| As a new source of germplasm, the Honglian cytoplasmic male sterile (HL-CMS) rice has been demonstrated to have great potential for hybrid rice production. A variety of hybrid rice cultivars derived from HL-CMS lines have been developed, which show good grain quality, high yields and enhanced resistance to disease. It is also thought that HL-CMS/Rf system, together with WA-CMS/Rf/and BT-CMS/Rf/systems in rice, is one of the suitable models in theoretic study on the molecular mechanism of heterosis and nucleolus-cytoplasmic interaction.Previous studies on HL-CMS/Rf system indicated that a mitochondrial gene named as orfH79, which is highly homologous to orf79 for BT-CMS, and an Rf/locus, Rf5, in the restorer line Miyang 23, which was located on the long arm of chromosome 10, might be responsible for the sterility and fertility-restoration in HL-CMS/Rf system. In this study on HL-CMS/Rf system, a sterile mitochondrial candidate gene, named HL-spl, and a novel fertility-restoring gene, designated Rf6(t), for Honglian gametophytic cytoplasmic male sterility in rice (Oryza sativa L.) were discovered.1. The DNAs isolated from the sterile lines, maintainer lines and restorer lines, as well as the lines with same nucleolus of Shijing B (SJB) and different cytoplasms, of HL-CMSIRf WA-CMSAR/and BT-CMS/Rf systems were used for RAPD analysis. There were eight CMS types-pecific or line-specific RAPD fragments, AJ-03/530, AB-18/750,0-12/520, AO-13/1600, U-18/2100, AB-05/200, OPN-09/700 and OPN-12/800 selected out of 264 RAPD primers. These eight RAPD fragments can be used as reliable PCR-based markers to identify three different types of three-line hybrid rice combinations derived from HL-CMS, WA-CMS and BT-CMS lines, respectively, and most of their parents. In addition, the SCAR markers, hlu-17 and hlu-18, derived from two of RAPD fragments, facilitate marker-aided selection (MAS) in the breeding of HL-CMS/Rf/system.The HL-CMS/Rf system-specific SCAR markers, hlu-17 and hlu-18, were further used for polymorphic analysis between mitochondrial genomic DNAs (mtDNAs) of sterile lines (CG-41A, ZSA and YTA), hybrids (C-M23, HL-2 and SY63), a maintainer line (YTB), and a restorer line (PC311), and one mitochondrial genome-associated candidate gene, HL-spl, for HL-CMS was screened. The HL-spl is a chimeric gene, which is composed of an abnormal atp6 gene and an unknown functional sequence. Sequence comparison showed that the abnormal atp6 gene has an 1141 bp deletion from 852 to 1994 bp in normal atp6 (N-atp6) gene. Moreover, 167 bp in the unknown functional sequence region of HL-spl is same as the 1062 to 1228 bp region in mitochondrial ribosomal protein L5 gene cDNA in rice. In addition, sequence comparison also indicated that the candidate gene HL-spl for HL-CMS has no identities with the previously identified mitochondrial gene orfH79.RT-PCR analysis showed that N-atp6 gene expression could be detected only in the anther tissue at mononuleate to dinucleate stages of YTB and 9311 with the fertile cytoplasm, whereas, it could not be detected in the anthers of YTA and HL-2 with HL-CMS cytoplasm at same stages. Sequence comparison demonstrated that the cDNA from abnormal atp6 gene expression has an 1141 bp deletion compared with that from N-atp6 gene expression. This result was same as that of the sequence analysis at mtDNA level at which 1141 bp deletion had been detected in abnormal atp6 gene. Gene expression analysis suggested that the abnormal function of the chimeric mitochondrial gene HL-spl might be responsible for HL-CMS.Using HL-spl as a marker, the distribution of it was studied in 36 rice materials,including 7 lines of cultivars and 29 wild rice materials with AA genome. HL-spl was detected only in some of the O. rufipogon wild rice materials, from which HL-CMS cytoplasm was originally found.2. Genetic study on fertility restoration of HL-CMS lines was carried out and two fertility-restoring loci for HL-CMS were discovered. Two HL-CMS lines, CG-41A and YTA, their maintainer lines CG-41B and YTB, and two fertility-restoring lines, MY23 and 9311, were used to construct four F, populations of CG-41A/MY23, CG-41 A/9311, YTA/MY23 and YTA/9311 and four testcross populations of CG-41A // CG-41B/MY23, YTA // YTB/9311, YTA // YTB/MY23 and YTA // MY23/9311 for this study. The F, plants obtained from the crosses between CMS lines and restorer lines, CG-41 A/MY23, CG-41A/9311, YTA/MY23 and YTA/93-11, appeared fertile, producing approximately 50% fertile pollens, and no sterile plants were identified in the selfing F2 populations, indicating that the tested HL-CMS lines were restored gametophytically. The segregation of fertility in three BCFi populations, CG-41A // CG-41B/MY23, YTA // YTB/9311 and YTA // YTB/MY23, was compatible with Mendelian inheritance of a single gene in each case, x2 value were 0.8812, 0.6480 and 1.1571, respectively. The Fi plants from the cross between the two restorer lines, MY23/9311, were completely male fertile. In the population obtained from the testcross YTA // MY23/9311, however, the segregation of fertile and sterile plants did not agree with expectations for the Mendelian inheritance of a single gene (where no sterile plants are expected), or with the 3:1 ratio expected for the independent segregation of two genes. The result suggested that there are two linked Rf loci, one from MY23, designated as Rf5 and another from 9311, designated as Rf6(t), which act on the HL-CMS cytoplasm. The calculated recombination frequency was 8.57%.3. To characterize the linkage relationship between the two i?/loci from MY23 and 9311, genetic linkage analysis was carried out using RAPD, SSR and SBD markers. Bulked segregation analysis (BSA) was used to detect the DNA polymorphism of those markers. Two RAPD markers, seven pairs of SSR markers and two SBD markers detected DNA polymorphisms between YTA and 9311, as well as in the sterile and fertilebulks from YTA // YTB/9311. Two RAPD markers and eight SSR markers showed DNA polymorphism between CG-41A and MY23, as well as in the sterile and fertile bulks from CG-41A // CG-41B/MY23. These polymorphic markers were then used on the large extremely sterile populations of 243 plants derived from the BCFi of YTA // YTB/9311 and 227 plants from the BCF, of CG-41A // CG-41B/MY23 for fine-scale mapping of the i?/loci. Linkage analysis was carried out, using those selected markers. Rf5 co-segregated with RM3150, and flanked by RM1108 and RM5373 with the genetic distance 0.9 cM and 1.3 cM on either side of the locus. Rf6(t) co-segregated with RM5373, flanked by RM6737, SDB01, RM6100, RM304 and RM6704 on one side, and by SDB07, RM5841 and OPN-12/800 on the other. The genetic distance from the Rf6(t) to RM6737, SBD01 and SBD07 was 0.4 cM. Fine mapping of Rf5 and Rf6(t) determined that these two i?/loci on the region of chromosome 10 for HL-CMS are different from Rf-1 for BT-CMS and Rf4 for WA-CMS, which were located on the same chromosome.4. Because the two Rf loci were located on chromosome 10, a BAC pooling strategy was employed to construct a contig containing the Rf6(t) locus present in 9311 first. Primer pairs derived from RM6737 and RM5373, SBD01 and SBD07, were used for PCR analysis of the 88 BAC DNA pools prepared from IR64 BAC library of genomic DNA. Five BAC clones were obtained by screening the BAC library using these primer pairs. The BAC clone 15C23 (105 kb) bears the markers SBD07, RM6737 and RM5373. The BAC clone 15C21 (90 kb) includes both RM5373 and SBD07, whereas 44E21 (105 kb) carries SBD07 only. The BAC clones 16J23 (60 kb) and 16J3 (55 kb) carry RM6373 and/or SBD01, respectively. These five BAC clones make up a contig of about 230 kb that spans the Rf6(t) locus in 9311. This result is consistent with the results of fingerprinting analysis and was confirmed by Southern blotting.To refine the physical map of Rf6(t) obtained from the IR64 BAC library, sequences of the fragments obtained by PCR with RM6737, SBD01, RM5373 and SBD07 were used for BLASTX searches in the 9311 genomic sequence database (http://btn.genomics.org.cn/rice/). The results revealed that four 9311 contigs carry these markers, respectively. Further BLASTX analyses using these marker-associated contigsequences to search for overlaps revealed that seven contigs could be assembled into a larger contig. By combining the BLASTX results with the physical mapping data, the physical distances between SBD01, RM5373 and SBD07 could be estimated to be about 30 kb and 36 kb, respectively. Therefore, the Rf6(t) present in 9311 lies within a segment of about 66 kb between the markers RM6373 and SBD07, which is 10 kb shorter than the distance delimited by the same markers in the sequence of Nipponbare chromosome 10.5. Several candidate genes in this region have been predicted by gene-annotation software (http://genes.mit.edu/GENESCAN). These include genes for a putative RNA-binding protein, a microtubule-severing protein and products of unknown function. Expression analysis of those genes was carried out using one-step RT-PCR and 3 seedling-specific expressed genes, 2 methyl jasmonate-induced genes and 3 anther tissue-specific expressed genes were screened. The results from expression analysis of candidate genes for HL-CMS have provided with very useful information to Rf gene cloning.
|
PDF Full Text Request