| Cotton is an important economic crop because its fiber is an important raw material for textile industry. The cultivated cotton includes two allotetraploid species: Gossypium hirsutum and G. barbadense. Upland cotton(G. hirsutum), has been widely grown in the world(more than 90% of the worldwide cotton planting area) for high yield and wide adaptability. The sea island cotton is famous for good fiber and disease resistance of Verticillium wilt. Interspecific hybridization combining their merits is a promising technology approach for cotton genetic improvement. Since reproductive isolation between G. hirsutum and G. barbadense, hybrid breakdown is widely found, such as growth period delay, hybrid sterility. Combination for objective traits has been a difficult work for the interspecific hybridization between G. hirsutum and G. barbadense. To dissect the genetic basis of reproductive isolation, three objects on this project were implemented in our study:(1) Identification and characterization of segregation distortion loci by reciprocal backcross populations and F2 populations.(2) QTL mapping for important agronomic traits on the segregation distortion chromosomes in the interspecific backcross populations.(3) The correlation analysis between genomic heterozygosity and hybrid breakdown in cotton.(1) Identification and characterization of segregation distortion loci by reciprocal backcross populations and F2 populationsIn order to dissect genetic mechanism of the segregation distortion(SD) on the segregation distortion chromosomes of cotton(Chr 2, Chr 16 and Chr 18), eight reciprocal backcross populations and two F2 populations were developed. The results showed that 21, 50 and 23 markers of distorted segregation were detected on chromosome 2, 16 and 18, respectively.By means of the EM(Expectation- maximization) method, 48 SDL were detected in ten populations; 10, 27 and 11 SDL were detected on chromosome 2, 16 and 18, respectively.The SDL 2.1 was detected in male-segregating population [(3E) E,(3E) 3] and female-segregating population E(E3) and(E3) F2 population, implying that the SD was caused by zygotic selection, suggesting that the heterozygote was preferred in the stages of the zygotic embryo under competition.The SDL 2.6, located on the 69.63 c M on the chromosome 2, was detected in male-segregating population [(E3) 3 and(3E) 3], not distorted in the reciprocal female-segregating backcrosses population, implying that the SD was caused by male gametic selection; meanwhile, this SDL was detected in the two F2 populations.The SDL16.14 was detected in female-segregating population [(E3) E,(E3) 3,(3E) E] and male-segregating population [E(E3), 3(E3) and 3(3E)], not distorted in the the two F2 populations, implying that the SD was caused by gametic selection.The SDL16.1 was detected in male-segregating population [E(3E) and 3(E3)], implying that the SD was caused by male gametic selection. Since the SDL was not distorted the reciprocal populations [3(3E) and E(E3)], suggesting that the SD was affected by stigma and embryo- sac effect.The SDL18.1 were distorted in the two male-segregating backcross populations [E(E3),(LOD = 6.06), E(3E)(LOD = 6.67)] and the(3E) F2 population. In the backcross populations the skew direction was towards E22 homozygotes, which imply that the pollens containing the E22 allele were preferred to the pollens containing the 3-79 allele under competition; but not distorted in the reciprocal female-segregating backcrosses populations [(E3) E and(3E) E], implying that the SD was caused by male gametic competition. Meanwhile, SDL18.1 was not distorted in other male-segregating backcrosses populations [3(E3) and 3(3E)], implying that the SD was affected by stigma and embryo- sac effect. In the(3E) F2 population, the skew direction was towards 3-79 homozygotes, which imply that the zygotic selection may be involved the SD, hat is, 3-79 homozygotes were preferred in the stages of the zygotic embryo.The SDL18.8 was detected in all the backcross populations and F2 populations, implying that the SD was caused by zygotic selection. The skew direction was towards heterozygotes in these populations, suggesting that the heterozygote were preferred under competition, as a consequence, zygotic selection such as the differentiation of zygote viability or heterosis should be the genetic mechanism for the SD.(2) QTL mapping for the important agronomic traits on the segregation distortion chromosomes of in the cotton interspecific backcross populationsThe four BC1F2 populations were derived from BC1F1 population [E(E3), E(3E),(E3) E,(3E) E] including a significant proportion of G. hirsutum genetic constitution. QTL mapping for yield component, earliness and fiber quality on the chromosomes of the segregation distortion in these backcross populations. The transgressive segregation for all studied traits was found in the four populations.Normal distribution test showed that seven fiber- related traits were accorded with normal distribution except Micronaire value in E(E3) population, but earliness- related traits were not accorded with normal distribution, suggesting that the genes for fiber- related traits are quantitative traits gene, nonetheless, earliness- related traits may have a genotype- environment interaction effect. Correlation analysis showed that fiber- related traits(except FL) were not significantly correlated between earliness- related traits.Base on composite interval mapping of Windows QTL Cartographer 2.5, a total of 47 QTL were found on these three chromosome(Chr 2, Chr 16, Chr 18), namely, 15 in(E3) E, 13 in(3E) E, 12 in E(E3), 7 in E(3E), with PVE, LOD ranging from 8.77% ~ 30.92%, 2.53 ~ 7.47. A total of 27 fiber- related QTL were found, namely, 5 for FL, 4 for FS, 2 for MIC, 4 for FLU, 3 for FM, 4 for SCI, 5 for SFR, with Additive, PVE, LOD ranging from- 18.98 ~ 12.73, 9.00% ~ 30.92%, 2.53 ~ 7.47. A total of 5 yield- related QTL were found, namely, 3 for BW, 2 for LP, 2 for LI, with Additive, PVE, LOD ranging from- 0.69 ~ 2.22, 8.77% ~ 17.35%, 2.66 ~ 3.57. A total of 13 earliness- related QTL were found, namely, 5 for SF, 4 for DBO, 4 for FBO, with Additive, PVE, LOD ranging from- 6.50 ~ 9.65, 9.40% ~ 19.40%, 2.56 ~ 5.04.Comparison analysis showed, most of the QTL in our study were consistent with the meta QTL, such as BW, FL, FLU, LP, MIC on chromosome 2, FL, FS, FU on chromosome 16 and BW, DBO, FL, FS, FLU, MIC on chromosome 18. The several QTL detected in our study was not found in meta QTL, such as DBO, LP on chromosome 16 and SF on chromosome 18.(3) The correlation analysis between genomic heterozygosity and hybrid breakdown in cotton(Gossypium)To research the correlations between genomic heterozygosity and hybrid breakdown in cotton interspecific populations, we developed reciprocal F2 populations. We have investigated hybrid breakdown of vegetative and reproductive traits in reciprocal F2 populations. To guarantee mutual independence among the genotyed markers, a total of 125 SSRs were randomly selected from the 2316 genome- wide loci according to the genetic distance in the interspecific genetic map previously constructed by the same parents, independent assortment between the loci must be validated by the check of LD analysis. Individual genomic heterozygosity was estimated after this check. We have investigated the relationship between genomic heterozygosity and vegetative and reproductive traits.Descriptive statistical parameters for each trait showed only two traits displayed a normal distribution, PH(p = 0.20) in(3E) F2 and SI(p = 0.69) in(E3) F2, others displayed non- normal distributions. The results demonstrated that hybrid breakdown may exist in the interspecific reciprocal F2 populations particularly with respect to reproductive traits such as infertility and bare seeds.Correlations analysis indicated correlations among the vegetative and reproductive traits in the(E3) F2 population were non- significant; however, this was not the case in the(3E) F2 population. The average MHI in each population was less than the expected value(0.5), indicating that the genomic composition from 3-79 was greater than that from E22, which suggests that the alleles from 3-79 may have a stronger selective advantage.Pearson‘s correlation analysis indicated MHI and BN(r = 0.54, p = 0) as well as between MHI and SW(r = 0.70, p = 0) were observed in the group(MHI < 0.5) from the(E3) F2 population.One- way analysis of variance indicated significant differences were observed between BN(F5, 134 = 4.925, p = 0.00036) and SW(F5, 55 = 10.59, p = 3.72e- 07) and individual genomic heterozygosity level in the(E3) F2 population. No significant differences were observed between plant traits and individual genomic heterozygosity levels in the(3E) F2 population. These results suggested that the G. barbadense cytoplasm background could exhibit better compatibility compared with that of G. hirsutum. Our results may therefore offer new insights into hybrid breakdown in allotetraploid cotton interspecific hybrids and may provide a fresh perspective on interspecific hybridization for the genetic improvement of cotton. |
PDF Full Text Request