Cloning Of Rice Hybrid Fertility Gene F5, And Development And Application Of High-Throughput Molecular Marker Technologies

Rice hybrid sterility is a kind of plant post-zygotic reproductive isolations. Hybrid sterility between subspecies indica and japonica in rice (Oryza sativa) is widespread, which brings a great difficulty in the utilization of heterosis of inter-subspecies. Wide compatibility varieties (WCVs), a special group of rice germplasm resources, are able to produce fertile hybrids when crossed to both indica and japonica. f5, located on rice chromosome 5, is a major QTL for hybrid pollen fertility and spikelet fertility in inter-subspecies crosses. One of the present studies is to clone f5 gene using map-based cloning method, with the help of genetic analysis, comparative sequencing, gene expression analysis and so on.Molecular marker technology is one of the most important tools in molecular biology research. Traditional molecular marker technologies, such as restriction fragment length polymorphism (RFLP) and simple sequence repeat (SSR), are of low-throughput and time-and cost-consuming, and also produce low-density genetic maps that are not able to provide enough information in QTL analysis and so on. Molecular marker technologies based on gene array or sequencing technology can provide high-throughput and massive molecular markers. The second study is to construct high-density genetic linkage maps based on single feature polymorphism (SFP) markers obtained from expression microarray data and SNP markers obtained from population sequencing data, respectively, using the most studied population in our lab, Zhenshan 97/Minghui 63 recombinant inbred lines (RILs). The main results are as follows:1. It was found that f5 gene influenced hybrid spikelet fertility and the effect sizes varied in different crosses, based on the analyses of the incomplete diallel crosses among 12 rice varieties, three-cross population of 02428/Nanjingll//Balilla, F2 populations derived from the crosses between indica and japonica, f5 near-isogenic line (NIL) ZS97(f5-Du/f5-ZS). We primarily mapped f5 to a 13 kb region of 78-91 kb on Nipponbare BAC clone P0008A07, using the population from the cross between NIL ZS97(f5-Du/f5-ZS) and Balilla, with the help of previous studies. The region has two candidate genes, the first ORF (ORF1) and the second ORF (ORF2). ORF1 encodes a hypothetical protein and ORF2 encodes an ankyrin repeat (ANK) containing protein.2. We performed the comparative sequencing of several rice varieties on the-100 kb genomic region of ORF1, ORF2 and the surroundings, and the result showed that there were three alleles at f5 locus, indica-like, japonica-like and neutral (wide-compatibility gene-like) alleles. Genomic sequences of the varieties with the same allele were consistent, while the sequences of different alleles were divergent, including many SNPs and InDels (insertions/deletions), in addition to different copy numbers and repeats. Analysis of gene expressions showed that the two genes, ORF1 and ORF2, were highly expressed in young panicles of both indica and japonica, but their expressions were not detected in the entire life cycle of WCVs.3. Analysis of the test-crosses between the recombinants of the NIL ZS97(f5-Du/f5-ZS) and the NIL Balilla(S5-Du/S5-Du) containing homozygous S5 alleles from Dular, showed that the spikelet fertilities of the plants from the recombinants flanking ORF1 and ORF2 were segregated as expected, while those between ORF1 and ORF2 were volatile and contradictory. And more, the genotypes of the recombinants between ORF1 and ORF2 in NIL Balilla(f5-NJ/f3-Ba) were disordered, and the physical positions of the markers located on Nipponbare genome were not consistent with their genetic positions based on recombination. It is inferred that the complex genetic mechanism of f5 may be associated with the complex sequence structure in f5 region.4. Using the 204 InDel markers, we constructed the whole genome genetic map of the Balilla/Nanjingll F2 population. We performed genome-wide analysis of the allele segregation distortion, pollen fertility and spikelet fertility. Five loci for segregation distortion,2 QTLs for pollen fertility (S-c and f5) and 3 QTLs for spikelet fertility (S-c,f5 and S5) were identified and verified.f5, the major QTL with the largest effect on both spikelet fertility and pollen fertility, had no significant allele segregation distortion in Balilla/Nanjingl 1 F2 population, as well as other F2 populations derived from the crosses between indica and japonica. 5. We used expression microarray data generated from seedling shoots of 110 RILs and the parents Zhenshan 97 and Minghui 63 at 72 h after germination to develop SFP markers and construct genetic map. We employed median polish, PAM clustering and Z-score formula to detect SFP markers. After filtering and processing, a total of 1632 SFP markers were obtained. After adding the 23 PCR-based markers to fill the gaps,1655 markers were obtained and were grouped into 601 recombination bins. The genetic map based on the bins was constructed and the total length was 1459 cM, with an average resolution of 2.43 cM per bin. Using this method we identified the exact parents of the RILs, Zhenshan 97 and Minghui 63, and develop a set of PCR-based markers to identify and distinguish them from the wrong parents.6. A total of 270,820 high quality SNPs were identified based on the genome-wide low-coverage (-0.06-fold) sequences of the 241 RILs generated using Illumina/Solexa new generation sequencing technology, yielding a genome-wide SNP density about 1 SNP/1.37 kb. We used the SNPs to genotype all the RILs. The RILs with SNP genotyping in disagreement with previous RFLP/SSR genotyping, the ones with unexpected high ratio of heterozygous genotypes, and the redundant RILs were excluded, resulting in the remaining 210 RILs. A total of 1619 recombination bins were obtained from the genotypes of the 210 RILs. A genetic linkage map based on the bins was constructed, which was 1625 cM in length, approximately 1.0 cM per bin, corresponding to 230 kb. Compared with the previous RFLP/SSR genetic map, the sequence-based SNP bin map was of high density and highly accurate. Using the SNP bin map, OsCl for rice apicule color, GS3 for grain length and GW5/qSW5 for grain width were accurately delimited to genomic regions<200 kb. In analyzing rice yield and yield-component traits, the SNP map detected more QTLs especially for grain weight, with precise map locations, demonstrating advantages in detecting power and resolution relative to the RFLP/SSR map.