Joint transcriptomic and metabolomic analysis provides new insights into drought resistance in watermelon (Citrullus lanatus)

Abstract

Introduction: Watermelon is an annual vine of the family Cucurbitaceae. Watermelon plants produce a fruit that people love and have important nutritional and economic value. With global warming and deterioration of the ecological environment, abiotic stresses, including drought, have become important factors that impact the yield and quality of watermelon plants. Previous research on watermelon drought resistance has included analyzing homologous genes based on known drought-responsive genes and pathways in other species.

Methods: However, identifying key pathways and genes involved in watermelon drought resistance through high-throughput omics methods is particularly important. In this study, RNA-seq and metabolomic analysis were performed on watermelon plants at five time points (0 h, 1 h, 6 h, 12 h and 24 h) before and after drought stress.

Results: Transcriptomic analysis revealed 7829 differentially expressed genes (DEGs) at the five time points. The DEGs were grouped into five clusters using the k-means clustering algorithm. The functional category for each cluster was annotated based on the Kyoto Encyclopedia of Genes and Genomes (KEGG) database; different clusters were associated with different time points after stress. A total of 949 metabolites were divided into 10 categories, with lipids and lipid-like molecules accounting for the most metabolites. Differential expression analysis revealed 22 differentially regulated metabolites (DRMs) among the five time points. Through joint analysis of RNA-seq and metabolome data, the 6-h period was identified as the critical period for watermelon drought resistance, and the starch and sucrose metabolism, plant hormone signal transduction and photosynthesis pathways were identified as important regulatory pathways involved in watermelon drought resistance. In addition, 15 candidate genes associated with watermelon drought resistance were identified through joint RNA-seq and metabolome analysis combined with weighted correlation network analysis (WGCNA). Four of these genes encode transcription factors, including bHLH (Cla97C03G068160), MYB (Cla97C01G002440), HSP (Cla97C02G033390) and GRF (Cla97C02G042620), one key gene in the ABA pathway, SnRK2-4 (Cla97C10G186750), and the GP-2 gene (Cla97C05G105810), which is involved in the starch and sucrose metabolism pathway.

Discussion: In summary, our study provides a theoretical basis for elucidating the molecular mechanisms underlying drought resistance in watermelon plants and provides new genetic resources for the study of drought resistance in this crop.

Keywords: RNA-Seq; WGCNA; drought; metabolomic; watermelon.

Figure 1

The net photosynthetic rate (Pn), transpiration rate (Tr), intercellular carbon dioxide (Ci) and stomatal conductance (Gs) in leaves at five time points before and after drought stress (0 h, 1 h, 6 h, 12 h and 24 h). The results are presented as the means ± SDs (n = 3, **P < 0.01).

Figure 2

(A) Correlation and cluster analysis of the 15 RNA-seq samples before and after drought stress in watermelon. (B) Venn diagram of DEGs at different time points and 0 h after drought treatment. (C) All DEG GO features are annotated and categorized. There are three levels of GO enrichment, and different levels are distinguished by different colors. The level of significance is expressed by the height of the column. The higher the column is, the smaller the P value and the more significant the enrichment is. (D) All DEG KEGG pathway annotations. The color represents the P value. The larger the -log10(P value) and the smaller the P value are, the more significant the pathway is.

Figure 3

With respect to clusters of DEGs and KEGG pathway enrichment analysis, the expression levels of DEGs are depicted as heatmaps, and the data were z score normalized during analysis; the larger the value was, the greater the expression level, with the highest being 2. The smaller the value is, the lower the expression level, with the lowest being -2.

Figure 4

(A) PCA of 25 metabolome samples of watermelon before and after drought stress. (B) The differences in the metabolome at different time points and before stress were analyzed via a Venn diagram. (C) Histogram of all different metabolic classifications. (D) Clustering and metabolite classification of all DRMs. The expression levels of metabolites are depicted in heatmaps, and the data were z score normalized during analysis; the larger the value, the greater the expression level, with the highest being 2. The smaller the value, the lower the expression level, with the lowest being -2.

Figure 5

(A) KEGG enrichment analysis of DEGs and DRMs. (B) Changes in metabolite levels in the starch and sucrose metabolism pathways. The expression levels of metabolites were drawn as heatmaps, and the data were z score normalized during analysis. The larger the value, the greater the expression level, with the highest being 2. The smaller the value, the lower the expression level, with the lowest being -2. (C) Starch and sucrose metabolism pathway gene expression patterns. The expression levels of genes are depicted in heatmaps, and the data were z score normalized during analysis; the larger the value, the greater the expression level, with the highest being 2. The smaller the value, the lower the expression level, with the lowest being -2. (D) Correlation network diagram of starch and sucrose metabolism pathway genes and metabolites.

Figure 6

(A) Changes in metabolite content in the photosynthesis pathway. The expression levels of metabolites were drawn as heatmaps, and the data were z score normalized during analysis; the larger the value, the greater the expression level, with the highest being 2. The lower the value, the lower the expression level, with the lowest being -2. (B) For the photosynthesis pathway gene expression patterns, the expression levels of metabolites are depicted in heatmaps, and the data were z score normalized during analysis; the larger the value, the higher the expression level, with the highest being 2. The smaller the value, the lower the expression level, with the lowest being -2. (C) Correlation network diagram of photosynthesis pathway genes and metabolites.

Figure 7

(A) Changes in the metabolite content of plant hormones (IAA, ABA, JA and SA), the expression levels of metabolites are depicted in heatmaps, and the data were z score normalized during analysis; the larger the value, the higher the expression level, with the highest being 2. The lower the value, the lower the expression level, with the lowest being -2. (B) Plant hormone ABA pathway gene expression patterns and the expression levels of genes are depicted as heatmaps, and the data were z score normalized during analysis; the larger the value, the higher the expression level, with the highest being 2. The lower the value, the lower the expression level, with the lowest being -2. (C) Plant hormone IAA pathway gene expression patterns and the expression levels of genes are depicted in heatmaps, and the data were z score normalized during analysis; the larger the value, the higher the expression level, with the highest being 2. The lower the value, the lower the expression level, with the lowest being -2. (D) Plant hormone JA and SA pathway gene expression patterns. The expression levels of genes are depicted in heatmaps, and the data were z score normalized during analysis; the larger the value, the higher the expression level, with the highest being 2. The lower the value, the lower the expression level, with the lowest being -2. (E) Correlation network diagram of plant hormone (IAA, ABA, JA and SA) pathway genes and metabolites.

Figure 8

(A) Hierarchical clustering tree of genes based on coexpression network analysis. (B) Heatmap of the significant correlations between the modules and Pn, Tr, Ci, Gs, ABA, SA and JA. (C) Gene coexpression network within the turquoise, brown and green modules.

Figure 9

Scatter plot of the correlation between the transcriptome and qRT−PCR gene expression levels.