Pollen Killer Gene S35 Function Requires Interaction with an Activator That Maps Close to S24, Another Pollen Killer Gene in Rice

Abstract

Pollen killer genes disable noncarrier pollens, and are responsible for male sterility and segregation distortion in hybrid populations of distantly related plant species. The genetic networks and the molecular mechanisms underlying the pollen killer system remain largely unknown. Two pollen killer genes, S24 and S35, have been found in an intersubspecific cross of Oryza sativa ssp. indica and japonica The effect of S24 is counteracted by an unlinked locus EFS Additionally, S35 has been proposed to interact with S24 to induce pollen sterility. These genetic interactions are suggestive of a single S24-centric genetic pathway (EFS-S24-S35) for the pollen killer system. To examine this hypothetical genetic pathway, the S35 and the S24 regions were further characterized and genetically dissected in this study. Our results indicated that S35 causes pollen sterility independently of both the EFS and S24 genes, but is dependent on a novel gene close to the S24 locus, named incentive for killing pollen (INK). We confirmed the phenotypic effect of the INK gene separately from the S24 gene, and identified the INK locus within an interval of less than 0.6 Mb on rice chromosome 5. This study characterized the genetic effect of the two independent genetic pathways of INK-S35 and EFS-S24 in indica-japonica hybrid progeny. Our results provide clear evidence that hybrid male sterility in rice is caused by several pollen killer networks with multiple factors positively and negatively regulating pollen killer genes.

Keywords: Oryza sativa; epistasis; pollen killer; reproductive isolation.

Figure 1

Genetic basis of hybrid male sterility genes and localization of the S35 locus. (A) Pollen phenotypes of NILs for S35, S24, and EFS genes associated with hybrid male sterility. Graphical genotypes (left) and pollen grains (right) of the NILs stained by I₂-KI are shown. Black and white bars denote IR24 (indica) and Asominori (japonica) chromosomes, respectively. The NIL carrying S35 alone (NIL-S35al) were fertile, but cointrogression of the S24 segment from IR24 dominantly activated S35, leading to pollen sterility in the Asominori genetic background (NIL-S35H). The S24-suppressor Efs gene from IR24 dominantly inactivated S24 (NIL-S24H+Efs). (B) Chromosome position of S35. Upper panel: Physical chromosome positions of the S35 locus and DNA markers (1cXXX). Lower panel: Diagram showing the recombination breakpoints and pollen phenotypes of the lines. All recombinant plants carried IR24 homozygous alleles for S24 to activate S35-i. The pollen fertility is shown as the mean (%) ± SD (n = 3–7).

Figure 2

Histological characterization of the sterile S35 heterozygotes. (A)–(C) Mature pollen grains stained with hematoxylin. (A) Pollen grains from Asominori. (B), (C) Pollen grains from the sterile S35 heterozygotes (NIL-S35H). White and black arrowheads indicate normal sperm cells and nuclei of abortive pollen grains, respectively. Pollen grains arrested at the bicellular stage (black arrowheads) and vacuolated pollen at the uninucleate stage are shown in (B). (C) shows that these two pollen grains from NIL-S35H were well filled and similar in size, but the pollen grain on the right has a single bold nucleus instead of two sperm cells as would be observed in normal pollen grains. (D)–(M) Transverse sections showing anther development of Asominori and the S35 heterozygote (NIL-S35H). Five stages of anther development in Asominori and NIL-S35H were compared. Transverse sections were stained with 0.05% Toluidine blue O. (D)–(H) are Asominori, and (I)–(M) are NIL-S35H. (D) and (I) the pollen mother cell stage, (E) and (J) the tetrad stage, (F) and (K) the young microspore stage, (G) and (L) vacuolated pollen stage, (H) and (M) the mature pollen stage. The solid black arrows indicate degenerated pollen grains. PMC, pollen mother cell; T, tapetal layer; Td, tetrad cell; Msp, microspore; Mp, mature pollen. Scale bars = 50 µm for (A)–(M).

Figure 3

Pollen fertility of 27 genotypes determined by three genes S24, S35, and EFS. Pollen fertility (right bar chart) for each genotype class (left panel) is expressed by the mean (%) ± SD (n = 3–10). Effective genes for pollen sterility are marked with red frames on the genotype panel. White and black bars in the bar chart represent the sterile and fertile pollen phenotypes, respectively. Micrographs of the pollen grains for each fertility class are shown in the bottom panel. Plants of each genotype were selected from self-pollinated progeny of NIL-S24+S35+Efs by genotyping using the linked markers mS2 for S24, 1c305 for S35, and 2c2015 for EFS. I: indica (IR24) homozygote. J, japonica (Asominori) homozygote; H, heterozygote; F, fertile; PS, partial-sterile; SS, semisterile; S, sterile.

Figure 4

Identification and localization of the S35-activator. (A) Graphical genotype of the near-isogenic line S24ILI. S24ILI was homozygous for S24 and carried a very small IR24 segment of less than 181-kb (mS1–mS3, see also Figure 4B) within a uniform Asominori background. (B) Pollen phenotype of the most informative recombinants and the NILs S24ILI/H. These lines carried Asominori homozygous alleles for both the EFS and S35 loci, representing the pollen sterility dependent exclusively on S24. Bars show the mean (%) with SD (n = 4–9). (C) Pollen fertility of a set of substitution lines covering the S24 region on chromosome 5. All the lines were heterozygous for S35. The white arrows in the photo indicate degenerated pollen grains. (D) Chromosome position of the S35-activator INK locus on chromosome 5. The diagram shows the recombination breakpoints of the obtained plants/lines. All plants/lines carried heterozygous alleles for S35 to evaluate the effect of the recombinant segment on the S35 phenotype. The pollen phenotype of each line was determined using selfed progeny (BC₄F₃) of the recombinant individuals (BC₄F₂) excluding 5R18. 5R18 was a single BC₅F₁ plant derived from a backcross of the recombinant BC₄F₂ plant carrying homozygous S35-i alleles. The 5R02 line had the Efs-i homozygous genotype that prevented the influence of heterozygous S24 on pollen fertility. The mean (%) ± SD (n = 3–10) are shown on the right.

Figure 5

Effect of the genetic background on pollen sterility. (A), (B) Frequencies of the S35 and S24 genotypes in the selfed progeny of the NILs with reciprocal genetic backgrounds. The stacked bar chart represents the frequency of the S35 genotype (A) and the S24 genotype (B) in the selfed progeny of the heterozygotes. Population size and chi-square values for the deviation from 1:2:1 are shown on the top. The numbers of plants in each genotype (black, IR24 homozygote; gray, heterozygote; white, Asominori homozygote) are shown inside the bars. The parental genotype is shown at the bottom. DNA markers 1c350 or 1c305 (for S35), and mS2 (for S24) were used for genotyping (the red line on chromosomes). The plant line ID is identical to that shown in Figure 3, Figure 4, and Figure 6. Parental plants #4, #6, #15, and 5R21 had sterile pollen. All other plants had fertile pollen. ** P < 0.01; *** P < 0.001; ns, not significant. (C) Graphical genotype (left) and pollen fertility (right) of the NILs that carried heterozygous segments harboring the S24 or S35 locus in either the Asominori or the IR24 genetic backgrounds. The NILs carried the appropriate genotype for the interacting partner gene (efs-j/efs-j for the S24 activation, or Ink-i/Ink-i for the S35 activation). Bars show the mean (%) with SD (n = 5).

Figure 6

Diagram showing the genetic interactions between hybrid male sterility loci. (A) Schematic representation of the chromosome position and genetic interactions of hybrid male sterility genes. Arrows and bars represent positive and negative regulation, respectively. Indica has the killer alleles for the S24 and S35 loci (S24-i and S35-i), and these genes kill the male gamete bearing the japonica alleles in heterozygous hybrid progeny. Activation of these pollen killer genes is dependent on the sporophytic factors EFS for S24 (B) and INK for S35 (C). Additional unknown suppressors/modifiers for S24 and S35 exist within the indica genome.