Haplotype Networking of GWAS Hits for Citrulline Variation Associated with the Domestication of Watermelon

Abstract

Watermelon is a good source of citrulline, a non-protein amino acid. Citrulline has several therapeutic and clinical implications as it produces nitric oxide via arginine. In plants, citrulline plays a pivotal role in nitrogen transport and osmoprotection. The purpose of this study was to identify single nucleotide polymorphism (SNP) markers associated with citrulline metabolism using a genome-wide association study (GWAS) and understand the role of citrulline in watermelon domestication. A watermelon collection consisting of 187 wild, landraces, and cultivated accessions was used to estimate citrulline content. An association analysis involved a total of 12,125 SNPs with a minor allele frequency (MAF) >0.05 in understanding the population structure and phylogeny in light of citrulline accumulation. Wild egusi types and landraces contained low to medium citrulline content, whereas cultivars had higher content, which suggests that obtaining higher content of citrulline is a domesticated trait. GWAS analysis identified candidate genes (ferrochelatase and acetolactate synthase) showing a significant association of SNPs with citrulline content. Haplotype networking indicated positive selection from wild to domesticated watermelon. To our knowledge, this is the first study showing genetic regulation of citrulline variation in plants by using a GWAS strategy. These results provide new insights into the citrulline metabolism in plants and the possibility of incorporating high citrulline as a trait in watermelon breeding programs.

Keywords: acetolactate synthase; citrulline; ferrochelatase; genome-wide association study; haplotype; watermelon.

Figure 1

Citrulline content in watermelon accessions. (A) Box plots showing the range, mean, and distribution of citrulline content in cultivars, landraces, and egusi types. (B) Violin-scaled contour map showing world geographic variation in citrulline content across the accessions.

Figure 2

Principal component analysis (PCA) based on the first two components showing the distribution of (A) cultivars, landraces, and egusi types; (B) and low, medium, and high citrulline content in 187 watermelon accessions by using 1410 single nucleotide polymorphisms (SNPs) generated by genotyping by sequencing. Each dot represents an accession. EV indicates the percentage of explained variance.

Figure 3

Genetic relationship between a set of 187 watermelon accessions. Neighbor-joining (NJ) tree constructed with 1410 high-quality SNPs explains most of the genetic structure of watermelon germplasm by (A) type and (B) citrulline content. Accessions in blue-, pink-, and red-colored clades are egusi-types, sweet watermelons, and landraces, respectively. In (B), blue, pink, and red clades represent low, medium, and high citrulline content, respectively.

Figure 4

Boxplots for citrulline content (mg/g) in flesh tissue at SNP S02_33508197 located in the intron of ferrochelatase (A) and S06_30991451 located in an exon of acetolactate synthase (B). Significant differences (based on the Kruskal–Wallis test) with p ≤ 0.01 and p ≤ 0.05 are marked with two (**) and one (*) asterisks respectively.

Figure 5

Linkage disequilibrium structure and implicated genomic regions for SNPs aligned with: (A,B) ferrochelatase; (C,D) acetolactate synthase.

Figure 6

(A) Citrulline content and expression of candidate genes in flesh of selected watermelon accessions. Data are means ± SD (n = 3). (B) Expression of acetolactate synthase (ALS) and (C) Ferrochelatase 1 genes in flesh. PI 526238 accession, with low citrulline content, was used as a calibrator for relative expression in (A). Gene expression was normalized to that of Actin 8. Data are means ± SD (n = 3).

Figure 7

Haplotyping and network analysis of (A) acetolactate synthase and (B) ferrochelatase on chromosomes 6 and 2, respectively.