Understanding the evolutionary forces in speciation is a central goal in evolutionary biology. Asian cultivated rice has two subspecies, indica and japonica, but the underlying mechanism of the partial reproductive isolation between them remains obscure. Here we show a presence-absence variation (PAV) at the Se locus functions as an indica-japonica reproductive barrier by causing hybrid sterility (HS) in indica-japonica crosses. The locus comprises two adjacent genes: ORF3 encodes a sporophytic pollen killer, whereas ORF4 protects pollen in a gametophytic manner. In F1 of indica-japonica crosses, pollen with the japonica haplotype, which lacks the sequence containing the protective ORF4, is aborted due to the pollen-killing effect of ORF3 from indica. Evolutionary analysis suggests ORF3 is a gene associated with the Asian cultivated rice species complex, and the PAV has contributed to the reproductive isolation between the two subspecies of Asian cultivated rice. Our analyses provide perspectives on rice inter-subspecies post-zygotic isolation, and will promote efforts to overcome reproductive barriers in indica-japonica hybrid rice breeding.
Detection and characterization of structural variants (SVs) has revolutionized the understanding of the landscape of genomic and phenotypic variation in different species1, 2. Presence-absence variation (PAV) is an extreme form of SV, in which a genomic segment containing one or more genes is present in some individuals but absent in others3. The unprecedented low cost of DNA sequencing and advances in high-quality genome assembly today make it possible to generate a wealth of data on PAVs in plant genomes4,5,6,7,8,9,10,11,12,13. Pangenomic and gene ontology analyses in crops consistently find that PAVs are enriched for abiotic stress and disease-response genes4,5,9,10,13. Increasing evidence has confirmed that PAV is one of the genetic factors underlying variation for agronomic traits such as submergence tolerance (conferred by Sub1A14, SNORKEL1 and SNORKEL215), phosphous-deficiency tolerance (conferred by Pstol16), anaerobic germination tolerance (conferred by OsTPP717), disease resistance (conferred by LRK118 and OsWAK112d6), and plant architecture (conferred by SLB1 and SLB219). In addition, fixation of complementary PAVs is believed to contribute to heterosis in hybrid breeding20.
Asian cultivated rice (O. sativa L.), which comprises the staple food for half of the global population, was domesticated from common wild rice (O. rufipogon Griff.) in Asia, and has two subspecies, O. sativa L. ssp. indica and O. sativa L. ssp. japonica (hereafter referred as indica and japonica, respectively)21. As two genetically diverged subspecies, indica and japonica varieties display distinct morphological, physiological and biochemical traits, reflecting the adaptive evolution of the two subspecies in divergent natural and human-influenced environments.
Various types of post-reproductive barriers exist between indica and japonica rice accessions22. Hybrid sterility (HS) is the most commonly found reproductive barrier in indica-japonica crosses, and is often cited as the criterion to classify indica-japonica subspecies23. Molecular genetic studies over two decades have revealed three mechanisms of HS in Oryza, which have been designated as the killer-target system, the killer-protector system, and the duplicate gametic lethal system24. These involve 9 single-locus HS loci (Sa25, Sc26, qHMS727, S528, S729, HSA130, ESA131, S132,33, pf12A34) and 3 two-locus HS pairs (DPL1/DPL235, S27/S2836, DSG1/DSG237). Despite extensive studies of these loci, we still lack efficient methods to fully overcome HS in inter-specific and inter-subspecific hybrid rice, and our understanding of the role of HS in genome evolution remains limited.
The yield potential of indica-japonica hybrid rice is estimated to be 15–30% higher than that of the indica intraspecific hybrid rice widely cultivated today. The utilization of strong heterosis in indica-japonica hybrids is one of the most promising future directions in rice breeding24,38,39,40, therefore, understanding indica-japonica reproductive isolation is important to unlock their breeding potential. Se/pf12 is a major locus conferring hybrid male sterility in indica-japonica hybrids, which was identified by a large number of studies using different populations34,41,42,43,44,45,46,47. Artificial disruption of pf12A gene at Se/pf12 locus successfully converts this hybrid sterility locus into a neutral allele34. However, the molecular mechanism of pf12A regulating rice HS remains elusive.
In this work, we show that a PAV at the Se locus functions as an indica-japonica reproductive barrier by causing HS in indica-japonica crosses. The complementary effect of two adjacent genes at Se locus within the PAV region induces hybrid male sterility (HMS), and thus reproductive isolation between indica and japonica. Evolutionary analyses of Se locus show that the PAV is an important contributor to the postzygotic isolation between the two subspecies, and it also supports the independent domestication of indica and japonica from different O. rufipogon populations. Based on these findings, effective approaches could be designed to break down reproductive barriers in indica-japonica hybrid rice breeding.
Identification of Se controlling HMS
In order to identify genes responsible for indica-japonica HS, near isogenic lines (NILs) were developed using japonica variety T65 as the recurrent parent and indica variety GLA4 as the donor parent. In the BC5F2 generation, a NIL, E9, was identified that had four substituted segments in the genetic background of T65 (Fig. 1a and Supplementary Fig. 1). T65, T65/E9 F1 hybrids and E9 exhibited similar plant morphology except for anther length (Fig. 1b and Supplementary Figs. 2 and 3a–c). Pollen staining using both I2-KI and TTC showed that male fertility was normal for T65 and E9; however, the T65/E9 F1 hybrid was semi-sterile, with a mix of fertile and defective pollen (Fig. 1c, d). Pollen grains of T65 and E9 were larger and spherical (Supplementary Fig. 3d, f–h), whereas T65/E9 F1 hybrids produced two kinds of pollen grains: normal, fertile pollen filled with abundant starch granules, and defective pollen filled with cytoplasm but without any starch granules (Supplementary Figs. 3e, i, j and 4a, b). Defective pollen showed thicker exines and intines than fertile pollen, suggesting that pollen wall formation was also affected in T65/E9 F1 hybrids (Supplementary Fig. 4c–e). Semi-thin sections of developing anthers indicated that half of the microspores were aborted in T65/E9 F1 hybrids starting from the bicellular pollen stage (St11) (Supplementary Fig. 5). At the tricellular pollen stage (St12), the aborted pollen resembled St10 vacuolated microspores, whereas normal mature pollen grains were densely stained and filled with starch granules (Supplementary Fig. 5).
Segregation analysis of 167 T65/GLA4 BC5F2 plants during E9 construction revealed severely distorted segregation of two SSR markers (RM19 and RM453) on chromosome 12, with significantly lower than the expected 25% homozygous T65 genotype, while SSR markers on chromosome 2, 3 and 7 showed normal 1:2:1 mendelian segregation (Supplementary Table 1). Test crosses with T65/E9 F1 hybrids as the male parent and T65 or E9 as female parents indicated that the male gametes derived from T65 were selectively aborted and T65 alleles were not transmitted to the progeny (Supplementary Table 2). Conversely, test cross with T65/E9 F1 hybrids as the female parent and E9 as the male parent presented 54:74 segregation ratio of T/E and E/E (Supplementary Table 2), implying that the female gametes carrying T65 allele is also affected, but at a relatively lower degree compared to male gametes, which is consistent with slightly lower spikelet fertility for T65/E9 F1 hybrid (Fig. 1e).
For 279 plants of T65/E9 F3 segregating population, spikelet fertility showed a normal distribution; however, a clear bimodal pattern with an apparent valley at 70–80% was observed for pollen fertility (Fig. 2a, b). Taking 70–80% as the cut-off region between semi-sterility and fertility, the population segregated for semi-sterility (<70%): fertility (>80%) at the ratio of 1:1 (χ2 = 0.13 < χ20.05,1 = 3.84), demonstrating monogenic inheritance for the trait. Our results indicated the substituted segment on chromosome 12 contained a gene locus, designated Se as in previous studies41,42, controlling HMS between T65 and E9, which was closely linked with RM19 and RM453 and appeared to be allelic to several previously identified HMS QTLs34,43,44,45,46,47.
Complex genomic structural variations at Se locus
Linkage analysis of pollen fertility and SSR markers using the above 279 T65/E9 F3 plants pinpointed the Se locus to an interval between IND9 and PSM448 (Fig. 2c). Through analysis of 7,274 F4 plants, Se was further restricted to the genomic region between markers PSM623 and IND48 (Fig. 2d and Supplementary Table 3). Sequence analysis revealed two genes (ORF1T65 and ORF2T65) in the 13.17-kb Se region in T65 (Fig. 2e). ORF1T65 is predicted to encode a 250-aa protein without any conserved domains and ORF2T65 is predicted to encode a 466-aa alpha/beta hydrolase family protein. Sequence comparison revealed nucleotide diversity for ORF1 and ORF2 between T65 and E9. Compared with ORF1E9, a 378-bp deletion in the last exon and 11 single-nucleotide polymorphisms (SNPs) resulted in a 126-aa deletion and 5 aa substitutions, respectively, in ORF1T65 (Supplementary Fig. 6). Although several indels and SNPs exist between ORF2T65 and ORF2E9, causing premature termination in the predicted ORF2E9 (Supplementary Fig. 7), the major difference in ORF2 is in the promoter region, which is completely different between the two alleles. ORF2E9 expression could not be detected in anthers or in any other tested tissues (Supplementary Fig. 8). In E9, a 28.45-kb insertion was found between ORF1E9 and the putative ORF2E9, containing two additional genes, ORF3 and ORF4 (Fig. 2e). ORF3 is predicted to encode a 596-aa superfamily II DNA and RNA helicase with a mitochondrial targeting signal (Supplementary Fig. 9), and ORF4 is predicted to encode a 523-aa protein with ankyrin repeats (Supplementary Fig. 10).
ORF3 encodes a sporophytic pollen killer
In order to identify the genes responsible for HMS, we used the CRISPR/Cas9 genome editing system to knock out T65 candidate genes in T65 or E9 candidate genes in E9, respectively. Unique target sites were selected in the coding region of each gene to make sure loss-of-function mutations were obtained. Three orf3 homozygous mutants (orf3-1 to orf3-3) and two homozygous mutants each for orf1T65, orf2T65, orf1E9 and orf1T65orf2T65 with different editing patterns, respectively, were selected and showed normal pollen viability, suggesting that ORF1, ORF2 and ORF3 are not essential for the development of male gametophytes (Fig. 3a, Supplementary Fig. 11a–d and Supplementary Tables 4, 5). The mutants were further crossed with T65 or E9 to produce mutant-type Se heterozygotes to test HMS. In contrast to the semi-sterile pollen exhibited by T65/E9 F1 hybrids and mutant-type F1 hybrids with orf1T65, orf2T65, orf1E9 or orf1T65orf2T65 alleles (Supplementary Fig. 11e–h and Supplementary Table 4), mutant-type F1 hybrids with the three orf3 alleles exhibited normal pollen fertility (Fig. 3b). Moreover, the F2 families showed the normal 1:2:1 segregation ratio, indicating that loss-of-function of ORF3 restores male fertility in Se heterozygotes and eliminates the segregation distortion (Fig. 3c). These results are in consistent with normal pollen fertility and 1:2:1 segragation of two heterozygous pf12A (allelic to ORF3) mutants34. To further study ORF3’s function, a genomic fragment containing ORF3 or ORF3 overexpression constructs were introduced into T65. In total, 34 out of 41 (genomic ORF3), and 3 out of 28 (ORF3 overexpression) transgenic lines contained detectable transgenes. Surprisingly, both the transgene-positive lines and transgene-negative lines showed normal pollen fertility (Supplementary Table 6). Expression analysis revealed that ORF3 expression in the transgene-positive lines was very low compared with that in T65/E9 F1 hybrids (Supplementary Fig. 12), suggesting that ORF3 is toxic to living cells without ORF4 and its expression should be strictly regulated under T65 genetic background. Direct evidence of ORF3’s killing effect was obtained from the transgenic callus experiment in which the selection frequency of Hyg-resistant callus was extremely low for ORF3 compared with the empty vector control (Supplementary Fig. 13a, b), and transgene-positive calli showed low expression of transgene (Supplementary Fig. 13c), indicating that the growth of T65 callus was retarded or inhibited when ORF3 was introduced. Similarly, the induction of ORF3 in yeast significantly inhibited the growth of the transformants (Supplementary Fig. 14). Our results indicated that ORF3 has a pollen-killing function and acts in a sporophytic manner.
ORF4 performs a pollen-protection function in a gametophytic manner
The selective abortion of T65-type pollen in T65/E9 F1 hybrids but not in T65 suggests that a pollen protector is most likely encoded in E9 to eliminate the pollen-killing effect of ORF3. To test this notion, we transformed intact ORF1E9, ORF2E9 and ORF4 genomic fragments from E9 into T65, respectively. Hemizygous transgenic T0 plants were further crossed with E9 to produce F1 hybrids carrying the pollen killer ORF3 (Fig. 4a). For F1 plants carrying the ORF4 transgene (T/E; ORF4/-), pollen fertility was restored to ~75% (Fig. 4a–c and Supplementary Table 7). By contrast, no effect on pollen fertility was observed in the transgene-negative plants (T/E; -/-) or F1 plants with either the ORF1E9 or ORF2E9 transgene (Fig. 4b and Supplementary Table 7). Moreover, plants in an F2 population derived from F1 plants with one hemizygous ORF4 transgene (T/E; ORF4/-) segregated as expected (χ2 = 1.42 < χ20.05,7 = 14.067) (Fig. 4d, e and Supplementary Table 8). Se-heterozygotes with homozygous ORF4 (T/E; ORF4/ORF4) were fully fertile, and F2 plants with genotypes of (T/T; ORF4/-) and (T/T; ORF4/ORF4) were recovered (Fig. 4d, e and Supplementary Table 8). In contrast, no (T/T) plants could be obtained in segregating population derived from T65/E9 F1 plants (Fig. 3c and Supplementary Table 9). These results indicated that ORF4 rescues T65-type pollen in a gametophytic manner. This notion was further supported by the partial restoration of pollen fertility (~75%) in hemizygous T0 ORF4 overexpression transgenic plants with a T65/E9 genotype (Supplementary Fig. 15 and Supplementary Table 9). In addition, we constructed orf4 mutants using the CRISPR/Cas9 genome editing system. Three hemizygous orf4 mutants were obtained, and produced semi-sterile pollen, as expected (Supplementary Fig. 16 and Supplementary Table 10). Furthermore, no orf4 homozygous mutant could be recovered in T1 progeny of these hemizygous orf4 plants (Supplementary Table 11). In contrast, orf4 homozygous mutants could be easily obtained when ORF3 was also mutated. As expected, orf3orf4 double mutants with T65/E9 genotypes showed normal pollen fertility (Supplementary Fig. 17).
Although both ORF3 and ORF4 localized in cytoplasmic foci (Supplementary Fig. 18), no direct interaction of ORF3 and ORF4 was detected using a yeast two-hybrid system (Supplementary Fig. 19). qRT-PCR analysis revealed that ORF3 and ORF4 are expressed broadly in different tissues and in all stages of anther development. The expression level in E9 was about twice as high as that in the T65/E9 F1 hybrids (Supplementary Fig. 20), reflecting the different copy numbers (one versus two) in these backgrounds.
Evolution of the Se locus
To trace the origin of Se locus, we BLAST searched ORF3 and ORF4 homologs in the Genebank database and found that full-length homologs of ORF3 can only be found in Oryza species, indicating it is a gene arose within the Oryza lineage. In contrast, homologs of ORF4 are widely distributed in Poaceae and they all share the conserved protein domain (Supplementary Fig. 21), suggesting its conserved function across grasses.
To trace the evolutionary history of Se locus in the Oryza genus, we first assayed the presence or absence of the locus by analyzing publicly available high-quality genome sequences of 59 O. sativa4, 6,48 and 1 O. glaberrima6, and PCR amplifying the region spanning part of ORF3 and ORF4 (Supplementary Fig. 22) in a large Oryza panel consisting of 787 individuals from 14 Oryza species. Consistent with the BLAST search analysis, target sequences were successfully identified from 363 out of 830 individuals of AA genome, but failed to be amplified from any individuals of GG (O. meyeriana, 3/3), FF (O. brachyantha, 2/2), EE (O. australiensis, 3/3), CC (O. eichingeri and O. officinalis, 5/5) and BB (O. punctata, 4/4) karyotypes (Fig. 5a and Supplementary Table 12). These results show that the Se locus is polymorphic in the AA rice species, but absent in Oryza species of other karyotypes, suggesting the emergence of the locus is specifically associated with the history of AA-genome Oryza species. Within AA species, the locus can be found in Asian cultivated rice species complex, including O. sativa (230/384), O. nivara (10/44) and O. rufipogon (107/342) and African cultivated rice species complex including O. glaberrima (7/7) and O. barthii (4/4) (Fig. 5a and Supplementary Table 12). Surprisingly, it is also present in the basal species of AA genome, O. meridionalis (5/5), but absent in the rest species including O. longistaminata (28/28) and O. glumaepatula (16/16) (Fig. 5a and Supplementary Table 12). Further sequence analyses on ORF3 in individuals containing the Se locus identified 14 different haplotypes, among which 13 haplotypes confer truncated form of ORF3 protein, and the haplotype conferring full length protein can only be found in Asian cultivated rice species complex (Supplementary Data 1). We tested the effect of truncations by verifying the function of a common truncation form (ORF3Hap-2) and found that its subcellular localization was disrupted (Supplementary Fig. 18) and the toxicity effect was abolished in transgenic experiments in yeast and rice callus (Supplementary Figs. 13, 14). These results show that the truncation renders ORF3Hap-2 non-functional and they also suggest that full-length sequence might be necessary for the ORF3 function, and if so, the functional ORF3 is confined to the Asian cultivated rice species complex, with a significant enrichment of accessions from O. sativa indica subspecies. In contrast, truncated haplotype at ORF4 is extremely rare, which is only found in one indica individual out of 770 accessions assayed in the panel consisting of O. sativa, O. nivara and O. rufipogon (Supplementary Data 2). A 369-bp sequence presence/absence variation was identified in the promoter of ORF4 in the panel (Supplementary Data 2), and the deletion of the sequence did not affect the subcellular location of ORF4 (Supplementary Fig. 23) but significantly reduced the expression level (Supplementary Fig. 24), suggesting it is a functional variant in the rice population. A total of 37 haplotypes of the Se locus were identified when ORF3 and ORF4 were considered together, and could be further divided into five Se types according to the function and origins of ORF3 and ORF4 (Fig. 5a and Supplementary Data 3). Types 1 to 3 exist in Asian cultivated rice species complex of O. sativa, O. nivara and O. rufipogon, with Type-1 and Type-2 as the two major types representing 27.55% and 64.19% of the rice accessions with the PAV region, respectively (Fig. 5a and Supplementary Tables 13, 14), while Type-4 and Type-5 Se are haplotypes of African cultivated rice species complex and O. meridionalis, respectively (Fig.5a and Supplementary Tables 13, 14). In sum, our analyses suggest that ORF3 is possibly formed in the Asian rice lineage, and its functional form is maintained in O. sativa indica subspecies in moderate frequency (74/236) (Fig. 5a and Supplementary Tables 13, 14). While in japonica, the majority varieties (132/148) have both ORF3 and ORF4 deleted (Fig. 5a and Supplementary Table 12). Therefore, the Se locus is an important isolation barrier between indica and japonica, as it is highly likely to cause hybrid incompatibility when individuals from the two subspecies interbreed.
PAVs represent a class of SVs that has been suggested to contribute to individual adaptation49, 50. Although PAVs have been mostly linked to roles in ecological and agronomic adaptation in rice4,7, little is known about possible involvement in speciation. In this study, we identified a PAV contributing to indica-japonica subspecies reproductive isolation through HS.
HS maintains the integrity of a species over time, reducing or directly restricting gene flow to or from other species23. Consistent with the complex nature of one-locus HS loci identified in previous studies25,26,27,28,29,30,31,32,33, the Se locus in the PAV region studied here comprises two adjacent genes, ORF3 and ORF4. In hybrids with functional ORF3 and ORF4, pollen lacking the PAV is aborted. Such pollen lacks protective ORF4 to counteract the pollen-killing effect of ORF3 (Fig. 5b). Interestingly, ORF3 expression in transgene-positive lines from two independent ORF3 complementation and one overexpression experiments was less than 1/6 of that in T65/E9 F1 hybrids (Supplementary Fig. 12), and no orf4 homozygous CRISPR/Cas9 mutants were obtained in T0 or T1 generation (Supplementary Tables 10, 11). These indicate that ORF3 is toxic to living cells without ORF4, and its expression should be strictly regulated. The toxicity of ORF3 is also supported by the growth inhibition in T65 callus transformed with ORF3 (Supplementary Fig. 13) and in yeast cells after ORF3 induction (Supplementary Fig. 14). However, one homozygous mutant containing a deletion in pf12B (allelic to ORF4) was identified under BL (pf12-NJ/NJ) genetic background34. One possibility is that this mutant pf12B allele might still preserve some weak function. Genetic verification of this mutant pf12B allele would clarify its functionality. Another possibility is that different copies of ORF4 homologs might exist in BL genetic background, and some ORF4 homologs might be functional, which also needs to be further studied.
Consistent with the predicted function of ORF3 as a superfamily II DNA and RNA helicase, ORF3 and ORF4 co-localized in cytoplasmic foci known as Processing bodies (P-bodies) and Stress Granules (SGs) (Supplementary Fig. 18b,e), which play critical role in mRNA metabolism, including processes such as RNA decay, storage and translation repression51,52. Plants deficient in P-body components display severe developmental perturbations such as postembryonic lethality and seedling lethality53. Correspondingly, transcriptomic analyses revealed that pf12A induced a drastically increased number of differentially expressed genes in the anthers at the mature pollen stage possibly due to aberrant base excision repair and RNA degradation in the gametes34. Although the exact role of ORF3 and ORF4 in cytoplasmic foci remained to be further studied, our results showed the ORF3 didn’t affect the formation of P-bodies and SGs (Supplementary Fig. 25). In addition to localization in cytoplasmic foci, ORF3 also showed localization in mitochondria (Supplementary Fig. 18g), suggesting the abortion of pollen might be partially due to impaired mitochondria function and possibly poor energy supply, as shown in male sterilities caused by mitochondria-expressed genes, such as orfH79, WA352, S27/S28 and qHMS7-ORF227,36,54,55. The induction of functional ORF3 in INVSc1 yeast cells significantly inhibited the growth of the transformants, while this growth inhibition was attenuated when ORF3Δ100, which encodes ORF3 without mitochondrion localization signal showing disrupted mitochondria localization (Supplementary Fig. 18h), was transformed (Supplementary Fig. 14). In addition, Hyg-resistant secondary calli were apparently propagated from Agrobacterium co-cultivated primary calli transformed with ORF3Δ100 (Supplementary Fig. 13a, b). In contrast to low expression of the transgene in ORF3 transgene-positive calli, high expression level of transgenes, although not as high as negative controls, could be obtained in ORF3Δ100 transgene-positive calli (Supplementary Fig. 13c), suggesting that the mitochondria and cytoplasmic foci localization of ORF3 are both essential for its function.
The PAV is identified between different subspecies of O. sativa which is AA-genome. The outgroups BB-genome O. punctata, CC-genome O. eichingeri, CC-genome O. officinalis, EE-gnome O. australiensis, FF-genome O. brachyantha and GG-genome O. meyeriana lack the PAV region, indicating the absence of PAV is the ancestral condition of the Se locus in Oryza species (Fig. 5a). The PAV is only found in AA genome of Oryza species, and remains polymorphic across different species in Asian and African cultivated rice species complexes, surprisingly, it is also found in basal species of AA genome O. meridionalis but not in other AA species including O. glumaepatula and O. longistaminata. This pattern could have been caused by recent gene introgression at the Se locus across different species or independent loss of the Se locus in O. glumaepatula and O. longistaminata. These results clearly show that the locus was formed within the AA genome species of Oryza genus, and particularly, ORF3 is only found within Asian cultivated rice species complex.
Monophyletic origin56, polyphyletic origin7,57, and japonica-to-indica introgression theories58 have been advanced to explain the origin and domestication of Asian cultivated rice. We found 89.19% japonica (japonica (I)) lack the PAV and 90.68% indica (indica (I), indica (II) and indica (IV)) contain the PAV (Fig. 5a and Supplementary Table 12). Moreover, the PAV is also segregating in the population of O. rufipogon, the presumed wild progenitor of domesticated rice. These patterns are consistent with the scenario that japonica and indica originated from different O. rufipogon populations respectively and showed different haplotype distributions at the locus as a result. Therefore, the haplotype diversity pattern at the Se locus is consistent with polyphyletic origin and independent domestication of indica and janopica from different ancestral O. rufipogon populations.
In indica (II)/japonica (I) F1 hybrids, no gametes were selectively aborted through Se function due to the lack of pollen killer ORF3, generating minor types such as indica (III) (9.32% indica) and japonica (II) (6.76% japonica) (Fig. 5). Through repetitive backcrosses of indica (I) with japonica (I) during artificial selection, the rare japonica (III) (4.05% japonica) could also be created (Fig. 5a). For example, pedigree analyses indicated that japonica (III) varieties WYG7 and LG31 were bred through repeatedly backcrossing indica varieties derived from IR54 or TDK with different japonica varieties (Supplementary Fig. 26). Our results demonstrate that the differentiation of two subspecies of O. sativa L. can be partially reversed through artificial selection, which is consistent with the genetic mixture of indica and japonica shown in the evolutionary study of S559.
Despite clear examples of PAV controlling agronomic traits in different species, PAV events have been proposed to be largely restricted to genes with signatures of reduced essentiality, so called dispensable genes, and are recent additions to plant genomes60. Our study presents a PAV contributing to reproductive isolation of the two subspecies of Asian cultivated rice. Although the origin and maintenance mechanisms of the PAV remain unclear for the moment, the near fixation of the PAV in indica suggests it might be favorable in environment adaptation and/or human selection during the indica domestication process. Future study of ORF3 and ORF4 in response to different environmental cues may help to elucidate its beneficial effects in adaptation.
Reproductive isolation hinders the utilization of strong hybrid vigor in indica-japonica crosses in rice. Besides major effect on HMS, Se was also found to affect female gamete development at a lower degree (Supplementary Table 2 and Fig. 1e), which is consistent with result of pf12A34. Therefore, Se is one of the most important target to overcome indica-japonica HS. In our study, Type-3 Se allele with non-functional ORF3 and full-length ORF4 was identified as a neutral haplotype in O. rufipogon (IV), O. nivara (IV) and indica (IV) (Fig. 5a and Supplementary Data 3), and could be crossed into current varieties to build natural hybrid-compatible lines. Alternatively, a neutral Se allele could be artificially generated by disrupting ORF3 using CRISPR/Cas9 knockout or base editing strategies. Through genetic modification of Se and other indica-japonica hybrid sterility loci such as S5, Sa and Sc, widely-compatible lines could be created to break down reproductive barriers in indica-japonica hybrid rice breeding, thus having important implications in agriculture.
Plant materials and growth conditions
O. sativa ssp. japonica T65, NIL E9 and T65/E9 F1 hybrid were used in gene cloning and functional analysis. The wild and cultivated rice species used for haplotype investigation and evolutionary analysis came from our laboratory or were provided by the International Rice Research Institute (Los Baños, Philippines), Guangxi Academy of Agricultural Sciences (Guangxi, China) and Yunnan Academy of Agricultural Sciences (Yunnan, China). All plant materials were grown at the experimental farm of South China Agricultural University, Guangzhou, China.
To analyze pollen fertility, panicles at the flowering stage were fixed in FAA fixative [89% (v/v) ethanol, 6% (v/v) formaldehyde and 5% (v/v) acetic acid], and more than 200 pollen grains stained with 1% potassium iodide (I2-KI) solution were observed with an optical microscope (BK6000, OPTEC, China) to obtain pollen phenotype. To assess pollen vitality, pollen grains at the anthesis stage were collected and stained with 0.1% 2,3,5-triphenyl-2h-tetrazolium chloride (TTC) solution for 30 min before microscopic examination. A stereo microscope (SZ810, OPTEC, China) was used to observe the spikelets and anthers. Spikelet fertility was measured as the seed-setting rate of 20 individual main panicles.
For scanning electron microscopy (SEM) and transmission electron microscopy (TEM) observations, mature anthers of T65, E9 and T65/E9 F1 hybrid were pre-fixed in 0.1 M sodium phosphate buffer containing 4% glutaraldehyde (pH7.0) at 4 °C overnight. For SEM, pre-fixed samples were rinsed 3 times with 0.1 M phosphate buffer (pH7.0), and then were post-fixed for 2 h with 1.5% osmium tetroxide (OsO4) in PBS (pH7.0). After rinsing with PBS (pH7.0) and ethanol dehydration (30% to 100%), samples were exchanged 3 times with isoamyl acetate. Then, the fixed samples were dried with a critical point drier and coated with gold. The anther and pollen grains were observed using a scanning electron microscope (EVO MA 15, ZEISS, Germany) with an acceleration voltage of 10 kV. For TEM, pre-fixed samples were rinsed 3 times with PBS (pH7.0), post-fixed with 1.5% OsO4 in PBS (pH7.0) for 2 hours, and rinsed again 3 times with PBS (pH7.0). The post-fixed anthers were subjected to ethanol dehydration (30% to 100%) and transferred gradually to 100% resin. Super-thin sections (60 nm) were collected on uncoated nickel grids, and double stained with uranyl acetate and citrate aqueous solution as described previously61. Images were obtained with a transmission electron microscope (Talos L120C, FEI, America) operated at 80 kV.
To observe the development of anther and pollen, anthers at different stages were pre-fixed in FAA fixative, and were then embedded using Technovit Embedding Kits (Technovit 7100, Germany). Semi-thin sections (2 μm) were prepared using a microtome (RM2235, Leica, Germany), followed by staining with 0.5% toluidine blue. The stained sections were imaged with an optical microscope (BK6000, OPTEC, China).
Fine mapping of Se locus and sequence analysis
A mapping population was generated from a cross between T65 and E9. Two hundred and seventy-nine plants from the F3 population were used to analyze the linkage between pollen fertility and SSR markers for primary gene mapping. Additional SSR and InDel markers were designed in the IND9-PSM448 interval based on the polymorphisms between japonica Nipponbare and indica 9311. Fine mapping was performed using 7274 plants from the F4 population, and the final region was further confirmed through genotypic segregation and phenotypic analysis of the progeny of key recombinants. The genomic DNA sequence of Se region was amplified from T65 and E9 using KOD-Plus-Neo DNA polymerase (TOYOBO, Japan) and assembled using SnapGene. All primers are listed in Supplementary Data 4.
Full-length cDNA cloning and qRT-PCR analysis
Total RNA from young panicles of T65 and E9 were extracted using TRIZOL reagent (Invitrogen, America) according to the manufacturer’s instructions. First-strand cDNA was reverse transcribed using a Hifair® III 1st Strand cDNA Synthesis Kit (YEASEN, China). The predicted coding regions of candidate genes were amplified from cDNA and sequenced. RACE assay was performed by nested PCR with the SMARTer RACE cDNA Amplification kit (Clontech, America) following the manufacturer’s instructions. The anthers in different stages (meiocyte stage, microspore stage, early-bicellular stage, late-bicellular stage, and tricellular stage) from T65, E9 and T65/E9 F1 hybrids, along with different tissues of E9 were selected for qRT-PCR assays according to a previous study62, and the expression values were normalized to UBQ5 transcript levels.
Constructs for rice transformation
For complementation analysis, four genomic fragments (7,641-bp, 8,082-bp, 8,463-bp and 10,755-bp) were amplified from E9 and cloned into the binary vector pCAMBIA1300 to obtain the ORF1E9CTL, ORF2E9CTL, ORF3CTL, and ORF4CTL constructs. All constructs were introduced into T65 using Agrobacterium-mediated transformation. Transformation of ORF3CTL was conducted twice in independent experiments, and 34 transgene-positive lines were obtained in total.
Mutations of Se were constructed using the CRISPR/Cas9 vector system as previously reported63, and target sites were designed by the web-based tool CRISPR-GE/targetDesign (http://skl.scau.edu.cn/)64. Constructs of orf1T65, orf2T65 and orf1T65orf2T65 were introduced into T65; orf1E9 and orf3 constructs were transformed into E9; and orf3orf4 was introduced into T65/E9 F1 hybrids. For generation of orf4 mutants, constructs of three orf4 with different target sites in the coding region of ORF4 (target-1: CCGAAGATTACGAATCTTGGACA; target-2: CATATCTCTCACTTCGCGTATGG; target-3: TGTTCGCCGATGCGCCGTAACGG) were transformed into E9 or T65/E9 F1 hybrids in four independent experiments (Supplementary Table 10), and two types of frameshift mutations in three independent hemizygous orf4 mutants were obtained.
Because we didn’t detect ORF2E9 expression in the tested materials (Supplementary Fig. 8), we didn’t construct CRISPR/Cas9 vector targeting ORF2E9 to generate orf2E9.
Overexpression constructs (ORF3OE and ORF4OE), in which coding sequence of E9 was driven by CaMV 35 S promoter, were transformed into T65 and T65/E9 F1 hybrids, respectively, to verify the killer function of ORF3 and the protector function of ORF4, respectively.
Rice transformations were performed by Edgene Biotechnology Co., Ltd. Wuhan, China.
For pollen sterility analysis and qRT-PCR analysis of single given T0 transgenic plant, standard deviation was calculated from the 3 plants generated from a single positive callus. Three or more independent T0 transgenic plants were used as biological replicates.
Agrobacterium-mediated callus transformation
Overexpression constructs of eGFP (empty vector), ORF3-eGFP, ORF3Δ100-eGFP (with 5’ sequence encoding N-terminal 100-aa of ORF3 deleted), ORF3HJX74-eGFP (ORF3Hap-2-eGFP) were transferred into Agrobacterium tumefaciens strain EHA105. Sterilized mature seeds were inoculated on UINB medium and cultured at 28 °C in the dark. After 7 days of incubating, the calli were collected and used for transformation. After 3 days of cocultivation with Agrobacterium at 25 °C in constant darkness on N6D-Acetosyringone (AS) medium (30 g/L sucrose, 10 g/L glucose, 0.3 g/L casamino acids, 4.39 g/L N6D medium salt mixture, 2.878 g/L L-proline, and 2.8 mg/L 2,4-D, 20 mg/L AS, pH 5.4), the calli were washed two or three times with sterile water and rinsed once with sterile water containing 500 mg/L carbenicillin. Then, the calli were cultured on UINB medium containing 50 mg/L hygromycin B (Hyg, Merck), 400 mg/L carbenicillin for 30 days at 32 °C in constant light. Selection frequency of Hyg-resistant callus represents data from 3 groups of independent experiments, with ~50 transformed calli per group. Six (for ORF3-eGFP) and more than ten (for others) Hyg-resistant calli were used to test the expression of transgenes.
Susceptibility assay with inducible ORF3 expression
S. cerevisiae INVSc1 transformed with a vector harboring coding sequence of ORF3 gene from E9 (pYES2-ORF3 and pYES2-ORF3Δ100), HJX74 (pYES2-ORF3HJX74) or empty vector (pYES2-CT) was pre-cultured overnight at 30 °C in SC-Ura liquid medium (Coolaber, China) containing 2% glucose (SCGlu-Ura medium). To induce the ORF3 gene driven by GAL1 promoter, the transformant was inoculated at an initial concentration equivalent to OD600 = 0.1 into 50 mL of SC-Ura containing 3% galactose and 2% raffinose (SCGal-Ura medium), and cultured for 92 h at 30 °C. The growth of yeast cells was monitored by OD600. In the case of the reference experiment without induction of the ORF3 gene expression, SCGal-Ura medium was replaced with SCGlu-Ura medium, and the culture time is 40 h.
Subcellular localization of ORF3 and ORF4
For subcellular localization, the ORF3 and ORF4 coding region (minus the stop codon) were cloned into the pRTVcGFP vector, and plasmids were prepared using the FinePure EndoFree Plasmid Mini Kit (GENFINE, China). Constructs of eGFP (empty vector), ORF3-eGFP, ORF3Δ100-eGFP, ORF3HJX74-eGFP, ORF4-eGFP (ORF4Hap-1-eGFP), ORF4Hap-2-eGFP, ORF4Hap-3-eGFP and ORF4Hap-12-eGFP were transformed or co-transformed with plasmid encoding the cytoplasmic foci marker AtTZF1-mCherry65 into rice protoplasts using the polyethylene glycol method. Following incubation for 14 to 20 h at 28 °C, fluorescence was observed using confocal microscopy (TCS SP8, Leica, Germany). Transformed protoplasts were stained with mitochondrion-specific dye MitoTracker Deep Red (Invitrogen, America) to view mitochondira signal.
Phylogenetic and evolutionary analysis of Se locus
To identify putative homologs of ORF3 and ORF4, full-length protein sequences were used to query the NCBI database (https://www.ncbi.nlm.nih.gov/). Sequence alignment and neighbor-joining phylogenetic tree construction were performed using MEGA7 software (bootstrap replication = 1000).
To trace the evolutionary history of the Se locus, genomic DNA was extracted from 787 rice samples, including 4 O. punctata, 2 O. eichingeri, 3 O. officinalis, 3 O. australiensis, 2 O. brachyantha, 3 O. meyeriana, 28 O. longistaminata, 16 O. glumaepatula, 5 O. meridionalis, 4 O. barthii, 342 O. rufipogon, 44 O. nivara, 6 O. glaberrima, and 325 landraces (consisting of 217 indica and 108 japonica) from Rice Diversity Panel 266. For all materials that contained the PAV region, ORF3 and ORF4 were sequenced and assembled using SnapGene. All of the sequences obtained and an additional 60 publicly available high-quality genome sequences4,6,48 were used for haplotype investigation and evolutionary analysis. All rice materials are listed in Supplementary Data 5.
Further information on research design is available in the Nature Portfolio Reporting Summary linked to this article.
The CDS and amino acid sequences of ORF1T65 [https://www.ncbi.nlm.nih.gov/nuccore/OR253652], ORF1E9 [https://www.ncbi.nlm.nih.gov/nuccore/OR253653], ORF2T65 [https://www.ncbi.nlm.nih.gov/nuccore/OR253654], ORF3 [https://www.ncbi.nlm.nih.gov/nuccore/OR253655] and ORF4 [https://www.ncbi.nlm.nih.gov/nuccore/OR253656] are available at GenBank under accession numbers OR253652 to OR253656, as well as in Figshare [https://doi.org/10.6084/m9.figshare.23690625]. Since expression of ORF2E9 could not be detected (Supplementary Fig. 8), we consider it presence as putative and thus we didn’t deposit its sequence to GenBank. ORF2E9 sequence used for alignment in Supplementary Fig. 7 was deduced from R498 reference genome annotation. Genealogical trees were deduced from the public available data on China Rice Data center (WYG7: [https://www.ricedata.cn/variety/varis/601512.htm]; LG31: [https://www.ricedata.cn/variety/varis/609156.htm]). To access the parental information of the variety, you can click the symbol “+” in the upper left corner of the description pages of WYG7 and LG31. The breeding history of the variety was inferred by continuously clicking the symbol “+” appearing next to the parental line. Source data are provided with this paper.
Hollox, E. J., Zuccherato, L. W. & Tucci, S. Genome structural variation in human evolution. Trends Genet 38, 45–58 (2022).
Yuan, Y., Bayer, P. E., Batley, J. & Edwards, D. Current status of structural variation studies in plants. Plant Biotechnol. J. 19, 2153–2163 (2021).
Saxena, R. K., Edwards, D. & Varshney, R. K. Structural variations in plant genomes. Brief. Funct. Genomics 13, 296–307 (2014).
Zhao, Q. et al. Pan-genome analysis highlights the extent of genomic variation in cultivated and wild rice. Nat. Genet. 50, 278–284 (2018).
Liu, Y. et al. Pan-genome of wild and cultivated soybeans. Cell 182, 162–176.e13 (2020).
Qin, P. et al. Pan-genome analysis of 33 genetically diverse rice accessions reveals hidden genomic variations. Cell 184, 3542–3558.e16 (2021).
Wang, W. et al. Genomic variation in 3,010 diverse accessions of Asian cultivated rice. Nature 557, 43–49 (2018).
Li, H. et al. Graph-based pan-genome reveals structural and sequence variations related to agronomic traits and domestication in cucumber. Nat. Commun. 13, 682 (2022).
Golicz, A. A. et al. The pangenome of an agronomically important crop plant Brassica oleracea. Nat. Commun. 7, 13390 (2016).
Montenegro, J. D. et al. The pangenome of hexaploid bread wheat. Plant J. 90, 1007–1013 (2017).
Schatz, M. C. et al. Whole genome de novo assemblies of three divergent strains of rice, Oryza sativa, document novel gene space of aus and indica. Genome Biol. 15, 506 (2014).
Lei, L. et al. Plant pan-genomics comes of age. Annu. Rev. Plant Biol. 72, 411–435 (2021).
Tao, Y., Zhao, X., Mace, E., Henry, R. & Jordan, D. Exploring and exploiting pan-genomics for crop improvement. Mol. Plant 12, 156–169 (2019).
Xu, K. et al. Sub1A is an ethylene-response-factor-like gene that confers submergence tolerance to rice. Nature 442, 705–708 (2006).
Hattori, Y. et al. The ethylene response factors SNORKEL1 and SNORKEL2 allow rice to adapt to deep water. Nature 460, 1026–1030 (2009).
Gamuyao, R. et al. The protein kinase Pstol1 from traditional rice confers tolerance of phosphorus deficiency. Nature 488, 535–539 (2012).
Kretzschmar, T. et al. A trehalose-6-phosphate phosphatase enhances anaerobic germination tolerance in rice. Nat. Plants 1, 15124 (2015).
He, G. et al. Haplotype variation in structure and expression of a gene cluster associated with a quantitative trait locus for improved yield in rice. Genome Res 16, 618–626 (2006).
Cardoso, C. et al. Natural variation of rice strigolactone biosynthesis is associated with the deletion of two MAX1 orthologs. Proc. Natl. Acad. Sci. 111, 2379–2384 (2014).
Gabur, I., Chawla, H. S., Snowdon, R. J. & Parkin, I. A. P. Connecting genome structural variation with complex traits in crop plants. Theor. Appl. Genet. 132, 733–750 (2019).
Londo, J. P., Chiang, Y.-C., Hung, K.-H., Chiang, T.-Y. & Schaal, B. A. Phylogeography of Asian wild rice, Oryza rufipogon, reveals multiple independent domestications of cultivated rice, Oryza sativa. Proc. Natl. Acad. Sci. 103, 9578–9583 (2006).
Ouyang, Y. & Zhang, Q. The molecular and evolutionary basis of reproductive isolation in plants. J. Genet. Genomics 45, 613–620 (2018).
Zhang, G. Prospects of utilization of inter-subspecific heterosis between indica and japonica rice. J. Integr. Agric. 19, 1–10 (2020).
Chen, R. et al. Rice functional genomics: decades’ efforts and roads ahead. Sci. China Life Sci. 65, 33–92 (2022).
Long, Y. et al. Hybrid male sterility in rice controlled by interaction between divergent alleles of two adjacent genes. Proc. Natl. Acad. Sci. 105, 18871–18876 (2008).
Shen, R. et al. Genomic structural variation-mediated allelic suppression causes hybrid male sterility in rice. Nat. Commun. 8, 1310 (2017).
Yu, X. et al. A selfish genetic element confers non-Mendelian inheritance in rice. Science 360, 1130–1132 (2018).
Yang, J. et al. A killer-protector system regulates both hybrid sterility and segregation distortion in rice. Science 337, 1336–1340 (2012).
Yu, Y. et al. Hybrid sterility in rice (Oryza sativa L.) involves the tetratricopeptide repeat domain containing protein. Genetics 203, 1439–1451 (2016).
Kubo, T., Takashi, T., Ashikari, M., Yoshimura, A. & Kurata, N. Two tightly linked genes at the hsa1 locus cause both F1 and F2 hybrid sterility in rice. Mol. Plant 9, 221–232 (2016).
Hou, J. et al. ESA1 is involved in embryo sac abortion in interspecific hybrid progeny of rice. Plant Physiol. 180, 356–366 (2019).
Koide, Y. et al. Lineage-specific gene acquisition or loss is involved in interspecific hybrid sterility in rice. Proc. Natl. Acad. Sci. 115, E1955–E1962 (2018).
Xie, Y. et al. An asymmetric allelic interaction drives allele transmission bias in interspecific rice hybrids. Nat. Commun. 10, 2501 (2019).
Zhou, P. et al. A minimal genome design to maximally guarantee fertile inter-subspecific hybrid rice. Mol. Plant 16, 726–738 (2023).
Mizuta, Y., Harushima, Y. & Kurata, N. Rice pollen hybrid incompatibility caused by reciprocal gene loss of duplicated genes. Proc. Natl. Acad. Sci. 107, 20417–20422 (2010).
Yamagata, Y. et al. Mitochondrial gene in the nuclear genome induces reproductive barrier in rice. Proc. Natl. Acad. Sci. 107, 1494–1499 (2010).
Nguyen, G. N. et al. Duplication and loss of function of genes encoding RNA polymerase III subunit C4 causes hybrid incompatibility in rice. G3 7, 2565–2575 (2017).
Guo, J. et al. Overcoming inter-subspecific hybrid sterility in rice by developing indica-compatible japonica lines. Sci. Rep. 6, 26878 (2016).
Guo, J. et al. Development of wide-compatible indica lines by pyramiding multiple neutral alleles of indica–japonica hybrid sterility loci. Front. Plant Sci. 13, 1–10 (2022).
Zhang, G. The next generation of rice: inter-subspecific indica-japonica hybrid rice. Front. Plant Sci. 13, 1–7 (2022).
Zhang, G., Lu, Y.-G., Zhang, H., Yang, J. & Liu, G. Genetic studies on the hybrid sterility in cultivated rice (Oryza sativa). IV. Genotypes for F1 pollen sterility. Acta Genet. Sin. 21, 34–41 (1994).
Zhu, W. et al. Preliminary identification of F1 pollen sterility gene S-e in Oryza sativa. J. South China Agric. Univ. 29, 1–5 (2008). (In Chinese).
Song, X., Qiu, S. Q., Xu, C. G., Li, X. H. & Zhang, Q. Genetic dissection of embryo sac fertility, pollen fertility, and their contributions to spikelet fertility of intersubspecific hybrids in rice. Theor. Appl. Genet. 110, 205–211 (2005).
Win, K. T. et al. Identification of two loci causing F1 pollen sterility in inter- and intraspecific crosses of rice. Breed. Sci. 59, 411–418 (2009).
Zhang, H. et al. A major locus qS12, located in a duplicated segment of chromosome 12, causes spikelet sterility in an indica-japonica rice hybrid. Theor. Appl. Genet. 123, 1247–1256 (2011).
Xu, P. et al. Mapping three new interspecific hybrid sterile loci between Oryza sativa and O. glaberrima. Breed. Sci. 63, 476–482 (2014).
Kubo, T., Yoshimura, A. & Kurata, N. Genetic characterization and fine mapping of S25, a hybrid male sterility gene, on rice chromosome 12. Genes Genet. Syst. 92, 205–212 (2017).
Li, F. et al. High-quality de novo genome assembly of Huajingxian 74, a receptor parent of single segment substitution lines. Rice Sci. 28, 109–113 (2021).
Tan, S., Zhong, Y., Hou, H., Yang, S. & Tian, D. Variation of presence/absence genes among Arabidopsis populations. BMC Evol. Biol. 12, 86 (2012).
Lye, Z. N. & Purugganan, M. D. Copy number variation in domestication. Trends Plant Sci. 24, 352–365 (2019).
Chantarachot, T. & Bailey-Serres, J. Polysomes, stress granules, and processing bodies: a dynamic triumvirate controlling cytoplasmic mRNA fate and function. Plant Physiol. 176, 254–269 (2018).
Ivanov, P., Kedersha, N. & Anderson, P. Stress granules and processing bodies in translational control. Cold Spring Harb. Perspect. Biol. 11, a032813 (2019).
Xu, J. & Chua, N.-H. Processing bodies and plant development. Curr. Opin. Plant Biol. 14, 88–93 (2011).
Luo, D. et al. A detrimental mitochondrial-nuclear interaction causes cytoplasmic male sterility in rice. Nat. Genet. 45, 573–577 (2013).
Peng, X. et al. The mitochondrial gene orfH79 plays a critical role in impairing both male gametophyte development and root growth in CMS-Honglian rice. BMC Plant Biol. 10, 125 (2010).
Molina, J. et al. Molecular evidence for a single evolutionary origin of domesticated rice. Proc. Natl. Acad. Sci. 108, 8351–8356 (2011).
Stein, J. C. et al. Genomes of 13 domesticated and wild rice relatives highlight genetic conservation, turnover and innovation across the genus Oryza. Nat. Genet 50, 285–296 (2018).
Huang, X. et al. A map of rice genome variation reveals the origin of cultivated rice. Nature 490, 497–501 (2012).
Mi, J. et al. Artificial selection in domestication and breeding prevents speciation in rice. Mol. Plant 13, 650–657 (2020).
Bush, S. J. et al. Presence–absence variation in A. thaliana is primarily associated with genomic signatures consistent with relaxed selective constraints. Mol. Biol. Evol. 31, 59–69 (2014).
Li, N. et al. The rice tapetum degeneration retardation gene is required for tapetum degradation and anther development. Plant Cell 18, 2999–3014 (2006).
Fang, C. et al. Identification of S23 causing both interspecific hybrid male sterility and environment-conditioned male sterility in rice. Rice 12, 10 (2019).
Ma, X. et al. A robust CRISPR/Cas9 system for convenient, high-efficiency multiplex genome editing in monocot and dicot plants. Mol. Plant 8, 1274–1284 (2015).
Xie, X. et al. CRISPR-GE: a convenient software toolkit for CRISPR-based genome editing. Mol. Plant 10, 1246–1249 (2017).
Han, J. et al. All-in-one: a robust fluorescent fusion protein vector toolbox for protein localization and BiFC analyses in plants. Plant Biotechnol. J. 20, 1098–1109 (2022).
McCouch, S. R. et al. Open access resources for genome-wide association mapping in rice. Nat. Commun. 7, 10532 (2016).
We thank Xiangdong Liu (South China Agricultural University, China), Yuntao Liang (Guangxi Academy of Agricultural Science, China) and Cuifeng Tang (Yunnan Academy of Agricultural Science, China) for providing wild rice materials, thank Bihai Shi and Bin Hu (South China Agricultural University, China) for the assistance with confocal microscopic observation, thank Donna E. Fernandez (University of Wisconsin-Madison, USA) for reading and editing the manuscript, and thank Bo Wang and Wenjuan Hu (South China Agricultural University, China) for the help in the work related to Agrobacterium-mediated callus transformation. This work was supported by National Natural Science Foundation of China (Grant No. 31571483 to Z.L.), Guangdong Provincial Natural Science Foundation (Grants No. 2019A1515012077, 2021A1515010446, and 2015A030313415 to Z.L.), Guangdong Provincial Key Laboratory of New Technology in Rice Breeding (Grant No. 2020B1212060047 to Z.L.), and start-up research grants from Science, Technology and Innovation Commission of Shenzhen Municipality to H.W.
Z.L., D.W., J.F., M.Z. and Z.M. are inventors on patent application (No. 202210609889.8) filed to the State Intellectual Property Office of China by South China Agricultural University that covers ORF3 and ORF4 sequences of the Se locus and their applications.
Peer review information
Nature Communications thanks the anonymous reviewers for their contribution to the peer review of this work. A peer review file is available.
Publisher’s note Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.
About this article
Cite this article
Wang, D., Wang, H., Xu, X. et al. Two complementary genes in a presence-absence variation contribute to indica-japonica reproductive isolation in rice. Nat Commun 14, 4531 (2023). https://doi.org/10.1038/s41467-023-40189-x