Genome-wide association study uncovers a novel gene responsible for rice seedling submergence tolerance

Abstract

Submergence tolerance is crucial for the direct seeding of rice, yet long-term domestication and breeding have inadvertently reduced the adaptability of cultivated rice to submergence stress. Here, we identify a nucleic acid excision repair protein-encoding gene qSHS5 as an essential regulator of seedling height under submergence through a genome-wide association study in 322 rice accessions. Disruption of qSHS5 in mutants resulted in seedling growth inhibition under submergence, while growth remains comparable to wild-type under normal conditions. This inhibition is primarily due to decreased cell number resulting from G1 phase cell cycle arrest. Further investigation showed that levels of reactive oxygen species (ROS), O₂ ^- and H₂O₂ significantly increased, and DNA damage was aggravated in qshs5 mutants under submergence. Additionally, we find the submergence-tolerant haplotype qSHS5^H4 has been progressively lost, while the elite haplotype qSHS5^H3 has been largely overlooked during the breeding of semi-dwarf and high-yield in rice. Importantly, we demonstrate that combining qSHS5^H3 with the semi-dwarfing haplotype SD1^H1 exhibited high yield without compromising submergence tolerance, offering significant potential for future breeding programmes targeting direct seeding cultivation. This study not only identifies a novel superior allele but also provides valuable insights for future improvement of rice cultivation, particularly under climate change-induced submergence for direct seeding.

Keywords: ROS; breeding improvement; natural variation; qSHS5; rice; submergence tolerance.

Figure 1

Identification of qSHS5 as the target candidate gene for the seedling height under submergence by GWAS in rice. (a) Manhattan plot and (b) quantile–quantile (Q‐Q) plot of GWAS for seedling height under submergence in the whole population with a threshold of −log₁₀ (P) = 8.29 (0.01 significance level). Red dashed lines indicate the close‐up of peak region shown in c. (c) Local Manhattan plot surrounding the peak site (6.00–6.25 Mb). The green, yellow, and blue dots respectively represent the GWAS results for the whole population, the indica subgroup, and the japonica subgroup. (d) Linkage disequilibrium heatmap for SNPs within a 250 kb surrounding peak site detected by GWAS. LD block is marked by a red triangle. The colour key (white to red) represents linkage disequilibrium values (R ²). The leading SNP located in the intron of candidate gene LOC_Os05g10980. (e, f) Violin plot of the seedling height under submergence from haplotype‐based association analysis of candidate genes with phenotypic variation. Different letters above the column indicate significant differences at the P < 0.05 level according to an analysis of variance (ANOVA) test. (h) Anaerobic expression of the three candidate genes in Nipponbare coleoptiles under submergence. (i) The expression of LOC_Os05g10980 during seed germination under submergence. Data are presented as mean ± SD, n = 3.

Figure 2

Characterization of the regulation of the seedling height by qSHS5 under submergence. (a) CRISPR target sites in qSHS5. The red character indicates the successful CRISPR‐Cas9 edited sites. (b) Representative images of the seed vigour of the WT and qshs5 mutants after 1 day of imbibition under normal conditions. Scale bar = 2 cm. (c) Germination rate of WT and qshs5 mutants under normal conditions. Data are presented as mean ± SD, n = 5. (d) Growth performance of WT and qshs5 mutants after 15 days of sowing under normal conditions. (e, i) Seedling performance of WT and qshs5 mutants after 9 days of sowing under normal (e) and submergence (i) conditions. Scale bar = 10 cm. (f, j) Comparison of seedling rate of WT and qshs5 mutants under normal (f) and submergence (j) conditions. Data are presented as mean ± SD, n = 3. (g, k) Dynamic changes in seedling height of WT and qshs5 mutants under normal (g) and submergence (k) conditions. Asterisks indicate significant differences between WT and mutants, determined by Student's t‐test: **, P < 0.01. (h) Submergence tolerance performance of WT and qshs5 mutants after 15 days of sowing under submergence. (l) qSHS5 expression levels in various tissues of WT, including root, sheath, and leaf at the seedling stage, as well as root, stem, leaf, sheath, and panicle at the spikelet stage. Data are presented as mean ± SD, n = 3. (m) Subcellular localization of the qSHS5‐GFP fusion protein in rice protoplasts.

Figure 3

Histological characterization and cell cycle analysis of qshs5 mutants. (a, d) Photographs of WT, qshs5‐Cr1, and qshs5‐Cr2 seedlings after 4 days of sowing under normal (a) and submergence (d) conditions. The white dashed line indicates the area on the leaf sheath selected for scanning electron microscopy (SEM) to observe cell morphology. Scale bar = 1 cm. (b, e) SEM images showing the inner epidermal cells of the leaf sheaths of WT, qshs5‐Cr1, and qshs5‐Cr2 under normal (b) and submergence (e) conditions. Scale bar = 100 μm. (c, f) Quantitative measurements of single‐cell length, cell width, and total cell number in the seedlings as shown in (a, d). (g, h) Nuclear DNA content of WT and qshs5 mutants shoot apex under normal (g) and submergence (h) conditions. (i, j) Quantification of the DNA profiles of WT and qshs5 mutants under normal (i) and submergence (j) conditions.

Figure 4

Determination of ROS contents and DNA damage status in WT and qshs5 mutants under normal and submergence conditions. (a, e) NBT staining of WT and qshs5 mutants under normal (a) and submergence (e) conditions. (b, f) DAB staining of WT and qshs5 mutants under normal (b) and submergence (f) conditions. (c, d, g, h) H₂DCFDA staining analyses of WT and qshs5 mutants under normal (c, d) and submergence (g, h) conditions. Fluorescence intensity was analysed using ImageJ. (i) Statistical analysis of H₂O₂, O₂ ⁻, APX activity, POD activity, SOD activity, and CAT activity under normal and submergence (blue background) conditions. Data are presented as mean ± SD, n = 3. Asterisks indicate significant differences between WT and qshs5 mutants according to Student's t‐test: **, P < 0.01; *, P < 0.05. (j) Four levels of DNA damage: 1%, 10%, 30%, and 50%. (k) Frequency distribution of DNA damage levels in WT and qshs5 mutants under normal and submergence (blue background) conditions; n ≥ 20.

Figure 5

Loss of the elite submergence‐tolerant haplotype of qSHS5 during modern japonica breeding. (a) Violin plots comparing the seedling height and relative seedling height under submergence across three qSHS5 haplotypes identified in the CDS region. (b) Schematic representation of four major haplotypes in promoter region (the 2 kb upstream) of qSHS5. Blue squares represent coding regions of the gene. ‘Num. of Sub‐I’ refers to the number of accessions in the indica, and ‘Num. of Sub‐J’ refers to the number of accessions in the japonica. (c) Expression analysis of qSHS5 under normal and submergence (blue background) conditions in randomly selected accessions representing different haplotypes (qSHS5 ^H1‐H4 ). Data are presented as mean ± SD, n = 3. (d, e) Violin plots comparing phenotypic traits among different qSHS5 haplotypes, showing plant height (d) and yield (e) at maturity. (f) Allele frequencies of qSHS5 haplotypes across four time periods in whole population, japonica, and indica. Varieties are categorized into four groups based on release dates: Native, <1971 (before 1970), 1971–1990 (inclusive), and > 1990 (post‐1990, 1991–2016). (g) Nucleotide diversity analysis and population differentiation analysis between different time groups of qSHS5. The y‐axis on top spectrum represents 200 kb window‐based π and Fst values.

Figure 6

Co‐selection of elite alleles of qSHS5 and SD1 in rice breeding. (a) Allele frequencies of SD1 haplotypes in the japonica across four different time periods. (b–d) Violin plots comparing phenotypic traits among different SD1 haplotypes, including the relative seedling height (b) under submergence, plant height at maturity (c), and yield at maturity (d). (e) Joint haplotype analysis of SD1 and qSHS5, indicating the distribution of different qSHS5 haplotypes within the three main SD1 haplotypes. (f) Allele frequencies of qSHS5 haplotypes under the fixed SD1 ^H2 background, shown across four time periods categorized by release dates. (g–i) Violin plots comparing phenotypic differences among the four major qSHS5 haplotype combinations with SD1 ^H1 or SD1 ^H2 , including plant height at maturity (g), yield at maturity (h), and RSH under submergence (i).

Figure 7

Coordinated artificial selection of qSHS5 and SD1 for enhanced submergence tolerance in japonica. (a) Under submergence stress, qSHS5 is significantly induced in wild‐type plants, where it plays a critical role in repairing DNA damage caused by the stress, thereby inhibiting excessive ROS accumulation. In contrast, qshs5 mutants fail to promptly repair damaged DNA, leading to ROS accumulation and further exacerbation of DNA damage. This cascade disrupts cell cycle progression and inhibits cell elongation, ultimately creating a detrimental feedback loop. Therefore, qSHS5 is essential for promoting seedling growth under submergence by facilitating efficient DNA repair and regulating ROS and DNA damage response pathways. (b) Native japonica varieties harbour diverse haplotype combinations of qSHS5 and SD1 ^H2 , forming a stable genetic unit that supports robust seedling establishment adaptation to submergence. However, in japonica, the tall stature and low yield associated with submergence tolerance were unfavourable traits, as they did not meet the requirements for lodging resistance and high productivity. As a result, artificial selection during japonica improvement favoured the semi‐dwarf and high‐yield haplotypes qSHS5 ^H1 and SD1 ^H1 , leading to the gradual loss of qSHS5 ^H4 . Without the protective function of qSHS5 ^H4 , the haplotypes qSHS5 ^H3 and SD1 ^H1 have been selected to enhance submergence tolerance. Future breeding strategies for japonica should focus on integrating qSHS5 ^H3 and SD1 ^H1 to simultaneously enhance submergence tolerance, maintain high yield, and reduce plant height. This combination would provide a competitive advantage for cultivated rice by balancing submergence adaptability with improved agronomic traits, ensuring both natural selection and sustainable production. Each haplotype is represented by a different coloured circle: orange circle (qSHS5 ^H1 ), deep blue circle (qSHS5 ^H3 ), green circle (qSHS5 ^H4 ), pink diamond (SD1 ^H1 ), and blue diamond (SD1 ^H2 ).