1 Institute for Crop Science and Plant Breeding, University of Kiel, Germany
2 Institute of Agronomy and Plant Breeding, University of Göttingen, Germany
3 Federal centre for Breeding Research on cultivated plants, Institute for Epidemiology and Resistance, Aschersleben, Germany
Four populations of different breeding structure (2x F2, 2x DH) were used for linkage analysis. A total of 318 AFLP markers could be assigned to 19 linkage groups of a previously established RFLP map of rapeseed based on the simultaneous mapping of RFLP markers in the two F2 populations. A considerable number of markers (15 - 72 %, depending on the population) did show deviations from the expected Mendelian segregation ratios, thus making it difficult to map these markers in the corresponding population. Comparative mapping in different populations may be a means to avoid this problem. In addition to the mapping of RFLP and AFLP markers QTL for resistance against TuYV (Turnip Yellows Virus) have been mapped. A major QTL was mapped in two populations, two additional QTL only in one of the F2 populations.
Keywords: RFLP, AFLP, Comparative Map, TuYV-Resistance, QTL, ELISA
The development of closely linked markers to economically important genes will be supported by the availability of dense linkage maps. RFLPs and Microsatellites are very laborious and expensive in development. AFLP markers in contrast are easy to produce, not expensive in development and nearly as reliable as RFLPs or Microsatellites. A further advantage is the high number of polymorphic marker fragments amplified by PCR in one reaction, resulting in a sufficient coverage of the genome using only a small number of primer combinations. Also QTL for the resistance against TuYV (Turnip-Yellowing-Virus) have been mapped in two populations. Virus infection of plant populations can reduce yield up to 25 % (Graichen and Schliephanke, 1996). Therefore, resistance breeding has become more and more important. This paper describes the localisation of a major resistance gene against TuYV-infections with the help of AFLP markers.
Marker data were evaluated in four populations. F2-Man and F2-Sam resulted from crosses of two doubled haploid lines derived from two winter rapeseed varieties (‘Mansholt´s Hamburger Raps’ and ‘Samourai’) with the donor of a resistance for Turnip Yellows Virus (R54). Each F2 population consisted of 100 individuals. The winter rapeseed doubled haploid lines were also the parents of a segregating doubled haploid population (DH-MS) of which 93 lines were included in the analysis. The full complement of 151 lines of this population had previously been used to develop an RFLP map (MS-map) of the rapeseed genome (Uzunova et al. 1995). Another segregating doubled haploid population (DH606) was developed from a cross of the modern winter rapeseed variety ‘Express’ with R54 (Express x R54). For marker analysis, 212 lines each showing a consistent expression of TuYV resistance across 10 individual plants (SD < 0.5), were selected from the primary population of 409 lines. Resistance screening was done in both F2-populations and their F3 progeny as well as DH606 by DAS-ELISA after inoculating the plants with infectious aphids (Mycus persiceae). F2-experiments were carried out in the greenhouse, F3-families and 10 individual plants per line of DH606 were evaluated in the field at BAZ Aschersleben, Germany.
Linkage analysis and QTL-mapping
RFLP markers known from the RFLP map of the rapeseed genome (Uzunova et al. 1995) were mapped in F2-Man and F2-Sam. AFLP loci were detected by a nonradioctive method according to the procedure described by Schondelmaier et al. (1996) on a LiCor-system, using the primer combination Pst I / Mse I. All markers were checked for segregation deviating from the expected mendelian ratios by Chi-square test (α=0.05). Mapping was done using the programs MAPMAKER/Exp. 3.0 (Lander et al. 1987) and JOINMAP 2.0 (Stam and Van Oijen, 1995) at LOD 4.0 and a maximum distance of 35 cM (Kosambi). QTL were mapped in DH606 with PlabQTL (Utz 1993) using a composite interval mapping model whereby cofactors were selected by the program and a LOD-threshold of 2.5 to identify significant QTL. Both F2-populations have been analysed using the program QTL-Cartographer V 1.13a (Basten et al. 1997). Thresholds for QTL detection have been determined by a permutation analysis depending on population characteristics in the latter case.
Fig. 1: Frequency distribution of phenotypic values for TuYV resistance in DH606 on the reduced subpopulation. Values near by 0 represent resistant plants.
PCR amplifications using 20 (DH-MS, F2-Man, F2-Sam) and 12 (DH606) primer combinations, resulted in the scoring of 530 AFLP bands in all four populations (DH-MS, F2-Man, F2-Sam and DH606). The mean number of polymorphic markers per primer combination was 8.7, varying from 1 to 17 markers per combination within the populations. Eight primer combinations revealed amounts of polymorphic markers below the mean and have been excluded from mapping in DH606. From 530 different polymorphic AFLP fragments 168 have been mapped in more than one population. Similar results concerning their assignment to the same linkage group were obtained for 124 markers of this group. Additional 150 markers have been mapped in one population only. A total of 212 markers has not yet been assigned to one of the existing linkage groups. Because of close relationships between the populations no marker could be mapped in all four populations. Ambiguities in marker order between different populations have been resolved by using RFLP markers as anchor markers in both F2-populations or, if RFLPs were not available, using AFLP markers mapped in more than one population. The degree of markers showing a non-mendelian segregation ratio ranged from 25 % in DH-MS to 71% in DH606 with the values of both F2-populations lying in-between at 49% and 65%. In DH606 32 markers (25%) revealed a variation of fragment intensities which has not been expected for a DH-population. Only 6 of these markers were segregating according to mendelian segregation ratios. However 20 of these markers were mapped to the expected linkage groups.
The phenotypic data of the selected lines of the DH606 population (mean values of 10 plants per line) exhibited a clear bimodal frequency distribution (Fig 1). QTL analysis on the data of ten offsprings and on mean values independently detected only one QTL for resistance against TuYV infection on linkage group MS17 of the rapeseed RFLP map (Uzunova et al., 1995). Other mapped genomic regions did not reveal significant QTL at threshold of LOD 2.5. The phenotypic variance explained by this QTL ranged from 30.0% to 39.5%, depending on replication, with LOD scores reaching a maximum of 15. The QTL was flanked by two relatively large AFLP fragments of 200 and 300 Bp. In the F2 population F2-Man a QTL could be mapped in the same region as in DH606. Contrary to DH606 two additional QTL could be localised on linkage groups MS5 and MS12. These linkage groups have been also established in DH606, but they do not represent regions equal to those harbouring the QTLs in F2-Man. The maximum LOD as well as the phenotypic variance explained by QTL on linkage group MS17 in F2-Man and DH606 were in close agreement for DH606 and F2-Man. In F2-Sam it was not possible to detect any significant QTL.
The number of markers which could be assigned to linkage groups was quite different between both F2-populations and DH-MS. This result was somewhat unexpected because the different parents of the three populations are genetically unrelated and an equal amount of markers was expected to be mapped in all populations. Instead of these findings, F2-Man revealed a larger number (224 AFLPs) of markers amplified with the same 20 primer combinations than F2-Sam. This might be due to the larger genetic distance of the old landrace „Mansholt’s Hamburger Raps“ revealed by other examinations (Knaak and Ecke, 1995). In contrast to this assumption the number of AFLP fragments derived from the winter rapeseed lines Man or Sam was nearly equal in DH-MS (48 % and 52 %). The same findings could be observed for the alleles of parents in DH606 and in both F2-populations too, although 18 % of the alleles derived from the resistance donor R54 could not be detected in one of the F2-populations. The assumption, that genetically different plants might have been used as resistance donors in populations was underlined by the fact that 21 markers out of 57 detecting an allele of R54 in DH606 could not be recovered in the F2-populations. In total, only 202 of 435 markers (AFLP and RFLP) have been mapped in at least two populations. Despite of this there was a sufficient number of anchor-markers to allow an alignment of the linkage maps. These aligned maps will serve as a starting point for upcoming mapping experiments because a larger number of polymorphic fragments has been evaluated in these four populations compared to a single population.
One problem occurring frequently in all four populations was the high proportion of markers segregating in a non mendelian way. Wagner et al. (1992) could show that disturbed segregation ratios should not affect mapping results. The presented mapping basically were in accordance with this statement but with a few exceptions. Some marker fragments with distorted segregation ratios mapped to different genome positions depending on the mapping population although the fragments were derived from the same parent. One explanation for this finding might be the duplication of complete genome regions possibly harbouring homologous loci with identical alleles (McGrath et al., 1990; Uzunova et al., 1994 and Osborne et al., 1997). This would result in at most nine different allele combinations depending on the degree of variation which has occurred since duplication: A1/A2; A1/H2; A1/B2; H1/A2; H1/H2; H1/B2 and B1/A2 B1/H2; B1/B2. These combinations allow the differentiation of five degrees of intensity for the fragment under optimal conditions. Identifying bands representing two independently segregating fragments is relatively easy in a DH-population. Problems occur in F2 populations. Here markers are scored codominantly by fragment intensities, heterozygous individuals will have only half the number of allele copies resulting in fragments with half the intensity. Multiple alleles having the same fragment length will then be identified as heterozygous or homozygous individuals with abnormal segregation ratios. One way to identify multiple markers is a χ2-test followed by genetic mapping of the markers. Diverging mapping results in different populations together with the results from the χ2-test have identified markers containing two or more independently segregating alleles. In the present study markers fulfilling both characteristics have been excluded from further comparative mapping or map alignments.
The maps which could be established in the different populations varied to a great extent in total genome length between DH and F2-populations and in total number of mapped markers. Maps of F2-populations harboured twice the number of markers mapped in DH-populations. Compared with the maps of F2-populations 7 identical out of 8 linkage groups in each DH-population are still left without markers. On the other hand, there are still some linkage groups left which could not be assigned to a chromosome of the MS-Map. Taking into account markers which are mapped multiple times in different populations a theoretical density of one marker each 6.5 cM is available. Within the four populations the value differs from 7.9 cM in F2-Man to 10.3 cM in F2-Sam. This is suitable for QTL-mapping approaches (Darvasi and Soller, 1994) but is fairly enough to find markers closely linked to specific genes for breeding purposes at this time.
Frequency distributions of phenotypic values for F2-plants of F2-Sam and F2-Man were not in agreement with the assumption of a monogenic inheritance of the TuYV-resistance. The QTL analysis presented here resulted in the detection of three significant QTL in F2-Man, but it was not possible to detect any significant QTL in F2-Sam. The first examination of the phenotypic mean values based on 10 offsprings per line for the complete DH606 population (N=409 lines) showed a trimodal frequency distribution which again implies a multigenic inheritance. Analysis of plants with phenotypic values close to the population mean exhibited a large variation of phenotypic values within the offspring derived from one DH-line. This finding was not consistent with the expectation of homogenous values for the offspring of one DH-Line. First assumption of a genetic variation within the offspring per line could be rejected by genetic analyses of 5 primer combinations on 109 offsprings of 14 DH-lines showing great phenotypic variation while all plants within one offspring were genetically uniform. The interpretation of these results led to the reduction of the complete population to those plants showing a more consistent behaviour. We are aware of the fact, that this procedure might result in a loss of variation within the population, possibly linked with minor factors. On the other hand, the calculated frequency distribution of the phenotypic mean values in the reduced population gave strong hints for monogenic inheritance of the resistance in this population. Therefore the variations within the offspring of one line may be due to experimental errors. Because the detected QTL did not explain the total phenotypic variance so far it will be important to look for additional factors influencing the expression of the virus resistance, although it might be difficult and erroneous to detect phenotypic variation. Despite these hints for a monogenic inheritance a QTL analysis was applied to the population resulting in only one sharp QTL on linkage group MS17. Position and explained phenotypic variance of this QTL was in close agreement with the findings in F2-Man and emphasises the chosen procedure to overcome possible erroneous interpretation of the phenotypic data. The detection of additional QTL in DH606 which are located in F2-Man might not have been possible because these genomic regions are not yet covered by AFLP markers in DH606. The importance of the QTL in DH606 on linkage group MS17 for a selection on TuYV resistance may be proven in different breeding material with known phenotype. The QTL is closely flanked by two AFLP-markers, which have a suitable fragment length to be converted into other PCR-based markers which are more easy to handle. Now we are on the way to clone specific fragments out of PAA-gels, that can be converted into suitable PCR-markers.
We thank N. Pinnow for excellent technical assistance, Norddeutsche Pflanzenzucht H.-G. Lembke KG Hohenlieth for providing us with plant material and the Foundation „Schleswig Holsteinische Landschaft“ for financial support.
1. Basten C.J., Weir B.S. and Zeng Z.B., 1997. QTL Cartographer: A reference manual and tutorial for QTL mapping. Dep. of Statistics, NC State Univ., Raleigh, USA.
2. Darvasi A. and Soller M., 1994. Optimum spacing of genetic markers for determining linkage between marker loci and quantitative trait loci. Theor. Appl. Genet. 89: 351 - 357.
3. Graichen K. and Schliephanke E., 1996. Auftreten, Symptome und Vektoren des Wasserrübenvergilbungsvirus (Syn. Westliches Rübenvergilbungsvirus) am Winterraps. Nachrichtbenbl. Deut. Pflanzenschutzd. 48.8/9: 186 - 191.
4. Knaak C. and Ecke W., 1995. Genetic diversity and hybrid performance in european winter oil-seed rape (Brassica napus L.). Proceedings of the ninth international rapeseed congress, Cambridge, United Kingdom: 110-112.
5. Lander E.S., Green P., Abrahamson J., Barlow A., Daly M.J., Lincoln S.E. and Newburg L., 1987. MAPMAKER: An interactive computer package for constructing primary genetic linkage maps of experimental and natural populations. Genomics 1: 174 - 181.
6. McGrath J.M., Quiros C.F., Harada J.J. and Landry B.S., 1990. Identification of Brassica oleracea monosomic alien chromosome addition lines with molecular markers reveals extensive gene duplication. Mol Gen Genet 223: 198 - 204.
7. Osborn T.C., Kole C., Parkin I.A.P., Sharpe A.G., Kuiper M., Lydiate D.J. and Trick M., 1997. Comparison of floweringtime genes in B. napus and Arabidopsis thaliana. Genetics 146: 1123 - 1129.
8. Schondelmaier J., Koch G. and Jung C., 1996. Die Einsatzmöglichkeiten von AFLPs in der Pflanzenzüchtung. Vortr. Pflanzenz. 33: 112 - 125.
9. Stam P. and Van Ooijen J.W., 1995. JoinMap (tm) version 2.0: Software for the calculation of geneic linkage maps. CPRO-DLO, Wageningen, NL.
10. Utz H.F. and Melchinger A.E., 1995. PLABQTL. A computer program to map QTL. Institut für Pflanzenzüchtung, Saatgutforschung und Populationsgenetik. Univ. Hohenheim, Stuttgart.
11. Uzunova M., Ecke W., Weissleder K. and Röbbelen G., 1995. Mapping the genome of rapeseed (Brassica napus L.) I. Construction of an RFLP linkage map and localization of QTLs for glucosinate content. Theor. Appl. Genet. 90: 194 - 204.
12. Wagner H., Weber W.E. and Wricke G., 1992. Estimating linkage relationship of isozyme markers and morphological markers in sugar Beet (Beta vulgaris L.) including families with distorted segregations. Plant Breeding 108: 89-96