GhMYB1 regulates SCW stage-specific expression of the GhGDSL promoter in the fibres of Gossypium hirsutum L.

Secondary cell wall (SCW) biosynthesis is an important stage of the cotton fibre development, and its transcriptional regulation is poorly understood. We selected the Gossypium hirsutum GDSL (GhGDSL) lipase/hydrolase gene (CotAD_74480), which is expressed during SCW biosynthesis (19 through to 25 days postanthesis; DPA), for study. T1 -transgenic cotton lines expressing the β-glucuronidase (gus) reporter under the control of a 1026-bp promoter fragment of GhGDSL (PGhGDSL ) showed 19 DPA stage-specific increase in GUS expression. 5' deletion indicated that the 194-bp fragment between -788 and -594 relative to the transcription start site was essential for this stage-specific expression. Site-directed mutagenesis of eight transcription factor binding sites within PGhGDSL demonstrated that the MYB1AT motif (AAACCA) at -603/-598 was critical for the 19 DPA-specific reporter gene expressions. Yeast one-hybrid (Y1H) analysis identified nine proteins, including GhMYB1 (CotAD_64719) that bound to the PGhGDSL promoter. Further, Y1H experiments using the 5' promoter deletions and individually mutated promoter motifs indicated that GhMYB1 interacted with PGhGDSL at MYB1AT sequence. GhMYB1 was expressed specifically in fibre from 19 DPA, overlapping with the sharp rise in GhGDSL expression, indicating that it could regulate GhGDSL during fibre development. Analysis of genes co-expressed with GhMYB1 showed that it potentially regulates a number of other 19-25 DPA-specific genes in networks including those functioning in the cell wall and precursor synthesis, but not the major polysaccharide and protein components of the fibre SCW. GhGDSL and its promoter are therefore potential tools for the improvement of cotton fibre quality traits.

Introduction

Cotton fibre is a single‐celled structure initiated on the epidermal surface of an ovule and matures over four developmental stages, namely initiation, elongation, SCW synthesis and maturation (Basra and Malik, 1984). A single mature cotton fibre consists of the thin outer primary cell wall (PCW), an inner thick SCW and a central lumen. The most desirable agronomical traits of cotton fibres are its length, strength and fineness. The SCW controls the strength and fineness of fibres through the extent of synthesis and deposition of cellulose, the main polysaccharide component of the SCW. Among the different developmental stages of the fibre, initiation is the best understood at the transcriptional level. Although several transcription factor types are reported to play a role in cotton fibre development, the MYB transcription factors are among the most prolific. GhMYB25, GaMYB2 and GhMYB109 have been shown to control steps in fibre initiation and elongation (Deng et al., 2011; Pu et al., 2008; Walford et al., 2011; Wang et al., 2004). Cotton fibre SCW biosynthesis is characterized by the expression of genes such as GhGlcAT1 (Wu et al., 2007) and GhRLK1 (Li et al., 2005b), but predominantly the stage‐specific cellulose synthases such as GhCesA4, GhCesA7 and GhCesA8 (Tuttle et al., 2015). Recently, it has been reported that GhKNL1 (KNOTTED1‐LIKE) is also expressed during SCW deposition (Gong et al., 2014). Although there are several genes that are expressed specifically during SCW synthesis, the transcription regulation of TFs involved in cotton fibre SCW biosynthesis still lacks detail investigation. However, MYB1 (R2R3‐MYB) has been shown to regulate SCW biosynthesis in Pinus taeda (Bomal et al., 2008) and cell wall integrity in fungi (Dong et al., 2015) so may be a potential candidate for a role in fibre SCW synthesis.

GDSL lipases are a very large gene family belonging to SGNH superfamily in plants (Akoh et al., 2004) and are conserved from microbes through to plants. More than a thousand members of GDSL lipase have been reported, but their functional roles in plant development and physiology are poorly explored. Trichome birefringence (TBR) in Arabidopsis possesses a GDSL esterase/lipase domain and plays a role in the cellulose deposition and the synthesis of SCW through the esterification of pectin compounds (Bischoff et al., 2010). In Pseudomonas aeruginosa, these lipases hydrolyse the esters of long‐chain fatty acids and acylglycerols (Wilhelm et al., 1999). In plants, GDSL lipases are involved in many developmental processes (Takahashi et al., 2009), seed germination morphogenesis (Clauss et al., 2008; Katavic et al., 2006; Kondou et al., 2008), defence (Kwon et al., 2009; Oh et al., 2005), abiotic and biotic stress (Hong et al., 2008; Kim et al., 2008) and responses to the hormones ethylene and auxin (Lee et al., 2009). GDSL lipases are expressed in cotton fibres (Nigam et al., 2013); however, their exact role in fibre development is not yet known.

Greater understanding of the genes, upstream regulatory elements and transcription factors that regulate fibre‐specific gene expression will be indispensable for enhancing fibre traits in the longer term. Genetic engineering for better fibres (longer, stronger and finer) will require stage‐specific promoters for manipulating those fibre properties. Several promoters have already been reported that regulate gene expression in cotton fibres during the initiation and elongation stages (Delaney et al., 2007; Hussey et al., 2011; Larkin et al., 1996; Li et al., 2002, 2009; Ma et al., 1995; Ni et al., 2008; Song and Allen, 1997; Wang et al., 2004; Wu et al., 2009). The promoter of GaRDL1, GhTUB1 and GhMYB25, for example, is active early during the initiation stage (Li et al., 2002; Machado et al., 2009; Wang et al., 2004), while the glucuronosyltransferase promoter (PGhGlcAT1) has been shown to be highly active during fibre elongation (Wu et al., 2007). The fibre‐specific lipid transfer protein (FSltp4) promoter has also been reported to be active during fibre elongation, and its upstream gene has been shown to be functional during the synthesis of fibre cutin (Delaney et al., 2007). GhACTIN1 is another gene expressed in fibre, and its promoter has been shown to be elongation‐specific (Li et al., 2002). Promoters for the genes, GhLTP3, GhDET2, GaMYB2 and GhMYB109, have also been reported to be active during initiation and some of them continue to be active until the elongation stage (Liu et al., 2000; Luo et al., 2007; Pu et al., 2008; Wang et al., 2004). There are other promoters, such as PGhRING1, that are active in all the stages of cotton fibre development, starting from initiation through to the SCW stage (Ho et al., 2010). However, there are no reports of a truly SCW stage‐specific promoter and its detailed molecular characterization.

The present work describes the cloning and characterization of promoter regulatory sequences of SCW biosynthesis stage‐specific GDSL lipase/hydrolase from cotton. Our analysis demonstrates that PGhGDSL is regulated by the GhMYB1 transcription factor that interacts with PGhGDSL at a MYB1AT motif. We also propose that MYB1 is involved in the regulation of a broader gene network that is expressed during the SCW biosynthesis from around 19 DPA, so these genes and promoters provide potential new targets for engineering to improve fibre quality attributes.

Results

GhGDSL has SCW stage‐specific expression during cotton fibre development

Our previous microarray data (Nigam et al., 2013) based on five genotypes and six cotton development stages suggested that a GhGDSL (Ghi.8746.2.A1_x_at) was preferentially expressed during SCW deposition stage (19 DPA and 25 DPA) of fibre development (Figure 1a). We decided to use GhGDSL promoter to delineate the potential gene regulatory networks that might controls SCW formation. The expression of GhGDSL was further verified by qRT‐PCR using cDNA samples prepared from 0, 3, 6, 9, 12, 15, 19 and 25 DPA stages fibres and from leaf, stem, root, buds and cotton boll coat to confirm its specificity. The qRT‐PCR showed fibre‐specific expression of GhGDSL with negligible expression in all the other tissues tested. Expression of GhGDSL was significantly higher at the 19 and 25 DPA stages (Figure 1b), similar to that observed in the microarrays. These results confirmed the fibre and SCW stage‐specific expression of GhGDSL. We also verified that the 19 and 25 DPA stages actually represent stages of SCW deposition by measuring the total cellulose content in each of the different fibre samples and saw significant cellulose deposition after 15 DPA (Figure S6). The phylogenetic analysis of GhGDSL (gene id CotAD_74480) with the known GDSL genes in the Arabidopsis genome demonstrated significant homology with the Arabidopsis APG2.ARATH GDSL lipase/hydrolase protein (Figure S7).

Figure 1

Sequence of the Gh promoter and validation of its stage‐specific expression. (a) Microarray expression profile (average of five genotypes: 703, 737, 783, 777 and 725) of Gh gene (Ghi.8746.2.A1_x_at) at the different cotton fibre development stages (0, 6, 9, 12, 19 and 25 DPA) using cotton Affymetrix chip data. (b) Quantitative RT‐PCR analysis of Gh gene expression in the cotton fibre development stages (0–25 DPA): the normalization of the gene expression was carried out using the cotton ubiquitin gene. The various stages of fibre development are indicated in DPA (day post anthesis). The asterisk indicates statistical analysis by t‐test (*P‐values <0.05 and **P‐values <0.01). The error bars represent ±SE (standard error) of three independent repeats. (c) Sequence analysis of P h : The putative cis‐regulatory elements (REs) are underlined. Transcription start site (TSS) is marked as +1 and –603/–598, –600/–594, –558/–553, –450/–445, –418/–413, –226/–219, –67/–62, –11/–6 represent the position of cis‐REs on the promoter.

We then cloned a 1026‐nucleotide (–951/+75 with respect to the ATG) long PGhGDSL promoter fragment using genome walking (Figures 1c, S4). The presence of the primers and the sequence of an overlapping part of the coding region of GhGDSL ensured that the cloned fragment belonged to the upstream promoter region of the same gene identified from the microarray. Genomewide BLAST against the Gossypium hirsutum genome (AD) further confirmed that both the promoter and the gene aligned at the same region on chromosome 4 (Table S1). The promoter fragment was fused to the gus gene in pBI101 to develop a reporter construct that was mobilized into Agrobacterium tumefaciens and used to generate several independent transgenic lines of cotton. We evaluated ten independent T1‐transgenic lines for both quantitative and histochemical expression of the gus reporter at different stages of fibre development. Histochemical staining of GUS indicated weak expression at 0 DPA as shown by weak blue staining of the ovules (Figure 2a). Expression increased gradually from 0 DPA and reached a maximum at 19 DPA (Figure 2a). The expression was still high even at 25 DPA, but was lower than at 19 DPA. Thin sections of stained ovules at 0, 6 and 19 DPA confirmed the SCW stage‐specific expression of GUS localized within the fibres and the ovule epidermis at 19 DPA. GUS staining was not observed in any other tissues, such as in seedling, root, leaf and bracts (Figure S1). Further, quantitative GUS expression in the ten independent lines, although showing minor differences among lines, was consistently highest at 19 DPA (Figure 2b), coinciding with the pattern of histochemical staining. None of the transgenic lines showed any significant expression in leaves used as controls (Figure 2b). We also developed independent transgenic cotton lines with gfp fused to the native GhGDSL promoter. There was weak GFP fluorescence at 0 DPA and significantly higher fluorescence at 19 DPA (Figure S2), consistent with the other measures of expression of this gene. Similar to the GUS reporter, no significant GFP expression was observed in any other tissues of the transgenic lines examined (not shown). The analyses in transgenic cotton further validated the SCW stage‐specific expression of GhGDSL and indicated that PGhGDSL should be a good candidate to explore the regulatory networks involved in SCW formation during fibre development.

Figure 2

GUS expression analysis in cotton fibre driven by Gh promoter. (a) Histochemical GUS staining of P h :gus expression in T1‐transgenic lines and nontransgenic control lines in individual bolls and ovules at the different development stages (0–25 DPA). GUS staining was absent in the nontransgenic cotton bolls and ovules (Scale bar = 0.635 cm). (b) Fluorometric GUS expression analysis of P h :gus in ten independent cotton transgenic lines: histochemical GUS analysis was performed as described by Jefferson et al. (1987). The expression values were normalized using the non‐transgenic control values. The results are the average of three biological replicates of T1 lines of the ten independent lines: the asterisk represents statistical t‐test between transgenic lines at different DPA compared to 0DPA (*P‐values <0.05 and **P‐values <0.01). The error bars represent ±SE of three independent biological replicates.

The cis‐regulatory elements controlling the SCW stage‐specific expression of GhGDSL

To delineate the cis‐regulatory sequences involved in regulating GhGDSL, we generated five progressive 5′ deletions of the native full‐length promoter PGhGDSL–951 (FLP) that were designated as: PGhGDSL–788 (D1), PGhGDSL–594 (D2), PGhGDSL–403 (D3), PGhGDSL–221 (D4) and PGhGDSL–35 (D5) (Figure 3a). Each of the promoter deletions was fused to the gus reporter gene and the constructs transformed into cotton. GUS activity was analysed in the ovules of 7–10 independent T1‐transgenic lines per construct at different fibre development stages. The first deletion construct D1 showed a significant reduction in GUS expression at 15 DPA, but much less at 19 DPA compared to the native PGhGDSL–951 (Figure 3b‐c) and was not significantly different to the full‐length promoter at other developmental stages. Interestingly, the second deletion construct D2 showed no significant change in the GUS expression at most developmental stages except for 19 DPA, where expression in the ovules of all the transgenic lines was significantly reduced relative to PGhGDSL–951. These results indicated that the cis‐regulatory elements, which are present between –788 and –594 nucleotides of the native promoter, regulate 19 DPA stage‐specific expression of GhGDSL. The expression of the D3 deletion construct was significantly lower at 15 and 19 DPA as compared to PGhGDSL–951. However, its expression at 19 DPA was significantly higher compared to the D2 construct. Similarly, the expression of the D4 construct was significantly lower than PGhGDSL–951 but higher than D2 at 19 DPA (Figure 3c). Interestingly, the expression of construct D4 was significantly higher than all the constructs, including the full‐length construct PGhGDSL–951 at 12 DPA. The final deletion construct D5 did not show any significant expression at any of the fibre development stages. However, this was expected considering that it contained only minimal promoter sequences. The results for D5 also confirmed that the regulatory region upstream of 110 nucleotide (–35/+75) had all the regulatory sequences necessary for 19 DPA fibre‐specific expression. Furthermore, the expression pattern of PGhGDSL–951, D2 and D5 was confirmed by staining the ovules of the transgenic lines at different development stages. The histochemical staining was concordant with the quantitative GUS estimation, confirming the 19 DPA elevated expression of PGhGDSL–951 (Figure 3b). D2 showed a reduced staining at all the development stages, especially at 19 DPA confirming that the deletion of the promoter up to –594 results in the loss of 19 DPA‐specific expression of GhGDSL. Ovules of the D5 construct did not show GUS staining at any stage.

Figure 3

Deletion analysis of the Gh promoter. (a) Schematics of P h deletions D1, D2, D3, D4 and D5 along with full‐length promoter. +1 represents TSS. Scale bar represents the length of promoter (–951/+75) from the start codon ATG. (b) Histochemical GUS analysis of the cotton T1‐transgenic lines carrying a deletion construct: the upper panel shows the GUS staining of a transgenic cotton flower and bolls of the native promoter reporter (P h ) during the development stages (0–25 DPA). The middle panel shows the GUS staining of a transgenic cotton flower, boll or ovule carrying P h –594:gus (D2) construct. The lower panel indicates that there was no staining in a transgenic cotton flower, boll or ovule with P h –35:gus (D5) construct. (c) Fluorometric GUS expression analysis of transgenic cotton lines carrying deletion constructs (D1‐D5). The histochemical GUS analysis was performed as described by Jefferson et al. (1987). The expression values are normalized with the control values. The results are the average of three biological replicates for each of 7–10 independent T1‐lines: statistical t‐test between deletion lines and full‐length promoter (*P‐values <0.05 and **P‐values <0.01). The error bars represent ±SE.

MYB1AT motif regulates the 19–25 DPA‐specific expression of GhGDSL

Computational analysis of conserved cis‐regulatory elements within PGhGDSL–951 using PLACE (www.dna.affrc.go.jp) revealed several conserved motifs in the promoter (Table S2), including some MYB binding motifs such as WAACCA (MYB1AT), CACATG (MYBCORE), GGATA (MYBST1), and bHLH and homeodomain binding motifs such as CACATG (MYCATRD22), TAAATGYA (L1BOX), CATTTG (MYC CONSENSUS) and CACATG (MYCATRD22) being present at –603/–598, –600/–594, –558/–553, –450/–445, –226/–219, –67/–62 and –11/–6 positions, respectively (Figure 1c). MYB transcription factors have been reported to play an important role in cotton fibre development, so we decided to examine their contribution by specifically mutating these MYB and other conserved motifs. Eight mutational constructs, designated as sdm1 to sdm8, were generated by site‐directed mutagenesis of native promoter PGhGDSL–951 (Figure 4a). Each of the mutant versions of the promoter was fused to the gus gene and transformed in cotton to produce multiple transgenic lines for each construct. At least 7–10 T1‐transgenic lines were evaluated for GUS expression in each construct and each of the different development stages of cotton fibre. None of the transgenic lines showed any significant change in quantitative GUS expression from 0 DPA to 12 DPA compared to the native promoter PGhGDSL–951 (Figure 4c). However, at 15 DPA, sdm2, sdm3, sdm7 and sdm8 mutations within the MYBCORE, MYBST1, MYCCONSENSUS and MYCATRD22, respectively, showed a significant reduction in GUS activity compared to PGhGDSL–951 (Figure 4c). The most striking differences were at the 19 DPA stage, where the sdm1 mutant of the MYB1AT motif almost completely abolished GUS activity. This result indicated that the SCW stage‐specific 19 DPA expression of GhGDSL was regulated through the MYB1AT motif. The sdm2, sdm7 and sdm8 mutants also showed a significant reduction in GUS activity compared to PGhGDSL–951 at 19 DPA, but their expression was still much higher than the sdm1 construct. In comparison with PGhGDSL–951, no major change in the expression of any of the site‐directed mutants was observed at 25 DPA (Figure 4c). Histochemical staining for GUS gave similar results to the quantitative analyses, and only weak staining of cotton bolls was seen for constructs, sdm1 and sdm7 (Figure 4b).

Figure 4

Mutation analysis of Gh promoter. (a) Schematics of mutated motifs present in P h : the site‐specific mutations in promoter designated as sdm1 (MYB1AT‐AACCCC), sdm2 (MYBCORE‐CACGTG), sdm3 (MYBST1‐GGACA), sdm4 (MYCATRD22‐CCCAGG), sdm5 (HEX‐motif‐ATGCCG), sdm6 (L1BOX‐TCAAATGT), sdm7 (MYCCONSENSUS‐CCTTGG) and sdm8 (MYCATRD22‐CCCATG) are highlighted in red. (b) Histochemical GUS analysis of T1‐transgenic cotton lines carrying mutated motifs (sdm1‐sdm8): GUS staining of a transgenic cotton bolls carrying native P h and each of the promoters with the mutated motifs such as MYB1AT, MYBCORE, MYBST1, MYCRD22, HEX‐motif, L1BOX‐TCAAATGT, MYC CONSENSUS and MYCATRD22 at 19 DPA. (c) Fluorometric GUS expression analysis of transgenic cotton lines carrying the promoter with mutated motif constructs (sdm1‐sdm8). The expression values are corrected with the control values. The results are the average of three biological replicates each of 7–10 independent T1 lines: statistical t‐test between the mutation lines (sdm1‐sdm8) and full‐length promoter (*P‐values <0.05 and **P‐values <0.01). The error bars represent ±SE.

GhMYB1 interacts with the MYB1AT motif of the GhGDSL promoter

We next identified putative promoter binding proteins that bind to PGhGDSL–951 using the yeast one‐hybrid (Y1H) system. PGhGDSL–951 was cloned upstream of the Aureobasidin resistance gene (AbA ) and used as a bait to screen for potential transcription factors that could activate this reporter in yeast. We used a cDNA library prepared from pools of different stages of fibre development (0, 3, 6, 9, 15, 19 and 25 DPA) fused to the Gal4 activation domain (AD) for screening. Positive interacting clones were selected on SD medium lacking leucine and containing AbA and the promoter binding proteins identified by sequencing the isolated plasmids from resistant colonies. Nine putative binding proteins (Table S3) that might interact with PGhGDSL–951 with a significant affinity, as confirmed by dilution (Figure 5a), were identified. Interestingly, out of the nine, three belong to MYB family, viz. an R2R3‐MYB, MYB10 and MYB1. As our previous deletion and mutational experiments clearly suggested that the MYB1AT motif regulates the 19 DPA stage‐specific expression of GhGDSL and GhMYB1 was here shown to bind to PGhGDSL–951, we confirmed this interaction through further experimentation. All of the promoter deletion constructs and the site‐directed mutagenesis constructs were cloned into the Y1H vectors and their interaction with GhMYB1 fused to Gal4‐AD assessed, as for the full‐length promoter. The D1 deletion showed similar binding to GhMYB1 as the native construct (Figure 5b). However, the successive deletions, D2, D3, D4 and D5, did not show any significant binding to GhMYB1 (Figure 5b). This observation clearly indicated that the binding site for GhMYB1 lies between –788 and –594. Mutants sdm2 to sdm8 did not show any significant difference in their binding compared to the native promoter (Figure 5c); however, the mutant of MYB1AT, that is sdm1, showed complete loss of binding, as no growth on selection medium was observed at any dilution tested (Figure 5c). Thus, we confirmed that GhMYB1 does interact with PGhGDSL–951 at its MYB1AT binding sites that lies between –788 and –594.

Figure 5

Yeast one‐hybrid interaction of nine TFs and the interaction analysis of GhMYB1 with deletion and mutated motif constructs (sdm1‐sdm8) of the Gh promoter. (a) The interaction analysis of nine TFs (RGA/GAI3, R2R3‐MYB, MYB1, GASL2, MYB10, CO‐like4, CNGC2, PDF1 and FDH) on SD/‐Leu/AbA agar plates. Columns are undiluted, 1:10‐, 1:100‐ and 1:1000‐fold dilutions, respectively. SD/‐Leu/AbA represents the selection of positively interacting clones on Leucine drop out medium with AbA antibiotic selection (100 ng/mL). (b) GhMYB1 interaction with P h deletions (D1, D2, D3, D4 and D5) construct. GhMYB1 binds between the −788 and −594 region of P h . Yeast colonies were not observed with D3–D5. Columns are undiluted, 1:10‐, 1:100‐ and 1:1000‐fold dilutions, respectively. (c) The interaction analysis of MYB1 with the promoter containing mutated promoter motifs confirmed that GhMYB1 binding with P h at the MYB1AT motif sequence AAACCA. No yeast colony was observed with the mutated motif sequence AAACCA (MYB1AT) of P h but binding was observed for all other mutated motifs.

GhMYB1 is expressed explicitly during SCW synthesis at 19 DPA stage of cotton fibre development

A transcription factor needs to be expressed at a particular development stage in order to regulate a gene specific to that stage. qRT‐PCR expression analysis was carried out for all nine of the identified putative promoter binding proteins using RNA from each of the different stages of fibre development. Many of these putative promoter binding proteins were more highly expressed during fibre development than in leaves (Figure 6). Only Co‐like4 was expressed at a significantly higher level in leaves than fibre. The transcription factors, such as RGA/GAI3, R2R3‐MYB, GASL2, CNGC2 and PDF1 (Figure 6), had high expression during 0 or 6 DPA so are fibre initiation stage‐specific genes. The FDH (Figure 6) showed significantly higher expression at 9 DPA indicating that it was elongation stage‐specific. Interestingly, only GhMYB1 showed a 19 DPA stage‐specific expression (Figure 6) similar to that seen for GhGDSL (Figure 1b). GhMYB1 is clearly co‐expressed with GhGDSL at the SCW stage consistent with its likely role in controlling SCW stage‐specific expression of GhGDSL and potentially other genes. It is also noteworthy that MYB10 had high expression at 19 DPA, but its expression appeared to be even higher at 25 DPA (Figure 6).

Figure 6

Quantitative real‐time PCR (qRT‐PCR) expression analysis of the nine putative promoter binding proteins identified from Y1H interaction with Gh promoter. Expression analysis of RGA/GAI3, R2R3‐MYB, MYB1, GASL2, MYB10, CO‐like4, CNGC2, PDF1 and FDH at different fibre development stages (0–25 DPA): gene expression was normalized with the cotton ubiquitin gene. The error bars represent ±SE of three independent biological replicates. Statistical t‐test is represented by the asterisk (*P‐values <0.05 and **P‐values <0.01).

GhMYB1 co‐expresses with SCW stage‐specific genes during fibre development

Genes that are co‐expressed with GhMYB1 during cotton fibre development from publically available RNA‐seq data of 0, 5, 10, 20 and 25 DPA were identified using the expression correlation network plug‐in available in Cytoscape version 2.8.1 (Shannon et al., 2003). A total of 118 positively co‐expressed genes (r ≥ 0.95) and 435 negatively co‐expressed genes (r ≤ −0.95) were identified (Appendix S1; Figure 7a). The expression levels (based on normalized read counts) of the positively and negatively co‐expressed genes at each fibre developmental stage were assessed (Figure 7b). Genes positively co‐expressed with GhMYB1 had significantly higher levels of expression during the later stages of fibre development, especially at 25 DPA, while those negatively co‐expressed with GhMYB1 had lower expression at 25 DPA relative to earlier time points (Figure 7b). Our results therefore suggest that GhMYB1 may be regulating genes expressed specifically at the SCW stages. We examined the gene ontology classifications of both the positively and negatively co‐expressing genes using MapMan visualization and statistical analysis tools. This revealed that the positively co‐expressing genes were enriched in the cell wall category including some genes involved in the synthesis of cell wall polysaccharides precursors (Figure 7c), although the major biosynthetic gene (like cellulose synthase) for the cell wall polysaccharides was not indicated (Appendix S1). The positively expressing genes were also enriched for genes involved in carbohydrate metabolism and secondary metabolism of isoprenoid and phenylpropanoid compounds (although again not the major pathway enzymes) that are minor components of cell wall (Figure S3). The presence of MYB1AT binding motif (Figure 7d) in the promoters (1000 nucleotide upstream of TSS) of the positively co‐expressed, negatively co‐expressed, and 150 randomly selected genes, as the control, was examined. The control data set showed a random frequency of promoters containing MYB1 motifs at 0.61, while positively and negatively regulated genes showed frequencies of 0.84 and 0.78, respectively, which were significantly higher than expected by chance. These results suggest that GhMYB1 could be up‐regulating the positively co‐expressed genes by binding to the MYB1AT motifs in their promoters. GhMYB1 might also be interacting with the negatively co‐expressed genes, although further detailed experimentations would be needed to confirm a repressive role for GhMYB1 during fibre development.

Figure 7

Gene co‐expression network analysis of Gh. (a) Unweighted gene co‐expression network of positive and negative partners of Gh in different fibre development stages. Circles (nodes) represent transcripts, while the lines (edges) represent a significant transcriptional interaction between Gh and the other transcripts. (b) Box plot showing the expression level of the positively and negatively co‐expressing partners of Gh at different fibre development stages. (c) MapMan analysis revealed that Gh co‐expresses positively with the genes involved in the biosynthesis of the cell wall and its precursors during the fibre SCW stages. (d) Graph showing the frequency of MYB1AT recognition motif (WAACCA) in the promoters of the positively and negatively interacting partners and a randomly selected set of expressed genes. The asterisks represent significant differences (Fisher's exact test, P < 0.05).

Discussion

SCW biosynthesis is a critical stage of fibre development, characterized by rapid deposition of cellulose and is known to influence fibre quality traits such as length, strength and micronaire in cotton. Despite the plethora of information available about cotton fibre initiation and its regulation, very little is known about the transcriptional regulation of SCW biosynthesis in cotton. To decipher the transcription regulation of the SCW, we cloned and characterized the SCW biosynthesis stage‐specific PGhGDSL in G. hirsutum. PGhGDSL was selected on the basis of our previous work (Nigam et al., 2013) on expression profiling of five genotypes of cotton at six different development stages, where GhGDSL was found to be one of the significantly up‐regulated genes during fibre SCW deposition stages (19 DPA and 25 DPA; Figure 1a). Quantitative RT‐PCR confirmed that GhGDSL is expressed at very high level during SCW stage (Figure 1b). GDSL lipase is known to play important roles in the cellulose deposition and the synthesis of the secondary cell wall in other plant species and also alters pectin composition by the esterification of pectin compounds in Arabidopsis (Bischoff et al., 2010).GDSL lipase has been found to be involved in several plant processes, but its exact role in cotton fibre development is still elusive. The fibre SCW stage‐specific expression of GhGDSL does implicate a possible role for this enzyme in SCW biosynthesis. The gus gene driven by PGhGDSL–951 in transgenic cotton lines showed a very high level of expression of GUS from 19 DPA and confirmed its SCW stage‐specific expression during cotton fibre development. The expression of PGhGDSL–951 began around 6 DPA and reached its maximum level at 19 DPA. No background expression of PGhGDSL–951 in any other floral part, or other tissues was observed in any of the transgenic lines, indicating that PGhGDSL is strictly fibre‐specific. Fibre‐specific expression (mostly initiation and elongation stage‐specific) of other promoters, such as GhGlcAT1 (Wu et al., 2007), FSltp4 (Delaney et al., 2007), GhACTIN1 (Li et al., 2005a), GhLTP6, GaMYB2 and GhMYB109 (Hsu et al., 1999; Pu et al., 2008; Wang et al., 2004), GhRING1 (Ho et al., 2010) GaRDL1, GhTUB1 and GhMYB25 (Li et al., 2002; Machado et al., 2009), has been reported previously. However, to the best of our knowledge, there are no other SCW‐specific promoters characterized in any detail so far in cotton. SCW stage‐specific expression of genes, such as GhRLK1 (Li et al., 2005b), GhKNL1 (Gong et al., 2014) and GhCesA (Fagard et al., 2000; Kim and Triplett, 2001), has been reported, but not studied in the way described here in transgenic cotton plants. The 5′ deletion experiments revealed that the major cis‐regulatory element that control 19 DPA‐specific expression of GhGDSL lie within the 194‐bp regions between –788 and –594 nucleotide upstream of the transcription start site, as deletion D2 completely abolished the expression of GhGDSL at 19 DPA (Figure 3b‐c). The deletion constructs D3 and D4 showed higher expression than D2, indicating that the absence of these upstream elements allows other downstream regulatory elements to become accessible, although these still need to be characterized in detail. The upstream regulator(s) also seem to be essential for the control of expression at 15 DPA and 12 DPA, because the expression of D2 is lower than the full‐length promoter at both these stages. At 25 DPA, however, the regulators within the 194 nucleotide region do not appear to be essential, as the expression of D2 is almost identical to that of the full‐length control at 25 DPA. The deletion experiments suggest that the 194 nucleotide region controls the expression of GhGDSL from elongation through to the SCW biosynthesis stage.

MYB transcription factors are known to play a major role during the different stages of cotton fibre development. GhMYB109, for example, is expressed during fibre elongation in G. hirsutum (Pu et al., 2008), while GhMYB25, GaMYB2 and GhMYB25 like have been identified to play key roles during initiation and early elongation stages (Deng et al., 2011; Walford et al., 2011; Wang et al., 2004). In the model plant Arabidopsis, it has been reported that MYBs play a role in trichome initiation, pattering, trichrome cell fate determination and are also involved in cell wall thickening (Oppenheimer et al., 1991; Zhong et al., 2008). Thus, we decided to examine the presence of conserved cis‐regulatory elements in the promoter of GhGDSL, particularly of MYB TF binding sites. PGhGDSL had eight potential cis‐acting promoter elements, ranging between 11 and 603 nucleotides upstream of the TSS. The independent mutagenesis of each of these motifs showed that the mutation of MYB1AT at 603 nucleotide led to a significant loss of expression of PGhGDSL at 15 and 19 DPA (Figure 4c). A mutation in the MYBCORE, that overlaps the MYB1AT elements, also resulted in a significant loss of expression of PGhGDSL at 15 and to a lesser extent at 19 DPA. However, the MYB1AT mutation seemed to completely abolish the expression of the promoter at 19 DPA, indicating that it is the major element responsible for the regulation of GhGDSL at that stages (Figure 4c). The MYB1AT element lies within the 194 nucleotide regulatory region identified by 5′ deletion. Thus, these observations support the regulation of PGhGDSL through the MYB1AT motif. The MYB1 recognition site was first reported in the promoter region of the ABA and drought responsive rd22 gene (Abe et al., 2003), and it was found to function as a cis‐acting element that regulates the expression of the rd22 gene. The fact that the MYB1AT motif regulates PGhGDSL led us to identify the transcription factors that bind to PGhGDSL. Y1H screening of a fibre‐specific expression library identified nine putative promoter binding proteins (Table S3; Figure 6). Six of these have their binding sites conserved within the PGhGDSL (Table S4) region, indicating that they interact specifically with it. Interestingly, GhMYB1 was identified as binding to PGhGDSL with a high affinity. Besides GhMYB1, two other MYB transcription factors, R2R3‐MYB and MYB10, also bound PGhGDSL with a high affinity, but the site(s) to which they bind need to be characterized further. We also identified other proteins, such as the repressor of gibberellic acid (RGA/GAI3) like factor, gibberellic acid stimulated like 2 (GASL2), cyclic nucleotide gated channel isoform 2 (CNGC2), the G. hirsutum protodermal factor 1 (GhPDF1), constans‐like 4 (GhCO‐like 4) and Fiddlehead‐like protein (FDH). Some of these, such as GhPDF1, are involved in cotton fibre initiation and elongation, while RGA/GAI3 like factors may play some role in gibberellic acid (GA) signalling in cotton fibre elongation, and GASL2 is involved in fibre elongation in response to GA signalling (Table S3).

As 5′ deletions and mutagenesis clearly indicated that MYB1AT is important in the 19 DPA‐specific expressions of PGhGDSL and Y1H also showed that GhMYB1 bound to this promoter, we further examined whether GhMYB1 interacts specifically with the MYB1AT element using Y1H. GhMYB1 failed to interact with any deletion beyond PGhGDSL–594, indicating that its binding site lies upstream of −594 (Figure 5b). Further, only the mutation in MYB1AT (sdm1) −603 relative to TSS resulted in a complete loss of interaction between GhMYB1 and PGhGDSL in the yeast (Figure 5c). Thus, both results strongly suggest that the regulation of PGhGDSL by GhMYB1 is due to its interaction with MYB1AT in the promoter. Further, unlike the other TFs that bound to PGhGDSL, GhMYB1 was expressed maximally from 19 DPA (Figure 6), suggesting that it not only regulates GhGDSL, but is potentially involved in the regulation of other genes expressed during the SCW stage of fibre development. Furthermore, we identified a suit of genes that are both positively and negatively co‐expressed with GhMYB1. Interestingly, the genes positively co‐expressed with GhMYB1 are generally expressed at significantly higher levels during the later stages of fibre development (20 DPA and 25 DPA; Figure 7b), while the genes that are negatively co‐expressed with GhMYB1 were more highly expressed at the initiation or elongation stages (Figure 7b). Some of the positively co‐expressed genes belong to the cell wall precursors functional categories (Figures 7c, S3), including a UDP glucose dehydrogenase that is required for pectin and hemicellulose synthesis and an alpha‐1,4‐glucan–protein synthase (UDP glucose forming) (UPTG, EC 2.4.1.112) that may have a role in cellulose synthesis and genes of the isoprenoid and phenylpropanoids pathways. Phenylpropanoids are an important but minor component of the secondary cell wall, and mature cotton fibre contains approximately 94% of cellulose and traces of what may be lignin like polymers (Fan et al., 2009). GhMYB1 is expressed precisely during the later stages of fibre development overlapping with SCW deposition, but the lack of the core cell wall polysaccharide biogenesis genes such as the SCW CesA, KORRIGAN and FLAs (Fasciclin‐like arabinogalactan proteins), etc., that are characteristics of most SCWs, suggests that GhMYB1 is regulating a specialized subcomponent of the SCW involving secondary metabolite synthesis, and stress hormone signalling‐related gene network. In accordance with the expression of positively and negatively co‐expressed genes with GhMYB1, we also identified that the promoters of both the groups of genes have significantly higher representation of the MYB1AT motif in their promoters (Figure 7d). A high occurrence of an MYB1AT motif in the positively co‐expressing genes with GhMYB1 implies that MYB1 is involved in the regulation of the genes that are similar to those we established for GhGDSL. However, there are a significant number of the negatively co‐expressing genes (435 genes), and interestingly, they also have a very high occurrence (0.78) of the MYB1AT motif. Thus, our results indicated that MYB1AT might be involved in the negative regulation of fibre initiation‐specific genes in the later stages of fibre development. However, the negative role of GhMYB1 in the regulation of initiation‐specific genes during the later stages of fibre development is just a speculation as at the moment we do not have any experimental evidence to support it.

In conclusion, the current study identified a SCW stage‐specific promoter PGhGDSL, which is activated during the later stages of fibre development. This promoter could be used for a variety of biotechnological applications in cotton. Furthermore, our study also indicates that a suit of other SCW biosynthesis stage‐specific genes are also regulated by GhMYB1, making it a target gene for the improvement of fibre quality traits in cotton varieties.

Plant materials

The cotton plants (G. hirsutum L. acc JKC725 a superior fibre quality genotype and Coker 312 genetic standard of G. hirsutum) were grown in a glasshouse at CSIR‐National Botanical Research Institute (NBRI), Lucknow, India. DNA and RNA were extracted from the cotton plants grown under standard field conditions. On the day of anthesis (0 DPA), cotton flowers were tagged and bolls were harvested for RNA isolation at 0, 6, 9, 12, 19 and 25 DPA stages. The harvested cotton tissues were quickly stored in ice, and the fibre from ovule was dissected followed by grounding in liquid nitrogen. An entire ovule was taken at 0 DPA for RNA isolation. Root and leaf samples were collected from 20‐day‐old seedlings.

Microarray and Real‐time PCR (RT‐PCR) analysis

The details are provided in the Appendix S2.

Genome walking and sequence analysis

The details are provided in Appendix S2.

Histochemical GUS staining and GUS assay

The details are provided in Appendix S2.

Microtomy and light microscopy

The details are provided in Appendix S2.

Fluorescence of GFP expressed in cotton fibre

The details are provided in Appendix S2.

Deletion analysis of promoter activity in transgenic cotton fibre

The sequence of full‐length PGhGDSL was used as a template for deletions. A forward primer with SalI and a reverse primer with BamHI restriction enzyme site were used to generate 5′ deletion constructs by PCR. The amplified PCR fragments were ligated into the plasmid pBI101 (Clontech, USA) with T4 DNA ligase (NEB) and the deletions constructs pBI‐PGhGDSL‐D:gus were named as D1 (pBI‐PGhGDSL–788:gus), D2 (pBI‐PGhGDSL–594: gus), D3 (pBI‐PGhGDSL–403:gus), D4 (pBI‐PGhGDSL–221:gus) and D5 (pBI‐PGhGDSL–35:gus). The transformation of the constructs in cotton was carried out using the embryo transformation method (Kumar et al., 2013). A forward primer 5′‐ATGATTGAACAAGATGGATTGCACG‐3′ (Npt–2 forward primer) and a reverse primer 5′‐TCAGAAGAACTCGT CAAGAAGGC‐3′ (Npt–2 reverse primer) were used to confirm the positive transgenic cotton lines by PCR (Figure S5).

Site‐directed mutagenesis (SDM) analysis of promoter activity in transgenic cotton

GeneArt® Site‐Directed Mutagenesis System (Cat No. A13282, Invitrogen) was used to introduce total eight mutations in the motifs (MYB1AT, MYBCORE, MYBST1, MYC AT RD22, HEXMOTIF, L1BOX, MYC CONSENSUS and MYCATRD22) of PGhGDSL at loci –603/–598, –600/–594, –558/–553, –450/–445, –418/–413, –226/–219, –67/–62 and –11/–6. The mutagenic forward and reverse primers (Table S5) were synthesized (Applied Biosystems, Fostercity, CA, USA) following the instructions (Invitrogen, Carlsbad CA, USA) and used to generate the mutagenic construct by PCR. The mutation constructs (sdm) were confirmed by a 96 capillary automated sequencer (ABI 3730 DNA Analyzer) using a vector specific T3 forward primer (5′‐AATTAACCCTCAC TAAAGGG‐3′) and a T7 reverse primer (5′‐GTAATACGACTCAC TATAGGGC‐3′). The confirmed plasmids were digested with restriction enzymes (SalI‐BamHI) and ligated into the plasmid pBI101 (Clontech, Terra Bella, CA, USA) using T4 DNA ligase (NEB). The mutagenic constructs were named as sdm1 to sdm8. All the constructs were transformed in cotton by embryo transformation method (Kumar et al., 2013). The forward primer 5′‐ATGATTGAACAAGATGGATTGCACG‐3′ (Npt–2 forward primer) and reverse primer 5′‐TCAGAAGAACTCGTCAAGAAGGC‐3′ (Npt–2 reverse primer) were used to confirm the positive transgenic cotton lines by PCR (Figure S5).

Transformation of cotton

The details are provided in Appendix S2.

Yeast one‐hybrid assay (Gold Y1H)

The details are provided in Appendix S2.

Gene co‐expression network analysis of MYB1AT

The details are provided in Appendix S2.

Pathway analysis of positively and negatively interacting partners of MYB

The metabolic pathways or cellular processes of positively and negatively interacting genes were analysed by MapMan software version 3.5.1 (http://gabi.rzpd.de/projects/MapMan/). An average statistical test followed by the Benjamini–Hochberg was used to identify the functional categories (BINSs, subBINs) enriched in these genes. This software was used to visualize the amplitudes of the changes in the expression of individual genes belonging to the metabolic pathways or cellular processes.

Conflicts of interest

The authors declare no conflicts of interest.

Figure S1 GUS expression analysis.

Figure S2 GFP expression analysis in cotton fiber driven by GhGDSL promoter.

Figure S3 MapMan analysis of the positive and negative interacting partners of MYBs.

Figure S4 Isolation of GhGDSL promoter by genome walking.

Figure S5 Screening the transgenic lines by PCR.

Figure S6 Estimation of cellulose content as described by Updegraff (1969).

Figure S7 Phylogenetic analysis of GDSL (gene id CotAD_74480) with Arabidopsis.

Table S1 BLAST result of PGhGDSL and GhGDSL gene.

Table S2 Putative motifs identified in PGhGDSL.

Table S3 Putative promoter binding proteins identified by Y1H system.

Table S4 Motif search for the TFs identified by Y1H.

Table S5 Primers used in this study.

Click here for additional data file.

Appendix S1 Positively and negatively co‐expressed genes.

Click here for additional data file.

Appendix S2 Supplementary materials and methods.

Click here for additional data file.