Figures
Abstract
COBRA-Like (COBL) genes, which encode a plant-specific glycosylphosphatidylinositol (GPI) anchored protein, have been proven to be key regulators in the orientation of cell expansion and cellulose crystallinity status. Genome-wide analysis has been performed in A. thaliana, O. sativa, Z. mays and S. lycopersicum, but little in Gossypium. Here we identified 19, 18 and 33 candidate COBL genes from three sequenced cotton species, diploid cotton G. raimondii, G. arboreum and tetraploid cotton G. hirsutum acc. TM-1, respectively. These COBL members were anchored onto 10 chromosomes in G. raimondii and could be divided into two subgroups. Expression patterns of COBL genes showed highly developmental and spatial regulation in G. hirsutum acc. TM-1. Of them, GhCOBL9 and GhCOBL13 were preferentially expressed at the secondary cell wall stage of fiber development and had significantly co-upregulated expression with cellulose synthase genes GhCESA4, GhCESA7 and GhCESA8. Besides, GhCOBL9 Dt and GhCOBL13 Dt were co-localized with previously reported cotton fiber quality quantitative trait loci (QTLs) and the favorable allele types of GhCOBL9 Dt had significantly positive correlations with fiber quality traits, indicating that these two genes might play an important role in fiber development.
Citation: Niu E, Shang X, Cheng C, Bao J, Zeng Y, Cai C, et al. (2015) Comprehensive Analysis of the COBRA-Like (COBL) Gene Family in Gossypium Identifies Two COBLs Potentially Associated with Fiber Quality. PLoS ONE 10(12): e0145725. https://doi.org/10.1371/journal.pone.0145725
Editor: Junkang Rong, Zhejiang A & F university, CHINA
Received: August 9, 2015; Accepted: December 8, 2015; Published: December 28, 2015
Copyright: © 2015 Niu et al. This is an open access article distributed under the terms of the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original author and source are credited
Data Availability: All relevant data are within the paper and its Supporting Information files.
Funding: This program was financially supported in part by the National Natural Science Foundation of China (31471539), the National High Technology Research and Development Program of China (863 Program) (2012AA101108-04-04), Key R & D program in Jiangsu Province (BE2015360), the Priority Academic Program Development of Jiangsu Higher Education Institutions (010-809001) and Jiangsu Collaborative Innovation Center for Modern Crop Production (No.10). The funders had no role in study design, data collection and analysis, decision to publish, or preparation of the manuscript.
Competing interests: The authors have declared that no competing interests exist.
Introduction
Cellulose, composed of long parallel linear β-1, 4-D-glucan chains, is the major component of the primary and secondary cell walls of plants. Studies on cellulose biosynthesis will not only facilitate an understanding of cell wall development but also open the possibility to increase the economic value of cotton fiber. Over the past decades, several genes essential for cellulose biosynthesis have been revealed, such as cellulose synthase (CESAs), KORRIGAN (KOR), chitinase-like genes (CTLs), fasciclin-like arabinogalactan genes (FLAs), NAC and MYB transcription factors [1–9]. Although these important findings have been made, the molecular mechanism for cellulose biosynthesis and deposition remains largely unknown.
Besides the importance of cotton fiber to the textile industry, it is also considered to be an ideal model to investigate the mechanism of cell elongation and cellulose deposition [10–11]. The fiber cell origins from the epidermis of ovules and is a highly elongated and thickened single-cell trichome with >90% crystalline cellulose in mature fibers. Fiber formation involves four distinct but overlapping stages [12–13]: initiation (-3 days post anthesis [DPA] to 3 DPA), elongation (3 DPA to 23 DPA), secondary cell wall synthesis (16 DPA to 40 DPA) and maturation (40 DPA to 50 DPA), which collectively determine the fiber yield and quality traits. Disorders of any stages will affect the quality traits of fiber such as fiber length, strength, micronaire, elongation and fiber uniformity [14–15]. As cellulose accounts for about 35% in primary cell wall and more than 90% in secondary cell wall of cotton fiber, it plays an important role in fiber elongation and secondary cell wall formation [16]. Especially during the secondary cell wall synthesis stage, almost pure cellulose is produced in cotton fiber cell. Thus cotton fiber provides an excellent model for mining and characterization of key genes related to biosynthesis and assembly of cellulose. On the other side, studies on cellulose biosynthesis in cotton fiber will also contribute to improving fiber quality and yield.
COBRA-Like genes (COBLs), which encode the glycosylphosphatidylinositol (GPI) anchored proteins, have been reported to play significant roles in the orientation of microfibrils and cellulose crystallinity status. The COBL family has been identified in Arabidopsis thaliana [17], Oryza sativa (BC1-Like) [18], Zea mays (BK2-Like) [19] and Solanum lycopersicum [20]. As a result, 12, 11, 9 and 17 members were found in these plants respectively. The COBL members are characterized by an N-terminal signal peptide, a carbohydrate-binding module (CBM), a highly conserved central cysteine-rich domain (CCVS) and a C-terminal domain including a GPI modification motif and a hydrophobic tail for targeting the protein to the endoplasmic reticulum. The COBL family is divided into two groups with the first group similar to COBRA and the second group to COBL7 in A. thaliana [17]. The mutant cobra in A. thaliana causes the defects in polar longitudinal expansion in root cells [21–23] and the mutant bc1l4 in O. sativa also shows a typical dwarf phenotype [24]. Similarly male gametophyte transmission in bc1l5 mutants is severely altered and specifically blocked [18] and the root hair of rth3 mutants in Z. mays initiates primordia but fails to elongate properly [25]. COBL genes are largely responsible for the secondary cell wall biosynthesis. The mutations cobl4 in A. thaliana, brittle culm 1 (bc1) in O. sativa and brittle stalk 2 (bk2) in Z. mays affect the mechanical strength of vascular bundles and have a significant reduction in cellulose content [26–28]. In addition, Liu et al. [29] demonstrated that COBL proteins could modulate cellulose structure by binding to cellulose and further affecting microfibril crystallinity. Whereas, it still needs to reveal the underlying molecular basis that how the COBLs proteins have functions in cellulose biosynthesis.
Recently the publically genomic information from three sequenced cotton species including the closest extant progenitor relatives for tetraploid cotton species, D-genome Gossypium raimondii [30], A-genome Gossypium arboreum [31], upland cotton genetic standard line Gossypium hirsutum acc. TM-1 [32], generates a solid foundation for characterizing gene families at a genome-wide level. Here we performed the first comprehensive investigation of the COBL gene family in three sequenced cotton species involving in the analysis of sequence phylogeny, genomic structure, chromosomal location and adaptive evolution. Further the expression patterns of COBL genes in G. hirsutum acc. TM-1 in various tissues/organs and different developmental stages of fiber development were elucidated by integrating RNA-Seq data and quantitative real-time PCR (qRT-PCR) analysis. Based on the co-expression analysis, correlation analysis and integration of quantitative trait loci (QTLs), we verified that two COBLs, GhCOBL9 and GhCOBL13 were significantly associated with fiber quality. This study will open up the possibility of exploring the use of COBLs to improve fiber quality in future cotton-breeding programs.
Materials and Methods
Database search and identification of COBL genes in Gossypium
The genomic database of three cotton species G. raimondii, G. arboreum and G. hirsutum acc. TM-1 were downloaded from http://www.phytozome.net/, http://cgp.genomics.org.cn and http://mascotton.njau.edu.cn/ respectively. The protein database of A. thaliana, O. sativa and Z. mays were obtained from The Arabidopsis Information Resource (TAIR: http://www.arabidopsis.org), the Rice Genome Annotation Project Database (RGAP release 7, http://rice.plantbiology.msu.edu/index.shtml) and http://www.phytozome.net/, respectively.
COBL domain (PF04833) was downloaded from Pfam (http://pfam.xfam.org/) and the genes shared with the domain could be obtained by taking PF04833 as the query to scan individually the three cotton protein database using the HMMER software version 3.0 [33] with the default parameters (E value < 0.01). Further, to confirm the genes above, gene prediction programs FGENESH (http://www.softberry.com/berry.phtml) and SMART (http://smart.embl-heidelberg.de/) were performed to detect the conserved motifs. Finally, the paralogs of COBLs in the three cotton species were confirmed by BLASTp.
Chromosomal locations and gene duplications
Chromosomal location of COBL genes in G. raimondii was performed using MapInspect software (http://www.plantbreeding.wur.nl/UK/ software_mapinspect.html). The nomenclature and description of the distribution of genes in chromosomes were derived from the map constructed by Zhao et al. [34]. In other two cotton species G. arboreum and G. hirsutum acc. TM-1, the nomenclature of COBL genes was following its ortholog in G. raimondii. DNAMAN (http://www.lynnon.com/) was conducted to calculate the sequence similarity among the COBL genes. Tandem duplicates were defined as genes separated by five or fewer genes and segmental duplicates were identified using the Plant Genome Duplication Database (http://chibba.agtec.uga.edu/duplication) [35]. The ratios of nonsynonymous to synonymous substitutions (Ka/Ks) were then analyzed to assess the selection pressure for the identified paralogous pairs [36].
Gene structure and phylogenetic analysis
MEGA 5.0 software (www.megasoftware.net) was used to construct the phylogenetic tree following the maximum likelihood method [37]. The parameters used were as follows: Test of phylogeny: Bootstrap method, No. of bootstrap replications: 1000, substitutions type: Amino acid, model/method: Jones-Taylor-Thornton (JTT) model, rates among sites: Uniform rates, gaps/missing data treatment: Complete deletion, ML heuristic method: Nearest-Neighbor-Interchange (NNI), initial tree for ML: Make initial tree automatically. The online ExPASy tool (http://www.expasy.org/tools/) and Gene Structure Display Server 2.0 (http://gsds.cbi.pku.edu.cn) were used to identify the physicochemical parameters [38] and the exon/intron organization. The alignment of the COBL family members was carried out with ClustalX 1.83 software [39].
Conserved motif annotation was performed using the MEME program [40]. If two or more motifs identified with SMART program represented the same domain and were located adjacently, they would be merged into one domain district. The parameters of the MEME program were as follows: Number of repetitions: Any, maximum number of motifs: 10, the optimum motif widths: between 4 and 50 residues. Furthermore the signal peptide, hydrophobic plot and GPI modification motif (ω site for dissociation) were predicted by the SignalP 4.1 Server (http://genome.cbs.dtu.dk/services/SignalP/), KYTE DOOLITTLE HYDROPATHY PLOT (http://gcat.davidson.edu/DGPB/kd/kyte-doolittle.htm) and the big-PI predictor (http://mendel.imp.ac.at/sat/gpi/gpi_server.html) [41] respectively.
Plant materials and DNA/RNA isolation
The field evaluations of TM-1 and Hai7124 were conducted in PaiLou experimental field, Nanjing Agricultural University, Jiangsu Province, China. All necessary permits for collecting TM-1 and Hai7124 were obtained from Nanjing Agricultural University, China. The field evaluations of natural population including 285 G. hirsutum and 4 G. barbadense cultivars or lines were conducted in the experiment field of three Ecological Stations of the Institute of Cotton Research, CAAS. All necessary permits for the field evaluations of these accessions were obtained from the Institute of Cotton Research, CAAS, China. All the field evaluations were not relevant to human subject or animal research. Therefore, they did not involve any endangered or protected species.
G. hirsutum acc. TM-1, the genetic standard line of upland cotton and G. barbadense cv. Hai7124 with superior fiber traits were used for gene cloning. G. hirsutum acc. TM-1 were also employed to carry out the expression analysis. Among them, vegetative tissues (roots, stems and leaves) were collected from two-week-old seedlings; Floral tissues (petals and anthers) were collected on the day of flowering; Fibers were sampled on the different days post anthesis. All samples were quick-frozen in liquid nitrogen and stored at -70°C before use.
Natural population including 285 G. hirsutum and 4 G. barbadense cultivars or lines (S1 Table) were collected mainly from China and some from foreign countries, which were available from the National Mid-term Genebank of the Institute of Cotton Research, Chinese Academy of Agricultural Sciences (CAAS). These accessions were grown in three Ecological Stations of the Institute of Cotton Research with three replicates: Kuche of Xinjiang province (northwestward cotton growing region), Anyang of Henan province (Yellow River valley cotton growing region) and Nanjing of Jiangsu province (Yangtze River valley cotton growing region) during 2007 to 2009. After harvesting of the plants, the fiber samples including three biological replicates were tested with HVI spectrum in the Supervision, Inspection and Test Center of Cotton Quality, Ministry of Agriculture in China. The analysis of fiber quality traits focused mainly on 2.5% fiber span length (FL), fiber strength (FS), fiber micronaire (FM), fiber elongation (FE) and fiber uniformity (FU).
Total DNA was extracted from the leaves of seedlings as described by Paterson et al. [42]. All samples were quantified using “one drop spectrophotometer OD-1000+” (OneDrop, Nanjing, China) and adjusted to a concentration of 20–60 ng/μL. Total RNA was isolated from all samples with CTAB [43] and the trizol of “TransScript One-Step gDNA Removal and cDNA Synthesis SuperMix” kit (TransGen, Nanjing, China) was used subsequently to obtain the first strand cDNA. The components of each reaction sample included 1 μg RNA, 1 μL of anchored oligo(dT)18 primer, 10 μL of 2*TS reaction mix, 1 μL of RT/RI enzyme mix, 1 μL of gDNA remover and additional ddH2O to give a total volume of 20 μL. All cDNA samples were stored at -30°C before use.
Primer design and gene cloning
Primers for PCR amplification of DNA or cDNA for sequencing were designed with Primer Premier 5 (http://www.premierbiosoft.com). Beacon Designer 7.91 software was used to design the primers for qRT-PCR based on the coding sequences of COBL members close to the 3’-UTR regions. Simultaneously, the specific primers for “Single Nucleotide Polymorphism (SNP)” of GhCOBL9 and GhCOBL13 were designed with the online software WebSNAPER-SNAP (http://pga.mgh.harvard.edu/cgi-bin/snap3/websnaper3.cgi) [44]. All the gene-specific primers were provided in S2 Table.
High-Fidelity ExTaq DNA Polymerase (TaKaRa Biotechnology Co. Ltd. China) was employed to conduct a standard PCR analysis following the manufacturer’s instructions with the amplification programs as follows: Pre-denaturation at 95°C for 5 min, 35 cycles of denaturation at 94°C for 45 s, annealing at 56°C for 45 s and extension at 72°C for 1 min/1 kb with a final extension at 72°C for 10 min. PMD19-T vector (TaKaRa Biotechnology Co. Ltd. China) and E. coli “strain” Top10 were used for transformation of target fragments. In order to obtain the sequences from both the A-subgenome and D-subgenome, at least 10 clones for each gene from each of the tetraploid species were picked randomly and sequenced with a minimum of three clones was used to determine the gene sequence in each duplicated copy.
Expression pattern analysis
The high-throughput RNA-sequencing data of G. hirsutum acc. TM-1 were employed from the accession codes SRA: PRJNA248163 in the National Center for Biotechnology Information (http://www.ncbi.nlm.nih.gov/) including the following tissues: leaves, roots and stems of two-week-old plants; petals and anthers dissected from whole mature flowers; ovules and fiber mixtures from plants -3, -1, 0, 1, 3 DPA and fibers from plants 5, 10, 20 and 25 DPA with three biological replicates. The expression levels of COBLs were calculated using the fragments per kilobase of exon model per million mapped reads (FPKM) method by using Cufflinks software with default parameters (http://cufflinks.cbcb.umd.edu/).
qRT-PCR was performed on an ABI 7500 real-time PCR system (ABI Biosystems, http://www.lifetechnologies.com) with the amplification programs as follows: Pre-denaturation at 95°C for 10 min, 40 cycles of denaturation at 95°C for 15 s, annealing at 60°C for 15 s and extension at 72°C for 15 s with a final extension at 72°C for 10 min. Each reaction sample was mixed with 10 μl of SYBR Green Master (Rox), 1.5 μL of cDNA, 5 μM of primers and 7.5 μL of ddH2O to give a total volume of 20 μL. For all real-time PCR reactions, three technical replicates were performed in each of the three biological experiments. The relative expression level was calculated using the 2-△CT method [45] and the expression level of the cotton histone3 (AF024716) gene was used as the endogenous control [46].
Pearson correlation analysis on expression pattern was used to calculate the correlation coefficient between RNA-seq and qRT-PCR detection, GhCESAs and GhCOBLs and different paralog pairs of GhCOBLs. For correlation analysis, the total expression of homeologous pairs in each gene from RNA-seq data was merged for matching the result from qRT-PCR detection using gene-specific primer pairs.
SNP/EcoTilling assays and QTLs integration for function prediction
The amplification program for SNP analysis by PCR was as follows: Pre-denaturation at 95°C for 10 min, 28 cycles of denaturation at 95°C for 15 s, annealing at 62°C for 30 s and extension at 72°C for 30 s followed by denaturing gel electrophoresis. For “EcoTilling” detection, several successive steps were performed following a method modified from that of Zeng et al. [47]. PCR amplification of the natural population (with annealing at 67°C for 30 s) was carried out twice to determine the density and specificity of targeted region with the same forward primer and two different reverse primers. Each sample was then mixed with DNA from the genetic standard line TM-1 in a 1:1 ratio and these were hybridized through 40 cycles of denaturation at 99°C for 10 min and annealing at 72°C with a decrease of 0.3°C per cycle. Finally CELⅠreaction was carried out for 30 min to cleave any mismatches between the query DNA and reference DNA. The denaturing gel electrophoresis after silver staining was used to pinpoint the cleaved fragments with the fragments sizes marked [48]. The cut DNAs were visible as bands and those with faster mobility than the full-length product were considered as polymorphisms. In addition, the DNA fragments around the polymorphic locus were cloned and sequenced to confirm the applicability and accuracy of the polymorphism detection.
Correlation analysis estimated by SPSS18.0 (http://www.spss.com.cn/) was conducted using the LSR (the least significant range) method to study significant difference of fiber quality traits between the two genotypes of targeted genes.
QTLs related to fiber quality traits (including FL, FS, FM, FE and FU) were retrieved and anchored to the G. hirsutum acc. TM-1 genome. Within ± 5 Mb interval flanking the QTLs, we integrated the QTLs with the prior target genes for association analysis using the MapChart program (www.joinmap.nl.).
Results
Genome-wide identification of COBL gene family in Gossypium
The whole genome sequence scaffolds of three sequenced cotton species (G. raimondii, G. arboreum and G. hirsutum acc. TM-1) were used for the genome-wide exploration of COBL family genes in Gossypium. Using the COBL protein database (PF04833), we searched the three protein database of cotton species by HMMER software. As a result, 19, 18 and 33 COBL members were identified respectively in G. raimondii, G. arboreum and G. hirsutum acc. TM-1 as shown in Table 1. From the phylogenetic view, one member in the diploid G. raimondii would be corresponding to one homologous gene in G. arboreum and two homologous genes (homeologs from A and D subgenome) in tetraploid G. hirsutum acc. TM-1. The inconsistency of the homologous genes among these three cotton species might result from the gene loss during their individually evolution process or assembly error in partial chromosomal regions, and need to be further confirmed. The related information of COBLs in Gossypium was summarized in Table 1.
Chromosomal distribution and gene duplication of COBL genes in G. raimondii
Two sequenced diploid genomes A genome G. arboreum and D genome G. raimondii were the closest extant progenitor relatives for tetraploid cotton species. Of them, the genome sequence in G. raimondii had been well assembled and annotated [30] and the collinearity of linkage groups between G. raimondii and G. hirsutum acc. TM-1 genome was obvious [32]. So we selected preferentially G. raimondii genome information to characterize the COBL family genes. By integrating 13 scaffolds of the G. raimondii genome (named as Chr. 1 to Chr. 13) with the reported high-density interspecific genetic map of allotetraploid cultivated cotton species [34], we obtained the corresponding relationship of orthologs between Chr. 1 to Chr. 13 in G. raimondii and D1 to D13 chromosomes in tetraploid cotton species. As a result, 19 candidate COBL genes were matched to 10 scaffolds of the G. raimondii genome. Based on the order of the homologs on chromosomes, we named the COBL members in G. raimondii as GrCOBL1 to GrCOBL19. Correspondingly, we also named COBL family genes GaCOBL1 to GaCOBL19 in G. arboreum and GhCOBL1 to GhCOBL19 in G. hirsutum acc. TM-1 following the orthologous relationship with the ones in G. raimondii (Table 1). The chromosomal distribution patterns of these GrCOBL genes were uneven. As in Fig 1, no GrCOBL was detected in the three chromosomes D4, D5 and D9, four GrCOBLs were found on the D12 chromosome, three on the D11 chromosome, two each were found on the D1, D3, D7 and D13 chromosomes, and only one each on the D2, D6, D8 and D10 chromosomes.
The chromosome numbers were consistent with the interspecific genetic map (D1 to D13) in allotetraploid cultivated cotton species [34] and the scaffolds (Chr.1 to Chr.13) in the genomic data of G. raimondii [30]. The nomenclature of COBLs was based on the order of the chromosomes in G. raimondii. Lines were drawn to connect the duplicated genes.
It was believed that whole genome duplication with genetic redundancy, chromosomal rearrangement and genome downsizing brought about a gene family a substantial expansion and evolutionary novelties [49]. The G. raimondii genome had undergone at least two rounds of genome-wide duplication [30]. To understand the mechanism of the COBL gene family expansion in G. raimondii, duplication events including tandem and segmental duplications were investigated by genome synteny analysis (Fig 1). Three gene pairs (GrCOBL4/GrCOBL5, GrCOBL12/GrCOBL13 and GrCOBL16/GrCOBL17) were detected as tandem duplication events, with 69.5%, 64.4% and 53.5% sequence similarity respectively. In addition ten gene pairs with high similarity that were located on segmental duplicated blocks were found to be a result of segmental duplication. All these paralogous pairs had ratios of nonsynonymous to synonymous substitutions (Ka/Ks) of less than 0.2 (Table 2) implying that COBL family genes in G. raimondii had mainly been subject to purifying selection pressure during the evolutionary process.
Phylogenetic and structural analysis of COBL genes
The overall protein sequences including COBLs in the eudicots (A. thaliana, G. raimondii) and monocots (O. sativa, Z. mays) were utilized to construct an unrooted tree to clarify the phylogenetic relationship of COBL genes from different species. Consistent with orthologs in A. thaliana and other species, the COBL family members were clustered into two subgroups Group Ⅰand Group Ⅱ, with phylogenetically related to AtCOBRA and AtCOBL7 in A. thaliana respectively. As in Fig 2, thirteen COBL genes from G. raimondii, seven from A. thaliana, eight from O. sativa and six from Z. mays were clustered in Group Ⅰ and six from G. raimondii, five from A. thaliana, three from O. sativa and three from Z. mays were in Group Ⅱ. The phylogenetic analysis indicated that the COBLs were descendants of an ancient duplication that occurred even before the separation of monocots and dicots.
Amino acid sequences were aligned using ClustalW and the phylogenetic tree was conducted using MEGA 5.0 software with the maximum likelihood method. Four different color fonts were represented for the four species.
The orthologs from three cotton species, G. raimondii, G. arboreum and G. hirsutum acc. TM-1 shared the same gene structure, thus, the COBLs in G. raimondii were used to give the insights into the diversity of the COBL family in Gossypium (Fig 3A and 3B). GrCOBLs in group Ⅰ had more introns than those in group Ⅱ. Besides, motif analysis (Fig 3C) displayed the conserved domains characterized by COBL gene family: the N-terminal signal peptide, the carbohydrate-binding module, a highly conserved CCVS motif and C-terminal domains that included the GPI modification motif and a hydrophobic tail, with the exception that three genes GrCOBL1, GrCOBL4 and GrCOBL7 lacked the N-terminal hydrophobic domain and GrCOBL1 lacked the CBM. Structurally the GrCOBLs in these two groups displayed the divergence at the N terminus. The members in Group Ⅱ contained an extra amino acid stretch after the N-terminal signal peptide and lead to the wide diversity in the length of amino acids ranging from 356 to 667 (Table 1).
A. The phylogenetic tree was conducted using MEGA 5.0 software with the maximum likelihood method. B. Exon/intron organization of GrCOBL members. Black boxes and black lines represented the exons and introns. The length of nucleotides was shown below. C. Schematic representation of the conserved motifs of GrCOBL proteins elucidated by MEME and SMART. Each motif was marked by the different colors and the ω-sites were showed in red font. The length of amino acids was shown below.
Moreover, multiple sequence alignments showed that hydrophobicity was ubiquitous for the COBL members, although the low sequence consistency of the N-terminal and C-terminal regions existed (S1 Fig). With the exception of the CCVS motif and the aromatic amino acids in the CBM, which were highly conserved, the majority of the residues showed characteristics that were specific to the two clustered groups. Besides, the ω-sites in GroupⅠ were close to the end of the C-terminal regions, while the ones in Group Ⅱ were close to the front of the C-terminal regions and were followed by a longer hydrophobic tail. This had not been reported previously and required to be further confirmed.
Expression patterns of COBL genes in G. hirsutum acc. TM-1
Based on RNA-Seq data of G. hirsutum acc. TM-1 including vegetative tissues (roots, stems and leaves), floral tissues (petals and anthers) and fibers in the development stage (0 DPA, 10 DPA and 20 DPA), we detected the expression patterns of COBL genes in G. hirsutum acc. TM-1. With the exception of GhCOBL14 (not detected in G. hirsutum acc. TM-1) and GhCOBL1, GhCOBL12 (FPKM<1.0 in all tested tissues), the expression patterns of other 15 GhCOBLs with distinguishable expression of homeologs could be grouped into four major classes (Fig 4). Four genes GhCOBL3, GhCOBL8, GhCOBL10 and GhCOBL16 in Class Ⅰ were highly expressed in all tissues examined with slight expression differences among the different samples. GhCOBL9 and GhCOBL13 in Class Ⅱ showed the predominant expression in the fiber tissue at 20 DPA. Five genes GhCOBL2, GhCOBL6, GhCOBL15, GhCOBL18 and GhCOBL19 in Class Ⅲ showed preferential expression in floral tissue with some expressed in fiber tissue. Five genes GhCOBL4, GhCOBL5, GhCOBL7, GhCOBL11 and GhCOBL17 in Class Ⅳ showed preferential expression in root or stem tissue. Generally, the similar expression pattern between homeologs in most COBLs were detected in different tissues in G. hirsutum acc. TM-1. Further, the gene-specific qRT-PCR was conducted to verify the expression of the 15 GhCOBLs in tetraploid cotton (S2 Fig) and the results of qRT-PCR were in agreement with the RNA-Seq data (S3 Fig). A correlation between the expression patterns in different tissues for the paralogous pairs was also detected (Table 2). The correlation coefficients in three COBL pairs (GhCOBL2/GhCOBL19, GhCOBL3/GhCOBL10 and GhCOBL4/GhCOBL16) showed the positive correlation, with the expression correlation of GhCOBL2/GhCOBL19 and GhCOBL3/GhCOBL10 greater than 0.5. However, other pairs had no clear correlation.
The eight different tissues of G. hirsutum acc. TM-1 were involved here: 1: root; 2: stem; 3: leaf; 4: petal; 5: anther; 6: 0 DPA ovule; 7: 10 DPA fiber; 8: 20 DPA fiber. Log2 (FPKM) indicated the different transcriptome profiling (FPKM: fragments per kb per million mapped reads). At and Dt were derived from the A-genome and D-genome progenitor in the tetraploid cotton.
To elucidate the expression profiles of COBL genes during fiber development, we further investigated RNA-Seq data of 18 GhCOBL genes (with GhCOBL14 not detected in G. hirsutum acc. TM-1) in nine samples involved in the fiber development stages of initiation (-3 DPA, -1 DPA, 0 DPA and 1 DPA), elongation (3 DPA, 5 DPA and 10 DPA) and secondary cell wall biosynthesis (20 DPA and 25 DPA) in G. hirsutum acc. TM-1 (Fig 5). With the exception of six genes GhCOBL1, GhCOBL2, GhCOBL4, GhCOBL7, GhCOBL11 and GhCOBL19, which were expressed at lower levels (FPKM<1.0), the remaining genes had a varied pattern of expression during the biosynthesis of cellulose. Of these genes, four COBLs (GhCOBL3, GhCOBL8, GhCOBL10 and GhCOBL16) were expressed at all the tested stages and maintained higher accumulation levels; Four COBLs (GhCOBL12, GhCOBL17, GhCOBL6 and GhCOBL18) and one COBLs (GhCOBL15) showed relatively high expression levels during the initiation stage of the fiber cell and elongation stage of the fiber cell respectively than that in other tested tissues; Three COBLs (GhCOBL5, GhCOBL9 and GhCOBL13) displayed higher accumulation during the secondary cell wall biosynthesis stage. The expression patterns of GhCOBLs in different tissues and fiber development stages displayed a dynamic change in the life cycle of fiber cells and cellulose synthesis.
The fiber developmental stages involved in initiation (from -3 DPA to 0 DPA), elongation (from 0 DPA to 10 DPA) and secondary wall biosynthesis (from 20 DPA to 25 DPA) were sampled for detecting the expression levels of COBL genes during the fiber development. The relative expression levels were shown as Log2 (FPKM). At and Dt were derived from the A-genome and D-genome progenitor in the tetraploid cotton.
Functional prediction of two COBL genes and their roles in fiber quality
To elucidate the relationship between COBL genes and fiber quality traits, we focused on the two COBL genes (GhCOBL9 and GhCOBL13) which preferentially expressed during the secondary cell wall stage of fiber development. Cellulose synthases (CESAs) were the key factors for cellulose biosynthesis [50] and there had been at least 15 CESA genes which had been classified into six groups in G. raimondii [30], so we investigated the expression correlations between the two COBLs and CESAs based on the RNA-Seq data during fiber developmental stages (S3 Table and Fig 6). The results showed that GhCOBL9/GhCOBL13 had nearly 100% correlation with the expression levels of GhCESA4A/B, GhCESA7A/B and GhCESA8B.
The X axis indicated the different fiber development stages and the Y axis indicated the expression levels.
To further unravel the potential roles of GhCOBL9/GhCOBL13 in fiber quality traits, we cloned the A and D subgenome of these two genes in G. hirsutum acc. TM-1 and their orthologs in G. barbadense cv. Hai 7124 (S4 Table). Based on the nucleotide polymorphisms of COBL9/COBL13 between TM-1 Dt and Hai7124 Dt, the SNP markers were developed successfully and confirmed using the SNP-PCR technology for COBL9 Dt and “EcoTilling” for COBL13 Dt respectively (S5 Table). For COBL9 Dt, the detected nucleotide polymorphism site was involved in a premature termination codon in TM-1 (TAA) but was not in Hai7124 (AAA) (S4 Fig). The SNP site also existed the similar difference within the G. hirsutum accessions (S1 Table) and the allele type “AAA” showed the favorable phenotypes related to the fiber quality traits such as FL, FM, FU (p<0.01) and FS, FE (p<0.05), especially with the longer and thinner fiber (Table 3). While for COBL13 Dt, the five detected SNPs were existed just between the G. hirsutum and G. barbadense but not within the G. hirsutum accessions (S1 Table; S4 Fig). Since the nucleotide polymorphism sites of COBL9 and COBL13 existed between TM-1 and Hai7124, we detected the expression levels of these two COBLs both in TM-1 and Hai7124 (Fig 7). COBL9 At and COBL13 At showed a high level accumulation from 10 DPA to 24 DPA both in TM-1 and Hai7124, even though the expression peak detected at 20 DPA in TM-1 but 24 DPA in Hai7124. However, compared to TM-1, COBL9 Dt and COBL13 Dt had an increasing accumulation from 10 DPA to 24 DPA in Hai7124 with peak at 24 DPA. In particular, COBL9 Dt in Hai7124 displayed a high level accumulation, whereas the COBL9 Dt in TM-1 was almost undetectable from 10 DPA to 24 DPA, which was consistent with significant correlation between COBL9 Dt and fiber quality traits. As for the contribution of COBL9/COBL13 from the A-subgenome to fiber qualities, we would further investigate in the future study.
The At and Dt were derived from A-subgenome and D-subgenome specific in tetraploid cotton species respectively. The Y axis indicated relative expression levels and the X axis indicated the different fiber development stages. “*” and “**” denoted the differences at P < 0.05 and P < 0.01 between TM-1 and Hai7124 respectively.
QTLs related to fiber quality traits were also employed to predict the roles of GhCOBL9 and GhCOBL13. By aligning the genome scaffolds in G. hirsutum acc. TM-1, GhCOBL9 Dt and GhCOBL13 Dt were mapped on the D8 and D11 chromosomes. Therefore, we further searched the QTLs flanking the target sites with ± 5 Mb intervals. As in Fig 8, both GhCOBL9 Dt and GhCOBL13 Dt were co-localized with the QTLs involved in several cotton fiber quality traits [51–59].
Fiber quality traits included fiber length (FL), fiber strength (FS), fiber micronaire (FM), fiber elongation (FE) and fiber uniformity (FU). Different symbols represented the QTLs with the references followed.
Discussion
COBL gene family was highly conserved in different species
The COBRA-Like gene family shared the CBM motif, CCVS motif, an N-terminal signal peptide and highly hydrophobic C terminus and had been identified in four species. These COBL genes could be classified into two groups GroupⅠand Group Ⅱ. In this study we systematically identified COBL members in three sequenced cotton species: 19 in G. raimondii, 18 in G. arboreum and 33 in G. hirsutum acc. TM-1. Consistent with other species, COBL family members in Gossypium were also classified into two groups, with Group Ⅱ carrying an amino acid stretch of unknown function after the N terminus and a longer hydrophobic tail following the ω-site.
Phylogenetic analysis showed that group Ⅰ and group Ⅱ contained the COBLs both from monocots and dicots (Fig 2), indicating that the COBL family members were descendants of an ancient duplication that occurred even before the separation of monocots and dicots. However, numerous duplications occurred after the divergence of monocots and dicots with more members of the COBL family in dicots than in monocots. In general, there was high sequence consistency of COBLs between G. raimondii and A. thaliana, however some exceptions detected (Fig 2). For example, GrCOBL5 and GrCOBL7 had relatively distant evolutionary relationships with the A. thaliana branch in Group Ⅰ, which might result from the additional duplication event that occurred in Gossypium, although A. thaliana and Gossypium both underwent the two common duplication events (β, α) [60–61]. Gene duplication in Gossypium might have played a crucial role in driving evolutionary novelty and increasing adaptation to new environments by functional diversification.
COBL gene family showed highly functional diversity in different tissues of Gossypium
In the study, COBL genes in G. hirsutum acc. TM-1 were found to have diverse expression patterns in vegetative tissues, floral tissues and developing fibers. As an example, GhCOBL2 and GhCOBL19 exhibited a unique and high expression pattern in the anther; while GhCOBL4 showed preferential expression in the stem (Fig 4). These results implied COBL genes played diverse functions in different cotton tissues. Actually, the diverse roles of COBLs had been elucidated in other plants. For instance, in A. thaliana, the functions of COBL genes were focused on two aspects. One represented by AtCOBRA was considered to have a significant impact on the orientation of cell expansion [62] and the other represented by AtCOBL4 was largely responsible for cellulose crystallinity status and secondary cell wall reconstruction [19]. In addition, it had been suggested that OsBC1L genes, the orthologs of AtCOBL4, performed a range of functions and participated in a various developmental processes in rice [29]. The functional divergence of COBLs could also be observed in the different developmental stages of fiber formation. For example, GhCOBL15 exhibited high transcript level in 10 DPA fibers but was low in 20 and 25 DPA fibers. Conversely, GhCOBL9 and GhCOBL13 were highly expressed in 20 and 25 DPA fibers but were low in 10 DPA fibers (Fig 6). These results indicated GhCOBL15 and GhCOBL9/GhCOBL13 might have the different functions in fiber developmental process. The different expression pattern of COBL genes could be indicative of their diverse functions in different tissues or developmental processes.
Interestingly, the functions of COBLs might be different from their phylogenetic clusters. We observed that the GhCOBLs with the differential expression patterns might come from the same phylogenetic group and the GhCOBLs with the similar expression patterns also might belong to the different phylogenetic groups (Figs 4 and 5). For example, four GhCOBL genes, GhCOBL8 and GhCOBL16 in group Ⅰ, and GhCOBL3 and GhCOBL10 in groupⅡ, showed similar expression patterns. Besides, AtCOBRA was expressed in most tissues in A. thaliana, while cobra mutant plants only had an altered phenotype in the root. GrCOBL8 and GrCOBL16 were clustered in the same branch as AtCOBRA and were preferentially expressed in a variety of tissues. Their functional mechanisms in cotton remained to be verified.
GhCOBL9 and GhCOBL13 potentially affected the cotton fiber quality
Cotton was an important raw material for the textile industry worldwide. During the development of cotton fiber, the deposition of cellulose determined the thickness of cell wall, which was the structural basis for fiber strength. Based on the reasons below, we could conclude that GhCOBL9 and GhCOBL13 play important roles during the stages of fiber formation.
Firstly, GhCOBL9 and GhCOBL13 were much higher expressed in fibers during the stages of secondary cell wall biosynthesis (Figs 4 and 5) and had a significantly high coexpression pattern with GhCESA4, GhCESA7 and GhCESA8 (Fig 6), which were involved in cellulose formation of secondary cell wall [63]. The functional roles of both GhCOBL9 and GhCOBL13 orthologs in other species had been revealed previously. irx6 (a T-DNA mutant of AtCOBL4) in A. thaliana exhibited abnormal secondary wall thickening and reduced stem strength [26] similar to the effect of BC1 in O. sativa [27] and BK2 in Z. mays [28]. It was reasonable that GhCOBL9 and GhCOBL13 which showed high amino acid sequence similarities to AtCOBL4 might play the similar role in cotton fiber secondary cell wall thickening.
Secondly, correlation analysis revealed that nucleotide polymorphism of GhCOBL9 Dt, similar to Hai7124 lineage, had significantly positive correlations with several fiber quality traits, seemed bring the favorable fiber quality in G. hirsutum (Table 3). Although no nucleotide polymorphism sites were detected for GhCOBL13 within G. hirsutum accessions, both of these two COBLs had higher expression levels in Hai7124 which showed the superior fiber quality than in TM-1 (Fig 7). Furthermore, it was demonstrated that GhCOBL9 and GhCOBL13 were co-located with the QTLs related to fiber quality reported previously (Fig 8). These findings collectively indicated GhCOBL9 and GhCOBL13 exerted an important role in fiber quality. Since modern agricultural mechanization requires higher fiber qualities in cotton, it is reasonable to make full use of GhCOBL9 and GhCOBL13 for improving the fiber quality and enhancing the durability of textiles by transgenic technology or molecular marker-assisted breeding.
Supporting Information
S1 Fig. Multiple sequence alignments of 19 COBL members in G. raimondii and AtCOBRA, AtCOBL7 in A. thaliana.
Multiple sequence alignments were carried out with ClustalX 1.83. The conserved motifs were marked by the boxes with different colors and the ω-sites were showed in red font. “*” denoted the aromatic amino acids in the CBM region.
https://doi.org/10.1371/journal.pone.0145725.s001
(TIF)
S2 Fig. Expression patterns of COBL members based on qRT-PCR analysis of G. hirsutum acc. TM-1.
The X axis indicated the different tissues and organs of G. hirsutum acc. TM-1 and the Y axis indicated relative expression levels of GhCOBL members. The cotton histone3 (AF024716) gene was used as the reference gene and the error bars were calculated based on three biological replicates using standard deviation (SD). 1: root; 2: stem; 3: leave; 4: petal; 5: anther; 6: 0DPA ovule; 7: 10DPA fiber; 8: 20DPA fiber.
https://doi.org/10.1371/journal.pone.0145725.s002
(TIF)
S3 Fig. Correlation analysis of expression pattern between the FPKM of expression profiling and relative expression level of qRT-PCR for GhCOBL members.
The X axis indicated the FPKM of expression profiling and the Y axis indicated relative expression level of qRT-PCR.
https://doi.org/10.1371/journal.pone.0145725.s003
(TIF)
S4 Fig. The different PCR fragments and sequencing analysis of the polymorphic sites between the two different allele types for COBL9 and COBL13 respectively.
A and B: Distinct fragments of GhCOBL9 and GhCOBL13 in the two different allele types revealed by denaturing gel electrophoresis via SNP-PCR (A) and EcoTilling analysis (B). “1–12” stand for the randomly selected individuals in G. hirsutum (including 1: 70-29-5, 2: Bao6716, 3: BaZhou5628, 4: ChangRong67-12, 5: Coker139, 6: GP137, 7: Ji91-12, 8: Zhong4612YaH, 9; HuBeiSongZiDaLing, 10: Zhong507145, 11: ZhongZhiBD89 and 12: ZhongZi10Hao) and “13–16” stand for the four lines in G. barbadense (including 13:E24-33891, 14: E24-33892, 15: Hai7124 and 16: Yinzi6022). “*” denoted that each DNA sample was mixed and hybridized with TM-1 (in 1: 1 ratio). C and D: Polymorphic sites of GhCOBL9 (C) and GhCOBL13 (D) were subsequently confirmed by sequencing and nucleotide polymorphisms were marked in orange shadows in each site.
https://doi.org/10.1371/journal.pone.0145725.s004
(TIF)
S1 Table. Information on 289 accessions used for correlation analysis of the targeted genes with fiber quality traits.
https://doi.org/10.1371/journal.pone.0145725.s005
(XLS)
S2 Table. Primers information for PCR amplification, qRT-PCR and EcoTilling analysis of COBL members.
The R' primer of EcoTilling was to conduct the second PCR amplification for amplifing the density and specificity.
https://doi.org/10.1371/journal.pone.0145725.s006
(XLS)
S3 Table. Pearson correlation coefficients of expression pattern between GhCOBLs and GhCESAs in different fiber developmental stages.
The pearson correlation coefficient: r>0: positive correlation; r<0: negative correlation; r = 0: no linear correlation. The genes in yellow shadow have the lowest expression level in all tested tissues in fiber developmental stages.
https://doi.org/10.1371/journal.pone.0145725.s007
(XLS)
S4 Table. Sequence information on COBL9 and COBL13 in cotton.
https://doi.org/10.1371/journal.pone.0145725.s008
(XLS)
S5 Table. Nucleotide polymorphism of COBL9 and COBL13 in cotton.
https://doi.org/10.1371/journal.pone.0145725.s009
(XLS)
Author Contributions
Conceived and designed the experiments: WZG. Performed the experiments: ELN XGS JHB YDZ XMD. Analyzed the data: ELN CZC CPC WZG. Contributed reagents/materials/analysis tools: WZG. Wrote the paper: ELN XGS WZG.
References
- 1. Beeckman T, Przemeck GK, Stamatiou G, Lau R, Terryn N, De Rycke R, et al. Genetic complexity of cellulose synthase a gene function in Arabidopsis embryogenesis. Plant Physiol. 2002;130(4):1883–1893. pmid:12481071
- 2. Ellis C, Karafyllidis I, Wasternack C, Turner JG. The Arabidopsis mutant cev1 links cell wall signaling to jasmonate and ethylene responses. Plant Cell. 2002;14(7):1557–1566. pmid:12119374
- 3. MacMillan CP, Mansfield SD, Stachurski ZH, Evans R, Southerton SG. Fasciclin-like arabinogalactan proteins: specialization for stem biomechanics and cell wall architecture in Arabidopsis and Eucalyptus. Plant J. 2010;62(4):689–703. pmid:20202165
- 4. Mitsuda N, Seki M, Shinozaki K, Ohme-Takagi M. The NAC transcription factors NST1 and NST2 of Arabidopsis regulate secondary wall thickenings and are required for anther dehiscence. Plant Cell. 2005;17(11):2993–3006. pmid:16214898
- 5. Nicol F, His I, Jauneau A, Vernhettes S, Canut H, Hofte H. A plasma membrane-bound putative endo-1,4-beta-D-glucanase is required for normal wall assembly and cell elongation in Arabidopsis. EMBO J. 1998;17(19):5563–5576. pmid:9755157
- 6. Sánchez-Rodríguez C, Estévez JM, Llorente F, Hernández-Blanco C, Jordá L, Pagán I, et al. The ERECTA receptor-like kinase regulates cell wall-mediated resistance to pathogens in Arabidopsis thaliana. Mol Plant Microbe Interact. 2009;22(8):953–63. pmid:19589071
- 7. Taylor NG, Howells RM, Huttly AK, Vickers K, Turner SR. Interactions among three distinct CesA proteins essential for cellulose synthesis. Proc Natl Acad Sci USA. 2003;100(3):1450–1455. pmid:12538856
- 8. Taylor NG, Laurie S, Turner SR. Multiple cellulose synthase catalytic subunits are required for cellulose synthesis in Arabidopsis. Plant Cell. 2000;12(12):2529–2540. pmid:11148295
- 9. Zhong R, Richardson EA, Ye ZH. The MYB46 transcription factor is a direct target of SND1 and regulates secondary wall biosynthesis in Arabidopsis. Plant Cell. 2007;19(9):2776–2792. pmid:17890373
- 10. Haigler CH, Betancur L, Stiff MR, Tuttle JR. Cotton fiber: a powerful single-cell model for cell wall and cellulose research. Front Plant Sci. 2012;3:104. pmid:22661979
- 11. Kim HJ, Triplett BA. Cotton fiber growth in planta and in vitro: models for plant cell elongation and cell wall biogenesis. Plant Physiol. 2001;127(4):1361–1366. pmid:11743074
- 12. Basara AS, Malik CP. Development of cotton fiber. Inter Rev Cyto. 1984;89:65–113.
- 13. Lee JJ, Woodward AW, Chen ZJ. Gene expression changes and early events in cotton fibre development. Ann Bot. 2007;100(7):1391–1401. pmid:17905721
- 14. Schubert AM, Benedict CR, Berlin JD. Cotton fiber development-kinetics of cell elongation and secondary wall thickening. Crop Sci. 1973;13(6):704–709.
- 15. Endler A, Persson S. Cellulose synthases and synthesis in Arabidopsis. Mol Plant. 2011;4:199–211. pmid:21307367
- 16. Arpat AB, Waugh M, Sullivan JP, Gonzales M, Frisch D, Main D, et al. Functional genomics of cell elongation in developing cotton fibers. Plant Mol Biol. 2004;54(6):911–929. pmid:15604659
- 17. Roudier F, Schindelman G, DeSalle R, Benfey PN. The COBRA family of putative GPI–anchored proteins in Arabidopsis. a new fellowship in expansion. Plant Physiol. 2002; 130(2):538–548. pmid:12376623
- 18. Dai XX, You CJ, Wang L, Chen GX, Zhang QF, Wu CY. Molecular characterization, expression pattern, and function analysis of the OsBC1L family in rice. Plant Mol Biol. 2009;71(4–5):469–48119. pmid:19688299
- 19. Brady SM, Song S, Dhugga KS, Rafalski JA, Benfey PN. Combining expression and comparative evolutionary analysis. The COBRA gene family. Plant Physiol. 2007; 143(1):172–187. pmid:17098858
- 20. Cao Y, Tang XF, Giovannoni J, Xiao FM, Liu YS. Functional characterization of a tomato COBRA-like gene functioning in fruit development and ripening. BMC Plant Biol. 2012;12(1):211.
- 21. Benfey PN, Linstead PJ, Roberts K, Schiefelbein JW, Hauser MT, Aeschabacher RA. Root development in Arabidopsis: four mutants with dramatically altered root morphogenesis. Development. 1993;119(1):57–7014. pmid:8275864
- 22. Hauser MT, Morikami A, Benfey . Conditional root expansion mutants of Arabidopsis. Development. 1995;121(4):1237–1252. pmid:7743935
- 23. Roudier F, Fernandez AG, Fujita M, Himmelspach R, Borner GH, Schindelman G, et al. COBRA, an Arabidopsis extracellular glycosyl-phosphatidyl inositol–anchored protein, specifically controls highly anisotropic expansion through its involvement in cellulose microfibrils orientation. Plant Cell. 2005;17(6):1749–1763. pmid:15849274
- 24. Dai XX, You CJ, Chen GX, Li XH, Zhang QF, Wu CY. OsBC1L4 encodes a COBRA-like protein that affects cellulose synthesis in rice. Plant Mol Biol. 2011;75(4–5): 333–345. pmid:21264494
- 25. Hochholdinger F, Wen TJ, Zimmermann R, Chimot-Marolle P, da Costa e Silva O, Bruce W, et al. The maize (Zea mays L.) roothairless 3 gene encodes a putative GPI-anchored, monocot-specific, COBRA-like protein that significantly affects grain yield. Plant J. 2008; 54(5): 888–898. pmid:18298667
- 26. Brown DM, Zeef LAH, Ellis J, Goodacre R, Turner SR. Identification of novel genes in Arabidopsis involved in secondary cell wall formation using expression profiling and reverse genetics. Plant Cell. 2005;17(8): 2281–2295. pmid:15980264
- 27. Li YH, Qian Q, Zhou YH, Yan MX, Sun L, Zhang M, et al. BRITTLE CULM1, which encodes a COBRA-like protein, affects the mechanical properties of rice plants. Plant Cell. 2003;15(9):2020–2031. pmid:12953108
- 28. Ching A, Dhugga KS, Appenzeller L, Meeley R, Bourret TM, Howard RJ, et al. Brittle stalk 2 encodes a putative glycosylphos phatidylinositol-anchored protein that affects mechanical strength of maize tissues by altering the composition and structure of secondary cell walls. Planta. 2006;224(5):1174–1184. pmid:16752131
- 29. Liu LF, Shang-Guan KK, Zhang BC, Liu XL, Yan MX, Zhang LJ, et al. Brittle Culm1, a COBRA-Like protein, functions in cellulose assembly through binding cellulose microfibrils, PLoS Genet. 2013;9(8): e1003704. pmid:23990797
- 30. Paterson AH, Wendel JF, Gundlach H, Guo H, Jenkins J, Jin D, et al. Repeated polyploidization of Gossypium genomes and the evolution of spinnable cotton fibres. Nature. 2012;492(7429):423–427. pmid:23257886
- 31. Li F, Fan G, Wang K, Sun F, Yuan Y, Song G, et al. Genome sequence of the cultivated cotton Gossypium arboreum. Nat Genet. 2014;46(6):567–572. pmid:24836287
- 32. Zhang TZ, Hu Y, Jiang WK, Fang L, Guan XY, Chen JD, et al. Sequencing of allotetraploid cotton (Gossypium hirsutum acc. TM-1) provides a resource for fiber improvement. Nat Biotechnol. 2015;33:531–537. pmid:25893781
- 33. Finn RD, Clements J, Eddy SR. HMMER web server: interactive sequence similarity searching. Nucleic Acids Res. 2011;39:29–37.
- 34. Zhao L, Lv YD, Cai CP, Tong XC, Chen XD, Zhang W, et al. Toward allotetraploid cotton genome assembly: integration of a high-density molecular genetic linkage map with DNA sequence information. BMC Genomics. 2012;13(1):539.
- 35. Lee TH, Tang HB, Wang XY, Paterson AH. PGDD: a database of gene and genome duplication in plants. Nucleic Acids Res. 2013;41:D1152–1158. pmid:23180799
- 36. Tang H, Bowers JE, Wang X, Ming R, Alam M, Paterson AH. Synteny and collinearity in plant genomes. Science. 2008;320:486–488. pmid:18436778
- 37. Cai CP, Niu EL, Du H, Zhao L, Feng Y, Guo WZ. Genome-wide analysis of the WRKY transcription factor gene family in Gossypium raimondii and the expression of orthologs in cultivated tetraploid cotton. Crop J. 2014;2(2–3):87–101.
- 38. Gasteiger E, Gattiker A, Hoogland C, Ivanyi I, Appel RD, Bairoch A. ExPASy: the proteomics server for in-depth protein knowledge and analysis. Nucleic Acids Res. 2003;31(13):3784–3788. pmid:12824418
- 39. Higgins DG, Thompson JD, Gibson TJ. Using CLUSTAL for multiple sequence alignments. Methods Enzymol. 1996;266:383–402. pmid:8743695
- 40. Bailey TL, Williams N, Misleh C, Li WW. MEME: discovering and analyzing DNA and protein sequence motifs. Nucleic Acids Res. 2006;34 Suppl 2:369–373.
- 41. Eisenhaber B, Bork P, Eisenhaber F. Prediction of potential GPI-modification sites in proprotein sequences. J Mol Biol. 1999;292(3):741–758. pmid:10497036
- 42. Paterson AH, Brubaker CL, Wendel JF. A rapid method for extraction of cotton (Gossypium spp.) genomic DNA suitable for RFLP or PCR analysis. Plant Mol Biol Rep. 1993;11(2):122–127.
- 43. Jiang JX, Zhang TZ. Extraction of total RNA in cotton tissues with CTAB-acidic phenolic method. Cotton Sci. 2003;15(3):166–167.
- 44. Drenkard E, Richter BG, Rozen S, Stutius LM, Angell NA, Mindrinos M, et al. A simple procedure for the analysis of single nucleotide polymorphisms facilitates map-based cloning in Arabidopsis. Plant Physiol. 2000;124(4):1483–1492. pmid:11115864
- 45. Livak KJ, Schmittgen TD. Analysis of relative gene expression data using real-time quantitative PCR and the 2-△△CT method. Methods. 2001;25(4):402–408. pmid:11846609
- 46. Xu YH, Wang JW, Wang S, Wang JY, Chen XY. Characterization of GaWRKY1, a cotton transcription factor that regulates the sesquiterpene synthase gene (+)-delta-cadinene synthase-A, Plant Physiol. 2004;135(1):507–515 pmid:15133151
- 47. Zeng CL, Wang GY, Wang JB, Yan GX, Chen BY, Xu K, et al. High-throughput discovery of chloroplast and mitochondrial DNA polymorphisms in brassicaceae species by ORG-EcoTILLING. PLoS One. 2012;7(11):e47284. pmid:23185237
- 48. Comai L, Oung K, Till BJ, Reynolds SH. Efficient discovery of DNA polymorphisms in natural populations by Ecotilling. Plant J. 2004;37(5):778–786. pmid:14871304
- 49. Zhang J. Evolution by gene duplication: an update. Trends in Ecology & Evolution 2003;18 (6): 292–8.
- 50. Richmond TA, Somerville CR. The cellulose synthase superfamily. Plant Physiol. 2000;124(2):495–498. pmid:11027699
- 51. Wang BH, Guo WZ, Zhu XF, Wu YT, Huang NT, Zhang TZ. QTL mapping of fiber quality in an elite hybrid derived–RILs in upland cotton. Euphytica. 2006;152(3):367–378.
- 52. Wang BH, Wu YT, Guo WZ, Zhu XF, Huang NT, Zhang TZ. Genetic dissection of heterosis for fiber qualities in an elite cotton hybrid grown in second–generation. Crop Sci. 2007;47:1384–1392.
- 53. Qin YS, Ye WX, Liu RZ, Zhang TZ, Guo WZ. QTL mapping for fiber quality properties in upland cotton (Gossypium hirsutum L.). Sci Agric Sin. 2009;42(12):4145–4154.
- 54. Chen H, Neng Q, Guo WZ, Song QP, Li BC, Deng FJ, et al. Using three overlapped RILs to dissect genetically clustered QTL for fiber strength on chro.24 in upland cotton. Theor Appl Genet. 2009;119(4): 605–612. pmid:19495722
- 55. Shen XL, Guo WZ, Zhu XF, Yuan YL, Yu JZ, Kohel RJ, et al. Molecular mapping of QTLs for qualities in three diverse lines in upland cotton using SSR markers. Mol Breed. 2005;15(2):169–181.
- 56. Yang C, Guo WZ, Zhang TZ. QTL mapping for resistance to vertieillium wilt, fiber quality and yield traits in upland cotton (Gossypium hirsutum L.). Mol Plant Breed. 2007;5(6): 797–805.
- 57. Ma X, Ding Y, Zhou B, Guo W, Lv Y, Zhu X, et al. QTL mapping in A-genome diploid Asiatic cotton and their congruence analysis with AD-genome tetraploid cotton in genus Gossypium. J Genet Genomics. 2008;35(12):751–62. pmid:19103431
- 58. Zhang TZ, Qian N, Zhu XF, Chen H, Wang S, Mei HX, et al. Variations and transmission of QTL alleles for yield and fiber qualities in upland cotton cultivars developed in China. PLoS One. 2013;8(2):e57220. pmid:23468939
- 59. Wang P, Zhu YJ, Song XL, Cao ZB, Ding YZ, Liu BL, et al. Inheritance of long staple fiber quality traits of Gossypium barbadense in G. hirsutum background using CSILs. Theor Appl Genet. 2012;124(8):1415–28. pmid:22297564
- 60. Zhang HB, Li YN, Wang BH, Chee PW. Recent advances in cotton genomics. Int J Plant Genomics. 2008;2008:742304. pmid:18288253
- 61. Renny-Byfield S, Gallagher JP, Grover CE, Szadkowski E, Page JT, Udall JA, et al. Ancient gene duplicates in Gossypium (cotton) exhibit near-complete expression divergence. Genome Biol Evol. 2014;6(3):559–571. pmid:24558256
- 62. Schindelman G, Morikami A, Jung J, Baskin TI, Carpita NC, Derbyshire P, et al. COBRA encodes a putative GPI-anchored protein, which is polarly localized and necessary for oriented cell expansion in Arabidopsis. Genes & Dev. 2001;5(9):1115–1127.
- 63. Betancur , Singh B, Rapp RA, Wendel JF, Marks MD. Phylogenetically distinct cellulose synthase genes support secondary wall thickening in Arabidopsis shoot trichomes and cotton fiber. J Integr Plant Biol. 2010;52(2):205–220. pmid:20377682