Background

MYB gene family members mainly regulate apoptosis, cell differentiation and proliferation. A- MYB, B- MYB and C-MYB are three groups of MYB gene family [1]. These three groups are largely present in vertebrates, fungi, slime molds and insects [2, 3]. For the very first time, MYB gene C1 was isolated from maize and found to regulate biosynthesis of anthocyanin [4]. MYB proteins play a key role in different developmental processes including trichome differentiation, biosynthesis of anthocyanin and flavonoids, floral symmetry, cell proliferation and control of cell cycles [5, 6]. In Antirrhinum majus, two dorsal petals of zygomorphic flowers were larger than ventral and lateral ones with single aborted stamen [7]. RADIALIS (RAD) belongs to MYB gene family. RAD is involved in creating the dorsal identity and regulating the domain activity of DIVIRICATA (DIV) (another transcription factors of MYB gene family) by restricting it to ventral regions of the flower [8]. RAD and DIV, interact with each other and control dorsoventral asymmetry of flowers [9].

In Arabidopsis, RAD and DIV gene families have 198 paralogs and they largely contribute in determining orthology across taxa. MYB genes contained three alpha helices. RAD and DIV have 64 members in Arabidopsis and are further classified into five subgroups. Among five subgroups, RAD falls in the I-box-like subgroup while R-R-type subgroups include DIV. RAD was the result of loss of the second MYB domain of DIV [10]. First two helices of RAD share more similarity with the N terminal MYB domain of DIV as compared to third helix and third helix include DNA binding domain of both RAD and DIV [11].

RAD5, is a member of Snf2 helicase family, encodes two domains including RING-finger domain and HIRAN domain. In yeast (Saccharomyces cerevisiae), RAD5 is the key component in the RAD5-dependent error free branch of postreplication repair. AtRAD5a and AtRAD5b share significant identity to RAD5 of yeast. AtRAD5a and AtRAD5b encodes same domain and these two genes have significant similarities with each other as compared to RAD5 of yeast. However, both AtRAD5a and AtRAD5b differ in their function [12]. On the basis of epistatic analysis, yeast DNA repair genes can be classified into three different groups, including RAD3, RAD52 and RAD6 groups. RAD3 group control nucleotide excision repair and also eliminate DNA damages caused by ultraviolet radiation [13]. However, the RAD52 group mainly functions in the repair of double-strand breaks induced by ionizing radiation and RAD6 group overcomes damages of DNA replication and also arrest replication forks [14]. These DNA repair groups encodes members of the DNA helicase-like-SNF2-gene family [15]. SNF2 family members undergo three main functions including binding of specific proteins to chromatin, destabilization of nucleosome structure and alteration of contact points between DNA and proteins [16, 17].

Cotton is an economic crop of the world and provides raw materials for the textile industry and edible oils [18]. G. hirsutum and G. barbadense are the two allotetraploid species of cotton and provide fiber, seed oil and protein meal to the industry [19]. G. hirsutum is famous for its high yield and moderate fiber qualities, while G. barbadense is important for superior quality fibers and accounts for 3% of the world’s cotton production [20]. Although some RAD genes have been previously reported in model organisms Arabidopsis and yeast, the function of RAD gene family in cotton has not been reported. In the present research, we systematically identified RAD genes in six different species and analyze their phylogenetic relationship, exon and intron structure, conserved protein motifs, biophysical properties, cis-acting elements of promoter regions, sequence logos, chromosomal localization, collinearity and miRNA target sites. Present study will increase our understanding about RAD gene family and provide insight into the molecular mechanisms of RADs in different Gossypium species.

Results

Identification of RAD genes in Gossypium species

In this study we used different bioinformatic analysis to identify RAD genes in various plant species. We identified 17 RAD genes in G. hirsutum, 16 genes in G. barbadense, eight genes each in G. herbaceum and G. arboreum and nine genes in G. raimondii. Furthermore, we also identified 10 RAD genes in Arabidopsis. SMART (http://smart.embl-heidelberg.de/), PROSITE (http://prosite.expasy.org/) and InterProscan 63.0 (http: //www.ebi.ac.uk/interpro/) were used for confirmation of RAD genes in different species. We found that among all the selected plant species, G. hirsutum had the highest number of RAD genes (Table S1) elucidating that during hybridization, GhRAD genes underwent polyploidization and experienced significant dupli- cation events. The biophysical properties of the GhRAD genes including number of amino acids, chromosomal position, isoelectric point, molecular weight, grand average of hydropathicity and protein localization were observed (Table S2). The number of amino acids of GhRAD proteins range from 885 (GhRAD1) to 1374 (GhRAD13). The isoelectric point (pI) of GhRAD genes range from 5.26 (GhRAD14) to (9.04) (GhRAD8). Molecular weight ranges from 76561.78 Da (GhRAD11) to 151381.67 Da (GhRAD13). Moreover, predicted subcellular localization indicated that out of 17 GhRAD genes 16 genes were located in the nucleus while only one gene was located in the cytoplasm.

Phylogenetic analysis, structural features, conserved domain and motifs analysis of GhRAD genes

To study the phylogenetic relationships of RAD gene family, 68 genes from six plant species were used to create a phylogenetic tree (Fig. 1). RAD genes were classified into five clades with 24 members in RAD-a clade, nine members in RAD-b, 14 members in RAD-c while RAD-d and RAD-e contained 18 and two members respectively. RAD-a with maximum number of genes contained RAD proteins from all six plant species while RAD-e, the smallest clade contained two genes from G. hirsutum and G. raimondii. All five clades contained genes from six species except RAD-e and one gene of Arabidopsis (AtRAD6) did not fall in any clade. Moreover, we created another phylogenetic tree to explore the evolutionary relationship between G. hirsutum and Arabidopsis (Fig. S1). Phylogenetic tree divided 17 G. hirsutum and 10 Arabidopsis RAD genes into three clades. Rad-c was the largest clade with 12 members, while Rad-b was the smallest one with four members. GhRAD9 did not fall in any clade, demonstrating that this genes was dissimilar from each other and it has some special function in species evolution.

Fig. 1
figure 1

Phylogenetic analysis of RAD proteins in six different species. Prefixes Gh, Gb, Ghe, Ga, Gr and At were used to identify RAD genes from G. hirsutum, G. barbadense, G. herbaceum, G. arboreum, G. raimondii and Arabidopsis

Full size image

Next to study the structural features of GhRADs, we analyzed the exon and intron structure, conserved domain and motifs using the GSDS program and MEME tool respectively. Multiple numbers of exons in GhRAD genes ranging from 3 to 24 were observed (Fig. S2). Maximum number of exons were present in GhRAD9 while GhRAD8 had minimum number of exons. Similarly, GhRAD9 contained the highest number of introns while GhRAD16 contained two introns. Number of introns ranged from 2 to 23 for GhRAD genes. GhRAD11 had the longest genomic sequence of more than 8 kb while GhRAD16 had the shortest genomic sequence of 4 kb. Next, we analyzed the conserved domain of GhRAD proteins. Results of conserved domain analysis showed that all GhRAD proteins contain SNF2_N superfamily domain (Fig. S3). Furthermore, we analyzed the motif distribution pattern of GhRADs and found that ten conserved motifs were identified in GhRAD proteins (Fig. S4). Results indicated that GhRAD genes had conserved gene structure and protein motif distribution pattern although the process of gene duplication took place a long time ago during evolution.

Chromosomal position information and synteny analysis

Chromosomal position of GhRADs exhibited that 17 GhRADs were distributed unevenly to 11 chromosomes. Out of 17 genes, six genes were mapped on At sub-genome, while nine GhRAD genes were mapped on Dt subgenome and two genes were present in the form of scaffolds (Fig. S5). This uneven distribution pattern of GhRAD genes on chromosomes illustrated the presence of genetic variability during the evolutionary process. More deeper insights showed that the maximum number of genes were mapped on chromosome D02. Most chromosomes contained a single GhRAD gene and no gene was located on chromosome A02 and A04 of At sub-genome as well as chromosome D03 of Dt sub-genome.

Collinearity analysis between orthologs of At and Dt sub genomes revealed that the different GhRAD and GbRAD loci showed significantly conserved pattern among At and Dt sub-genomes (Fig. 2 A and B and Table S3, S4). Our results indicated that At and Dt sub-genomes are orthologous in the A genome (G. arboreum) or D genome (G. raimondii). Upland cotton (G. hirsutum) was produced after hybridization between G. arboreum and G. raimondii. In our study, 112 and 15 orthologous/paralogous gene pairs were identified in G. hirsutum and G. barbadense respectively.

Fig. 2
figure 2

Collinearity analysis of GhRAD and GbRAD genes. Orthologous/paralogous gene pairs were indicated by blue color lines

Full size image

Ka/Ks ratios of GhRAD and GbRAD genes were calculated by Ka/Ks calculator 2.0 (Table S3, S4). In the course of evolution, duplicated gene pairs underwent non-functionalization, neofunctionalization and sub-functionalization [21]. In G. hirsutum 74 gene pairs exhibited Ka/Ks values below 0.5, illustrating that GhRAD genes bear purifying selection pressure while 15 genes showed Ka/Ks ratio above 0.5 and only three genes have Ka/Ks ratio of less than 0.1. While in G. barbadense 15 genes exhibited Ka/Ks ratio of less than 0.5. Taken together, our results indicated that purification selection pressure has a great contribution in maintaining the functional divergence of GhRAD and GbRAD genes.

Sequence logos and cis-acting elements

We created the protein sequence logos of Arabidopsis, G. hirsutum and G. raimondii to reveal the conservation of RAD genes during the process of evolution. Sequence logos provide valuable information about sequence similarities of selected plant species. Results indicated that protein sequence logos were conserved among all observed plant species at most sites within and between species across the N and C terminals (Fig. 3).

Fig. 3
figure 3

Sequence logos for conserved amino acid residues in G. hirsutum, G. raimondii and Arabidopsis. Protein sequence logos were created using online software WEBLOG

Full size image

Next, we identified different cis-acting elements of GhRAD genes. Process of gene expression and transcription can be well studied by analyzing the promotor cis-acting elements present in specific gene. The 17 putative GhRAD promoter regions possessed typical core cis-acting elements that were related to different phytohormones like MeJA, gibberellin, auxin, salicylic acid and abscisic acid (Fig. 4). Most of the GhRAD genes promoter shared numerous elements for plant growth and development such as seed specific regulation, meristem and endosperm expression and cell cycle regulation. Additionally, we found that many drought responsive elements, flavonoid biosynthesis, stress responsive elements, light and low temperature responsive elements, anaerobic induction, anoxic specific inducibility and wound responsive elements were also present in GhRADs promotor regions.

Fig. 4
figure 4

Analysis of promoter cis-acting elements of GhRAD genes. Cis-acting elements were confirmed using PlantCARE database. Cis-acting elements were represented by different colour boxes

Full size image

Expression profile of GhRAD genes in different tissues and hormonal stresses

Analysis of gene expression patterns is helpful to predict the biological functions of genes. We downloaded the RNA-seq data from NCBI. We investigated the gene expression pattern of GhRADs in various tissue samples and a heatmap was generated (Fig. S6). As shown in figure, GhRAD1 and GhRAD10 showed high transcript level in almost all selected tissues demonstrating that these two genes have multiple biological functions and they have some special functions related to fiber development while GhRAD6 and GhRAD15 were poorly expressed in some tissues showing that these two genes might be pseudo genes and have limited role in cotton growth and development. Next to validate the results of RNA-seq data, we performed qRT-PCR analysis in eight different tissues including root, leaf, stem, flower and 0, 5, 10 and 20 DPA fiber (Fig. 5). Results indicated that most of the GhRAD genes showed preferential expression in root except GhRAD11 whose expression was preferentially high in 20 DPA fiber, indicating its potential role in fiber elongation and biosynthesis of secondary cell wall. GhRAD2, GhRAD3, GhRAD4 and GhRAD11 showed high expression in flower tissues demonstrating that these RAD genes might function in reproductive development.

Fig. 5
figure 5

Tissue specific expression patterns analysis of GhRAD genes. Tissues were collected at different development stages such as root, stem, leaf, flower and 0, 5, 10 and 20 DPA fiber

Full size image

Next, we inspected the transcript level of GhRADs under three phytohormonal treatments (BL, JA and IAA) at 1, 3, 6 and 12 h of treatment (Fig. 6). Results revealed that most of the GhRAD genes showed up and downregulated expression patterns in different hormones with different time points. GhRAD1, GhRAD2 and GhRAD9 showed upregulated expression in all three selected phytohormonal stresses, illustrating that these genes might respond to various hormone signaling pathways. GhRAD1 showed significantly upregulated expression at 1, 3, 6 and 12 h of GA treatment, indicating its potential role in GA signaling pathway. Moreover, transcript level of GhRAD12 was preferentially high at 1, 3 and 6 h of IAA treatment, suggesting that GhRAD12 might respond positively to IAA signal. Taken together our results illustrated that GhRAD genes were important in several phytohormone signaling pathways.

Fig. 6
figure 6

Expression pattern analysis of GhRAD genes under different phytohormonal stresses including BL, GA and IAA at 1, 3, 6 and 12 h of treatment by qRT-PCR analysis

Full size image

Prediction of miRNA target sites

In plants, microRNAs (miRNAs) help in regulation of gene expression by activating mRNA translational repression [22]. To explore miRNA mediated post-transcriptional regulation of GhRAD genes, coding sequences of 17 GhRAD genes for putative target sites were searched with the help of psRNATarget server. Results indicated that GhRAD1 was targeted by Ghi-miRN1438 with a site in SNF2_N superfamily domain (Fig. 7 and Table S5). Similarly, GhRAD2, GhRAD3 and GhRAD4 were targeted by Ghi-miRN1496, Ghi-miR1440 and Ghi-miR2111b respectively. Further, Ghi-miR2950a, Ghi-miR390a, Ghi-miR390b and Ghi-miR7495 regulated GhRAD9, GhRAD10, GhRAD11 and GhRAD12 respectively, with SNF2_N superfamily domain. These results suggested that GhRAD genes can be regulated by different miRNAs.

Fig. 7
figure 7

The miRNA-mediated target sites of GhRAD genes. Green color boxes indicated ORFs (open reading frames) of GhRADs. SNF2_N superfamily domain was represented by orange box while miRNA complementary sites with the nucleotide positions of RAD genes were indicated by red color

Full size image

Discussion

Cotton is an important economic crop of the world and contributes largely to the textile industry. Different gene families such as GGPPS [23], LOG [24], BES1 [25], GH3 [26], GSK [27], GATL [28], GhPHD [29] and GhAA1 [30] have been identified in cotton, but there is no published research about RAD gene family in cotton. Previous studies reported that RAD51 participate in DNA repair, homologous recombination and genome stability. Members of the recA_RAD51 family perform functions that have differentiated during evolution [31]. Human RAD51 is a homologue of the Escherichia coli RecA protein and function in recombinational repair of double-stranded DNA breaks. Mutations of RAD51 reduce repair of double-stranded DNA breaks. Loss of RAD51 function result in an elevated mutation rate and accumulation of DNA damage therefore increase risk of cancer in human [32]. In Antirrhinum, RADIALIS gene is associated with regulation of floral asymmetry [33]. Induction of recessive mutation in RADIALIS gene produce symmetrical floral morphology. The Arabidopsis RAD gene family contains four members, including RSM1, RSM2, RSM3 and RSM4, RADIALIS-LIKE SANT/MYB 1–4). Arabidopsis RAD-like (RSM) genes might function in developmental process that is not necessarily relevant to the floral architecture [34] [35].

Gossypium genus contain 45 diploid and six tetraploid species [36, 37]. Two allotetraploid species including, G. hirsutum and G. barbadense arisen through interspecific hybridization of A and D genome progenitors [19]. Polyploidy occurred around 1–2 million years ago and produced allotetraploid species [38]. G. hirsutum and G. barbadense are the two oldest allotetraploid species of cotton [39] [40]. Here, we comprehensively identified RAD genes in different plant species to identify the diverse functions of RAD gene family during the process of plant growth and development. Our study will provide basic information about biological and functional mechanisms and could be helpful for further research of RAD gene family in different cotton species.

GhRAD genes show evolutionary conservation

In this study, we identified RAD genes from Arabidopsis and five Gossypium species. Interestingly, the number of GhRAD genes was higher as compared to GaRAD and GrRAD genes, possibly because G. hirsutum is an allotetraploid crop resulting from hybridization of A and D genome progenitors. In upland cotton, both At and Dt sub-genome donors are orthologous relatives, leading to the GhRAD gene duplication [41]. Because of duplication and doubling of GhRAD genes, the numbers of GhRAD genes were equal to the total number of GaRAD and GrRAD genes. Phylogenetic analysis naturally classified RAD genes into five clades from RAD-a to RAD-e with 24 genes in RAD-a clade and two genes in RAD-e clade. Most of the RAD genes from diploids and allotetraploids were distributed closely in the phylogenetic tree, indicating that tetraploid cotton was produced after the hybridization of A and D genomes progenitors. From the phylogenetic tree we observed that RAD-a, RAD-b and RAD-d clades were ancient groups containing RAD genes from all selected plant species. Whereas, RAD-e clade might be more advanced containing RAD genes from only two selected plant species including G. hirsutum and G. raimondii, however, RAD-c clade lacks the genes from Arabidopsis.

Analysis of sequence logos of G. hirsutum, G. raimondii and Arabidopsis also supported these findings as the results of sequence logos showed significantly conserved pattern of amino acid across N and C terminals. Each conserved amino acid showed similar positions in three different species.

Interaction between transcription factors and promoter binding sites is important for gene regulation at the transcriptional level and transcriptional regulation is mainly responsible for regulation of gene expression. Different external signals activated inducible promoters and cis-acting elements in promoter regions are specific and consistent, for example, cis-acting elements of AuxRE were generally present in auxin-induced promoters regions [42], similarly AT-rich cis-acting elements, I-box, G-box, and GT1-motif were generally present in the light-induced promoters [43,44,45]. Likewise, CACG and CATGTG cis-acting elements were found in drought-induced promoters [46]. In our study, we extracted the upstream promoter fragments of candidate genes. Phytohormone (IAA, SA, ABA, GA and MeJA) responsive cis-acting elements were present in GhRAD genes promotors. Previous studies have shown that phytohormones maintain the integrity of plant cell walls. MeJA mediates gene expression in responses to plant injury [47]. Auxin regulates cell walls by induction of cell wall looseness [48]. In strawberry, ethylene promotes fruit softening and ripening by regulating the synthesis of pectin [49]. Stress responsive cis-acting elements help plants to respond quickly to stress and improve plant resistance to environmental stresses by activating different stress related genes. GhRAD promoter regions contain different cis-acting elements related to phyto- hormones, plant growth and development, seed specific regulation, meristem and endosperm expression and cell cycle regulation. In the present study, we found the uneven distribution of GhRADs on the chromosomes of At and Dt sub-genome. Uneven distribution of GhRAD genes indicated that during evolution and hybridization, these genes might have experienced gene duplication.

Gene structure is important to predict gene evolution [50]. Here, we found that the number of exons ranged from 3 to 24 while the number of introns ranged from 2 to 23. We observed that length of intron is significantly different, suggesting that during functional diversification of GhRADs intron length might play major roles. It has been reported that gene structure is associated with the evolution of different plant species [51]. The evolutionary analysis within the RAD genes showed that most of the RAD genes were greatly conserved during evolution. As RAD genes introns were not lost during evolution and at the early expansion stages of evolution, these genes diverged, whereas, over evolutionary time, other genes lost their introns [52], indicating that group specific genes may have similar functions. Insertion/deletion events might be the reason for structural differences of exon/intron and may help to identify different evolutionary mechanisms [53]. Gene families with fewer number of introns are considered to be advanced gene families [54] while gene families showing more introns have acquired some novel functions during the evolutionary process. Furthermore, motif distribution pattern of GhRAD genes was conserved indicating that the proteins with similar kind of motif distribution pattern might have some specific functions in cotton growth and development.

Diverse expression patterns of GhRADs, prediction of their miRNA target sites and selection pressure

In the recent years, different researches have been done to identify the function of RAD genes in different plant species but no research has been conducted to find the functional mechanisms of RAD genes in cotton. As our results indicated that promotor of GhRADs contain elements related to phytohormones, plant growth and development, seed specific regulation, meristem and endosperm expression and cell cycle regulation, so to check the potential regulatory functions of GhRADs during cotton growth and development, we evaluated the transcript patterns of 17 GhRAD genes using qRT-PCR. Results indicated that many GhRAD genes showed high expression in roots. However, GhRAD2, GhRAD3, GhRAD4 and GhRAD11 showed significantly high expression in flower tissues. Results indicated that GhRAD genes might be playing a positive role in root and flower development. All GhRAD genes showed poor expression in fiber development stages except GhRAD11 had relatively high expression at 20 DPA fiber.

Different research studies indicated that phytohormones respond positively to abiotic stress response [55]. Auxin response factors (ARFs) activate auxin responsive genes and ARFs are the potential mediators that help the plants to respond to adverse environmental conditions [56]. Ethylene response factors of AP2/ARF transcription factor family play a positive function in plant growth, disease resistance and phytohormone response [57]. In this study, analysis of GhRADs for three different phytohormones indicated that GhRAD genes can be regulated by the exposure of phytohormones such as BL, IAA and JA. GhRAD1 and GhRAD2 showed upregulated expression for observed phytohormones, suggesting that these genes may have a positive role in phytohormones signaling.

Previous studies reported that miRNAs have significant functions in plant growth and development, regulation of cell growth and different metabolism associated with cotton fiber development and also respond to different environmental stresses including cold, heat and drought [58]. In the past few years, different miRNAs have been identified in cotton and they are preferentially expressed in different vegetative tissues as well as during abiotic stresses [59, 60]. For instance, 32 families of miRNA were found to be differentially expressed in cotton ovules and leaves. With the help of RNA-seq data of cotton leaf and ovule, a total of 65 families of cotton miRNA have been identified [61]. In our study we identified different GhRAD genes targeted by various miRNA. Precisely, Ghi-miRN1496, Ghi-miR1440, Ghi-miR2111b, Ghi-miR2950a, Ghi-miR390a, Ghi-miR390b and Ghi-miR7495 were the miRNAs targeting most of GhRAD genes. Our results will lay the foundation for future research about biological functions of GhRAD miRNAs and also their target sites in G. hirsutum.

Next we calculated the Ka/Ks ratios of GhRADs and GbRADs. Genes with Ka/Ks = 1.0 are considered as pseudogenes as a result of neutral selection, Ka/Ks less than1.0 demonstrates the tendency of duplicated genes for purifying selection, while Ka/Ks ratios greater than 1.0 exhibits positive selection of accelerated evolution. Ka/Ks values of most GhRAD and GbRAD were less than 1 indicating that during the long-term evolutionary process, RAD gene family experienced purifying selection pressure.

Conclusions

In the present study, we conducted a comprehensive genome wide analysis to identify phylogenetic relationship, sequence conservation and biological functions of RAD genes. Phylogenetic analysis divided RAD genes into five clades on the basis of sequence homology. Sequence logos, exon-intron structure, protein motifs and conserved domain analysis indicated that GhRAD genes were highly conserved during evolution with the uneven distribution on the chromosomes. Collinearity and selection pressure analysis indicated the expansion of RAD gene family in G. hirsutum and G. barbadense with purifying selection pressure. Promoter cis-acting elements analysis indicated the presence of several plant growth and stress related regulatory elements. Tissue specific expression pattern analysis of GhRAD genes indicated that some GhRAD genes showed significantly high expression in roots and flowers, except GhRAD11 had maximum expression at 20 DPA fiber. Further, GhRADs showed regulatory response under three phytohormonal stresses indicating candidate genetic material for cotton breeding against different stresses.

Methods

Identification of RAD gene family

The database of cotton species including (G. hirsutum, G. raimondii, G. arboreum, G. barbadense and G. herbaceum) were acquired from the Cotton FGD (https://cottonfgd.org/) [62]. The Arabidopsis protein sequences of RAD genes were acquired from Arabidopsis Information Resource, version 10 (TAIR 10) (http://www.arabidopsis.org). The AtRAD genes were used as queries to identify the RAD genes in other plant species. SMART (http://smart.embl-heidelberg.de/) and Interproscan 63.0 (http://www.ebi.ac.uk/InterProScan/) were used to confirm RAD protein sequences [63, 64]. Furthermore, Pfam: (http://pfam.janelia.org/) and hidden Markov model (HMM) was used to identify the conserved domain of RADs [65]. Moreover, to predict the biophysical properties of RAD genes, ExPASy-ProtParam tool (http://us.expasy.org/tools/protparam.html) was used. Subcellular localization was predicted by CELLO v2.5 server [66].

Phylogenetic analysis and sequence logos

For phylogenetic analysis of RAD gene family, Clustal W was used for multiple sequence alignment [67] with the default settings and the phylogenetic tree was generated by iTOL [68] with 1000 bootstrap replicates. For sequence logos, RAD proteins of Arabidopsis, G. hirsutum and G. raimondii were aligned with Clustal W [69] and protein sequence logos were created using online software WEBLOG with default parameters [70].

Chromosomal location, gene structure and analysis of protein motif

Chromosomal position of GhRADs was determined by cotton genome annotation data (ftp://ftp.bioinfo.wsu.edu/species/Gossypium hirsutum/NAU-NBI_G) and extracted gff3-files was used in MapInspect software (https://mapinspect.software.informer.com/) to map genes to their corresponding chromosomes. For gene structure analysis of GhRADs, Gene Structure Display Server 2.0 (GSDS) was used (http://gsds.cbi.pku.edu.cn/index.php) [71]. Online program MEME (http://meme-suite.org) [72] was used to determined conserved motifs of GhRAD gene family and finally motifs were displayed by TBtools [73].

Promoter cis-acting elements, synteny, Ka/Ks ratio and transcriptome data analysis

For analysis of promotor cis-acting element, 2-kb upstream promotor regions were downloaded from Cotton FGD website [62] and subjected to Plant CARE Database [74]. During collinearity, analysis orthologous/paralogous data were obtained by the previously described methods [75, 76] and circos was used to generate the figure [77]. Non-synonymous (Ka) and synonymous (Ks) divergence level ratios were calculated by aligning duplicated gene pair protein sequences in Clustal X 2.0, after which they were translated into complementary DNA (cDNA) sequences using the PAL2NAL program (http://www.bork.embl.de/pal2nal/). Finally, Ka and Ks values were calculated with the help of the CODEML program using the PAML package. For expression pattern analysis of GhRADs, RNA seq data (https://www.ncbi.nlm.nih.gov/pmc/articles/PMC4482290/) was used and heatmap was generated by using TBtools [73].

Prediction of miRNA target sites of GhRAD genes

Plant MicroRNA database (http://bioinformatics.cau.edu.cn/PMRD/) [78], miRBase (http:// www.mirbase.org/) [79] and the Cotton EST database (http://www.ncbi.nlm.nih.gov/nucest) were used to obtain MicroRNA sequences of G. hirsutum. Predictions of miRNA target sites were done with the help of psRNATarget server (http://plantgrn.noble.org/psRNATarget/home) as described previously [67].

RNA extraction and RT-qPCR analysis

In this research, the CRI24 variety was used for gene expression analysis. Total RNA was extracted by collecting cotton samples from the field. For tissue specific expression pattern analysis, plant samples such as root, stem, leaf, flower and 0, 5, 1 and 20 DPA fiber were used. Similarly, for hormonal stresses, cotton seedlings were subjected to JA, BL and IAA treatment for 1, 3, 6 and 12 h. RNA extraction was done with the help of RNAprep Pure Plant Kit (Tiangen, Beijing, China), after that cDNA was synthesized by using the Prime-Script®RT reagent kit (Takara, Dalian, China). Green qPCR SuperMix was used to conduct qRT-PCR assay and GhHis3 (GenBank accession no. AF024716) was used as internal control. PCR reaction was done in three replications. Results of qRT-PCR were calculated as described previously [80]. Primers used for RT-qPCR analysis were presented in (Table S6).