Identification of RR genes in cotton

The RR amino acid sequences reported in Arabidopsis and rice were used as query sequences. A total of 93, 94, 48, 49, and 27 genes were confirmed as RR family members in G. hirsutum (Gh), G. barbadense (Gb), G. raimondii (Gr), G. arboreum (Ga) and Theobroma cacao (Tc). The cotton and Theobroma cacao RR genes were named according to their location on chromosomes (Additional file 1: Table S2). The number of RR genes in the two allotetraploid cotton lines was almost twice that in the two diploid cotton lines. Their number also was more than that of other plant species (Table. 1), indicating that the RR gene family in cotton has undergone expansion during evolution. The number of amino acids, molecular weight (MW) and isoelectric point (pI) were calculated based on the predicted protein sequences. All 284 genes in cotton encoded proteins ranging from 826 (GhRR91) to 70 (GrRR3) amino acids, with protein pIs varying from 10.36 (GrRR3) to 4.58 (GhRR67) and MWs varying from 88.16 (GhRR91) kDa to 8.42 (GrRR3) kDa (Additional file 1: Table S1). The subcellular localization prediction of cotton response regulator proteins showed that most of the Type-A RR and Type-B RR proteins were located in the nucleus, and a few were located in the cytoplasm and extracellular space. The Type-C RRs of the four cotton cultivars were all located in the cytoplasm, chloroplast and nucleus. There are a total of 23 Type-B PRR proteins in cotton, all of which are located in the nucleus. In addition to localization in the nucleus, Type-Clock PRR also is found in organelles such as chloroplasts and mitochondria. These results indicate that cotton response regulators not only participate in biological processes in the nucleus but also participate in cell activities in the cytoplasm and organelles.

Table 1 Summary of the RR gene numbers identified in plantsPhylogenetic analysis of RR genes family in cotton

To investigate the evolutionary relationships among the RR genes from four Gossypium species, Theobroma cacao, Arabidopsis thaliana, Glycine max and Oryza sativa, we constructed a phylogenetic tree in MEGA 7.0 using the NJ method. We utilized 238 orthologous and paralogous RR protein sequences from Theobroma cacao, Arabidopsis thaliana, G. hirsutum, Glycine max and Oryza sativa to determine the evolutionary history (Fig. 1A). The phylogenetic tree showed that the plant RR gene family could be grouped into five subfamilies, namely, Type-A RR, Type-B RR, Type-C RR, Type-B PRR and Type-Clock PRR, where Type-B RR is divided into three subfamilies, i.e., Type-B RRI, Type-B RRII and Type-B RRIII (Fig. 1A).

Fig. 1

The phylogeny trees of RR genes family. A Phylogenetic relationship of the 238 identified RR genes from 5 plant species. Theobroma cacao, Arabidopsis thaliana, G. hirsutum, Glycine max and Oryza sativa. B Phylogenetic relationship of the 317 identified RR genes from four cotton species and Arabidopsis. The two neighbor joining (NJ) phylogeny trees constructed using MEGA 7.0 software. Bootstrap values above 50% from1000 replicates are shown at each node

The largest clade was Type-B RR (Table 1). These results demonstrated that the Type-B RR subfamily is an ancient group of RR genes having the highest number of RR members from almost all plant species. All the subfamilies are composed of monocot and dicot species. Based on the phylogenetic tree, we found that RR genes in G. hirsutum were more closely related to those in Theobroma cacao than other plant species because they always naturally clustered closely to each other in the phylogenetic tree. In most cases, one Theobroma cacao RR gene corresponded to two homologous RR genes in G. hirsutum. Statistics found that some Theobroma cacao RR genes corresponded to at least two G. hirsutum RR genes, some corresponded to four, some corresponded to six, and at most one cocoa RR gene corresponded to eight cotton RR genes. For example, TcRR27 and GhRR14, GhRR61, GhRR80, GhRR32, GhRR79, GhRR31, GhRR78 and GhRR30 in Type-A RR are in one branch, TcRR20 and GhRR10, GhRR57, GhRR18, GhRR63, GhRR73, GhRR25, GhRR9 and GhRR56 in Type-B RR are in one branch, and TcRR6 and GhRR71, GhRR23, GhRR89, GhRR40, GhRR3, GhRR51, GhRR38 and GhRR87 in Type-B PRR are in one branch. This shows that ancient cotton experienced a double multiplication event when separated from grapes and cocoa, forming an ancient tetraploid. After cotton polyploidization, the genome is unstable, resulting in a large number of chromosome rearrangements, DNA segmental inversions and DNA segmental losses, such as Type-B RRI branches GhRR29, GhRR77, GhRR28, GhRR76, GhRR45, GhRR1, GhRR46 and GhRR2, which do not correspond to Theobroma cacao but cluster with Glycine max. Compared to the other dicot species, the phylogenetic analysis showed that the gene families of G. hirsutum and G. barbadense have experienced significant expansion because the number of RR genes of G. hirsutum and G. barbadense is almost more than double the number of RR genes, and due to the conserved functions, they tend to cluster into the same subgroup. In addition, it is a Type-B RR response regulator designated OsRR22 with a Tos17 insertion in the homozygous state, indicating that the loss-of-function mutation in OsRR22 is responsible for the hst1 salinity tolerance phenotype [24]. GhRR10, GhRR57, GhRR63, GhRR93, GhRR44, etc., in the Type-B RR family have the closest homology relationship with OsRR22, and they are in the same evolutionary branch. This indicates that they may have the same function and might be used as candidate genes for responsible salinity tolerance.

To investigate the evolutionary and orthologous relationships among different cotton species, the NJ phylogenetic tree was constructed using protein sequences of the cotton RR and AtRR families (Fig. 1B). On the basis of the previous AtRR gene family, the phylogenetic tree of all cotton RR proteins also was grouped into five clearly defined subfamilies, each of which contained proteins from Arabidopsis, both the diploid and allotetraploid cotton species. In almost every orthologous group, there is one copy in diploid cotton species and two copies in allotetraploid cotton species (one is from the Arabidopsis subgenome and the other is from the Dt subgenome). The clustering results further confirmed that two allotetraploid cotton species, G. hirsutum and G. barbadense, were the result of hybridization and doubled between the ancestors of two diploid cotton species, G. arboreum and G. raimondii [42].

Chromosomal localization of cotton RR genes

To clearly understand the chromosome distribution of the RR gene family, we drew a chromosome distribution map of 284 cotton RR gene family members (Additional file 2: Fig. S1). The results show that in G. arboreum, 49 RR genes are distributed on the remaining 12 chromosomes except for chr02, of which chr05 has the most genes, with a total of eight GaRR11, GaRR12, GaRR13, GaRR14, GaRR15, GaRR16, GaRR17 and GaRR18 genes. The RR genes were evenly distributed on 13 chromosomes in G. raimondii. In G. hirsutum and G. barbadense, except for AD1-A03 and AD2-A02, no RR gene distribution and all other chromosomes were distributed, and the RR genes distributed on the fifth chromosome were the most abundant. This shows that the RR genes were duplicated after the fifth chromosome was doubled in the genome. In addition, there were paired parallel homologous genes between the corresponding subgenomes of the GhRR and GbRR family genes in the A subgenome and the D subgenome. In AD1-A10 and AD1-D10, and in AD1-A11 and AD1-D11, there were five pairs of parallel homologous genes in the four pairs of subgenomes, AD2-A12, AD2-D12, AD1-A12 and AD1-D12. AD1-A01 and AD1-D01, and AD1-A13 and AD1-D13, are two pairs of subgenomes. There were two pairs of parallel homologous genes in AD1-A06 and AD1-D06, and in AD2-A06 and AD2-D06, there was one pair of parallel homologous genes in the two pairs of subgenomes, AD2-A08 and AD2-D08, and AD2-A010, and there were four pairs of parallel homologous genes in the two pairs of subgenomes of AD2-D10. There were three pairs of parallel homologous genes in the pair of subgenomes AD2-A09 and AD2-D09.

Collinearity analysis for cotton RR genes

The gene family refers to the genes derived from the same ancestor, a set of genes composed of two or more copies produced. The evolution of gene families arises through three processes, namely, whole-genome duplication, segmental duplication and tandem duplication [56]. The evolutionary relationships between the RR gene families of diploid and tetraploid cotton species were evaluated based on chromosomal distances, coverage and resemblance between their members, which was used to identify segmental tandem repeats and whole-genome duplication. The genes that lie on the same chromosomal block (e-value < 1e−5) were categorized as tandemly duplicated, while the remaining genes from the same genomes were considered segmental; other genes from different genomes and subgenomes of four Gossypium species were allocated in whole genome duplication [57]. Combined analysis of RR genes from G. hirsutum, G. barbadense, G. arboreum and G. raimondii was performed to analyze the gene duplications and syntenic relationships among them. G. barbadense (Gb) and G. hirsutum (Gh) genes were duplicated between G. raimondii (Gr) and G. arboreum (Ga), which indicates the origination of both tetraploid genomes from diploid genomes (A and D) during polyploidization (Additional file 1: Table S2). A total of 1049 orthologous/paralogous gene pairs were identified, of which 239 pairs were predicted in segmental duplication to form paralogous gene pairs within the GhAt/GhDt, GbAt/GbDt, A2/A2 and D5/D5 subgenomes, while 810 orthologous gene pairs experienced whole genome duplication. No tandem duplication gene pairs were found for RR genes family. From these results, we presumed that orthologous/paralogous gene pairs were generally raised from WGD before polyploidization was involved in the evolutionary process (Fig. 2).

Fig. 2

Syntenic relationships among RR genes of two diploid (G. arboreum and G. raimondii) and two allotetraploid (G. hirsutum and G. barbadense) cotton. The chromosomes of G. arboreum, G. raimondii, G. hirsutum and G. barbadense were shown with pink, blue, green and Purple colors, respectively

Structure analysis of cotton GhRRs

We analyzed the gene structure, motif and protein domains of the RR genes identified in upland cotton (Fig. 3). Exon/intron analysis and phylogenetic tree results show that the RR gene families all have exons and introns, and genes in the same subfamily usually have surprisingly similar exon/intron structures. Except for GhRR47 in the Type-Clock PRR subfamily and GhRR56 and GhRR31 in the Type-A RR subfamily, the three gene intron sequences are longer than those in the same subfamily (Fig. 3B). These results indicate that genes with different gene structures may have special biological functions. A total of 10 motifs were identified by the online program MEME and named Motif 1–10. As shown in Fig. 3C, except for Motif 1, which is a widely distributed RR domain, as expected, the same subfamily of RR gene members usually has a common motif, indicating that their functions are similar. We also found that some protein motifs existed in multiple members of the five classes. However, others are specific to a particular class. For example, 10 motifs exist only in one branch of the Type-B RR type. In addition to the three motifs of GhRR47 and GhRR56, Type-C RR has only two motifs for the rest of the genes. The Type-A RR gene contains only four motifs. The unique motifs in different subfamilies may represent the conservation and specific functions of the GhRR gene family (Fig. 3C).

Fig. 3

Comparison of the gene structure, conserved protein motifs and domains in RR genes on the G. hirsutum. A The NJ phylogenetic tree was constructed based on the full-length sequences of G. hirsutum RR proteins using MEGA 7.0 software. Details of subfamilies are shown in different colors. B Exon–intron structure of G. hirsutum RR genes. Blue boxes indicate exons; black lines indicate introns. C The motif composition of G. hirsutum. The motifs, numbers 1–10, are displayed in different colored boxes. D Schematic representation of the conserved domains in G. hirsutum RR proteins. The Rec domains, Myb and CCT domains are highlighted by red boxes, green boxes and watermelon red boxes, respectively. The length of DNA genomic or protein can be estimated using the scale at the bottom

Fig. 4

Analysis of RR genes promoter and its expression pattern under different stresses. A Phylogenetic tree of GhRRs. B Cis-elements in promoters of GhRRs. C Expression pattern of GhRRs under different stresses

Through the analysis of protein domains, it was found that 93 GhRR family gene members all have response regulator receiver domains. In addition, Type-Clock PRR not only contains a response regulator receiver domain but also contains a unique CCT motif on the C side. The CCT (CONSTANS, CO-like and TOC1) domain is a highly conserved basic module of 43 amino acids that is found near the C-terminus of plant proteins and is often involved in light signal transduction. The CCT domain contains a putative nuclear localization signal within the second half of the CCT motif and has been shown to be involved in nuclear localization and probably also has a role in protein–protein interaction [58]. Type-B PRRs and Type-B RRs also have a special MYB-like DNA-binding domain.

Promoter cis-elements analysis of GhRRs and their roles in gene expression under different abiotic stresses

The cis-acting element regulates gene expression and is closely related to the biological function of genes. We selected a 2000 bp 5′-flanking region upstream of the transcription start site of each RR gene in G. hirsutum and identified and analyzed them with the PlantCARE database. At the same time, we performed analysis of differentially expressed genes under cold, heat, salt and PEG stress for various durations (1 h, 3 h and 6 h) using RNA-seq. In the 93 RR gene promoter regions, 30 types of growth and development cis-elements, 14 types of phytohormone response cis-elements and nine types of abiotic or biotic stress cis-elements were identified (Fig. 4; Additional file 1: Table S4). It is worth noting that among the nine cis-elements of abiotic or biotic stress, analysis of 15 RR gene promoter cis-elements revealed that the MYB binding site was involved in drought inducibility. For example, GhRR78 and GhRR92 are classified as Type-A RRs, GhRR23, GhRR51 and GhRR71 are classified as Type-B PRRs, and GhRR50 and GhRR82 are classified as Type-Clock PRRs. These seven genes are all highly expressed under RNA-Seq stress. There were also 116 MYB motifs (which play a vital role in the regulation of auxin-regulated genes), 82 ABRE motifs (involved in abscisic acid responsiveness), 41 CGTCA motifs (regulatory elements involved in MeJA responsiveness), 32 LTRs (involved in low-temperature responsiveness), 37 W-boxes (wounding and pathogen response sites), 22 TC-rich repeats (involved in defence and stress responsiveness) and 122 EREs (ethylene-responsive elements) in the RRs that were highly expressed under other stresses (Fig. 4C).

Tissue expression patterns of RR genes in upland cottons

Through the analysis of the transcriptome data of different tissues in upland cotton (root, stem, leaves, flowers, ovules and fiber) [51] (Additional file 3: Fig. S2; Additional file 1: Table S7), the expression patterns in each subfamily were basically the same, and the expression patterns between subfamilies were quite different. However, most of the RR genes present a constitutive expression pattern. A small number of genes, such as the Type-C RR subfamily, were expressed at low levels or even not in the tissue group. The expression levels of the Type-Clock PRR subfamily genes in the initial stages of root and fibre development (0 DPA, 1 DPA) were significantly higher than those of other subfamilies. Type-Clock PRR subfamily members GhRR4, GhRR50, GhRR37, GhRR86, GhRR26 and GhRR74 have high expression levels in various tissues and cotton fibre development stages. This result implies that the abovementioned genes have unique functions and are relatively conserved during the evolution of cotton. This phenomenon also can be found in the Type-B RR branch.

Expression analysis of RR genes under salt stress and drought stress

To further ascertain whether GhRR gene family expression levels were related to abiotic stress, referring to the functional analysis of the rice RR gene family, Arabidopsis and other crops have been studied for RR gene function, with RNA-seq data for abiotic stress analysis. We selected three Type-A RR genes, GhRR7, GhRR17 and GhRR27, three Type-B RR genes, GhRR1, GhRR4 and GhRR28, and four Type-Clock PRR genes, GhRR26, GhRR34, GhRR41 and GhRR86, for a total of 10 genes as our research objects. Real-time fluorescence quantitative expression analysis was performed. To analyze the expression types of the selected genes more clearly and intuitively, we selected five varieties: salt-tolerant material Zhong9807, salt-sensitive material ZhongJ0102, drought-tolerant material ZhongH177, drought-sensitive material ZhongS9612 and genetics standard TM-1. We detected the relative expression levels of 10 genes at different time periods after treatment with 12% PEG6000 (Fig. 5) and 200 mM NaCl (Fig. 6).

Fig. 5

Expression levels of 10 GhRRs in 12% PEG6000 drought stresses. Error bars represent SD of three independent experiments

Fig. 6

Expression levels of 10 GhRRs in 400 mM NaCl stresses. Error bars represent SD of three independent experiments

The qPCR results showed that the expression levels of each gene in different time periods and different materials were different. The expression trend of each gene under salt stress and drought stress also was inconsistent. The expression level of GhRR1 in salt-sensitive materials after 1 h and 6 h treatments was significantly higher than that of the control. In salt-tolerant materials, with the prolonged stress time, the expression level showed a trend of first decline and then increase, and the expression level was significantly lower at 3 h. Compared with the control, the expression at 9 h and 12 h showed a very significant difference. In the standard line TM-1, the expression level of GhRR1 first increased and then decreased, reached the highest level at 9 h, and decreased slightly at 12 h, but it was significantly higher than that of the control. The expression level of GhRR1 after drought treatment was similar to that after salt treatment, and the expression level was the lowest at 3 h, which was consistent with the abiotic stress RNA-Seq data of GhRR1. In contrast to GhRR1, after salt treatment, the expression level of GhRR4 showed an upwards trend in salt-tolerant materials. After drought treatment, the expression level showed a trend of first declining and then increasing. Regardless of salt stress or drought stress, GhRR4 showed an upwards trend in salt-sensitive materials. The material is the same as the expression in TM-1, showing a fluctuating state. Under salt treatment, the expression of GhRR7 in the salt-tolerant materials increased significantly at 3 h, and the expression decreased in the salt-sensitive materials. The expression trends of the three materials were basically the same in the other three time periods. In contrast, after drought treatment with GhRR7, except for the expression of the three materials at 3 h, the expression levels of salt-tolerant materials and salt-sensitive materials were always opposite, and the expression level of TM-1 fluctuated. This indicates that GhRR7 may be involved in the regulation of drought tolerance in cotton.

Under salt stress, the expression patterns of GhRR17 and GhRR27 were the same. As the amount of time increases, the expression trends of the three materials are consistent. At 1 h and 12 h after drought treatment, the expression levels of drought-sensitive and drought-tolerant materials were significantly opposite. Quantitative data show that GhRR17 also may participate in cotton drought regulation. In contrast to the previous genes, GhRR26 had different expression trends in different materials under salt stress and drought stress. The trends of salt-tolerant materials and salt-sensitive materials were the same at 9 h, but the other time periods were significantly opposite. In drought-tolerant and drought-sensitive materials, the expression trend remained constant at 1 h, and the expression trends in other time periods also showed a significant opposite trend. It is speculated that GhRR26 not only participates in the regulation of cotton drought tolerance but also participates in the regulation of cotton salt tolerance. Under salt stress, the expression trend of GhRR28 was the same for the salt-tolerant and salt-sensitive materials at 3 h. Under drought stress, the expression level of GhRR28 under drought-tolerant materials showed at first a decline and then an upwards trend, while under drought-sensitive materials, the expression level of GhRR28 showed an expression trend that first increased and then decreased. Compared with photos, there was a very significant difference, which indicates that GhRR28 may respond to drought stress in the early stage of stress. The expression level of GhRR34 was not high under drought stress and salt stress in the early stage, and the expression trend in each material was basically the same. The expression trend of GhRR41 under drought stress and salt stress was basically the same, except that the expression trend of salt-tolerant materials and salt-sensitive materials was opposite except at 9 h. It is speculated that GhRR41 may participate in cotton salt stress and drought stress after stress. Under salt stress, the expression of GhRR41 showed the same trend in salt-tolerant and salt-sensitive materials. Under drought stress, the expression trends of GhRR86 were drought-sensitive materials and drought-tolerant materials at 1 h and 3 h, and the expression trends were the same in the later period, which indicated that GhRR86 might be involved in cotton drought regulation in the early stage of drought stress. This is basically the same as the heat map made by RNA-Seq data.

Protein interaction network

To analyze the function of the cotton RR proteins, we used gene homology to analyze the protein interaction network of the RR genes in Arabidopsis thaliana through software. Through the fully studied Arabidopsis RR proteins, we can infer the large part of the regulatory network involved in cotton RR proteins [54, 59] (Fig. 7). performing a protein family search, we found the DNA-binding transcription factor activity (COG5641) of cotton RR proteins and other related regulatory pathways, such as participation in plant drought, high salt, low temperature, hormones (ET/ABA/GA/auxin), trauma, serine/threonine protein kinase signal transduction mechanisms [60] for pathogen response and cell cycle regulation, and other signal transduction processes. Two-component phosphor lay intermediate involved in MAP kinas cascade regulation and signal transduction histamine kinase. This shows that the function of cotton RR members depends on the two-component signal transduction system and serine/threonine protein kinase signal transduction.

Fig. 7

Interaction network of RR proteins. A Interaction network of RR proteins families. The red letters represent RR proteins signaling pathway. B Interaction network of GhRRs proteins with other proteins. The black and red letters represent AtRR proteins and cotton RR proteins, respectively. C Interaction network of GhRR7 proteins with other proteins. The black and red letters represent AtRR proteins and GhRR7 proteins, respectively. D Interaction network of GhRR41 and GhRR26 proteins with other proteins. The black letters represent AtRR proteins, red letters represent GhRR41 and GhRR26 proteins, respectively. E Interaction network of GhRR28 proteins with other proteins. The black and red letters represent AtRR proteins and GhRR28 proteins, respectively

By multiple sequence searches (Fig. 7B), all Type-C RR proteins interacted with AHP family proteins. Deserving of special attention is that almost all Type-B PRRs interact with other proteins by interacting with the CRY2 (cryptochrome-2) protein. In Arabidopsis, CRY2 can perceive low blue light (LBL) and respond by directly contacting two bHLH transcription factors, PIF4 and PIF5, at chromatin on E-box variant 5'-CA[CT]GTG-3' to promote their activity and stimulate specific gene expression to adapt global physiology (e.g., hypocotyl elongation and hyponastic growth in low blue light) [61]. In response to blue light irradiation, it triggers the nuclear accumulation of ROS and finally regulates plant abiotic stress. It is speculated that the cotton Type-B PRR gene also has the function of responding to abiotic stress. In addition, through the multisequence search method, we also found that GIGANTEA protein (GI), phytochrome interaction factor 7 (PIF7), phytochrome A (PHYA) and phytochrome B (PHYB) proteins participate in circadian clock signal regulation through light conversion. Arabidopsis histidine phosphotransfer protein (AHP) is involved in the phosphate relay signal transduction of histidine (His) to aspartate (Asp) [62].

We found that the proteins that interact with GhRR7 are mainly histidine kinases such as WOL (histidine kinase 4) [63], Arabidopsis thaliana histidine phosphotransfer proteins (AHPs) [64] and phytochrome B (PHYB) [62], which participate in the photoconversion and induce an array of morphogenetic responses. This implies that the GhRR7 protein not only participates in the regulation of the histidine protein kinase pathway but also participates in the transfer of light signals. Unlike GhRR7, GhRR28 not only interacts with AHPs but also with nonsymbiotic haemoglobin 2 (HB2), which may not function as an oxygen storage or transport protein but might act as an oxygen sensor or play a role in electron transfer, possibly to a bound oxygen molecule. GhRR28 has a low affinity for O(2), belongs to the plant globin family, and interacts with the UX/IAA transcriptional regulator family protein SHY2 in a regulatory pathway. Aux/IAA proteins are short-lived transcription factors that function as repressors of early auxin response genes at low auxin concentrations [65]. This indicates that the GhRR28 protein is involved in more regulation in plants. Both GhRR41 and GhRR26 belong to the Type-Clock PRR subfamily. Through protein comparison, we found that Arabidopsis PRR5 and PRR7, which are in the same branch as the GhRR41 and GhRR26 proteins in the phylogenetic tree, passed the homeodomain-like superfamily protein PCL1 [66] and LHY [67], Galactose oxidase/kelch repeat superfamily protein ZTL [68], protein EARLY FLOWERING 4 (ELF4) [69], LOV KELCH PROTEIN2 (LKP2) [68], CIRCADIAN CLOCK ASSOCIATED 1 (CCA1) [70], FLAVIN-BINDING, KELCH REPEAT, F-BOX 1 (FKF1) [71] and other proteins involved in the interaction of plant biological clock regulation and function. It is speculated that cotton GhRR41 and GhRR26 proteins also are involved in regulating the response function of the cotton circadian clock.

Drought tolerance of cotton enhanced after GhRR7 gene silencing

Combined with the fluorescence quantitative results and RR protein interaction network, we selected the GhRR7 gene for further research. Based on the interaction network analysis and differential expression patterns under different abiotic stresses, we hypothesized that GhRR7 is potentially important in the regulation of stress responses. To verify our hypothesis, we adopted a VIGS approach to knock down the expression of GhRR7 using TRV vectors, TRV: GhRR7. TRV: CLA was used as the positive control (Fig. 8A).

Fig. 8

Silencing of GhRR7 compromised cotton drought stress tolerance. A The positive control plants, control and target gene-silenced before 20% PEG6000 stress treatment. B Relative expression levels of control plants and target gene-silenced. C The expression levels of GhAHK3, GhAHP5 and GhPHYB genes in GhRR7 silenced plants. D POD activity between control plants and target gene-silenced under mock and 20% PEG6000 stress treatment. E SOD activity between control plants and target gene-silenced under mock and 20% PEG6000 stress treatment The expression levels of GhAHK3, GhAHP5 and GhPHYB genes in GhRR7 silenced plants. F CAT activity between control plants and target gene-silenced under mock and 20% PEG6000 stress treatment. G H2O2 amount of control plants and target gene-silenced under mock and 20% PEG6000 stress treatment. The data represent the means_SE from three independent experiments, t-tests: *p < 0.05, **p < 0.01

Two weeks after VIGS, when albino seedlings appeared in the positive control, qRT-PCR was used to determine the expression levels in the leaves of the TRV: GhRR7 and TRV: 00 control plants. The results show that the transcript levels of GhRR7 were significantly reduced following two weeks of VIGS (Fig. 8B). We treated the empty and silent plants with 20% PEG6000 and found that the silent plants exhibited a phenotype of wilting and water loss. In the next 12 h, the control plants (TRV: 00) displayed severe wilting and yellowing of leaves compared to the target gene-silenced plants under drought stress. Plants have a complete antioxidant enzyme system, which can eliminate ROS generated after drought stress. These antioxidant enzyme activities can directly reflect the stress resistance of plants. We tested the activity of POD, SOD, CAT and contents of H2O2 in the target gene-silenced plants and control plants (Fig. 8D–G). The results showed that compared with control plants, the activity of POD, SOD and CAT in the target gene-silenced plants’ enzyme activity were significantly increased, the contents of H2O2 in the arget gene-silenced plants were significantly depressed compared with those in the TRV: 00 plants, thereby increasing the removal of ROS and improving the drought tolerance of cotton.

To further explore the function of the GhRR7 protein, we selected three TCS system genes, GhAHK3, GhAHP5 and GhPHYB, to design specific primers (Additional file 1: Table S1) based on the predicted results of the GhRR7 protein regulatory network. The expression levels of three genes in the TRV: GhRR7 target gene-silenced plants were detected (Fig. 8C). The results showed that, compared with control plants, the expression levels of the GhAHK3 genes did not change significantly. The expression level of GhAHP5 in the target gene-silenced plants was significantly lower than that in the control plants, and the expression level of GhPHYB in the target gene-silenced plants was significant. It shows that GhRR7 may regulate the expression of downstream GhAHP5 gene and participate in the response of cotton to drought stress.