Insights into the Genomic Regions and Candidate Genes of Senescence-Related Traits in Upland Cotton via GWAS

Abstract

Senescence is the last stage of plant development and is controlled by both internal and external factors. Premature senescence significantly affects the yield and quality of cotton. However, the genetic architecture underlying cotton senescence remains unclear. In this study, genome-wide association studies (GWAS) were performed based on 3,015,002 high-quality SNP markers from the resequencing data of 355 upland cotton accessions to detect genomic regions for cotton senescence. A total of 977 candidate genes within 55 senescence-related genomic regions (SGRs), SGR1-SGR55, were predicted. Gene ontology (GO) analysis of candidate genes revealed that a set of biological processes was enriched, such as salt stress, ethylene processes, and leaf senescence. Furthermore, in the leaf senescence GO term, one candidate gene was focused on: Gohir.A12G270900 (GhMKK9), located in SGR36, which encodes a protein of the MAP kinase kinase family. Quantitative real-time PCR (qRT-PCR) analysis showed that GhMKK9 was up-regulated in old cotton leaves. Overexpression of GhMKK9 in Arabidopsis accelerated natural leaf senescence. Virus-induced gene silencing (VIGS) of GhMKK9 in cotton increased drought tolerance. These results suggest that GhMKK9 is a positive regulator and might be involved in drought-induced senescence in cotton. The results provide new insights into the genetic basis of cotton senescence and will be useful for improving cotton breeding in the future.

Keywords: GWAS; GhMKK9; candidate gene; genomic region; senescence; upland cotton.

Figure 1

Frequency distributions of the mean values of SPAD. (A) The mean value of SPAD in the FBP period. (B) The mean value of SPAD in the BOP period. (C) The mean value of diurnal variation of SPAD.

Figure 2

GO enrichment analysis of candidate genes associated with cotton senescence.

Figure 3

GWAS identification of candidate gene in the SGR36. (A) Manhattan plot (upper) and LD heat map (lower) of SGR36. (B) Gene structure and haplotypes of the candidate gene GhMKK9. (C) Phenotypes of different haplotypes. There are 158 accessions for Hap1 and 197 accessions for Hap2. Asterisks indicate significance levels (*** p < 0.001); ns, not significant.

Figure 4

Functional analysis of the candidate gene GhMKK9. (A) Expression of GhMKK9 in young and old cotton leaves by qRT-PCR. (B) Phenotypes of empty control (CK) and VIGS cotton plants (pTRV2-GhMKK9) under drought stress. After four weeks, the CK and VIGS cotton plants were treated with water shortage for 7 days. (C) Expression levels of GhMKK9 in the CK and VIGS cotton plants. (D) SPAD value of the CK and VIGS plants under drought stress. (E) Phenotypes of six-week-old WT and transgenic Arabidopsis plants (OE7 and OE14). (F) Expression levels of GhMKK9 in the WT and transgenic Arabidopsis plants. (G,H) Expression levels of senescence-marked genes AtSAG12 and AtCAT2 in the WT and transgenic Arabidopsis plants. Asterisks indicate significance levels (*** p < 0.001, ** p < 0.01, and * p < 0.05).

Figure 5

Distribution of senescence-related genomic regions at chromosomes from this and previous studies. Red vertical bars represent genomic regions from this study. Blue vertical bars represent genomic regions from previous study. Black triangles indicate SGRs located within ~1 Mb of the genomic regions reported by previous study.