Zilong Guo, Shouchuang Wang, Feng Zhang, Denghao Xiang, Jun Yang, Dong Li, Baowei Bai, Mingqiu Dai, Jie Luo, Lizhong Xiong 

Stress Biol.; 2024 Jan 23; 4(1):6. doi: 10.1007/s44154-024-00150-4.

Abstract

Plants orchestrate drought responses at metabolic level but the genetic basis remains elusive in rice. In this study, 233 drought-responsive metabolites (DRMs) were quantified in a large rice population comprised of 510 diverse accessions at the reproductive stage. Large metabolic variations in drought responses were detected, and little correlation of metabolic levels between drought and normal conditions were observed. Interestingly, most of these DRMs could predict drought resistance in high accuracy. Genome-wide association study revealed 2522 significant association signals for 233 DRMs, and 98% (2471/2522) of the signals were co-localized with the association loci for drought-related phenotypic traits in the same population or the linkage-mapped QTLs for drought resistance in other populations. Totally, 10 candidate genes were efficiently identified for nine DRMs, seven of which harbored cis-eQTLs under drought condition. Based on comparative GWAS of common DRMs in rice and maize, representing irrigated and upland crops, we have identified three pairs of homologous genes associated with three DRMs between the two crops. Among the homologous genes, a transferase gene responsible for metabolic variation of N-feruloylputrescine was confirmed to confer enhanced drought resistance in rice. Our study provides not only genetic architecture of metabolic responses to drought stress in rice but also metabolic data resources to reveal the common and specific metabolite-mediated drought responses in different crops.

See https://pubmed.ncbi.nlm.nih.gov/38253937/

Figure 3. Integration of dmGWAS, dpGWAS, and drQTLs. A Circos plot showing dmGWAS loci, co-localized drQTLs from previously reported studies, and co-localized dpGWAS loci of PAR_R and GPAR_R. The five tracks (from inner to outer) represent the number of significant SNPs for each association signal, significance (−log₁₀ P) of dmGWAS signals, co-localized drQTLs from previously reported studies (partially shown), the number of DRMs associated with each association signal (shown by heat map), and co-localized dpGWAS loci of PAR_R and GPAR_R (partially shown), respectively. B Association network of dmGWAS and dpGWAS. In the network, each node represents an annotated DRM or a phenotypic trait of evaluating DR capacity (measured by optical phenotyping facility and manual measurements in our previous study); each edge represents that a DRM and a phenotypic trait are connected by a co-localized locus. Different colors represent different metabolic groups. Each purple circle in the center of the network represents a specific phenotypic trait.