Comparative evolutionary and developmental dynamics of the cotton (Gossypium hirsutum) fiber transcriptome

Abstract

The single-celled cotton (Gossypium hirsutum) fiber provides an excellent model to investigate how human selection affects phenotypic evolution. To gain insight into the evolutionary genomics of cotton domestication, we conducted comparative transcriptome profiling of developing cotton fibers using RNA-Seq. Analysis of single-celled fiber transcriptomes from four wild and five domesticated accessions from two developmental time points revealed that at least one-third and likely one-half of the genes in the genome are expressed at any one stage during cotton fiber development. Among these, ~5,000 genes are differentially expressed during primary and secondary cell wall synthesis between wild and domesticated cottons, with a biased distribution among chromosomes. Transcriptome data implicate a number of biological processes affected by human selection, and suggest that the domestication process has prolonged the duration of fiber elongation in modern cultivated forms. Functional analysis suggested that wild cottons allocate greater resources to stress response pathways, while domestication led to reprogrammed resource allocation toward increased fiber growth, possibly through modulating stress-response networks. This first global transcriptomic analysis using multiple accessions of wild and domesticated cottons is an important step toward a more comprehensive systems perspective on cotton fiber evolution. The understanding that human selection over the past 5,000+ years has dramatically re-wired the cotton fiber transcriptome sets the stage for a deeper understanding of the genetic architecture underlying cotton fiber synthesis and phenotypic evolution.

Figure 1. Number of genes differentially expressed during fiber development within and between wild and domesticated cottons.

Left: Representative images of individual seeds with attached fiber are presented for domesticated (top) and wild (bottom) accessions. Right: Number of differentially expressed genes in developing cotton fiber within and between wild and domesticated cottons (RPKM≧1, FDR<0.05, fold-change≧1.5). For example, between two developmental stages within domesticated cottons, 362 genes were up-regulated at 10 dpa, whereas 762 genes were more highly expressed at 20 dpa. Similarly, between wild and domesticated cottons at 10 dpa, 1,844 genes were up-regulated in domesticated cottons, while 1,737 genes were more highly expressed in wild cottons.

Figure 2. The phylogeny of cellulose synthase (CesA) and cellulose synthase-like (CSL) genes.

Inside and outside circles show differential expression between wild and domesticated cottons at 10 and 20-regulation at domesticated cottons, while blue one indicates up-regulation at wild cottons during domestication. Grey circle shows no differential expression, while white circle designates no expression (zero read count).

Figure 3. Phylogenetic relationship of ACTIN and their gene expression patterns at 10 dpa.

(A) Phylogenetic relationship of G. raimondii ACTIN genes. Red and blue dots on the node indicate up- or down regulation at domesticated cottons, respectively. Asterisk (*) shows ACTIN homologues previously studied. (B) Expression patterns of differentially expressed ACT genes. Bar denotes standard error.

Figure 4. Expression patterns of genes related to phenylpropanoid pathway (A) and their actual expression levels (B) in wild and domesticated cottons.

Blue text in (A) indicates up-regulation in wild cottons relative to domesticated cottons at 10 dpa, and bar in (B) denotes standard error. PAL, phenylalanine ammonium lyase; C4H, cinnamate-4-hydroxylase; 4CL, 4-coumaroyl-CoA synthase; CHI, chalcone isomerase; F3H, flavonol 3-hydroxylase; FLS, flavonol synthase; DFR, dihydroflavonol-4-reductase; ANS, anthocyanin synthase, LDOX, leucoanthocyanidin dioxygenase; ANR, anthocyanidin reductase; COMT, F5H, CCR, CCD.

Figure 5. Number of TFs differentially expressed between wild and domesticated cotton.

Blue and red asterisk (*) represent over-represented TFs in wild and domesticated cottons relative to their counterparts at both developmental time points, respectively (proportion test; P<0.05).

Figure 6. Carbohydrate and fatty acid metabolisms, focusing on cell wall biosynthesis –.

Red text and line indicate up-regulated genes or pathways in domesticated cottons relative to wild cottons at 10 dpa, while blue text and lines show up-regulated genes or pathways in wild cottons compared to domesticated cottons at 10 dpa. Purple text indicates that the genes were up- or down-regulated during domestication. Blue and pink shaded boxes show enriched pathways in wild and domesticated cottons compared to their counterparts, respectively. ADPG, ADP-Glucose (ADPG); BGAL, β-galactosidase; BGLU, β-1,3-glucosidase; BXL, β-xylosidase; CER3/WAX2, fatty acid hydroxylase superfamily; CER4/FAR3, fatty acid reductase 3; DAHP, 3-deoxy-D-arabino-heptulosonate-7-phosphate; DHAP, dihydroxyacetone phosphate; E4P, erythrose-4-phosphate; F6P, Fructose-6-Phosphate; GAE, UDP-D-glucuronate-4-epimerase; G1P, Glucose-1-Phosphate; G3P, Glyceraldehyde-3-phosphate; G6P, Glucose-6-Phosphate; KCR, beta-ketoacyl reductase; KCS, 3-ketoacyl-CoA synthase; LTP, lipid transfer protein; PE, pectinesterase; PEP, phosphoenolpyruvate; PG2, polygalacturonase 2; PHS2, alpha-glucan phosphorylase 2; PL, Pectate lyase; R5P, Ribose-5-phosphate; SUS, sucrose synthase; UGD, UDP-glucuronate decarboxylase; UGDH, UDP-glucose-6-dehydrogenase; UGP, UDP-glucose pyrophosphorylase; VLCFA, very long chain fatty acids; WBC1, ATP-binding cassette transporter white-brown complex homolog protein 1; XTH, Xyloglucan endotransglycosylase.