Identification of qPSR7-2 as a Novel Cold Tolerance-Related QTL in Rice Seedlings on the Basis of a GWAS

Abstract

Rice is the primary source of food for more than half of the global population. Accordingly, improving the cold tolerance of rice is vital for ensuring food security. In this study, a new cold tolerance-related QTL in rice (qPSR7-2) was detected on chromosome 7 following a genome-wide association study involving 173 japonica rice germplasm resources. The fine mapping of this locus identified Os07g0541800 as a candidate gene associated with qPSR7-2. This gene encodes a cysteine-rich receptor-like kinase. The functional verification of Os07g0541800 involving transgenic plants indicated that qPSR7-2 positively regulates rice cold tolerance at the seedling stage. The examination of the cold tolerance of 984 germplasm resources from the 3000 Rice Genomes Project at the seedling stage and their respective haplotypes at qPSR7-2 revealed that the proportion of favorable haplotypes in germplasm resources increased as the latitude increased. More than 90% of the rice varieties cultivated in Europe and Japan appear to carry qPSR7-2, implying that qPSR7-2 may mediate the acclimation of rice to low-temperature stress. The findings of this study will further clarify the molecular networks regulating rice cold tolerance, while also providing researchers and breeders with new genetic resources and information relevant for developing cold-tolerant rice varieties.

1. Introduction

Rice (Oryza sativa L.) is one of the most important food crops worldwide, serving as a staple food for half of the global population. Because rice originated in tropical and subtropical regions, it is much more sensitive to low temperatures than wheat (Triticum aestivum L.) and barley (Hordeum vulgare L.). More specifically, rice plants are damaged by ambient temperatures below 15 °C [1]. To satisfy the increasing demand for food, rice-growing regions have spread from tropical and subtropical areas to high-altitude and high-latitude areas. However, there have recently been frequent instances of extreme weather conditions, with cold stress damaging crops in nearly 24 countries worldwide, including in East Asia (e.g., China, Japan, and North Korea). Cold-induced damages have decreased the annual global rice yield by approximately 3–5 million tons, thereby threatening the stable production of rice worldwide [2,3]. Therefore, there is a critical need for identifying cold tolerance genes and elucidating the associated regulatory molecular mechanisms relevant for breeding cold-tolerant rice varieties and ensuring food security.

The cold tolerance of rice is a complex genetic trait controlled by multiple genes. Over the past 20 years, nearly 300 quantitative trait loci (QTLs) related to low-temperature stress responses have been identified on 12 rice chromosomes following analyses of genetic mapping populations and natural populations [4], but only 15 cold tolerance genes have been cloned (i.e., qLTG3-1, COLD1, qCTS-9, GSTZ2, LTG1, Ctb1, CTB4a, qPSR10, OsWRKY53, OsWRKY63, HAN1, OsLTPL159, OsbZIP54, bZIP73, and OsSAPK6) [1,5,6,7,8]. These genes encode diverse proteins, including membrane proteins, kinases, and transcription factors with molecular functions. For example, the membrane protein Cold1 and kinase-like transcription factors are important for establishing kinase cascade signaling pathways and activating defense gene expression [9]. In rice, OsWRKY transcription factors generally negatively regulate cold tolerance. Earlier research confirmed that OsWRKY53 controls rice cold tolerance at the booting stage by negatively regulating the gibberellin content in anthers [8], whereas OsWRKY63–OsWRKY76–OsDREB1B negatively regulates rice cold tolerance via a transcriptional regulatory chain reaction [1]. In addition, the cell wall composition can influence the cold tolerance of rice plants by modulating the formation and transmission of cold stress-related signaling molecules. Moreover, OsLTPL159 may limit the toxic effects of reactive oxygen species and enhance cellulose deposition in the cell wall, while also promoting the accumulation of osmoregulatory substances, maintaining chloroplast integrity, and enhancing the cold tolerance of young rice seedlings [6]. Furthermore, COG2 can inhibit SNP1A expression under low-temperature stress conditions, increase the pectin content, maintain cell elongation and differentiation, and protect against the detrimental effects of low-temperature stress [10]. Therefore, the molecular mechanism underlying rice cold tolerance involves various regulatory pathways that form an intricate network.

Cysteine-rich receptor-like kinases (CRKs), which are also called domain of unknown function 26 (DUF26) receptor-like kinases, belong to a receptor-like kinase subfamily. The rice genome includes multiple copies of genes encoding CRKs. The 49 CRK-encoding genes that have been identified are distributed on chromosomes 1, 5, 7, 10, 11, and 12, but 30 of these genes are present in clusters in the same region of chromosome 7 [11]. Previous studies indicated that ALS1 is a CRK that negatively regulates rice blast resistance [12], whereas OsCRK6 and OsCRK10 positively regulate rice resistance to blight [13]. Unfortunately, there has been limited research on whether CRK-encoding genes affect rice cold tolerance. In this study, qPSR7-2 (containing the Os07g0541800 gene) on chromosome 7 was identified as a new rice cold tolerance-related QTL on the basis of a genome-wide association study (GWAS). The functional characterization of Os07g0541800 revealed that the encoded CRK family protein positively regulates the cold tolerance of rice plants. The findings of this study will provide researchers with new cold tolerance-related genetic resources as well as the theoretical basis for breeding and cultivating novel rice varieties able to withstand low-temperature stress.

2. Materials and Methods

2.1. Materials and Evaluation of Cold Tolerance

The 173 conventional japonica rice varieties used in this study were collected from various geographical locations and sorted (Table S1 and Figure S1A). The 3000 Rice Genomes Project (3KRGP) germplasm resources were provided by the International Rice Research Institute. Additionally, location, sequencing, and SNP data (average sequencing depth of 15×) were obtained from 3KRGP (https://snp-seek.irri.org (accessed on 7 April 2023)) [14]. The Os07g0541800 gene knockout rice material was purchased from Hangzhou Baige Biotechnology Co., Ltd. CRISPR/Cas9 constructs were generated using a CRISPR/Cas9 kit (Biogle) containing CRISPR plasmid BGK032 in the japonica rice (Oryza sativa) “Zhonghua 11” (ZH11) background. The genome sequencing primers, sgRNA sequence, and mutation type were shown in Tables S1 and S3. The Os07g0541800-overexpressing material was provided by Wuhan Biorun Biotechnology Co., Ltd. using overexpression plasmid pBWA(V)HS in ZH11 background. Details regarding the primers used for plasmid construction in this study are provided in Table S2. The conventional rice materials, transgenic lines, and gene knockout lines mentioned above were planted in Yangzhou, Jiangsu, and Sanya, Hainan, China. Phenotypic analysis was conducted on the transgenic materials and gene knockout materials using the F₃ generation materials.

The cold tolerance of the study materials was assessed. Briefly, rice seeds were germinated under normal conditions, and the resulting seedlings were allowed to grow to the four-leaf stage, after which the total number of seedlings was recorded. The seedlings were then placed in an incubator set at 4 °C with an 8 h dark:16 h light (light intensity of 4000 lx) cycle for a 48 h cold treatment. Then, plants were allowed to recover under normal conditions for 7 days at 28 °C with an 8 h dark:16 h light (light intensity of 4000 lx) cycle. The rice plant survival rate (PSR) was calculated using the following formula: PSR (%) = (number of viable rice seedlings after the cold treatment/number of normal rice seedlings before the cold treatment) × 100. The cold tolerance of each material was evaluated using three biological replicates. More than 20 rice seedlings were screened in each biological replicate per variety. The 24-hole seedling-raising plate was used for rice growth in this study.

2.2. Whole-Genome Sequencing Data Analysis

After the aforementioned 173 conventional japonica rice materials and sequencing data were collected and sorted [15], genomic DNA was extracted from the seedlings at the three-leaf stage according to the CTAB method. The genomic libraries were sequenced using the Illumina X-Ten sequencer (sequencing depth of 15.0× to 30.9×). The quality of the raw reads was first assessed using FastQC (v.0.10.1) and then the reads were filtered using Trimmomatic (v.0.36) to eliminate adapter sequences and low-quality sequences. The genome sequence of 173 varieties is available under NCBI with the BioProject ID: PRJNA756648. The clean reads were aligned to the Nipponbare reference genome sequence (IRGSP-1.0) using BWA-MEM (v.0.7.13). The sequence alignments were processed using SAMtools (v.1.3.1). The GATK-HaplotypeCaller module (v.4.4) was used to detect single nucleotide polymorphisms (SNPs) and insertions/deletions. The high-quality sequence variants were screened and retained according to the following criteria: QD > 2.0, FS < 60.0, and MQ > 20.0. In contrast, the following criteria were used to detect low-quality sequence variants, which were deleted: (1) missing rate > 80%; (2) heterozygous genotype frequency > 5% or more than double the frequency of the minor homozygous alleles; (3) deviation from the Hardy–Weinberg equilibrium proposed by GATK (excess heterozygosity < 1 × 10⁻⁵).

2.3. Analysis of the Population Structure

The VCF file obtained from method 2.2 was converted into the ped/map format for population structure analysis using Plink software (v.5.0). Population structure analysis was performed using Admixture software (v1.3) to calculate its CV-error and obtain its optimal K-value when the CV-error is smallest. The R software was used to plot images. The Plink software (v.5.0) was used for the principal component analysis, and the GraphPad Prism (v.8.0) was used to visualize. Tassel (v.5.2.82) was used to analyze the evolutionary relationships among the sequenced varieties, whereas iTOL (v.6) (https://itol.embl.de (accessed on 7 April 2023)) was used to visualize phylogenetic trees.

2.4. Genome-Wide Association Study

The mixed linear model of Tassel (v.5.2.82) was used for the GWAS. The criteria for screening SNP markers were as follows: MAF > 0.05 and missing rate < 20%. The GWAS threshold p-value was determined according to the Bonferroni correction method. The threshold p-value was calculated as 0.01/n, where n is the number of SNPs used in the GWAS. Manhattan plots and QQ plots were prepared using CMplot (v.1.0.1) from the R package. Linkage disequilibrium (LD, calculated as r²) in the study was also calculated using Tassel (v.5.2.82). To identify the genomic interval containing candidate genes related to cold tolerance, the region closely linked with an SNP (r² ≥ 0.6) was analyzed. Genes located directly in or within 50 kb (genome-wide average distance of LD decay to r² = 0.2) around the confidence intervals were selected as candidate genes for the GWAS loci.

2.5. Analysis of Candidate Gene Expression Patterns

A quantitative real-time polymerase chain reaction (qRT-PCR) analysis was performed to examine the changes in the expression of the candidate genes due to the cold treatment. Specifically, 200 mg of fresh rice leaves was collected at 0 h (i.e., before the cold treatment) and at 48 h (i.e., after the cold treatment). Total RNA was extracted from the leaf samples using the RNAprep Pure Plant Kit (Tiangen, Beijing, China, DP441) and then reverse transcribed to cDNA using the HiScript II 1st Strand cDNA Synthesis Kit (Vazyme, Nanjing, China, R211). The qRT-PCR analysis was performed using the Bio-Rad CFX96 Real-time PCR system, with OsActin (Os04g0177600) selected as the internal reference gene. The candidate gene expression analysis was completed using three biological replicates. Details regarding the primers used in real-time polymerase chain reactions are provided in Table S2.

2.6. Data Analysis

The significance of intergroup differences was investigated by one-way ANOVA and Tukey’s honestly significant difference (HSD) mean-separation test using IBM SPSS Statistics 26.0 software. The significance of difference between the two groups of data was assessed by Student’s t test using Microsoft Office Excel software.

3. Results

3.1. Genome-Wide Association Study of Cold Tolerance QTLs

A total of 1,487,651 SNP markers distributed throughout the genome were obtained after the 173 japonica rice materials were re-sequenced and reads were aligned to the Nipponbare reference genome (MAF > 0.05). The analysis of the population structure showed that the tested germplasm resources could be divided into five subgroups (CV error value was lowest when K = 5) (Figure 1A and Figure S1B). In addition, the principal component analysis results were basically consistent with the population structure (Figure 1B). The diversity in PSR among the 173 materials that underwent the 48 h cold treatment at 4 °C at the four-leaf stage revealed significant differences in cold tolerance. Specifically, the PSR was greater than 50% for 158 materials, whereas it was less than 30% for eight materials (Jin 9540, Handao 442, Zhonghan 502, Liaojing 21, R109, Tianjin 28-1, Wujing 15, and Jiahe 218) (Figure 1C). On the basis of the genotypic and phenotypic data, a major cold tolerance-related QTL [qPSR7-2; −log₁₀(P) = 9.6, R² = 0.3257] was identified on chromosome 7 following the GWAS, which was performed using −log₁₀(P) > 6 as the threshold (Figure 1D). In addition, other genes related to cold tolerance such as DREB1F and qLTG3-1 have also been identified in the GWAS.

3.2. Fine Mapping and Functional Verification of the Cold Tolerance QTL qPSR7-2

According to the GWAS results, the qPSR7-2 candidate interval was narrowed to 21.36–21.42 Mb, which is within the linkage disequilibrium region (Figure 2A). Using the Bonferroni-corrected significance threshold [−log₁₀(P) > 6], we identified 36 significantly correlated SNP sites (Figure 2B). The specific locations of these significant SNP sites are indicated in Figure 2C. There were six candidate genes in the candidate region which were all functionally annotated and identified as CRK-encoding genes. These genes included Os07g0541400 (CRK6) and Os07g0541500 (CRK10), which are closely associated with disease resistance [13]. The changes in the expression levels of these six candidate genes in the cold-tolerant variety Nipponbare exposed to cold stress were subsequently analyzed. The results showed that the expression levels of five of the six candidate genes were down-regulated in response to the cold treatment. The exception was Os07g0541800, which had a significantly up-regulated expression level (Figure 2D). In the cold-sensitive indica variety R6547, the expression levels of several candidate genes, including Os07g0541800, were down-regulated (Figure 2D). Therefore, Os07g0541800 was preliminarily identified as a candidate gene for qPSR7-2.

To explore whether Os07g0541800 influences rice cold tolerance, Zhonghua 11 lines in which Os07g0541800 was knocked out or overexpressed were generated. The lines were then grown and incubated at 4 °C for 48 h to evaluate their cold tolerance. Compared with the PSR of the wild-type control (71.6%), the PSR was significantly lower for the two knockout lines (37.2%) (Figure 2E) but significantly higher for the three overexpressing lines (89.2%) (Figure 2F). These results suggest Os07g0541800 is the functional gene of qPSR7-2 that positively regulates the cold tolerance of rice.

The genotyping of Os07g0541800 was conducted using the significantly correlated SNP loci [−log₁₀(P) > 6]. The analysis detected two significant SNP loci in exon 7 of Os07g0541800 that resulted in amino acid substitutions in the 173 varieties; these two loci were interlinked (Figure 2G). The genotyping results revealed two main haplotypes (Hap1 and Hap2), with 92% of the 173 varieties carrying Hap1. Moreover, the PSR of the varieties with Hap1 was significantly higher than that of the varieties with Hap2 (Figure 2G). Therefore, Hap1 was identified as the favorable genotype in the 173 varieties.

3.3. qPSR7-2 Is Mainly Distributed in Rice Materials from High-Latitude Areas

The 3KRGP database was used to analyze the qPSR7-2 haplotype in global rice germplasm resources. There were four main SNP loci in Os07g0541800 exon regions 1 and 7, all of which resulted in amino acid substitutions. Additionally, they were divided into five haplotypes. Among these haplotypes, Hap1, Hap2, and Hap5 were more common than Hap3 and Hap4 in the examined germplasm (Figure 3A). The cold tolerance of the 984 3KRGP rice germplasm resources was assessed. The PSR was significantly higher for the germplasm carrying Hap1 and Hap2 than for the germplasm carrying Hap3, Hap4, and Hap5, implying that Hap1 and Hap2 are favorable genotypes (Figure 3C). The analysis of the distribution of the different haplotypes in indica and japonica subpopulations showed that Hap3, Hap4, and Hap5 were mainly distributed in the indica subpopulation, whereas the distribution of Hap1 and Hap2 was relatively complex. The main japonica type carrying Hap1 was temperate rice, but the main japonica type carrying Hap2 was tropical rice (Figure 3B). In addition, the proportion of germplasm resources carrying Hap1 and Hap2 increased as the latitude increased. Specifically, the proportions of Hap1 and Hap2 were as high as 90% in Europe and Japan. This suggests that Hap1 and Hap2 were selected during the domestication of cold-tolerant rice varieties in high-latitude regions (Figure 3D).

4. Discussion

Rice is a staple food for half of the global population. Improving the cold tolerance of rice, especially in high-latitude and high-altitude regions, is critical for ensuring stable rice production. Rice cold tolerance is an extremely complex trait mediated by multiple genes. To date, only a few cold tolerance genes have been identified and characterized, but the related molecular mechanisms and signaling network pathways have not been thoroughly studied. The cold tolerance of plants reportedly depends on the production and transduction of cold-related signaling molecules and appropriate changes to the expression of specific genes [16]. For example, cold stress signals (Ca²⁺, reactive oxygen species, alternative splicing, and hormonal changes) can be transduced from organelles to the nucleus to modulate stress response-related gene expression [17]. Therefore, as the upstream regulatory factors of cold tolerance signaling pathways, cold stress signal receptors on cell membranes will need to be more comprehensively investigated to further clarify the mechanism underlying cold tolerance [18]. Receptor kinases, which are one of the many stress signal receptors, are important for the formation and transmission of cold stress signals necessary for plant responses to low-temperature stress. However, there are relatively few reports describing cold tolerance signaling pathways involving kinases. The CTB4a gene, which is associated with cold tolerance at the booting stage, was cloned and revealed to encode a chloroplast-localized, leucine-rich repeat receptor kinase. Earlier research demonstrated that CTB4a interacts with AtpB to enhance the cold tolerance of rice by mediating the increase in the ATP content under cold stress conditions [19]. Unfortunately, how other types of receptor kinases regulate cold tolerance remains unknown. In this study, qPSR7-2 was identified as a new QTL related to the cold tolerance of rice seedlings. It was detected in a japonica population on the basis of a GWAS. One of the candidate genes in this QTL encodes a CRK. The preliminary functional annotation of this gene suggests the encoded CRK positively regulates rice cold tolerance, but how this gene responds to cell–cell or environmental signals and the related ligands in the CRK signaling pathway will need to be examined in future studies.

Interestingly, we detected two significant SNP loci (position 21408137 and position 21408170) in exon 7 of Os07g0541800 that resulted in amino acid substitutions in the 173 varieties. The genotyping results revealed two main haplotypes (Hap1 and Hap2) according to the two significant SNPs (Figure 2G), whereas in the 3KRGP germplasm resources, it could be divided into five haplotypes according to four SNP loci with two additional SNPs (position 21408233 and position 21408969) compared with 173 varieties. The 173 varieties were composed of cultivated japonica rice from China and Japan (Figure S1, Table S1), whereas the 3KRGP rice germplasm resources were composed of wild rice and cultivated rice including indica and japonica rice from all over the world with higher genetic diversity. Therefore, we speculated that the two additional SNPs (position 21408233 and position 21408969) might have been selected and mostly fixed in the genome during the domestication of cultivated japonica varieties in China and Japan.

Exploiting cold tolerance genes is an affordable and effective way to improve crop varieties susceptible to low-temperature stress. The breeding of crops with increased tolerance to cold conditions should ideally not affect crop yield, quality, disease resistance, and other traits. Therefore, the synergistic enhancement of multiple traits is one of the objectives of breeding programs. However, diverse traits are usually controlled by different gene regulatory networks and there are interactions between genes. Cold tolerance is associated with the yield, growth period, and plant type to some extent. For example, a mutation to the semi-dominant gene LGS1, which controls the grain length, grain width, and 1000-grain weight, may disrupt the OsGRF4–miR396 stress response network. Thus, the rice grain length, width, and weight may increase by enhancing the cold tolerance of rice at the seedling stage [20]. The expression of the rice growth-related gene Ghd8 can delay the onset of the heading stage, while also increasing cold tolerance [21]. A previous study showed OsSAPK6 phosphorylates the plant-type regulatory protein IPA1, resulting in the accumulation of IPA1 and the activation of OsCBF3 expression, thereby enhancing rice cold stress resistance [22]. Hence, when selecting genes to improve rice cold tolerance, their effects on other traits should be considered. Meanwhile, the cold tolerance of rice is a complex genetic trait controlled by multiple genes, and cold tolerance-related genes such as qLTG3-1 and COLD1 have been cloned, so in the domestication process of cold-tolerant varieties, we should consider the enhancement effect of multiple gene aggregation on the cold stress tolerance ability of rice. In addition to studying the molecular mechanism underlying the cold tolerance mediated by qPSR7-2, the effects of this QTL on the yield and other traits should be investigated. Doing so will elucidate the molecular network regulating rice cold tolerance, while also providing the theoretical basis for the synergistic improvement of stress resistance and disease resistance to further increase the rice grain yield.

5. Conclusions

In this study, a GWAS was performed and a new cold tolerance-related QTL in rice (qPSR7-2) was identified on chromosome 7. The fine mapping and identification of candidate genes indicated that Os07g0541800 was the most likely candidate gene involved in the cold tolerance phenotype related to qPSR7-2. This gene encodes a CRK. Moreover, the functional verification involving transgenic plants showed that qPSR7-2 positively regulates cold tolerance at the rice seedling stage. The evaluation of the cold tolerance of 984 3KRGP germplasm resources at the seedling stage revealed that the proportion of favorable haplotypes at qPSR7-2 in rice germplasm increases as the latitude increases, with more than 90% of the rice accessions in Europe and Japan carrying these haplotypes. This implies that favorable haplotypes at qPSR7-2 were selected during the spread of rice cultivation to higher latitudes contributing to the low-temperature tolerance of rice.