Fruit characteristics of sweet watermelon are largely the result of human selection. Here we report an improved watermelon reference genome and whole-genome resequencing of 414 accessions representing all extant species in the Citrullus genus. Population genomic analyses reveal the evolutionary history of Citrullus, suggesting independent evolutions in Citrullus amarus and the lineage containing Citrullus lanatus and Citrullus mucosospermus. Our findings indicate that different loci affecting watermelon fruit size have been under selection during speciation, domestication and improvement. A non-bitter allele, arising in the progenitor of sweet watermelon, is largely fixed in C. lanatus. Selection for flesh sweetness started in the progenitor of C. lanatus and continues through modern breeding on loci controlling raffinose catabolism and sugar transport. Fruit flesh coloration and sugar accumulation might have co-evolved through shared genetic components including a sugar transporter gene. This study provides valuable genomic resources and sheds light on watermelon speciation and breeding history.
The genus Citrullus contains seven extant species4. The only diecious and most morphologically unique species, Citrullus naudinianus, is commonly found in sub-Saharan Africa3,5. Citrullus ecirrhosus and Citrullus rehmii are adapted to a desert environment and are endemic to southern Africa. Citrullus colocynthis is grown for its medicinal properties and seed oil, and is widely distributed in northern Africa and southwestern and central Asia2. The wild form of Citrullus amarus can be found in southern Africa, and the cultivated types are grown throughout the Mediterranean region, where they are used for jam and animal fodder and as a source of water5. Citrullus mucosospermus is mainly grown for seed consumption and is now distributed in western Africa5. In addition, C. colocynthis, C. amarus and C. mucosospermus have been used in breeding programs to identify new sources of disease and pest resistance for the improvement of sweet watermelon3.
Windows XP Sweet 62 Fr ISO
Download File: https://protabliho.blogspot.com/?file=2vFaE9
Phylogenetic relationships between the Citrullus accessions were inferred using 89,914 SNPs at fourfold degenerate sites. The placement of the seven species in the phylogenetic tree (Fig. 1b) was largely consistent with the previously reported phylogeny4,5, with the most morphologically distinct Citrullus species, C. naudinianus, sister to the other six species, followed by C. colocynthis and C. rehmii. However, C. ecirrhosus was sister to C. amarus, C. mucosospermus and C. lanatus, instead of being most closely related to C. amarus, as proposed previously4. Two C. lanatus accessions collected in Sudan (PI 481871 and PI 254622) were placed in the deepest branch of the C. lanatus clade (Fig. 1b), supporting the idea that the primitive watermelons from Sudan and neighboring countries of northeastern Africa may be the closest to the progenitor of the sweet watermelon2,5,10. Twelve accessions were clustered into unexpected species groups and were therefore excluded from downstream analyses (Supplementary Table 5).
The variation map at single-base resolution empowered GWAS for seven important fruit quality traits in watermelon (Supplementary Table 6). In total, 43 association signals were identified, of which eight overlapped with previously identified QTLs. A peak strongly associated with flesh sweetness (measured by soluble solid content (SSC)) was identified within the previously reported QTL, QBRX2-1 (ref. 9), which harbors the sugar transporter gene ClTST2 (Cla97C02G036390, Fig. 2a). Two additional regions strongly associated with flesh sweetness were found on chromosome 10 (Fig. 2b), in agreement with previous GWAS and QTL studies13,14. These two regions contained the sucrose synthase gene Cla97C10G194010 and the raffinose synthase gene Cla97C10G196740, which contribute to the synthesis of sucrose and raffinose, respectively, the dominant metabolites transported in watermelon fascicular phloem tissues15. Two signals significantly associated with flesh color were detected on chromosomes 2 and 4 (Fig. 2c), with the one on chromosome 4 overlapping with the flesh color QTL FC4.1 (ref. 16) and harboring a lycopene β-cyclase gene (LCYB, Cla97C04G070940). In total, 14 signals associated with fruit shape were detected, with the strongest signal near the ClFS1 (Cla97C03G066390) gene, which is known to control fruit elongation17, and overlapping with fruit shape QTLs Qfsi3, FSI3.1 and FSI3.2 (refs. 9,16) (Fig. 2d). Three peaks highly associated with rind color and rind stripe were found on chromosomes 4, 6 and 8, corresponding to the rind trait loci, Dgo, S and D, respectively18 (Fig. 2e,f). Candidate genes in these peaks included Cla97C08G161570, which encodes a chloroplastic 2-phytyl-1,4-beta-naphthoquinone methyltransferase that is required for the conversion of 2-phytyl-1,4-beta-naphthoquinol to phylloquinone in photosystem I (ref. 19), and Cla97C04G068530, which encodes a magnesium-chelatase subunit H involved in chlorophyll synthesis20. The strongest signal associated with rind stripe was found in a WD40-repeat gene, Cla97C06G126710. In total, 13 regions were found to be associated with seed coat color (Fig. 2g). The strongest associated SNP on chromosome 3 overlapped with the seed coat color QTL qrc-c8-1 (ref. 21) and was located in Cla97C03G057100, which encodes a polyphenol oxidase that polymerizes o-quinones to produce black, brown or red pigments22.
Qfwt2-1 and Qfwt3 overlapped with both speciation and domestication sweeps (Fig. 3a,b). Another fruit weight QTL, Qfwt5-1, was found to be under selection only during watermelon improvement (Fig. 3c). Qfwt2-2 and Qfwt5-2 were not found in domestication or improvement sweeps, indicating their potential in the future improvement of sweet watermelon fruit size.
Selection for non-bitter fruits probably occurred during the initial domestication of sweet watermelon. Among the 374 Citrullus accessions evaluated for flesh bitterness, all nine C. colocynthis and 25 C. amarus accessions produced bitter fruits, whereas 12 of 16 C. mucosospermus accessions and all 324 C. lanatus landraces and cultivars accessions had non-bitter fruits (Supplementary Fig. 6). The previously identified bitterness QTL, qbt-c1-1 (ref. 21), contains the ClBt gene (Cla97C01G003400), which encodes a basic helix-loop-helix transcription factor and is homologous to the cucumber bitterness regulatory genes CsBt and CsBl (refs. 25,26). The genomic region near ClBt was highly differentiated between C. lanatus landraces and C. mucosospermus (Supplementary Fig. 7), and genetic diversity was substantially higher in C. mucosospermus than in C. lanatus landraces (Supplementary Fig. 8). At the SNP site leading to a premature stop codon of ClBt associated with non-bitterness25 (Chr01:3,216,322C to T), all C. colocynthis, C. amarus and C. mucosospermus accessions carrying bitter fruits had the homozygous bitter allele (C), whereas the homozygous non-bitterness allele (T) was found in all 12 non-bitter C. mucosospermus and all C. lanatus accessions, suggesting that this non-bitterness allele arose in the progenitor of C. lanatus and is fixed in sweet watermelons. Interestingly, the expression of ClBt was not detectable in the fruit flesh and rind (Supplementary Table 15). Exploring the public RNA-seq datasets in the Cucurbit Genomics Database27 revealed that the expression of ClBt was detected in the leaf but not in the flower, fruit, seed or root tissues, suggesting that the mechanisms by which ClBt regulates watermelon fruit bitterness may be complicated.
QTLs Qfru2-3, QBRX2-1 and QBrix6, which are known to control fruit flesh sugar content9,24, overlapped with domestication sweeps (Fig. 3b). Based on the identified improvement sweeps, QBRX2-1 was also under selection during the breeding of modern cultivars, whereas QBrix6 was probably only selected for during domestication (Fig. 3c and Supplementary Table 15). In addition, a sweetness locus, FCE10.1 (ref. 14), was found in the improvement sweeps (Fig. 3c). Several fruit quality-related genes were found in these and other sweeps, indicating their potential contribution to the aromatic flavor, texture and nutritional profiles of cultivated watermelon fruit (Supplementary Note).
Different from wild and primitive watermelons that produce pale-colored mature fruit, sweet watermelons produce abundant carotenoids in fruit flesh and accumulate them in chromoplasts during ripening, leading to a spectrum of flesh colors such as red, orange and yellow31. Phytoene synthase (PSY) is the first rate-limiting enzyme in the carotenogenesis pathway and defines the size of the carotenoid pool32. A PSY1 gene, Cla97C01G008760, within the flesh color QTL qFC.1 (ref. 33) was highly expressed in fruit flesh and its expression levels positively correlated with increased lycopene accumulation during fruit ripening23 (Supplementary Fig. 13a). Cla97C01G008760 was located in a genomic region that was highly differentiated between C. lanatus landraces and C. mucosospermus (Supplementary Figs. 7,13b). These results suggest that the regulation of PSY1 expression might contribute to the transition from pale-colored to red, orange or yellow flesh by increasing total carotenoid production in the ripening fruit of sweet watermelon. 2ff7e9595c
Comments