Cotton plants (G. raimondii and G. hirsutum cv. TM-1) were cultivated in the field under standard agronomic conditions at Hangzhou Normal University, in Hangzhou, China. Roots, stems and leaves were collected from 15-day-old seedlings. Petals, anthers and stigmas were harvested from flowers at 0 days post-anthesis (DPA). Fibers were detached gently from the ovules at 10 to 25 DPA. The collected materials were immediately frozen in liquid nitrogen and stored at –80°C prior to use.

The Arabidopsis thaliana wild-type (WT) was a Columbia-0 accession and the JA-deficient mutant dde2-2 was derived from Columbia-0. Seeds were first vernalized at 4°C for at least 2 days and then germinated in an illuminated growth chamber (22°C, 16-/8-h light/dark cycle regime) in soil or on agar plates (after surface sterilization of the seeds) containing half-strength Murashige and Skoog (MS) salts.

The reference genome of diploid cotton G. raimondii was downloaded from the Joint Genome Institute (JGI) Phytozome.¹ The TCP protein sequences of Arabidopsis were downloaded from NCBI² and used as reference sequences for building a Hidden Markov Model (HMM) to identify the G. raimondii TCP proteins using HMMER 3.1.³ All the putative TCPs were further subjected to SMART⁴ and InterPro⁵ analyses to identify their conserved domains as previously described (Long et al., 2020). The Compute pI/Mw tool⁶ was used to calculate the molecular weight (Mw) and theoretical isoelectric point (pI). Multiple sequence alignments were performed using Clustal X (version 1.83) (Thompson et al., 1997). The phylogenetic tree was generated from the aligned sequences, using the neighbor-joining method in MEGA6, with 1000 bootstrap replicates (Tamura et al., 2011).

Transcriptome data of G. hirsutum cv. TM-1 were downloaded from SRA databases (Accession code: PRJNA248163) (Zhang et al., 2015). Various tissues (root, stem, leaf, petal, pistil, stamen, ovule, and fiber) and abiotic stress (cold, hot, drought and salt stress for 1, 3, 6, and 12 h) transcriptome datasets were employed. The heat map was generated according to the method described previously (Gao et al., 2018). The reads per kilobase of transcript per million mapped reads values representing the expression levels of TCPs were collected from the “TM-1” cultivar transcriptome data downloaded from NCBI. The gene expression data were analyzed using the Genesis 1.8.1 program to generate heat maps. The cluster analysis, which was developed using the K-means method on the expression profiles of all 37 TCP genes, was also performed using the Genesis program.

Total RNA was extracted from 0.1 g fresh weight samples with the OmniPlant RNA Kit (DNase I) CW2598 (CWBIO, Beijing, China) according to the manufacturer’s instructions. HiFiScript gDNA Removal cDNA Synthesis Kit CW2582 (CWBIO, Beijing, China) was used to achieve first-strand cDNA synthesis from approximately 1 μg of total RNA. qRT-PCR was performed using the iQ SYBR Green Supermix (Bio-Rad, Hercules, CA, United States) and run on the ABI Prism 7000 system (Applied Biosystems, Foster City, CA, United States). GhUB7 and AtACT2 were used as the reference housekeeping genes in cotton and Arabidopsis, respectively. The relative expression level of the target genes was normalized against the reference housekeeping genes (Long et al., 2019). The error bars represent the standard deviation of three biological replicates. The primers used in the qRT-PCR are listed in Supplementary Table 1.

The open reading frame of GrTCP11 was amplified from G. raimondii fiber cDNA with primers GW-GrTCP11F and GW-GrTCP11R (Supplementary Table 2), and inserted into pB7WG2D,1 with the CaMV 35S promoter to generate the overexpression vector via the Gateway BP and LR reactions. The construct was electroporated into the GV3101 strain of Agrobacterium tumefaciens and used for Arabidopsis transformation (Clough and Bent, 1998).

For root hair length measurements, seedlings from surface-sterilized seeds were grown on plates held upright on half-strength MS medium, as previously reported (Hao et al., 2012). Root hairs on 7-day-old seedlings were photographed using an inverted microscope (Nikon Eclipse Ti, Tokyo, Japan), and measured with ImageJ software⁷. At least ten roots were measured for each independent experiment. Data were analyzed by Student’s t-test.

One-month-old seedlings (500 mg fresh weight) were extracted with 80% cold methanol (v/v) by shaking overnight at 4°C. Each sample was extracted twice to ensure adequate extraction. The organic phase was evaporated to dryness with N₂ and the residue dissolved in 0.4 mL methanol. The samples were stored at –20°C before being assayed. JA was quantified using an Applied Biosystems 4000Q-TRAR high-performance liquid chromatography–tandem mass spectrometry (HPLC–MS/MS) system with JA (Sigma, St. Louis, MO, United States) as the external standard. Three biological replicates were performed.

Surface-sterilized Arabidopsis seeds of the WT, 4-1 and dde2-2 were grown on plates held upright on half-strength MS medium plates with or without 50 μM methyl jasmonate (MeJA) at 22°C under a 16-h light/8-h dark photoperiod. The effect of MeJA on root growth inhibition was scored after 12 days of growth. Root lengths were measured using a ruler. At least 20 roots were measured for each independent experiment. Data were analyzed by Student’s t-test.

The TCP transcription factor family is a group of plant-specific transcription factors that have versatile functions in diverse aspects of plant development, including cotton fiber initiation and elongation (Hao et al., 2012; Wang et al., 2013). Based on an extensive genome-wide survey, we isolated a total of 37 putative non-redundant TCP genes in diploid G. raimondii (Table 1 and Supplementary Table 3), different from previous report of 38 non-redundant TCP genes identified from diploid G. raimondii (Ma et al., 2014). The deduced protein sequences alignment showed that the counterpart of GrTCP16 was reduced in our identification. Compared with Cotton_D_gene_10033515, the counterpart of GrTCP16 had the identical amino acid sequences, except for a truncation of 136 amino acids at the C-terminus. So, the counterpart of GrTCP16 was discarded as a redundant sequence of Cotton_D_gene_10033515 in our identification. The identities of other TCP proteins exhibit a good correspondence, excluding slight differences between the six proteins (GrTCP1, GrTCP6, GrTCP18a, GrTCP20c, GrTCP22, and GrTCP23) and their counterparts.

The expression levels of the 37 putative TCP genes in different tissues and in response to abiotic stress were analyzed using previously published transcriptome data to understand the possible roles of TCPs (Zhang et al., 2015). The reads per kilobase of transcript per million mapped reads values are listed in Supplementary Tables 4, 5. These values were used to create a heat-map of TCP genes’ expression. As indicated in Figure 1, substantial diversity is manifested in the tissue expression profile of TCP members. A few genes (with Gene ID of Cotton_D_gene_10006656, Cotton_D_gene_10017874, Cotton_D_gene_10030346, and Cotton_D_gene_10030663) were expressed at very low level across all 20 tissues, which suggested that they might be primarily expressed under particular conditions. More than half of the remaining TCP genes were predominantly expressed in different stages of cotton fiber development, which suggested that these genes might play important roles in fiber development. For example, the genes with Gene ID of Cotton_D_gene_10006406, Cotton_D_gene_10033147 or Cotton_D_gene_10008391 were preferentially expressed during the stages of fiber initiation, elongation or SCW synthesis, respectively. Two genes (with Gene ID of Cotton_D_gene_10000500 and Cotton_D_gene_10027048) had high expression in floral organs but low expression in all other tissues, which indicated they might specifically regulate the reproductive development of cotton. Additionally, the gene with Gene ID of Cotton_D_gene_10031549 was constitutively expressed in every tissue tested at very high level, which implied its possible roles at multiple development stages (Figure 1). However, most TCP members are not very sensitive to abiotic stresses, including cold, hot, drought and salt. For example, four genes (with Gene ID of Cotton_D_gene_10025316, Cotton_D_gene_10030346, Cotton_D_gene_10017874, and Cotton_D_gene_10030663) were not detected in any treatment, which suggested that these genes might not be involved in the stress responses (Figure 2).

Due to the expression profiles analysis, GrTCP11 (Gene ID: Cotton_D_gene_10006406) was cloned for further study as one of the fiber-specific expressed genes. Phylogenetic analysis was used to investigate the evolutionary relationship between the TCP proteins from G. raimondii and Arabidopsis. GrTCP11 was found to belong to class I TCP proteins as a unique homolog of AtTCP11 (Figure 3A). The amino acid sequence alignment showed that the identity of GrTCP11 with AtTCP11 full-length amino acids was 38%, and their TCP domains were 86% identical. Like most class I TCP proteins, the position 15 of the TCP domain in GrTCP11 is arginine, which is different from the threonine in AtTCP11 (Figure 3B). The coding sequence of GrTCP11 is 813 bp in length and encodes a putative polypeptide of 270 amino-acid residues with a calculated molecular weight of 29.6 kDa and an isoelectric point of 9.20. Based on the alignment between the coding sequence and the genomic DNA sequence, GrTCP11 was found to consist of two exons and a 366-bp intron. qRT-PCR analysis showed that the transcript level of GrTCP11 was the highest in anthers, while its transcript level was lower in roots, stems, leaves, stigmas and petals. qRT-PCR also showed that GrTCP11 was expressed in cotton fibers at different developmental stages. It had the highest transcript level at 0 DPA and 15 DPA, and could not be detected at other stages of fiber development (Figure 3C).

To verify the function of GrTCP11, we generated ectopic GrTCP11-overexpressed transgenic Arabidopsis plants, to bypass the difficulty and delay associated with conventional cotton transformation. Several independent transformants were isolated and the expression levels were measured using qRT-PCR. Five independent T₃ transgenic homozygous lines (3-1, 4-1, 5-4, 8-3, and 10-4) with different expression levels were further analyzed (Figure 4A). Overexpression of GrTCP11 in Arabidopsis inhibited root hair elongation. The mean length of the root hairs was significantly shorter in transgenic overexpressing lines 4-1 (171.2 μm), 8-3 (332.6 μm), 10-4 (448.3 μm), 5-4 (507.2 μm) and 3-1 (568.1 μm) than the root hairs of WT plants (650.3 μm) (Figures 4B,C). Overexpression of GrTCP11 in Arabidopsis also influenced the timing of flowering. The flowering time of the WT plants was earlier than in the transgenic lines (Figure 5). But there was no difference in their maturity time (Supplementary Figure 1).

Previous studies indicated that TCP transcription factors could regulate the development of cotton fibers and Arabidopsis root hairs through mediating phytohormone biosynthesis and signal transduction (Hao et al., 2012; Wang et al., 2013). Therefore, the expression levels of genes involved in phytohormone biosynthesis and response were determined in transgenic overexpressing lines and WT plants by qRT-PCR. As shown in Figure 6, many genes associated with JA biosynthesis (AtLOX4 and AtAOC3) and response (AtMYC2, AtJAZ1, AtJAZ2, and AtERF1) were significantly down-regulated in GrTCP11-overexpressed transgenic lines, compared with the WT plants. However, there was no significant difference in the expression levels of genes associated with other phytohormones, such as AtACO2, AtETR1, AtCTR1, AtTAA1, AtPIN1, AtIAA3, AtITP1, AtAHK3, and AtARR1 (Supplementary Figure 2).

Due to changes in the expression of JA biosynthesis and signaling-related genes, it was hypothesized that GrTCP11 might affect root hair development by regulating JA concentrations. To test this hypothesis, we extracted and determined the concentration of JA in transgenic lines overexpressing GrTCP11, dde2-2 and WT plants, using HPLC–MS/MS. The results showed that the concentrations of JA were significantly lower in the overexpressed lines, compared with the WT controls. In particular, the JA concentration in line 4-1, which had the lowest expression level of JA-related genes, was similar to that in the mutant dde2-2 (Figure 7).

To further test the effect of GrTCP11 on JA biosynthesis in Arabidopsis, 12-day-old seedlings of WT, 35S:GrTCP11 transgenic line 4-1 and dde2-2 were treated with 50 μM MeJA. As shown in Figures 8A,B, before MeJA treatment, 4-1 and dde2-2 were similar to WT plants in terms of root length. In the presence of 50 μM MeJA, the root growth of 4-1 was as insensitive to MeJA as was dde2-2. However, WT plants were hypersensitive to MeJA, with significantly shorter roots.

Several TCPs have been shown to function in the control of trichome and root hair development (Hao et al., 2012; Vadde et al., 2018; Wang et al., 2019, 2020). In the present study, a cotton class I TCP transcription factor GrTCP11 was characterized, which was homologous to AtTCP11. Heat-map and qRT-PCR analysis showed that GrTCP11 was preferentially expressed during the stages of fiber initiation and SCW synthesis, rather than in the fiber elongation stage, which indicated that GrTCP11 had a spatio-temporal regulatory effect on cotton fiber development. Ectopic overexpression of GrTCP11 in Arabidopsis led to shorter root hair length. Previous studies have shown that there may be similar regulatory mechanism operating between the elongation of cotton fibers and Arabidopsis root hairs (Hao et al., 2012). We speculated that GrTCP11 might also inhibit cotton fiber elongation.

Plant hormones play vital roles in fiber and root hair development, as previously described. The expression levels of genes associated with JA biosynthesis and response (AtLOX4, AtAOC3, AtJAZ1, AtJAZ2, and AtMYC2) were significantly down-regulated in the transgenic lines compared with the WT controls, in association with a decrease in JA concentration. Changes in JA concentration were further validated with respect to JA sensitivity and flowering time. It has been reported that MeJA inhibits root growth in Arabidopsis, whereas the sensitivity to MeJA can be alleviated in some JA-deficient mutants (Staswick et al., 1992; Lorenzo et al., 2004). Compared to the wild type, the transgenic line 4-1 exhibited reduced sensitivity of root growth to 50 μM MeJA, similar to that of the JA-deficient mutant dde2-2. Jasmonic acid has been implicated in regulating flowering time (Pak et al., 2009). The flowering times of transgenic lines 4-1 and 8-3 were later than the other lines, which might be associated with the decreased JA concentration in these transgenic lines. AtERF1, a common downstream responsive element of the jasmonate and ethylene pathways, was also significantly suppressed by GrTCP11 (Lorenzo et al., 2003). However, no significant difference was observed in the expression levels between the transgenic lines and WT plants of other hormone-related genes, involved in the biosynthesis and response of ethylene, auxin or cytokinin. The above results revealed that GrTCP11 might inhibit fiber and root hair elongation by directly suppressing JA biosynthesis and response. JA is a crucial factor which suppresses cotton fiber initiation, with an appropriate concentration promoting fiber elongation as described above. Therefore, the expression pattern of GrTCP11 is associated with fiber initiation and elongation. Overall, spatio-temporal specific GrTCP11 may be an important regulator for cotton fiber development through negative regulation of JA biosynthesis and response.

Cotton is one of the most important economic crops in the world, as its main product, cotton fibers, are the main raw material for the natural textile industry. Previous studies had revealed that transcription factors were widely involved in the regulation of cotton fiber development (Samuel Yang et al., 2006; Hande et al., 2017). Most of the plant-specific R2R3-MYBs were shown to regulate various developmental stages in cotton fiber, such as GhMYB109, GhMYB25 and GhMYB25-like (Pu et al., 2008; Machado et al., 2009; Walford et al., 2011). A R3-MYB transcription factor GhCPC delayed fiber initiation and inhibited early elongation by a potential CPC-MYC1-TTG1/4 complex. GhMYC1 could bind to the E-box cis-elements and the promoter of GhHOX3, which suggested that GhHOX3 may be downstream gene of the regulatory complex (Liu et al., 2015). GbTCP and GhTCP14 were shown to play vital roles in fiber initiation and/or elongation through regulating JA- and auxin-related genes, respectively, opening up research into the role of TCP transcription factors on cotton fiber development (Hao et al., 2012; Wang et al., 2013). Genome-wide analysis of the TCP family in a number of cotton varieties showed that they might be involved in different physiological processes of cotton fiber development. It was found that GhTCP22 and GhTCP14a could interact with several transcription factors related to cotton fiber growth and development, such as GhSLR1, GhGL3, GhARF6, GhTTG1, GhMYB23, and GhMYB25. In addition, GhTCP14a could also interact with GhEIN3, GhBZR1, and GhMYB25-Like proteins (Li et al., 2017). GhTCP4 interacted with GhHOX3 to coordinate fiber cell elongation and SCW biosynthesis, two events that are key to cotton fiber traits (Cao et al., 2020). Overexpression of GbTCP4 increased root hair length, root hair and trichome density, and the lignin content by binding directly to the AtCPC and AtCAD5 promoters in Arabidopsis (Wang et al., 2019). GbTCP5 regulated root hair development and SCW formation by binding to the promoters of the AtGL3, AtEGL3, AtCPC, AtMYB46, AtLBD30, AtCesA4, AtVND7, AtCCOMT1, and AtCAD5 genes to upregulate their expression in Arabidopsis (Wang et al., 2020). Here, we found that GrTCP11 possibly regulate fiber development by inhibiting JA biosynthesis and response in fiber initiation and SCW synthesis. The results indicated that TCP transcription factors possibly regulated fiber development by interacting with other fiber-related transcription factors, or binding to promoters to activate or inhibit their expressions. Protein complexes or targeted genes may regulate fiber development by affecting the homeostasis of phytohormones.

A growing number of TCP transcription factors have been characterized and confirmed to be widely involved in the regulation of plant growth, architecture and development. TCP transcription factors have some evolutionarily conserved roles in a range of plant species, such as regulation of branching, floral symmetry and leaf development (Danisman, 2016). For example, TEOSINTE BRANCHED1 in maize and its orthologs in rice, Arabidopsis, pea and poplar have a conserved function of suppressing branching (Doebley et al., 1997; Aguilar-Martinez et al., 2007; Braun et al., 2012; Muhr et al., 2016). On the other hand, there are suggestions that TCP proteins may have some new evolutionary roles in different species. Two closely related Arabidopsis TCP transcription factors, TCP14 and TCP15, were shown to be redundant in promoting cell proliferation in internodes and trichomes, and repressing cell proliferation in leaves and flower tissues (Kieffer et al., 2011). They were also necessary for seed germination by gibberellin-dependent activation (Resentini et al., 2015). AtTCP15 could modulate gynoecium development by participating in a feedback loop that helps to adjust the balance between auxin levels and cytokinin responses (Lucero et al., 2015). However, their homologous proteins in cotton had new biological functions. GbTCP, a protein orthologous to AtTCP15 in Sea-island cotton, promoted cotton fiber and Arabidopsis root hair elongation, as well as plant branching by directly activating the biosynthesis of and response to JA and then affecting downstream complex signal regulatory networks (Hao et al., 2012). Ectopic expression of GhTCP14 from upland cotton, which is homologous to AtTCP14, promoted the differentiation and elongation of trichomes and root hair cells through alteration of auxin homeostasis (Wang et al., 2013). AtTCP11 influences the growth of leaves, stems and petioles as well as pollen development in Arabidopsis (Viola et al., 2011). In the current work, a homolog of AtTCP11, named GrTCP11, was characterized in G. raimondii. High transcript abundance in anthers implied that GrTCP11 may have the same conserved function in pollen development as did AtTCP11. Our results showed that GrTCP11 may affect fiber and root hair elongation by directly inhibiting JA biosynthesis and response. The above research indicated that TCPs share some common functions but have also partially diverged, evolving distinct roles in different species through different regulatory mechanisms.