaps some of AID’s most important recent achievements in enhancing agricultural biotechnology capacity within developing countries have been at the level of the national agricultural research programs. One example of such bilateral support was the first National Conference on Plant and Animal Biotechnology in Kenya, held from 25 Feb to 3 Mar 1990, which was jointly sponsored by AID, the Kenya Agricultural Research Institute, and a number of other donors. AID’s bilateral assistance efforts in agricultural biotechnology have encompassed a variety of crops, including the traditionally supported cereals, and have recently included more commercially important crops such as the estate and plantation crops. Such efforts have specific relevance for this workshop on biotechnology for banana improvement. Recently, AID’s Office of Agriculture, based in Washington, DC, has further strengthened its programmatic support for agricultural biotechnology with the development and award of a new project in plant biotechnology for developing countries. The project, which is entitled “Agricultural Biotechnology for Sustainable Productivity” (ABSP) will be discussed in specific detail later in this document. However, it is important to recognize that the design of this project derived from a number of inputs and value decisions which formed the basis of the project’s philosophical intent and programmatic content (Cohen 1991).

National Research Council Panel Report In part, the design of the ABSP project was structured after recommendations advanced by a National Research Council (NRC) panel of the US National Academy of Sciences entitled “Plant Biotechnology Research for Developing Countries”. This panel, which included members from private industry, universities, Collaborative Research Support Programs (CRSPs), IARCs, developing-country programs, and the US Department of Agriculture, was convened at the request of AID’s Office of Agriculture to consider priorities in plant biotechnology research that could provide near-term (3-5 years) JA Chambers, JI Cohen 213

benefit to agriculture in developing countries. Further contributions to the design of the ABSP project originated from correspondence between the Office of Agriculture and individual AID missions, which were asked to respond to an extensive series of questions regarding the utility of this type of project for their host countries. The NRC panel’s report, entitled “Plant Biotechnology Research for Developing Countries”, was issued in September 1990 (USNRC 1990). It elucidated a distinct need for a technical assistance program that would encourage and assist developing-country scientists in the development and application of biotechnology focused on critical agricultural constraints of local importance. A number of general institutional and technical priority areas were identified which could be used as a basis for the development of more specific recommendations after detailed discussion and consultation with local experts most familiar with local problems, needs, and opportunities. Institutional priorities identified included three areas: (a) biosafety: AID should assist developing countries to implement and monitor appropriate biosafety regulations; (b) intellectual property: AID should participate in the development of policies to promote international cooperation in intellectual property rights (IPRs); and (c) human resource development and networking: AID should enhance biotechnology capabilities through doctoral and postdoctoral fellowships, and nondegree training for plant biotechnologists. Thus, the report emphasized that the progress and success of any new funding initiative was predicated on support for the institutional, as well as technical dimensions of biotechnology. The technical recommendations were divided into three major categories. These categories outlined areas of technical opportunity which AID should consider in the design of a new project in plant biotechnology. They are presented in order of degree of general panel support, increasing scale of complexity, and the types of research capacity that are prerequisite to achieving results with more complex techniques. These research categories are thus those that appeared to present the greatest chance of ensuring that applications of biotechnology contribute to agricultural research in the near term. The first category includes: (a) tissue culture, micropropagation, and transformation: support the building of developing-country capacity in plant tissue culture technologies which can augment conventional plant improvement programs, including micropropagation, cell selection, embryo rescue and haploid techniques, protoplast fusion and regeneration; (b) micropropagation: develop capacity to use micropropagation to produce virus-free planting material of forest, plantation, fruit, vegetable, and tuber crops; and (c) crop transformation: support the development of transformation and regeneration techniques for crops of importance to developing countries, such as cassava and millet. The second category involves plant disease and pest control and includes the following: (a) Bacillus thuringiensis (Bt) strain and gene identification: assist developing countries to identify strains and clone genes of Bt, which are effective against insect pests important to tropical areas. While there are many companies working on this worldwide, insufficient effort has been focused on tropical pests; (b) antiviral strategies: support development of antiviral strategies to combat plant viruses that attack beans, cassava, sweet potatoes, groundnut, and tropical fruits and vegetables; and (c) pathogen diagnostics and probes: 214 Donor Assistance in Plant Biotechnology for the Developing World

support research to develop DNA probes, as well as antisera and monoclonal antibody probes for plant bacteria, fungi, and viruses that attack crops of importance in the developing world. The final category, genetic mapping of tropical crops, summarized interest in developing specific genetic maps for many of the major crop plants. It included the following guidance: (a) assist the CGIAR and developing-country crop breeders to acquire capacity to use RFLP maps in plant breeding of rice, maize, sorghum, cowpea, and other crops where these maps are becoming available; and (b) to undertake research on primitive germplasm, especially woody perennial plants.

Value Decisions Several value decisions were also factored into the project’s design process. These value decisions, which were developed and agreed upon by technical staff within the Office of Agriculture and then reviewed by staff within AID missions and national research programs, include (1) support for field-proven, applied research programs which are integrated with conventional agricultural research and sustainable agriculture approaches; (2) mutuality of benefit to US and developing-country participants; (3) support for and programmatic inclusion of both private nonprofit (university, land-grant institutions) and private-for-profit organizations; (4) pursuit of equitable intellectual property rights through patents and licensing arrangements which recognize the contributions of germplasm from the USA and developing countries and technical input from all participating countries; and (5) support for testing and evaluation of recombinant products in accordance with USDA/APHIS regulations and host-country approval mechanisms. The incorporation of these value decisions into the project’s design was based on an understanding of the technology’s potential and limitations for developing-country agriculture, a recognition of changing institutional trends and their influence on AID’s traditional involvement in development-based, agricultural research, and a growing level of experience among staff which derived from the Agency’s initial projects in biotechnology. A few of these value decisions, and their role in shaping the project, are detailed below. Integrated research in agricultural biotechnology Agricultural biotechnology may be defined as the science which focuses on cellular and molecular biology and the new techniques deriving from these disciplines targeted towards the improvement of the genetic make-up and management of crops and livestock. Continuing investment in this technology derives from the perception that, when integrated with conventional, science-based techniques, the alternative tools of biotechnology provide additional opportunities to address contemporary challenges for sustainable increases in agricultural production. The tools of biotechnology, when specifically provided to plant breeders, present opportunities for increased efficiency, targeted modifications, and reliability in crop production, while simultaneously ensuring increased profitability and environmental compatibility (Schneiderman 1990). However, a JA Chambers, JI Cohen 215

key component to the ultimate success of this technology is that these alternative tools, when applied at the cellular and molecular level, must be integrated with conventional research to assure product delivery to producers and consumers. Integration of this technology with conventional research programs is also essential if objectives are to be achieved in a timely manner. Ultimately, relationships between biotechnologists and plant breeders may take several forms. The least desirable scenario would be that where biotechnology programs lack any formal collaboration with breeders and develop and release materials accordingly. More acceptable situations may involve truly collaborative interactions, where plant breeders are involved at the project’s inception and throughout, thus establishing an interdisciplinary team approach, or may involve opportunistic interactions in which an initial collaboration between breeders and biotechnologists may be absent but is later formulated based on positive results from early trials. Breeders may then be more integrally involved in the evaluation and further development of genetically altered materials. The most favored approach, and that which was judged as one criterion for selection during the ABSP project review process, was based on establishing such integration as a function of initial program development (Cohen et al. 1988a). In addition, the project’s emphasis was targeted towards ongoing integrated research programs which had a field-proven history of success. This implied that the project would not provide funds to establish new biotechnology centers or to move scientiﬁc capacity out of ongoing agricultural research programs. Detailed developing-country studies and appraisals of US institutional capabilities confirmed this approach and increased interest in the forthcoming ABSP project (Hall, Klausmeier 1988). Further integration of biotechnology with sustainable agricultural concerns was considered essential. This would permit a more detailed examination of contributions from biotechnology to sustainable agricultural practices. Such potential contributions derive from seed-based technologies which could provide more effective and less chemically-dependent control of weeds, insects, and disease (Hauptili 1990).

Mutuality of benefit By design, the ABSP project was initially intended to derive mutual benefits from collaborative research between the US public- and private-sector institutions and effective partners in developing countries. In so doing, the enhancement of self- sustaining partnerships was envisioned, in which each country’s participants would derive economic benefit from returns on the original investment in research and capital. Such a focus on economic return is unique to a limited number of programs supported elsewhere in AID’s portfolio. However, the ABSP project was developed as the Agency’s first agricultural project in which research would be geared, whenever possible, towards a realization of economic return and mutual commercial benefit. These issues were first explored by AID in conjunction with the US private sector, IARCs, and national programs in April 1988 at a symposium entitled “Strengthening Collaboration in Biotechnology: International Agricultural Research and the Private Sector” (Cohen 1989). This symposium was sponsored by AID to encourage dialogue on 216 Donor Assistance in Plant Biotechnology for the Developing World

these issues and was followed by workshops directed at speciﬁc aspects of international collaboration in agricultural biotechnology. At that time it became apparent that there were many opportunities for AID to develop commercially-directed biotechnology research of mutual benefit to both industrialized and developing-country partners (Sawyer 1989). Recognizing the value of mutual beneﬁt also corresponds with guidance developed for implementing the CRSP projects under the authority of the Title XII Amendment (BIFAD 1985).

Private-sector collaboration The design of the ABSP project established a significant role and programmatic support for the private-for-profit biotechnology sector as well as AID’s traditional research partners, the US university community. The decision to include the private sector in a development-oriented plant biotechnology project derived from a series of recent, recognizable trends that have been changing the face of agricultural research in the developed world, and in the USA in particular. To date (Jan 1992), most biotechnology research in agriculture has been conducted in and targeted towards the markets of industrialized countries, with practical applications limited to those problems and scenarios unique to commercial agriculture. Increasingly, the problem for donors, such as AID, has become one of technology access due to the increasing privatization and ownership of valuable biotechnological methods and materials by the private sector in developed countries. This is, in part, due to the capacity of the new technological tools to shorten the historical lag time between basic biological discoveries and product development (USNRC 1987). This trend has been further stimulated by the growth of the venture-capital industry in the USA and its support for entrepreneurial, academic-like companies (Bull et al. 1982). In addition, decreased Federal funding for agricultural research in the public sector (especially the US university community) has been a catalyst for greater collaboration between public institutions and private biotechnology companies. This collaboration has been specifically aided by two legislative changes in the USA: (1) a 1985 US Patent Office decision which held a plant to be patentable, thus permitting the protection of a single trait in a particular plant (Bent 1989); and (2) the Federal Technology Transfer Act which authorized government-supported agencies and institutions to collaborate contractually with private companies, thus bridging the gap between basic research in the public sector and more applied, product-oriented research in the private sector (Tallent 1989). As a result, universities are now frequently patenting promising research products for their eventual license or sale to private companies. In return, they have been financially rewarded through royalty agreements with private industry which are utilized to sustain their continued research efforts. The net effect of the above trends and changing institutional relationships has been AID’s decreased access to its traditional partners (the universities) and an expanded role for its nontraditional partners (the private companies) in agricultural research. Thus, the ABSP project includes mechanisms to engage significantly the technical and managerial resources of both the US public and private sectors in one integrated program. JA Chambers, JI Cohen 217

Prior to the completion of the ABSP project design, AID’s Office of Agriculture had benefited from the award of a precedent-setting grant to the Plant Sciences Division of the Monsanto Agricultural Company. Following extensive negotiations, a grant agreement was signed between AID and Monsanto which would support the improvement of agronomically important root and tuber crops for Africa through the utilization of Monsanto’s proprietary technologies in plant biotechnology. Specific focus of the project is on the development of virus resistance through molecular transformation of cassava, sweet potato, and yams using virus coat protein technologies developed in collaboration with R Beachy. The project supports the placement of two senior African scientists for a 3-year period at Monsanto’s laboratory in St Louis. The project’s ultimate goal is to export improved root and tuber germplasm to Africa, after compliance with USDA/APHIS requirements, for further testing, prebreeding and cultivar development. The establishment of this novel agreement between AID and a major US agricultural company provided some historical perspective for the development of the ABSP project in that it illustrated the need for a comprehensive and integrated endeavor in the area of development-assisted transfer of biotechnology. Issues related to the field-testing of recombinant plants (biosafety) were raised early in the development of this agreement. The participating African scientists will gain experience in both state- of-the-art technologies as well as acquiring valuable expertise in US procedures for field- testing of genetically transformed germplasm. In addition, due to the highly proprietary nature of the technology in question, issues related to protection and ownership of intellectual property were at the forefront of the negotiation process. Under the terms of the agreement, Monsanto will grant “a royalty-free license to use in Africa any proprietary technology (including disarmed Agrobacterium strains, intermediate vectors, selectable markers, promoters, other expression and transformation technology and coat protein-based virus tolerance technology) developed by African scientists embodied within any cassava, yam, or sweet potato germplasm developed in this program” (USAID 1990). The specific advantages for all parties concerned with this grant are listed in Table 1. In addition, AID’s experience with the Monsanto agreement provided an invaluable resource for the design of future collaborative efforts with private companies which are a significant component of the ABSP project.

Intellectual property protection The 1980 landmark Supreme Court decision in Diamond vs Charkrabarty, which held a microorganism to be patentable and, in so doing, ruled in favor of the protection of animate nature whose existence resulted from human intervention, has directly favored the growth of the biotechnology industry in the United States (Beier, Strauss 1985). Since then, numerous biotechnology patents have been filed in the USA by both universities and industry alike, thus catalyzing the inclusion of initiatives related to intellectual property rights (IPRs) in some of AID’s development-assistance efforts. The design of the ABSP project, which provides substantial support for IPRs, reflects the recognized awareness for effective coordination of IPR initiatives in technology transfer programs to developing countries. This awareness stems from a desire equitably to 218 Donor Assistance in Plant Biotechnology for the Developing World

Table 1. Advantages derived from the AID and Monsanto Corporation Agreement (Cohen, Chambers 1992). a) Secures specific rights for specific proprietary technologies used in accordance with conditions of the grant. b) Places senior African scientists within a commercial laboratory setting to become familiar with commercial orientation, concerns, and decisions. c) Familiarizes African scientists with testing protocols and regulatory compliance. d) Transformation provides all parties with gene expression data essential to expand research efforts. e) Establishes precedence for donor grants to commercial entities and gain acceptance for opportunities within commercial legal establishment. f) Establishes relationship between national program, private sector, donor, and IARC for biotechnology research and product development. protect the sources of and contributions to the resulting inventions, whether they are rooted in proprietary technology contributed by advanced laboratories or indigenous germplasm contributions and “know-how” provided by developing-country collaborators. Thus, a key component of agreements filed under the ABSP project will be an initial understanding and delineation of the IPRs which will accrue to participants. It may be difficult to enforce violations of proprietary agreements which result from unauthorized use of germplasm or technologies and litigation may not always be possible or warranted, particularly in developing countries that lack significant legal infrastructure. Nonetheless, advanced laboratories, both public and private, will differentiate between those institutions which present lesser or greater risk. This is an important point, for once a negative reputation has been formulated, opportunities for further collaboration will decrease as other universities or commercial firms become aware of the problem. Those who violate legal agreements may gain in the short term, but may eventually suffer as new material is developed and no longer released to a particular center or national program.

Biosafety Integrating biotechnology, and specifically genetic engineering, into development- assisted agricultural research requires an additional consideration of biosafety procedures surrounding the transfer of technology to developing countries, which often lack appropriate regulatory infrastructure. Complications surrounding the establishment of such infrastructure in the developing world are due to a variety of influences. Foremost among these is the diverse array of players involved in agricultural biotechnology research and testing (Cohen et al. 1988b). These include the IARCs; donor agencies which provide support to IARCs and national programs; governments and research institutions of developing countries; governments, research institutions, and private industry of industrialized nations; and various environmental and public JA Chambers, JI Cohen 219

advocacy concerns. Each of these groups has a specific agenda and an intrinsic definition of what constitutes safe and appropriate use of biotechnology products in the environment. Furthermore, the continuing “product vs process debate” has figured dramatically in the role of regulatory oversight with respect to biotechnology products. The perceived power or specificity associated with particular genetic modification technologies is directly related to an increase in regulatory concern or oversight and continues to fuel this debate. Nonetheless, despite the confusion and complications in the regulatory arena, enhancement of host-country capabilities in biosafety, including national abilities to evaluate, approve, and monitor specific requests for testing, is essential if these countries are fully to participate in and benefit from the exchange of technology. Recent AID initiatives in biosafety reflect the Agency’s awareness of a need to increase indigenous capacity in coordination with its sponsorship of technical programs. AID has adopted a “case-by-case” approach to projects warranting close regulatory involvement. In addition to domestic concerns and regulatory responsibilities, AID must now consider how grant recipients will effect contained testing conducted in its client countries. Coordination between the principal investigator of each grant, as well as with AID, has been rigorously pursued to ensure that applicable regulations and oversight procedures are undertaken (Cohen, Chambers 1990). As an example, since 1986 the Office of Agriculture includes, as a standard provision of grants supporting recombinant DNA research, specific language regarding regulatory compliance and approval from US regulatory agencies (Jones 1987). Thus, there is a formal written requirement by the Office that any risks posed by the research be fully addressed in the context of US government regulations prior to obtaining host-country approval. In addition, this approval must be obtained in writing, as must that of the AID in-country mission. Further support of the Agency’s commitment to this issue is seen in its 1987 establishment of the Standing Committee on Biotechnology which functions to provide: (1) an agency forum for advice on issues surrounding biotechnology and on the development of mechanisms to address those issues that are technical, regulatory, and programmatic in nature; and (2) an entity within AID which will provide liaison with US government scientific and regulatory agencies principally concerned with supporting and regulating biotechnology (USAID 1987). The Standing Committee serves a central purpose in the review and dissemination of information and the coordination of domestic and host-country regulations affecting biotechnology research and offers specific procedural guidance for AID projects. By establishing this committee, AID recognized the role which biotechnology research occupied in its development programs, and also addressed the Agency’s need safely and responsibly to monitor and implement the products of its supported research. This initiative has been continued with the establishment of the ABSP project, which provides significant financial resources in support of regulatory activities which are necessary to safely move any genetically engineered products developed by the ABSP research effort into the field. 220 Donor Assistance in Plant Biotechnology for the Developing World

A New AID Initiative: the ABSP Project The US$6 million, 6-year ABSP project was authorized in 1991 by AID’s Bureau for Research and Development to “mutually enhance U.S. and LDC institutional capacity for the use and management of biotechnology to develop environmentally compatible, improved germplasm”. The project provides direct support for research and development activities in the form of grants to US public and private sectors and developing-country institutions for research to address agricultural production constraints in AID’s client countries. It provides indirect support for private-sector involvement (both in the USA and developing countries) in agricultural research through activities designed to increase host-country capability in the competitive management of biotechnology. This indirect support is evidenced through the project’s planned activities in intellectual property, biosafety, networking, and the establishment of downstream commercial linkages for product development. As such, the ABSP project is multidisciplinary: it supports (1) human resource development via long-term, postdoctoral training in advanced US institutions, both public and private; (2) constraint-oriented, as opposed to commodity-based, biotechnology research and development within private and public institutions; (3) activities, either through internships or expert consultation, in IPR and biosafety, to increase host-country capacity; (4) increased host-country capacity in the management of biotechnology through commercially-oriented seminars in private US biotechnology firms; (5) networking through regional and international conferences, trade association memberships within the Association for Biotechnology Companies, newsletters, directories, developing-country laboratory support, and database linkages; (6) product development via an incentive-based system for funding field-testing and commercialization activities for promising products of laboratory research. A peer-review process for submitted proposals was utilized in determining the award for the project. A Request for Application (RFA) to US public and private institutions was issued in May 1991. The external review, conducted by the National Research Council, established individual rankings for the public- and private-sector proposals. Based on this ranking and on a financial analysis conducted by AID’s procurement office, the cooperative agreement for the project was awarded to Michigan State University (MSU) as the lead institution and management entity in a consortium of universities which also includes Texas A&M and Cornell. J Dodds, formerly of the International Potato Center, is the project’s managing director. The university research component is under the coordination of M Sticklen, of MSU. The private-sector subgrant was awarded to DNA Plant Technology Corporation (DNAP), based in New Jersey. Proprietary agreements will be established for all aspects of joint research programs conducted between the various universities, developing-country collaborators, and private companies. J Barton, a professor at Stanford Law School, has been retained as chief counsel to advise the project on intellectual property issues. The technical focus of the university research component is based on the use of recombinant DNA technology and crop transformation techniques to genetically engineer resistance to a variety of crop pests and pathogens in agronomically important JA Chambers, JI Cohen 221

crops. The expected outputs for this research will be to increase developing-country crop production through the availability of genetically resistant germplasm, and ultimately to decrease dependence on chemical inputs. Specific research activities include: (1) genetic engineering of potato for tuber moth resistance; (2) genetic engineering of sweet potato for weevil resistance; (3) genetic engineering of maize for stem borer resistance; and (4) genetic engineering of cucurbits for potyvirus resistance. Host-country collaborators initially identified include Kenya, Ecuador, and Indonesia, although additional collaborators will also be pursued in an attempt to broaden the project’s global relevance. The technical focus of the private-sector subgrant to DNAP is under the direction of M Söndahl. As opposed to the MSU/Texas A&M/Cornell technical program, which is strongly oriented towards the use of rDNA technologies, the DNAP program concentrates on the use of cell culture-based techniques of micropropagation to increase yields and improve quality of several tropical crops varieties. Under the terms of the agreement, DNAP will collaborate with a private company in Costa Rica, Agribiotecnología de Costa Rica, to utilize DNAP’s proprietary technology for (1) the development of high-frequency embryogenic cell cultures for targeted crops as a prelude to developing large-scale micropropagation systems; (2) the genetic improvement of elite, but currently limited, planting stock; and (3) the utilization of the technology and planting material as the base for commercial production of targeted crops in Costa Rica. The four crops targeted by this agreement include banana, pineapple, coffee, and ornamental palms. Ancillary activities on the project’s agenda include support for short-term internships for host-country participants in intellectual property and biosafety at US legal and regulatory institutions, respectively; inclusions of IARCs in the project’s network, which will sponsor three regional and one international conference; support for ABC trade association memberships for all project participants; and, funding for field-testing, product development initiatives, and economic impact studies. To date (Jan 1992), response to the project has been very positive among AID in- country missions, international agricultural research centers, and national programs in developing countries. Expansion of the project’s activities, through supplemental funding by individual AID missions, is currently under development and may result in additional research agreements between institutions not presently supported under the core agreement.

References

BEIER FK, STRAUSS J. 1985. Patents in a time of rapid scientific and technological change: inventions in biotechnology. Pages 15-35 in Biotechnology and Patent Protection (Beier FK, Crespi RS, Strauss J, eds). Paris, France: Organization for Economic Co-operation and Development. BENT SA. 1989. Patenting genes that encode agriculturally important traits. Pages 109-122 in Intellectual Property Rights Associated with Plants. Madison, WI, USA: American Society of Agronomy. BIFAD (Board for International Food and Agricultural Development). 1985. Guidelines for the collaborative research support programs under Title XII of the International Development and Food Assistance Act of 1975. Washington DC, USA: Agency for International Development. 222 Donor Assistance in Plant Biotechnology for the Developing World

BULL AT, HOLT G, LILLY MD. 1982. Inportant issues affecting developments in biotechnology. Pages 57-63 in Biotechnology: International Trends and Perspectives. Paris, France: Organization for Economic Co- operation and Development. COHEN JI (ed.). 1989. Strengthening Collaboration in Biotechnology: international agricultural research and the private sector. Washington DC, USA: Agency for International Development. COHEN JI. 1991. Value decisions responsive to international initiatives in plant biotechnology. (Paper presented at the CCSA Plenary Session: Ethical and Policy Issues in Global Crop Science, convened by G Heichel.) Crop Science (in press). COHEN JI, CHAMBERS JA. 1990. Biotechnology and biosafety: perspective of an international donor agency. Pages 378-394 in Risk Assessment in Genetic Engineering: Environmental Release of Organisms (Strauss HS, Levin M, eds). New York, USA: McGraw-Hill. COHEN JI, CHAMBERS JA. 1992. Industry and public sector cooperation in biotechnology for crop improvement. Pages 17-30 in Biotechnology and Crop Improvement in Asia (Moss JP, ed.). Patancheru, India: International Crops Research Institute for the Semi-Arid Tropics. COHEN JI, JONES KA, PLUCKNETT DL, SMITH NJ. 1988a. Regulatory concerns affecting developing nations. Bio/Technology 6:744. COHEN JI, PLUCKNETT DL, SMITH NJH, JONES KA. 1988b. Models for integrating biotechology into crop improvement programs. Bio/Technology 6:387-392. HALL P, KLAUSMEIER WH. 1988. Opportunities to commercialize life science applications in less developed countries: a strategic plan. Arlington, VA, USA: Resources Development Foundation. HAUPTILI H, KATZ D, THOMAS BR, GOODMAN R. 1990. Biotechology and crop breeding for sustainable agriculture. Pages 141-156 in Sustainable Agricultural Systems (Edwards CA, Lal R, Madden P, Miller RH, House G, eds). Akeny, Iowa, USA: Soil and Water Conservation Society. JONES KA. 1987. A.I.D.’s activities in biotechnology: regulatory considerations. Contract DAN-1406-0-7008-00. Washington DC, USA: Agency for International Development. PARTON TR. 1990. Issues affecting Technical Change and its impact on the International Agribusiness Environment. Board on Agriculture, National Research Council. Washington DC, USA: National Academy Press. SAWYER RL. 1989. Building bridges of research collaboration through the CGIAR system of centers. Pages 13-19 in Strengthening Collaboration in Biotechnology: International Agricultural Research and the Private Sector (Cohen J, ed.). Washington DC, USA: Agency for International Development. SCHNEIDERMAN HA. 1990. Innovation in agriculture. Pages 97-106 in Technology and Agricultural Policy. Board on Agriculture, National Research Council. Washington DC, USA: National Academy Press. TALLENT W. 1989. A new era of government-industry cooperation in research and development. Pages 423-428 in Strengthening Collaboration in Biotechnology: International Agricultural Research and the Private Sector (Cohen JI, ed.). Washington DC, USA: Agency for International Development. USAID (US Agency for International Development). 1987. Charter for agency Standing Committee on Biotechno- logy. Washington DC, USA: the Agency. USAID (US Agency for International Development). 1990. Grant Agreement DAN-4109-G-00-0085-00. Improving cassava for African farmers through biotechnology. Washington DC, USA: USAID Contracts Ofﬁce. USNRC (US National Research Council). 1987. Committee on a National Strategy for Biotechnology in Agriculture, Board on Agriculture. Pages 51-144 in Agricultural Biotechnology: Strategies for National Competitiveness. Washington DC, USA: National Acedemy Press. USNRC (US National Research Council). 1990. Plant Biotechnology Research for Developing Countries. Washington DC, USA: National Academy Press. 44 pp. WALGATE R. 1990. Finding genes for better crops. Pages 6-22 in Miracle or Menace? Biotechnology and the Third World. London, UK: Panos Institute. Part 4

Recommendations Working Group 1 225

Nonconventional Strategies for Producing Banana and Plantain Resistant to Pathogens and Pests

Working Group 1

Introduction The potential for new technologies was readily apparent to workshop participants. Maps of the Musa acuminata and M. balbisiana genomes may well be produced within the next 2 years, and cultivars transformed for useful traits could be available in 3 years. It was stressed, however, that these new strategies be viewed as complements and not replacements for conventional breeding. It was also stressed that reasonable optimism be used when anticipating results with the new techniques. Since many of these strategies (Table 1) are complicated and expensive, signiﬁcant or rapid progress may be impeded by technical difﬁculties and funding constraints.

Construction and Use of a Genetic Linkage Map of the Musa Genome for the Improvement of Bananas and Plantains Although considerable progress has been made in breeding bananas resistant to some of their most important pathogens, very little is known about the genetic inheritance of these or other characters. With the advent of new molecular genetic marker technologies such as RFLPs and RAPDs, it should be possible, as has been the case for many other crops, to construct a genetic linkage map of the banana genome. A publicly available map will serve to mark important agronomic traits and will promote more rapid progress in existing and future breeding programs. It will also provide a unique tool for studying and manipulating characters such as sterility arising from structural heterozygosity (very common in the diploid bananas) or genic factors, and parthenocarpy, a character that has probably played a major role in the domestication of diploid bananas. The construction of a genetic map requires that certain conditions be fulfilled: distinguishable alleles should exist at as many loci as possible; crosses for the generation of segregating populations of adequate size should be feasible; methods for the measurement and analysis of characters of interest (such as resistance to pathogens) should be available. 226 Nonconventional Strategies for Producing Banana and Plantain Resistant to Pathogens & Pests

Table 1. Priorities for the biotechnological improvement of banana and plantain Target Application —————————————— ——————————————————— Importance1 Diagnosis2 Control, other3 ———————— ———————— —————————————— Local Export Approach Need Approach Potential Diseases caused by: Fungi Sigatoka H HRAPD Y markers H complex toxin(s) H various genes M Fusarium wilt H M RAPD? Y markers H various genes M Viruses Banana bunchy top H D MonoAb, N Coat protein gene? M cDNA Banana bract mosaic M D Potyvirus Y Coat protein gene H Banana streak L D PolyAb, N Coat protein gene? M cDNA Cucumber mosaic L D MonoAb, N Coat protein gene H PolyAb, cDNA Abaca mosaic L D Potyvirus Y Coat protein gene H Abaca bunchy top L D cDNA Y Coat protein gene? M Other agents Bacteria L L ? Y Vector interactionsM Banana weevil M L - - Markers H BT gene(s)? M Other gene(s) M Nematodes M L Develop’g Y Markers H Various genes M Other targets: Ripening, H M - - AntiPG M postharvest AntiACC M 1Importance of target for locally consumed and exported bananas are high (H), medium (M), or low (L). Other targets were discussed (D) but determined to be relatively unimportant. 2Approaches for diagnosis include the use of: randomly amplified polymorphic DNAs (RAPDs); monoclonal antibodies (MonoAb); polyclonal antibodies (PolyAb); and complementary DNA (cDNA), to detect disease agents. Broad-spectrum kits, which recognize 95% of all known potyviruses, may be useful for detecting the banana bract mosaic and abaca mosaic viral agents. Whether the need is yes (Y) or no (N), work on an approach is based on the status of development for the approach and the specific agent. 3Biotechnological approaches to control various diseases or otherwise to improve banana and plantain include: the identification of resistance markers that would assist breeding efforts; the use of host-specific toxin(s) produced by Mycosphaerella fijiensis to select resistant variants; and transformation of the host with various genes. Potentially useful genes for transformation include those which produce: chitinases, β1,3-glucanases, chitin-specific lectins, and nikkomycins for diseases caused by fungi; viral coat proteins for diseases caused by viruses; toxins produced by Bacillus thurinigensis, chitinases, and other oral and lytic enzymes for banana weevil and/or nematodes; and antisense polygalacturonases (antiPG) and inhibitors of ethylene production (antiACC) that extend the postharvest life of fruit. Since none of these strategies has been tested for banana and plantain, the potential for success with the various target:approach combinations was evaluated based on current success with other plants. Working Group 1 227

To date sufﬁcient data are available to indicate that the DNA of many diploid banana accessions is highly polymorphic and that the phenomenon may well be a general one. Crosses between some of the accessions (see González de León’s paper in this volume) generate populations from which a saturated linkage map could be constructed. Marker-assisted selection in banana breeding is therefore a realistic goal which will require the creation and/or strengthening of collaborative projects between breeding programs and laboratories applying the new technologies. Furthermore, the relatively small size of the banana genome should facilitate advances in the construction of a physical map which should permit the isolation and characterization of specific banana genes for use through nonconventional breeding methodologies. Recommendations Work on the construction and use of a saturated linkage map of the Musa acuminata and M. balbisiana genomes should be continued and enhanced through increased support of collaborative projects. Immediate action for achieving these goals should include the following. • The development of a wide variety of segregating populations according to speciﬁc breeding objectives. • The construction of a publicly available high-density core map at a level of saturation that will suit marker-assisted selection strategies. Long-term objective: • The construction of a physical map using long-range mapping strategies (Cosmid and YAC libraries) which should facilitate the identification and eventual isolation of genes of interest.

Transformation An efficient system for transforming bananas will be needed before this important crop can be genetically engineered. The essential components of such a system are: (1) an efficient DNA delivery technique (for instance, biolistics); (2) an effective selection system (for instance, antibiotic or herbicide resistance); (3) promoters to drive high expression in Musa; and (4) an efficient plant-regeneration system. Preliminary results presented at this workshop indicate that progress has been made in each of the areas. However, it must be stressed that considerable research and development is required before an efficient Musa transformation system is available. Several useful genes exist that could be incorporated into banana and plantain, particularly pest and disease resistance genes (Table 1), and new genes will become available within a year. Recommendations • The development of an efficient Musa transformation system should be given very high priority. 228 Nonconventional Strategies for Producing Banana and Plantain Resistant to Pathogens & Pests

• The incorporation of virus resistance, particularly coat protein-mediated resistance, should be the ﬁrst priority for the incorporation of useful traits into Musa. In this instance, incorporation of genes for cucumber mosaic virus coat protein could serve as a useful model system. In addition, development of strategies for conferring resistance to banana bunchy top virus should be pursued. • Resistance to banana borer weevil, mediated by genes from Bacillus thurinigensis for toxin production should be investigated. • Other potentially useful genes, including those for fungal resistance (chitinase, nikkomycin, etc.) and extended postharvest life of fruit, should be investigated.

Identifying and Generating Resistance New and more efficient techniques are required for identifying and generating resistance to plant diseases and pests. These may include apparent host-speciﬁc toxins, such as that produced by Mycosphaerella fijiensis, elicitors for host-resistance responses in cases of diseases caused by nematodes and other fungi, and irradiation to induce new sources of resistance.

Virus Indexing One of INIBAP’s primary activities is to facilitate the exchange or distribution of germplasm. Within this activity is the inherent risk of distribution of viruses in tissue- culture material. Early in 1992 there was no information regarding the transmission of viruses through Musa seeds. To reduce the possibility of distributing virus-infected material, INIBAP has implemented a virus indexing program by which accessions moving through the transit centers are tested for virus infection in two indexing centers. However, no one center has a collection of all the known Musa viruses or the accessibility to implement the range of appropriate diagnostic technologies. Virus detection techniques continue to be developed and improved as “new” viruses of Musa are being described. It is therefore essential that those developments are incorporated in the Musa virus testing program to increase the efﬁciency and security of the germplasm distribution program. Recommendations • It is recommended that INIBAP establish at least one central facility for Musa virus testing and that: • The facility have the capacity to test for all known Musa viruses. • The facility develop expertise in the application of contemporary virus diagnostic technology including Elisa (using monoclonal and polyclonal antisera), molecular hybridization, and the polymerase chain reaction. • The facility obtain and maintain in-vivo cultures of all known Musa viruses, to be used as positive controls. Working Group 1 229

• The facility monitor new developments in diagnostic technology and incorporate these technologies into the testing program as appropriate. • The facility monitor and investigate reports of “new” viruses in Musa and incorporate tests for these viruses when available. • Groups developing diagnostic tests for Musa viruses make these available to the central facility. • INIBAP encourage research on the characterization and diagnosis of Musa viruses.

Understanding Pathogen Diversity The importance of understanding pathogen diversity is well understood. Current research on variation in Mycosphaerella ﬁjiensis, Fusarium oxysporum f.sp. cubense, Radopholus similis, and BBTV should continue. Up to early 1992 work on the fusarium wilt pathogen has progressed furthest. Pathogenic and genetic diversity in F. oxysporum f.sp. cubense is extensive and data have accumulated indicating that the performance of even some tolerant banana clones varies from one location to another. Thorough research has just begun on the BLS burrowing nematode and banana bunchy top pathogens. There is evidence that there exists diversity in each that affects the performance of important cultivars. Recommendations • INIBAP-assisted efforts to collect F. oxysporum f.sp. cubense in Southeast Asia, a center of origin for Musa acuminata and its hybrids, with M. balbisiana, should continue. Minor support should be made available for maintenance of the Homestead collection of this pathogen and for the distribution of accessions it contains. Techniques for directly and indirectly accessing pathogenicity in all the agents would be useful. The development of such techniques should receive high priority.

Screening Methods Tissue culture, somaclonal variation, and irradiation techniques are being used to generate variants and mutations that may have resistance to serious pathogens of Musa. The detection of useful germplasm generated by these methods is handicapped by the lack of rapid, reliable screening techniques for determining disease reactions. Field- testing is costly and time-consuming. Laboratory and greenhouse screening methods would be extremely useful and more appropriate. Recommendation • Research to develop rapid, reliable techniques for screening Musa germplasm for reaction to Sigatoka and fusarium wilt diseases to detect useful resistance, should be encouraged. This would involve a continuation of investigations in the role of toxins as possible screening agents for detecting resistance to Sigatoka disease. Work to 230 Nonconventional Strategies for Producing Banana and Plantain Resistant to Pathogens & Pests

develop a rapid, reliable test for detecting resistance to fusarium wilt is also seen as very important.

Cooperation between Biotechnologists and Plant Breeders It is recognized that new biotechnological or nonconventional approaches to plant breeding offer many useful techniques and opportunities for success. However, they cannot be successful in a vacuum. Close continued cooperation must exist between biotechnologists and conventional plant breeders if society is to be the beneﬁciary of this technology. Working Group 2 231

In-Vitro Strategies for Musa

Working Group 2

Introduction The Working Group on this topic acknowledged the signiﬁcant impact and progress that has been made in Musa propagation and improvement in recent years using in-vitro techniques. The Group also wished to emphasize that these techniques are complementary to more established techniques for propagation and improvement of bananas and plantains.

In-Vitro Techniques for Germplasm Handling

Micropropagation Micropropagation is a useful technique for the mass propagation of disease- and pest- free planting material. Techniques are currently well established and applicable to a wide range of genotypes, but there is a continual need to upgrade and develop more efﬁcient and cost-effective strategies. Recommendation • It is recommended that studies involved with the development of more efficient micropropagation methods that address the issues of somaclonal variation, and of more rapid multiplication, should be supported. Also, INIBAP and other organizations should endeavor to strengthen in-country capability for in-vitro culture in developing countries, including the training of personnel.

Conservation In-vitro culture has become an important means of conserving Musa germplasm. Within a complementary conservation strategy, alongside field gene banks, seed and pollen storage, in-vitro techniques can help achieve secure, efficient, accessible and sustainable conservation of the Musa gene pool. Techniques for medium-term conservation through slow growth are relatively well developed. Progress has been made with long-term in-vitro storage of germplasm through cryopreservation, but there is still room for improvement. 232 In-Vitro Strategies for Musa

Recommendations • There is a need to establish a rational basis for management of in-vitro germplasm collections. • Research is needed better to understand the behavior of cultures over extended periods of slow-growth storage, particularly with respect to variability. • Continued research is required in the area of cryopreservation to improve the technique to include a range of culture systems, and to deﬁne conditions that are applicable to a diverse range of germplasm.

Germplasm exchange Considerable advantages are conferred by the safe movement of pathogen-tested germplasm by in-vitro methods. Detailed guidelines have been formulated by FAO, IBPGR, and INIBAP. However dissemination of germplasm could be accelerated. Recommendations • Research and information is needed to develop simple, sensitive, and reliable diagnostic methods for the early detection of banana and plantain diseases, particularly for BBTV and other obscure pathogens. These techniques and adequate quarantine facilities are needed in developing countries. • The FAO/IBPGR/INIBAP guidelines should be updated as new knowledge in diagnostics is acquired. •INIBAP should endeavor to enhance the efficiency of its germplasm exchange program by introducing recently developed diagnostics to the indexing centers.

Collecting The establishment of an in-vitro germplasm collections is a big undertaking that frequently involves collecting and transporting suckers to laboratories many kilometers away or in different countries. Apart from logistical problems, soil-bearing propagules increase the risk of introducing serious pests and diseases. Recommendation • There is a need to build upon the promising preliminary research that has been carried out on in-vitro collecting for Musa in order to reﬁne techniques and make them more widely available.

In-Vitro Techniques for Genetic Improvement In-vitro techniques have enormous potential for the genetic improvement of bananas and plantains by overcoming barriers to the full utilization of the gene pool. Many techniques are available and have been used successfully in Musa. Some of the techniques available include: embryo rescue; meristem culture; cell and protoplast culture; haploidy; mutagenesis; and chromosome doubling. Those requiring particular attention were identiﬁed by the Working Group. Working Group 2 233

Haploid methods Effective strategies have been developed to produce haploid material for the production of homozygous clones to facilitate and accelerate practical breeding programs and to develop a greater understanding of the genetics of Musa. Progress has been made on a number of fronts. Recommendation • Research into induction of haploidy should continue and, at this early stage in application of the techniques, it is proposed to convene a meeting with groups interested or currently working in this area.

Cell culture Simple and adaptable cell-culture techniques are being developed that have application for: (a) rapid clonal propagation; (b) transformation; and (c) in-vitro selection. Advances in this area have been very rapid and there is room further to streamline techniques and extend the use of these techniques for a wider range of cultivars. Recommendations • There is a need further to develop somatic embryogenesis techniques for the mass propagation of desirable clones. The scale-up and automation of techniques necessary to reduce the costs of production further should be investigated. In addition, field-testing of plants regenerated from cell culture should be investigated. • It is imperative that research should be supported to develop cell-culture techniques to enhance the selection and regeneration of transformed plants. Considerable opportunities for utilizing novel gene products in improvement are now at hand. The stable integration of these genes into whole plants is the goal, and research in this area is needed to develop efficient and applicable techniques. The Group also recommends that a meeting be convened to facilitate interaction with groups involved in Musa transformation strategies, and it strongly supports research in this area. • In-vitro screening and regeneration of plants from selected cell lines has been identified as a means of obtaining plants with tolerance of/resistance to pathogens and pests, with or without the use of mutagenic procedures to increase variation. Further research is needed to refine these approaches to plant improvement.

Protoplast culture There are bottlenecks in conventional breeding, and recent positive results with protoplast culture have demonstrated their use as a means of bringing together plants of various ploidies and with different genomic and cytoplasmic backgrounds to produce novel combinations. Although the use of protoplast technology is usually more technically difﬁcult, its uses should not be ignored. 234 In-Vitro Strategies for Musa

Recommendation • Research is still needed on the development of protoplast fusion techniques to produce somatic hybrids and on electroporation of single-cell systems for use in transformation studies. These techniques need to be extended to cultivars of agronomic value.

Somaclonal Variation: detection and characterization The Group acknowledges the occurrence of somaclonal variation as a serious impediment in the in-vitro handling and improvement of Musa. Somaclonal variants are usually inferior; they can result in serious economic losses to the farmer and can impact negatively on the work of scientists. Somaclonal variation is linked to both inherent and culture-related factors. In certain cases and applications, particularly micropropagation, inherent factors might largely be controlled by the choice of more genetically stable cultivars and explants. Culture-related factors may be manipulated and culture management strategies developed to reduce the risk of somaclonal variation. However, the knowledge base is presently inadequate and more research is urgently needed. The Group identiﬁed the following research priorities for concurrent attention covering methods for the early detection of somaclonal variation and the development of a better understanding of the factors affecting somaclonal variation and its causes and origins. Recommendations • The Group recommends the undertaking of a survey of somaclonal variants in Musa. This will include the identification and characterization of the more significant somaclonal variants in order to understand the magnitude of the problem and focus on the development of detection methods. •To facilitate the characterization of somaclonal variation, the Group recommends the compilation and distribution of a set of standardized morphological descriptors of somaclonal variants of bananas and plantains. The Group recommends that these descriptors could be prepared and disseminated through INIBAP’s Information Services. • Methods for the early detection and characterization of somaclonal variants are needed and may be based on a combination of morphological, biochemical, and molecular markers. It is recommended that a concerted research effort be made to develop appropriate and efﬁcient strategies for early detection. • The Group recommends the undertaking of research into the factors influencing somaclonal variation. This should be integrated with research on early markers in order to provide opportunities to accelerate research and give timely feedback to improve culture procedures. • In the course of all of the above research, the fundamental issues of the causes and origins of somaclonal variation should be addressed. Working Group 2 235

Coordination and Dissemination of Information on In-Vitro Techniques and Applications In order that Musa in-vitro technologies may have greater impact, information must be collected and disseminated in a timely fashion to groups engaged in this work. This is particularly true of the need for information in developing countries. The establishment of regional networks with an extension focus for growers is also important in this regard. The Group acknowledges the potential of using INIBAP’s information facilities to contribute to a better ﬂow of information on biotechnology. 236 Technology Transfer

Technology Transfer

Working Group 3

Introduction The task assigned to this Group of 35 workshop participants was to explore ways of facilitating biotechnology transfer to scientists working in banana- and plantain- producing countries. This topic was developed in a preworkshop paper entitled “A Strategy for Action”, which explored the workshop objectives, organization, and expected results. The Group used the nominal group technique (NGT) of Delbecq et al. (1975), which comprises a highly structured brainstorming session, to develop responses to the single question: What activities are needed to facilitate access to newer biotechnologies for banana and plantain genetic improvement for producing countries? Following a brief training session in NGT for the participants, the Group identified 178 items related to the target question. Following a period of clarification, a few items were combined to eliminate clear duplication. The items were then prioritized and, through this balloting process, 12 items were identified as collectively important: 1. Financial support. 2. Training. 3. Develop laboratory infrastructure. 4. Bi-directional scientific exchange. 5. Link biotechnology laboratories to breeding programs. 6. Provision of documents. 7. Identify targets with defined value. 8. Funding for collaborative research. 9. Appoint INIBAP biotechnology coordinator. 10. Networking of international programs and IARCs. 11. Link research laboratories in producing countries to biotechnology laboratories. 12. Transfer technology. Following prioritization, a few items were eliminated from the list of 178 as being either not serious or inappropriate for technical or policy reasons. These were removed by consensus and totaled only four items. The remainder of this report organizes the responses according to the categories identified through prioritization, conceptually similar items being merged (items ranked 5, 10, and 11; items ranked 1 and 8). An additional topic dealing with policy, research administration, and public information was created to deal with the remainder of the items not associated with the foregoing groupings. Working Group 3 237

Financial support (items 1, 8) Access to ﬁnancial support for biotechnology is needed to fund collaborative research that will apply biotechnology to the improvement of banana and plantain. Additionally, there is a need to provide access to consumable materials; that might include lower prices through research subsidies or some other mechanism. There is also a need to provide logistical support for biotechnology research that would include services such as germplasm exchange and access to genetic materials (e.g., cDNA, YAC libraries, etc.). It was proposed that foreign aid in developed countries should increase to 0.6% of their gross national product and that these countries should give consideration to funding research on banana and plantain biotechnology in other developed-country laboratories to support the application of biotechnology in producing countries. Another suggestion was to place a tax on pesticides to raise funding for biotechnology research. It was also noted that increased funding is needed for research projects on banana virus, control, and elimination of viruses while exchanging Musa germplasm.

Training There appeared to be general recognition of the need to intensify training efforts as they relate to the application of biotechnology to the improvement of banana and plantain. This might be done through the creation of an international training center that would undertake responsibilities for preparing teaching programs and materials, training the trainers, and offering regional and inter-regional training courses. There will be a need to target the trainees. That might include organized training of experts in local laboratories. The establishment of biotechnology degree programs in producing-country universities was also suggested. This might have the effect of stimulating Ph.D. research on banana biotechnology in national programs. Other responsibilities of a training effort might include refresher training programs and follow-ups to training that would trouble-shoot related issues. Consideration should be given to creating a banana/plantain collection for training purposes and the organization of a training package for instrument maintenance.

Develop laboratory infrastructure Numerous suggestions were made to develop laboratory infrastructure in a broad sense. Such infrastructure development could include the establishment of reference laboratories and collections that would support the application of biotechnology to banana and plantain improvement. Infrastructure development might also include the elaboration of simple methods, nonhazardous techniques, and low-cost technology that could be used in producing-country research laboratories. Additionally, providing laboratory safety procedures and advice on relevant biotechnology instrumentation could be useful. Moreover, standardized techniques would be useful for interlaboratory comparisons of results. 238 Technology Transfer

It was proposed that an improved nursery infrastructure would facilitate research activities, as would improved quarantine facilities. It was proposed that producing-country quarantine regulations should be evaluated for their impact on the importation of biotechnology research supplies. Access to biological materials could be increased through facilitated exchanges of materials such as DNA, plasmids, cDNA, YAC libraries, and through the establishment of collections of Musa pests and pathogens.

Bi-directional scientiﬁc exchange The participants recommended that attention be given to facilitating bi-directional scientist exchange in the application of biotechnology to banana and plantain improvement. This could help create mutual projects with joint responsibilities, and thus strengthen research activities. The dimensions of such exchanges should include visits among producing-country scientists, as well as producing-country scientist visits to advanced laboratories. This could be enhanced if fellowships were made available for producing-country scientists to support these exchanges. Additionally, there is a need for an increase in visits by developed-country scientists to producing-country research laboratories. This could be assisted by adopting improved communications between overseas experts and producing-country scientists. Scientiﬁc exchanges could also be encouraged by holding periodic biotechnology exhibitions or science fairs in producing countries. Finally, to support partnerships, producing-country scientists could be invited to serve as editors of biotechnology journals.

Research networking (items 5, 10, 11) Several dimensions were identified by the participants for networking research programs. These included the linking of: national programs to IARCs; producing-country laboratories to biotechnology laboratories in developed countries; and biotechnology laboratories to breeding programs. On this last point, it was noted that there seemed to be a need to develop the acceptance of biotechnology materials and products by breeders and agronomists to facilitate adoption of this technology. To realize the potential of the biotechnologies for the improvement of productivity and stability of banana and plantain, the new technologies should be integrated into plant breeding programs, and satisfy the goals and needs of breeders. Biotechnology research in institutes of regional relevance needs to be strengthened. This might be done by holding small international meetings in producing countries and by developing interdisciplinary groups to address speciﬁc topics. There is an apparent need to encourage collaboration between university research scientists and agricultural ministries in producing countries. The establishment of a tropical biotechnology network in developed-country research laboratories could be of assistance in facilitating the application of biotechnology to Working Group 3 239

banana and plantain improvement. Encouraging research laboratories to work with producing-country research laboratories on tropical crop problems would also be of beneﬁt. In this respect, there may be a need to support these linkages through model negotiations (e.g., memoranda of understanding) for producing-country laboratories to engage in work with developed-country research laboratories. In addition, some consideration should be given to developing common banana and plantain biotechnology research programs between the European Community and the USA. Other combinations of developed-country laboratories might also be considered. Efforts should given to promoting the role of women in the application of biotechnology to banana and plantain improvement. Efforts should be given to discourage brain drains (i.e., the loss of talented scientists from producing countries).

Provision of documents The application of biotechnology to banana and plantain improvement could be facilitated through research support activities that would distribute certain kinds of information as documents. For instance, there is a need to catalogue and disseminate information on existing biotechnology and the impact of those applications vis-à-vis banana and plantain improvement. Information on biotechnology and the identiﬁcation of needs of producing countries should be developed and disseminated widely. The publication of biotechnology manuals in the languages of user countries was seen as a desirable activity. The results of biotechnology research pertaining to banana and plantain improvement should be published in producing-country journals. In addition, publications that simplify the application of biotechnology in producing countries, and information on the progress of large-scale production of germplasm evolved by biotechnology, should be produced. These might take the form of newsletters or some other type of broad information distribution. Efforts should be made to document the genetic diversity of the Musa gene pool available in the producing countries, to evaluate their important characteristics, and to disseminate existing information on collections of Musa so that this information could be made available to others. The creation of databases to support biotechnology research should be explored.

Identify targets with deﬁned value There is a need to set research priorities correctly. This would assist in targeting research that would address clearly stated objectives as to what is to be transferred as technology and who would benefit. There is a need to evaluate objectively the expectations of biotechnological research and its relevance to national goals. It was proposed that reviews of biotechnology programs in producer countries be conducted and that encouragement be given to conducting problem-oriented research. Encouragement should also be given to linking biotechnology to the development of 240 Technology Transfer

sustainable agricultural production. This was seen as linking “ecologically sound” practises through biotechnology research. Encouragement should be given to applying biotechnology to local nonexport banana and plantain varieties to improve their productivity and quality, for local and export markets. Some consideration should be given to ensuring that technologies to be transferred are realistic to the needs of producer countries, and they should be verified through independent, expert appraisal. It was proposed that technology transfer projects be peer-reviewed, especially to ensure relevance for eventual adoption. Research on impact assessment (perhaps through cost-benefit analysis) of current banana/plantain research could be useful. Additionally, fact-ﬁnding missions focusing on biotechnology-relevant problems should be encouraged. These might include a survey of problems in producer countries, and the convening of regional/national workshops for identifying research needs. There was a recognized need for sensitivity to, and understanding of, the limitations of biotechnology transfer for producing countries. This should include an understanding of the socioeconomic constraints of the end-user, and perhaps an agreement of donors supporting banana and plantain research to be more receptive to producing-country decisions on what is wanted. Documentation of the costs arising from a failure to increase Musa productivity should help to establish the need for the application of biotechnology to banana and plantain improvement. Moreover, clear assessment of the losses that occur from field infestations should provide further evidence. It was proposed that the successful completion of a few biotechnology projects could be helpful in establishing expectations and acceptance. Successful experiences could be advertised to promote broader acceptance of this new technology. Contract research should be undertaken on speciﬁc objectives related to the transfer of biotechnology for banana and plantain improvement. Some examples of such targeted research might be: - Development of transformation methods. - Postharvest preservation and storage. -Improved pest control. -Mechanisms of resistance. -Diagnostic kits. - Bioprocessing technology. -Biomass conversion. - Alternative uses for plant products. - Research-enabling technologies. -Value-added technologies. It was also proposed that an inventory be made of unsuccessful technology transfer projects to learn from these experiences. Donors should be advised that technologies other than biotechnology may sometimes be more appropriate. A policy statement might be issued that would caution against accepting “biotechnology at any price”. Working Group 3 241

Finally, it was proposed that biotechnology products must not be imposed on producing countries and that some organization should assume responsibility for monitoring activities to make sure that this does not happen.

Appoint INIBAP biotechnology coordinator It was proposed that INIBAP appoint a biotechnology coordinator. Responsibilities of this proposed position might include the coordination of activities to improve research efficiency and to stimulate greater support for biotechnology research from funding agencies. These responsibilities might be undertaken through the creation of an international advisory center to support the network.

Transfer of technology In this section, consideration was given to activities that would facilitate the transfer of technology to the user level (as opposed to transferring technology between research laboratories). Efforts should be made to encourage private enterprise in banana- and plantain- producing countries to engage in Musa biotechnology. This might be done through the stimulation of venture capital and the fostering of long-term support for biotechnology from local industries. By involving local commercial entities in the transfer process, adoption could be encouraged as the technology emerges. Markets for the products of biotechnology should also be developed and these could include the industrial sector. Additionally, by strengthening relationships between the public and private sectors, technology transfer could be enhanced. This might entail subsidies for biotechnology products, especially during early market development. Government support for research and development of biotechnology products may also need to be considered. Certain support activities could also help the transfer of technology. By providing legal expertise and access to information, issues of intellectual property rights could be resolved and thus foster technology transfer. Encouragement of open markets, free from protectionism, could also foster the transfer of technology. To encourage the application of biotechnology to banana and plantain improvement, some consideration should be given to the equitable distribution of royalties that represent shared contributions and investments. This might include recognition of proprietary know-how in producing countries and sensitivity to the fact that bananas and plantains are staple food crops in some areas. Another area of support that could facilitate the transfer of technologies to commercial products would be programs to ensure environmental safety of biotechnology products. Biosafety regulations should be based on rational science and some thought should be given to providing assistance to national programs in need of developing biosafety regulations and/or guidelines. In a related way, regulatory delays in developed countries (e.g., USA) could send the wrong signals to producing countries and thus cause delays in deploying biotechnology there. 242 Technology Transfer

A need was recognized to develop better distribution systems to make the products of biotechnology available to the end-user. Such a system should have provisions for the rapid dissemination of biotechnology products, and mechanisms for supporting these products that might be aided through funding from donor countries to help develop such new markets. It was proposed that appropriate training be given to agricultural extension workers on the usefulness of successful biotechnology products, for use by interested growers, and that the products of biotechnology be demonstrated in the field to show success and encourage adoption. Local markets should be prepared in advance for the products of biotechnology to enhance acceptance. This might be supported visually by the preparation of a logo for biotechnology products, and the use of slogans such as “Buy Biotech”. Additionally, farmers from producing countries should be encouraged to become familiar with the expected products of biotechnology research. This should include education about not only the advantages, but any attendant risks, of this new technology, and might be done through local grower groups and the exchange of scientists and farmers between producing countries and nonproducing countries. It could also include television interviews with farmers and politicians on subjects pertaining to biotechnology research applications for banana and plantain improvement. Nonproducing countries should be encouraged to eat more and/or different types of bananas to help expand markets and thus create opportunities for the application of biotechnology to banana and plantain improvement. Activities relating to policy, research administration, and public information There is an apparent need for an educational program for decisionmakers, research administrators, and donors. The educational program should include the benefits, as well as the problems, of biotechnology. This activity might be supported through the use of a donor consultant group on biotechnology. There is also a need to educate the developed world population on the food value of bananas in many producing regions. This would help to educate the electorate in donor countries, thus providing motivation and increased political will on the donor side. To address this, it was proposed that donors be encouraged to hire biotechnology experts and thus improve access to information. Such action might be supplemented by lobbying donors for research program needs, and perhaps assisting in the placement of pro- biotechnology people in “high places”. As a corollary activity, it was proposed that effort be given to encouraging political motivation for biotechnology research for banana and plantain improvement in producing countries. Some specific recommendations included the suggestion that international aid agencies purge politics from their decisionmaking process and work to avoid administrative bureaucracy. It was proposed that the applications of biotechnology to benefit banana- and plantain- producing countries be placed on the UNCED agenda and that draft legislation be prepared that would be helpful to producing countries establishing a biotechnology policy. Working Group 3 243

It was proposed that nongovernmental organizations (NGOs) be identiﬁed to explain that biotechnology is a “good thing”. Also, consumers could be targeted through advertising to explain the environmental advantages of biotechnology. This would help to educate the biotechnology users. A suggestion was made to develop a biotechnology vocabulary less frightening to the public and to begin the process of consumer acceptance by educating consumers on the products and beneﬁts likely to be derived from biotechnology research applied to banana and plantain. This could be supported with the previously proposed slogan “Buy Biotech”. Films and videos could be developed to help provide public information on this subject. Such an effort could address the “fear” of biotechnology and might also consider utilizing a well-known personality to act as a global spokesperson for this technology.

Conclusions A rich assortment of suggestions were developed to provide guidance on how to facilitate the transfer of biotechnology for the genetic improvement of banana and plantain. The organization of the material in this report was intended to report faithfully the intention of the Group participants, and any misrepresentation should be considered unintentional and solely the responsibility of the facilitators. Given the enthusiastic participation of Working Group 3, it should be reasonable to conclude that there is both a recognized need and strong support for deliberate activities targeted at many levels, from the farmer to the scientist, the politician to the general public, to facilitate the transfer of biotechnology, not only to research laboratories in producing countries, but to foster global exchange for its eventual application to improve production of banana and plantain.

Reference DELBECQ AC, VONDEN VEN AH, GUSTAFSON DH. 1975. Group Techniques for Program Planning: a guide to nominal group and delphi processes. Glenview, IL, USA: Scott, Foreman and Co. Annexes Acronyms and Abbreviations 247

Annex 1: Acronyms and Abbreviations

AAcc cultivated Musa acuminata diploid AAw wild Musa acuminata diploid ABSP Agricultural Biotechnology for Sustainable Development (USAID project) ACIAR Australian Centre for International Agricultural Research AID Agency for International Development (USA) AP-PCR arbitrarily primed PCR ASPNET Asian and South Paciﬁc Network (INIBAP) BAP benzylaminopurine BBTV banana bunchy top virus BM basal medium CATIE Centro Agronómico Tropical de Investigación y Enseñanza (Costa Rica) cDNA complementary DNA CE crude extract CEC Commission of the European Communities CGIAR Consultative Group on International Agricultural Research CGIS Central Germplasm Information System (INIBAP) CIRAD Centre de coopération internationale en recherche agronomique pour le développement (France) CNPMF Centro Nacional de Pesquisa de Mandioca e Fruticultura Tropical (Brazil) CP coat protein CPMR coat protein-mediated resistance CRSP Collaborative Research Support Program (USA) CTA Technical Centre for Agricultural and Rural Cooperation (the Netherlands) CTAB cetyl-trimethylammonium bromide cv, cvs cultivar, cultivars DMSO dimethylsulphoxide DNA deoxyribonucleic acid dsDNA double-stranded DNA dsRNA double-stranded RNA ECS embryogenic cell suspension EEC European Economic Community EIA enzyme immunoassay EM electron microscopy EMBRAPA Empresa Brasiliera de Pesquisa Agropecuária EMS ethylmethane sulphonate FHIA Fundación Hondureña de Investigación Agrícola FOC Fusarium oxysporum f.sp. cubense FWI French West Indies GNP gross national product 248 Acronyms and Abbreviations

IAEA International Atomic Energy Authority (Austria) IARC international agricultural research center IARPCB International Association of Research on Plantains and Cooking Bananas IBPGR International Board for Plant Genetic Resources (Italy) ICTA International College of Tropical Agriculture (Trinidad) IEF isoelectrofocusing IEM immunoelectron microscopy IITA International Institute of Tropical Agriculture (Nigeria) IMTP International Musa Testing Program (INIBAP) INIBAP International Network for the Improvement of Banana and Plantain (France) IPM integrated pest management IPR intellectual property rights IRFA Institut de recherches sur les fruits et agrumes, CIRAD (France) kb kilobase KU Leuven Katholieke Universiteit Leuven (Belgium) MF Mycosphaerella fijiensis MGES Musa Germplasm Exchange System (INIBAP) NEP naked-eye polymorphism NRC National Research Council (USA) ORF open reading frame PBIP Plantain and Banana Improvement Program (IITA) PCR polymerase chain reaction PDA potato dextrose agar PRX peroxidase QDPI Queensland Department of Primary Industries (Australia) QTL quantitative trait loci QUT Queensland University of Technology, Australia R&D research and development RAPD random amplified polymorphic DNA RAPiD DNA amplification fingerprinting rDNA ribosomal DNA RFLP restriction fragment length polymorphism RMP reference mapping population RNA ribonucleic acid rRNA ribosomal RNA Sat satellite (e.g., Sat RNA) ssDNA single-stranded DNA ssRNA single-stranded RNA TAC Technical Advisory Committee (of the CGIAR) USAID United States Agency for International Development USDA United States Department of Agriculture UV ultraviolet VLP virus-like particle YAC yeast artificial chromosome Participants 249

Annex 2:

Participants

A Abdelnour Esquivel, CATIE, PO Box Biotechnology (QUT), 2434 Brisbane, 17, 7170 Turrialba, Costa Rica 4001 Queensland, Australia P Acuna Chinchilla, Corporación E De Langhe, INIBAP, Bât.7, Parc Bananera Nacional, PO Box 6504- Scientifique Agropolis, 34397 1000, San José, Costa Rica Montpellier Cedex 5, France O Arias, Agribiotecnología de Costa Rica D de Waele, Plant Genetic Systems NV, S.A., P.O. Box 25-4001, Alajuela, Jozef Plateaustraat 22, 9000 Gent, Costa Rica Belgium CJ Arntzen, Department of M Dufour, CATIE/CIRAD-IRCC, Apartado Biochemistry, Texas A&M University, Postal 11, 7170 Turrialba, Costa Rica College Station, Texas 77843-2128, JV Escalant, CATIE/CIRAD-IRFA, PO USA Box 104, 7170 Turrialba, Costa Rica F Bakry, CIRAD-IRFA, Station de A Fareck, Vitronov Inc., 35 Ballantyne Neufchâteau, 97130 Sainte Marie, Terrace, Dorval, Quebec H9S 3E4, Guadeloupe Canada C Fauquet, ORSTOM Research Scientist, J Barba, Laboplant, Ave 12 de octubre Co-Director ILTAB, Division of Plant 1035/of. 701, Quito, Ecuador Biology, MRC7, 10666 North Torrey L Bernier, Bureau Régional Pines Rd, La Jolla, CA 92037, USA de Coopération Scientifique P Ganashan, Plant Genetic Resources et Technique, French Embassy in Center, Department of Agriculture, Costa Rica, Apartado 10177-1000, PO Box 59, Gannotuwa Peradeniya, Costa Rica Sri Lanka V Broomes, National Agricultural D González de León, CIRAD-BIOTROP, Research Institute, Mon Repos, East PO Box 5035, 34032 Montpellier Coast Demerara, Guyana Cedex 1, France IW Buddenhagen, Agronomy R Haïcour, Université Paris XI Orsay, Department, University of California, Bât. 360, Laboratoire de Davis, CA 95616, USA Morphogenèse végétale JA Chambers, US Agency for expérimentale, 91405 Orsay Cedex, International Development, Bureau France for Research and Development, RB Horsch, Monsanto Company, Mail Office of Agriculture, Zone GG4J, 700 Chesterfield Village Room 406-A SA-18, Washington DC Parkway, St Louis, MO 63146, USA 20523-1809, USA RD Huggan, INIBAP, Bât.7, Parc JL Dale, Queensland University of Scientifique Agropolis, 34397 Technology, Centre for Molecular Montpellier Cedex 5, France 250 Participants

R Jaramillo, INIBAP-LAC CATIE, B Neuenschwander, PO Box 60, Turrialba, Costa Rica, IRCC/CATIE/University of Zurich, c/o PO Box 4824-1000, San José, Apartado 143, CATIE, Turrialba, Costa Rica Costa Rica DR Jones, INIBAP, Bât.7, CL Niblett, University of Florida, Parc Scientiﬁque Agropolis, Plant Pathology Department, 34397 Montpellier Cedex 5, France PO Box 110680, Florida, USA G Kahl, Johann Wolfgang Goethe FJ Novak, IAEA Laboratories, Plant University, Siesmayerstr.70, Breeding Unit, PO Box 100, 6000 Frankfurt/Main, Germany A-2444 Seibersdorf, Austria K Kamate, Université Nationale, BD Oakes, Vitronov Inc., 35 Ballantyne Faculté des Sciences et Techniques Terrace, Dorval, Quebec H9S 3E4, (FAST), 22 BP 582, Abidjan 22, Canada Côte d’Ivoire RO Quezada, CATIE Representative, DK Koumba, Projet CIAM, BP 2183, Fray Cipriano de Utrera, Centro de Libreville, Gabon Los Heroes, PO Box 374-2, Santo SL Kurtz, Twyford Plant Labs. Inc., Domingo, Dominican Republic 15245 Telegraph Rd, Santa Paula, W Parrott, University of Georgia, CA 93060, USA Department of Crop and Soil RDR MacKenzie, US Dept of Agriculture, Sciences, Athens, GA 30602-7272, Cooperative State Research Service, USA Suite 330 Aerospace Building, M Perea-Dallos, National University, 901 D St S.W., Washington DC Biology Department, Science Faculty, 20250-2200, USA PO Box 23227, Bogotá, Colombia AM Mailu, Kenya Agricultural Research RC Ploetz, University of Florida, Tropical Institute, PO Box 57811, Nairobi, Research and Education Center, Kenya 18905 SW 280th St, Homestead, K Matasumoto, CENARGEN/EMBRAPA, FL 33031, USA S.A.I.N. Parque Rural 70770, PO Box P Ramirez Fonseca, CIBCM, Universidad 02372, Brasilia-DF, Brazil de Costa Rica, Apartado 7353-1000, X Mourichon, Lab. Pathologie Végétale, Costa Rica CIRAD-IRFA, BP 5035, O Reuveni, Agricultural Research 34032 Montpellier, France Organization, The Volcani Center, L Muller, PO Box 336-1007, Centro Bet Dagan, PO Box 6, Israel Colon, San José, Costa Rica JG Rivas Ducca, PO Box 3597, LC Navarro Mastache, Centro de 1000 San José, Costa Rica Investigación Cientíﬁca de Yucatán H Rodriguez, CATIE, PO Box 127, A.C., Apartado 87 Cordemex, Turrialba, Costa Rica 97310 Merida, Yucatán, Mexico F Rosales, FHIA, PO Box 2067, G Ndamage, Institut des Sciences San Pedro Sula, Honduras Agronomiques du Rwanda, P Rowe, FHIA, PO Box 2067, San Pedro PO Box 138, Butare, Rwanda Sula, Honduras Participants 251

P Rubaihayo, Makerere University, FA Taha, Agriculture College, Cairo Department of Crop Science, University, 23 El-Etehad Square, PO Box 7062, Kampala, Uganda Maadi, Cairo, Egypt D Shaw-Wisdom, Scientiﬁc Research C Teisson, CIRAD, BP 5035, Council, PO Box 350, Kingston 6, 34032 Montpellier Cedex 1, France Jamaica H Tezenas du Montcel, INIBAP, Bât.7, M Smith, Queensland Department of Parc Scientiﬁque Agropolis, Primary Industries (QDPI), 34397 Montpellier Cedex 5, France Maroochy Horticultural Research N Von Mende, Rothamsted Experimental Station, PO Box 5083, SCMC Station, Harpenden, Herts, Nambour, Queensland 4560, AL5 2JQ, UK Australia D Vuylsteke, International Institute of MR Söndahl, DNA Plant Technology Tropical Agriculture, Oyo Road, PMB Corporation (DNAP), 2611 Branch 5320, Ibadan, Nigeria Pike, Cinnaminson, NF 08077, USA LA Withers, IBPGR, Via delle Sette G Strobel, Montana State University Chiese 142, 00145 Rome, Italy College of Agriculture, Dept of Plant K Wright-Platais, CGIAR Secretariat, Pathology, Bozeman, MT 59717-0314, The World Bank, 1818 H Street N.W., USA Washington, DC 20433, USA C Suarez Capello, FUNDAGRO, PO Box AB Zamora, University of the 4076, Guayaquil, Moreno Bellido 127 Philippines, Los Baños, PO Box 3720, y Ave. Amazonas, Quito, Ecuador College, Laguna, Philippines R Swennen, Katholieke Universiteit S Zok, Institut de la Recherche Leuven, Laboratory of Tropical Crop Agronomique, Tissue Culture Husbandry, Kardinaal Mercierlaan Laboratory, IRA Ekona, PO Box 25, 92, B-3001 Heverlee, Belgium Buea, Cameroon Layout and production: Louma productions, Montpellier, France Printing: Imprimerie Joubert, Nice, France