22. An FRLP-based genetic analysis of photoperiod sensitive male sterlity in rice

Ai-Fa ZHANG¹, B.S. SHEN¹, X.K. DAI¹, M.H. MEI¹, M.A. Saghai MARROF² and Z.B. Li¹

1) Biotechnology Center, Huazhong Agricultural University, Wuhan 430070, China

2) Department of Crop and Soil Environmental Sciences, Virginia Polytechnic Institute and State University, Blacksburg, VA 24061, USA

Photoperiod sensitive male sterile (PSMS) rice possesses tremendous potential for hybrid breeding. Utilization of PSMS for the development of two-line hybrids has become one of the major objectives of many rice breeding programs (Yuan 1992). The objective of this study is to determine the locations of the genes for PSMS in the molecular map, with the long-term goal to develop selective markers to facilitate the transfer of the genes in breeding programs and to isolate the PSMS genes.

A cross was made in 1990 using 32001S (a PSMS line developed by transferring the PSMS gene from Nongken 58S) as the female parent, and Ming Hui 63 (a normal rice cultivar) as the male parent. A large F\2\ population (>1200 individuals) was planted in the field annually in the last three years (1991-1993) to determine the inheritance of PSMS. The segregation ratio, which fluctuated from one year to another, indicated that PSMS in this population is conditioned by recessive alleles at 2-3 loci.

We devised a two-step approach for mapping the genes for PSMS: (i) identifying the chromosomal segments linked to the genes with bulked extremes, and (ii) determining the map locations of the genes using recessive individuals. Two bulks were made: the first bulk (bulk F) consisted on an equi-mix of DNA from 30 highly fertile plants, and the second bulk (bulk S) was made of an equi-mix of DNA from 58 highly sterile plants. These two bulks together with the two parents were digested with 6-21 restriction enzymes and probed with about 400 clones, mostly from the Cornell RFLP linkage map (Tanksley et al. 1992). A total of about 90 markers, covering >90% of the Cornell Map, detected polymorphisms between the parents with at least one enzyme.

For a marker that is closely linked to a gene for PSMS, bulk F would show the bands from both parents in the autoradiography, since, based on the assumption that a sterile individual is completely recessive, a fertile individual would be genetically of any of the possible multilocus genotypes except the one of complete recessive. But, in bulk S, the paternal band would be absent or very faint depending on the distance of the marker from the gene, because all the recessive alleles are from the maternal parent, and appearance of the paternal bands in the sterile individual s would be the result of recombination. In this way, markers from three regions, located on chromosomes 1, 3 and 7, respectively, were identified that are likely to be linked to the genes for PSMS.

To determine the map locations of PSMS genes, the 58 sterile plants were analyzed individually with a total of 25 probes from the three chromosomal regions. The recombination frequency (rf) between a marker and a gene for PSMS was calculated with data from these 58 plants (assuming them to be recessive at all loci) using a maximum likelihood estimator (Allard 1956).

The rf values for markers on chromosome 7 agreed well with the distances given in the Cornell map. We thus concluded that there is a major locus for PSMS on chromosome 7 located about 4 map units from RG477, and 16 map units from RG511. However, the rf values calculated for markers on chromosomes 1 and 3 were not quite consistent with the map distances in the Cornell map. There are two explanations for such inconsistency: (i) An analysis of 112 random F\2\ individuals from the same population indicated significant segregation distortion for markers in both of these two regions. (ii) The effects of fertility caused by these two loci are small, and consequently, some of the 58 sterile plants were not homozygous for male sterile alleles at these two loci.

To ascertain the existence of loci for PSMS in these chromosome regions, an additional sample of about 220 individuals from the same F\2\ population was assayed with 6 markers, two from each region. An one-way analysis of variance (AOV) performed using marker genotypes as the group effects showed that the locus on chromosome 7 had the largest effect followed by the one on chromosome 3, and the effect of the locus on chromosome 1 was barely significant. When a three-way AOV was performed using one marker from each of these three regions (RG811, RG191 and RG477), the effect of the locus on chromosome 1 (marked by RG811) became insignificant. These analyses established the existence of two major loci for PSMS on chromosomes 3 and 7, respectively. We refer to, in a descending order of contribution, the locus on chromosome 7 as Pms1, and the one on chromosome 3 as Pms2. The approximate map locations of these two loci in the RFLP linkage map are illustrated in Fig. 1.

To determine the amounts of intra- and inter-locus genetic effects of these two loci, a two-way AOV was performed on the basis of marker genotypes at the RG191 and RG477 loci (Table 1). The genetic parameters were estimated for

Fig. 1. The locations of the two loci for photoperiod sensitive male sterility, Pms1 and Pms2, in relation to RFLP markers. The exact map distances of Pms2 from the markers cannot be decided in the present analysis. The relative magnitudes of the effects detected by RG191 and RG348 indicated that Pms2 is located at about 2/5 of the interval between these two markers.

Table 1. A two-way analysis of variance of fertility based on marker genotypes at the RG191 and RG477 loci

================================================
Effect       df       ms         F         p
================================================
1: RG477     2       32548     90.70     0.000
2: RG 191    2        4758     13.26     0.000
1x2          4        1578      4.40     0.002
Error      258         359
================================================

Table 2. The average seed-setting rates on unbagged heads for each of the two-locus genotypes as marked by the two RFLP loci on chromosomes 3 and 7, based on a two-locus model. Allele 1 is from 32001S and allele 2 from Ming Hui 63. The numbers in parentheses are numbers of individuals in different genotypic classes

============================================================
                             RG477
              ==================================  Margin
                 11          12          22
============================================================
RG191   11    12.9 (39)    61.6 (20)  72.1  (14)   52.03
        12    35.9 (43)    73.1 (60)  75.3  (23)   64.36
        22    50.6 (11)    72.7 (36)  74.5  (21)   67.58
Margin        33.84        70.11      74.28
============================================================

these two loci based on marker genotypes using a two-locus model (Table 2). These analyses indicated that the main effect of Pms1 was about three times larger on the average than that of Pms2. There was also considerable amount of interaction between these two loci, and the change in fertility caused by substituting an allele at one locus depended largely on the genotypes of the other locus.

References

Allard, R. W. 1956. Formulas and tables to facilitate the calculation of recombination values in heredity. Hilgardia 24: 235-278.

Tanksley, S.D., M. Causse, T. Fulton, N. Ahn, Z. Wang, K. Wu, J. Xiao, Z. Yu, G. Second and S. McCouch. 1992. A high density molecular map of the rice genome. RGN 9: 111-115.

Yuan, L.P. 1992. (ed) Current Status of Two-line Hybrid Rice Resarch. Agricultural Press, Beijing.