JL SLth ~ .«(■

D o c t o r a l T h e s O- 2 0 0 ^ ^ ,

F a c u l t y o f L a n d s c a p e P ] HOR^MdL^tTRE _ rr//°>: k >u O - e * and A gricultural Sciengb + r/„ T

Genetic Diversity and Oil Quality of Guizotia Cass. (Asteraceae)

Y o h a n n e s P e t r o s f l i p L*^ t 0 O t3 tf I

Genetic Diversity and Oil Quality of Guizotia Cass. (Asteraceae)

Yohannes Petros Faculty of Landscape Planning, Horticulture and Agricultural Science Department of Plant Breeding and Biotechnology Alnarp

Doctoral Thesis Swedish University of Agricultural Sciences Alnarp 2008 Acta Universitatis agriculturae Sueciae 2008:85

Cover: Inflorescences of niger

ISSN 1652-6880 ISBN 978-91-86195-18-2 © 2008 Yohannes Petros, Alnarp Tryck: SLU Service/Repro, Alnarp 2008 Petros, Y. 2008. Genetic Diversity and Oil Quality of Guizotia Cass. (Asteraceae). Doctoral Dissertation. ISSN 1652-6880, ISBN 978-91-86195-18-2

Genetic diversity of G. abyssinica (L.f) Cass., G. scabra (Vis.) Chiov. ssp. scabra; G. scabra (Vis.) Chiov. ssp. schimperi (Sch. Bip.) Baagoe; G. villosa Sch. Bip., G. zavattarii Lanza and G. arborescens (I. Friis) collected from Ethiopia were studied using Inter Simple Sequence Repeat (ISSR) markers. Higher genetic diversity was revealed among the individuals belonging to the same population than among the populations of the different regions. Overall, greater variation was observed between the niger populations originating from Wollo and Hararghe on the one hand and those from the rest of the regions on the other. Among the wild G uizotia species, G. scabra ssp. schimperi was found to be closer to G. villosa than to any of the wild G uizotia taxa. Likewise, G. zavattarii and G. arborescens are found to be more closely related to each other than to the rest of the wild taxa. Based on the ISSR results, revision of the previous classification that placed G. scabra ssp. schimperi as a sub species of G. scabra was suggested. Both the field evaluation of agronomic characters as well as the ISSR analysis revealed variation among the niger populations grown in different regions of the country. Based on the agronomic characters, it was observed that the niger populations obtained from Wollo and Hararghe are of the early maturing types while the accessions originating from the rest of the regions are mostly of the late maturing types. The early maturing and the late maturing niger types differ in many of their agronomic characters notable among which are days to flower initiation, days'to 50% flowering, plant height and seed size. A niger breeding experiment was undertaken in an attempt to elevate the oleic acid content in the seed oil. The objective to increase the oleic acid content in niger seed oil from what it is today, approximately 5-11% in the “wild type” niger of Ethiopian origin to over 80% in the strains improved for oleic acid content has been achieved. The increase in the oleic acid content of the seeds has been gradual. Niger strains that are true breeding for high oleic acid content of over 80% were obtained after three rounds of selection and breeding.

Keywords:Africa, Ethiopia, Fatty acid, Genetic diversity, Guizotia, ISSR, niger, oleic acid

Author’s address: Yohannes Petros, Department of Plant Breeding and Biotechnology, Swedish University of Agrcultural Sciences (SLU),P.O.Box 101, SE-230 53 Alnarp, Sweden. E-mail: [email protected] Dedication

To A chatholic nun, the late Sister Helen My uncle, the late Mikael Bartholome and My mother, the late Yowanna Bartholome.

”Take heed that ye despise not one of these little ones; for their angels do always behold the face of my father which is in heaven” M at 18:10 Contents

List of Publications 7

1 Introduction 9 1.1 Background History 9 1.2 Botanical description 10 1.3 Geographical distribution 10 1.4 Cytology 11 1.5 Domestication of niger 1 2 1.6 Economic importance of niger 13 1.7 Agronomic considerations 14 1.8 Oil content of niger 15 1.9 Genetic diversity 1 7 1.10 Breeding and biotechnology 1 8 2 Objectives of the study 20 3 Materials and methods 21 3.1 The plant material 21 3.2 ISSR diversity 22 3.2.1 DNA extraction and amplification 22 3.3 Field trial 22 3.4 Fatty acid profile 23 3.5 Data analysis 23 4 Results and discussion 25 4.1 Genetic diversity of the Guizotia species 25 4.1.1 Genetic distance and gene flow 26 4.2 Phenotypic diversity of niger 30 4.3 Comparing the molecular and phenotypic diversity of niger 34 4.4 Breeding for high oleic acid in niger 36 5 Concluding remarks 39 6 References 41 7 Acknowledgement 47 List of Publications

This thesis is based on the work contained in the following papers, referred to by Roman numerals in the text:

I Petros, Y., Merker, A. & Zeleke, H. 2007. Analysis of genetic diversity of Guizotia abyssinica from Ethiopia using inter simple sequence repeat markers. Hereditas 144, 18-24.

II Petros, Y., Merker, A. & Zeleke, H. 2008. Analysis of genetic diversity and relationships of wild Guizotia species from Ethiopia using ISSR markers. Genetic Resources and Crop Evolution 55, 451-458.

III Petros, Y., Zeleke, H. & Merker, A. Quantitative trait variation of Guizotia abyssinica (L.f) Cass. Collected from Ethiopia. Submitted

IV Petros, Y., Carlsson, A.S., Stymne, S., Zeleke, H., Fait, A-S. & Merker, A. Developing high oleic acid Guizotia abyssinica (L.f). Cass, by plant breeding. Plant Breeding Accepted pending revision

Papers I—II are reproduced with the permission of the publishers.

7 1 Introduction

1.1 Background History

The genus Guizotia belongs to the family Asteraceae (Compositae), tribe Heliantheae, sub tribe Coreopsidinae. The taxonomic revision of the genus was done by Baagoe (1974) who reduced the number of species to six, five of which grow in Ethiopia. These are G. abyssinica (L.f) Cass., G. scabra (Vis) Chiov. ssp. scabra; G. scabra (Vis) Chiov. ssp. schimperi (Sch. Bip.) Baagoe; G. Villosa Sch. Bip., G. Zavattarii Lanza; G. Arborescens (I. Friis) and G. Jacksoni (S. Moore) J. Baagoe.

G. abyssinica (niger in English), being the only cultivated member of the tax o n is the most important economically. Niger is also the only species that is found outside Africa (Dagne and Heneen, 1992). To date, there is an unresolved controversy regarding the taxonomic position of G. scabra ssp. schimperi vis-a-vis G. abyssinica. The argument is whether G. scabra ssp. schimperi is more closely related to G. abyssinica than to G. scabra ssp. scabra as classified by Baagoe (1974). To this effect the bulk of the evidence points to the close similarity of G. scabra ssp. schimperi to G. abyssinica than to G. scabra ssp. scabra and the suggestion that G. scabra ssp. schimperi be given a specific status on its own (Murthy et al., 1993; Dagne, 1994a; Geleta et al., 2007a; Petros et al., 2008). Even after the publication of the revision work on the taxonomy of the genus by Baagoe in 1974, some authors continued to refer to the two sub species as G. scabra and G. schimperi (Murthy et al., 1995). The most recent workers supporting the independent specific status of G. scabra ssp. schimperi base their argument on molecular similarity, cytology, crossability and phytogeography as well as on morphological similarities (Dagne, 1994a; M urthy et al., 1993; M urthy et al., 1995; Geleta et al., 2007a; Petros et al., 2008). The controversy regarding the taxonomic status of G. scabra ssp. Schimperi is still continuing and will probably continue until further revision of the taxon that takes into account the cytology, molecular diversity as well as morphological features is undertaken.

9 Triacylglycerols are the major components of the neutral lipid class making up 89.7-91.9% of the neutral lipids in niger seed oil (Ramadan and Morsel, 2002). There are four major fatty acids in niger seed oil. These include two main unsaturated fatty acids, linoleic acid (18:2) and oleic acid (18:1) and two major saturated fatty acids, palmitic acid (16:0) and stearic acid (18:0). The abundance of these major fatty acids as well as some minor fatty acids reported to be present in small or trace amounts varies greatly from one sample to the other. Dagne andjohnsson (1997) reported the abundance of linoleic acid as ranging from 65.7-68%, oleic acid, 5.4-7.5%, palmitic acid 9.6-10% and stearic acid 7.6-8.1% while Ramadan and Morsel 2003c reported 63%, 11%, 17% and 7% for the percentage composition of linoleic, oleic, palmitic and stearic acids respectively in niger seed oil.

Though the variation in the oil content of niger cannot be accounted for by the location or the climatic condition of the area as indicated by Dutta et al. (1994), variation in the fatty acid profile, however, can be attributed to several factors such as the area of origin of the material, the climatic condition of the area and more importantly genetic variation. In this regard, niger samples from India are reported to yield relatively high oleic acid. It is also reported that higher temperatures particularly during the period of flowering and fruit set would favour the production of increased oleic acid levels in sunflower (Rondanini et al., 2003). Thus higher temperatures during the reproductive phase of the plant would tilt the balance of the oleic /linoleic ratio towards the production of higher levels of oleic acid in the seed. The increase in the oleic acid content of the seed in this case is compensated for by a relatively reduced level of linoleic acid. In general cooler temperatures favour the production of linoleic acid where it is reported to rise as high as 85% in niger while higher temperatures have the reverse effect on the production of the acid. Linoleic acid is the predominant fatty acid among the neutral lipids as well as the glycolipid fraction of the oil (Ramadan and Morsel, 2003a, b).

The major phospholipid bound fatty acids in niger are linoleic, palmitic and oleic acid (Ramadan and Morsel, 2003b). Among the polar lipids in niger, phospholipids contain relatively high proportions of palmitic acid, the percentage increase of which is compensated by a relative reduction in the proportion o f linoleic acid (Dutta et al., 1994). Dutta et al (1994) also reported -tocopherol as the predominant tocopherol in niger seed oil as well as sterols of which -sitosterol predominates occurring at 2000pg/g oil. The high tocopherol content of niger seed oil is of great dietary significance because tocopherols act as antioxidants particularly in light of the high

16 proportion of the poly unsaturated fatty acid (18:2) which renders the oil prone to oxidative deterioration. The oxidative stability of niger seed oil containing higher percentage of linoleic acid even than those reported for sunflower and safflower (Gecgel et al., 2007; Martinez et al., 1993), can be attributed to the tocopherol content that serves as antioxidant in the oil. It is the crude niger seed oil that is traditionally consumed in Ethiopia (Getinet and Shamia, 1996). This natural oil contains minor constituents that would be removed from the oil at several stages in the refining process. These minor constituents such as tocopherols, and phospholipids have antioxidant activity (Ramadan and Morsel, 2002). Phytosterols are known to be hypocholestrolemic and the consumption of oils rich in these antioxidants is believed to provide protection against cancer, cerebrovascular and cardiovascular diseases (Ramadan and Morsel, 2002).

Teklewold and Wakjira (2004) studied the pattern and rate of seed dry weight and oil accumulation in two improved varieties, Fogera and Kuyu. They observed that the critical duration for the accumulation of oil in niger seed was between 15 to 35 days after anthesis initiation where the oil accumulation increases from 11.68 to 40.6%. They pointed out that the amount of oil would tend to slightly decrease until harvesting date thereby leading to a reduction in the oil content and the seed dry weight. Thus, timely harvest is recommended to maximize the produce of niger seed oil.

1.9 Genetic Diversity

Molecular studies revealing the genetic diversity of niger are indeed scanty. The few genetic diversity studies so far done using DNA markers reveal that the genetic variability of the Guizotia species is high both within and among populations (Petros, 2007, 2008). Random Amplified Polymorphic DNA (RAPD) (Geleta et al., 2007b) and Amplified Fragment Length Polymorphism (AFLP) (Geleta et al., 2008) also revealed high genetic diversity for G. abyssinica populations from Ethiopia. The genetic variability as a function of percent polymorphic loci as well as the Shannon-Weaver diversity indices was reported to be higher for G. abyssinica than the other wild Guizotia species (Petros et al., 2007, 2008). The within population genetic diversity among all the Guizotia populations is reported to be higher than the among population genetic diversity (Geleta et al., 2007b, c, 2008; Petros et al., 2007, 2008). Geleta et al (2007d) made a molecular phylogenetic analysis of the Guizotia species based on chloroplast DNA

17 sequences and suggested the transfer of the genus from its present sub tribe Coreopsidinae to the sub tribe Milleriinae. They also indicated that the perennial forms within the genus might have been the first to evolve.

1.10 Breeding and Biotechnology

Niger is completely out crossing and highly self incompatible. The self incompatibility of niger is homomorphic and of the sporophytic type (Prasad, 1990; Almaw and Teklewold 1995). In self incompatible forms of niger the pollen fails to germinate normally but instead twists and coils over the stigmatic papillae (Prasad, 1990). Guizotia species can interbreed normally in nature producing FI hybrids and even more so when they happen to grow in close proximity with each other (Dagne, 1994a). The FI interspecific hybrids show quantitative characters that are intermediate between the parents but express the qualitative characters of the male parent (Murthy et al., 1993). Performance of interspecific hybrids is a valuable indicator of the feasibility of transferring valuable agronomic traits from the wild or weedy species to the cultivated one (Murthy et al., 1993).

Adda et al (1994a) produced diploid plants from anther cultures of niger that showed significant variation with regards to some traits of agronomic importance. These were self compatible short plants with large head size. This finding is all the more important as self incompatibility is the main problem on the way of the realization of the full reproductive potential of niger (Adda et al., 1994b). Reduced height coupled with high yield is in fact the most desirable character if the production of niger is to be amenable to mechanization such as the use of combine harvester. Adda et al. (1994b) also developed a protocol for the regeneration of niger from the cotyledons through somatic embryogenesis.

The single most important objective to be considered when undertaking niger improvement programs is to increase the amount of oil per unit area of land (Teklewold and Wakjira, 2004). This objective embodies two facets each of which lead to the same ultimate goal but can be carried out in two different directions. These are either increasing the seed yield by breeding for high seed yielding varieties or developing varieties of niger with high oil content. As it stands today, niger is a low yielding crop whose cultivation is plagued by a number of critical drawbacks. Among these factors that severely limit the realization of the full production potential of the crop are indeterminate growth habit leading to non-synchronous maturity ultimately leading to seed shattering and a severe loss of yield and self incompatibility

18 that, among other things, contributes to the reduction in seed yield. The height of the plant is also an important agronomic attribute that needs thoughtful consideration because at its present tall stature it is very unlikely that mechanized fanning can be practiced in the production of niger. In order to exploit the full potential of the crop, single headed dwarf types need to be developed (Getinet and Sharma, 1996). There is, however, ample opportunity for researchers in the area to develop niger varieties with the desired agronomic qualities because the Ethiopian niger is inherently diverse with a wealth of genetic variability. The Ethiopian germplasm collection contains the three known maturity groups. As there is a host of strikingly contrasting features particularly between the early maturing ‘Bunigne’ and the late maturing ‘Abat’ niger, there appears to be ample resources present for the improvement of the crop by selection and breeding for desirable agronomic qualities from among the niger populations growing in the country. The early and late maturing types of niger differ with respect to a number of important characters such as heads per plant, seeds per head, seed weight, plant height, days to maturity and number of primary branches. The niger populations of Ethiopia, thus, present rich genetic variability to be utilized for the development of niger varieties with ideal agronomic characters.

The other direction that niger improvement program could follow but which also aid to achieve the same ultimate goal is to breed for increased oil content in niger seed. To date, there are few works done on the oil content of niger seeds from Ethiopia (Ramadan and Morsel, 2003a, 2003b; Dagne and Johnsson, 1997; Dutta et al., 1994) and it appears that little or no work is done to improve the oil content in niger seed, at least not to the knowledge of the present author. Various authors reported varied oil contents in niger seed (Ramadan and Morsel, 2003a, 2003b; Dagne and Johnsson, 1997; Dutta et al., 1994). This variation in the oil content of the Ethiopian niger is an encouraging reality as the presence of variability by itself is indicative of the prospect to improving the oil content of the crop.

19 2 Objectives of the study

The genus Guizotia becomes important as one of the species, G. abyssinica is an important oil crop. Studying the genetic diversity of the Guizotia species and in particular of G. abyssinica lays the foundation for future work on the improvement of the crop plant. As all the taxa within the genus are related, some qualities of agronomic importance can be transferred from the wild or weedy species to the cultivated species through biotechnological manipulation. The study on oil quality is indeed crucial as it aids in the understanding of the diversity with regards to the oil content and oil quality pertinent to different populations of niger in Ethiopia. The present study was therefore, undertaken in an attempt to achieve the following objectives:

To reveal the molecular genetic diversity among and within the different populations of niger growing in Ethiopia using inter simple sequence repeat markers.

To elucidate the phenotypic variation with regards to the agronomic performance of niger populations collected from different regions in Ethiopia.

To assess the genetic diversity and relationships among the wild Guizotia species using inter simple sequence repeat markers.

To study the fatty acid profile of niger in an attempt to develop varieties that are true breeding for a rare fatty acid or a common fatty acid in unique proportions in the oil.

To increase the content of oleic acid in niger seed oil by plant breeding.

20 3 Materials and Methods

3.1 The plant material

The plant materials were collected from different regions in Ethiopia. For G. abyssinica, plants on a fanner’s field are regarded as a single population while for the wild Guizotia species, plants growing in the same locality are considered to belong to the same population. Seeds from individual plants were sampled separately in all instances. G. abyssinica was collected twice during the course of the study in 2003 and 2005. Because of uneven distribution, some taxa like G. abyssinica and G. scabra ssp. schimperi are collected from wider area because they have a wide distribution in the country, while others like G. arborescens and G. zavattarii are collected only from one region as they have a very restricted distribution in the country. Each taxon used in the study along with the region of collection is presented in table 1.

Table 1. The Guizotia materials used in the study along with the regions of collection in Ethiopia. (The full name of each region is given on paper II)

Axa Regions o f origin Ha W o Gn Gj Sh Ji W e Ar Ba 11 Si Ka G. abyssinica X XXXX XX X G. scabra ssp. schimperi

XXX XXXX XXXX G. scabra ssp. scabra XX X G. villosa X X G. zavattarii X G. arborescens X scabra ssp scabra exhibiting total genetic diversity of 0.1867 and 0.2117 respectively. In all the Guizotia taxa investigated, the within population genetic diversity (Hs) was found to be higher than the among population genetic diversity (Dst). The within population genetic diversity for all the Guizotia taxa ranged from 0.1391 for G. villosa to 0.3738 for G. abyssinica with G. scabra ssp. schimperi, G. scabra ssp. scabra and G. zavattarii possessing Hs of 0.1534, 0.1576 and 0.1615 respectively. In fact G. abyssinica had one of the lowest among population genetic diversity and the highest within population genetic diversity among the Guizotia species. Thus, as it is generally true for all out crossing species, most of the genetic diversity of the Guizotia species is accounted for by the variability among the individuals of a population than among the populations. The small Gst value amounting to 0.0918 for G. abyssinica is also an evidence of the greater within population variation than among them. The small Gst value for G. abyssinica indicates that there is only a small genetic differentiation among the niger populations growing in Ethiopia. This could be assumed to be the result of the high degree of out crossing resulting in a relatively high gene flow among the different populations of the country.

4.1.1 Genetic Distance and Gene flow

There seems to be a strong correlation between the genetic distances and the geographic distances between the populations of the taxa. The standard genetic distance and the Nei (1972) genetic distance was found to be least between populations from a region than between populations from different regions. Thus for G. abyssinica, the highest genetic distance (0.3261) was between populations of Tiyyo and Koladi from Jimma and Wollo regions respectively, while the lowest distance was between Kombolcha and Gerado both from Wollo region. The standard genetic distance for the regions’ populations of G. abyssinica ranged from 0.0281 between populations of Wollo and Gojam to 0.1148 between niger populations of Hararghe and Jimma. Among the wild Guizotia, the standard genetic distance ranged from 0.1188 (between G. scabra ssp. schimperi and G. villosa) to 0.2740 (between G. scabra ssp. schimperi and G. arborescens). Thus, the genetic distance between populations reflects more or less the level of similarity and the degree of relatedness of populations. In terms of species relatedness, the ISSR analysis results indicate that G. scabra ssp. schimperi is more closely related to G. villosa with a genetic distance of 0.1188 between them than to any other wild Guizotia taxa. Similarly G. zavattarii is more closely related to G. scabra

26 ssp scabra and G. arborescens with a genetic distance of 0.1819 and 0.1716 from G. scabra ssp scabra and G. arborescens respectively.

KomboScha Gerado S*d

Hlfasam Rufael • Tareta Haro Asfachew Kara > Shirka Kotkotuma Az«so Zuria T /H aim anot Yaya Worku Awabei Dacha

Ayno Kane Jeiko Ale Gobessa Damasa Jirata Kobo Soyama Tiyyo

Figure 1. UPGMA clustering pattern of 37 populations of Guizotia abyssinica.

The mean Shannon-Weaver diversity index for G. abyssinica data was 0.8841 while the indices for the wild Guizotia species ranged from 0.5791 (G. arborescens) to 0.7373 (G. scabra ssp. schimperi). The average heterozygosity of the wild Guizotia species ranged from 0.1867 for G. villosa to 0.2410 for G. scabra ssp. schimperi. The high genetic diversity exhibited by the Guizotia taxa can be ascribed to the out crossing mode of their reproduction and the greater capability for the dispersal of their pollen. The Guizotia are pollinated by insects mainly by bees. As indicated by Xiao and Gong (2006), it is mainly the ability for pollen dispersal that leads to higher gene flow.

The amount of gene flow estimated as Nm = 0.5(1 -Gst)/Gst was found to be (0.23716) for G. abyssinica. Among the wild taxa, it ranged from 0.8849 for G. scabra ssp. schimperi to 4.5760 for G. zavattarii with G.

27 scabra ssp. schimperi and G. scabra ssp scabra showing Nm values of 1.5473 and 1.6782 respectively. As the Nm is indicative of the number of migrants per generation (Jian et al. 2004), it indicates that G. abyssinica has a high gene flow in comparison to most of the other Guizotia species though there is some variation with regards to the amount of gene flow among the G. abyssinica populations of the different regions. Among the different populations of Guizotia abyssinica grown in the different regions in Ethiopia, it was observed that the niger populations from Wollo had the highest gene flow with an Nm of 3.8439 and those from Gojam had the least Nm of 1.9461. On the other hand all the wild Guizotia taxa had Nm values lower than that of G. abyssinica except for G. zavattarii. The lowest Nm (0.8849) was recorded for G. scabra ssp schimperi and the highest Nm (4.576) for G. zavattarii. As N m is a measure of gene flow, and therefore, migration of individuals/pollen, it follows that the Nm value for a given species is affected by the geographic distance between populations as well as human interference. The low Nm value for G. scabra ssp. schimperi is a reflection of the countrywide distribution of this taxon and therefore the extent of the geographical distance separating individual populations of the taxon. Likewise, among the populations of G. abyssinica, the low Nm for the populations of Gojam indicates that the sampling areas in Gojam are indeed far apart compared to the collection sites of the niger populations of Wollo.

28 Figure 2. Clustering pattern of 45 populations of wild Guizotia generated by UPGMA cluster analysis

While the ISSR technique was able to clearly discriminate among the various wild Guizotia taxa from Ethiopia, it did not, however, reveal clear cut identification when it comcs to the niger populations from the different niger growing areas in Ethiopia. As observed, the interspecific standard genetic distance of the wild Guizotia taxa ranged from 0.1716 to 0.2740 compared to the inter-regional standard genetic distance for niger populations which only ranges from 0.0281 to 0.1148. As a result,

29 the UPGMA dendrogram based on the standard genetic distances was not able to discriminate among the niger populations growing in different area of the different regions in the country. The clustering pattern indicated on figure 1, seems to have been influenced more by the type of niger grown in the regions than by ecogeographical factors. This is evidenced from the dendrogram where the populations were grouped into two main clusters. The first cluster contains mainly the populations from Wollo and Hararghe and the second cluster containing the populations from the rest of the regions. It is known that fanners in Wollo and Hararghe grow the early maturing ‘Bunigne’ niger in contrast to those in Wellega, Shewa, Gojam and Jimma who mainly grow the late maturing ‘Abat’ niger (Getinet and Shamia, 1996). It is highly probable that the clustering pattern of the niger populations reflect both the varietal as well as the ecological factors. This is further confirmed from the UPGMA dendrogram produced for the regions of origin where two major clusters one containing the eastern and the northern regions and the other containing the rest of the regions is produced.

Among the wild Guizotia species the interspecific genetic distance was sufficiently large to group them into distinctly separate clusters on the UPGMA dendrogram (Figure 2). Based on the genetic distances and the clustering pattern obtained, it is the authors’ opinion that G. scabra ssp schimperi be treated as a separate species, because, it is in fact more closely related to G. villosa than to G. scabra ssp. scabra. This opinion is shared by other authors as well whose argument is based on molecular as well as cytogenetic evidences (Geleta et al., 2007a; Murthy et al., 1993, 1995).

4.2 Phenotypic Diversity of niger (paper III)

Thirty six accessions that were previously used for genetic diversity studies using ISSR were also used for the field evaluation of agronomic characters of niger. The results indicate wide variation among the different accessions with regards to several of the quantitative characters investigated. The variations concern traits of agronomic importance such as days to flower initiation, days to 50% flowering, seed size,

30 number of flower heads per plant, number of seeds per head, yield per plant, yield per plot, plant height and the extent of branching.

It was observed that the thirty six accessions originating from the eight niger growing regions of the country fall into two main groups, the variability being contributed mainly by days to flower initiation, days to 50% flowering, plant height, number heads per plant, number primary branches and yield per plot. The results confirmed the notion that there are two main types of niger growing in Ethiopia that greatly differ in the duration to maturity (Almaw and Teklewold, 1995; Getinet and Sharma, 1996). The present study revealed that the niger populations originating from Hararghe and Wollo have many agronomic characteristics in common that are in stark contrast to those from the other regions of the country. Notable among these characters contributing to the observed variations among the accessions are primarily plant height, days to flower initiation, days to 50% flowering, number of heads per plant and 1000 seed weight.

-wl-l I v\1-2 -sh-2'ha-5 ->vW - wl-5 --ha-1 wl-3 -ha-3 -ha-4 -ha-2 -gn-1 -gn-2 -ar-4 -sh-5 -gn-3 -wg-2 -gM - «j>-3 -wg-5 -ar-1 -ar-2 -wfi-4-gn-4 -sh-4 -jm-2-g.t-3 ■rH -Ri-5 -jm-3 -jm-1 -wg-1 -ar-3 -gj-2 -sh-1-sh-3

Figur 3. UPGMA Clustering Pattern of the niger populations used in the study

Inter-regional variations among the accessions were based on the regional means for each character. Accordingly, the first four principal components accounted for about 91% of the total variation among the

31 regions populations. Eigenvectors from the first, second, third and fourth principal components axes accounted for about 43, 25, 16 and 7% of the total variation respectively. The most important variables of the first principal components axis were days to flower initiation, days to 50% flowering, number of heads per plant, plant height, yield per plot and 1000 seed weight, while in the second principal components axis, number of primary branches, number of secondary branches, number of heads per plant and 1000 seed weight were the most important variables responsible for the variations observed among the accessions.

Days to flower initiation, days to 50% flowering, plant height and the number of primary branches contributed most to the variations observed in the first principal components axis for the niger populations grown at Haramaya university while number of heads per plant, number of primary and secondary branches, yield per plant and 1000 seed weight happened to be the most important variables of the first principal components axis for the populations grown at Hima experimental site.

These characters affect the time taken for the completion of the plant’s life cycle as well as the productivity in terms of seed yield. Thus, all the accessions from Wollo and Hararghe are of the early maturing type while most of the accessions from the other regions, especially those from Gojam, Wellega and Jimma are of the late maturing type. Thus, the Ethiopian niger germplasm consists mainly of these two types (Almaw and Teklewold, 1995), though there is a third type of niger that is frost resistant (Getinet and Shamia, 1996). The ‘Bunigne’ niger grown in Wollo, Hararghe and north east Shewa are early maturing small plants producing more flower heads per individual plants, and with relatively smaller seeds, while the ‘Abat’ types are late maturing tall plants with fewer flower heads and producing relatively larger seeds compared to the early maturing types. There is significant positive correlation of plant height to the duration to 50% flowering. Thus the earliest ‘Bunigne’ accession (sh-1) took 63 days to flower while the latest ‘Abat’ accession (gj-2) took 109 days to flower. The accessions from Wollo and Hararghe had average heights of 130.3cm and 133.8cm respectively while the ‘Abat’ accessions from Gojam and Wellega were found to be taller on the average even than the other ‘Abat’ accessions from the rest of the regions measuring 146.7cm and 150.8cm respectively.

Based on the principal components analysis, scatter plots were produced for the accessions obtained from the different regions of the country. The accessions form two separate clusters, the Wollo and Hararghe accessions being grouped together. The UPGMA dendrogram for the accessions (Figure 3) also depicted a similar pattern producing two main clusters, one of which mainly consists of the accessions from Wollo and Hararghe while the second cluster consists of the accessions from the other regions. Likewise, the UPGMA dendrogram based on the regional means for each character produced two main clusters placing Wollo and Hararghe together in a separate cluster. The second major cluster for the regions consists of two distinct sub clusters where Gonder and Shewa were grouped together in a sub cluster and Gojam, Wellega and Arsi likewise placed in the second sub cluster. Jimma, however, remained solitary in the second sub cluster connected only distantly to the two sub clusters of the second major cluster. This indicates that the niger grown in Jimma, though of the ‘Abat’ type is , however, sufficiently different from the other ‘Abat’ accessions grown in the other regions. It has been observed from table 2 (paper III) that some of the unique features of the Jimma populations separating them from the other accessions might have been the extent of primary and secondary7 branching and the number of seeds per head. The average number of primary and secondary branches of the accessions from Jimma exceeds the number recorded for the other accessions studied. Thus, based on the study, the Jimma populations of niger could be described as late maturing plants that are highly branching and producing fewer seeds per head.

Niger populations growing in Ethiopia harbour rich genetic diversity. Previous investigations using molecular markers have shown that great variation is indicated within the same population growing on a field as well as among different populations growing in different regions of the country (Geleta et al., 2007b, 2008; Petros et al., 2007). The variations existing among the niger populations in the different regions of the country are to a large extent attributed to the inherent variability in important agronomic characters existing between the different strains of niger, whether early maturing or late maturing, and to a lesser extent to ecogeographic factors unique to the various regions of origin. The present study has indicated these important characters differentiating between these two niger strains. These important agronomic traits inherent in the two strains and responsible for producing such striking contrast between them relate to the number of days to flower initiation and 50% flowering both of which determine the duration to maturity. Equally very important traits marking the phenotypic variation between the two types of niger are the number of flower heads per plant, the extent of branching and seed size. As only very little work is done so far to improve the oil and seed yield of niger, it is believed that the future

33 holds encouraging prospects in this regard for the improvement of important agronomic qualities of this crop. The ideal niger plant needs to be early maturing with short and sturdy stems, for which candidates could be sought among the ‘Bunigne’ types. It also needs to have larger seed size producing greater yield in terms of oil and seed for which the ‘Abat’ types could serve as good starting materials.

3 Comparing the molecular and phenotypic diversity of niger (paper I & III)

The phenotypic characterization of niger for agronomic traits and the molecular diversity analysis using ISSR markers produced results that are more or less in harmony, though not totally congruent. Two major clusters with smaller sub clusters are produced in both the molecular analysis and the field evaluation. In both cases, the accessions from Wollo and Hararghe were included in the first major cluster. In fact the first major cluster produced for the phenotypic characters of niger wholly consisted of the accessions from these two regions except for one accession from Shewa (sh-2) which had paired with sh-3 in the ISSR data. It can be observed that the same accessions grouped in the same cluster in the ISSR analysis were placed in two separate clusters in the dendrogram for the field data. Some accessions occurring in the first sub cluster of the second major cluster of the field data were grouped in the first cluster of the ISSR data. The first major cluster of the ISSR data consisted of nineteen accessions all of which occur in the two clusters of the field data except for two populations each from Arsi and Gojam. Conversely, there are twenty one accessions included in the first two clusters of the field data. Two accessions each from Arsi and Wellega and an accession each from Shewa and Gonder were not included in the first major cluster of the UPGMA dendrogram produced for the ISSR data. An accession from north-east Shewa (sh-1) which was the earliest to flower both at the Haramaya and Hirna field experimental sites was grouped together with the Wollo and Hararghe accessions in the dendrogram produced for the ISSR data but remained an outlier in the dendrogram for the field trial. The duration to flowering as well as other important agronomic parameters suggest that this accession indeed belongs with the Wollo and Hararghe accessions. None of the Jimma accessions were grouped in the cluster containing the Wollo and Hararghe accessions in both the dendrograms produced for the ISSR and the field data. Based on the phenotypic trait evaluation of the field data, they were grouped together with the accessions from Gojam. This clustering pattern of the Jimma accessions along with the accessions from Gojam is convincing when one considers that the accessions from these two regions have a number of agronomic features in common notable among which is the duration to flowering. Taking the regions of origin of the materials in retrospect, the molecular analysis and the phenotypic evaluation of agronomic characters more or less agree. The UPGMA dendrograms for the regions in both analyses produced two major clusters. Jimma was an outlier in the dendrogram for the ISSR data and was not grouped with any of the regions in the field data. Wellega was grouped with Gojam in the analysis of the field data but with Arsi based on the ISSR data. All these three regions, Gojam, Wellega and Arsi are known to grow the late maturing types of niger.

The dendrograms for the field and ISSR data are produced using the Euclidean and the standard genetic distances respectively. The variation in the pattern of clustering in the two data analyses might have been the result of these two different methods of analysis as well as the different data generated from these two completely isolated works. What can be asserted from these two analyses results is that the field evaluation of the phenotypic characteristics of niger and the genetic analysis using ISSR markers revealed slightly different patterns for the diversity of niger populations growing in Ethiopia.

As generally, the revelation of molecular diversity is manifested in phenotypic expression, there can be a host of traits for which one accession may differ from the other. Thus, all the accessions used for the present study do not group together solely on the basis of the region of origin. This indicates that, while eco-geographic feature of their respective regions is a factor, there seems to be other factors, namely, material transfer between neighbouring regions or difference in the type of strains grown in different regions that play roles in assigning the accessions to the different clusters. It is also evident that the late maturing types of niger are more varied compared to the early maturing ones.

35 4.4 Breeding for high oleic acid in niger (paper IV)

The oleic acid content in the nine materials selected for breeding ranged from 17-30.6%. This was increased to about 35.2% on the average after the first round of breeding. It became clear, however, that the oleic acid content from a single cross and indeed even from the same flower head were not uniform in the percent oleic acid content but showed great variation. Although the difference in the oleic acid content between the lowest and the highest in the starting material was only about 14%, this variation, however, increased in the progeny seeds to about 79% after the first harvest. Whereas no seed used as the starting material for the breeding experiment had oleic acid content above 31%, it was observed that after the first round of breeding a number of progeny seeds had percent oleic acid higher than 50% with one seed’s oleic content equalling 83%. The increase in the oleic acid content of progeny seeds continued producing more than 20% of the seeds with oleic acid content of greater than 80% after the second round of breeding with an average of about 53.2%. Even then, there was great variation in the percent oleic acid content among the individual seeds. Only those seeds with oleic acid content of more than 80% were selected for the third ro u n d of breeding, after which almost all progeny seeds had oleic acid content of 80% or more

The study showed the possibility of increasing the oleic acid content in niger seed oil by repeated selection and breeding. In the present study it was made possible to increase the oleic acid in niger seed oil from 1 approximately 8.4% in the ordinary niger materials from Ethiopia to over 80% after selection and breeding in a controlled environment at Alnarp (Fig. 4). It was also observed that plants whose oleic acid content is approximately > 79% are true breeding for the high oleic trait 60000

50000-

40CiCK)

30000-

20000-

tO 10 mfn win Figure 4. Fatty acid profile of niger. A. ordinary niger material. B. high oleic acid yielding niger.

The mode of inheritance of the high oleic trait in niger or the number of alleles involved in determining the oleic/linoleic ratio is not yet known. It is reported that in the sunflower mutant ‘Pervenets’ the interaction of three alleles control the oleic acid content in the seeds (Martinez et al 1989). Schuppert et al (2006) found out that the increase in the sunflower variety Pervenets is partly caused by a mutant allele ‘OF exerting incomplete dominance. Thus it is possible that more than one allele may be involved in the production of the high oleic acid content in niger. The number of genes and alleles involved in the expression of the high oleic trait in niger is presently the subject of investigation at SLU, Alnarp.

37 The changes in the biosynthetic pathway resulting in the accumulation of oleic acid in niger seed also needs to be investigated. As the biosynthetic pathway for the production of oleic and linoleic acids entail series of desaturation steps leading from 18:0 to 18:1 and from 18:1 to 18:2, it is envisaged that the phenomenon leading to the accumulation of oleic acid in niger may be due to either of the following reasons. 1. increased activity of 9 desaturase enzyme, the result of which is increased production of oleic acid from stearic acid (18:1). 2. Reduction or loss of activity of the 12 desaturase enzyme that utilizes oleic acid in the production of linoleic acid. The net effect of these activities is the preponderance of triolein as the predominant triacylglycerol (TAG) molecular species. 5 Concluding remarks

Niger populations growing in Ethiopia harbour rich genetic diversity. There is great variation among the populations growing in different regions and localities as well as among individuals of the same population. This variability could be made into good use by breeding for niger plants with the desired agronomic characters. The niger ideotype should be an early maturing short plant with strong stems whose flowering and maturity are synchronized. It is imperative that the plant produces superior yield in terms of seed or oil.

The early maturing and late maturing niger plants differ with regards to several traits such as the span of their life cycle, height and seed size. The late maturing niger types are also known to have higher oil content and seed yield than the early maturing ones. In fact, there is great variation even among the late maturing niger types themselves with respect to many characters of agronomic importance. Future attempts to improve the seed or oil yield of niger need to exploit this diversity to develop the ideal niger variety with the desired agronomic qualities that are well suited to the environmental conditions of the areas of cultivation.

Niger strains with exceptionally high oleic acid have been developed in the present study. There could as well be a host of other valuable traits waiting yet to be explored as the plant is very little studied. Future research trend in this regard should aim at the development of hybrid varieties that are high yielding both in terms of oil content and oleic acid content.

High genetic diversity was observed in the wild Guizotia species both within and among the populations of the respective species. As reported by several workers, niger shares more similarity with G. scabra ssp. schimperi and G. villosa than with the other members of the taxa. As these taxa are known to form inter-specific hybrids with relative ease, there is a good likelihood of transferring important agronomic qualities from the wild species to the cultivated one.

Some of the Guizotia species such as G. scabra ssp. schimperi are widely distributed in the country and are not facing any threat for their survival, at least not in the immediate future. Others like G. arborescens on the other hand, are limited to a small area in the country and are at a greater

39 risk due to human activities in the area, hence the urgent need for conservation and preservation of the germplasm.

40 6 References

Adda, S., Reddy, T.P. & Kishor, P.B.K. 1994a. Androclonal variation in niger (Guizotia abyssinica Cass.). Euphytica 79, 59-64. Adda, S., Reddy, T.P. & Kishor, P.B.K. 1994b. Somatic embryogenesis and organogenesis in Guizotia abyssinica. In Vitro Cellular and Developmental Biology-Plant 30, 104-107. Ajebade, S.R., Weeden, N.F. & Chite, S.M. 2000. Inter simple sequence analysis of genetic relationships in the genus Vigna. Euphytica 111, 47-55. Almaw, G. & Teklewold, A. 1995. An agronomic and seed quality evaluation of noug (Guizotia abyssinica Cass.) germplasm in Ethiopia. Plant Breeding 114, 375-376. Ammiraju, J.S., Dahlakia, B.B. & Santra, D.K. 2001. Identification of inter simple sequence repeat (ISSR) markers associated with seed size in wheat. Theoretical and Applied Genetics 102, 726-732. Assefa, K., Merker, A. & Teffera, H. 2002. Qualitative trait variation in tef [Eragrostis tef (Zucc.) Trotter]. Euphytica 127, 399-410. Assefa, K., Merker, A. & Tefera, H. 2003a. Multivariate analysis of diversity of tef [Eragrostis tef (Zucc.) Trotter] germplasm from western and southern Ethiopia. Hereditas 138, 228- 236. Assefa, K., Merker, A. & Tefera, H. 2003b. Inter simple sequence repeat (ISSR) analysis of genetic diversity in tef [Eragrostis tef (Zucc.) Trotter], Hereditas 139, 174-183. Ayana, A. & Bekele, E. 1999. Multivariate analysis of morphological variation in Sorghum (Sorghum bicolor (L.) Moench) germplasm from Ethiopia and Eritrea. Genetic Resources and Crop Evolution. 46, 273-284. Baagoe, J. 1974. The Genus Guizotia (Compositae): A taxonomic revision. Botanisk Tidsskrift 69, 1-39. Bekele, E., Geleta, M., Dagne, K., Jones, A.L., Barnes, I., Bradman, N. & Thomas, M.G. 2007. Molecular phylogeny of genus Guizotia (Asteraceae) using DNA sequences derived from ITS. Genetic Resource and Crop Evolution 54, 1419-1427. Bomet, B. & Branchard, M. 2001. Non anchored inter simple sequence repeat (ISSR) markers: reproducible and specific tools for genome fingerprinting. Plant Molecular Biology Reports 19, 209-215. Burton, J.W., Wilson R.F. & Brim C.A. 1983. Recurrent selection in Soybeans. IV. Selection for increased oleic acid percentage in seed oil. Crop Science. 23, 744-747. Chapman, K.D., Austin-Brown, S., Sparace, S.A., Kinney, A.J., Ripp, K.G., Pirtle, I.L. & Pirtle R.M . 2001. Transgenic cotton plants with increased seed oleic acid content. Journal of the American Oil Chemists Society 78, 941-947. Dagne, K. 1994a. Meiosis in interspecific hybrids and genomic interrelationships in Guizotia Cass. (Compositae). Hereditas 121, 119-129. Dagne, K. 1994b. Cytology, phylogeny and oil quality of Guizotia Cass. (Compositae). Ph.D. thesis, Addis Abeba university, Ethiopia.

41 Dagne, K. 1994c. Mitotic and meiotic behaviour of B-chromosome of Guizotia scabra (Vis.) Chiov. ssp. scabra (Compositae). Caryologia 48, 35-45. Dagne, K. 1995. Karyotypes, C-banding and nucleolar numbers in Guizotia (Compositae). Plant Systematics and Evolution 195, 121-135. Dagne, K. 2001. Cytogenetics of new Guizotia Cass. (Compositae), interspecific hybrids pertaining to genomic and phylogenetic affinities. Plant Systematics and Evolution 230, 1-

11. Dagne, K. & Heneen, W.K. 1992. The karyotypes and nucleoli of Guizotia abyssinica (Compositae). Hereditas 117, 73-83. Dagne, K. andjonsson, A. 1997. Oil content and fatty acid composition of seeds of Guizotia Cass. (Compositae). Journal of the Science o f Food and Agriculture 73, 274-278. Dagne, K., Cheng, B. & Heneen, W.K. 2000. Number and sites of rDNA loci of Guizotia abyssinica (L.f) Cass. As determined by florescence in situ hybridization. Hereditas 132, 63- 65. DISPAN (1993) Genetic distance and phylogenetic analysis. Copyright by Tatsuya and the Pennsylvania state university Dutta, P.C., Helmersson, S., Kebedu, E. & Appelqvist, L.A. 1994. Variation in lipid composition of niger seed (Guizotia abyssinica Cass.) Samples collected from different regions in Ethiopia. Journal o f the American Oil Chemists Society 71, 839-843. Fehr, R.W . 2007. Breeding for modified fatty acid composition in soybean. Crop Science. 47, S-72-S-87. Fernandes-Martinez, J., Jimenez, A., Dominiguez, J., Garcia, J.M., Garces, R. & Mancha, M. 1989. Genetic analysis of the high oleic acid content in cultivated sunflower (Helianthus annus L.). Eupytica, 39-51. Fernandez-Martinez, J., del Rio, M. & de Harp, A. 1993. Survey of safflower (Carthamus tinctorius L.) germplasm for variants in fatty acid composition and other seed characters. Euphytica 69, 115-122. Friis, I. 1971. A new species of Guizotia (Compositae) from north east tropical Africa. Norwegian Journal of Botany 18, 231-234. Fuller, G., Kohler G.O. & Applewhite, T.H. 1996. High oleic safflower oil: A new stable edible oil .Journal of the American Oil Chemists Society 43, 477-481. Gecgel, U., Demirci, M. & Esendal, E. 2007. Fatty acid composition of the oil from developing seeds of different varieties of safflower (Carthamus tinctorius L.) Journal of the American Oil Chemists Society 84, 47-54. Geleta, M., Bryngelsson, T., Bekele, E. & Dagne, K. 2007a. Comparative analysis of genetic relationship and diagnostic markers of several taxa of Guizotia Cass. (Asteraceae) as revealed by AFLPs and RAPDs. Plant Systematics and Evolution 265, 221-239. Geleta, M., Bryngelsson, T., Bekele, E. & Dagne, K. 2007b. Genetic diversity of Guizotia abyssinica (L.f) Cass. (Asteraceae) from Ethiopia as revealed by random amplified polymorphic DNA. Genetic Resources and Crop Evolution 54, 601-614. Geleta, M., Bryngelsson, T., Bekele, E. & Dagne, K. 2007c. AFLP and RAPD analyses of of genetic diversity of wild and/or weedy Guizotia (Asteraceae) from Ethiopia. Elereditas 144, 53-62.

42 Geleta, M. 2007d. Genetic Diversity, Phylogenetics and Molecular Systematic of Guizotia Cass. (Asteraceae). Ph.D thesis. Swedish University of Agricultural Sciences (SLU), Sweden. Geleta, M., Bryngelsson, T., Bekele, E. & Dagne, K. 2008. Assessment of genetic diversity of Guizotia abyssinica (L.f) Cass. (Asteraceae) from Ethiopia using Amplified fragment length polymorphism (AFLP). Plant Genetic Resources 6, 41-51. Genet, T. & Belete, K. 2000. Phenotypic diversity in the Ethiopian noug germplasm. African Crop Science Journal 8, 137-143. Gessesse, A. 1997. The use of nug meal as a low-cost substrate for the production of alkaline protease by the alkaliphilic Bacillus sp. AR-009 and some properties of the enzyme. Bioresource Technology 62, 59-61. Getinet, A. & Sharma, S.M. 1996. Niger. Guizotia abyssinica (L.f) Cass. Promoting the conservation and use o f underutilized and neglected crops. 5. International Genetic Resources Institute. Gatersleben, Germany. Hailu, F., Merker, A., Singh, H., Belay, G. & Johansson, E. 2006. Multivariate analysis of diversity of tetraploid wheat germplasm from Ethiopia. Genetic Resources and Crop Evolution 53, 1089-1098. Hiremath, S.C. and Murthy, H.N. 1986. The structure, stability and meiotic behaviour ofB- chromosoine in Guizotia scabra (Vis.) C hiov. ssp. scabra (Compositae). Caryologia 39, 397- 402. Hiremath, S.C. and Murthy, H.N. 1988. Domestication of niger (Guizotia abyssinica). Euphytica 37, 225-228. Hiremath, S.C. and Murthy, H.N. 1992. Cytogenetical studies in Guizotia (Asteraceae). Caryologia 45, 69-82. Iruela, M., Rubio, J. & Cubera, J.J. 2002. Phylogenetic analysis in the genus Cicer and cultivated chickpea using RAPD and ISSR markers. Theoretical and Applied Genetics 104, 643-651. Jian, S., Tang, T. & Zhong, Y. 2004. Variation in inter simple sequence repeat (ISSR) in mangrove and non mangrove populations of Heritiera littoralis (Sterculiaceae) from China and Australia. Aquatic Botany 79, 75-86. Joshi, S.P., Gupta, V.S. & Agarwal, R.K. 2000. Genetic diversity and phylogenetic relationship as revealed by inter simple sequence repeat (ISSR) polymorphism in the genus Oryza. Theoretical and Applied Genetics 100, 1311-1320. Kandel, H. and Porter, P. (eds) 2002. Niger: Guizotia abyssinica (L.f.) Cass. Production in northwest Minnesota. University of Minnesota extension service. Kandel, H.J., Porter, P.M., Johnson, B. L., Henson, R.A., Hanson, B.K., Weisberg, S. & LeGare, D.G. 2004. Plant population influences niger seed yield in the Northern Great Plains. Crop Science 44, 190-197. Kinney, A.J. 1994. Genetic modification of the storage lipids of plants. Current Opinion in Biotechnology 5, 144-151. Lagercrantz, U., Ellegren, H. & Andersson, L. 1993. The abundance of various polymorphic microsatelite motifs differs between plants and vertebrates. Nucleic Acids Research 21, 1111— 1115.

43 Martin, J.P. & Sanchez-Yelemo, M.D. 2000. Genetic relationships among species of the genus Diplotaxis (Brassicaceae) using inter simple sequence repeat markers. Theoretical and Applied Genetics 101, 1234-1241. Mathur, R. K. and Gupta, S. C. 1992. Discriminate function analysis in niger (Guizotia abyssinica Cass.). Crop Research. 5, 164-165. Montagnon, C. and Bouharmont, P. 1996. Multivariate analysis of phenotypic diversity of Coffea Arabica. Genetic Resources and Crop Evolution. 43, 221-227. Mugendi, J.B., C.A. Sims, D.W . Gorbet, and S.F O ’Keefe 1998. Flavor stability of high oleic peanuts stored at low humidity. Journal of the American Oil Chemists Society 75, 21-25. Murthy, H.N., Hiremath,S.C. & Salimath, S.S. 1993. Origin, evolution and genome differentiation in Guizotia abyssinica and its wild species. Theoretical and Applied Genetics 87, 587-592. Murthy, H.N., Hiremath,S.C. & Pyati, A.N. 1995. Genomic classification in Guizotia (Asteraceae). Cytologia 60, 67-73. Nagaoka, T. & Ogihara, Y. 1997. Applicability of inter simple sequence repeat polymorphisms in wheat for use as DNA markers in comparison to RFLP and RAPD markers. Theoretical and Applied Genetics 94, 597-602. Nayakar, N. Y. 1976. Genetic variability and heritability of six quantitative characters in niger (Guizotia abyssinica Cass.). Mysore Journal of Agriculture Sci 10, 553-558. Nei, M. 1972. Genetic distance between populations. American Naturalist 106, 283-292. Nei, M. 1973. Analysis of gene diversity in subdivided populations. Proceedings of the National Acadamy of Sciences of the United States of America 70, 3321-3323. Okuley, J., Lightner, J., Feldmann, K., Yadav, N., Lark, E. & Browse, J. 1994. Arabidopsis FAD2 gene encodes the enzyme that is essential for polyunsaturated lipid synthesis. The Plant Cell 6, 147-158. Petros, Y. 1988. Anatomical studies o f oil bearing species of Compositae, Vernonia galamensis (Cass.) Less. And Guizotia abyssinica (L.f) Cass. M.Sc thesis Addis Ababa University, Ethiopia. Petros, Y., Merker, A. & Zeleke, H. (2007. Analysis of Genetic Diversity of Guizotia abyssinica (L.f.) Cass, from Ethiopia using Inter Simple Sequence Repeat markers. Hereditas 144, 18-24 Petros, Y., Merker, A. & Zeleke, H. 2008. Analysis of genetic diversity and relationships of wild Guizotia species from Ethiopia using ISSR markers. Genetic Resources and Crop Evolution 55, 451-458. Pradhan, K., Mishra, R.C. & Paikary, R.K. 1995. Genetic variability and character association of niger. Indian Journal of Genetics 55, 457-459. Prasad, V. 1990. Pollen tube growth and site of incompatibility reaction in niger (Guizotia abyssinica Cass.). Current Science 59, 466-468. Prevost, A. & Wilkinson, M.J. 1999. A new system of comparing PCR primers applied to ISSR fingerprinting of potato cultivars. Theoretical and Applied Genetics 98, 107-112. Qian, W., Ge, S. & Hong, D-Y. 2001. Genetic variation within and among populations of a wild rice Oryza granulate from China detected by RAPD and ISSR markers. Theoretical and Applied Genetics 102, 440-449.

44 Rajesh, P.N., Sant, V.J., Gupta, V.S., Muehlbauer, F.J. & Rajekar, P.K. 2002. Genetic relationships among annual and perennial wild species of Cicer using inter simple sequence repeat (ISSR) polymorphism. Euphytica 129, 15-23. Ramachandran, T.K. & Menon, P.M. 1979. Pollination mechanism and inbreeding depression in niger. Madras Agricultural Journal 66, 449-454. Ramadan, M.F. & Morsel, J.T. 2002. Proximate neutral lipid composition of niger. Czech Journal of Food Science 20, 98-104. Ramadan, M.F. & Morsel, J.T. 2003a. Analysis of glycolipids from black cumin (Nigella sativa L.), coriander (Coriandrum sativum L.) and niger (Guizotia abyssinica Cass.) oilseeds. Food Chemistry 80, 197-204. Ramadan, M.F. & Morsel, J.T. 2003b. Phospholipid composition of niger (Guizotia abyssinica Cass.) seed oil. Lebensm.-Wiss. U.-Technol 36, 273-276. Ramadan, M.F. & Morsel, J.T. 2003c. Determination of the lipid classes and fatty acid profile o f niger (Guizotia abyssinica Cass.), seed oil. Phytochemical Analysis 14, 366-370. Reddy, M.P., Sarla, N & Siddiq, E.A. 2002. Inter simple sequence repeat (ISSR) polymorphism and its application in plant breeding. Euphytica 128, 9-17. Riley, K.W. & Belayneh, H. 1989. Niger - In: Oil crops of the world, (eds Robbelen, G., Downey, R.K. & Ashri, A.), McGrow-Hill, New York. pp. 394-403. Rohlf, F.J. 2000. NTSYSpc Numerical Taxonomy and Multivariate Analysis System. Version 2.1. Exeter software, New York, USA. Rondanini, D., Savin, R. & Hall, A.J. 2003. Dynamics of fruit growth and oil quality of sunflower (Helianthus annus L.) exposed to brief intervals of high temperature during grain filling. Field Crops Research 83, 79-90. SAS Institute. 2004. JMP 5.1.1 Users Guide. SAS Institute Inc. Cary, NC, USA. Schnurbusch, T., Mollers, C. & Becker, H.C. 2000. A mutant of Brassica napus w ith increased palmitic acid content. Plant breeding 119, 141-144. Schuppert, G.F., Tang, S., Slabaugh, M.B & Knapp, S.J. 2006. The sunflower high oleic acid mutant Ol carries variable tandem repeats of FAD2-1, a seed specific oleoyl-phosphatidyl choline estrase. Molecular Breeding 17, 241-256. Seegeler, C.J.P. 1983. Oil plants in Ethiopia, their taxonomy and agricultural significance. C enter for Agricultural Publishing and Documentation. Wageningen pp. 122-146. Singh, B. & Patra, G.J. 1989. Character association and path coefficients of quantitative traits in niger (Guizotia abyssinica Cass.) Indian Journal of Agricultural. Science. 59, 442-445. Slatkin, M. & Barton, N.H. 1989. A comparison of three indirect methods for estimating average levels of gene flow. Evolution 43, 1349-1368. Stymne, S. & Appelqvist, L.A. 1980. The biosynthesis oflinoleate and -linolenate in homogenates from developing soya bean cotyledons. Plant Science Letters 17, 287-294. Sudupak, M.A. 2004. Inter and intra species inter simple sequence repeat (ISSR) variations in the genus Cicer. Euphytica 135, 229-238. Takagi, Y. & Rahman, S.M.1996. Inheritance of high oleic acid content in the seed oil of soybean mutant M23. Theoretical and Applied Genetics 92, 179-182. Teklewold, A. and Wakjira, A. 2004. Seed filing and oil accumulation in noug (Guizotia abyssinica (L.f.) Cass. Sinet: Ethiopian Journal of Science. 27(1): 25-32.

45 Teklewold, A. & Alemayehu, N. 2002. Studies on the floral characteristics of niger seed (Guizotia abyssinica (L.f.) Cass.). Journal of Applied Botany 76, 163-167. Tikunov, Y.M., Khrustaleva, L.I. & Karlov, G.I. 2003. Application of ISSR markers in the genus Lycopersicuni. Euphytica 131, 71-80. Tompkins, C. & Perkins E.G. 2000. Frying performance of low linolenic acid soybean oil. Journal of the American Oil Chemists Society 77, 223-229. Urie, A.L. 1985. Inheritance of high oleic acid in sunflower. Crop Science. 25, 986-989. Vanhala, T.K., Van Rijn, C.P.E., Buntjer, J., Stam, P., Nevo, E., Poorter, H. & Van Eeuwijk, F.A. 2004. Environmental, phenotypic and genetic variation of wild barley (Hordeum spontaneum) from Israel. Euphytica 137, 297-309. Vijayan, K. & Chatteijee, S.N. 2003. ISSR profiling of Indian cultivars of mulberry (Morns ssp.) and its relevance to breeding program . Euphytica 131, 53-63. Virk, P.S., Zhu, J. & Newbury, H.J. 2000. Effectiveness of different classes of molecular marker for classifying and revealing variation in rice (Oryza sativa) germplasm. Euphytica 112, 275-284. Voelker, T. & Kinney, A.J. 2001. Variation in the biosynthesis of seed storage lipids. Annual Review of Plant Physiology and Plant Molecular Biology 52, 335-61. Warner, K. & Knowlton S. 1997. Frying quality and oxidative stability of high oleic com oil. Journal of the American Oil Chemists Society 74, 1317-1322. Warner, K., Orr, P. & Glynn M. 1997. Effects of fatty acid composition of oils on flavor and stability of fried foods. Journal of the American Oil Chemists Society 74, 347-356. Wilson, R.F. & Rinne, R.W . 1976. Alteration of soybean oil composition by plant breeding. Journal of the American Oil Chemists Society 53: 595-597. Xiao, L-Q. & Gong, X. 2006. Genetic differentiation and relationships in Cycas balansae complex (Cycadaceae) and its conservation implications. Annals of Botany 97, 807-812. Yao, M.Z., Chen, L. & Liang, R. 2008. Genetic diversity among tea cultivars from China, Japan and Kenya revealed by ISSR markers and its implication for potential selection in tea breeding programmes. Plant Breeding 127, 166-172 Yeh, F.C., Yang, R.C. & Boyl, T. 1999. Popgene version 1.31. Microsoft window based freeware for population genetic analysis. University of Alberta, Edmonton. Zietkiewicz, E., Rafalski, A. & Labuda, D. 1994. Genome finger-printing by simple sequence repeat (SSR)-anchored polymerase chain reaction amplification. Genomics 20, 176-183.

46 Acknowledgements

My heart-felt gratitude to the following persons who were very concerned about my work and deeply involved with me.

Professor Arnulf Merker, my main supervisor for his guidance and advice throughout the course of my studies, for his kindness and responsibility in all of my works. I owe him so much for his concern about my wellbeing as well as for his vital contribution to all the achievements in my studies.

Doctor Anders, S. Carlsson, my co-supervisor from whom I have learned a lot about laboratory analyses of lipids and who has always been available whenever I needed help. His gentle and kind advice and above all his critical suggestions were indispensable for the successful completion of my works on fatty acids.

Professor Sten Symne who felt it his responsibility to help me finish my experiment on oleic acid with success. Who was deeply involved in the work and highly enthusiastic about the outcome, whose advice, critical suggestions and constant encouragement proved absolutely essential for the successful completion of my last experiment.

Doctor Habtamu Zeleke, my Ethiopian Supervisor for his help when I conducted the field trial, for his concern and encouragement.

I am very grateful to Ann-Sofie Fait who showed me DNA extraction and gel electrophoresis and who crossed the plants throughout the winter and autumn of 2008 while I was away. I would like to express my deepest gratitude to Anders Smolka who helped me in from the cold when I broke my leg in 2007 and to Asa Ekman who gladly provides help with the GC on every occasion I needed. I would also like to express my heart-felt gratitude to Doctor Mulatu Geleta and Monica Lotfinia, who have been of great help during the times I was a patient because of a broken leg. I used to obtain their unreserved help whenever I needed it, for which I am very grateful. My deepest gratitude goes to Professor Waheeb Heneen and Iris Heneen for their kind hospitality and the joyous and relaxing times we had together. I am greatly indebted to Faris Hailu who helped me in the analysis of my results. I greatly appreciate his sincerety, patience and profound knowledge.

47 I would like to extend my gratitude to Professor Olof Hellgren and Doctor Goran Nilsson who helped me when I raised my plants in the Biotron. I am also highly indebted to Goran Olsson and Karl-Johan Bergstrand for their help in the greenhouse. I am very thankful to Helen Lindgren, Susanne Hjerdin, Annelie Ahlman, Marissa Prieto-Linde, Jonas Hansson and Britt Green who gladly offered their help whenever I needed. I am especially grateful to all the people I have known here in the department of Plant Breeding and Biotechnology. I would like to extend my sincere thanks to the good time I had with you all. I would like to express my gratitude to the Swedish International Development Agency, Sida-SAREC for financing this project, and to Haramaya University for giving me the opportunity to study in Sweden. I am particularly grateful to Professor Belay Kassa, Doctor Asha Yahya, And Doctor Chemeda Fininsa. I would like to thank Doctor Adugna Wakjira and Doctor Ketema Belete for providing materials and advice. I would like to express my thanks to the former Ethiopian PhD students in Sweden, Sissay M enkir, Esayas Aga, Jemma Haji, Hussein Hamda, Genet Birmeta, Mengistu Urge, Tiliye Feyissa, Yoseph Mekasha &Yitbarek W /H. I am very grateful to Wondossen Bedilu, Abdulfetah Abdulahi, Yemane Aklil and Haimanot Bizuneh who helped me in the collection of the plant material and to Melaku Zewde who managed the plants during the field trial. Finally I would like to thank my relatives Dr. Bisrat Mikael, Theodros Mikael, Dire Mishel, Eyasu Mishel, Maria Petros, Tony Petros, Lule Mikael and my aunts Martha Rufael and Rosa Poli.

48

pa-1

Hereditas 144: 18-24 (2007)

Analysis of genetic diversity of Guizotia abyssinica from Ethiopia using inter simple sequence repeat markers

YOHANNES PETROS1, ARNULF MERKER1 and HABTAMU ZELEKE2 1Swedish University o f Agricultural Sciences, Alnarp, Sweden 2Alemaya University, Dire Dawa, Ethiopia

Petros, Y., Merker, A. and Zeleke, H. 2006. Analysis of genetic diversity of Guizotia abyssinica from Ethiopia using inter simple sequence repeat markers. — Hereditas 144: 18-24. Lund, Sweden. elSSN 1601-5223. Received August 10, 2006. Accepted November 28, 2006

Within and among population genetic diversity of 37 Guizotia abyssinica populations from Ethiopia were analyzed using inter simple sequence repeats (ISSRs). Five primers amplified a total of 118 genomic DNA fragments across a total of 370 individuals of which 106 were polymorphic (89.83%). The average number of polymorphic bands per primer was 21.2. More bands were generated by primer UBC 888 (BDB(CA)7. The total genetic diversity (Ht) and the coefficient of genetic differentiation (Gst) were 0.4115 and 0.0918 respectively, while the within population genetic diversity (Hs) and the among population genetic diversity(Dst) were 0.3738 and 0.03776 respectively suggesting more variability within the populations than among them. The standard genetic distances between the G. abyssinica populations of the eight regions ranged from 0.0281 (between Wollo and Gojam) to 0.1148 (between Jimma and Hararghe). Generally, the standard genetic distances are smaller between populations of neighboring regions and highest between those of Jimma and the other regions, ranging from 0.0696 (between Jimma and Shewa) to 0.1148 (between Jimma and Hararghe). The ISSR based UPGMA clustering using the standardized genetic distances matrix also placed populations from neighboring regions closer than those from farther apart areas, while the UPGMA clustering by regions based on the standard genetic distances produced three clusters following the proximity and the contiguity of the regions. The mean Shannon Weaver diversity indices for the populations of the eight regions ranged from 0.8197 (Jimma) to 0.9176 (Hararghe), with a mean of 0.8841 for the whole material.

Yohannes Petros, Swedish University o f Agricultural Sciences, Department o f Crop Science, SE-230 53 Alnarp, Sweden. E-mail: [email protected]

Guizotia abyssinica (L.f) Cass, (niger in English) in Ethiopia based on phytogeographic, morphological belongs to the family Asteraceae (Compositae), tribe and cytogenetic evidences. Heliantheae, subtribe Coreopsidinae. The taxonomic Niger is the only cultivated species of the genus revision of the genus was made by B a a g o e (1974) Guizotia. It is a diploid plant with a chromosome who reduced the number of species to six. The major number of 2n=2x=30 ( D a g n e 1994). The plant growing areas of niger as edible oil seed are Ethiopia height varies depending on the environmental condi­ and the Indian sub continent (M u r t h y et a!. 1993), tions of the growing area. It grows to an average though it is reported to be grown as a minor oil seed height of 1.2 m, though a height of 2.1 m is also crop also in Sudan, Tanzania, Uganda and Malawi observed depending on the growing condition. Niger (R il e y and B e l a y n e h 1989). In Ethiopia, it is is an annual plant with a hollow stem. It is highly cross cultivated mainly in Gojam, Shewa, Wellega and pollinated and self incompatible ( H i r e m a t h and Gonder regions and to a lesser extent in Jimma, M u r t h y 1986; A d d a et al. 1994). Thus growers Wollo, Arsi and Hararghe regions (G e t in e t and may be recommended to have bee hives close to the S h a r m a 1996). growing area of the crop to facilitate cross pollination It is asserted that Guizotia abyssinica has its center by bees. It is a low yielding crop with low fertility of diversity in Ethiopia (B a a g o e 1974; M u r t h y et al. needs. The low yield may be attributed to the self 1993), and believed to have its centcr of origin also in incompatibility nature of the crop and the low in p u t Ethiopia (H ir e m a t h and M u r t h y 1988). M u r t h y condition under which it is generally grown. et al. (1993) showed the similarity and the homologous According to G e t in e t and S h a r m a (1996), the nature of the genomes of G. abyssinica and G. scabra niger populations in Ethiopia fall into three maturity ssp schimperi based on cytological studies and as­ groups referred to as ‘Bunigne’ niger, ‘Mesno’ niger serted that G. abyssinica might have evolved from G. and Abat’ niger. Bunigne is the early maturing type scabra ssp schimperi in northern Ethiopia through with a shorter growing period of about four months selection and cultivation of large achene mutants. (July to October), while Abat niger takes about seven H ir e m a t h and M u r t h y (1988) also suggested the months to mature (June to December). Mesno niger is evolution of G abyssinica from G scabra ssp schimperi late maturing (September to February) but frost

DOI: 10.1111/j.2007.0018-0661.01969.x Hereditas 144 (2007) Analysis of genetic diversity of Guizotia using ISSRs 19 resistant unlike the other two (G e t in e t a n d S h a r m a MATERIAL AND METHODS 1996). The plant material and DNA extraction Generally, th e re is not m u ch w o rk done on niger. An in-depth treatment of its taxonomy and distribution The plant material used in the study include 10 w as done by B a a g o e (1974). Cytological studies w ere individuals of Guizotia abyssinica (L.f) Cass, from done by several workers (H irem ath and M urthy each of the 37 populations (accessions) collected from 1986; Dagne and Heneen 1992; Hirem ath and eight niger growing regions in Ethiopia (Table 1, M u r t h y 1992; Dagne et al. 2000; D a g n e 2001). Fig. 1). As niger is grown only in localized areas and However, to date there is no report of using ISSR in much smaller scales in some of the regions, only 3, 4 markers to study th e genetic diversity o f niger. ISSR and 5 populations were collected from Jimma, Arsi markers, like any other PCR-based marker, are rapid and Hararghe respectively. Twenty populations were and require only small amount of the template DNA. collected from each of the main niger growing regions Each marker system has its own advantages and (Gojam, Gonder, Wellega, Shewa and Wollo. How­ disadvantages. RAPD’s lack reproducibility (V irk et ever, only five populations that were physically located al. 2000; BoRNETand B ranchard 2001), a n d AFLP farther apart from each of these regions were selected has high operational cost (Prevost and W ilkinson for analysis. This was meant to make the sample size 1999). Microsatelites, though highly polymorphous, of all the regions match for a better comparison of the require prior knowledge of the genomic sequence to develop specific primers and are thus limited Table 1. Regions and site coordinates of the G. to economically important plants (B ornet and abyssinica populations studied. B r a n c h a r d 2001). Inter sim ple sequence re p eat markers with low cost and low labor requirement Region Population code Site coordinates Altitude(m) but with high reliability have been developed Shewa Asfachew 9°53'N, 39°5'E 1009 since 1994 (Zietkiewicz et al. 1994). ISSR Kobo 8°42'N, 38°15'E 2155 amplification does not require genome sequence Soyama 8°25'N, 37°53'E 1900 information but produce highly polymorphic patterns Yaya 9°42'N, 38C49'E, 2669 (Zietkiew icz et al. 1994; Nagaoka and Ogihara W orku 9°45'N, 38°46'E 2674 Jimma Tiyyo 7°52'N, 37°16'E 1760 1997; Prevost and W ilkinson 1999). They seem to D acha 7°53'N, 37°17'E 1748 have the reproducibility of SSR’s and the usefulness of Ayno 7°52'N, 37°17'E 1725 RAPD’s (Bornet and B ranchard 2001), a n d th u s Wellega Kane 9°3'N, 36°29'E 2176 combine the advantages of SSR a n d AFLP a n d th e Qawissa 8°58'N, 36°29'E 2240 utility of RAPD. ISSR markers have been used to Damasa 8°59'N, 36°30'E 2260 Ale 8°57'N, 36°29'E 2252 determine the genetic diversity of Eragrostis tef Jirata 9°2'N, 36°29'E 2142 (Assefa et al. 2003), Cicer (Sudupak 2004) a n d Arsi Gobessa 7°36'N, 39°31'E 2374 wild rice (Q ian et al. 2001). It was also applied to Jelko 7°26'N, 39°32'E 2352 the study of genetic relationships and phylogenetic Tareta 7°35'N, 39°33'E 2336 analysis of various crop plants (Joshi et al. 2000; Shirka 7°36'N, 39°34'E 2334 Wollo Koladi 10°52'N, 39°49'E 2374 M artin and Yelemo 2000; Iruela et al. 2002). Sedeko 10°29'N, 39°56'E 1551 ISSR is a technique that is gaining wide acceptance in Kom bolcha 10°59'N, 39C46'E 1767 the area of plant improvement. Its utility in crop Gerado 11°45'N, 39°37'E 1913 improvement by plant breeding makes use of the fact Libso 11°34'N, 39°40'E 1662 G onder Tamo 12°6'N, 39°46'E 1894 that certain DNA markers are closely linked to Anguabo 11°56'N, 37°48'E 1947 important agronomic traits (Reddy et al. 2002). Zuria 12°22'N, 37 33E 1942 Thus it has been widely used to identify markers T/H 12°32'N, 37°26'E 1895 associated with different qualities in crop plants such Azezo 12°29'N, 37:27'E 1898 as disease tolerance and seed size (A m m iraju et al. Gojam Kotkotum a 11°27'N, 37°14'E 2021 Awabel 10°13'N, 38°8'E 2466 2001). Yabesh 10°37'N, 37°31'E 2097 The present work is an attempt to study the genetic Nifasam 10°17'N, 37°48'E 2463 diversity among the niger populations grown in Rufael 11°30'N, 37°24'E 1793 different regions of Ethiopia using ISSR markers in Hararghe Makanisa 8°53'N, 40'43'E 1714 an effort to provide some information for future M akana 8°54'N, 40=46'E 1702 Kara 8°52'N, 40e40'E 1752 research that might be aimed at improving some of Haro 8°52'N, 40°37'E 1747 the agronomic traits of niger for efficient utilization Bareda 8°51'N, 40C38'E 1746 and conservation. 20 Yohannes Petros et al. Hereditas 144 (2007)

Fig. 1. Geographycal map of Ethiopia indicating the regions of origin of the niger samples used in the study.

diversity of the regions populations. The plant geno­ dTTP), 2% formamide, 0.2 pM primer, 0.05 U pl~* mic DNA was extracted following the CTAB (cetyl- of Taq DNA polymerase and deionized water to trimethyl ammonium bromide) method as applied by make up the reaction volume. Amplification of DNA A s s e f a et al. (2003). was performed in a GENE AMP PCR thermocycler (HITACHI Ltd, Tokyo, Japan), programmed for the PCR amplification and electrophoresis following temperature profiles: 1 min of initial PCR was performed by means of five ISSR primers denaturation at 94‘C followed by 40 cycles, each that were selected out of fifteen tested (Table 2). The consisting of a denaturation step at 94°C for 1 min, primer selection was based on the degree of poly­ an annealing step at 55°C for 2 min, and an morphism and the distinctness of the bands they extension step at 72°C for 2 min, with a final produced when tested on a sample set. Each primer extension at the end of the 40 cycles at 72CC for was tested for reproducibility for different PCR 5 min. Products were electrophoresed in polyacryla­ products of DNA samples of the same population mide gels supplied by Amersham Pharmacia Biotech and separate runs on polyacrylamide gels and AB, along with two lanes of size markers. Fragments the ones producing consistent DNA fragments across were visualized by silver staining on the Hoefer the different samples and PCR runs were selected. Automated gel stainer (Pharmacia Biotech). DNA The PCR reaction mix was a 25 pi volume contain­ fragment sizes were estimated by comparing the ing 10 ng of genomic DNA, 1 x PCR buffer (10 mM DNA bands with a 100 base pair ladder marker Tris-HCl, pH 8.3, 50 mM KC1), 2m M MgCl2, loaded in the peripheral wells of the gel on either side 0.2 mM of the dNTP’s (dATP, dCTP, dGTP, of the sample wells.

Table 2. ISSR primers used in the analysis and the number of bands obtained along with the mean Shannon Weaver diversity index (H') and the polymorphism information content (PIC) for each.

Primer code Primer sequence Bands generated H' (mean +se) PIC (mean+se)

Tot Poly

UBC 834 (AG)g YT 26 23 0.950+0.012 0.466 + 0.008 UBC 841 (GA)s YC 25 25 0.971+0.008 0.480 + 0.006 UBC 866 (CTC)5 19 18 0.969+0.008 0.479+0.006 UBC 878 (G G A T)4 20 17 0.938+0.017 0.459 + 0.011 UBC 888 (BDB(CA)7 28 23 0.964+0.007 0.476 + 0.005

*Y =pyrimidine (C or T), B =non A (C, G or T), D =non C (A, G or T), tot=total bands generated, poly=number of polymorphic bands Hereditas 144 (2007) Analysis o f genetic diversity o f Guizotia using ISSRs 21

Data analysis Table 3. Genetic diversity and the Shannon weaver diversity index for the niger populations o f each region. The bands were recorded as present (1) or absent (0), and assembled in a data matrix. POPGENE software Region Gst Ht Hs Dst H' 1.31 (Y e h et al. 1999) was utilized to generate the single population gene frequencies and the grouped Wolo 0.1143 0.3735 0.3308 0.0427 0.9102 Gonder 0.1806 0.3583 0.2936 0.0647 0.8827 population gene frequencies as well as the N e i (1972) Gojam 0.2090 0.3375 0.2968 0.0407 0.9062 genetic distances matrix between the populations from Hararghe 0.1743 0.3810 0.3146 0.0664 0.9176 the 0, 1, data matrix. The resulting single population Shewa 0.1915 0.4088 0.3305 0.0783 0.9080 gene frequencies were used to construct an unweighted Jimma 0.1738 0.3703 0.3059 0.0644 0.8197 pair group method using the arithmetic average Wollega 0.1609 0.3559 0.2987 0.0573 0.8627 Arsi 0.1896 0.3688 0.2989 0.699 0.8656 (UPGMA) phenogram for the populations using a software package, Genetic Distances and Phylogenetic Analysis (D isp a n 1993). D isp a n (1993) was also used populations Tiyyo and Koladi from Jimma and Wollo for the analysis of the grouped population gene regions respectively. The standard genetic distances for frequencies to generate the standard genetic distance the regions’ populations ranged from 0.0281 (between matrices among the eight regions’ populations and the Wollo and Gojam) to 0.1148 (between Jimma and resulting distance matrices used to construct an Hararghe) (Table 4). The average heterozygosity unweighted pair group method using the arithmetic ranged from 0.3632 (Wellega) to 0.4187 (Shewa), while average (UPGMA) phenogram for the populations in the estimates of the mean Shannon Weaver diversity the eight regions. The N e i (1973) genetic diversity index for the data was 0.8841. When analysed by parameters; the total genetic diversity (Ht), the within region, the mean Shannon weaver diversity indices population genetic diversity (Hs), the among popula­ ranged from 0.8197 for Jimma to 0.9176 for Hararghe tion genetic diversity (Dst) and the coefficient of (Table 4) genetic differentiation (Gst), were analyzed with the A UPGMA dendrogram based on pairwise compar­ same software package from the single population ison of genetic distances identified three major clusters gene frequencies computed by POPGENE software (Fig. 2). Populations from the northern and eastern 1.31 (Y e h et al. 1999). The mean Shannon weaver regions (Wollo, Gonder, Gojam and Hararghe) were diversity indices were calculated following the proce­ grouped together in one cluster, while the Shewa dure of A s se f a et al. (2002). (central), Wellega (western) and Arsi (southern) po­ pulations formed a separate cluster. Two populations RESULTS from south west Shewa and a population from Jimma formed a small cluster removed from the major two. ISSR analysis using five primers produced a total of The clustering pattern for the region’s populations 118 scorable fragments of which 106 (89.83%) were based on the standard genetic distances also reveal the polymorphic. Of the five primers used in the experi­ formation of three clusters with Gojam, Gonder, ment, UBC 888 (BDB(CA)7 produced more bands Wollo and Hararghe forming one cluster, and Wellega, while UBC 841 (GA)8YC produced more polymorphic Shewa and Arsi the other, while Jimma forms a bands. The size of the bands produced by these separate cluster of its own (Fig. 3). primers ranged from about 50 bp to about 1500 bp. The coefficient of gene differentiation (Gst) for all loci was 0.0918, while the estimate of the total genetic DISCUSSION diversity (Ht) was found to be 0.4115. The among population genetic diversity (Dst) was 0.0378, and the ISSR analysis is a powerful tool for assessing the within population genetic diversity (Hs) was 0.3738. genetic diversity in G. abyssinica. This fact is evi­ The within population genetic diversity is highest for denced by the detection of high level of polymorphism the populations from Wollo region, while the lowest in this species using ISSR primers. In the present study among population genetic diversity was observed the three genetic diversity indices; percent poly­ among the Gojam populations (Table 3). The amount morphic loci, average heterozygosity and the Shannon of gene flow among these populations, estimated as Weaver diversity indices proved that the genetic N m =0.5(1 —Gst)/Gst was found to be 2.3716. diversity in the niger populations of Ethiopia is indeed The measure of genetic distance for all populations high. Of the total gene diversity (Ht) which was used in the study (N e i 1972) was lowest (0.0503) 0.4115, the within population genetic diversity ac­ between the populations kombolcha and Gerado, both counts for about 90.8%, while the among population from Wollo region, and highest (0.3261) between genetic diversity takes only a small fraction of the total 22 Yohannes Petros et al. Hereditas 144 (2007)

Table 4. Standard genetic distances of the G. abyssinica populations analysed by region.

Region Wollo Gonder Gojam H ararghe Shewa Jimma Wellega Arsi

Wollo 0 G onder 0.0308 0 Gojam 0.0281 0.0289 0 Hararghe 0.0394 0.0388 0.0336 0 Shewa 0.0428 0.0589 0.0440 0.0499 0 Jimma 0.1007 0.1018 0.1005 0.1148 0.0696 0 Wellega 0.0669 0.0589 0.0682 0.0716 0.0522 0.0505 0 Arsi 0.0491 0.0508 0.0503 0.0468 0.0391 0.0951 0.0355 0

Ht. This is characteristic of cross pollinating species. 2.0488 respectively (data not shown). This becomes In fact niger is highly cross pollinating and very interesting when one considers the geographic self incompatible (G e t in e t a n d S h a r m a 1996, distance between the collection sites from these H ir e m a t h a n d M u r t h y 1986). The amount of regions with lower Nm The samples from Gojam gene flow among these populations was estimated as and Shewa were very much removed from each other Nm =2.3716. A s N m is indicative of the number of within their respective regions as compared to those m ig ra n ts (J ia n et al. 2004), it suggests that the average with relatively higher gene flow values (Table 1) number of migrants per generation (Nm) between the While the general trend in the UPGMA clustering is niger populations of Ethiopia included in the present that of grouping populations by region of origin and stu d y is 2.3716. About similar figure is obtained by proximity of geographic location of the collection using the formula of S l a t k in a n d B a r t o n (1989) fo r sites, not all populations, however, belonging to the the amount of gene flow. Gene flow is found to be same region were grouped together in the same cluster. highest for the populations from Wollo region (Nm = Thus the population ‘Awabel’ from Gojam is included 3.8439) and lowest for the populations from Gojam in the cluster containing Shewa, Wellega and Arsi, followed by those from Shewa with N m o f 1.9461 a n d while two populations from Arsi (Shirka and Tareta)

Fig. 2. Dendrogram illustrating the clustering pattern of 37 populations o f Guizotia abyssinica generated by UPGMA cluster analysis of 106 ISSR markers. Hereditas 144 (2007) Analysis of genetic diversity o f Guizotia using ISSRs 23

Gonder is reported to be of the ‘Bunigne’ type, while those in Wellega and Gojam of the ‘Abat’ type (G e t in e t and S h a r m a 1996). The standard genetic distance between the Jimma Hararghe populations and those from the rest of the regions are

Shewa large indicating a sort of genetic isolation of the Jimma population from those of other regions. Indeed in the dendrogram clustering (Fig. 3), the Jimma populations formed a separate cluster. Looking at the geographic location of Jimma region would explain this fact. Jimma is situated between Wellega and Fig. 3. Clustering by region based on the standard genetic Shewa, but the niger growing areas of Jimma are not distances for the niger populations from Ethiopia contiguous with the niger growing areas of Wellega and Shewa as expanses of two great arid valleys were included in the cluster containing the populations separate it from these regions. On the north west the of the northern and eastern regions. This may be due Diddessa valley (Diddessa desert as it is commonly to either or both of the following reasons: called) separates Jimma from Wellega and on the north east side the Ghibe valley (also called the Ghibe - the continuity of the niger growing regions of desert) separates Jimma from Shewa. Thus, it is no Ethiopia which makes possible the transfer of wonder that the niger populations from Jimma are seed materials from one region to the other. In removed from the others on the basis of the dendro­ fact, samples collected from the border of a gram grouping maintaining higher genetic distances region are closer by proximity of geographic with all the populations from the other regions. location to the adjoining region than to the Overall our findings demonstrated that there is collection sites of the same region. For example, variation within the niger populations of Ethiopia the population Awabel collected from southern and that ISSR markers would be useful for the Gojam is grouped with Yaya and Worku collected assessment of genetic diversity and phylogenetic from northern Shewa bordering Gojam. It is relationships of this species. thought that though the Abay valley separates northern Shewa from southern Gojam material Acknowledgements - We are indeed very grateful for those transfer between these two areas is highly likely. who contributed to the realization and completion of this - The existence of three strains of niger that are study. To the Institute of Biodiversity conservation and Research, the Swedish Univ. of Agricultural Sciences (SLU) differentiated on the basis of the duration to and the Alemaya Univ. Special thanks go to Dr. Kebebew maturity identified as ‘bunigne’, ‘abat’ and Assefa, Dr. Ketema Belete, Mr. Faris Hailu and Mr. Mulatu ‘mesno’. Geleta who contributed either in the provision of materials or in the analysis of the data. In fact, the field observation of these same samples reveals the existence of niger strains that are widely different in the duration to flowering and maturity. REFERENCES According to the field observation all samples from Adda, S., Reddy, T. P. and Kishor. P.B.K. 1994. Somatic Wollo and Hararghe are of the early maturing type, embryogenesis and organogenesis in Guizotia abyssinica. and only one sample from Shewa (Asfachew) and two - In Vitro. Cell. Dev. Biol. 30P: 104-107. samples from Gonder (Tamo and Anguabo) are of the Ammiraju, J. S. S., Dhlakia, B. B. and Santra, D. K. 2001. early maturing types, while the samples from Gojam, Identification of inter simple sequence repeat (ISSR) markers associated with seed size in wheat. - Theor Wellega, Jimma Arsi and Shewa are of the Abat (late Appl. Genet. 102: 726-732. maturing) types. The clustering in the dendrogram Assefa, K.. Merker. A. and Teffera, H. 2002. Qualitative (Fig. 2) seems, to a certain extent to follow this trend trait variation in Tef [Eragrostis tef (Zuc.) Trotter]. while the general trend is that of clustering the - Euphytica 127: 399-410. Assefa, K., Merker, A. and Teffera, H. 2003. Inter simple populations by their region of origin. sequence repeat (ISSR) analysis of genetic diversity in Tef Thus it appears that the ‘Bunigne’ populations of a (Eragrostis tef (Zucc.) Trotter). - Hereditas 139: 174— region are grouped differently from the ‘Abat’ popula­ 183. tions of the same region in the dendrogram. However, Baagoe, J. 1974. The Genus Guizotia (Compositae). A taxonomic revision. - Bot. Tidskr. 69: 1-39. the confirmation of this assertion needs further Bornet, B. and Branchard, M. 2001. Non anchored inter investigation. Most of the niger grown in Wollo and simple sequence repeat (ISSR) markers: reproducible and 24 Yohannes Petros et al. Hereditas 144 (2007)

specific tools for genome fingerprinting. - Plant Mol. Murthy, H. N , Hiremath, S. C. and Salimath, S. S. 1993. Biol. Rep. 19: 209-215. Origin, evolution and genome differentiation in Guizotia Dagne, K. 1994. Meiosis in interspecific hybrids and abyssinica and its wild species. - Theor. Appl. Genet. genomic interrelationships in Guizotia Cass. - Composi- 587-592. tae)-H ereditas 121: 119-129. Nagaoka, T. and Ogihara, Y. 1997. Applicability of inter Dagne, K. 2001. Meiotic properties of induced autopoly­ simple sequence repeat polymorphisms in wheat for use ploid Guizotia abyssinica (L.f) Cass. - J. Genet. Breed. as DNA markers in comparison to RFLP and RAPD 55: 11-16. markers. - Theor. Appl. Genet. 94: 597-602. Dagne, K. and Heneen, W.K. 1992. The karyotypes and Nei, M. 1972. Genetic distance between populations. - Am. nucleoli of Guizotia abyssinica (Compositae). - Hereditas N at. 106: 283-292. 1 17: 73-83. Nei, M. 1973. Analysis of gene diversity in subdivided Dagne, K., Cheng, B. and Heneen, W. K. 2000. Number and populations. - Proc. Natl Acad. Sci. USA 70: 3321- sites of rDNA loci of Guizotia abyssinica (L.f) Cass, as 3323. determined by florescence in situ hybridization. - Her­ Prevost, A. and Wilkinson, M. J. 1999. A new system of editas 132: 63-65. comparing PCR primers applied to ISSR fingerprinting DISPAN 1993. Genetic distance and phylogenetic analysis. of potato cultivars. - Theor. Appl. Genet. 98: 107-112. © by Tatsuya and the Pennsylvania State Univ. Qian, W., Ge, S. and Hong, D-Y. 2001. Genetic variation Getinet, A. and Sharma, S. M. 1996. Niger. Guizotia within and among populations of a wild rice Oryza abyssinica (L.f) Cass. Promoting the conservation and granulata from China detected by RAPD and ISSR use of underutilized and neglected crops. 5. - Int. Genet. markers. - Theor. Appl. Genet. 102: 440-449. Resour. Inst., Gatersleben, Germany. Reddy, M. P., Sarla, N. and Siddiq, E. A. 2002. Inter simple Hirem ath, S. C. and Murthy, H. N. 1986. The structure, sequence repeat (ISSR) polymorphism and its application stability and meiotic behavior of B-chromosome in in plant breeding. - Euphytica 128: 9-17. Guizotia scabra (Vis.) Chiov. ssp scabra (Compositae). Riley, K. W. and Belayneh, H. 1989. Niger. - In: Robbelen, - Caryologia. 39: 397-402. G., Downey, R. K. and Ashri, A. (eds), Oil crops of the Hirem ath, S. C. and Murthy, H. N. 1988. Dom estication of World: their breeding and utilization. McGraw-Hill, pp. niger {Guizotia abyssinica). - Euphytica 37: 225-228. 394-403. Hirem ath, S. C. and Murthy, H. N. 1992. Cytogenetical Slatkin, M. and Barton, N. H. 1989. A com parison of three studies in Guizotia (Asteraceae). - Caryologia 45: 69-82. indirect methods for estimating average levels of gene Iruela, M., Rubio, J. and Cubera, J. J. 2002. Phylogenetic flow. - Evolution 43: 1349-1368. analysis in the genus Cicer and cultivated chickpea using Sudupak, M. A. 2004. Inter and intra species inter simple RAPD and ISSR markers. - Theor. Appl. Genet. 104: sequence repeat (ISSR) variations in the genus Cicer. 643-651. - Euphytica 135: 229-238. Jian, S., Tang, T. and Zhong, Y. 2004. Variation in inter Virk, P. S., Zhu, J. and Newbury, H. J. 2000. Effectiveness of simple sequence repeat (ISSR) in mangrove and non­ different classes of molecular marker for classifying and mangrove populations of Heritiera littoralis (Sterculia- revealing variation in rice (Oryza sativa) germplasm. ceae) from China and Australia. - Aquat. Bot. 79: 75- - Euphytica 112: 275-284. 86 . Yeh, F. C., Yang, R. C. and Boyl, T. 1999. Popgene version Joshi, S. P., Gupta, V. S. and Agarwal, R. K. 2000. Genetic 1.31. Microsoft window based freeware for population diversity and phylogenetic relationship as revealed by genetic analysis. - Univ. of Alberta, Canada. inter simple sequence repeat (ISSR) polymorphism in the Zietkiewicz, E., Rafalski, A. and Labuda, D. 1994. Genome genus Oryza. - Theor. Appl. Genet. 100: 1311-1320. fingerprinting by simple sequence repeat (SSR)- anchored Martin, J. P. and Sanchez-Yelemo, M. D. 2000. Genetic polymerase chain reaction amplification. - Genomics 20: relationships among species of the genus Diplotaxis 176-183. (Brassicaceae) using inter simple sequence repeat mar­ kers. - Theor. Appl. Genet. 101: 1234-1241. II Genet Resour Crop Evol (2008) 55:451^458 DOI 10.1007/s 10722-007-9251-4

RESEARCH ARTICLE

Analysis of genetic diversity and relationships of wild Guizotia species from Ethiopia using ISSR markers

Yohannes Petros • Arnulf Merker • Habtamu Zeleke

Received: 6 January 2007 / Accepted: 7 May 2007 / Published online: 14 July 2007 © Springer Science+Business Media B.V. 2007

Abstract Genetic relationships and diversity of 45 The unweighted pair group method using the arithme­ Guizotia populations each consisting of ten individuals tic average clustering of the five taxa using the standard and belonging to five taxa of the genus Guizotia were genetic distances produced two clusters, with G. analyzed using Inter Simple Sequence Repeat (ISSR) scabra ssp. schimperi and G. villosa occurring in one markers. Five ISSR primers generated a total of 145 cluster and G. scabra ssp. scabra, G. arborescens and scorable bands across the 450 individuals used for the G. zavattarii together in the other cluster. The study study. The percent polymorphic loci for the taxa ranged reveals that G. scabra ssp. schimperi is more closely from 68.2 (G. arborescens) to 88% (G. scabra ssp. related to G. villosa than to G. scabra ssp. scabra. schimperi), with G. scabra ssp. scabra, G. zavattarii and G. villosa following G. scabra ssp. schimperi in Keywords Gene flow • Genetic diversity • this order with respect to the abundance of percent Guizotia ■ Polymorphic loci • UPGMA polymorphic loci. The Shannon-Weaver diversity indices (H'), for the five taxa also followed a similar pattern, with G. scabra ssp. schimperi exhibiting the Introduction highest H' (0.7373) and G. arborescens the least (0.5791), while H' for G. scabra ssp. scabra, G. villosa The genus Guizotia belongs to the family Asteraceae and G. zavattarii were 0.7313, 0.6620 and 0.6564, (Compositae), tribe Heliantheae, subtribe Coreopsid- respectively. The least genetic distance (0.1188) was inae. The taxonomic revision of the genus was made observed between G. scabra ssp. schimperi and by Baagoe (1974) who reduced the number of species G.villosa, revealing closer genetic relationships of to six. The genus Guizotia consists of six species and the two species with each other than with the others, two subspecies five of which are found in Ethiopia and the highest genetic distance (0.2740) was observed (Baagoe 1974). These are G. abyssinica (L.f.) Cass., between G. scabra ssp. schimperi and G. zavattarii. G. scabra (Vis.) Chiov. ssp. scabra; G. scabra (Vis.) Chiov. ssp. schimperi (Sch. Bip.) Baagoe; G. villosa Sch. Bip., G. zavattarii Lanza; Y. Petros (E3) • A. Merker Department of Crop Science, Swedish University Guizotia arborescens (I. Friis) and G. jacksonii of Agricultural Sciences, Box 44, Alnarp 230 53, Sweden (S. Moore) J. Baagoe. All the six species grow in East e-mail: [email protected] Africa. The chromosome number of all Guizotia species as reported by Dagne (1994a) is 2x = 2n = 30. H. Zeleke Alemaya University of Agriculture, P.O. Box 138, The distribution of the Guizotia species in Africa Dire Dawa, Ethiopia and in Ethiopia varies greatly. Some of the species

^ Springer 452 Genet Resour Crop Evol (2008) 55:451-458 such as G. villosa, G. arborescens, G. zavattarii and However, to date little work has been done on the G. jacksonii are restricted in their distribution, while genetic relationships of the Guizotia species. In fact others like G. scabra ssp. scabra, G. scabra ssp. no work has been done on the genetic diversity and schimperi and G. abyssinica cover a relatively wide relationships of the Guizotia species using ISSR area in east Africa with a greater concentration in markers. The utility of ISSR to study genetic Ethiopia. G. arborescens is a component of the relationships and identify taxa within a genus has natural vegetation in the Imatong mountains of been shown by several workers (Ajebade et al. 2000; Uganda and Sudan and southern Ethiopia (Friis Rajesh et al. 2002; Tikunov et al. 2003; Vijayan and 1971; Baagoe 1974; Hiremath and Murthy 1988). Chatterjee 2003; Sudupak 2004). ISSR is a new G. zavattarii is endemic to mount Mega in southern approach of direct DNA analysis developed in 1994 Ethiopia and the Huri hills in northern Kenya (Dagne (Zietkiewicz et al. 1994). The technique relies on the 1994a; Baagoe 1974; Hiremath and Murthy 1988). fact that eukaryotic genomes contain an abundance of The distribution of G. scabra ssp. scabra covers a Simple Sequence Repeats (SSR) (Lagercrantz et al. wide range extending from east Africa to Nigeria 1993). The SSR sequences are used as primers in with a distributional gap in the rain forest of Congo, polymerase chain reactions (PCR) to amplify regions while G. scabra ssp. schimperi is a common weed of between SSR loci (Zietkiewicz et al. 1994). cultivated crops and widely distributed in Ethiopia Inter Simple Sequence Repeat analysis of G. (Hiremath and Murthy 1988; Dagne 1994a). G. abyssinica populations collected from eight niger villosa is endemic to the northern part of Ethiopia growing regions of Ethiopia was done by Petros et al. and G. jacksonii is endemic to mount Kenya, (2007). It was shown that the technique reveals high Aberdares and Mt. Elgon in Uganda and Kenya polymorphism within and among the niger populations (Dagne 1994a; Hiremath and Murthy 1988). of Ethiopia. In fact the technique was able to discrim­ Murthy et al. (1995) tried to revise the present inate between the different strains of niger that were taxonomic classification of the genus that was done by identified based on the duration to flowering and Baagoe (1974). Baagoe’s classification of the genus maturity. Geleta et al. (2007) also studied the genetic was based on morphological similarity. Murthy et al. diversity of niger populations from Ethiopia as (1995) concluded that classifying G. scabra ssp. revealed by random amplified polymorphic DNA and schimperi as a subspecies of G. scabra was not sound reported a high degree of polymorphism among and on the basis of their karyomorphological studies. within the populations. As G. abyssinica is the most While the existing classification of the genus is based important oil crop in Ethiopia, the present work on morphological similarities, karyomorphological becomes worthwhile because attempts to improve its studies, however, reveal that G. scabra ssp. schimperi agronomic or oil quality, whether it be through is more closely related to G. abyssinica than to G. conventional plant breeding or genetic engineering scabra ssp. scabra (Dagne 1994b; Murthy et al. 1993, would, to a certain extent rely on the knowledge of its 1995). Similar conclusion was made by Hiremath and wild relatives. The present work is therefore an attempt Murthy (1992) who suggested that G. abyssinica, G. to study the genetic diversity and relationships among scabra ssp. schimperi and G. villosa are more closely five of the Guizotia taxa growing in Ethiopia using related as they have symmetrical karyotypes unlike G. ISSR markers envisaging that the information obtained zavattarii, G. scabra ssp. scabra and G. jacksonii from this investigation might be used by future which have asymmetrical karyotypes. Furthermore, investigators in their endeavor to improve the agro­ Dagne (1994b) showed that G. scabra ssp. schimperi, nomic quality of niger. G. scabra ssp. scabra, G. villosa and G. abyssinica are closely related because of the high level of crossabil­ ity among them. His studies revealed that chromo­ Materials and methods some homology is higher between G. scabra ssp. schimperi and G. abyssinica than between any other The plant material and DNA extraction two taxa he studied. The controversy, however, on the taxonomic status of G. scabra ssp. schimperi still The plant materials used in the study were collected remains. during November-December 2003. The materials

Springer Genet Resour Crop Evol (2008) 55:451^-58 453 include 27 populations of G. scabra ssp. schimperi the degree of polymorphism and the distinctness of collected from 11 regions in the country (Table 1). the bands they produced when tested on a sample set. Eight populations of G. villosa collected from Gojam The PCR reaction mix was a 25 pi volume containing and Gonder, seven populations of G. scabra ssp. 10 ng of genomic DNA, lx PCR buffer (10 mM Tris- scabra from Jimma, Illubabor and Wellega, two HC1, pH 8.3, 50 mM KC1), 2 mM MgCl2, 0.2 mM of populations of G. zavattarii from Sidamo and one the dNTP’s (dATP, dCTP, dGTP, dTTP), 2% Form- population of G. arborescens from Kaffa. While amide, 0.2 pM primer, 0.05 U/pl of Taq DNA every effort was made to include all the wild Guizotia polymerase and deionized water to make up the taxa growing in Ethiopia in the study, not all were, reaction volume. Amplification of DNA was per­ however equally represented in the collection. This formed in a GENE AMP PCR thermocycler (HIT­ anomaly stems from the fact that some of the species ACHI Ltd, Tokyo, Japan), programmed for the such as G. villosa, G. zavattarii and G. arborescens following temperature profiles: 1 min of initial are highly restricted in their distribution while others denaturation at 94°C followed by 40 cycles, each like G. scabra ssp. scimperi are found growing all consisting of a denaturation step at 94°C for 1 min, an over the country. Thus, 27 populations of G. scabra annealing step at 55°C for 2 min, and an extension ssp. schimperi were collected while G. scabra ssp. step at 72°C for 2 min, with a final extension at the scabra was represented by seven populations from end of the 40 cycles at 72°C for 5 min. Products were three of the western regions of the country where they electrophoresed in polyacrylamide gels supplied by occur. G. villosa, G. zavattarii and G. arborescens on Amersham Pharmacia Biotech AB, along with two the other hand were very much localized in their lanes of size markers. Fragments were visualized by distribution. G. villosa grows only in certain areas in silver staining on the Hoefer Automated gel stainer Gojam, Gonder and Tigray in the north and collection (Pharmacia Biotech). DNA fragment sizes were was made from Gojam and Gonder. G. zavattarii estimated by comparing the DNA bands with a 100 grows in very restricted areas covering small areas base pair ladder marker loaded in the peripheral wells near Yabello town and Mt. Mega in the south. As this of the gel on either side of the sample wells. species has a continuous distribution over a small area, the authors considered the Yabello material as a Data analysis single population and that from Mt. Mega as another. G. arborescens on the other hand, has the most The bands were recorded as present (1) or absent (0), restricted distribution of them all, growing near Omo and assembled in a data matrix. POPGENE software Nada in the south west. In fact all the G. arborescens 1.31 (Yeh et al. 1999) was utilized to generate the materials used by Friis (1971) to describe the species single population gene frequencies and the grouped came from this area around Mount Maigudo. Thus, population gene frequencies as well as Nei’s (1972) the inclusion of only a single population of G. genetic distances matrix between the populations arborescens and only two populations of G. zavattarii from the 0, 1, data matrix. The resulting single is due to their highly localized distribution. population gene frequencies were used to construct The achenes (seeds) collected from Ethiopia were an unweighted pair group method using the arithme­ grown in a greenhouse at the department of crop tic average (UPGMA) phenogram for the populations science of the Swedish University of Agricultural using a software package, Genetic Distances and Sciences (SLU). Phylogenetic Analysis (DISPAN 1993). DISPAN The plant genomic DNA was extracted following (1993) was also used for the*analysis of the Grouped the CTAB (cetyltrimethyl ammonium bromide) population gene frequencies to generate the standard method as applied by Assefa et al. (2003). genetic distances matrix between the five Guizotia taxa studied and the resulting distance matrix used to PCR amplification and electrophoresis construct an (UPGMA) phenogram for the five taxa. Nei’s (1973) genetic diversity parameters; the total Polymerase chain reaction was performed by means genetic diversity (Ht), the within population genetic of five ISSR primers that were selected out of 15 diversity (Hs), the among population genetic diver­ tested (Table 2). The primer selection was based on sity (Dst) and the coefficient of genetic differentiation

4^ Springer 454 Genet Resour Crop Evol (2008) 55:451-458

Table 1 Region of collection, altitude, site coordinates (description) of the collection sites of the Guizotia species used in the study

No Name Region Pop. code Site coordinates Altitude (m)

1 Guizotia scabra ssp. schimper Hararghe gspl 09°01'01N, 040°53'47E 2,384 2 Guizotia scabra ssp. schimper Hararghe gsp2 09°58'06N, 040350'17E 2,147 3 Guizotia scabra ssp. schimper W. Shewa gsp3 08°38'25N, 038°09'35E 2,292 4 Guizotia scabra ssp. schimper W. Shewa gsp4 08°45'43N, 038°18'20E 2,114 5 Guizotia scabra ssp. schimper W.Shewa gsp5 08°42'10N, 038°15'04E 2,153 6 Guizotia scabra ssp. schimper N. Shewa gsp6 09°45'09N, 038=28' 13E 2,283 7 Guizotia scabra ssp. schimper Jimma gsp7 08°09'17N, 037°31'43E 1,728 8 Guizotia scabra ssp. schimper Jimma gsp8 08°03'39N, 037°22'10E 1,891 9 Guizotia scabra ssp. schimper Illubabor gsp9 08°22'23N, 036°12'38E 2,167 10 Guizotia scabra ssp. schimper Illubabor gsplO 08°23'14N, 035°59'10E 1,618 11 Guizotia scabra ssp. schimper Wellega gspll 09°02'56N, 036°23'50E 2,007 12 Guizotia scabra ssp. schimper Wellega gspl 2 09°01'43N, 036°38'1 IE 2,245 13 Guizotia scabra ssp. schimper Arsi gspl 3 07°49'08N, 039°08'31E 2,276 14 Guizotia scabra ssp. schimper Arsi gspl 4 07° 13'10N, 039°14'31E 2,517 15 Guizotia scabra ssp. schimper Bale gspl 5 07°12'47N, 040°30'48E 2,177 16 Guizotia scabra ssp. schimper Bale gspl6 07°00'18N, 039°28556E 2,411 17 Guizotia scabra ssp. schimper Gojam gspl 7 10°21'15N, 037°12'12E 2,394 18 Guizotia scabra ssp. schimper Gojam gspl 8 10°22'16N, 037°36'24E 2,249 19 Guizotia scabra ssp. schimper Gojam gspl 9 10°31'25N, 037°31'17E 2,097 20 Guizotia scabra ssp. schimper Gojam gsp20 10°41'32N, 037°12'31E 2,006 21 Guizotia scabra ssp. schimper Gonder gsp21 12°39'16N, 037°29'14E 2,246 22 Guizotia scabra ssp. schimper Gonder gsp22 12°48'1 IN, 037°19'19E 2,682 23 Guizotia scabra ssp. schimper Gonder gsp23 12°40'27N, 037°30'10E 2,429 24 Guizotia scabra ssp. schimper Gonder gsp24 12°43'14N, 037°29'30E 2,690 25 Guizotia scabra ssp. schimper Well gsp25 11°43'06N, 038°55'08E 2,522 26 Guizotia scabra ssp. schimper Well gsp26 11°43'50N, 038 55'32E 2,375 27 Guizotia scabra ssp. schimper Sidamo gsp27 06°04'27N, 038°13'27E 2,328 28 Guizotia villosa Gojam Gv28 H°27'01N, 037°23'57E 1,922 29 Guizotia villosa Gojam Gv29 11°22'37N, 037°24'38E 2,220 30 Guizotia villosa Gojam gv30 iri7'10N , 037°28'49E 2,210 31 Guizotia villosa Gojam gv31 11°21'54N, 037°25'19E 2,296 32 Guizotia villosa Gojam gv32 11°17'28N, 037°2834E 2,247 33 Guizotia villosa Gonder gv33 H °53’01N, 037°39'54E 1,825 34 Guizotia villosa Gonder gv34 12°11'40N, 037°40'53E 1,827 35 Guizotia villosa Gonder gv35 12°13'43N, 037°37'44E 1,894 36 Guizotia scabra ssp. scabra Jimma gsc36 07°59'55N, 037°26'22E 1,929 37 Guizotia scabra ssp. scabra Jimma gsc37 08°09'47N, 037°32'09E 1,711 38 Guizotia scabra ssp. scabra Jimma gsc38 07°40'43N, 036°53'59E 1,819 39 Guizotia scabra ssp. scabra Illubaborl gsc39 07°56'19N, 036°30'44E 1,751 40 Guizotia scabra ssp. scabra Illubabor gsc40 08°00'44N, 036°28'26 1,706 41 Guizotia scabra ssp. scabra Wellega gsc41 09°05'59N, 036' 58'26E 1,849 42 Guizotia scabra ssp. scabra Wellega gsc42 09°04'10N, 036°54'51 1,767 43 Guizotia arborescens Kaffa gar43 Omo Nada, on the hill 27 km from the town 2,325-2,420 44 Guizotia zavattarii Sidamo gz44 Yabelo, 3-7 km from the town on the road to Konso 2,000-2,050 45 Guizotia zavattarii Sidamo gz45 3 km from Mega to Moyale 1,700-1,750

Springer Genet Resour Crop Evol (2008) 55:451^458 455

Table 2 ISSR primers used in the analysis and the number of bands obtained along with the mean Shannon-Weaver diversity index (H')

Primer code Sequence Bands generated H' (mean ± se) PIC (mean ± se)

Tot poly

UBC 834 (AG)g y t 31 29 0.7300 ± 0.0475 0.3373 ± 0.0094 UBC 841 (GA)g YC 33 32 0.7258 ± 0.0483 0.3447 ± 0.0096 UBC 866 (CTC)S 27 26 0.6693 ± 0.0480 0.3298 ± 0.0112 UBC 878 (GGAT)4 25 23 0.5997 ± 0.0577 0.2972 ± 0.0126 UBC 888 [BDB(CA]7 29 28 0.7809 ± 0.0486 0.3971 ± 0.0097

Y pyrimidine (C or T), B non A (C, G or T), D non C (A, G or T), tot total number of bands, poly number of polymorphic bands

(Gst), were analyzed with DISPAN software (1993), Table 3 Genetic diversity and the Shannon-Weaver diversity from the single population gene frequencies com­ index for the Guizotia species investigated puted by POPGENE software 1.31. The mean Taxa Gst Ht Hs Dst H' Shannon-Weaver diversity indices were calculated following the procedure of Assefa et al. (2002). G. scabra ssp. 0.3610 0.2401 0.1534 0.0867 0.7373 schimperi G. villosa 0.2551 0.1867 0.1391 0.0476 0.6619 G. scabra ssp. 0.2552 0.2117 0.0541 Results 0.1576 0.7313 scabra G. zavattarii 0.0985 0.1791 0.1615 0.0176 0.6564 Polymorphism and genetic diversity

Five ISSR primers amplified a total of 145 scorable zygosity ranged from 0.1867 (G. villosa) to 0.2410 bands (Table 2). The highest number of bands was (G. scabra ssp. schimperi). scored for G. scabra ssp. schimperi of which 88% were polymorphic, followed by G. scabra ssp. scabra Genetic distance and gene flow with 83.5% polymorphic bands. The least polymor­ phism (68.2%) was observed in G. arborescens. The The standard genetic distance between the G. scabra coefficient of genetic differentiation for the G. scabra ssp. schimperi populations was found to be least ssp. schimperi, G. scabra ssp. scabra, G. villosa and between populations from adjacent areas and greatest G. zavattarii populations were 0.3610, 0.2552, between those distant from each other (data not 0.2551 and 0.0985, respectively, while the total shown). Thus, the standard genetic distance for the G. genetic diversity was 0.2401, 0.1868, 0.2117 and scabra ssp. schimperi populations ranged from 0.1791 for G. scabra ssp. schimperi, G. villosa, G. 0.0017 between populations 1 and 2 (gshl & gsh2) scabra ssp. scabra and G. zavattarii, respectively. both from Hararghe region to 0.1754 between The within population genetic diversity was observed population 18 and population 27 (gshl8 and gsh27) to be higher than the among population genetic from Gojam and Sidamo, respectively. Likewise the diversity for G. scabra ssp. scabra, G. scabra ssp. standard genetic distance for the G. villosa popula­ schimperi, G. villosa, and G. zavattarii (Table 3). The tions ranged from 0.0125 (between gv 31 and gv 32) most diverse among its populations was G. scabra both from Gojam to 0.0866 (between gv 28 and gv ssp. shimperi with a Dst (among population genetic 34) from Gojam and Gonder, respectively, while for diversity) of 0.0867 followed by G. scabra ssp. G. scabra ssp. scabra it ranged from 0.097 (between scabra, G. villosa and G. zavattarii with Dst of gsc 36 and gsc37) to 0.1076 (between gsc 36 and gsc 0.0540, 0.0476 and 0.0177 respectively. The mean 42). The standard genetic distance between the two Shannon-Weaver diversity indices for the taxa ranged populations of G. zavattarii was 0.0214. The Nei’s from 0.5791 (G. arborescens) to 0.7373 (G. scabra (1972) genetic distance also followed a similar trend ssp. schimperi) (Table 3), while the average hetero- for all the individuals of the species (data not shown).

Springer 456 Genet Resour Crop Evol (2008) 55:451—458

The standard genetic distance among the five taxa G. zavattarii populations, estimated as Nm = 0.5(1 - ranged from 0.1188 (between G. scabra ssp. Gst)/Gst was found to be 0.8849, 1.5473, 1.6782 and schimperi and G. villosa) to 0.2740 (between 4.576, respectively. G. scabra ssp. schimperi and G. zavattarii) (Table 4). The UPGMA dendrogram (Fig. 1) based on The amount of gene flow among G. scabra ssp. pairwise comparison of genetic distances produced schimperi, G. villosa, G. scabra ssp. scabra and five separate clusters, one for each of the five taxa.

Table 4 Standard genetic G. scabra G. villosa G. scabra G. zavattarii distances of the Guizotia ssp. schimperi ssp. scabra species used for the study G. villosa 0.1188 G. scabra ssp. 0.1899 0.2729 scabra G. zavattarii 0.2535 0.2602 0.1819 G. arborescens 0.2740 0.2555 0.2553 0.1716

Fig. 1 Clustering pattern of 45 populations of wild Guizotia species generated by UPGMA cluster analysis, gsh - G. scabra ssp. schimperi, gsc - G. scabra ssp. scabra, gv - G. villosa, gar - G. arborescens, gz - G. zavattarii

Springer Genet Resour Crop Evol (2008) 55:451—458 457

The extent of genetic relatedness between popula­ was found to be greater and the amount of gene flow tions of a species is revealed as populations from less among the populations of G. scabra ssp. nearby areas are placed nearer to each other forming schimperi populations than among the populations smaller clusters in the major clusters containing the of G. scabra ssp. scabra, G. villosa and G. zavattarii specific taxa. The genetic affinities and relatedness of populations. In fact the higher range of variation in the five taxa under consideration can also be observed the standard genetic distance among the G. scabra from the UPGMA clustering based on the standard ssp. schimperi populations is the result of the greater genetic distances between the five taxa (Fig. 2). Thus distance maintained by population 27 (gsh 27) from two clusters, one consisting of G. scabra ssp. the others attaining a distance of 0.1754 with schimperi and G.villosa and the other consisting of population 18 (gsh 18). The standard genetic distance G. arborescens, G. zavattarii and G. scabra ssp. between gsh 27 and the rest of the G. scabra ssp. scabra can be observed. schimperi populations ranged from 0.0488 with population 26 (gsh 26) to 0.1754 with gsh 18. This population is even more distant from all the popula­ Discussion tions of the other taxa studied. Thus, the highest standard genetic distances between gsh 27 and the The high polymorphism observed in G. scabra ssp. other taxa were 0.2677, 0.4081, 0.44574 and 0.4044 schimperi is attributed to the countrywide distribution between G. villosa, G. scabra ssp. scabra, G. of this species and the wide distance separating the arborescens and G. zavattarii respectively, revealing collection sites of its populations included in the its affinity with the G. scabra ssp. schimperi popu­ study covering most of the regions in the country lations and its distance from the other Guizotia taxa. (Table 1). G. scabra ssp. scabra is the second most The standard genetic distance between the Guizo­ polymorphic. In fact it is also the second most widely tia taxa studied based on the grouped gene frequency distributed in Ethiopia among the five Guizotia taxa reveal that the distance is least between G. scabra studied. The least polymorphic of the five taxa ssp. schimperi and G. villosa (0.1188) and highest studied was G. arborescens followed by G. zavattarii. between G. scabra ssp. schimperi and G. zavattarii This can likewise be explained by the lesser extent of (0.2740). The UPGMA dendrogram based on the distribution of these two species in the country. standard genetic distances for the five species also Overall G. scabra ssp. schimperi is more variable elucidate this fact (Fig. 2). On the dendrogram G. with respect to percent polymorphic loci, total gene scabra ssp. schimperi and G. villosa were clustered diversity, the mean Shannon-Weaver Diversity Index together, while G. zavattarii, G. arborescens and G. and the range of standard genetic distance among its scabra ssp scabra were grouped together. population. Our analysis reveals that G. scabra ssp. schimperi It may be interesting to consider the amount of and G. villosa share more features in common than gene flow among the populations of these taxa in the with any of the other three species. Thus based on the light of the genetic distances among the populations amount of bands shared, and the standard genetic of the respective taxa. Overall, the genetic distance distance, G. scabra ssp. schimperi is more closely related to G. villosa than to G. scabra ssp. scabra. Apart from their similarity at the molecular level, these two species share many other characteristics including similar karyotypes (Hiremath and Murthy 1992), herbaceous growth habit, being annual plants and both being common weeds in fields under cultivation, very often occurring together in the same field. It has already been shown by Dagne (1994b) Fig. 2 Unweighted pair group method using the arithmetic that these species can form interspecific hybrids in average clustering of the five Guizotia taxa based on the areas where they grow together. standard genetic distance gsh— G. scabra ssp. schimperi, gsc — G. scabra ssp. scabra, gv— G. villosa, gar— G. Based on this study, the authors recommend the Arborescens, gz— G. zavattarii revision of the previous classification that merged the

*0 Springer 458 Genet Resour Crop Evol (2008) 55:451-458 two taxa as sub species of G. scabra. The authors also Geleta M, Bryngelsson T, Bekele E, Dagne K (2007) Genetic suggest that G. scabra ssp. schimperi merits its own diversity of Guizotia abyssinica (L.f) Cass. (Asteraceae) from Ethiopia as revealed by random amplified poly­ specific status as the result of the present investiga­ morphic DNA. Genet Resour Crop Evol 54:601-614 tion and earlier reports (Murthy et al. 1993; Getinet Getinet A, Sharma SM (1996) Niger. Guizotia abyssinica (L.f.) and Sharma 1996). However, the classification of Cass. Promoting the conservation and use of underutilized G. scabra ssp. schimperi as a subspecies of G. abyss­ and neglected crops 5. IPGRI Rome, IPK, Gatersleben Hiremath SC, Murthy HN (1988) Domestication of niger inica, as suggested by some authors (Murthy et al. (Guizotia abyssinica). Euphytica 37:225-228 1995), may not be a solution because the Interna­ Hiremath SC, Murthy HN (1992) Cytogenetical studies in tional Rules of Botanical Nomenclature would not Guizotia (Asteraceae). Caryologia 45:69-82 support the merger of a wild species with a cultivated Lagercrantz U, Ellegren H, Andersson L (1993) The abundance of various polymorphic microsatelite motifs differs be­ one. tween plants and vertebrates. Nucleic Acids Res 21:1111— 1115 Acknowledgements The authors would like to extend their Murthy HN, Hiremath SC, Salimath SS (1993) Origin. Evo­ gratitude to the Swedish University of Agricultural Sciences lution and genome differentiation in Guizotia abyssinica and the Alemaya University because of whose cooperation this and its wild species. Theoret Appl Genet 87:587-592 study was made possible. We are very grateful to Dr Kifle Murthy HN, Hiremath SC, Pyati AN (1995) Genomic classi­ Dagne of the Addis Abeba University for his encouragement fication in Guizotia (Asteraceae). Cytologia 60:67-73 and providing important materials. Special thanks also go to Dr Nei M (1972) Genetic distance between populations. Am Nat Kebebew Assefa, Mr Faris Hailu and Mr Mulatu Geleta who 106:283-292 were always available for help and advice. Nei M (1973) Analysis of gene diversity in subdivided popu­ lations. Proc Natl Acad Sci USA 70:3321-3323 Petros Y, Merker A, Zeleke H (2007) Analysis of genetic References diversity of Guizotia abyssinica (L.f.) Cass, from Ethiopia using inter simple sequence repeat markers. Hereditas 144:18-24 Ajebade SR, Weeden NF, Chite SM (2000) Inter simple se­ Rajesh PN, Sant VJ, Gupta VS, Muehlbauer FJ, Ranjekar PK quence analysis of genetic relationships in the genus (2002) Genetic relationships among annual and perennial Vigna. Euphytica 111:47-55 wild species of Cicer using inter simple sequence repeat Assefa K, Merker A, Teffera H (2002) Qualitative trait varia­ (ISSR) polymorphism. Euphytica 129:15-23 tion in Tef [Eragrostis tef (Zucc.) Trotter]. Euphytica Sudupak MA (2004) Inter and intra species inter simple se­ 127:399—410 quence repeat (ISSR) variations in the genus Cicer. Eu­ Assefa K, Merker A, Teffera H (2003) Inter simple sequence phytica 135:229-238 repeat (ISSR) analysis of genetic diversity in tef (Era­ Tikunov YM, Khrustaleva LI, Karlov GI (2003) Application of grostis tef (Zucc.) Trotter). Hereditas 139:174-183 ISSR markers in the genus Lycopersicum. Euphytica Baagoe J (1974) The genus Guizotia (Compositae): a taxo­ 131:71-80 nomic revision. Botanisk Tidsskrift 69:1-39 Vijayan K, Chatterjee SN (2003) ISSR profiling of Indian Dagne K (1994a) Cytology, phylogeny and oil quality of Gui­ cultivars of mulberry (Morus ssp.) and its relevance to zotia Cass. (Compositae). Ph.D. thesis, Addis Abeba breeding program. Euphytica 131:53-63 University, Ethiopia Yeh FC, Yang RC, Boyl T (1999) Popgene version 1.31. Dagne K (1994b) Meiosis in interspecific hybrids and genomic Microsoft window based freeware for population genetic interrelationships in Guizotia Cass. (Compositae). analysis. University of Alberta, Edmonton Hereditas 121:119-129 Zietkiewicz E, Rafalski A, Labuda D (1994) Genome finger­ DISPAN (1993) Genetic distance and phylogenetic analysis, printing by simple sequence repeat (SSR)-anchored copyright by Tatsuya and the Pennsylvania State Uni­ polymerase chain reaction amplification. Genomics versity 20:176-183 Friis I (1971) A new species of Guizotia (Compositae) from north east tropical Africa. Nor J Bot 18:231-234

Springer I l l Quantitative trait variation of Guizotia abyssinica (L.fCass. Collected from Ethiopia

1Yohannes Petros , Arnulf Merker1, Habtamu Zeleke2 1 Swedish University of Agricultural Sciences, Box 101, Alnarp, Sweden 2 Haramaya University, P.O.Box 138, Dire Dawa, Ethiopia

Abstract

Thirty six accessions o f Guizotia abyssinica (niger in English) collected from eight niger growing regions of Ethiopia and grown at Haramaya University and Hirna trial sites were evaluated for phenotypic diversity. Ten quantitative traits were used to characterize the accessions at the two sites. The first four principal components based on the regional means of each trait accounted for about 91% of the total variation observed among the accessions. Eigenvectors from the first, second, third and fourth principal components axes accounted respectively for about 43, 25, 16 and 7% of the total variation. In the first principal components axis that explained about 43% of the regional variation; plant height, days to 50% flowering, days to flower initiation, number of inflorescence per plant, 1000 seed weight, yield per plot and number of primary branches were the most important variables for grouping the accessions into clusters. The UPGMA dendrogram produced using the plot means of each accessions clearly separated the two main types of niger grown in Ethiopia into two main clusters. It is observed that all the accessions originating from Wollo and Hararghe and one accession from Shewa belong to the early maturing Bunigne niger type while the other accessions originating from the other six regions of Ethiopia are more of Abat type. The prospects of improving the crop are also discussed.

Yohannes Petros, Swedish University of Agricultural Sciences, Department of Crop Science, Box 101, Se-230 53 Alnarp, Sweden, e-mail: yohannes [email protected]

1 Introduction

Guizotia abyssinica (niger in English) is the most common oil crop in Ethiopia. It was more or less unknown in most parts of the world as an oil crop except in Ethiopia and India. Recently, however, it has been reported to be grown in other countries as well including Sudan, Tanzania, Malawi and Uganda (Baagoe 1974). Hiremath and Murthy (1988) indicated the domestication of niger to be earlier than 3000 B.C through selection and cultivation of large achene mutants of G. scabra ssp. schimperi in northern Ethiopia (Murthy et al. 1993). Ethiopia is thought to be the center of origin of this species because none of the wild Guizotia species are known to grow in India (Hiremath and Murthy 1988). Cytological studies also show that G. scabra ssp. schimperi share more karyomorphological similarities with niger than any other member of the genus as well as greater crossability producing fully fertile hybrids (Murthy et al. 1995; Dagne 1994). Hiremath and Murthy (1988) found that the diploid chromosome number 2n = 30 is characteristic of all Guizotia species.

In Ethiopia niger is cultivated mainly in Gojam, Shewa, Gonder and Wellega under rain fed conditions (Getinet and Sharma 1996). It grows best on poorly drained heavy clay soils (Almaw and Teklewold 1995). It is a sh ort day plant with indeterminate growth habit and completely out crossing, a fact evidenced by lack of any seed set when the plants were grown in isolation (Kandel et al. 2004; Prasad 1990). The self incompatibility in niger is of the sporophytic type (Prasad 1990). Honey bees are the primary agents for cross pollination (Ramachandran and Menon 1979). The Ethiopian niger is relatively taller and with greater seed yield than the Indian cultivars (Getinet and Sharma 1996). The extent of branching of the individual plants depends on the seeding rate and the plant population (Kandel et al. 2004). It usually branches profusely when the plants are spaced farther apart than when they are grown closer together. In Ethiopia, niger is well adapted to marginal soils and is highly tolerant to water logging (Adda et al. 1994; Teklewold and Alemayehu 2002). Application of nitrogen fertilizers are known to enhance vegetative growth but does not increase the seed yield (Almaw and Teklewold 1995), though combination of Potassium and Calcium is known to increase the oil content by enhancing the transformation of carbohydrates into fats (Teklewold and Wakjira 2004). Nitrogen fertilizers in excess of 30 kg/ha negatively affects the seed yield by causing the plants to lodge due to extensive vegetative growth (Getinet and Sharma 1996). Adda et al. (1994) pointed out that the inherent low yield of niger cultivars is partly attributed to the self

2 incompatible nature of the crop and lodging and stressed the need to develop shorter plants with stronger stems. It is also indicated by Ramachandran and Menon (1979) that there is considerable reduction in the number of branches and capitulla of the selfed progenies leading to an inbreeding depression of yield.

Today, there is a need to develop high yielding varieties of niger in Ethiopia. One of the factors contributing to the low seed yield in niger is its indeterminate growth habit where the flowering process is not synchronized, ultimately leading to shattering of the seed and yield reduction. An individual plant can have heads at different stages of development. At any one time a plant can have mature heads while some are at the bud stage and some just flowering. So improvement in this regard calls for selecting plants with determinate growth (Teklewold and Wakjira 2004), in order that the flowering as well as maturity would be more synchronous to limit the loss of yield through shattering. Singh and Patra (1989) also indicated the possibility of yield improvement through selection based primarily on seed size and the number of heads/plant. Selection, however, would be more efficient when the indices involve combination of characters rather than being based on a single component trait (Mathur and Gupta 1992). Mathur and Gupta (1992) suggested high relative selection efficiency when the indices consisted of yield, heads/plant and harvest index. There seems to be greater scope for improvement of niger through selection as the plant is inherently variable with respect to many characters such as: the number of days to flowering and maturity, plant height, number of heads per plant, seed size and number of seeds per head (Singh and Patra 1989). The occurrence of three types of niger that vary in the span of their life cycle in Ethiopia: Bunigne niger an early maturing type, Abat niger, a late maturing one and Mesno niger late maturing but frost resistant (Getinet and Sharma 1996), would give the breeder a wider scope for selection. These three types of maturity groups have their own positive and negative attributes. The desirable characteristics ofbunige are its shorter life cycle and the short stature of the plant but it is a low yielder (Pradhan et al. 1995). Abat niger on the other hand is relatively high yielding but it is also late maturing and taller while Mesno niger is frost tolerant but like Abat niger late maturing (Getinet and Sharma 1996).

Recent molecular work on niger has shown that the plant possesses rich genetic variability (Geleta et al. 2007; Petros et al. 2007). The wealth of genetic variability of niger presents a challenge to researchers to exploit this diversity to improve its agronomic quality. Teklewold and Wakjira (2004) indicated that during niger improvement programs, the primary aim of

3 researchers need to be increasing the oil yield per unit area which is also in part a function of its seed yield. This calls for developing a high seed yielding variety coupled with high oil content in the seeds.

Kandel et al. (2004), when describing the niger varieties grown in the United States, indicated the development of an early maturing variety called ‘early bird’. He described this early maturing variety as being short, with more uniform flowering and little or no shattering and therefore producing greater seed yield than the late maturing ones. The early maturing varieties in Ethiopia and India, however, are reported to have lesser seed yield than the late maturing ones (Getinet and sharma 1996; Singh and Patra 1989).

To date, very little work is done regarding improvement in seed yield and oil content of niger. Moreover, more needs to be done to eliminate yield reducing attributes such as indeterminate growth habit, seed shattering, lodging and self incompatibility. It is understood that phenotypic variation is a reflection of the interaction of the genotype and the environment (Vanhala et al. 2004). As improvement of agronomic traits in crop plants depends on variability and heritability of desired characters (Nayakar 1976), the present study focusing on the phenotypic diversity among niger populations grown in Ethiopia is believed to lay a foundation for future scientists who may endeavor to improve the agronomic quality of this little understood oil crop. In this study, we employed principal components and cluster analyses, the utility of which have already been shown by previous workers assessing the diversity in germplasm collection (Assefa et al. 2003; Ayana and Bekele 1999; Hailu et al. 2006; Montagnon and Bouharmont 1996).The present study thus becomes important as phenotypic variability is the basis for future endeavor to improve the agronomic qualities of this crop.

Materials and methods

The Plant Material The plant materials used in the study were collected during November- December 2003 from eight niger growing regions in Ethiopia (Fig. 1). The specimens used in the study include five populations each from Gojam, Shewa, Hararghe, Wollo and Wellega and four populations each from Gonder and Arsi and three populations from Jimma (Table 1, Figure 1). The plants were grown in two experimental stations, Haramaya University (9° 24’N, 42° 2’E) and Hirna (9° 13’N , 41° 6’E) in eastern Hararghe region, Ethiopia, located about 200 km apart. The plants were sown on July 5 and July 7, 2005 in Haramaya University and Hirna respectively.

4 ►

Figure. 1. Geographical map of Ethiopia indicating the regions of origin of the niger samples used in the study.

Simple lattice was used for the experimental design with two replications in each of the field stations. The plants were grown in a 2x4m plots with spacing of 20cm and 40cm between individual plants and rows respectively and lm spacing between the plots. The two replications in each of the two sites were separated by a space of 2 meters. Each plot was designed to carry one hundred plants in ten rows of ten plants each. Thus, three to six seeds were dropped at places 20cm apart in a row at first and later thinned when the plants produced their second true leaves to make the spacing between each plant in a row of uniformly 20cm. Two rounds of weeding by hand were done at both trial sites throughout the growing period. Fertilizer was not applied in both the sites because there was no need for it, as the soils at both sites were sufficiently rich to support the growth of niger and

5 T able 1: Region and site coordinates of the G. abyssinica populations studied

Region Code Locality Site coordinate Altitude(m) Shwa sh-1 Asfachew 9° 53’N, 39° 5’E 1009 sh-2 Kobo 8° 42’N, 38° 15’E 2155 sh-3 Soyama 8° 25’N, 37° 53’E 1900 sh-4 Yaya 9° 42’N, 38° 49’E 2669 sh-5 W orku 9° 45’N, 38° 46’E 2674 Jim m a jm -1 Tiyyo 7° 52’N, 37° 16’E 1760 jm -2 Dacha 7° 53’N, 37° 17’E 1748 jm -3 Ayno 7° 52’N, 37° 17’E 1725 Wellega wg-1 Kane 9° 3’N, 36° 29’E 2176 w g-2 Qawissa 8° 58’N, 36° 29’E 2240 w g-3 Damasa 8° 59N, 36° 30’E 2260 w g-4 Ale 8° 57’N, 36° 29’E 2252 w g-5 Jirata 9° 2’N, 36° 29’E 2142 Arsi Ar-1 Gobessa 7° 36’ 39° 31’E 2374 ar-2 Jelko 7° 26’N, 39° 32’E 2352 ar-3 T areta 7° 35’N, 39° 33’E 2336 ar-4 Shirka 7° 36’ 39° 34’E 2334 W ollo wl-1 Koladi 10° 52’N, 39° 49’E 2374 wl-2 Sedeko 10° 29’N, 39° 56’E 1551 wl-3 Kombolcha 10° 59’N, 39° 46’E 1767 wl-4 Gerado 11° 45’N, 39° 37’E 1913 wl-5 Libso 11° 34’N, 39° 40’E 1662 G onder g n -l Tam o 12° 6’N, 39° 46’E 1894 gn-2 A nguabo 11° 56’N, 37° 48’E 1947 gn-3 Zuria 12° 22’N, 37° 33’E 1942 gn-4 T/H 12° 32’N, 37° 26’E 1895 Gojam g j-1 Kotkotuma 11° 27’N, 37° 14’E 2021 gj-2 Awabel 10° 13’N, 38° 8’E 2466 gj-3 Yabesh 10° 37’N, 37° 31’E 2097 gj-4 Nifasam 10° 17’N, 37° 48’E 2463 gj-5 Rufael 11° 30’N, 37° 24’E 1793 Hararghe ha-1 Bakanisa 8° 53’N, 40° 43’E 1714 ha-2 M akana 8° 54’N, 40° 46’E 1702 ha-3 Kara 8° 52’N, 40° 40’E 1752 ha-4 Haro 8° 52’N, 40° 37’E 1747 ha-5 Bareda 8° 51’N, 40° 38’E 1746

also as niger plants are known to require little fertility levels (Almaw and Teklewold 1995).

In the absence of well established descriptors for niger, the characters that were assumed with some degree of reliability to contribute to the diversity of niger were selected. The traits that were studied are flower initiation, 50% flowering, plant height, heads/plant, seeds/head, number of primary branches, number of secondary branches, yield/plant, yield/plot and 1000 seed weight. Some of these traits were also used by other authors working on niger (Nayakar 1976; Singh and Patra 1989; Pradhan et al. 1995). Flower initiation was recorded when the first flower opens in a given plot and 50% flowering when approximately half of the flower buds open on approximately 50% of the plants on a given plot. Plant height is measured at maturity from the ground level to the tip of the top most bud. Ten sample plants from each plot were randomly selected and all data recorded for all the 10 sample plants for all character in each plot. All statistical treatment of the data was made on the means of the 10 sample plants for each character, the exceptions being flower initiation, 50% flowering and yield/plot which were made on plot basis.

Data Analysis

Patterns of variation among the accessions in both the field experimental sites were analyzed using the NTSYS-pc software package (Rohlf 2000) and the JMP statistical software (SAS institute 2004). An unweighted pair group method using arithmetic averages (UPGMA) dendrogram was constructed for the accessions using the NTSYS software. The NTSYS was also used to depict the regional clustering pattern based on regional means. Phenotypic means of each character were standardized to a mean of zero and standard deviation of unity before subjecting the data to cluster analysis. Principal components analysis was performed using JMP statistical software. In order to study the regional pattern of variation, principal components analysis was performed using the regional means of each trait and eigenvalues and eigenvectors calculated from the principal components axes using the correlation matrix produced for the regional means of each character. Plot means were used for the principal components analysis for the two experimental sites.

Results

The results indicate a wider range of variation among the accessions for days to flower initiation and days to 50% flowering as well as seed size and the number of flower heads per plant. The earliest was an accession from Shewa (sh-1) which took, on the average 51.5 and 53.5 days to flower

7 initiation at Haramaya and Hirna respectively and 69 and 63 days to 50% flowering at Haramaya and Hirna respectively (data not shown).

T able 2. Mean values for the traits used in the analysis for Haramaya and Hima sites and average for the regions population. (Fl-flower initiation, FPF- 50% flowering, PB-primary branch, SB-secondary branch, HPP-heads per plant,PH- plant height, SPH-seeds per head, YPPn-yield per plant, YPPo-yield per plot, TSW-1000 seed weight, HU-Haramaya University, Hir- Hima

Traits W ol G on H ar She Jim W ei Ars Goj FI. H U 57.2 65.1 59.5 64.9 73.7 70.8 65.8 77.7 FI. Hir. 61.6 69.3 62.6 68.1 79.7 73.1 71.1 79.5 Ave. 59.4 67.2 61.1 66.5 76.7 72.0 68.4 78.6 FPF H U 81.2 85.8 82.3 89.2 96.8 95.9 92.4 103.1 FPF Hir. 75.1 88.3 75.4 86.4 97.2 90.7 87.5 100.6 Ave. 78.2 87.0 78.9 87.8 97.0 93.3 89.9 101.9 PB.HU 10.4 13.1 10.9 13.3 13.0 12.6 13.3 13.3 PB .H ir 16.1 15.1 17.1 15.2 18.3 16.4 18.1 16.3 Ave. 13.3 14.1 14.0 14.3 15.7 14.5 15.7 14.8 SB H U 28.6 32.0 29.0 27.7 29.3 29.1 32.8 32.7 SBHir. 76.9 75.0 92.0 60.8 96.7 68.8 83.9 74.2 Ave. 52.8 53.5 60.5 44.3 63.0 49.0 58.3 53.5 H PP H U 120.5 112.4 113.1 110.2 94.3 102.4 111.3 106.4 H PP H ir 248.3 215.6 273.9 215.7 266.2 222.9 261.0 226.5 Ave. 184.4 164.0 193.5 163.0 180.3 162.7 186.1 166.5 PH H U 117.0 132.3 112.9 124.2 132.3 138.7 123.5 136.7 PH H ir 150.5 154.2 147.8 146.4 143.1 162.9 150.6 156.8 Ave. 133.8 143.2 130.3 135.3 137.7 150.8 137.0 146.7 SPH H U 21.0 20.6 23.1 23.5 19.5 22 24.8 22.7 SPH Hir 21.6 19.5 19.7 20.0 22.0 22.8 21.6 22.3 Ave. 21.3 20.1 21.4 21.8 20.8 22.4 23.2 22.5 Y PPn H U 6.0 6.6 6.2 7.1 5.9 7.6 6.6 5.7 Y PPn H ir 4.3 3.8 8.2 6.3 7.1 6.7 6.3 5.5 Ave. 5.1 5.2 7.2 6.7 6.5 7.1 6.4 5.6 Y PPo H U 500.0 433.0 475.0 532.9 489.9 514.9 524.1 440.6 YPPo Hir 532.4 315.8 449.7 437.7 367.1 402.2 374.2 308.4 Ave. 516.2 374.4 462.3 485.3 428.5 458.6 449.1 374.5 T SW H U 3.0 3.5 2.9 3.3 3.0 3.6 3.2 3.2 TSW Hir 2.5 2.9 2.6 2.6 2.5 3.2 2.6 2.8 Ave. 2.7 3.2 2.8 2.9 2.7 3.4 2.9 3.0

The latest of the accessions was a Gojam accession (gj-2) which took on the average about 83 and 109 days for flower initiation and 50% flowering respectively (data not shown). On the average, the Wollo and Hararghe samples started flowering on the 59 and 61 days respectively (Table 2). These were the earliest to begin flowering followed by Shewa, Gonder and

8 Arsi which took approximately 66.5, 67 and 68 days respectively to flower initiation. The Wellega, Jimma and Gojam populations were observed to be late flowering, taking on the average approximately 72, 77 and 79 days to I begin flowering. Among all the accessions, those grown at Haramaya had their flower initiation 3-6 days earlier than those grown at Hirna. Days to 50% flowering also followed a similar trend with Wollo and Hararghe samples taking 78 and 79 days respectively followed by Shewa, Goder and Arsi with 87, 88 and 90 days respectively to flower. The Wellega, Jimma and Gojam accessions took 93, 97 and 102 days respectively to reach the 50% flowering stage. Secondary branches and number of flower heads per plant are consistently higher for the populations grown at Hirna, while seed size is higher for the poplations grown at Haramaya. Plant height is positively correlated with the duration to flowering (data not shown). Thus samples originating from Wollo and Hararghe on the average are > respectively about 134 and 130cm tall while samples from the rest of the regions ranged from 135 cm (Shewa) to 151 cm (Wellega) (Table 2)

The UPGMA dendrogram with Eucledian distance produced two main clusters each of which comprised of several sub clusters (Fig 2). The UPGMA clustering grouped the accessions into two main clusters. Those accessions originating from Wollo and Hararghe formed one cluster and the accessions from the rest of the regions formed a second main cluster while two accessions from Shewa (sh-1 and sh-3) and one accession from Gojam (gj-2) were not grouped in either of the two main clusters. The first cluster consisted wholly of Wollo and Hararghe samples with only one sample from Shewa (sh-2) occurring among them. The accessions originating from Wollo and those from Hararghe more or less separated from each other in the sub clusters except for one Hararghe accession (ha-5) which is closely placed with an accession from Wollo (wl-2). On the other hand, it is evident that the second main cluster did not discriminate among the accessions from the different regions of origin. Here one of the sub clusters consisted mainly of samples from Gonder, Wellega and Arsi while the second sub cluster comprised mainly of samples from Gojam and Jimma.

i

9 ------w H ____I «t-2 1 ha-5 ------sh-2 ------\\t-4 ------wi-5 ------ha-1 ----- wK? ------ha-3 ------ha-4 ------ha-2 ------gn-1 ------gii-2 ------ar-4 ------sh-5 ------gn-3 ------wg-2 ------g H ------wg-3 J-----"g -5 '----- ar-1 ------ar-2 ------gn-4 ------"g-4 ------sh-4 ------gj-3 ------jm-2

I------.------1------.------1------1------.------1------1------1------1------>------1------1------1------1------1------1------1------> -sh-1 | 1.96 1.65 L34 1.03 0.73 Distance

Fig. 2. UPGMA Clustering Pattern of the niger populations used in the study

The UPGMA clustering that was produced using the regional means for each of the characters also grouped the regions into two main clusters (Fig. 3). Wollo and Hararghe are grouped in one cluster while Gonder, Shewa, Jimma, Wellega, Gojam and Arsi are placed in the other cluster. In the second regional cluster, Jimma remained solitary only distantly connected to the sub cluster containing Wellega, Gojam and Arsi. The Wellega and Gojam populations seem to be more similar maintaining the least distance between them than with the rest of the regions (Fig 3).

On the basis of the regional means for each trait, the first four principal components accounted for approximately 91% of the total variation observed among the niger populations grown in the eight niger growing regions of Ethiopia (Table 5). Eigenvectors from the first, second, third and fourth principal components axes accounted for about 43, 25, 16 and 7% of the total variation respectively.

10 T ab le 3. Eigenvalues and eigenvectors for the first four principal components for the G. abyssinica populations grown at Haramaya T raits EIGENVALUES PCI P C ” PC3 PC 4 Flower initiation 0.40826 -0.32481 -0.15045 0.13572 50% flowering 0.42717 -0.21457 -0.20339 0.13702 Primary branch 0.33310 -0.25877 0.38657 0.01447 Secondary branch 0.21707 -0.19625 0.59855 0.25113 Heads per plant -0.27606 0.13857 0.59133 0.15656 Plant height 0.44704 0.04586 -0.03608 -0.14837 Seeds per head 0.17959 0.3687 -0.02187 0.6935 Yield per plant 0.26932 0.51490 0.04964 -0.09758 Yield per plot 0.24275 0.54956 -0.08160 0.05552 1000 seed weight 0.23374 0.13628 0.27209 -0.68490 Eigenvalue 3.7747 1.9113 1.3684 1.0050 % of variance explained 37.747 19.113 13.684 10.050 % o f total variance 37.747 56.859 70.543 80.593

It is also observed from the relative magnitude of the eigenvectors from the first principal components axis, based on the regional means that: plant height, days to 50% flowering, days to flower initiation, number of heads per plant, and seed size (1000 seed weight) were the most important characters contributing to the variation observed among the niger populations studied. Plant height, days to 50% flowering, days to flower initiation, number of primary branches, number of flower heads per plant and yield per plant were also the most

T able 4. Eigenvalues and eigenvectors for the first four principal components for the G. abyssinica populations grown at Hirna

Traits Eigenvectors ------£2------P C I PC2 PC3 PC4 Flower initiation -0.19783 0.57839 -0.04673 -0.21270 50% flowering -0.28268 0.53423 -0.07573 -0.15902 Primary branch 0.36826 0.34733 0.22930 -0.05548 Secondary branch 0.46935 0.18810 -0.09318 0.34915 Heads per plant 0.51078 0.10407 0.24920 0.12159 Plant height -0.2776 0.04440 0.28023 0.76616 Seeds per head 0.0227 0.21797 0.63541 -0.16667 Yield per plant 0.30169 0.07451 -0.05532 -0.08644 Yield per polt 0.02365 -0.39648 0.45527 -0.38721 1000 seed weight -0.30627 0.03127 0.41966 0.13273 Eigenvalue 2.8701 2.3810 1.4051 0.9641 % of variance explained 28.701 23.810 14.051 9.641 % o f total variance 28.701 52.510 66.561 76.202

11 T able 5. Eigenvalues and the magnitude of eigenvectors in the principal components axes for each of the characters of G. abyssinica populations grown in Ethiopia based on regional means.

Eigen vectors Traits P C I PC2 PC3 PC 4 PC 5 Flowerinitiation 0.42091 0.25605 -0.04934 -0.08518 -0.27736 50% flowering 0.43284 0.22193 -0.00036 -0.17869 -0.26064 Primary branch 0.24060 0.49797 0.11301 0.01207 -0.12200 Second, branch -0.11258 0.53326 -0.28162 0.22501 0.33540 Heads per plant -0.36304 0.37420 0.00691 -0.3397 0.37908 Plant height 0.43587 -0.16033 0.02159 0.04383 0.31748 Seeds per head 0.11596 0.14238 0.60427 -0.52933 0.43104 Yield per plant -0.00730 0.19449 0.57923 0.69055 -0.16471 Yield per plot -0.33829 -0.12291 0.43828 -0.14349 -0.29137 1000 seed wt. 0.33941 -0.33750 0.11061 0.35962 0.43063 Eigenvalue 4.3210 2.4819 1.6039 0.7077 0.5090 % variance 43.210 24.819 16.039 7.077 5.090 explained % of total variance 43.210 68.028 84.068 91.144 96.234 important variables for the variation observed among the niger populations grown at Haramaya experimental site (Table 3). From the second principal components axis, on the other hand, secondary branches, primary branches, heads per plant as well as seed size were the most important variables for the region’s populations, which same traits become the most important variables of the first principal components axis for the niger populations grown at Hirna experimental site (Table 4). Seeds per head, yield per plant and yield per plot were also important classification variables for the regions’ populations in the third principal component axis, while heads per plant, secondary and primary branching at Haramaya and seeds per head, yield per plot and thousand seed weight at Hirna are also important classification variables in the third principal components axis. The distribution of the accessions in a two dimensional space (fig. 4) and three dimensional space (fig. 5) based on the first two and the first three principal components respectively also depict the separation of the Wollo and Hararghe accessions from those originating from the rest of the regions.

12 Discussion

The results obtained from both the Haramaya and Hirna trial sites indicate that there is a significant variation among the niger populations grown in the different regions in Ethiopia with respect to the characters studied. All accessions from Wollo and Hararghe are sufficiently different from the niger accessions collected from the rest of the six regions. This variation in the quantitative traits holds regardless of the difference in the location of the two trial sites. The Hararghe and Wollo accessions are early maturing, shorter plants producing on the average more inflorescence per plant. It appears that the number of inflorescences per plant depends on the number of secondary branches and these two traits are significantly positively correlated (data not shown). Even though the Hirna populations produced more flower heads than those grown at Haramaya (Table 2), the plot yield of the niger population grown at Haramaya is higher than those at Hirna. This anomaly can be explained by the seed size of the Haramaya samples which were consistently larger than those of Hirna. Sterility and loss of seed through shattering may be other likely reasons for the reduction in the yield of the Hirna samples.

w g

1J1 Coefficient

Fig. 3. UPGMA clustering pattern for the regions of origin.

13 The UPGMA clustering patterns both for the accessions and the regions clearly followed the assertion that there are different types of maturity groups of niger growing in Ethiopia (Getinet and Sharma (1996). This is also in conformity with an earlier study made on the material using ISSR markers (Petros et al. 2006). While the clustering pattern for the accessions confirm the existence of different types of niger in Ethiopia, it did not, however, discriminate among the same maturity group originating from different regions. This is particularly so when it comes to the second major cluster (fig 2). The late maturing Abat niger is clearly separated from the early maturing Bunigne niger by forming a separate major cluster. But the Abat accessions originating from the different niger growing regions in Ethiopia occur intermingled with each other in the second cluster of the dendrogram in Fig 2. This may be explained in part by gene flow due to transfer of materials among the niger growers of neighboring regions. In the UPGMA clustering for the regions Wollo and Hararghe form a separate cluster from the others. Actually it is reported that these regions grow the early maturing Bunigne niger (Getinet and Sharma 1996). Though these two regions are not contiguous, it may be assumed that the farmers of the Hararghe highlands might have obtained the seeds from Wollo or the vice versa. It is also observed (data not shown) in both the Haramaya and Hirna experimental sites that all the samples that originated from Wollo and Hararghe flowered much earlier than all the rest of the samples from the other regions with the exception of one sample from Shewa (sh-1) which proved to be the earliest to flower in both sites. In the cluster depicting the regional pattern of variation (Fig. 3), the accessions obtained from Jimma, though of the Abat type seem to be distant from the other Abat accessions from the other regions even though they were comfortably placed with the others in the dendrogram depicting the pattern of variation for the accessions though non of the Jimma accessions paired with any of the other (fig. 2).

14 Fig. 4.Two dimensional distribution in space of the niger accession collected from eight niger growing regions of Ethiopia.

On the other hand the accessions from Wellega and Gojam seem to be more similar. This similarity can be discerned from the longer duration they took to flower, their larger seed size and taller plant stature as compared to the accessions from the other six regions.

The variation occurring within a region’s populations as well as the close similarity of populations of different regions could be explained by gene flow through material transfer among the fanners of neighbouring regions and neighbouring niger growing areas which might have been the case for niger populations of Wellega, and those grown in Gojam, that appear to be more related to each other than to any of the populations of other regions (Fig 3). The Jimma populations, however, are relatively only distantly related to the accessions of the other regions. Though the niger grown in the region is of the Abat type, Jimma, However, is not paired with any one of the Abat growing regions of Ethiopia though it is distantly paired with the cluster containing Wellega, Gojam and Arsi. This indicates that though the niger populations of Jimma belong to the Abat type, there exists variation that might have been built up as the result of centuries of isolation merely due to its geographical location that restricted the gene flow to and from the neighbouring niger growing regions. This fact is also supported by other studies (Petros et al. 2007). Though Jimma is located adjacent to Wellega and Shewa, the niger growing areas of Jimma are, however, not contiguous with the niger growing areas of these two regions as geographical barriers in the form of two massive valleys, the Diddessa valley and the Ghibe valley separate them from the niger growing areas of these two adjacent regions.

0.91

Fig. 5. Three dimensional depiction of the clustering pattern of the niger accessions used in the study

Niger populations growing in Ethiopia harbour rich genetic variation. The variability observed in the present study could be explained primarily by the inherent variation between the two main types of niger, a late maturing

16 Abat niger and an early maturing Bunigne niger. The principal components analysis based on the regional means for the traits under investigation (table 5) also revealed that plant height, days to flower initiation, days to 50% flowering and seed size were the most important traits for the classification of these accessions into clusters. In fact, the variation observed between the Bunigne niger and Abat niger relate to plant height, duration to maturity, and seed size (Getinet and Sharma 1996). Thus individual Bunigne samples in the present study are early maturing short plants, producing more flower heads than the Abat types but not superior to Abat niger in yield. As self incompatibility, lodging, indeterminate flowering and seed shattering are the major problems in niger cultivation today, it is indicated by earlier investigators that improvement in this regard calls for the development of niger varieties that are early maturing with short and strong stems to avoid lodging, and that flower synchronously to avoid or minimize seed loss through shattering and with increased seed and oil yield per unit area (Adda et al. 1994; Teklewold and Wakjira 2004). In this regard the niger populations growing in Wollo and Hararghe would be good candidates as starting materials to develop early maturing plants with shorter stems. Moreover, as the earliest accession in this study is one from NE Shewa (sh.l), we recommend that future investigators also focus on the niger populations of that area to develop very early varieties. However, niger cultivation has a lot of problems and a lot remains to be done to develop among other things self compatible and high oil yielding varieties.

Acknowledgements

The authors would like to express their sincere gratitude to Dr. Adugna Wakjira of Holeta Research Station for providing the parameters used for the field evaluation of niger and to Mr. Faris Hailu for his unreserved help in the analysis o f the field data.

References

Adda, S., Reddy, T. P .and Kishor, P. B. K. 1994. Somatic embryogenesis and organogenesis in Guizotia abyssinica. In Vitro Cell Dev Biol 30P: 104-107. Almaw, G. and Wold, A. T. 1995. An agronomic and seed quality evaluation of noug (Guizotia abyssinica Cass.) germplasm in Ethiopia. Plant Breeding 114: 375-376. Assefa, K., Merker, A. and Tefera, H. 2003. Multivariate analysis of diversity of tef (Eragrostis tef (Zucc.) Trotter) germplasm from western and southern Ethiopia. Hereditas 138: 228- 236.

17 Ayana, A. and Bekele, E. 1999. Multivariate analysis of morphological variation in Sorghum (Sorghum bicolor (L.) Moench) germplasm from Ethiopia and Eritrea. Genet. Resour. Crop Evol. 46: 273-284. Baagoe, J. 1974. The Genus Guizotia (Compositae): A taxonomic revision. Botanisk Tidsskrift 69: 1-39. Dagne, K. 1994. Meiosis in interspecific hybrids and genomic interrelationships in Guizotia Cass. (Compositae). Hereditas 121: 119-129. Geleta, M., Bryngelsson, T., Bekele, E. et al. 2007. Genetic diversity of Guizotia abyssinica (L.f) Cass. (Asteraceae) from Ethiopia as revealed by random amplified polymorphic DNA. Genet. Resour. Crop Evol. 54: 601-614. Getinet, A. and Sharma, S. M. 1996. Niger. Guizotia abyssinica (L.f) Cass. P rom oting the conservation and use of underutilized and neglected crops. 5. International Genetic Resources Institute. Gatersleben, Germany. Hailu, F., Merker, A., Singh, H. et al. 2006. Multivariate analysis of diversity of tetraploid wheat germplasm from Ethiopia. Genet. Resour. Crop Evol. 53: 1089-1098. Hiremath, S. C. and Murthy, H. N. 1988. Domestication of niger (Guizotia abyssinica). Euphytica 37: 225-228. Kandel, H. J., Porter, P. M., Johnson, B. L. et al. 2004. Plant population influences niger seed yield in the Northern Great Plains. Crop Sci. 44: 190-197. Mathur, R. K. and Gupta, S. C. 1992. Discriminate function analysis in niger (Guizotia abyssinica Cass.). Crop Res. 5(1): 164-165. Montagnon, C. and Bouharmont, P. 1996. Multivariate analysis of phenotypic diversity of Coffea Arabica. Genet. Resour. Crop Evol. 43: 221-227. Murthy, H. N., Hiremath, S. C. and Salimath, S. S. 1993. Origin, Evolution and genome differentiation in Guizotia abyssinica and its wild species. Theor. Appl. Genet. 87: 587- 592. Murthy, H. N., Hiremath, S. C. and Pyati, A. N 1995. Genomic classification in Guizotia (Asteraceae). Cytologia 60 : 67-73. Nayakar, N. Y. 1976. Genetic variability and heritability of six quantitative characters in niger (Guizotia abyssinica Cass.). M ysore J Agric Sci 10: 553-558 Petros, Y., Merker, A. and Zeleke, H.(2007. Analysis of Genetic Diversity of Guizotia abyssinica (L.f.) Cass, from Ethiopia using Inter Simple Sequence Repeat markers. Hereditas 144: 18-24 Pradhan, K., Mishra, R. C. and Paikary, R. K. 1995. Genetic variability and character association of niger. Indian J. Genet. 55(4):457-459. Prasad, V. 1990. Pollen tube growth and site of incompatibility reaction in niger (Guizotia abyssinica Cass.). Current Science 59: 466-468. Ramachandran, T. K. and Menon, P. M. 1979. Pollination mechanism and inbreeding depression in niger. Madras Agric. J. 66(7): 449-454. Rohlf, F. J. 2000. NTSYSpc Numerical Taxonomy and Multivariate Analysis System. Version 2.1. Exeter software, New York, USA. SAS Institute. 2004. JMP 5.1.1 Users Guide. SAS Institute Inc. Cary, NC, USA. Singh, B. and Patra, G. J. 1989. Character association and path coefficients of quantitative traits in niger (Guizotia abyssinica Cass.) Indian J. Agric. Sci. 59(7): 442-445.

18 Teklewold, A. and Wakjira, A. 2004. Seed filing and oil accumulation in noug (Guizotia abyssinica (L.f.) Cass. Sinet: Ethiop. J. Sci. 27(1): 25-32. Teklewold, A. and Alemayehu, N. 2002. Studies on the floral characteristics of niger seed (Guizotia abyssinica (L.f.) Cass.). Journal o f Applied Botany 76: 163-167. Vanhala, T. K., Van Rijn, C. P. E., Buntjer, J. et al. 2004. Environmental, phenotypic and genetic variation of wild barley (Hordeum spontaneuni) from Israel. Euphytica 137: 297- 309.

19 IV Developing high oleic acid Guizotia abyssinica (L.f) Cass, by plant breeding

Yohannes Petros1 , Anders Carlsson1, Sten Stymne*, Habtamu

Zeleke2, Ann-Sofie Fait1 Arnulf Merker1

Swedish University of Agricultural Sciences, Box 101, Alnarp, Sweden ‘ Haramaya University, P.O.Box 138, Dire Dawa, Ethiopia Corresponding author, E-mail: [email protected]

Abstract

Oleic acid content was increased from approximately 5-11% to 80-86% in Guizotia abyssinica (L. f.) Cass, materials from Ethiopia after repeated selection and breeding. Achenes collected from Ethiopia were screened for elevated oleic acid content by half seed technique. The starting materials for breeding were nine plants selected from among 272 seeds analyzed and having an average oleic acid content of approximately 21% which is considered to be high compared to the 5-11% reported earlier for niger materials of Ethiopian origin. It was observed that the oleic acid content steadily increased after each round of selection and breeding. The percent oleic acid in the seed oil increased to an average of 35.2% and 53.4% after the first and second round of breeding respectively and ultimately to over 80% after the third round of breeding. It was also observed that the percent oleic acid in the oil stabilizes and the plants breed true when the oleic acid content in the parental seeds was above 79%.

Key words: Guizotia abyssinica — niger — high oleic acid.

1 Introduction

Guizotia abyssinica (niger in English) is a little known oilseed crop. The major niger growing countries of the world are Ethiopia and India. Niger is also reported to be grown in West Indies, East Africa and the United States (Kandel and Porter 2002). In Ethiopia 50-60% of the edible oil requirement of the country is met by niger seed oil, whereas only about 2% of the edible oil requirement is provided by niger seed in India (Riley and Belayneh 1989, Dutta et al. 1994, Hiremath and Murthy 1988).

The oil content of niger is variously reported as 29-39% (Dutta et al 1994), 30-35% (Kandel and Porter 2002) and 42-44% (Dagne and Johnsson 1997). Dutta et al. (1994) reported that the Ethiopian niger seed oil contains more than 70% linoleic acid, whereas, Dagne and Johnsson (1997) reported 66- 69% linoleic acid. In all the works so far done on the fatty acid composition of niger, linoleic acid is unequivocally the dominant fatty acid present in niger seed oil followed by palmitic, oleic and stearic acids (Dutta et al. 1994, Ramadan and Morsel 2003, Dagne and Johnsson 1997). The percentage of oleic acid in the Ethiopian niger seed oil was reported to be in the range of 6-11% (Dutta et al. 1994), 5.4-7.5% (Dagne and Jonsson 1997). It is indicated that the oil content and the fatty acid profile may vary depending on the origin of the material and the maturity level of the seeds (Riley and Belayneh 1989, Stymne and Appleqvist 1980).

The quality of oil and its suitability for a particular purpose, be it for industrial use or for human consumption depends on the proportion of the different fatty acids it contains. Oils where linoleic acid is the predominant fatty acid are reported to have poor shelf life whereas those with high oleic acid content are more stable. For cooking oils, it becomes imperative that oils be suitable for the kind of cooking they are intended for (Mugendi et al. 1998, Warner and Knowlton 1997, Fehr 2007). It is also reported that the presence of high proportions of linoleic acid in oils positively contributes to the frying flavor intensity, a fact which also renders its keeping quality poorer (W arner et al. 1997). Thus, in deep fat frying, which has now become a common way of preparing food, the oil should be optimal for the frying flavor intensity of the food and the frying time of the oil (Mugendi et al. 1998, Warner and Knowlton 1997, Fehr 2007).

The high oleic sunflower oil with about 78% oleic acid is shown to have greater frying stability of the oil and oxidative stability of the food (Warner et al. 1997). Warner et al. (1997) also reported that there is significant positive correlation between increase in the level of linoleic acid in the oil and decreasing oxidative stability of the food. Because the oxidative stability of an oil and its shelf life have an inverse relationship to the amount of polyunsaturated fatty acids it contains, it becomes necessary to lower the degree of unsaturation in order to decrease the oxidative rancidity of oils. For the most part this is done by blending the oil with one having high oleic acid content or by hydrogenation. (Mugendi et al.1998, Burton et al.1983). Hydrogenation of oils, although it increases the thermal stability and resistance to atmospheric oxidation, is however, known to produce unwanted positional isomers (Tompkins and Perkins 2000, Wilson and Rinne 1976, Fehr 2007). Trans-esterification of oils by hydrogenation is indicated to pose adverse health effects, and recently attempts have been made to genetically modify the degree of unsaturation in oils through genetic engineering (Kinney 1994, Chapman et al.2001). It has been shown that oils whose oleic/linoleic ratio has been modified through genetic engineering exhibited greater frying stability (Warner and Knowlton 1997, Tompkins and perkins 2000). Genetically engineered food crops, however, are not appealing for the most part to the public and its use for human consumption is still controversial in many countries around the world. Thus, to circumvent the ethical, public health as well as economic problems presented by the chemical and genetic modification of oils, modification of the proportion of the fatty acids towards the desired composition by plant breeding remains the best alternative to date. To this end, several oil crop varieties have been developed by plant breeders including high oleic acid sunflower, High oleic acid safflower and high oleic acid soybean (Urie 1985, Fuller et al. 1996, Wilson and R inne 1976, Burton et al. 1983). As both oleic acid and linoleic acid are produced by the same desaturation pathways of 18:1 to 18:2 and 18:2 to 18:3 (Stymne and Appleqvist 1980, Voelker and Kinney 2001), modifying the fatty acid composition of niger by reversing the oleic/linoleic ratio towards elevated proportion of oleic acid in the seed by plant breeding is believed to increase the oxidative stability of the oil. It is also envisaged that developing high oleic varieties of niger is indeed an important objective as it is reported that intake of diet with high oleic acid in the oil would reduce the low density lipoprotein cholesterol in b lo o d plasma in addition to increasing its shelf life and oxidative stability which same characters are the most sought after in developing any crop seed for deep fat frying (Takagi and Rahman 1996). Increasing the percentage of oleic acid in the oil of niger would inevitably lead to the reduction in the percentage of linoleic acid in the seed oil, as there is an inverse relationship in the inheritance of these two fatty acids in niger.

3 The present work is an attempt to increase the percentage of oleic acid in niger seed oil through selection and breeding. We are reporting, in the present paper, an increase of oleic acid >80% of the total fatty acid in niger seed oil not heretofore reported. It is believed that the study would contribute to the efforts already underway to enhance some of the desirable qualities in niger by plant breeding.

Materials and Methods

The plant material

The plant materials were collected from niger growing regions in Ethiopia during November and December 2005. Seeds from single plants were collected in farmer’s fields. Bulked seeds from each of 87 plant progenies from 78 fields in four regions were analyzed for content of oleic acid. In the progenies of nine plants with elevated levels of oleic acid, the half seed method was used. The nine individual seeds with the highest levels of oleic acid (Table 1) were planted and used in further crossing. The breeding program was initiated in the Autumn of 2006 with these nine plants; four plants from Wellega, three plants from Jimma and two plants from Gojam.

Table 1. Region of origin, site coordinates and the percentage composition of oleic acid of the niger materials collected from Ethiopia.

C ode R egion Site coordinates % oleic acid W45-1 Wellega 9° 22’N, 36° 24’E 22 W45-2 Wellega 9° 22’N, 36° 24’E 30.6 W45-3 Wellega 9° 22’N, 36° 24’E 29.6 W 46 W ellega 9° 22’N , 36° 24’E 17 G37 Gojam 11° 31’N, 37° 27’E 16.8 G65 Gojam 11° 41’N, 37° 28’E 17.4 J6-1 Jim m a 7° 45’N, 37° 11’E 20.2 J6-2 Jimma 7° 45’N, 370 ll’E 19.2 ...... Jimma 7° 45’N, 37° ll’E 17.9 The screening procedure

Seeds from the various accessions were soaked in water on filter paper in Petri dishes. The husk was removed to expose the embryo. The cotyledons were cut in half and the part with the miniature embryo left on the Petri dish while the other half is analyzed for its oleic acid content by gas chromatography. Where the oleic acid content of the half seed turns out to be high, the other half seed is planted in the green house. Screening of the seeds for high oleic acid was done after harvest of each round of breeding.

Breeding procedure

The design of the experiment was a simple one, the objective being to obtain niger progeny that are breeding true for high oleic acid. During the first round of breeding (November 2006 to March 2007), half seeds known to have relatively high oleic acid were planted in the Biotron at SLU, Alnarp, the day time temperature adjusted to 25°C and the night temperature adjusted to 18°C. Cross pollination was carried out by dusting the pollen from a plant onto the inflorescence of another plant. Mature inflorescences from all the plants were harvested in March 2007 and screened for the high oleic trait by the half seed technique. Only seven individuals which showed markedly high oleic acid percentage were selected for planting in the second round of breeding (April 2007 - Sept 2007). These had percent oleic acid content of 57-83% (Table 2). Crosses were made in all possible combinations including reciprocal crosses in all instances. These were harvested in September 2007 and the screening for the high oleic phenotype resumed by the half seed technique. Only those known to have high percentage composition of oleic acid (> 79%) were planted for the third round of breeding (October 2007 — March 2008). These were harvested in March 2008 and analyzed the same way for the oleic acid content.

Analytical procedure

For the bulk materials samples containing 18 seeds were homogenized in 3.75ml Methanol:Chloroform (v/v 2:1) and lml of 0.15M acetic acid, and the homogenate transferred into a screw cap tube. The homogenizer was rinsed with 1.25ml chloroform and transferred to the screw cap tube, to which was added 1.25ml water. This was thoroughly mixed, centrifuged

5 and the bottom phase taken. This was evaporated under Nitrogen and methylated with 2ml of 0.1M Sodium Hydroxide in methanol and heated for 15 minutes at 90°C. To this was added, 3ml hexane, 2ml water and lOOfil of the internal standard, methyl heptadecanoate and mixed thoroughly before centrifugation for 3 minutes at 2000 rpm. Seventy microliters of the top hexane phase was transferred to the GC vials. The half seeds were directly methylated by the addition ofO.lM Sodium Hydroxide in methanol and the fatty acid methyl esters extracted with hexane.

Separation of the fatty acid methyl esters was done by gas chromatography (Shimadzu GC-17A, Kyoto, Japan) with a flame ionization detector (FID). A WCOT fused Silica capillary column (50m x 0.32mm), CP-wax 58 (FFAP)-CB (Cromopack, Middelburg, the Netherlands) was used. The temperature was held at 160°C for 0.5min and increased at a rate of 3°C/min to 250°C and held there for 0.5min before coming to 265°C at a rate of 25°C/min and held there for 2.4minutes. The injection port and the detector were held at 275°C and 280°C respectively.

Results

The oleic acid percentage in the total oil was observed to steadily increase after each round of selection and breeding. The starting material consisting of nine plants had an average oleic acid content of around 21% with the individual seed’s oleic acid content ranging from 17-30.6% (Table 1). It was observed that after the first round of breeding the average oleic acid content in the progeny has increased to about 35.2%. The individual seeds from these crosses, however, exhibited wide range of variation from 4-83% (Table 2.A). Whereas only 1% of the seeds analyzed after the first round of breeding had oleic acid content >80%, after the second round of breeding, however, more than 20% of the seeds were observed to have oleic acid percentage >80. The progeny seeds from all possible combinations of crosses including reciprocal crosses in the second round of breeding showed a wide range of variation with respect to the oleic acid content of individual seeds. Nevertheless, the proportion of progenies with high oleic acid content significantly increased and the average oleic acid content of seeds was found to be 53.2%, though the oleic acid content of individual seeds ranged from 14% to 86.3% (Table 2.B).

6 Table 2. Mean values and range of the major fatty acids in niger obtained after the first (A), second (B) and third (C) round of breeding.

Percentage composition of the major fatty acids

Palmitic Stearic Oleic Linoleic N o. Crosses seeds Mean Range Mean Range Mean Range Mean Range

A 16.8x17 27 6.6 5.5-7.7 4.7 3.6-6.2 38.8 22.1-53.2 47.5

.9 32.6-64.9

17x17.4 21 7.1 5.8-8.5 5.7 3.5-6.6 42.3 16.2-83.0 42.0 4.0-69.7 19.2x22 5 6.1 5.1-6.3 4.5 4.1-5.2 43.5 26.8-62.7 43.6 24.9-59.9

22x20.2 21 7 5.7-8.6 4.7 3.2-6.2 29.2 5.8-59.8 56.6 25.8-78.2

29.6x30 29 7.3 5.9-10.2 6.3 4-9.8.0 30.1 4.0-57.4 53.5 25.1-77.8 .6

B 83x62 15 6 4.8-7.5 4.4 2.9-6.3 64.8 39.3-86.3 22.8 2.7-49.7 83x61 39 7 5.3-8.6 5.1 3.6-6.9 39.8 15.9-70.3 46.2 15.0-70.3

61x62 3 6.4 5.4-7.8 6.3 5.5-7.3 40.5 19.6-54.5 44.6 31.7-64.7

61x59 13 6 4.7-6.8 6.1 3.9-7.1 52.0 25.0-84.6 33.7 4.5-48.0 61x57 8 7.2 6.3-8.6 5.9 4.7-8.5 45.5 21.6-65.5 39.1 19.5-59.6

83x58 30 6 4.5-7.3 4.9 3.5-6.4 67.9 48.4-83.8 19.1 3.4-38.2 83x57.7 23 6.8 5.1-8.6 4.1 2.5-5.2 60.5 28.5-84.6 26.9 3.3-57.9

57.7x62 9 6 5.4-6.8 4.3 3.2-5.3 46.7 24.6-61.2 41.1 25.8-63.0 61x57.7 14 7.3 6.0-8.8 5.4 4.4-6.9 38.6 14.0-65.4 46.6 19.5-71.5 58x57.7 13 6.2 4.7-7.3 5.4 3.9-6.6 55.7 25.2-84.3 30.9 4.5-60.1 57.7x57 32 5.9 4.0-7.3 5.0 3.3-7.1 51.9 19.6-86.1 35.5 3.7-67.6 83x57 14 5.8 4.6-7.7 4.6 3.2-6.2 65.4 30.8-85.4 22.3 3.6-55.9

C 82x85 16 5.2 4.6-5.9 3.4 2.9-4.2 83.6 77-87 6.0 2.9-12.2 85x83 8 4.9 4.6-5.6 3.4 2.3-4.7 83.1 81.1-85.2 6.6 3.3-8.9 84x85 6 5.3 4.9-5.8 4.1 3.2-4.7 82.7 81.5-84.8 5.9 3.9-8.2

86x84 5 5.6 5.2-6.3 3.8 3.1-5.2 81.8 78.1-83.9 6.9 3.9-11.0 85x85 2 5.2 5.1-5.3 3.5 3.1-3.8 84.3 83.6-85.0 5.3 5.1-5.5 82x83 3 5.6 5.2-5.9 3.9 3.4-4.5 80.9 78.1-82.5 7.4 5.3-9.7 82x86 3 4.6 4.1-5.3 3.5 2.8-4.3 85.2 83.6-86.3 4.7 3.7-6.4 81x79 2 4.3 3.9-4.6 3.4 2.8-3.9 82.0 79.7-84.2 9.0 7.9-10.1 82x79 3 5 4.7-5.2 3.5 3.2-3.7 83.2 82.9-83.8 6.6 6.2-6.9 79x85 3 5 4.9-5.2 3.6 3.1-4.1 83.4 82.8-84.2 6.1 5.0-6.7

79x83 2 4.1 3.9-4.3 4.1 3.6-4.5 82.3 81.9-82.6 7.4 6.4-8.3

7 No significant variation was observed between the reciprocal crosses. All materials selected for the third round of breeding had oleic acid content in the range of 79-86%.There has been a dramatic increase in the average percent oleic acid after the third round of breeding. As expected, the vast majority of the progeny seeds from this round of breeding contained oleic acid >80% (Table 2.C). It was observed that progeny seeds obtained from parents having >79% oleic acid content were highly stable and yielded progeny seeds with >80% oleic acid content. It was also noted that as the oleic acid content in the seed oil of niger increased from approximately 8.4% in the wild type niger collections (data not shown) to approximately 80 — 86% in the high oleic material developed, there has been a corresponding decrease in the content of palmitic acid from 7.9% to 5.3% which has a significant negative correlation value of -0.994 (P 0.006). The chromatograms for the wild type niger materials from Ethiopia and that for the high oleic acid niger developed in this study are presented in figure 1.

8 Figure 1. Chromatographic depiction of the fatty acid profile of niger seed oil. A- wild type niger. B- high oleic acid niger.

Discussion

The present study shows that oleic acid content in the seeds of niger is heritable and can be increased by repeated selection and breeding. It was observed that parental plants whose oleic acid content is approximately >79% are true breeding for the high oleic trait, whereas those with oleic acid content < 70% yield all ranges of oleic acid content (Table 2). Presently, a study is underway to determine the inheritance of high oleic trait in niger. From our present study, it may not be possible to point out what changes in the biochemical pathway led to the accumulation of 18:1 in

9 the seed oil of niger but Okuley et al. (1994) indicated that mutation at the FAD2 locus of Arabidopsis led to the reduction in the production of polyunsaturated fatty acids. As there can be all combinations of triacylglycerol (TAG) molecular species in niger seed oil namely, trilinolein, monooleyl-dilinolein, dioleyl-monolinolein and triolein in the original starting materials with relatively high oleic acid content relative to the common niger materials, it is thought that selection and breeding increased the oleic acid moieties in the TAG and ultimately rendered triolein to be the predominant TAG molecular species in the oil. Thus, selection for high oleic acid seeds and breeding led to the increase of the percent oleic acid on the average after every round of breeding until the oleic acid content stabilized at oleic acid levels of >80%. Increase in oleic acid content in niger inevitably leads to proportional decrease in the content of PUFA’s particularly linoleic acid. This is also the case for many of the oil crops including Brassica napus (Schnurbusch 2000) and cottonseed (Chapman et al. 2001). Schnurbusch (2000) indicated that increase in the palmitic acid content of Brassica napus is accompanied by decrease in oleic acid and the oil content. Whether increase in the oleic acid content of niger is accompanied by an increase in its oil content needs to be investigated further.

It has been reported that the abundance of PUFA’s particularly linoleic acid enhances frying flavor intensity of foods due to the presence of volatiles like 2,4-decadieenal which is the oxidation product of linoleic acid whereas an increase in the content of oleic acid in oils increases the stability of the oil (Warner et al.1997). The high oleic materials of niger developed in the present study presents ample opportunity for breeders to develop plants with optimal fatty acid composition for healthy human consumption by maintaining a balance between oleic and linoleic contents of the oil to increase the stability of the oil without unduly compromising its frying flavor intensity.

References

Burton, J.W., R.F. Wilson and C.A Briml983. Recurrent selection in Soybeans. IV. Selection for increased oleic acid percentage in seed oil. Crop Sci. 23:744-747. Chapman, K. D., Austin-Brown, S., Sparace, S. A., Kinney, A. J., Ripp, K. G., Pirtle, I. L. and Pirtle R. M. 2001. Transgenic cotton plants with increased seed oleic acid content. JAOCS 78: 941-947. Dagne, K. andjonsson, A. 1997. Oil content and fatty acid composition of seeds of Guizotia Cass. (Compositae). J. Sci. Food Agric. 73: 274-278.

10 Dutta, P. C., Helmersson, S., Kebedu, E., and Appelqvist, L. A. 1994. Variation in lipid composition of niger seed (Guizotia abyssinica Cass.) Samples collected from different regions in Ethiopia. JAOCS 71: 839-843. Fehr, R.W . 2007. Breeding for modified fatty acid composition in soybean. Crop Sci. 47: S- 72-S-87. Fuller, G., G.O. Kohler and T.H Applewhite 1996. High oleic safflower oil: A new stable edible oil. J.Am.Oil Chem.Soc. 43: 477-481. Hiremath, S. C. and Murthy, H. N. 1988. Domestication of niger (Guizotia abyssinica). Euphytica 37: 225-228. Kandel, H. and Porter, P. (eds) 2002. Niger: Guizotia abyssinica (L.f.) Cass. Production in northwest Minnesota. University of Minnesota extension service. Kinney, A. J. 1994. Genetic modification of the storage lipids of plants. Current Opinion in Biotechnology 5: 144-151. Mugendi, J.B., C.A. Sims, D.W. Gorbet, and S.F O ’Keefe 1998. Flavor stability of high oleic peanuts stored at low humidity. J. Am. Oil Chem. Soc 75: 21-25. Okuley, J., Lightner, J., Feldmann, K., Yadav, N., Lark, E. And Browse, J. 1994. Arabidopsis FAD2 gene encodes the enzyme that is essential for polyunsaturated lipid synthesis. The Plant Cell 6: 147-158. Ramadan, M. F. and Morsel, J. T. 2003. Determination of the lipid classes and fatty acid profile of niger (Guizotia abyssinica Cass.), seed oil. Phytochem. Anal. 14: 366-370. Riley, K. W. and Belayneh, H. 1989. Niger - Robbelen, G., Downey, R. K. and Ashri, A. (eds), Oil crops of the world: Their breeding and utilization. McGrow-Hill, pp. 394-403. Stymne, S. and Appelqvist, L. A. 1980. The biosynthesis of linoleate and -linolenate in homogenates from developing soya bean cotyledons. Plant Science Letters 17: 287-294. Schnurbusch, T., Mollers, C. and Becker, H. C. 2000. A mutant o f Brassica napus w ith increased palmitic acid content. Plant breeding 119: 141-144. Takagi, Y. and S.M Rahman 1996. Inheritance of high oleic acid content in the seed oil of soybean mutant M23. Theor. Appl. Genet. 92: 179-182. Tompkins, C. and E.G. Perkins 2000. Frying performance of low linolenic acid soybean oil. J. Am. Oil Chem. Soc. 77: 223-229. Urie, A.L. 1985. Inheritance of high oleic acid in sunflower. Crop Sci. 25: 986-989. Voelker, T. and Kinney, A. J. 2001. Variation in the biosynthesis of seed storage lipids. Annu. Rev. Plant Physiol. Plant Mol. Biol. 52: 335-61. Warner, K. and S. Knowlton 1997. Frying quality and oxidative stability of high oleic corn oil. J. Am. Oil Chem. Soc. 74: 1317-1322. Warner, K., P. Orr and M. Glynn 1997. Effects of fatty acid composition of oils on flavor and stability of fried foods. J. Am. Oil Chem. Soc. 74: 347-356. Wilson, R.F. and R. W. Rinne 1976. Alteration of soybean oil composition by plant breeding. J. Am. Oil Chem. Soc. 53: 595-597.

11 A cta U niversitatis A griculturae Sueciae

D octoral Thesis No. 2 0 0 8 :8 5

Genetic diversity studies of Guizotia species from Ethiopia using ISSR markers showed variation among the different populations studied. Field evaluation of agronomic characters of niger also showed great variation among the populations originating from the different regions and localities in the country. The diversity is reflected in important agronomic traits affecting duration of life cycle and seed yield. Oleic acid content in niger seed oil was raised to a new level of over 80% after repeated selection and breeding.

Yohannes Petros recieved his graduate education at the Department of Plant Breeding and Biotechnology, SLU, Alnarp. He received his BSc and MSc degrees from Addis Abeba University, Ethiopia.

Acta Universitatis Agriculturae Sueciae presents doctoral theses from the Swedish University of Agricultural Sciences (SLU).

SLU generates knowledge for the sustainable use of biological natural resources. Research, education, extension, as well as environmental monitoring and assessment are used to achieve this goal.

Online publication of thesis summary: http://epsilon.slu.se/eng/index.html

ISSN 1652-6880

ISDN 978-91-86195-18-2