Genome-wide identification and expression analysis of CIPK genes in diploid cottons

Recognition and identification of CIPK family genes in cotton

To predict the members of the CIPK gene family in cotton, BLAST analysis was performed using the protein sequences of the CIPK gene families of A. thaliana, rice, and maize as the query sequences and the sequence data of the D and A genomes of the diploid cottons, G. raimondii and G. arboreum L., respectively; the E-value was set at 0.0001. Further confirmation was done using the screened protein sequences present in the protein database (http://www.expasy.ch/prosite/). The protein sequences of the N-terminal SNF kinase and C-terminal NAF domains were selected as the subjects for further analysis. NAF is the specific domain of CIPK. The selection method used was as described earlier by Albrecht et al. (2001).

Phylogenetic analysis of CIPKs

The protein sequences of A. thaliana CIPKs were downloaded from the A. thaliana genome database (http://www.arabidopsis.org/), those of rice were retrieved from the rice genome database (http://rice.plantbiology.msu.edu/) and National Center of Biotechnology Information (NCBI; Oryza sativa subsp japonica), and those of maize were downloaded from the maize genome database (http://www.maizegdb.org/gbrowse/maize) and NCBI. The protein sequences of the CIPKs in Populus trichocarpa were downloaded from the genome database of P. trichocarpa (http://www.ncbi.nlm.nih.gov/). According to the naming conventions reported in the literature, CIPKs of A. thaliana, rice, maize, and P. trichocarpa were named as AraCIPK, OsCIPK, ZmCIPK, and PtCIPK, respectively. Multiple sequence alignments were performed on the sequences of the proteins encoded by all the CIPKs using ClustalW2. The DNAMAN software (http://www.lynnon.com/) was used for sequence alignment. The neighbor-joining method was used for the construction of the phylogenetic tree and the MEGA software (http://www.megasoftware.net/) was used for inferring the phylogenetic tree.

Prediction of the isoelectric point (pI), molecular weight, and subcellular localization of proteins encoded by CIPK family genes

The theoretical pI and molecular weight of CIPK were calculated using the ExpASy protein server (http://web.expasy.org/compute_pi/). The average isotopic weight of one molecule of protein was expressed in daltons (Da). The subcellular localization of each CIPK was analyzed using the CELLO v2.5 server.

Chromosome location and gene structure analysis

The sequences of CIPK cDNAs were obtained from the D genome of G. raimondii and the A genome of G. arboreum L., respectively. The complete D genome of G. raimondii and A genome of G. arboreum L. were compared using the cDNA sequence as a query to determine the chromosomal location of each CIPK; the CIPK cDNA and the corresponding genomic DNA sequences were compared to identify the exon/intron structure of the gene (http://gsds.cbi.pku.edu.cn/).

Transcriptome expression analysis of CIPKs

The expression of CIPKs was analyzed in three tissues, namely mature leaves, 0-DPA (day post-anthesis) ovules, and 3.0-DPA ovules of G. raimondii and G. arboreum L. The transcriptome sequencing data were obtained from the NCBI Sequence Read Archive (SRA); the registration numbers of the G. raimondii samples were SRX111367, SRX111365, and SRX111366, and the original sequence number of G. arboreum L. was SRA150181. To evaluate the expression levels of CIPKs, the sequence reading aligned on the sequence of CIPK was converted to RPKM (Mortazavi et al., 2008) and the equation used for the evaluation was as follows:

where C refers to the reading length aligned uniquely on the transcript, N is the total reading length aligned uniquely on one specific sample, and L is the base of the transcript.

Expression analysis of CIPKs under different abiotic stress conditions

The seedlings of diploid cottons, G. raimondii and G. arboreum L. cv. Shixiya 1, were planted using the sand culture method. The true leaves, stems (hypocotyls), and roots of the seedlings were harvested and minced after abiotic stress treatments (low temperature: 4.0°C, 24 h, salt stress: 150 mM NaCl, 24 h, drought: relative water content in sand was reduced to approximately 5.0%) at the trefoil stage. Subsequently, the samples were quickly frozen in liquid nitrogen and stored at -80°C for RNA isolation. Untreated controls samples were also obtained. Total RNA was extracted from the collected tissues (Carra et al., 2007), and its concentration and purity was determined by Nanodrop2000 nucleic acid analyzer (Thermo, America). The RNA was reverse transcribed into cDNA using PrimeScript RT reagent kit with gDNA eraser (TaKaRa, China). Primer Premier 5.0 (PREMIER Biosoft) was used to design fluorescent quantitative primers (Table S1). Quantitative real time PCR (qRT-PCR) was performed using actin gene (GenBank accession No. AY305733) as a reference. The PCR programs were set as follows: 94°C for 30 s, followed by 40 cycles of 94°C for 5.0 s, 55°C for 34 s, and 72°C for 34 s. The assays were designed with three biological and three technical replicates. Relative quantitative analysis of the target genes was performed using the 2^-ΔΔCt method.

Identification of CIPK gene family members in the complete genome of cotton

The CBL-CIPK complex significantly affects the expression of target proteins, thereby, influencing various metabolic activities of plants. BLAST analysis was conducted using the CIPK sequences from A. thaliana, rice, and maize as the query sequences and the A and D genomes of cotton as the reference genomes. The results revealed that there were 39 CIPKs in the A genome (Table 1). On the basis of their order of presence on the chromosomes (from chromosome 1 to 13), the genes were named as GaCIPK1-GaCIPK39, with the first two letters indicating the corresponding species. There were 41 CIPKs in the D genome, which were similarly named as GrCIPK1-GrCIPK41. As demonstrated in Table 1, there were significant differences in the number of amino acid residues (AAs) present in the various CIPKs, which ranged from 196 to 1396. Most of the CIPKs contained between 412 and 480 AAs. GrCIPK37 was the smallest with 196 AAs because it lacked the NAF structure. The pIs ranged from 5.58 to 9.42. Except for GrCIPK2, GrCIPK15, GrCIPK22, GrCIPK30, and GaCIPK16, these five genes were slightly acidic, and the other 75 CIPKs were alkaline, similar to CIPKs of rice, Sorghum, and A. thaliana. Rice contains 30 CIPKs; except for OsCIPK1 and OsCIPK8, which were acidic, the remaining CIPKs were reported to be alkaline (Xiang et al., 2007). Of the 31 CIPKs reported in Sorghum, the proteins of all, except SbCIPK9, were reported to be alkaline (Li et al., 2010). The dicotyledonous A. thaliana contains 25 CIPKs. AtCIPK1, AtCIPK5, and AtCIPK16 were reported to be acidic whereas the others were alkaline (Li et al., 2010). The data indicate that the ancestors of CIPKs in plants were alkaline, whereas the acidic CIPKs evolved later. The molecular weight of CIPKs ranged from 21.92 to 151.57 kDa, with the majority of proteins being in the range from 41.59 to 58.75 kDa. GrCIPK37 had the minimum molecular weight of 21.92 kDa. The molecular weights of GaCIPK36, GaCIPK16, and GrCIPK30 were 81.96, 151.06, and 151.57 kDa, respectively. Two CIPKs, on each of the D (GrCIPK4 and GrCIPK6) and A (GaCIPK14 and GaCIPK17) genomes, were predicted to be localized to the mitochondria. In contrast, the proteins of five CIPKs on the D genome (GrCIPK8, GrCIPK19, GrCIPK30, GrCIPK33, and GrCIPK38) and four CIPKs on the A genome (GaCIPK16, GaCIPK21, GaCIPK28 and GaCIPK36) were predicted to localize to the nucleus. The remaining CIPKs were localized in the cytoplasm, suggesting that the subcellular localization of the proteins of CIPK gene family members in cotton was diverse. It is, therefore, suggested that their corresponding functions should also have diversity, and hence, their subcellular location needs further experimental verification. Previous studies reported that PsCIPK was localized in cytosol and outer membrane in leguminous plants, which was confirmed by immunofluorescence and confocal microscopy (Mahajan et al., 2002). AtCIPK1 could be located in the cell membrane, cytoplasm, and nucleus.

Functional domain analysis of the CIPK family members in cotton

Functional domain analysis of the identified CIPK family members was conducted (http://prosite.expasy.org/) and the results revealed that all the protein sequences of CIPKs in the D and A genomes contained the protein kinase domain (PS50011). Except for GrCIPK37 in the D genome, all CIPKs contained the NAF domain (PS50816). Although GrCIPK37 had no NAF domain, Wang et al. (2012) annotated it as a CIPK; therefore, we also listed it as a CIPK family member in the present study and analyzed it with the other members. Proteins of 41 CIPKs in the D genome and 37 CIPKs (except for GaCIPK6 and GaCIPK12) in the A genome contained the serine/threonine protein kinase active site (PS00108). Except for GrCIPK2, GrCIPK7, and GrCIPK22 in the D genome and GaCIPK7, GaCIPK12, and GaCIPK15 in the A genome, all CIPKs encoded proteins with a protein kinase ATP binding domain (PROTEIN_KINASE_ATP, PS00107). All the CIPKs in the D and A genomes, except for GaCIPK12, contained a proton acceptance locus (ACT_SITE). There were some differences in the distribution of CIPKs and a number of CIPK family members contained five domains; it is speculated that these five domains have different roles in CIPK function. In addition, protein encoded by GrCIPK12 in the D genome and GaCIPK34 in the A genome contained the ribosomal protein L29 marker (RIBOSOMAL_L29, PS00579) and GrCIPK24 contained an EF-hand type calcium-binding domain (EF_HAND_1, PS00018). In contrast, the zinc finger ring type domain (ZF_RING_2, PS50089) appeared in GaCIPK36 but was not found in other CIPK family members. It is speculated that these genes experienced directional evolution in structure resulting from changes in the environment. These structures might have special functional properties, which would need further experimental verification.

Distribution of CIPK gene family members in the complete genome of cotton

The information about the position of genes on chromosomes provides important evidence for the study of evolution and the function of a gene family. Combined with the chromosome information of the cotton A and D genomes and the locations of the CIPKs on the chromosomes, the distribution map of CIPKs on the chromosomes was prepared (Figure 1). As shown in Figure 1A, 38 of the 41 CIPKs were mapped to 11 chromosomes of the D genome of cotton. In comparison, distribution of the remaining three CIPKs was as follows: GrCIPK39, GrCIPK40, and GrCIPK41 were located on scaffold84, scaffold131, and scaffold163, respectively. None of the three genes was located on the corresponding chromosomes. Chromosomes 2 and 12 had no CIPKs; conversely, chromosome 10 had the highest number of CIPKs (seven), followed by chromosome 9 with six CIPKs. Chromosomes 1 and 6 had five CIPKs and chromosomes 4 and 8 had one. As shown in Figure 1B, all CIPKs in the A genome of cotton were located on the chromosomes and were found on the 11 chromosomes; however, neither chromosome 2 nor chromosome 3 contained any CIPK. Chromosome 8 contained the maximum number (eight) of CIPKs, followed by chromosome 11 with five genes. In contrast, chromosomes 12 and 13 contained only one CIPK. The remaining chromosomes contained 2-4 CIPKs each. In the D genome, GrCIPK6 and GrCIPK7 were distributed on chromosome 3 in the form of a cluster. GrCIPK9 and GrCIPK10 were located on chromosome 5 in a close-linking mode. These two gene clusters contained two gene pairs, GrCIPK6/GrCIPK9 and GrCIPK7/GrCIPK10, which had fewer intron types and similar functions. Furthermore, GaCIPK8 and GaCIPK9 were closely linked on chromosome 5 on the A genome, and GaCIPK15 and GaCIPK14 were distributed on chromosome 7 in a cluster, both of which contained two gene pairs, GaCIPK8/GaCIPK15 and GaCIPK9/GaCIPK14, with fewer intron types and similar functions. This distribution pattern might have been caused by the substitution and insertion in chromosomes. A gene family is created by the random amplification of genes, resulting in the formation of gene clusters. The scattered distribution of the members of a gene family on many chromosomes is most likely caused by partial-fragment replication of chromosome regions (Schauser et al., 2005). Compared to other eukaryotes, plants have a higher gene duplication rate (Wei et al., 2014). The results of previous studies showed that gene duplication and separation of the latter stages are the two main objectives of evolution (Chothia et al., 2003), resulting in the diversity of gene family members. Wang et al. (2012) demonstrated that in G. raimondii, complete genome replications occurred at least two times. The scattered distribution mode of the CIPK genes in the complete chromosome might reflect a series of complete genomes, chromosomes, and large-fragment duplication events with typical characteristics of the D and A chromosome genomes. Duplication of genes leads to diversity of gene functions, which plays a very important role driving the evolution of new features, organ differentiation, and better adaptation to changes in environments (Flagel and Wendel, 2009). Phylogenetic analysis reveals that G. arboreum L. and G. raimondii were derived from a common ancestor dating about 5.0 mya, that the genomes of these two species exhibited a high degree of collinearity at the chromosome level, and that the number of genes and sequences were very similar (Li et al., 2014). For example, there were 13 CIPKs on chromosomes 9 and 10 of the D genome, and 13 CIPKs on chromosomes 8 and 11 of the A genome, which might also be the result of gene duplication or partial segment duplication over the long evolutionary history of the cotton genome.

Figure 1

Chromosome distribution of the CIPK genes in the cotton genome. A. CIPK genes in Gossypium raimondii; B. CIPK genes in Gossypium arboreum L.

Genetic structural analysis of the CIPK gene family members in cotton

Gene structure analysis is important in the research on gene evolution. According to the number of introns, the CIPK gene family members of the D and A genomes of cotton were divided into two categories (as shown in Figure 2), one with more than 10 introns and the other with fewer than three introns. In contrast, the category containing less than three introns had 26 members in the D genome, accounting for 63.41% of the members. Except for GrCIPK13, which contained two exons, all the 25 members contained only one exon. The category with more than 10 introns comprised 15 members in the D genome. GrCIPK15 contained only 11 exons and GrCIPK30 contained 20 exons. The rest of the CIPK members had between 12 and 14 exons each. The situation in the A genome was similar to that in the D genome. The category with less than three introns comprised 25 members, accounting for 64.10% of the members, of which GaCIPK4, GaCIPK8, and GaCIPK38 contained two exons and the remaining 22 members contained one exon. The category with more than 10 introns contained 14 members and more than 10 exons. GaCIPK12 had only 11 exons. In contrast, GaCIPK16 had the highest number (20) of exons. It is speculated that the number of exons of each CIPK member in the D and A genomes of cotton was similar during the evolutionary process of CIPK gene structure, which was relatively conserved, suggesting that the functions of these genes might also be consistent. The results of previous studies showed that the insertion of small fragments of DNA could change the function of the gene, and that the gene could disappear by natural selection (Long et al., 1995). CIPK members, GrCIPK30 and GaCIPK16, of the cotton D and A genomes, respectively, had the highest number (up to 20) of exons, with large differences in the exon length, suggesting that there were significant changes in the structures or functions of these two genes in the evolutionary process of cotton. In addition, this indicates that there were relatively stable exon-intron pairs in either the D or A genome during the evolutionary process of cotton. Compared to the introns, the exons were more vulnerable to selective pressures from the external environment. The structures of the exons in many replication-type gene families were generally conserved; therefore, the differences in the structures of exons and introns caused by insertion-deletion events could be used to predict the evolutionary history of gene families (Lecharny et al., 2003).

Figure 2

Intron-exon structure analysis of the CIPK gene family in cotton. Group I and Group II, CIPK genes in Gossypium raimondii; Group III and Group IV, CIPK genes in Gossypium arboreum L.

Gene pair analysis of the CIPK gene family

The results of cluster analysis conducted on the D genome revealed that 41 members of CIPK in the D genome could be divided into 2 categories, one with no or very few introns and the other with several introns (Figure 3A). Similar results were also obtained with the A genome (Figure 3B). In addition, there were 13 pairs of homologous genes in the D and A genomes (Table S2).

Figure 3

Phylogenetic analysis of the CIPK proteins in the D and A genomes in cotton. A. Phylogenetic trees of the CIPK proteins in D genome. B. Phylogenetic trees of the CIPK proteins in A genome. C. Phylogenetic trees of the CIPK proteins in D and A genomes.

The cluster analysis of the D and A genomes (Figure 3C) also illustrated that 36 CIPKs in the D genome were paired with 36 CIPKs the A genome (Table S2). The formation of such gene pairs might have been caused by duplication events. The sequence consistency of proteins encoded by the CIPK gene pairs of the D genome was from 65.53% (GrCIPK7/GrCIPK10) to 92.2% (GrCIPK25/GrCIPK40), with an average of 82.51% consensus in the sequences. By comparison, the proteins encoded by the CIPK gene pairs of the A genome had a sequence consistency from 65.53% (GaCIPK8/GaCIPK15) to 92.71% (GaCIPK3/GaCIPK29), with an average consensus of 82.79%. In contrast, the consensus in the sequences of corresponding proteins encoded by CIPK family gene pairs of the D and A genomes ranged from 89.05% (GrCIPK14/GaCIPK36) to 99.73% (GrCIPK2/GaCIPK3), with an average consensus of 97.47%. Of these pairs, seven showed >99.00% consensus; these pairs were as follows: GrCIPK32/GaCIPK13, GrCIPK26/GaCIPK22, GrCIPK37/GaCIPK39, GrCIPK6/GaCIPK14, GrCIPK34/GaCIPK26, GrCIPK40/GaCIPK35, and GrCIPK2/GaCIPK3. These results indicate that the interspecies genetic relationship between the CIPK gene family members of cotton is stronger than the intraspecies relationship. They also suggest that the common ancestor of G. arboreum L. and G. raimondii formed the CIPK gene family during the evolutionary process. After the differentiation of the two species about 5.0 mya from the same progenitor, the CIPK gene family experienced linear synchronous evolution in the two species. These comprehensive analyses demonstrate that a gene duplication event had a significant effect on the amplification of the CIPK gene family members in cotton, and that the amplification of the CIPK members might be the result of a high-stress environment. Cotton is chronically affected by an adverse environment during the growth period, which promotes the widespread amplification of genes with similar functions. Moreover, the newly formed CIPK genes must have evolved with the corresponding CBL genes. These functional limitations might have helped in the evolution of the current CBL-CIPK signaling network (Kolukisaoglu et al., 2004).

Phylogenetic analysis of the CIPK gene families of cotton and four other species

To test the evolutionary relationship between the CIPK gene family members in cotton and those in Arabidopsis, rice, maize, and poplar, the amino acid sequences of all CIPK members from these species were compared and a phylogenetic tree was constructed (see Material and Methods). The analysis of the phylogenetic tree revealed that CIPK of the D and A genomes of cotton were mostly clustered together (Figure 4). The relationship between G. raimondii and G. arboreum L. was closer than between the other four species. In addition, many members of the CIPK gene family of cotton and poplar clustered together, suggesting that these two species were evolutionarily close; this was followed by the evolutionary closeness of cotton with Arabidopsis (Table S3). These genes are similar in structure and it is predicted that their functions might also be similar. There were more family members in cotton than in Arabidopsis, rice, and poplar plants, suggesting a specific linear amplification of the gene family in cotton. Whether these additional members of the gene family have additional functions, as well, or whether they are only produced because functional redundancy requires further experimental verification.

Figure 4

Phylogenetic analysis of the CIPK proteins in cotton, Arabidopsis, rice, maize, and poplar. The species used for the construction of the phylogenetic tree are Gr, Gossypium raimondii; Ga, Gossypium arboreum L.; At, Arabidopsis; Os, rice; Zm, maize; and Pt, poplar.

CIPK expression analysis based on the transcriptome sequencing

Transcription analysis has been used for identifying the protein-coding genes during the annotation of the D and A genomes (Wang et al., 2012; Li et al., 2014). In this study, the sequence read lengths of the matched CIPK genes were converted to RPKM values and used to predict the gene expression levels. The data were retrieved from NCBI and at least 20 nucleotide sequences were selected and used in this study.

The abundance of transcripts in the cotton leaves at the beginning of flowering stage, and in 0 and 3.0 DPA ovules was detected (Figure 5). The results showed that about 9.76% of the CIPK genes (GrCIPK12, GrCIPK18, GrCIPK30, and GrCIPK31) in the D genome were expressed in the leaves (Figure 5A). Of these, the expression of GrCIPK18 and GrCIPK31 in the leaves was much higher than that of GrCIPK12 and GrCIPK30, indicating that these two genes play an important role in leaf development and that GrCIPK18 and PtCIPK13 are homologous. Nine CIPK genes (21.95%) were expressed in the 0-DPA ovules and eleven (26.83%) were expressed in the 3.0-DPA ovules. Compared to the expression in the 0-DPA ovules, the expression of GrCIPK4, GrCIPK7, GrCIPK8, GrCIPK30, GrCIPK31, GrCIPK36, and GrCIPK40 in the 3.0-DPA ovules was increased by 1.19-, 1.09-, 1.02-, 3.02-, 3.87-, 2.82-, and 1.78-times, respectively, indicating that these genes play a major role in fiber elongation. GrCIPK12, GrCIPK18, GrCIPK23, and GrCIPK26 were all downregulated with different extent in 3.0-DPA ovules compared with that in 0-DPA ovules, indicating that these genes play important roles in the initiation of fiber formation in cotton. Of the CIPK genes in the A genome, approximately 41.03% (sixteen) were expressed in the leaves, with five genes being highly expressed in the leaves of G. arboreum L. (Figure 5B). GaCIPK8, GaCIPK14, GaCIPK22, GaCIPK33, GaCIPK39, GaCIPK8, and OsCIPK2 were observed to be homologous. In contrast, 35.90% (fourteen) of the CIPK genes were expressed in the 0-DPA ovules, with five genes, GaCIPK4, GaCIPK8, GaCIPK14, GaCIPK33, and GaCIPK39, being highly expressed. One-third (13) of the CIPK genes were expressed in the 3.0-DPA ovules in G. arboreum L., and two of these, namely GaCIPK8 and GaCIPK39, were highly expressed. Compared to the expression in 0-DPA ovules, only GaCIPK2 was significantly upregulated (1.02-times) in the 3-DPA ovules, indicating that this gene plays a main role in fiber elongation. Compared to the expression in the 0-DPA ovules, the expression of three CIPK genes, namely GaCIPK7, GaCIPK14, and GaCIPK32, was downregulated in the 3.0-DPA ovules by 1.95-, 1.54-, and 1.27-times, respectively, indicating that these three genes play important roles in the initiation of cotton fiber. The expression of another nine CIPK genes (GaCIPK4, GaCIPK8, GaCIPK10, GaCIPK15, GaCIPK20, GaCIPK28, GaCIPK31, GaCIPK33, and GaCIPK39) in the 0- and 3.0-DPA ovules was similar, suggesting that they all play important roles in the initiation and development of fiber. GaCIPK22 and GaCIPK25 were expressed only in the leaves, and GaCIPK4 was expressed only in the 0-DPAovules, indicating that the expression of these three genes was tissue specific. GaCIPK22 and GrCIPK26 belong to homologous gene pairs, and GaCIPK22 was abundantly expressed only in the leaves. The expression of GrCIPK26 in the 0-DPA ovules was high, suggesting that homologous genes might also have functional diversity.

Figure 5

Transcript abundance analysis of the CIPK genes in mature leaves, 0-DPA and 3.0-DPA ovules in Gossypium raimondii and Gossypium arboreum L. The transcript expressions were calculated using the RPKM method (Mortazavi et al., 2008).

Expression analysis of the CIPK genes in diploid cottons under different stress conditions

To investigate the anti-adversity function of CIPK genes in cotton, their expression patterns were studied in different tissues of diploid cottons exposed to different stress conditions (low temperature, drought, and salt stress). The results showed that 80 CIPK genes were expressed in the roots, stems, and leaves of diploid cottons, but the expression levels of most of the genes were different in the roots, stems, and leaves after exposure to abiotic stresses (Figure 6).

Figure 6

Expression of the CIPK genes in different tissues of diploid cotton under different abiotic stress conditions. Gene expression analysis of three different tissues of diploid cotton species was launched under different stresses with qRT-PCR method. Red color represented genes upregulated and the expression folds were greater than or equal to 5; Pale red color represented the genes upregulated with expression folds was greater than 1 and less than 5; black color represented the genes with expression amount was almost unchanged; green color represents genes downregulated and the expression folds were less than 1.

Sixteen, 28, and 14 CIPK genes were upregulated in the roots, stems, and leaves of G. raimondii, respectively and 3, 11, and 22 genes were downregulated, respectively, after 4°C treatment. In the roots, stems, and leaves of G. arboreum L. 29, 15, and 22 CIPK genes were upregulated and 10, 19, and 16 genes were downregulated, respectively, after 4°C treatment. However, the expression levels of CIPK genes varied. Moreover, similar expression patterns of CIPK genes were observed under both drought and salt stress. In different tissues, the number of CIPK genes up- or downregulated was different, showing tissue specificity. The results also showed that 11 CIPK genes in G. raimondii and G. arboreum L. were upregulated in the leaves, stems, and roots simultaneously after low temperature treatment (Table S4). In addition, 12 and 15 CIPK genes were simultaneously upregulated in the leaves, stems, and roots in both the diploid cotton species after drought and salt treatment, respectively. These results suggest that these genes could be induced by low temperature, drought, and salinity, and thus, play an important role in the resistance of cotton to the imposed stress. Moreover, differences in the expression of CIPK genes under different stress conditions reveal specificity in the expression.

Analysis of anti-adversity function of CIPK genes in cotton showed that 9 (GrCIPK20/21/23/31, GaCIPK15/19/21/22/35), 14 (GrCIPK6/7/21/22/31/36/37, GaCIPK3/5/6/7/10/14/15), and 7 (GrCIPK22/34/39 GaCIPK32/33/34/36) genes showed upregulation in leaves, stems, and roots, respectively, after low temperature, drought, and salt stress, indicating that these genes could respond to various adversities. Therefore, we speculate that these CIPK genes could respond to many other stresses, and therefore, play a vital role in stress resistance.