Seed abortion caused by the combination of two duplicate genes in the progeny from the cross between Oryza sativa and Oryza meridionalis

Abstract

Seed development is an essential phenomenon for all sexual propagative plant species. The functional allele at SEED DEVELOPMENT 1 (SDV1) or SEED DEVELOPMENT 2 (SDV2) loci is essential for seed development for Oryza sativa and Oryza meridionalis. In the present study, we performed fine mapping of SDV1, narrowing down the area of interest to 333kb on chromosome 6. Haplotype analysis around the SDV1 locus of O. meridionalis accessions indicated that they shared the DNA polymorphism, suggesting that they have a common abortive allele at the SDV1 locus. Linkage analysis of the candidate SDV2 gene showed that it was located on chromosome 4. The candidate SDV2 was confirmed using a population in which both the SDV1 and SDV2 genes were segregating. The chromosomal region covering the SDV1 gene was predicted to contain 30 protein-coding genes in O. sativa. Five of these genes have conserved DNA sequences in the chromosomal region of the SDV2 gene on chromosome 4, and not on chromosome 6, of O. meridionalis. These results suggest that these five genes could be candidates for SDV1, and that their orthologous genes located on chromosome 4 of O. meridionalis could be candidates for SDV2.

Introduction

Cultivated rice (Oryza sativa) is not only a staple food for many people, but is also an extremely important grain from a food security perspective. In the process of rice breeding, AA genome wild rice is considered an effective gene pool because it is closely related to cultivated rice and is highly diverse (Brozynska et al. 2016, Henry et al. 2010). However, reproductive barriers often occur in crosses between cultivated and wild rice. There have been numerous reports of reproductive barriers in crosses between O. sativa and its closely related wild species, and these reports provide very important information for rice breeding and for understanding the mechanisms of reproductive barriers. Oryza meridionalis is a wild rice with an AA genome native to Australia, offering a unique evolutionary path compared to other AA genome species (Brozynska et al. 2017, Cheng et al. 2002, Stein et al. 2018, Zhang et al. 2014). Reproductive barriers have been reported in a number of crosses between O. meridionalis and other AA genomic species (Banaticla-Hilario et al. 2013, Ichitani et al. 2022, Juliano et al. 2005, Li et al. 2018, Naredo et al. 1997, 1998). In our previous study, we observed that no homozygote of O. meridionalis alleles for the loci closely linked with HD1 on chromosome 6 appeared in progenies from crosses between O. meridionalis and O. sativa (Toyomoto et al. 2019). This was caused by abortion during seed development, indicating that a gene controlling seed development (SEED DEVELOPMENT 1, SDV1) was located on chromosome 6. O. sativa carries the dominant functional fertile allele Sdv1-s (s stands for sativa). O. meridionalis is homozygous for the recessive abortive allele sdv1-m (m stands for meridionalis) (Fig. 1). However, O. meridionalis is able to self-fertilize and produce seeds, suggesting that it carries a functional fertile gene at other loci probably deriving from gene duplication. A putative duplicative gene of SDV1, tentatively named SEED DEVELOPMENT 2 (SDV2), was proposed (Toyomoto et al. 2019): O. meridionalis and O. sativa should carry the functional allele and the nonfunctional allele at the SDV2 locus, respectively. We named these respective alleles Sdv2-m and sdv2-s (Fig. 1).

Fig. 1.

The genetic model of seed fertility controlled by SDV1 and SDV2, the development of experimental lines for fine-mapping SDV1, mapping of SDV2, verification of SDV2 and survey of gametic distortion. Long and short vertical bars respectively denote chromosomes 4 and 6. Blue and pink segments respectively depict chromosomal segments from O. sativa and O. meridionalis. The chomosomal locations of SDV1 and SDV2 are indicated by horizontal bars on the chromosomes. The number in parenthesis indicates the segregating genes: 1 and 2 denote SDV1 and SDV2, respectively. Dotted X mark indicates that the genotype is aborted during seed development.

The existence of the SDV2 gene has remained untested. Seed development is an essential phenomenon for all sexual propagative plant species: All Oryza species should carry at least one functional allele of the SDV1 or SDV2 locus. AA genome genus Oryza comprises seven species, O. sativa, O. glaberrima, O. barthii, O. rufipogon, O. glumaepatula, O. meridionalis, and O. longistaminata. Annual accessions of O. rufipogon are referred to as O. nivara by some researchers (Stein et al. 2018, Zhang et al. 2014). O. rufipogon is the progenitor of O. sativa, and O. barthii is the progenitor of African rice O. glaberrima. The history of duplication and loss of function of SDV1 and SDV2 will shed light on the controversial AA-genome Oryza speciation (Htet et al. 2022, Ohtsubo et al. 2004, Stein et al. 2018, Yin et al. 2015, Zhang et al. 2014). In this study, we first performed fine mapping and haplotype analysis of SDV1. Then we performed linkage analysis of SDV2. Finally, we tested if the two genes follow duplicate gene action, and searched for their candidate genes, utilizing their chromosomal locations to determine how these genes cause reproductive barriers and the genetic basis of these genes.

Materials and Methods

Plant material

In this study, we used an O. meridionalis accession W1297 and a rice cultivar ‘Taichung 65’ (T65) for linkage analysis of SDV1 and SDV2. W1297 was collected in Darwin, Northern Territory, Australia, and was provided by the National Institute of Genetics in Mishima, Japan. O. sativa comprises two subspecies, japonica and indica. T65 is a japonica cultivar used frequently in the study of rice genetics as a recurrent parent of chromosome segment substitution lines, near-isogenic lines, and the study of induced mutation (Ichitani et al. 2014, Sano 1990, Yamagata et al. 2019, Yoshimura et al. 2010). Using T65 genetic background with incorporated W1297 chromosomes, we developed several experimental lines for analyzing SDV1 and SDV2 genes. To avoid confusion, number(s) in parenthesis was added after the general generation name to refer to target gene(s), as shown in Fig. 1: 1 and 2 denote SDV1 and SDV2, respectively.

Cultivation was conducted according to the methods described by Ichitani et al. (2014). The germinated seeds were sown in seedling boxes in a greenhouse. Approximately two weeks after sowing, the seedlings were transferred out of the greenhouse. Approximately 30 days after sowing, the seedlings were transplanted to the paddy fields in Experimental Farm of the Faculty of Agriculture, Kagoshima University. The applied fertilizers were 4, 6, and 5 g/m², respectively, for N, P₂O₅, and K₂O. The plant spacing was 15 × 30 cm. Seeds were sown in May to June, and seedling transplanting to paddy fields were conducted in late June to July from 2017 to 2023.

For the haplotype analysis, we used nineteen accessions of O. meridionalis, four accessions of Oryza rufipogon from New Guinea, two accessions of O. rufipogon from Australia, fourteen accessions of O. rufipogon from Asia, and nine accessions of O. sativa (Table 1). All the wild rice accessions except Jpn1 and Jpn2 were distributed from the National Institute of Genetics supported by the National Bioresource Project, MEXT, Japan. Jpn1 and Jpn2 were collected in Australia with the permission from the Queensland government, under the EcoAccess program (Sotowa et al. 2013). World Rice Core collection (WRC) by Kojima et al. (2005) comprises 69 accessions based on a genome-wide RFLP polymorphism survey of 332 accessions of O. sativa. A total of eight WRC lines were selected so that they cover the six groups classified by 40 Indel markers (Ichitani et al. 2016).

Table 1.

Haplotypes of DNA markers around SDV1 gene

Accession name	Origin	Genotype of DNA markera
		KGC6_11.53	KGC6_11.67	KGC6_11.71	KGC6_11.732	KGC6_11.736	KGC6_11.737	KGC6_11.74	KGC6_11.84	KGC6_11.87
Oryza meridionalis
W1297	Australia	W	W	W	W	W	W	W	W	W
W1299	Australia	W	W	W	W	W	W	W	W	W
W1300	Australia	W	W	W	W	W	W	W	W	W
W1627	Australia	W	W	W	W	W	W	W	W	W
W1631	Australia	W	W	W	W	W	W	W	W	W
W1635	Australia	W	W	W	W	W	W	W	W	W
W1638	Australia	W	W	W	W	W	W	W	W	W
W2069	Australia	W	W	W	W	W	W	W	W	W
W2071	Australia	W	W	W	W	W	W	W	W	W
W2077	Australia	W	W	W	W	W	W	W	W	W
W2079	Australia	W	W	W	W	W	W	W	W	W
W2080	Australia	W	W	W	W	W	W	W	W	W
W2081	Australia	W	W	W	W	W	W	W	W	W
W2100	Australia	W	W	W	W	W	W	W	W	W
W2103	Australia	W	W	W	W	W	W	W	W	W
W2105	Australia	W	W	W	W	W	W	W	W	W
W2112	Australia	W	W	W	W	W	W	W	W	W
W2116	Australia	W	W	W	W	W	W	W	W	W
Jpn2	Australia	W	W	W	W	W	W	W	W	W
Oryza rufipogon in Australia
W2109	Australia	T	T	T	T	T	T	T	T	T
Jpn1	Australia	T	T	T	T	T	T	T	T	T
Oryza rufipogon in New Guinea
W1230	Indonesia	T	T	T	T	T	T	T	T	T
W1235	Indonesia	W	W	W	W	W	W	W	W	W
W1239	Indonesia	W	W	W	W	W	W	W	W	W
W1236	Papua New Guinea	T	T	T	T	T	T	T	T	T
Oryza rufipogon in Asian Countries
W0106	India	T	T	T	T	T	T	T	T	T
W0107	India	T	T	T	T	T	T	T	T	T
W0120	India	T	T	T	T	T	T	T	1	T
W0137	India	N	T	T	T	T	T	T	T	T
W0630	Burma	T	T	T	T	T	T	T	1	T
W1294	Philippines	N	T	T	T	T	T	T	T	T
W1551	Thailand	T	T	T	T	T	T	T	T	T
W1681	India	T	T	T	T	T	T	T	T	T
W1690	Thailand	T	T	T	T	T	T	T	T	T
W1807	Sri Lanka	1	T	T	T	T	T	T	T	T
W1865	Thailand	T	T	T	T	T	T	T	T	T
W1866	Thailand	T	T	T	T	T	T	T	T	T
W1945	China	T	T	T	T	T	T	T	T	T
W2014	India	N	T	T	T	T	T	T	T	T
Oryza sativa
WRC01	Japan	T	T	T	T	T	T	T	T	T
WRC02	India	T	T	T	T	T	T	T	T	T
WRC21	Myanmar	T	T	T	T	T	T	T	T	T
WRC37	India	T	T	T	T	T	T	T	T	T
WRC43	China	T	T	T	T	T	T	T	T	T
WRC50	America	T	T	T	T	T	T	T	T	T
WRC51	Japan	T	T	T	T	T	T	T	T	T
WRC98	China	2	T	T	T	T	T	T	T	T
Taichung 65	Taiwan	T	T	T	T	T	T	T	T	T

^a T and W represent the same banding pattern as T65 and W1297, respectively.

Numbers indicate a banding pattern other than that found in T65 and W1297.

N indicates that the band did not appear.

Trait evaluation

Seed fertility was evaluated by collecting 50 seeds from the upper side of each of three panicles, following Toyomoto et al. (2019). Seemingly sterile seeds were dehusked to see if sterility occurred before or after fertilization.

DNA analysis

DNA extraction from leaves was performed according to Toyomoto et al. (2019). PCR mixtures, cycles, electrophoresis, DNA staining, and gel image recordings were conducted according to Ichitani et al. (2014). Antmap (Iwata and Ninomiya 2006) was used to draw the linkage map of SDV2. The Kosambi function was used to calculate the map distance.

DNA markers

Most published PCR-based DNA markers for Oryza species are based on an O. sativa genome sequence such as ‘Nipponbare’ (Os-Nipponbare-Reference-IRGSP-1.0 (Kawahara et al. 2013)) and ‘93-11’ (Yu et al. 2002). However, many discrepancies in DNA sequence between the genome of Nipponbare (IRGSP 1.0) and that of O. meridionalis accession W2112 (GCA_000338895) (Stein et al. 2018) lead to failure in amplification from the O. meridionalis genome when using O. sativa genome-based DNA markers (Toyomoto et al. 2019). Our strategy for designing codominant DNA markers for mapping SDV1 and SDV2 genes was to select insertions/deletions (indels) ranging from 5 to 100 base pairs that were shared by the two O. meridionalis accessions, W2112 (Stein et al. 2018) and IRGC105298 (Zhang et al. 2014), and not shared by the following AA genome species sequences, O. sativa cv. Nipponbare (Kawahara et al. 2013), O. sativa cv. Taichung 65 (Wei et al. 2016), O. rufipogon accession W1943 (Stein et al. 2018), O. nivara accession IRGC100897 (Stein et al. 2018), two O. longistaminata accessions W1413 and W1508 (Ohyanagi et al. 2016), O. glumaepatula accession GEN1233 (Stein et al. 2018) and O. barthii accession IRGC105608 (Stein et al. 2018) with the aid of TASUKE+ (Kumagai et al. 2019). Based on sequence similarity surrounding the indels between the Nipponbare (Os-Nipponbare-Reference-IRGSP-1.0) and O. meridionalis W2112 (GCA_000338895.3) genome assemblies, the selected indels were further screened. Primer design followed Busungu et al. (2016). Mismatches between primer sequences and genomic sequences were checked using Oryzabase BLAST (https://shigen.nig.ac.jp/rice/oryzabase/locale/change?lang=en) (Yamazaki et al. 2010) and MBKBASE BLAST (Peng et al. 2020). A single base mismatch within the primer sequence was accepted.

Results

Fine mapping of SDV1

Our previous study showed that SDV1 is located between two indel markers, KGC6_10.09 and KGC6_22.19 (approximately 12 Mb) (Toyomoto et al. 2019). To narrow down the candidate region for SDV1, we first selected 115 recombinants between the two DNA markers KGC6_10.09 and KGC6_22.19 from 4625 individuals in a BC₄F₄(1) population. Four plants were recombinants between KGC6_10.09 and KGC6_13.00. Next, we designed 10 new DNA markers located between the two DNA markers (Supplemental Table 1), and applied them to the four plants (Table 2). Because homozygotes for W1297 at SDV1 locus are aborted during seed development (Toyomoto et al. 2019), the SDV1 locus should not be located near any DNA markers for which genotypes were homozygous for W1297 allele in the four plants (Table 2, Fig. 1). The experimental result indicates that the candidate region for the SDV1 locus was narrowed down to 333 kb encompassed by KGC6_11.53 and KGC6_11.87.

Table 2.

Genotypes of informative recombinants BC₄F₄(1) for DNA marker loci closely located with SDV1

BC4F4(1) Recombinant	Genotype at DNA Markera
	KGC6_10.09	KGC6_10.87	KGC6_11.53	KGC6_11.67	KGC6_11.71	KGC6_11.732	KGC6_11.736	KGC6_11.737	KGC6_11.74	KGC6_11.84	KGC6_11.87	KGC6_12.02	KGC6_13.00	KGC6_22.19
1	W	W	H	H	H	H	H	H	H	H	H	H	H	H
2	H	H	H	H	H	H	H	H	H	H	H	H	W	W
3	H	H	H	H	H	H	H	H	H	H	W	W	W	W
4	W	W	W	H	H	H	H	H	H	H	H	H	H	H

^a W and H respectively denote homozygotes for W1297 and heterozygotes.

Haplotype analysis around SDV1

The genotypes of 48 accessions for the nine DNA markers located around the SDV1 locus are presented in Table 1. As described in the Materials and Methods, the DNA markers in Supplemental Table 1 were designed to detect the DNA polymorphism between two O. meridionalis accessions and the other AA genome Oryza species. All the O. meridionalis accessions, W1235 and W1239 shared the same banding patterns of DNA markers located in the candidate SDV1 region. It has been proposed that the W1235 and W1239 accessions in New Guinea might have been misclassified as O. rufipogon, and that they probably are O. meridionalis (Htet et al. 2022, Yin et al. 2015) : these two accessions belonged to a same cluster based on nuclear SSR markers as O. meridionalis (Yin et al. 2015), and shared a large intron in PolA1 gene with O. meridionalis (Htet et al. 2022). These results suggest that O. meridionalis accessions in Australia and New Guinea share the sdv1-m allele in common, though O. meridionalis and O. rufipogon accessions in Australia share chloroplast DNA polymorphisms (Sotowa et al. 2013), and they share the same habitats in Australia (Henry 2019).

Mapping of the candidate of SDV2

Because segregation distortion and seed abortion caused by sdv1-m were observed in the T65 genetic background, those caused by sdv2-s should be observed when the SDV1 locus is fixed for sdv1-m. We therefore developed experimental lines for mapping SDV2 gene (Fig. 1). Among thirty nine BC₁F₁ plants, ten latest heading plants were selected: they were expected to be heterozygote for SDV1 locus because SDV1 is closely linked with HD1 locus, which mainly controlled heading date in the progeny from the cross between T65 and W1297, and W1297 carries late heading allele at HD1 (Toyomoto et al. 2019). Approximately 80 BC₁F₂(2) plants for each BC₁F₁ plant were genotyped for the DNA markers linked with SDV1. The eight BC₁F₂(2) lines showed segregation distortion of the DNA markers explainable by SDV1 gene segregation: no homozygotes of W1297 allele at KGC6_12.02 appeared. This suggests that they became homozygous for sdv2-s by one cycle of backcrossing. One BC₁F₁ plant proved to be recombinant between HD1 and SDV1: it was fixed for T65 allele of the DNA markers linked with SDV1 gene. The BC₁F₂(2) from one BC₁F₁ plant showed the segregation of 28 T65 homozygotes: 40 heterozygotes: 9 W1297 homozygotes at KGC6_12.02 locus, skewed from the expected Mendelian ratio 1:2:1 (χ² = 9.493, P = 0.009). However, the existence of W1297 homozygotes suggests that SDV2 gene could be also segregating in this population. Among 9 BC₁F₂(2) plants homozygous of W1297 allele at KGC6_12.02 locus, four semi-sterile BC₁F₂(2) plants were selected by visual observation of seed fertility (Fig. 1). This selection was based on the idea that highly fertile plants were expected to be fixed for Sdv2-m allele for SDV2 locus, and highly sterile plants were expected to be segregating for other loci affecting seed fertility. BC₁F₃(2) in Fig. 1 were subjected to further genetic analysis. The bulked DNA from ten plants from each BC₁F₃(2) lines were genotyped for 80 DNA markers covering all the 12 rice chromosomes: no homozygotes of T65 alleles were detected for some DNA markers on chromosomes 4, 5, and 8. BC₁F₄(2) lines derived from BC₁F₃(2) plants heterozygous for these markers were subjected to a preliminary linkage analysis with a small number of plants, showing that SDV2 could be located on chromosome 4. Then we repeatedly selected possible heterozygotes of the SDV2 from BC₁F₄(2) to BC₁F₆(2) generation with the aid of DNA markers and seed fertility data. The repeated selfing process could make the other loci homozygous, which was expected to reduce the variance of seed fertility caused by genes other than SDV2.

The BC₁F₇(2) generation showed a bimodal distribution of seed fertility (Fig. 2). No homozygotes of T65 allele were present in consecutive DNA markers from KGC4_19.55 to KGC4_21.33, supporting that SDV2 candidate gene is linked with these DNA markers (Supplemental Table 2). The four DNA markers, KGC4_20.55, KGC4_20.67, KGC4_21.00, and KGC4_21.33 cosegregated with one another. The genotypes of these four markers explained the variation of seed fertility: The homozygotes of W1297 allele were distributed toward high seed fertility, and the heterozygotes were distributed toward low seed fertility (Fig. 2). Several recombinants between the four DNA marker set and the other DNA markers were subjected to a progeny test to determine if homozygotes of T65 allele appeared or not (Table 3): All the recombinants supported cosegregation of the candidate SDV2 with the four DNA markers. In the BC₁F₇(2) generation, the segregation ratio of W1297 homozygote: heterozygote: T65 homozygote is 49:80:0, which is fitted to the expected ratio 1:2:0, the gene model that homozygotes of T65 allele at SDV2 locus are aborted (Fig. 1).

Fig. 2.

Frequency distribution of seed fertility of BC₁F₇(2) population. Two classified genotypes were assessed for KGC4_21.00 as indicated: grey, homozygotes for W1297; and white, heterozygotes. No homozygotes for T65 appeared in this population.

Table 3.

Segregation of DNA markers of the progeny of the informative recombinants in BC₁F₇(2) around SDV2 locus

BC1F7(2) Recombinant	Genotype at the DNA markera							Seed Fertility	Segregation in BC1F8(2) generationb
	KGC4_19.95	KGC4_20.12	KGC4_20.55	KGC4_21.00	KGC4_21.33	KGC4_21.54	KGC4_22.07		KGC4_20.12	KGC4_21.00	KGC4_22.07
1	W	W	W	W	W	H	H	0.88			15:19:13
2	H	H	H	H	H	T	T	0.69		10:36:0
3	W	W	H	H	H	H	H	0.71		10:17:0
4	W	H	H	H	H	H	H	0.57	17:24:0

^a W, H, and T respectively denote homozygotes for W1297, heterozygotes and homozygotes for T65.

^b Number of homozygotes for W1297, heterozygotes and homozygotes for T65 are shown in this order.

If the SDV2 gene is the counterpart of SDV1 gene, lower fertility of heterozygotes than homozygotes of W1297 allele at SDV2 locus can be explained by abortion after fertilization, the same gene model as SDV1 (Toyomoto et al. 2019). Sterile seeds of BC₁F₇(2) plants were dehusked to see if abortion occurred before or after fertilization. The proportion of seeds aborted after fertilization for heterozygotes for KGC4_21.00 was higher than that for homozygotes of the W1297 allele (Fig. 3). This experimental result also supports that SDV2 is located on chromosome 4, closely linked with the DNA marker KGC4_21.00.

Fig. 3.

Scatter diagram between seed fertility (the ratio of fertile seeds) and the ratio of seeds aborted after fertilization in BC₁F₇(2) population. Two classified genotypes were assessed for KGC4_21.00 as indicated: solid circles, homozygotes for W1297; and open circles, heterozygotes.

Based on the inference that the genotype of candidate SDV2 gene was the same as KGC4_21.00, a linkage map of candidate SDV2 gene was constructed (Fig. 4). The candidate SDV2 gene is located on the long arm of chromosome 4.

Fig. 4.

Linkage map showing SDV2 gene on the long arm of chromosome 4. A: RFLP framework map of chromosome 4 modified from Harushima et al. (1998). B: Linkage map of the SDV2 gene constructed from the BC₁F₇(2) population (n = 128). C: The comparative physical map of the duplicated region of chromosomes 4 and 6. Blue arrows indicate genes annotated by Rice Genome Annotation Project. LOC_Os06g20500 and LOC_Os04g34670 are paralogous genes to each other. LOC_Os06g20470 and LOC_Os06g20530 are also paralogous genes to each other. Green arrows indicate DNA sequences on chromosome 4 homologous to the genes on chromosome 6. The direction of the arrowhead indicates the strand of DNA sequences shown in Supplemental Table 3. DNA markers, genes and/or DNA sequences located near each other on Nipponbare pseudomolecules are connected by dotted lines. Homologous DNA sequences to each other on chromosomes 4 and 6 are connected by long dashed lines.

Verification of SDV2

If the gene closely links with the above four DNA markers on chromosome 4 is the true SDV2, the combination of sdv1-m/sdv1-m, the homozygote of W1297 allele at SDV1 locus, and sdv2-s/sdv2-s, the homozygote of T65 allele at SDV2 locus, should not be found in a population in which both SDV1 and SDV2 genes are segregating (Fig. 1). We developed such experimental population: a heterozygote at candidate SDV2 locus was selected with the aid of DNA marker in a BC₁F₅(2) population, and was backcrossed to T65 as pollen donor to develop BC₂F₁(1, 2) population. This population showed a bimodal distribution of seed fertility, which is explainable by the genotype of KGC4_21.00, a DNA marker cosegregating with candidate SDV2 (Fig. 5). This population shared the genotype Sdv1-s/sdv1-m (Fig. 1). The fact that homozygotes of T65 allele at KGC4_21.00 showed low seed fertility, and heterozygotes showed high seed fertility supports our genetic model.

Fig. 5.

Frequency distribution of seed fertility of BC₂F₁(2) population. Two classified genotypes were assessed for KGC4_21.00 as indicated: grey, homozygotes for T65; and white, heterozygotes.

Then we examined the BC₂F₂(1, 2) population derived from BC₂F₁(1, 2) plants heterozygous both at KGC6_11.74, a cosegregating DNA marker of SDV1, and KGC4_21.00, a cosegregating DNA marker of candidate SDV2. The combination of genotypes of the both DNA markers of 1,373 plants were examined. Among the possible nine genotype combinations, the homozygote of W1297 allele at KGC6_11.74 combined with the homozygote of T65 allele at KGC4_21.00 did not exist (Table 4, Fig. 1). This result verifies that SDV2 gene is tightly linked with a DNA marker KGC4_21.00 located on chromosome 4.

Table 4.

Segregation of DNA markers KGC4_21.00 and KGC6_11.74 of BC₂F₂(1, 2) generation

		Genotype of KGC4_21.00a			Sum	χ2 (1:2:1)
		T	H	W
Genotypes of KGC6_11.74 a	T	137	190	70	397	P < 0.001
	H	211	384	108	703	P < 0.001
	W	0	218	55	273	P < 0.001
Sum		348	792	233
χ2 (1:2:1)		P < 0.001	P = 0.258	P = 0.205

^a T, H, and W denote homozygotes for T65, heterozygotes, and homozygotes for W1297, respectively.

The segregation ratio of KGC4_21.00 were highly distorted in all the three genotypes at the KGC6_11.74 locus (Table 4). The distortion of KGC4_21.00 under homozygotes of T65 allele and heterozygotes at the KGC6_11.74 locus cannot be explained by the combination of SDV1 and SDV2 genes. A BC₃F₁(1, 2) plant heterozygous at KGC6_11.74 and KGC4_21.00 was reciprocally backcrossed to T65 to develop a BC₄F₁(1, 2) generation (Fig. 1). When T65 was used as the egg donor, the genotype ratio of KGC4_21.00 deviated significantly from the expected ratio 1:1 (Table 5). This suggests that segregation distortion was caused by a gene causing male gamete abortion linked with SDV2: pollen carrying the W1297 allele were aborted (Table 5). The combination of the two genotypes KGC6_11.74 and KGC4_21.00 fitted to the ratio assuming their independent inheritance for reciprocal crosses (Table 5). If SDV1 and SDV2 gametophyically interacted to cause pollen sterility or egg sterility, the ratio of plants with heterozygotes for KCG4_21.00 and homozygotes for KGC6_11.74 or plants with heterozygotes for KGC6_11.74 and homozygotes for KCG4_21.00 would be reduced from the expected ratio assuming their independent inheritance, as seen in DPL1 and DPL2 genes (Mizuta et al. 2010). Explanation of these genes are in Discussion section. The independent inheritance of KGC6_11.74 and KCG4_21.00 suggests that the genes on SDV1 and SDV2 do not interact gametophytically.

Table 5.

Segregation of BC₄F₁(1, 2) at KGC6_11.74 and KGC4_21.00

		T65 as pollen donor					T65 as egg donor
		KGC4_21.00a		Sum	χ2 (1: 1)		KGC4_21.00a		Sum	χ2 (1: 1)
		T	H				T	H
KGC6_11.74a	T	10	11	21			22	4	26
	H	6	14	20			19	2	21
	Sum	16	25		P = 0.16		41	6		P < 0.001
χ2 (1: 1)				P = 0.875					P = 0.466
χ2 for independence between
KGC6_11.74 and KGC4_21_00				P = 0.248					P = 0.550

^a T and H denote homozygotes for T65 and heterozygotes, respectively.

Search for candidate gene of SDV1

We narrowed down the chromosomal location of SDV1 to approximately 333 kbp between the DNA markers KGC6_11.53 and KGC6_11.87. According to Rice Genome Annotation Project (http://rice.uga.edu/) (Kawahara et al. 2013), this region is predicted to contain 30 protein-coding genes (Table 6). BLAST search for orthologous genes in O. meridionalis genome was performed using MEGABLAST optimized for highly similar sequences (https://blast.ncbi.nlm.nih.gov/Blast.cgi), with the genomic sequence of predicted genes used as query, and the genomic sequence O. meridionalis accession W2112 (coded GCA_000338895.3) used as subject. Most of the genes ranging from LOC_Os06g20130 to LOC_Os06g20430, and LOC_Os06g20570 and LOC_Os06g20610 have conserved DNA sequences in chromosome 6 of W2112 in the same order. The four genes, LOC_Os06g20470, LOC_Os06g20500, LOC_Os06g20530 and LOC_Os06g20550, have conserved DNA sequences in the candidate chromosomal region of SDV2 gene on chromosome 4, not on chromosome 6 of W2112 (Table 6, Supplemental Table 2). These results suggest that these four genes could be candidate genes for SDV1, and that their orthologous genes located on chromosome 4 of W2112 could be the candidate genes for SDV2.

Table 6.

Predicted genes located between KGC6_11.53 and KGC6_11.87, and conserved sequences in O. meridionalis genome^a

Predicted genesb				Conserved sequences in O. meridionalis genomec
Name	Protein	Location		Chromosome	Location
		(kb)d			(kb)a
LOC_Os06g20130	expressed protein	11549
LOC_Os06g20140	aspartic proteinase nepenthesin-1 precursor, putative, expressed	11552		6	11637
LOC_Os06g20150	peroxidase precursor, putative, expressed	11559		6	11645
LOC_Os06g20180	expressed protein	11575		6	11649
LOC_Os06g20190	aspartic proteinase nepenthesin-2 precursor, putative, expressed	11581		6	11658
LOC_Os06g20200	gibberellin receptor GID1L2, putative, expressed	11586
LOC_Os06g20240	latency associated nuclear antigen, putative, expressed	11619		2	28908
LOC_Os06g20260	expressed protein	11627
LOC_Os06g20270	hypothetical protein	11631
LOC_Os06g20300	expressed protein	11657		6	11680
LOC_Os06g20310	expressed protein	11673		6	11753
LOC_Os06g20320	peptidyl-prolyl cis-trans isomerase, FKBP-type, putative, expressed	11676		6	11761
LOC_Os06g20330	hypothetical protein	11683
LOC_Os06g20340	dual specificity protein phosphatase, putative, expressed	11687		6	11769
LOC_Os06g20354	PPR repeat domain containing protein, putative, expressed	11693		6	11771
LOC_Os06g20370	microtubule associated protein, putative, expressed	11701		6	11779
LOC_Os06g20380	expressed protein	11706
LOC_Os06g20390	expressed protein	11712		6	11786
LOC_Os06g20400	DHHC zinc finger domain containing protein, expressed	11722		6	11796
LOC_Os06g20410	BAH domain containing protein, expressed	11734		6	11807
LOC_Os06g20420	expressed protein	11734
LOC_Os06g20430	BPI/LBP family protein At3g20270 precursor, putative, expressed	11741		6	11816
LOC_Os06g20450	hypothetical protein	11747		8	17376
LOC_Os06g20470	expressed protein	11752		4	19942
LOC_Os06g20480	expressed protein	11761
LOC_Os06g20500	tRNA-splicing endonuclease positive effector-related, putative, expressed	11787		4	19891
LOC_Os06g20530	expressed protein	11795		4	19942
LOC_Os06g20550	expressed protein	11829		4	19900
LOC_Os06g20570	glycosyltransferase, putative, expressed	11844		6	11823
LOC_Os06g20610	seven in absentia protein family domain containing protein, expressed	11867		6	11835

^a The sequence of O. meridionalis accession W2112 (GCA_000338895.3). The detailed assembly information is shown in Supplemental Table 4.

^b http://rice.uga.edu/.

^c Homogous DNA sequences detected by NCBI megablast.

^d Approximate location in Nipponbare genome (IRGSP 1.0 pseudomolecule) (kb). The detailed assembly information is shown in Supplemental Table 4.

We further examined the candidate chromosomal region of SDV1, using the genome sequences of eight AA genome Oryza species (Supplemental Tables 3, 4). We found that the each of the five genes, LOC_Os06g20470, LOC_Os06g20480, LOC_Os06g20500, LOC_Os06g20530 and LOC_Os06g20550, were duplicated and located nearby on chromosomes 6, and also located on chromosome 4 with a few exceptions in the three species, O. sativa, O. rufipogon and O. nivara (Supplemental Table 3). LOC_Os06g20470 and LOC_Os06g20530 are examples of a pair of duplicate genes on chromosome 6 (Fig. 4, Supplemental Table 3). The paralogous sequences of O. sativa on chromosome 4 are located within SDV2 candidate region (Fig. 4, Supplemental Table 3). On the other hand, the orthologous sequences of the four species, O. meridionalis, O. glaberrima, O. barthii and O. glumaepatula, are located only on chromosome 4. The order and orientation of the conserved gene sequences on chromosome 4 are shared by all eight species except that O. longistaminata shares the order and orientation of the above orthologous sequences with the other seven species on chromosome 11 (Supplemental Table 3).

The five genes located upstream and downstream of the duplicated region in Supplemental Table 3 were selected because they are single copy genes in O. sativa, and histories of their orthologous genes could be discussed simply. The seven species, O. sativa, O. rufipogon, O. nivara, O. meridionalis, O. glaberrima, O. barthii, and O. glumaepatula, share the same orders and orientations on the five upstream genes on chromosome 6. O. longistaminata carries these genes in the opposite orientation and order on chromosome 3. As for the five downstream genes, the all eight AA genome Oryza species share them in the same order and orientation on chromosome 6.

These data strongly suggest that the duplication and loss of the candidate chromosomal region of SDV1 of an O. sativa cultivar Nipponbare and O. meridionalis accession W2112 are not confined to the two accessions but can be extended to the history of AA genome Oryza species. They also support the above inference that the five genes, LOC_Os06g20470, LOC_Os06g20480, LOC_Os06g20500, LOC_Os06g20530 and LOC_Os06g20550, are the candidate SDV1 genes. Among them, only LOC_Os06g20500 is predicted to have a specific function described as tRNA-splicing endonuclease positive effector-related, putative, expressed, and the other four genes are described to encode expressed proteins. According to Rice Genome Annotation Project (http://rice.uga.edu/) (Kawahara et al. 2013), LOC_Os06g20500 is expressed in all ten tissues including seed-5 days after pollination (DAP), seed-10 DAP, embryo-25 DAP and endosperm-25 DAP. On the other hand, the other four candidate genes are not expressed in these tissues.

Discussion

In this study, we narrowed down the chromosomal location of SDV1 to approximately 333 kbp between the DNA markers KGK KG C6_11.53 and KG C6_11.87 on chromosome 6. The chromosomal region is highly conserved among many O. meridionalis accessions, which is consistent with linkage analysis of SDV1. We also performed linkage analysis of the candidate of SDV2, and found that it is located between KGC4_20.12 and KGC4_21.54, co-segregating with KGC4_20.55, KGC4_20.67, KGC4_21.00 and KGC4_21.33 on chromosome 4. The candidate chromosomal location of SDV2 proved to be correct in an experiment using a population in which both SDV1 and SDV2 genes were segregating. The chromosomal region covering SDV1 gene was predicted to contain 30 protein-coding genes in O. sativa. Among them, five genes had conserved DNA sequences in the candidate chromosomal region of SDV2 gene on chromosome 4, not on chromosome 6, of the genome of an O. meridionalis accession W2112. These results suggest that these five genes could be candidate genes for SDV1, and that their orthologous genes located on chromosome 4 of W2112 could be candidate genes for SDV2. The gene expression data suggested that the candidate can be confined to LOC_Os06g20500. However, the gene has two paralogous genes in the rice genome Os-Nipponbare-Reference-IRGSP-1.0, LOC_Os04g34670 located on the candidate SDV2 region of chromosome 4 and LOC_Os03g32526 located on chromosome 3. They are also expressed in in all ten tissues including seed-5 days after pollination (DAP), seed-10 DAP, embryo-25 DAP and endosperm-25 DAP. Further linkage analysis on SDV1 and SDV2, DNA sequence comparison of candidate chromosomal regions between chromosomes 4 and 6 among AA genome Oryza species and fine-tuned gene expression study targeting the above candidate genes will contribute to identification of SDV1 and SDV2 genes.

Based on Supplemental Table 3, three phylogenetic trees of the eight species can be drawn (Fig. 6). Because the Oryza genome has undergone whole genome duplication (Salse et al. 2008, Yu et al. 2005), we first assumed that hypothetical ancestor carried functional SDV1 and SDV2 genes (Fig. 6). Fig. 6A reflects phylogeny from genomic data by Zhang et al. (2014) and Stein et al. (2018). At least three mutation events including gene loss on the SDV1 locus and one mutation on the SDV2 occurred in this model. Mizuta et al. (2010) detected paralogous hybrid incompatibility genes, DOPPELGANGER1 (DPL1) and DOPPELGANGER2 (DPL2), cause loss of germination of pollen in intraspecific cross in O. sativa. Independent disruptions of DPL1 and DPL2 occurred in indica and japonica, respectively. Pollen carrying two defective DPL alleles became nonfunctional and did not germinate, suggesting an essential role for DPLs in pollen germination. DPL1 corresponds to LOC_Os01g15448 located at 8.65 Mb on chromosome 1, and DPL2 gene corresponds to LOC_Os06g08510 located at 4.20 Mb on chromosome 6. Loss-of-function mutations of DPL1 genes emerged multiple times in indica and its wild ancestor O. rufipogon, and the DPL2 gene defect is specific to japonica cultivars. Bikard et al. (2009) showed that in a cross between two Arabidopsis thaliana accessions Columbia-0 and Cape Verde Island accession Cvi-0, the LD1.1 locus on chromosome 1 and LD1.5 locus on chromosome 5 interact epistatically to control recessive embryonic lethality, and this was explained by divergent evolution between paralogs of essential duplicated genes. This genetic model is same as that of SDV1 and SDV2. These reports suggest that Fig. 6A assuming multiple mutations is not improbable. Fig. 6B assumes the least number of mutations at the SDV1 and SDV2 loci: The above mutation event occurred once at each locus. This tree is different from Fig. 6A reflecting the evolution of the whole genome. It seems improbable that duplication and loss of SDV1 and SDV2 genes occurred independent of the evolution of the whole genome. Fig. 6C assumes that the hypothetical ancestor carried only the functional SDV2 gene, and that duplication of SDV2 turning into SDV1 and loss of function of the original SDV2 gene occurred in the common ancestor of O. sativa, O. rufipogon and O. nivara after divergence from the common ancestor of O. glaberrima and O. barthii. Remarkably it seems that both gene duplication and loss occurred in one linage in an evolutionarily short period. However, this hypothesis is consistent with the evolution of the whole genome, and assumes the least mutation events. Further investigation of SDVs would deepen our knowledge of the duplication phenomenon and speciation.

Fig. 6.

Working hypotheses for the evolution of functionally duplicate genes, SDV1 and SDV2, in AA genome Oryza species. A, a model based on phylogeny from genomic data by Zhang et al. (2014) and Stein et al. (2018), assuming that the hypothetical ancestor carried Sdv1, the functional allele at SDV1 locus, and Sdv2, the functional allele at the SDV2 locus. Mutation from Sdv1 to the abortive allele sdv1 including gene loss at the SDV1 locus occurred once, and that from Sdv2 to the abortive allele sdv2 on SDV2 locus occurred at least three times. O. longistaminata is excluded because Zhang et al. (2014) and Stein et al. (2018) did not test this species. B, a model assuming least mutations of the SDV1 and SDV2 loci, and that the hypothetical ancestor carried both Sdv1 and Sdv2. The above mutation event occurred once at each locus. C, a model based on phylogeny from genomic data by Zhang et al. (2014) and Stein et al. (2018), assuming that the hypothetical ancestor carried only Sdv2. The duplication of SDV2 turning into SDV1 and loss of function of the original SDV2 occurred in the common ancestor of O. sativa, O. rufipogon and O. nivara after divergence from the common ancestor of O. glaberrima and O. barthii.

There have been many reports of hybrid sterility genes found in interspecific crosses among AA genome Oryza species located close to SDV1 and SDV2. Supplemental Table 5 shows isolated or DNA-marker tagged genes causing sterility found in distant crosses located on chromosomes 4 and 6. Among the genes above, Cif₂, Su-Cif and cim are located close to HD1 (Koide et al. 2008, Matsubara et al. 2003), hence SDV1, on chromosome 6. Matsubara et al. (2003) reported unidirectional cross-incompatibility detected in advanced generation of backcrossing between O. rufipogon accessions and O. sativa accessions. The near-isogenic line of the line named T65wx (japonica type) carrying an alien segment of chromosome 6 from O. rufipogon gave a reduced seed setting only when crossed with T65wx as the male. The four causal genes, Cif₁, Cif₂, cim and Su-Cif, are involved in this phenomenon (Koide et al. 2008). This unidirectional cross-incompatibility controlled by these genes occurred during the development of hybrid seeds. Though the morphology and gene location are similar between the unidirectional cross-incompatibility (Matsubara et al. 2003) and seed abortion by SDV1 and SDV2 genes, their genetic models are different: homozygotes of loss of function genes can not develop into mature seeds in our model. The location of S6 is close to that of SDV1. However, these genes are distinct judging from the gene function.

As for the genes on chromosome 4, locations of S9 and DUPLICATED GAMETOPHYTIC STERILITY 1 (DGS1) are close to that of SDV2. S9 was detected in intraspecific cross among O. sativa (Zhao et al. 2006), causing female gamete abortion. DGS1 gene was detected in the cross between O. sativa and O. nivara (Nguyen et al. 2017). Pollen fertility was controlled by gametophytic gene combination on DGS1 and DUPLICATED GAMETOPHYTIC STERILITY 2 (DGS2) loci: pollen carrying O. nivara allele at DGS1 and O. sativa allele at DGS2 is sterile. DGS1 is located on chromosome 4, and DGS2 is located on chromosome 7. The two genes encode protein homologous to DNA-dependent RNA polymerase (RNAP) III subunit C4 (RPC4). The combination of loss of the functional allele at DGS1 and DGS2 caused hybrid pollen sterility gametophytically. However, S9 and DGS1 were distinct from SDV2, judging from the gene segregation pattern.

In BC₂F₂(1, 2) population and BC₄F₁(1, 2) generation using T65 as egg donor, genotypes of KGC4_21.00 was significantly skewed from the expected Mendelian ratio (Tables 4, 5). No gametophytic interaction genes linked with KGC4_21.00 and KGC6_11.74 was observed. These data suggest that a gene causing male gamete abortion carrying the allele from W1297 is linked with SDV2. Such a gene might have been fixed in BC₁F₇(2) for mapping SDV2 (Fig. 1). The recombination value of the gene and KGC4_21.00 is roughly estimated to be 0.128 (6/47). The chromosomal locations of SDV2 (Supplemental Table 2, Fig. 4), S9 (Zhao et al. 2006) and DGS1 (Nguyen et al. 2017) (Supplemental Table 5) suggest that the gene causing male gamete abortion is different from S9 and DGS1.

Isolation of SDV1 and SDV2 genes will contribute to elucidation of speciation of AA genome Oryza species, monitoring possible ongoing hybridization between O. rufipogon and O. meridionalis in Australia, molecular biology of seed development and rice breeding: Useful genes located close to sdv1-m gene on chromosome 6 in O. meridionalis genome will be efficiently incorporated into O. sativa genome in combination with Sdv2-m on chromosome 4. Supplemental Table 3 suggests that this strategy could be applied to using O. barthii, O. glaberrima and O. glumaepatula as useful gene donors.

Author Contribution Statement

K.I. conceptualized and developed the methodology. D.T., Y.S., and M.U. validated the research. D.T., Y.S., M.U., and K.I. conducted the formal analysis and investigation. S.T., T.S., R.H., R.I., and K.I. provided the resources. D.T., Y.S., M.U., and K.I. curated the data. D.T. and K.I. prepared the original draft. D.T., M.U., S.T., T.S., R.H., R.I., and K.I. reviewed and edited the writing. D.T., Y.S., M.U., and K.I. were responsible for the visualization. K.I. supervised the project. K.I. and R.I. administered the project and acquired funding. All authors have read and consented to the published this version of the manuscript.

Acknowledgments

We are grateful to the National Institute of Genetics supported by the National Bioresource Project, MEXT, Japan, for their kind provision of seeds of wild rice accessions listed in Table 1 except Jpn1 and Jpn2. We gratefully acknowledge the Genebank of National Institute of Agrobiological Sciences for their kind provision of WRC lines. This research was funded by JSPS KAKENHI Grant Numbers JP16H05777 and 21K18118 from the Japan Society for the Promotion of Science. We thank Noboru Takasaki, Rieko Nomura, and all the members of the Plant Breeding Laboratory, Kagoshima University, Japan, for their technical assistance.

Literature Cited

•  Banaticla-Hilario,  M.C.N.,  K.L.  McNally,  R.G.  van den Berg and  N.R.S.  Hamilton (2013) Crossability patterns within and among Oryza series Sativae species from Asia and Australia. Genet Resour Crop Evol 60: 1899–1914.
•  Bikard,  D.,  D.  Patel,  M.  Le,  V.  Giorgi,  C.  Camilleri,  M.J.  Bennett and  O.  Loudet (2009) Divergent evolution of duplicate genes leads to genetic incompatibilities within A. thaliana. Science 323: 623–626.
•  Brozynska,  M.,  A.  Furtado and  R.J.  Henry (2016) Genomics of crop wild relatives: Expanding the gene pool for crop improvement. Plant Biotechnol J 14: 1070–1085.
•  Brozynska,  M.,  D.  Copetti,  A.  Furtado,  R.A.  Wing,  D.  Crayn,  G.  Fox,  R.  Ishikawa and  R.J.  Henry (2017) Sequencing of Australian wild rice genomes reveals ancestral relationships with domesticated rice. Plant Biotechnol J 15: 765–774.
•  Busungu,  C.,  S.  Taura,  J.I.  Sakagami and  K.  Ichitani (2016) Identification and linkage analysis of a new rice bacterial blight resistance gene from XM14, a mutant line from IR24. Breed Sci 66: 636–645.
•  Cheng,  C.,  S.  Tsuchimoto,  H.  Ohtsubo and  E.  Ohtsubo (2002) Evolutionary relationships among rice species with AA genome based on SINE insertion analysis. Genes Genet Syst 77: 323–334.
•  Harushima,  Y.,  M.  Yano,  A.  Shomura,  M.  Sato,  T.  Shimano,  Y.  Kuboki,  T.  Yamamoto,  S.Y.  Lin,  B.A.  Antonio,  A.  Parco et al. (1998) A high density rice genetic map with 2275 markers using a single F₂ population. Genetics 148: 479–494.
•  Henry,  R.J. (2019) Australian wild rice populations: A key resource for global food security. Front Plant Sci 10: 1354.
•  Henry,  R.J.,  N.  Rice,  D.L.E.  Waters,  S.  Kasem,  R.  Ishikawa,  Y.  Hao,  S.  Dillon,  D.  Crayn,  R.  Wing and  D.  Vaughan (2010) Australian Oryza: Utility and conservation. Rice (NY) 3: 235–241.
•  Htet,  A.H.,  S.  Makabe,  H.  Takahashi,  P.A.  Samuel,  Y.I.  Sato and  I.  Nakamura (2022) A large deletion within intron 20 sequence of single-copy PolA1 gene as a useful marker for the speciation in Oryza AA-genome species. Breed Sci 72: 267–273.
•  Ichitani,  K.,  D.  Yamaguchi,  S.  Taura,  Y.  Fukutoku,  M.  Onoue,  K.  Shimizu,  F.  Hashimoto,  Y.  Sakata and  M.  Sato (2014) Genetic analysis of ion-beam induced extremely late heading mutants in rice. Breed Sci 64: 222–230.
•  Ichitani,  K.,  S.  Taura,  M.  Sato and  T.  Kuboyama (2016) Distribution of Hwc2-1, a causal gene of a hybrid weakness, in the World Rice Core collection and the Japanese Rice mini Core collection: its implications for varietal differentiation and artificial selection. Breed Sci 66: 776–789.
•  Ichitani,  K.,  D.  Toyomoto,  M.  Uemura,  K.  Monda,  M.  Ichikawa,  R.  Henry,  T.  Sato,  S.  Taura and  R.  Ishikawa (2022) New hybrid spikelet sterility gene found in interspecific cross between Oryza sativa and O. meridionalis. Plants 11: 378.
•  Iwata,  H. and  S.  Ninomiya (2006) AntMap: constructing genetic linkage maps using an ant colony optimization algorithm. Breed Sci 56: 371–377.
•  Juliano,  A.B.,  M.E.B.  Naredo,  B.R.  Lu and  M.T.  Jackson (2005) Genetic differentiation in Oryza meridionalis Ng based on molecular and crossability analyses. Genet Resour Crop Evol 52: 435–445.
•  Kawahara,  Y.,  M.  de la Bastide,  J.P.  Hamilton,  H.  Kanamori,  W.R.  McCombie,  S.  Ouyang,  D.C.  Schwartz,  T.  Tanaka,  J.  Wu,  S.  Zhou et al. (2013) Improvement of the Oryza sativa Nipponbare reference genome using next generation sequence and optical map data. Rice (NY) 6: 4.
•  Koide,  Y.,  M.  Ikenaga,  Y.  Shinya,  K.  Matsubara and  Y.  Sano (2008) Two loosely linked genes controlling the female specificity for cross-incompatibility in rice. Euphytica 164: 753–760.
•  Kojima,  Y.,  K.  Ebana,  S.  Fukuoka,  T.  Nagamine and  M.  Kawase (2005) Development of an RFLP-based rice diversity research set of germplasm. Breed Sci 55: 431–440.
•  Kumagai,  M.,  D.  Nishikawa,  Y.  Kawahara,  H.  Wakimoto,  R.  Itoh,  N.  Tabei,  T.  Tanaka and  T.  Itoh (2019) TASUKE+: A web-based platform for exploring GWAS results and large-scale resequencing data. DNA Res 26: 445–452.
•  Li,  J.,  J.  Zhou,  P.  Xu,  X.  Deng,  W.  Deng,  Y.  Zhang,  Y.  Yang and  D.  Tao (2018) Mapping five novel interspecific hybrid sterility loci between Oryza sativa and Oryza meridionalis. Breed Sci 68: 516–523.
•  Matsubara,  K.,  Khin-Thinder and  Y.  Sano (2003) A gene block causing cross-incompatibility hidden in wild and cultivated rice. Genetics 165: 343–352.
•  Mizuta,  Y.,  Y.  Harushima and  N.  Kurata (2010) Rice pollen hybrid incompatibility caused by reciprocal gene loss of duplicated genes. Proc Natl Acad Sci USA 107: 20417–20422.
•  Naredo,  M.E.B.,  A.B.  Juliano,  B.R.  Lu and  M.T.  Jackson (1997) Hybridization of AA genome rice species from Asia and Australia I. crosses and development of hybrids. Genet Resour Crop Evol 44: 17–23.
•  Naredo,  M.E.B.,  A.B.  Juliano,  B.R.  Lu and  M.T.  Jackson (1998) Taxonomic status of Oryza glumaepatula Steud. II. hybridization between new world diploids and AA genome species from Asia and Australia. Genet Resour Crop Evol 45: 205–214.
•  Nguyen,  G.N.,  Y.  Yamagata,  Y.  Shigematsu,  M.  Watanabe,  Y.  Miyazaki,  K.  Doi,  K.  Tashiro,  S.  Kuhara,  H.  Kanamori,  J.  Wu et al. (2017) Duplication and loss of function of genes encoding RNA polymerase III subunit C4 causes hybrid incompatibility in rice. G3 (Bethesda) 7: 2565–2575.
•  Ohtsubo,  H.,  C.  Cheng,  I.  Ohsawa,  S.  Tsuchimoto and  E.  Ohtsubo (2004) Rice retroposon p-SINE1 and origin of cultivated rice. Breed Sci 54: 1–11.
•  Ohyanagi,  H.,  T.  Ebata,  X.  Huang,  H.  Gong,  M.  Fujita,  T.  Mochizuki,  A.  Toyoda,  A.  Fujiyama,  E.  Kaminuma,  Y.  Nakamura et al. (2016) OryzaGenome: Genome diversity database of wild Oryza species. Plant Cell Physiol 57: e1.
•  Peng,  H.,  K.  Wang,  Z.  Chen,  Y.  Cao,  Q.  Gao,  Y.  Li,  X.  Li,  H.  Lu,  H.  Du,  M.  Lu et al. (2020) MBKbase for rice: An integrated omics knowledgebase for molecular breeding in rice. Nucleic Acids Res 48: D1085–D1092.
•  Salse,  J.,  S.  Bolot,  M.  Throude,  V.  Jouffe,  B.  Piegu,  U.M.  Quraishi,  T.  Calcagno,  R.  Cooke,  M.  Delseny and  C.  Feuillet (2008) Identification and characterization of shared duplications between rice and wheat provide new insight into grass genome evolution. Plant Cell 20: 11–24.
•  Sano,  Y. (1990) The genic nature of gamete eliminator in rice. Genetics 125: 183–191.
•  Sotowa,  M.,  K.  Ootsuka,  Y.  Kobayashi,  Y.  Hao,  K.  Tanaka,  K.  Ichitani,  J.M.  Flowers,  M.D.  Purugganan,  I.  Nakamura,  Y.  Sato et al. (2013) Molecular relationships between Australian annual wild rice, Oryza meridionalis, and two related perennial forms. Rice (NY) 6: 26.
•  Stein,  J.C.,  Y.  Yu,  D.  Copetti,  D.J.  Zwickl,  L.  Zhang,  C.  Zhang,  K.  Chougule,  D.  Gao,  A.  Iwata,  J.L.  Goicoechea et al. (2018) Genomes of 13 domesticated and wild rice relatives highlight genetic conservation, turnover and innovation across the genus Oryza. Nat Genet 50: 285–296.
•  Toyomoto,  D.,  M.  Uemura,  S.  Taura,  T.  Sato,  R.  Henry,  R.  Ishikawa and  K.  Ichitani (2019) Segregation distortion observed in the progeny of crosses between Oryza sativa and O. meridionalis caused by abortion during seed development. Plants 8: 398.
•  Wei,  F.J.,  Y.C.  Tsai,  H.P.  Wu,  L.T.  Huang,  Y.C.  Chen,  Y.F.  Chen,  C.C.  Wu,  Y.T.  Tseng and  Y.C.  Hsing (2016) Both Hd1 and Ehd1 are important for artificial selection of flowering time in cultivated rice. Plant Sci 242: 187–194.
•  Yamagata,  Y.,  K.T.  Win,  Y.  Miyazaki,  C.  Ogata,  H.  Yasui and  A.  Yoshimura (2019) Development of introgression lines of AA genome Oryza species, O. glaberrima, O. rufipogon, and O. nivara, in the genetic background of O. sativa L. cv. Taichung 65. Breed Sci 69: 359–363.
•  Yamazaki,  Y.,  S.  Sakaniwa,  R.  Tsuchiya,  K.  Nonomura and  N.  Kurata (2010) Oryzabase: an integrated information resource for rice science. Breed Sci 60: 544–548.
•  Yin,  H.,  M.  Akimoto,  R.  Kaewcheenchai,  M.  Sotowa,  T.  Ishii and  R.  Ishikawa (2015) Inconsistent diversities between nuclear and plastid genomes of AA genome species in the genus Oryza. Genes Genet Syst 90: 269–281.
•  Yoshimura,  A.,  H.  Nagayama,  Sobrizal,  T.  Kurakazu,  P.L.  Sanchez,  K.  Doi,  Y.  Yamagata and  H.  Yasui (2010) Introgression lines of rice (Oryza sativa L.) carrying a donor genome from the wild species, O. glumaepatula Steud. and O. meridionalis Ng. Breed Sci 60: 597–603.
•  Yu,  J.,  S.  Hu,  J.  Wang,  G.K.  Wong,  S.  Li,  B.  Liu,  Y.  Deng,  L.  Dai,  Y.  Zhou,  X.  Zhang et al. (2002) A draft sequence of the rice genome (Oryza sativa L. spp. indica). Science 296: 79–92.
•  Yu,  J.,  J.  Wang,  W.  Lin,  S.  Li,  H.  Li,  J.  Zhou,  P.  Ni,  W.  Dong,  S.  Hu,  C.  Zeng et al. (2005) The genomes of Oryza sativa: a history of duplications. PLoS Biol 3: e38.
•  Zhang,  Q.J.,  T.  Zhu,  E.H.  Xia,  C.  Shi,  Y.L.  Liu,  Y.  Zhang,  Y.  Liu,  W.K.  Jiang,  Y.J.  Zhao,  S.Y.  Mao et al. (2014) Rapid diversification of five Oryza AA genomes associated with rice adaptation. Proc Natl Acad Sci USA 111: E4954–E4962.
•  Zhao,  Z.,  C.  Wang,  L.  Jiang,  S.  Zhu,  H.  Ikehashi and  J.  Wan (2006) Identification of a new hybrid sterility gene in rice (bi Oryza sativa L.). Euphytica 151: 331–337.