Plant Breeding 126, 487—491 (2007) doi:10.1111/j.1439-0523.2007.01393.x Ó 2007 The Authors Journal compilation Ó 2007 Blackwell Verlag, Berlin

Genetic diversity of Brassica carinata with emphasis on the interspecific crossability with B. rapa

Y. Jiang1,2, E. Tian1 , R. Li1 , L. Chen1 and J. Meng1 1National Key Laboratory of Crop Genetic Improvement, Huazhong Agricultural University, Wuhan 430070, China; 2Present address: Institute of Crop Science, Anhui Academy of Agricultural Sciences, Hefei 230031, China; 3Corresponding author, E-mail: [email protected] With 3 figures and 1 table Received 23 March, 2006/Accepted 9 February, 2007 Communicated by A.-M. Chevre

Abstract strated that a hexaploid hybrid (AABBCC), derived from an Brassica carinata (BBCC, 2n ¼ 34) possesses many good agronomic interspecific cross between B. carinata and B. rapa (AA, 2n ¼ characters and, to provide more genetic information about this species 20), could be used as a bridge hybrid (Li et al. 2004). By and utilize it further, 110 accessions of B. carinata were tested for crossing the hexaploid hybrid with B. napus,itwaspossibleto genetic variation by using 233 amplified fragment length polymorph- create a pentaploid hybrid (AABCC). This pentaploid hybrid ism markers, so that a dendrogram could be constructed. To combine wasthenusedtocreateavariantofB. napus whose DNA the good traits of B. carinata and B. rapa (AA, 2n ¼ 20) with those of comprised half the A genome of B. rapa and half the B. napus (AACC, 2n ¼ 38) which is the major rapeseed variety in CgenomeofB. carinata (Li et al. 2004, 2006). The crossability China, interspecific crosses between these two species have been made between B. carinata and the three basic species of Brassica and which have resulted in 276 doubled hybrids (AABBCC). These B. napus is very low (Stebbins 1958). Several interspecific hexaploids combine positive characters and can be used for crossing with B. napus. Brassica carinata as the female parents exhibited crosses involving B. carinata have been made, which have extensive differences in interspecific crossability and 20 high crossabil- resulted in the transfer of some desirable traits from B. carinata ity accessions were identified. The Poisson Regression Model analysis to other Brassica species (Choudhary et al. 2000, Rahman result (P < 0.0255) indicated that such differences were due to the 2001, Tonguc and Griffiths 2004). Therefore, exploring geno- genotype of the accessions. Accessions with high crossability could types of B. carinata that have high interspecific crossability promote gene flow within the genus Brassica. would be useful in order to promote gene flow within the genus Brassica. Key words: Brassica carinata — Brassica napus — genetic This investigation has three objectives. First, to evaluate the variation — amplified fragment length polymorphism — genetic diversity of 110 accessions of B. carinata. Secondly, to interspecific cross — gene pool create a new bridge germplasm (AABBCC) for transferring genetic material from B. carinata and B. rapa to B. napus. Ethiopian mustard (also called Abyssinian Mustard or Abys- Thirdly, to screen genotypes of B. carinata for high crossability sinian Cabbage Brassica carinata, BBCC, 2n ¼ 34) possesses to B. rapa. many desirable agronomic characteristics that are rare in the other Brassica oil crops: heat and drought tolerance, disease and pest resistance, and a yellow-seeded germplasm (Gugel Materials and Methods et al. 1990, Bayeh and Gebre Medhin 1992, Yitbarek 1992). Its Plant materials: One hundred and seven accessions of Brassica. importance to breeders has increased because of rapid carinata, mainly of Ethiopian origin, were obtained from the improvement in its seed quality (Getinet et al. 1994, Velasco Centre for Genetic Resources, Wageningen, the Netherlands et al. 2003, Teklewold and Becker 2005). As a result, its genetic (CGN; Supplementary Table 1). An additional three accessions diversity has prompted considerable interest among Brassica (88-216845, 80-139, 88-219790) of B. carinata were obtained from breeders. Alemayehu and Becker (2002) analysed 13 morpho- Germany (Supplementary Table 1). Twenty-nine accessions of logical and seed quality characteristics of 36 accessions of B. rapa were collected, mainly from various locations in China B. carinata and found extensive genetic variability. Warwick and grown for at least four generations in Wuhan, China (Supplementary Table 2). et al. (2006) used amplified fragment length polymorphism (AFLP) to evaluate the patterns and extent of genetic diversity Field trial: All accessions were sown in the field station of the of B. carinata. Exploring the genetic diversity of B. carinata Huazhong Agricultural University, Wuhan in September 2003. Each may provide more information for plant breeders. accession was planted in three rows with 10 plants per row. To ensure Interspecific hybridization is a useful way to transfer pollination by B. rapa throughout the flowering period of B. carinata, desirable traits from one species to another, and has enriched the seeds of B. rapa were sown 43 days and 55 days after sowing the the gene pool of Brassica crops (Allard 1960, Prakash and accessions of B. carinata, respectively. When the plants flowered in Chopra 1998, Liu 2000). Over the past few decades, numerous March 2004, buds from three or four plants of each B. carinata efforts to widen the genetic basis of B. napus (AACC, 2n ¼ 38) accession were picked and randomly fertilized with pollen of B. rapa. which is narrow when compared to B. rapa and B. carinata At harvest, the crossability of different parents was analysed. In 2005, have been undertaken. In previous research, it was demon- four accessions (CGN03941, CGN03949, CGN03995 and CGN04004)

www.blackwell-synergy.com 488 Jiang, Tian, Li, Chen and Meng of B. carinata with high interspecific crossability and one accession AFLP procedure: DNA samples were prepared from young leaves (CGN03931) with low crossability to B. rapa were selected for crossing of five plants of each accession of B. carinata to estimate the with B. rapa to assess the environmental effect on the interspecific genetic variation amongst accessions. DNA was extracted using a crossability. To make the triploid interspecific hybrids fertile, the modified Cetyltrimethyl Ammonium Bromide (CTAB) procedure putative hybrids were treated with 0.1% (w/v) colchicine. (Doyle and Doyle 1990), concentration and purity were measured

Fig. 1: Dendrogram generated with 233 amplified fragment length polymorphism markers on 110 accessions of Brassica carinata. The scale shows SM similarity coefficients and the dendrogram was developed using the Unweighted Pair Group Method with Arithmetic mean clustering procedure Genetic diversity of Brassica carinata 489 by a Beckman spectrophotometer (Beckman Coulter Inc., Fuller- Interspecific crossability ton, CA, USA). AFLP–PCR was performed according to the In total, 409 interspecific combinations were made. Eventually, method described by Vos et al. (1995), with minor modifications. 13 000 seeds were obtained from 40,000 cross-pollinated buds Genomic DNA (200 ng) was restricted with EcoR1 (2 U/200 ng DNA) and Mse1 (2 U/200 ng DNA) for 5 h. After ligation for involving 375 combinations. As in most cases B. carinata and 10 h (1 U ligase/200 ng DNA), the preamplification was carried B. rapa could be distinguished easily by their morphology, it out in a Perkin-Elmer 9600 thermocycler (Foster City, CA, USA). was not difficult to identify true hybrid plants. Few of the The pre-PCR products were diluted 1 : 20 with deionized water hybrids required verification by molecular marker analysis and and used for a second PCR with selective primers. The PCR chromosome counting because they resembled their maternal products were separated on a 6% polyacrylamide gel followed by parents. Eventually, more than 5000 hybrid plants were silver staining. generated from 326 interspecific combinations (Supplementary Table 1). However, colchicine treatment doubled the number TM Data analysis: To facilitate scoring across gels, a GeneRuler 50 bp of chromosomes in only 23.3% of the hybrid plants, which AFLP DNA ladder (MBI Fermentas, GmbH, Germany) was used. resulted in 276 combinations of hexaploid hybrid plants. Polymorphic fragments between 50 and 400 bp in size were scored as The ICIs of accessions of B. carinata and B. rapa differed either present (1) or absent (0). DNA markers were analysed by the considerably (Fig. 2). Some accessions of B. carinata yielded Unweighted Pair Group Method with Arithmetic mean. A dendro- more than 80 hybrid seeds per 100 interspecific pollinated buds, gram was constructed by software NTSYS-PC 2.1 (Rohlf 2002) using SM coefficients (Sokal and Michener 1958). The SM coefficient was whereas some produced none. A set of data selected from the calculated with the formula: random interspecific combinations met the requirement of PRM where the male parent, of B. rapa, ÔBaijian 13Õ, was crossed to SMij ¼ðn10 þ n01Þ=ðn11 þ n10 þ n01 þ n00Þ; nine female parents each involving three or four individual plants. The Value/DF of Deviance and Pearson chi-square were where n was the number of bands in common between accessions 11 0.21 and 0.16, respectively, which indicated that PRM was fitting i and j,n was the number of bands present in i and absent in j,n 10 01 the data set for the analysis. The PRM analysis result (v2 ¼ was the number of bands absent in i and present in j, and n00 was the number of bands both absent in i and j. 17.48, df ¼8, P < 0.0255) showed that the ICI for each female Bootstrap analysis was performed with the Phylogenetic Analysis, parent was statistically significant, indicating that such differ- using Parsimony (PAUP) software (Swofford 1999) to evaluate the ences were caused by the genotype of the accessions. robustness of the nodes of the dendrogram. 90 80 Crossability analysis: The interspecific crossability index (ICI) 70 between B. carinata and B. rapa was defined as the number of 60 hybrid plants obtained from 100 cross-pollinated buds, an ICI value 50 40 bigger than 20 being considered an accession with high crossability. 30 After the ICI was determined, the Poisson Regression Model (PRM) 20 in the GENMODE procedure of the Statistical Analysis System (SAS Crossability index 10 0 Institute 1999, Park 2005) was used to compare the differences in 1 11 21 31 41 51 61 71 81 91 101 111 121 131 ICI for each female parent, where the female parents were included Accession no. as factors and likelihood ratio tests were used to determine the appropriate model. Fig. 2: The interspecific crossability index of Brassica carinata and B. rapa. The accession number (x-axis) corresponds to the numbers listed for B. carinata in Supplementary Table 1 and B. rapa in Results Supplementary Table 2 Genetic variation amongst accessions of B. carinata 1500 Morphological and agronomic characteristics varied amongst the B. carinata accessions. The colour of young stems or leaves 1400 differed, especially along the main leaf vein. Some plants 1300 showed purple colouring which made them easily distinguish- 1200 able from the green plants. Bolting and flowering times of the accessions ranged between 90 and 212 days. Some of the 1100 accessions were seriously damaged by frost when the plants 1000 were bolting in the early winter. Two-hundred and thirty-three AFLP markers were gen- 900 erated from the 110 accessions of B. carinata with 36 primer 800 pairs, of which 11 produced more than seven polymorphic 700 bands (data not shown). The genetic similarity coefficients varied from 0.46 (CGN03951–CGN03989) to 0.88 600

(CGN03953–CGN04004 and CGN04032–CGN04036) and 500 a dendrogram based on them was obtained (Fig. 1). 3300 3400 3500 3600 3700 3800 3900 4000 4100 4200 4300 Bootstrap analysis showed that only four of 109 nodes in the dendrogram had values higher than 0.5 (data not shown). Fig. 3: The distribution of 104 accessions of Ethiopian origin of Brassica carinata. The y-axis and the x-axis denote the latitude and the There were no significant relationships between accessions longitude, respectively. The solid triangles denote the 20 accessions originating from the same region that could indicate gene flow with high crossability to B. rapa, of which 16 were located in the south- between accessions from Ethiopia. east region of the Great Rift Valley (broken line) 490 Jiang, Tian, Li, Chen and Meng

Table 1: Results from crossing five Brassica carinata accessions with B. rapa in the year 2005

Accession Number of Number of Number of number Accession name crossing combinations pollinated buds true hybrid plants ICI

CGN03941 Gommenzer (Amh.) 18 1089 874 80.3 (40)1 CGN03949 Gommenzer (Amh.) 28 1617 1081 66.9 (28) CGN03995 Gommenzer (Amh.) 62 2540 2506 98.7 (80) CGN04004 Gommenzer (Amh.) 13 848 535 63.1 (47) CGN03931 Gommenzer (Amh.) 20 1272 93 7.3 (6)

1The number in brackets is the interspecific crossability index (ICI) value from 2004.

Twenty accessions with ICI values above 20 were selected Sharma and Gill 1983, Poysa 1990). Similar findings have been (Supplementary Table 1). Of these, 16 originated from the reported in Brassicaceae species (Ripley and Beversdorf 2003). south-east region of the Great Rift Valley, Ethiopia. Further- Earlier reports highlight the importance of the maternal more, 13 of these accessions were geographically located genotype in this species (Meng et al. 1992, Meng and Lu between 850–920N and 3950–4205E (Fig. 3). The ICI data of 1993, Liu and Meng 1995). Choudhary et al. (2000) reported 2004 and 2005 (Table 1) matched well, with a high correlation that the crossability between B. carinata and B. rapa is very coefficient (r ¼ 0.88), indicating that the basis of the variation low, especially when the latter is used as the female parent. In a of its interspecific crossability was mainly determined previous report, two cultivars of B. rapa with a high genetically. interspecific crossability involving three accessions of B. carinata and nine cultivars of B. rapa were selected. It was not possible to identify a cultivar of B. carinata with good Discussion crossability due to the limited number of accessions being The genetic diversity detected in this trial is wider than that evaluated (Meng et al. 1992). In the present experiment, highly previously reported by Warwick et al. (2006) and may be due crossable accessions were selected and these lines could be used to a larger sample (110 accessions). This diversity could be as bridge materials for transferring the desirable characteristic useful for breeders when utilizing genotypes from different to other Brassica species and vice versa. In contrast to most genetic subclusters in their breeding programmes. other interspecific combinations where cross-pollinated The low support value of the bootstrap analysis may be siliques were developed between 15 and 20 days after pollin- partly due to the wide range of genetic diversity within most of ation, the compatible interspecific combinations usually pro- the open-pollinated accessions. This is supported by observa- duced fully developed seeds within each cross-pollinated tions on morphology and by molecular marker analysis. Low silique. This implies that accessions of B. carinata allow support values can arise when a relatively small number of foreign pollen to germinate on their stigmas and enable pollen polymorphic bands and dominant markers that may exhibit tubes to penetrate their styles. However, only cultivars with a homoplasies are used for constructing the dendrogram high crossability have the mechanism(s) to enable the hybrid (Koopman et al. 2001). Homoplasies in AFLP data sets can embryos to develop fully. Further studies on the genetic basis be caused either by scoring non-sequential identical fragments of interspecific crossability, to increase our understanding of of equal length or bands that represent codominant loci as the mechanism, will be undertaken. dominant loci. Nevertheless, genetic variability studies have shown consistency between AFLP-based phenetic dendro- grams and pedigree or phylogenetic data, which suggests that Acknowledgements homoplasies in AFLP data are relatively rare (Hill et al. 1996, The authors thank the Centre for Genetic Resources, the Netherlands Lombard et al. 2000, Seefelder et al. 2000, Koopman et al. (CGN) and Dr Wancang Sun for kindly providing the accessions of B. 2001). carinata and B. rapa and Miss Yuexiu Liu for reading the manuscript. Previous studies on the cytogenetics and molecular genetics The study was supported by the 948 Program of the Ministry of Agriculture and Nature Science Foundation of Hubei Province of of Brassicaceae species have revealed significant differences China. between the A-, B- and C-genomes from the three diploid species and the three amphidiploid species (McGrath and Quiros 1990). A synthetic rapeseed (B. napus, AACC, 2n ¼ 20) References generated by combining half of the A-genome of B. rapa (AA) Alemayehu, N., and H. C. Becker, 2002: Genotypic diversity and and half of the C-genome of B. carinata (BBCC) is very patterns of variation in a germplasm material of Ethiopian mustard different from the original B. napus and exhibits strong (Brassica carinata A. Braun). Genet. Resour. Crop Evol. 49, 573—582. heterosis when crossed to the wild-type B. napus (Li et al. Allard, R. W., 1960: Principles of Plant Breeding. John Wiley and 2006). Bridge hexaploids (AABBCC) combine the positive Sons, Inc., New York. characteristics of B. rapa and B. carinata and are essential for Bayeh, M., and T. Gebre Medhin, 1992: Insect pests of noug, linseed broadening the genetic variation of B. napus. and Brassica in Ethiopia. In: Oilseeds Research and Development in Interspecific reproductive isolation is the main mechanism Ethiopia, 174—178. Proceedings of the First National Oilseeds, for speciation and specific maintenance (Griffiths et al. 2002). Workshop, 3–5 December 1991, Addis Abeda, Ethiopia. However, genetic leaks in the isolation system have occasion- Choudhary, B. R., P. Joshi, and S. Ramarao, 2000: Interspecific hybridization between Brassica carinata and Brassica rapa. Plant ally enabled interspecific gene flow and promoted the evolution Breed. 119, 417—420. of species. Incompatibility with related species in some Doyle, J., and J. Doyle, 1990: Isolation of plant DNA from fresh genotypes of wheat and tomato have been overcome when tissue. Focus 12, 13—15. the varieties are used as female parents (Snape et al. 1979, Genetic diversity of Brassica carinata 491

Getinet, A., G. Rakow, J. P. Raney, and R. K. Downey, 1994: Ripley, V. L., and W. D. Beversdorf, 2003: Development of self- Development of zero erucic acid Ethiopian mustard through an incompatible Brassica napus: (I) introgression of S-alleles from interspecific cross with zero erucic acid Oriental mustard. Can. Brassica oleracea through interspecific hybridization. Plant Breed. J. Plant Sci. 74, 793—795. 122, 1—5. Griffiths, A., W. Gelbart, R. Lewontin, and J. Miller (eds) 2002: Rohlf, F. J., 2002: NTSYS-pc. Numerical Taxonomy and Multivariate Modern Genetic Analysis, 619—648. Freeman and Company, New Analysis System, Version 2.1. Applied Biostatistics, New York. York. SAS Institute, 1999: SAS Online Doc Ò, version 8.0, Cary, NC, USA. Gugel, R. K., G. Sguin-Swartz, and A. Petrie, 1990: Pathogenicity of Seefelder, S., H. Ehrmaier, G. Schweizer, and E. Seigner, 2000: Genetic three isolates of Leptosphaeria maculans on Brassica species and diversity and phylogenetic relationships among accessions of hop, other crucifers. Can. J. Plant Pathol. 12, 75—82. Humulus, as determined by amplified fragment length polymorph- Hill, M. H., M. Witsenboer, R. Vos Zabeau, R. Kesseli, and ism fingerprinting compared with pedigree data. Plant Breed. 119, R. Michelmore, 1996: PCR-based fingerprinting using AFLPs as a 257—263. tool for studying genetic relationships in IMCIUCU spp. Theor. Sharma, H. C., and B. S. Gill, 1983: Current status of wide Appl. Genet. 93, 1202—1210. hybridization in wheat. Euphytica 32, 17—31. Koopman, W. J. M., M. J. Zevenbergen, and R. G. van den Berg, Snape, J. W., V. Chapman, J. Moss, C. E. Blanchard, and T. E. Milier, 2001: Species relationships in Lactuca s.l. (Lactuceae, Asteraceae) 1979: The crossability of wheat varieties with Hordeum bulbosum inferred from AFLP fingerprints. Am. J. Bot. 88, 1881—1887. crossability. Heredity 42, 291—298. Li, M., W. Qian, J. Meng, and Z. Li, 2004: Construction of novel Sokal, R. R., and C. D. Michener, 1958: A statistical method for Brassica napus genotypes through chromosomal substitution and evaluating systematic relationships. U. Kansas Sci. Bull. 38, elimination using interploid species hybridization. Chromosome 1409—1438. Res. 12, 417—426. Stebbins, G. L., 1958: The inviability, weakness and sterility of Li, M., X. Chen, and J. Meng, 2006: Intersubgenomic heterosis in interspecific hybrids. Adv. Genet. 9, 147—215. rapeseed production with a partial new-typed Brassica napus Swofford, D. L., 1999: PAUP: Phylogenetic Analysis Using Parsi- containing subgenome Ar from B. rapa and Cc from B. carinata. mony. Version 4. Sinauer, Sunderland, Mass. Crop Sci. 46, 234—242. Teklewold, A., and H. C. Becker, 2005: Comparison of phenotypic and

Liu, H., 2000: Genetics and Breeding in Rapeseed, 82—177. Chinese molecular distances to predict heterosis and F1 performance in Agricultural University Press, Beijing, China. Ethiopian mustard (Brassica carinata A. Braun). Theor. Appl. Liu, X., and J. Meng, 1995: Embryology research on crossabilty of Genet. 112, 752—759. Brassica napus and B. carinata. Acta. Agron. Sin. 21, 385—391. Tonguc, M., and P. D. Griffiths, 2004: Transfer of powdery mildew Lombard, V., C. R. Baril, R. Dubreuil, R. Blouet, and D. Zhang, 2000: resistance from Brassica carinata to Brassica oleracea through Genetic relationships and fingerprinting of rapeseed accessions by embryo rescue. Plant Breed. 123, 587—589. AFLP: consequences for varietal registration. Crop Sci. 40, Velasco, L., J. M. Ferna´ndez-Martı´nez, and A. De Haro, 2003: 1417—1425. Inheritance of increased oleic-acid concentration in high erucic-acid McGrath, J. M., and C. F. Quiros, 1990: Generation of alien Ethiopian mustard. Crop Sci. 43, 106—109. chromosome addition lines from synthetic Brassica napus: morphol- Vos, P., R. Hogers, and M. Bleeper (eds) 1995: AFLP: a new technique ogy, cytology, fertility, and chromosome transmission. Genome 33, for DNA fingerprinting. Nucleic Acids Res. 21, 4407—4414. 374—383. Warwick, S. I., R. K. Gugel, T. Mcdonald, and K. C. Falk, 2006: Meng, J., and M. Lu, 1993: Genotype effects of Brassica napus on its Genetic variation of Ethiopian mustard (Brassica carinata) germ- behavior after pollination with B. juncea. Theor. Appl. Genet. 87, plasm in western Canada. Genet. Resour. Crop Evol. 53, 297—312. 238—242. Yitbarek, S., 1992: Pathological research on noug, linseed, gomenzer Meng, J., J. Wu, and P. Wang, 1992: Crossability of Brassica carinata and raeseed in Ethiopia. In: Oilseeds Research and Development in with B. rapa and B. juncea. J. Huazhong Agric. Univ. 11, 203—207. Ethiopia, 151—161. Proceedings of the First National Oilseeds Park, H. M., 2005: Regression Models for Event Count Data Using Workshop, 3–5 December 1991, Addis Abeba, Ethiopia. SAS, STATA, and LIMDEP. Available at: http://www.indiana.edu/ statmath/stat/all/count/count4.html. Poysa, V., 1990: The development of bridge lines for interspecific gene Supplementary Materials transfer between Cycopersicon esculentum and L. peruvianum. Theor. The following material is available online at http://www.blackwell- Appl. Genet. 79, 187—192. synergy.com/ Prakash, S., and V. L. Chopra, 1998: Introgression of resistance to Table 1: Accessions of B. carinata obtained from the Netherlands and shattering in Brassica napus from Brassica juncea through nonho- Germany and their crossability with B. rapa mologous recombination. Plant Breed. 101, 167—168. Table 2: Accessions of B. rapa obtained from China and Sweden and Rahman, M. H., 2001: Production of yellow-seeded Brassica napus their crossability with B. carinata through interspecific crosses. Plant Breed. 120, 463—472.