2016 Volume 66 Issue 3 Pages 367-371
In order to know the genetic nature of hybrid sterility further, three populations, a BC4F2 population derived from Oryza nivara crossed with Yundao 1, a BC4F2 population derived from O. rufipogon crossed with Yundao 1, and a BC10F1 population derived from a cross between O. barthii and Dianjingyou 1 were developed, respectively. Three hybrid sterility QTLs, qHS-6a, qHS-6b, and qHS-6c, detected from those three populations, were mapped into the region between RM190 and RM510, RM190 and RM3414, RM190 and RM587 on chromosome 6, respectively. These QTLs showed collinearity, and explained 88.24%, 61.52%, 44.46% of the phenotypic variance in pollen fertility and 80.60%, 35.20%, 29.01% of the phenotypic variance in spikelet fertility, respectively. In all three crosses, the gametes carrying Yundao 1 or Dianjingyou 1 alleles were eliminated by gametes carrying the wild species alleles. Comparison of the location and the mode of gene action of three QTLs correspond to the S1 locus indicates a common and conserved hybrid sterility locus in AA genome specie playing an important role in reproductive barriers in Oryza. Fine mapping of these QTLs would lead to understand the micro-collinearity and evolutionary relationship among Oryza species.
Rice is one of the most important staple food crops and feeds more than half of the world population. In recent years, rice yield became plateau due to the narrow genetic basis of parental materials (Tanksley and McCouch 1997). Two cultivated rice species Oryza sativa and O. glaberrima domesticated from wild species, only retained 10–20% of genetic diversities in wild species (Zhu et al. 2007). Besides the two cultivated species, six wild species belonging to AA genome are the most accessible genetic resources for the improvement of cultivated rice.
But hybrid sterility hinders the transfer of useful traits from wild species to cultivated species. Mapping hybrid sterility locus and understanding the genetic nature of hybrid sterility can help to break reproductive barrier and introgress the useful alleles from AA genome wild species for rice breeding. Up to now, several sterility loci have been reported, and S1 gene, an eliminator for both male and female gametes in the presence of modifier genes (Sano 1990), was found to be responsible for hybrid sterility in the cross combination between O. glaberrima and O. sativa. The genetic pattern of S1 was explained by the one locus sporo-gametophytic interaction model (Sano et al. 1979). S10 inducing both male and female gametes abortion in a cross between indica and japonica was mapped into the similar position to S1 on chromosome 6 (Sano et al. 1994) and some studies indicated that S1 and S10 were allelic to each other (Heuer et al. 2003, Zhu et al. 2005).
In this study, three different interspecific populations were developed, including BC4F2 populations for O. nivara (Acc. 102176) and O. rufipogon (Acc. 106138), and a BC10F1 population derived from a cross of japonica variety Dianjingyou 1 with O. barthii (Acc. 105987). QTLs were identified for hybrid sterility using these mapping populations and SSR markers.
Yundao 1 and Dianjingyou 1 are japonica irrigated variety from Yunnan province, P. R. China. O. nivara (IRGC102167), O. rufipogon (IRGC106138), O. barthii (IRGC105987) were introduced from the International Rice Research Institute (IRRI). O. nivara (IRGC102167) and O. rufipogon (IRGC106138) as donor and male parents, were crossed with Yundao 1. Two self-fertilized populations were developed, 2012H2E836 (Acc. 102176/Yundao 1/4/ Yundao 1, BC4F2) and 2012H2E840 (acc. 106138/Yundao 1/4/Yundao 1, BC4F2). The population size are 219 and 163 plants, respectively. O. barthii (IRGC105987) as donor and male parent, was crossed with Dianjingyou 1, and 116 individuals of a back-crossed population (2010H3E855, Acc. 105987/Dianingyou 1/10/Dianjingyou 1, BC10F1) were developed. All plant materials were grown in the paddy field at the Winter Breeding Station, Yunnan Academy of Agricultural Sciences (YAAS) located in Sanya, Hainan Province, P. R. China. Field management followed essentially the agricultural practice in this district.
PhenotypingPollen fertility was determined using anthers collected 1 to 2 days before anthesis and stored in 70% ethanol (Doi et al. 1998). Six anthers of a spikelet was stained with 1% I2-KI solution, three views were observed by light microscope and pollen grains were classified as typical abortion, spherical abortion, stained abortion and fertile grain. Pollen fertility was the percentage of fertile grains in all pollen grains. The spikelet fertility was determined for each plant by counting fertile and sterile spikelets on the upper half of three panicles after mature.
DNA extraction and PCR amplificationDNA was extracted from fresh leaves of each plant following the method of Edwards et al. (1991). The SSR markers were selected based on published molecular map of rice (McCouch et al. 2002). Amplification reactions were performed in 10 ul containing 10 ng DNA template, 1 × buffer, 0.5 μmol/L of each primer, 50 μmol/L of dNTPs and 0.5 U of Taq polymerase. PCR programmed as 94°C for 5 min, followed by 32 cycles of 94°C 30 s, 55°C 30 s, 72°C 40 s, final extension of 7 min at 72°C. The products of PCR amplification were electrophoresed on 8% polyacrylamide gel in 1 × TBE.
QTL analysisFor the QTL analysis, linkage maps of the three different populations were constructed using MAPMAKER 3.0 with a minimum LOD score of 3.0 (Lander et al. 1987). The QTL responsible for pollen fertility and spikelet fertility was identified by the interval mapping analysis using the QTLCARTOGRAPHER software package (Basten et al. 1998). The LOD threshold significance level was determined from 1,000 permutation tests, as implemented by the QTL Cartographer (Churchill and Doerge 1994).
In the two BC4F2 populations, the pollen fertility and spikelet fertility showed continuous and bimodal distribution (Fig. 1). In the 2012H2E836 population derived from O. nivara, peak values of pollen fertility were observed at 50% and 100% respectively, and peak values of spikelet fertility were observed at 50% and 80%, respectively (Fig. 1A). In the 2012H2E840 population derived from O. rufipogon (BC4F2), peak values of pollen fertility were observed at 50% and 100%, respectively, and peak values of spikelet fertility were observed at 50% and 80%, respectively (Fig. 1B). In the 2010H3E855 population derived from O. barthii (BC10F1), pollen fertility and spikelet fertility showed bimodal and trimodal distribution, respectively. Peak values of pollen fertility were observed at 60% and 100%, respectively, but peak values of spikelet fertility were observed at 20%, 50% and 100%, respectively (Fig. 1C). Correlation analysis showed that the pollen fertility was significantly correlated with the spikelet fertility in all three populations (Table 1). This result indicated that major QTL control both male and female gametes sterility simultaneously in all three populations.
Frequency distribution of pollen and spikelet ferility in the three populations, 2012H2E836 (IRGC102167/Yundao 1/4/Yundao 1, BC4F2), 2012H2E840 (IRGC106138/Yundao 1/4/Yundao 1, BC4F2), 2010H3E855 (IRGC105987/Dianjingyou 1/10/Dianjingyou 1, BC10F1), respectively.
| Populations | Cross combinations | Donor parent | Generation | Population size | Coefficient of correlation | Level of significance |
|---|---|---|---|---|---|---|
| 2012H2E836 | IRGC102167/Yundao 1/4/Yundao 1 | O. nivara | BC4F2 | 219 | 0.79** | 0.01 |
| 2012H2E840 | IRGC106138/Yundao 1/4/Yundao 1 | O. rufipogon | BC4F2 | 163 | 0.59** | 0.01 |
| 2010H3E855 | IRGC105987/Dianjingyou 1/10/Dianjingyou 1 | O. barthii | BC10F1 | 116 | 0.68** | 0.01 |
A set of 452 SSR markers distributed in rice genome were used to detect polymorphism between parents. In the 2012H2E836 population, 13SSR markers were polymorphic and distributed on chromosome 6, 8, 11, respectively. Genotyping was carried out on 219 individuals besides pollen fertility and spikelet fertility investigation, one major QTL for pollen fertility and spikelet fertility was detected on chromosome 6, designated as qHS6-a. qHS6-a was restricted to the region between the SSR markers RM190 and RM510, explaining 88.24% of the phenotypic variance in pollen fertility and 80.60% of the phenotypic variance in spikelet fertility (Table 2). In the 2012H2E840 population, 44 SSR markers were polymorphic and distributed on chromosome 1, 2, 3, 5, 6, 7, 9, 10, 11, 12, respectively. Genotypes of 163 individuals were investigated besides pollen fertility and spikelet fertility. QTL analysis indicated that a QTL for pollen fertility and spikelet fertility close to RM190 on chromosome 6, designated as qHS6-b, was mapped into the region between the SSR markers RM190 and RM3414, explaining 61.52% of the phenotypic variance in pollen fertility and 35.20% of the phenotypic variance in spikelet fertility. In the 2010H3E855 population, four SSR markers on chromosome 6 were polymorphic, genotypes and phenotypes of 116 individuals were investigated. QTL analysis indicated that a QTL for pollen fertility and spikelet fertility close to RM190, designated as qHS6-c, was located in the region between the SSR markers RM190 and RM587, explaining 44.46% of the phenotypic variance in pollen fertility and 29.01% of the phenotypic variance in spikelet fertility. qHS6-a, qHS6-b and qHS6-c were all mapped close to RM190 on chromosome 6 (Fig. 2), where the known S1 was mapped (Sano 1990, Zhu et al. 2005).
| Populations name | QTL | Flanking marker | Trait | LOD | Variance explained (%) | Additive effect |
|---|---|---|---|---|---|---|
| 2012H2E836 | qHS6-a | RM190-RM510 | Pollen fertility | 80.52 | 88.24 | −0.61 |
| Spikelet fertility | 56.19 | 80.60 | −0.23 | |||
| 2012H2E840 | qHS6-b | RM190-RM3414 | Pollen fertility | 30.47 | 61.52 | −2.84 |
| Spikelet fertility | 15.35 | 35.20 | −1.68 | |||
| 2010H3E855 | qHS6-c | RM190-RM587 | Pollen fertility | 13.95 | 44.46 | 26.15 |
| Spikelet fertility | 7.72 | 29.01 | 42.09 |
Co-linear analysis among qHS6-a, qHS6-b, qHS6-c and S1 locus on chromosome 6.
A common segregation distortion was found in the mapping region for all three populations (Table 3). Both 2012H2E836 and 2012H2E840 are BC4F2 populations in Yundao 1 background. Segregation ratio of Yundao 1 homozyge, heterozyge, and wild rice homozyge did not follow 1:2:1 ratio, which indicates that for qHS6-a and qHS6-b, the interaction between Asian cultivated rice allele and wild rice allele leads to the partial abortion of male and female gametes carrying the allele of cultivar in the heterozygotes. However, a number of homozygotes of Yundao 1 existed in the two BC4F2 populations indicated that qHS6-a/qHS6-b wild alleles did not eliminate qHS6-a/qHS6-b Yundao 1 alleles completely (Table 3). For qHS6-c, the number of homozygotes of cultivated rice Dianjingyou 1 at RM190 was significantly lower than that of heterozygotes (Table 2), which indicated that most cultivar alleles were eliminated by the wild alleles, too. Thus, most gametes carrying Yundao 1 or Dianjingyou 1 alleles were eliminated by gametes carrying the wild species alleles in all three populations derived from O. nivara, O. rufipogon, and O. barthii. These phenomena were explained by one-locus sporophyte-gamete interaction model (gamete elimination) (Sano 1990). Meanwhile, in all 3 populations, QTLs for pollen grain fertility were all detected in the same interval as spikelet fertility (Table 2), which again indicates that same QTL controls both pollen grain fertility and spikelet fertility.
| Populations | Generation | Marker | No. of genotypea | Ratio | χ2 | ||
|---|---|---|---|---|---|---|---|
| AA | Aa | aa | |||||
| 2012H2E836 | BC4F2 | RM190 | 24 | 80 | 115 | 1:2:1 | 57.73 |
| 2012H2E840 | BC4F2 | RM190 | 7 | 56 | 100 | 1:2:1 | 79.67 |
| 2010H3E855 | BC10F1 | RM190 | 4 | 112 | – | 1:1 | 100.55 |
We identified that hybrid sterility QTLs in the populations derived from O. nivara, O. rufipogon, and O. barthii were restricted to RM190 and RM510, RM190 and RM3414, RM190 and RM587 about 4.27 cM, 4.47 cM, 0.8 cM region, respectively (Fig. 2). There existed significant collinearity among them since all on chromosome 6 and linked to RM190.
S1 is the most important hybrid sterile locus responsible for both pollen and spikelet sterility in rice (Sano 1990). By now, S1 locus was reported in many different hybrid combinations, including O. sativa × O. glaberrima, O. sativa × O. longistaminata (Chen et al. 2009, Sano et al. 1979), but has not been reported from other AA genome relatives (O. nivara, O. rufipogon, and O. barthii). In this study, we mapped three hybrid sterility QTLs; they seem to be the major loci controlling the male and female gametes sterility in the crosses between O. sativa and O. nivara, O. sativa and O. rufipogon, O. sativa and O. barthii. The gametes carrying Yundao 1 or Dianjingyou 1 alleles were eliminated by gametes carrying the wild species alleles, this phenomenon was explained by one-locus sporophyte-gamete interaction model (Sano 1990). Moreover, these three QTLs explained 88.24%, 61.52%, 44.46% of the phenotypic variance in pollen fertility and 80.60%, 35.20%, 29.01% of the phenotypic variance in spikelet fertility, respectively. In addition, S1 locus was mapped into a 27.8 kb region in chromosome 6 near RM190; S1 and RM190 are 0.18 cM apart (Garavito et al. 2010). Comparison of location and the mode of gene action, qHS6-a, qHS6-b and qHS6-c are likely to correspond to the S1 locus (Fig. 2). They had good co-linear and might be orthologous loci, which means a common and conserved hybrid sterility locus, exited in AA genome species, is highly possible orthologous before divergence of these species from their common ancestor.
Similarly, good co-linear relationships were observed between S22 from O. glumaepatula and S29(t) from O. glaberrima on chromosome 2 (Hu et al. 2006), between S21 from O. glaberrima and O. rufipogon, and S23 from O. glumaepatula on chromosome 7, between S39(t) from O. glaberrima and S36 from O. nivara on the end of short arm of chromosome 12 (Doi et al. 1999, Miyazaki et al. 2007, Sobrizal et al. 2000, Xu et al. 2014), indicating that those loci for interspecific hybrid sterility were conserved in AA genome species, and might play an important role in species maintenance and reproductive isolation.
Hybrid sterility genes for improving interspecific hybrid sterilityIn interspecific hybridization, hybrid inviability and sterility are major obstacles for the utilization of closely related species in breeding programs, impairing the exploitation of the rich genetic diversity found within the genus Oryza. Breaking the reproductive barriers is necessary to transfer the favorable genes from rice relatives to the cultivated rice. There are two ways to over interspecific hybrid sterility. One is to use the interspecific neutral allele or wide compatibility allele, which does not cause gamete abortion in the hybrids. But unfortunately, the neutral alleles, except for S6, have no any effect on overcoming interspecific hybrid sterility (Heuer et al. 2003, Koide et al. 2008). Another way is to introgress the hybrid sterile allele from wild rice into the recurrent parent (cultivated rice) as the introgression lines (ILs), then it might be more accessible to overcome the hybrid sterility by crossing between the ILs as the genetic bridge parents and the wild rice (Tao et al. 2003). Previous report confirmed that O. sativa lines carrying S1-g allele from O. glaberrima can be used as bridge parents to significantly improve the fertility of hybrids between O. glaberrima and O. sativa (Deng et al. 2010). In this study, three major QTLs were identified. Introgression of qHS6-a, qHS6-b, qHS6-c into the cultivated rice as the bridge parents will help to overcome interspecific hybrid sterility between O. sativa and O. nivara, O. rufipogon, O. barthii, respectively, and to introgress the favorable genes from wild rice to cultivated rice.
This research was supported by National Natural Science Foundation of China (Grant Nos. U1502265, U1036605, 31000704, 31201196), Yunnan Provincial National Science Foundation, China (Grant Nos. 2013FA056, 2011FB118, 2015HB079).
