Genomic structural variation-mediated allelic suppression causes hybrid male sterility in rice

Abstract

Hybrids between divergent populations commonly show hybrid sterility; this reproductive barrier hinders hybrid breeding of the japonica and indica rice (Oryza sativa L.) subspecies. Here we show that structural changes and copy number variation at the Sc locus confer japonica-indica hybrid male sterility. The japonica allele, Sc-j, contains a pollen-essential gene encoding a DUF1618-domain protein; the indica allele, Sc-i, contains two or three tandem-duplicated ~ 28-kb segments, each carrying an Sc-j-homolog with a distinct promoter. In Sc-j/Sc-i hybrids, the high-expression of Sc-i in sporophytic cells causes suppression of Sc-j expression in pollen and selective abortion of Sc-j-pollen, leading to transmission ratio distortion. Knocking out one or two of the three Sc-i copies by CRISPR/Cas9 rescues Sc-j expression and male fertility. Our results reveal the gene dosage-dependent allelic suppression as a mechanism of hybrid incompatibility, and provide an effective approach to overcome the reproductive barrier for hybrid breeding.

Fig. 1

Cloning of Sc reveals variations in genomic structure and copy number. a Pollen phenotypes of T65, E5, and their F₁. Arrows indicate smaller, sterile pollen grains (scale bars, 50 μm). The self-fertilized F₂ family shows a severely distorted segregation of the alleles (P  < 0.001 for the chi-square test). b Fine-mapping of Sc using F₂ plants of the T65/E5 cross. The markers indicate the positions (kb) in the BAC of the japonica cultivar (cv) Nipponbare (Nip). c Genomic configurations of the Sc alleles in japonica lines (from GenBank except for T65) and indica lines (determined in this study). The Sc-i allele variants have three or two tandemly duplicated segments, each containing the gene homologous to Os03g0247300. T1 (2016 bp) and T2 (9087 bp) are transposon insertions (a DDE-type and a Rim2/Hipa-type, respectively). The Sc-i paralogous genes have recombinant promoter/upstream sequences different from that of Sc-j. d Fiber-FISH images of the MH63 BAC (~ 180 kb) and the E5 chromatin (genomic) DNA (gDNA). For the BAC fiber-FISH, Probes I and II were labeled for red fluorescence, and the whole BAC DNA was labeled for green fluorescence. The labeled Probe I (or Probe II) was mixed with the labeled BAC for the hybridization. For fiber-FISH of E5 gDNA, Probes I and II were labeled for red and green fluorescence, respectively, and mixed for hybridization. Scale bars, 2 μm. Note that the BAC and chromatin DNA fibers for the same segments had different extension levels, similar to the previous observations. FF full fertility, SS semi-sterility

Fig. 2

Expression of the Sc alleles. a Expression of Sc-i and Sc-j in the developing anthers of E5, T65, and their F₁ by qRT-PCR using a common primer pair (P3/P8, Supplementary Fig. 1) for all the alleles/Sc-i paralogs (except for Sc-is). The values (means of three biological replicates with s.e.m.) were normalized to OsActin 1 transcript levels. b qRT-PCR analysis of Sc-j (using P3/P8) in separated pollen and anther wall cells (at early-bicellular stage) of T65. Data are shown as means ± s.e.m. (n = 3). c Different Sc protein abundances in anthers (at stages of uninucleate late-microspore to early-bicellular pollen) of the parents and the F₁, detected with anti-Sc antibodies

Fig. 3

Functional analysis of Sc-j. a qRT-PCR analysis of relative Sc-j expression levels (using the Sc-j-specific primers P1/P2) in the anthers and isolated pollen (at early-bicellular stage) of the F₁ (T65/E5) and T65, calculated as the ratios of the transcript per copy Sc-j to the transcript of OsActin 1 (two copies), because there is only one Sc-j copy in the F₁. The data are shown as means ± s.e.m. (n = 3), and * and ** indicate significance at P  < 0.05 and P  < 0.01, respectively, by two-tailed Student’s t test (the same for Fig. 3c). b Percentages of cDNA clones from Sc-j, Sc-ia, and Sc-ib (Sc-ib1/ib2) of the F₁ anthers of two stages indicated. c Expression of Sc-j in the early-bicellular stage anthers of two T₁ (T65 background) plants carrying the hemizygous antisense-transgene of Sc-j. The detected transcript levels in Anti-1 and Anti-2 could be mainly from the pollen grains (~ 50%) without the antisense-transgene. d Semi-sterility of three T₁ plants carrying the hemizygous antisense-transgene. WT (T-), a T₁ segregant without the antisense-transgene. Scale bars, 50 μm

Fig. 4

Genomic editing of the Sc-i paralogs rescues Sc-j expression and male fertility in the hybrids. a Targeted editing of the sites (arrowed) in E5 produced a plant (E5-d1) with deletion of one of the ~ 28-kb segments (Supplementary Fig. 6), and editing another site (underlined) specific to Sc-ib1 and Sc-ib2 of E5 identified two plants (E5-ed1 and E5-ed2) with mutated Sc-ib1 and Sc-ib2 and intact Sc-ia (Supplementary Fig. 10). The SNP (blue) in Sc-ib1 and Sc-ib2 formed the PAM (NGG) required for genome editing. Primers P9a-P9e combined with the Sc-i-specific P10 were used for qRT-PCR of the intact and mutated Sc-i transcripts show in (c). b Improvement of pollen fertility (sterile pollen grains arrowed) in the edited F₁ plants from crossing E5-d1, E5-ed1, and E5-ed2 with T65. Scale bars, 50 μm. c Expression levels (in the meiocyte stage anthers) of the Sc-i paralogs in the F₁ (T65/E5), and the intact Sc-ia and mutated Sc-ib1/Sc-ib2 in the edited F₁ plants. Data are shown as means ± s.e.m. (n = 3). d Expression levels of per copy Sc-j (qRT-PCR using P1/P2) in the early-bicellular stage anthers of the intact and edited F₁ plants and T65. The data are shown as means ± s.e.m. (n  = 3), and * and ** indicate significance at P  < 0.05 and P  < 0.01, respectively, by two-tailed Student’s t test

Fig. 5

A proposed model for the molecular genetic mechanism of Sc-mediated HMS. In a japonica–indica hybrid, the Sc-i paralogs are highly expressed in sporophytic cells and the product may interact with the Sc-j promoter region. This might cause transcriptional-repressive epigenetic modification(s) (red asterisk) in the Sc-j promoter region. This putative modification state (Sc-j*) could be retained, through meiosis, in the male gametophytes, thus causing allelic suppression of Sc-j expression and the consequent Sc-j-specific pollen abortion (HMS) and mTRD. Only two of the four pollen grains from a meiocyte, and one Sc-j copy and one set of Sc-i copies in the bicellular and tricellular pollen grains, are shown