Foxtail millet (Setaria italica (L.) P. Beauv) CIPKs are responsive to ABA and abiotic stresses

CBL-interacting protein kinases (CIPKs) have been shown to regulate a variety of environmental stress-related signalling pathways in plants. Foxtail millet (Setaria italica (L.) P. Beauv) is known worldwide as a relatively stress-tolerant C4 crop species. Although the foxtail millet genome sequence has been released, little is known about the functions of CIPKs in foxtail millet. Therefore, a systematic genome-wide analysis of CIPK genes in foxtail millet was performed. In total, 35 CIPK members were identified in foxtail millet and divided into four subgroups (I to IV) on the basis of their phylogenetic relationships. Phylogenetic and gene structure analyses clearly divided all SiCIPKs into intron-poor and intron-rich clades. Cis-element analysis subsequently indicated that these SiCIPKs may be involved in responses to abiotic stimuli, hormones, and light signalling during plant growth and development, and stress-induced expression profile analysis revealed that all the SiCIPKs are involved in various stress signalling pathways. These results suggest that the CIPK genes in foxtail millet exhibit the basic characteristics of CIPK family members and play important roles in response to abiotic stresses. The results of this study will contribute to future functional characterization of abiotic stress responses mediated by CIPKs in foxtail millet.


Introduction
It is well known that abiotic stresses such as drought, salinity and low temperature severely affect the yield and quality of crops. To cope with such adverse conditions, plants have developed elaborate and systematic mechanisms during their evolution [1]. Many CBL-interacting protein kinases (CIPKs) are involved in the response and adaptation to stresses [2]. The genes encoding these serine/threonine protein kinases constitute an important, widespread multigene family in the plant kingdom [3]. Sequence analysis has indicated that CIPK proteins have a relatively conserved N-terminal kinase domain and a less conserved C-terminal regulatory domain [4]. A highly conserved 24-amino acid region within the C-terminal regulatory domain of CIPK designated as the NAF domain (Pfam No. PF03822) is required for interaction with CBLs during exposure to stress [5]. a1111111111 a1111111111 a1111111111 a1111111111 a1111111111 various physiological functions in response to multiple stress stimuli, especially abiotic stresses.
Foxtail millet (Setaria italica (L.) Beauv) is a C 4 graminaceous crop species that originates from China, has a long (approximately 7,000 years) history of cultivation and is widely planted in northern China and in other Asian countries [35]. Foxtail millet is a hardy cereal known for its superior tolerance to drought, salinity, and diseases; high nitrogen-and water-use efficiencies; and nutritional properties [36]. The potential abiotic stress tolerance of foxtail millet has motivated the research community to study the molecular mechanisms of this species [37]. Although the foxtail millet genome sequence has been completed and released [35,38], the CIPK gene family in foxtail millet has not been fully characterized. Therefore, it is very important to identify CIPK genes in foxtail millet and evaluate their different responses to abiotic stresses. In this study, we identified 35 CIPK genes from the foxtail millet genome and analysed their genetic organization, gene structure and conserved motifs, evolution, and stress-related cis-elements. We also evaluated the gene expression profiles of the SiCIPKs under ABA and abiotic stresses. These systematic analyses revealed that the CIPKs in foxtail millet exhibit the basic characteristics of CIPK family members and may be important contributors to the superior tolerance of this species. The SiCIPK genes reported in this study will enrich our knowledge of CIPKs in the plant kingdom, and our results lay a foundation for revealing the functions and mechanisms of stress responses mediated by the CBL/CIPK pathway in foxtail millet.

Genome-wide identification of CIPK genes in foxtail millet
To identify the genes encoding CIPK proteins in foxtail millet, the 8.3X assembled version (V2.2) of the foxtail millet genome was downloaded from Phytozome V12.1 (https://genome. jgi.doe.gov) [38]. The CIPK genes were identified via the methods described by Hu [3]. Each CIPK candidate sequence was examined for the presence of the NAF domain and protein kinase domain necessary for consideration as a member of the CIPK family.

Protein properties and sequence analyses
The protein parameters were evaluated, and sequence analysis was performed according to the method described by Tang [39]. Prediction of the molecular weight (MW) and isoelectric point (pI) of each SiCIPK protein was conducted using the online tool compute pI/Mw (http://web.expasy.org/compute_pi/). Motif prediction of SiCIPKs was carried out using MEME software (http://meme-suite.org/tools/meme) [40]. The maximum number of domains was set to 15, the width of functional domains was set to 6-60, and the other parameters were set to the default values. The motifs identified by MEME software were annotated on the PRO-SITE website (http://prosite.expasy.org/). The Gene Structure Display Server (GSDS 2.0, http://gsds.cbi.pku.edu.cn/) was then used to analyse the SiCIPK gene structure by comparing cDNAs with the corresponding genomic sequences [41].

Plant materials, growth conditions and treatments
Seeds of the foxtail millet cultivar Yugu 1 were obtained from the Millet Research Institute, Shanxi Academy of Agricultural Sciences, sown in pots and grown in a controlled incubator (16-h light/8-h dark photoperiod, 28˚C day/20˚C night temperature, 60-70% humidity) [12]. After reaching the three-leaf stage, the seedlings were transferred and precultured for 24 h in half-strength Hoagland's solution with aeration. Stress treatments were applied by transferring the seedlings to the same solution containing 20% polyethylene glycol (PEG) (MW of 6,000), 250 mM NaCl or 100 μM (+)-cis, trans-ABA [12]. For cold treatment, the seedlings were transferred to a growth chamber set at 4˚C. Whole plants under a time series treatment (0, 3, 6, 12 and 24 h; three biological replicates) were harvested, immediately frozen in liquid nitrogen and stored at -80˚C for RNA extraction.

Real-time quantitative PCR (RT-qPCR)
Total RNA was extracted from the samples via RNAiso reagent (TaKaRa, Dalian, China) according to the manufacturer's instructions. Expression of the SiCIPK genes in response to PEG, salt, cold and ABA was examined by RT-qPCR analysis via SYBR Green qPCR Master Mix (TaKaRa, Dalian, China) in conjunction with a LightCycler 480 instrument (Roche Diagnostics, Penzberg, Germany). PCR was performed according to the manufacturer's protocol. Each 20-μL amplified PCR mixture comprised the following components: 10 μL of 2X SYBR Green qPCR Master Mix (BBI, Shanghai, China), 2 μL of cDNA template, 0.4 μL of each primer (10 μM), and 7.2 μL of sterile distilled water. The PCR procedure consisted of 95˚C pre-denaturation for 3 min, followed by 45 cycles of 95˚C denaturation for 7 s, 57˚C annealing for 10 s and 72˚C extension for 15 s. Gene-specific primers were designed with Primer Premier 5.0 software on the basis of the nonconserved sequence of each SiCIPK gene. The β-actin gene of foxtail millet (Seita.7G294000) was used as an internal control for normalization, as described by Xu [45]. All primers used are listed in S1 Table. RT-PCR was performed three times for each biological sample, and the relative expression of the SiCIPKs under the different stresses was calculated according to the relative 2 -ΔΔCt method [46]. The values are given as the means ± SEs of three different experiments with three replicate measurements. Statistical analysis was performed via Student's t test.

The foxtail millet genome encodes 35 CIPK genes
We searched for candidate CIPK genes in the foxtail millet genome according to the methods described by Hu [3]. After the genes were systematically analysed, 35 SiCIPKs were identified as candidate CIPK genes. To facilitate recording, we named these genes according to their corresponding locations in the genome. Like other CIPKs, all identified SiCIPKs contain a conserved N-terminal catalytic kinase domain and a C-terminal regulatory domain (S1 Fig). In addition, all SiCIPKs possess an activation domain within the N-terminal sequence and a highly conserved NAF domain within the C-terminal sequence (S1 Fig). The proteins range in size between 402 and 983 amino acids. The relative MWs vary from 45.51 to 112.42 kD, and most (77.14%) have high pIs (pI > 7.0). Detailed information concerning various CIPK parameters is provided in Table 1.

Genomic location, phylogeny and genetic organization of the SiCIPK gene family
The SiCIPK genes were mapped to all nine foxtail millet chromosomes. However, SiCIPK35 was located in scaffold 47 because of a lack of complete chromosomal information. As shown in Fig 1, the SiCIPK genes are distributed across the 9 chromosomes, and similar to that which occurs for other CIPK family members, the SiCIPK genes are unevenly distributed across the genome [26,39]. Among all the chromosomes, chromosome 5 contains the most (7) SiCIPK genes (20%), while chromosomes 2 and 3 each contain 6 genes (~17%), and at least one gene is located on chromosome 6 (~2.9%). On the basis of their phylogeny and gene structure, the 35 CIPK family members were divided into two clades with four subgroups (subgroup I to subgroup IV) (Fig 2A). Among the subgroups, subgroup I contains 4 members; the largest one, subgroup II, comprises 16 members. Subgroups III and IV contain 11 and 4 members, respectively. Exon/intron organization analysis further revealed that the SiCIPK genes can be clearly divided into an intron-poor clade (< 3 introns per gene, subgroups I and II) and an intronrich clade (> 10 introns per gene, subgroups III and IV). In subgroups I and II, only

Conserved motif and evolution analysis of SiCIPK
To better understand the structural characteristics of the SiCIPK genes and provide clues to their functions, the conserved motifs of the SiCIPKs were analysed. We found 15 main conserved motifs, with all SiCIPK proteins having four core motifs: motifs 1, 4, 7 and 9 ( Fig 3B). According to PROSITE, motifs 1, 2 and 3 are annotated as protein kinase domains. Motif 9 harbours the core NAF (Pfam No. PF03822) residues and is annotated as the NAF domain. The kinase domain and the NAF domain are the two major domains typical of CIPK proteins [4]. The NAF domain is necessary for binding between CBL and CIPK proteins, and formation    of a stable complex of CBL and CIPK proteins is a prerequisite for the CBL-CIPK network to participate in the regulation of plant abiotic stress responses [5]. Analysis of the conserved motifs demonstrated the conserved structure of the members of the SiCIPK gene family (Fig 3  and S2 Table), which is consistent with the conserved nature of CIPK family members in other plant species [47,48]. To determine the origin and evolution of SiCIPKs, we constructed a phylogenetic tree of foxtail millet and other species (Arabidopsis, rice, maize, sorghum, moss, Picea, green algae and fungi) via the software ClustalW 2.0 and MEGA 6.0 on the basis of fulllength protein sequences. The analysis revealed that all CIPKs in the nine species could be divided into two clades and four subclades (I to IV) (Fig 4), which is consistent with the evolutionary analysis of SiCIPK in foxtail millet (Fig 2). The CIPKs in these plants evolved into their present states through continuous expansion of four subgroups, which is also consistent with the general consensus concerning the evolution and development of plant CIPKs [49,50]. In addition, we found that the angiosperm (rice, Arabidopsis, maize, sorghum, and foxtail millet) and gymnosperm (Picea) CIPK proteins among the nine plant species could be divided into intron-poor and intron-rich clades (Fig 4). These results revealed that the intron-poor/rich clades of CIPKs may have originated before the divergence of gymnosperms and angiosperms approximately 300 Myr ago [51].

Cis-regulatory elements and expression profile analysis of SiCIPKs in response to stress
Functional analysis of many identified CIPK family genes has demonstrated their important role in response to abiotic stress in plants [52,53]. To determine the possible responses of SiCIPKs to abiotic stress, we examined the promoter regions of SiCIPKs for typical stressresponsive elements. We found a large number of stress-responsive elements within these promoter regions, including ABA-responsive elements (ABREs), MYB-binding sites (MBSs) involved in drought inducibility, defence and stress-responsive elements (TC-rich repeats) and low-temperature-responsive (LTR) elements (S3 Table). Additional stress-responsive ciselements within SiCIPK gene promoter regions were found, including C-repeat/DRE, motif IIb, WUN-motif, CE3, fungal elicitor (Box W1), auxin (TGA-element) and endosperm expression (GCN4 motif and Skn-1 motif) elements, as well as those essential for anaerobic induction (ARE) (S4 Table). To verify the results of the cis-element analysis and further reveal the responses of SiCIPK genes to different stresses, the expression of all putative SiCIPK genes was measured via real-time PCR after PEG, salt, cold or ABA treatment. Among all the SiCIPKs, two (SiCIPK12 and -14) were not detected because of their low expression. However, the expression of all the other 33 SiCIPK genes was upregulated in response to at least one stress; specifically, the expression of 32, 28, 33 and 33 genes was upregulated under ABA, PEG, cold and salt treatments, respectively (S5 Table). Our results clearly suggest that different SiCIPK genes exhibit different responses to stress (Figs 5 and 6). Furthermore, the transcript levels of some SiCIPK genes did not change obviously under certain treatments (e.g., SiCIPK1, -25 and -32 under ABA treatment; SiCIPK1, -8, -25, -28, -30, and -32 under PEG treatment; and SiCIPK1 under cold treatment). The expression of these genes varied little, and their level fluctuated around that of the controls (S6 Table). The expression of some genes was induced mainly by one or two stresses. For example, the expression of SiCIPK1, -25, -28 and -32 was strongly induced by only cold stress, and that of SiCIPK8 was induced by cold and salt stress but not by ABA or PEG (S5 Table). These experimental data indicate that the CIPK family members in foxtail millet play important roles in stress responses.

SiCIPK genes are involved in numerous common regulatory systems or cross-talk triggered by different stresses
To further investigate the responses of SiCIPKs to different stresses, we considered SiCIPKs to be stress-induced genes when their expression levels were more than twofold those of the controls for at least one time point. Among the 33 SiCIPK genes detected, the expression of 22, 26, 28 and 31 was induced by PEG, ABA, salt and cold treatments, respectively (Fig 5). We also found some SiCIPK genes to be responsive to multiple stresses. Among the 22 PEG-inducible genes, the expression of 20, 20, and 21 was also induced by ABA, cold and salt, respectively, and among the 28 salt-inducible genes, the expression of 25, 21, and 26 was also induced by ABA, PEG and cold, respectively. Similarly, among the ABA-and cold-inducible genes, the expression of a large number was induced by the other three treatments. Remarkably, the expression of 17 genes was induced by all four stress treatments (PEG, salt, cold and ABA) evaluated in this study (Fig 5). These results indicate that SiCIPK genes may be involved in numerous common regulatory systems or cross-talk triggered by different stresses. As many abiotic stresses lead to dehydration of plant cells and osmosis imbalance at the cellular level [54], the expression patterns of some stressrelated genes in plants may overlap after exposure to various abiotic stresses such as drought, salt and cold as well as ABA. This phenomenon suggests that ABA and these stress signals share some elements in their signalling pathways and that these elements interact with each other to achieve a new cellular equilibrium [55,56]. Our RT-PCR results also support this inference.

Discussion
CIPKs are highly conserved and widely present in the plant kingdom. CIPK family members have been identified and characterized in many plant species [26,39,47,48,57,58], and a large number of CIPK genes have been reported to be involved in the response to complex environmental stimuli in higher plants [59]. Although the foxtail millet genome has been sequenced and released [35,38], there are very few reports of CIPKs in foxtail millet, especially in terms of responses to abiotic stress. In this study, 35 CIPK genes were identified in the foxtail millet genome and were mapped onto the nine foxtail millet chromosomes. Like CIPK family members in other plant species [4], SiCIPKs contain a conserved N-terminal catalytic kinase domain, a C-terminal regulatory domain and highly conserved NAF domain (S1 Fig). Phylogenetic analysis divided the 35 CIPK family members of foxtail millet into two clades, including four subclades: I, II, III and IV (Fig 2A). One clade contains subclades I and II, and the other clade contains subclades III and IV. In addition, exon/intron organization analysis confirmed this classification. All SiCIPK genes were clearly divided into two distinct groups: an intron-poor group and an intron-rich group. The intron-poor group (< 3 introns per gene) included subgroups I and II, and the intron-rich group (> 10 introns per gene) included subgroups III and IV (Fig 2B). Gene structure analysis of the SiCIPKs supported the phylogenetic groupings of the SiCIPK family, and the intron-rich/poor pattern of the CIPK family in foxtail millet is similar to that in other species [60,61]. Subsequent multiple sequence alignment analysis of CIPKs from nine representative plant species also revealed that all CIPKs could be divided into two groups (intron-poor and intron-rich groups) and four subgroups (I to IV) (Fig 4). This finding further confirmed the results of our multiple sequence alignment, conserved motif, and phylogenetic and organization analyses of SiCIPKs in foxtail millet (Figs 2 and 3). Moreover, phylogenetic analysis of the CIPKs from the nine species supported the amplification models of the plant CIPK gene family originating before the genome duplication within the four subgroups [49,50,60]. The divergence of angiosperms and gymnosperms occurred approximately 300 Myr ago [51], and we found many intron-poor and intron-rich CIPK genes in both angiosperms (rice, Arabidopsis, maize, sorghum, and foxtail millet) and gymnosperms (Picea) (Fig 4). This phenomenon revealed that the intron-poor/rich clade grouping may have originated before the separation of gymnosperms and angiosperms. These data indicated that intron gain and loss events have played important roles in the evolution of the CIPK family. Evolutionary analysis revealed many homologous gene pairs in foxtail millet and sorghum, maize, and rice, indicating a close relationship among these crop plant species, which is consistent with the current understanding of plant evolutionary history. In addition, the detection of homologous CIPK gene pairs with high sequence similarity from different species suggested that these genes may have similar functions in adaptation or evolution. There are many reports of highly homologous genes with similar molecular and biological functions in plants. For example, MdCIPK6L, MdSOS2, ZmCIPK16, and SiCIPK24, which are orthologous to AtCIPK24/AtSOS2, reportedly increase plant salt tolerance, with functions similar to those of AtSOS2 [9][10][11][12]. AtCIPK8, an orthologue of MeCIPK8, is involved in regulating the low-affinity phase of the primary nitrate response [21]. AtCIPK21, which is highly similar to MeCIPK21, has a positive role in both osmotic and salt stress responses [62]. This evidence suggests possible roles for SiCIPKs and provides clues for future functional studies of SiCIPK genes in response to stress.
In general, the presence of a cis-element indicates that a gene may be involved in the response to the corresponding signal [63]. Abiotic stresses such as drought, salinity, low temperature, heat and wounding can lead to an increase in ABA in plant cells, and it is well known that ABA plays an important role in response to abiotic stress in plants [64][65][66]. Our cis-element analysis revealed that many stress-responsive cis-elements, such as ABRE, MBS, TC-rich repeats, LTR, C-repeat/DRE, motif IIb, and WUN-motif elements, are widely distributed within the promoter regions of SiCIPK members (S3 Table). Further stress expression profile analysis revealed that the expression of 33 SiCIPK genes was induced by at least one stress. The expression patterns and characteristics of these 33 genes after exposure to stress were diverse, with the expression of a few genes being induced by one stress (SiCIPK1, -25, -28 and -32 were strongly induced by cold) and the expression of most genes being induced by multiple stresses. Compared with those of the controls, some gene expression levels under a specific stress treatment did not change obviously, while the expression levels of some genes did change considerably (S6 Table, Figs 5 and 6). Furthermore, we found that the expression of 17 SiCIPK genes was simultaneously induced by the four stresses applied, indicating that almost half of the SiCIPK genes are involved in common regulatory systems or cross-talk triggered by different stresses. In addition, the expression of most of the SiCIPK genes was strongly induced under salt and cold treatments, but that of only a few SiCIPK genes was strongly induced under ABA and PEG treatments (Fig 6), indicating that foxtail millet seedlings are more sensitive to lowtemperature and salt stresses than to ABA and PEG stresses. Thus, when foxtail millet seedlings are subjected to various adverse conditions, they will preferentially perceive and initiate responses to low temperature and salt. These results are consistent with the characteristics of foxtail millet, which is highly resistant to both salt and low-temperature stresses [36]. In addition, some hormone response elements were found in the promoter regions of all SiCIPK members, including GARE-motif, AuxRR-core, CGTCA-motif, TGACG-motif, ERE, TGAelement, TGA-box, TCA-element, SARE, P-box, TATC-box, ABRE, and motif IIb elements (S4 Table). Pathogen infection usually leads to increases in salicylic acid (SA), jasmonic acid (JA) and ethylene (ET) levels in cells, and these hormones are thought to play major roles in biotic stress responses [67]. The large number of plant hormone response elements within the SiCIPK promoter regions further suggested that SiCIPKs might play important roles in plant biotic stress responses. As a crop species that prefers short days and warm climates, foxtail millet is sensitive to photoperiod [68], and the discovery of many light-responsive elements suggests that SiCIPKs likely play very important roles in response to light. In short, the large variety of cis-elements widely distributed within the SiCIPK genes and the results of the expression profile analysis indicated that foxtail millet CIPK family members are likely involved in many biological processes, including both abiotic and biotic stress responses and hormone signalling during plant growth and development. The results of this experiment not only verified the predictions for stress-related cis-elements but also further confirmed that SiCIPK genes are involved in biological responses to stress and greatly contribute to the ability of foxtail millet to cope with adverse conditions. Although the functions of these SiCIPKs must still be analysed in detail, our work provided clues for future research on the functions of SiCIPK genes.
We also noticed that the long N-terminal sequences of SiCIPK29 and SiCIPK30 render them the two longest SiCIPK genes (Figs 2 and 3). Sequence comparison revealed that SiCIPK29 and SiCIPK30 are highly homologous, with 81.36% sequence identity. Moreover, functional domain analysis revealed that, compared with other SiCIPKs, these proteins contain two copies of the N-terminal functional domain. Thus, their long sequence appears to be due to repetition of the N-terminal kinase domain, and we speculate that these two long genes may have evolved into their current states via replication of the N-terminal sequence. Regardless, no notable differences were found between them and the other SiCIPK genes with respect to cis-elements and expression under stress. We speculate that these genes may have additional functions in other signalling pathways. Remarkably, the expression of seven genes (SiCIPK6, -8, -10, -19, -21, -24 and -34) was highly upregulated and was higher (� 20-fold) in the plants under stress than in the untreated control plants (Fig 7). Analysis of these highly induced expression profiles during stress treatments revealed several types of different expression profiles. In one type, which included SiCIPK6, -8, -21, -26, and -34 under cold treatment, gene induction was rapid and transient in response to stress treatment, reaching a maximum at 24 h. In the second type (including SiCIPK10, -19, and -21 under ABA treatment; SiCIPK21 and -24 under salt treatment; and SiCIPK19 under PEG treatment), gene expression was induced after stress, peaked at 12 h (SiCIPK10 under ABA treatment peaked at 3 h), and then decreased. In the third group, which contained only SiCIPK24 under PEG treatment, gene expression was induced after stress, peaked at 3 h, decreased from 3 to 6 h, and then increased again at 24 h. These types represent the typical expression profiles of SiCIPK genes whose expression is induced under stress conditions. Together, these results indicate that members of the SiCIPK gene family exhibit stimulus-specific and time-dependent responses. It has been reported that the products of stress-inducible genes can be classified into two groups. The first group consists of functional proteins or proteins that probably directly protect against environmental stresses. The second group consists of regulatory proteins involved in further regulation of signal transduction and in the expression of genes that probably function in stress responses [69,70]. Notably, SiCIPK genes belong to the second group, and they may be involved in further signal transduction and downstream gene expression in response to stress. Stress-inducible genes have been used to improve the stress tolerance of plants by gene transfer [71,72]. It is important to analyse the functions of stress-inducible SiCIPK genes, not only to understand the molecular mechanisms of stress tolerance and the responses of higher plants but also to improve the stress tolerance of crops by genetic manipulation.
In summary, using sequence alignment methods, we identified 35 CIPK genes from the graminaceous C 4 crop species foxtail millet. The members of the SiCIPK family have similar characteristic parameters and share high sequence identity with other plant CIPK genes according to protein parameter prediction, functional domain analysis, sequence alignment, and evolutionary analysis. Promoter cis-element analysis revealed that a large number of cis-elements related to stresses, hormones, light, and other growth and development processes were found in the promoter regions of the SiCIPK gene family members. qRT-PCR validation experiments revealed that the expression of 33 SiCIPK genes was induced by different stresses and that the expression of 17 SiCIPK genes was induced by four stresses simultaneously. Moreover, we found that foxtail millet seedlings are more sensitive to low-temperature and salt stresses than to ABA and PEG stresses. We also found that the expression of seven SiCIPK genes was highly induced by different stresses (S2 Fig). These results show that SiCIPKs exhibit the basic characteristics of CIPK family members and play important roles in response to abiotic stresses. It has been reported that intracellular Ca 2+ concentrations temporarily fluctuate when plants are exposed to drought, salt, low temperature and ABA. As a Ca 2+ sensor, CBL first perceives this change and interacts with its downstream CIPK to form a CBL/CIPK complex. The CBL/ CIPK complex then phosphorylates downstream target proteins, relays various dynamic calcium signals and regulates related physiological processes. This activity causes a series of intracellular biochemical reactions, which enable plants to resist or adapt to various stresses [5]. The CBL/CIPK network is very complex. One CBL can interact with multiple CIPKs, and one CIPK can interact with multiple CBLs [23]. Accordingly, our future research will focus on elucidating the molecular mechanism of the interaction between SiCBLs and the highly induced SiCIPKs. Further identification of key components of the CBL/CIPK signalling pathway and analysis of related functions will allow a better understanding of how foxtail millet responds to abiotic stresses via the CBL/CIPK signalling system. SiCIPKs whose expression is strongly induced by stress treatments. The relative expression levels of SiCIPKs were quantitatively calculated according to the 2 -ΔΔCt method. Gene expression was measured at six time points (0 h, 1 h, 3 h, 6 h, 12 h and 24 h). The Y-axis indicates the relative expression level, and the error bars represent the standard deviations calculated on the basis of three technical replicates for each biological duplicate. The asterisks indicate significant differences between the treatment and control according to Student's t test ( �� p < 0.01 and � p < 0.05). https://doi.org/10.1371/journal.pone.0225091.g007

Conclusions
In this study, 35 SiCIPK genes were identified in foxtail millet, a C 4 Gramineae crop species known for its outstanding stress tolerance. Our systematic analysis revealed that SiCIPKs have basic characteristics similar to those of CIPK family members identified from other plant species, and we found that SiCIPK genes play important roles in stress responses. The results of this study increase the number of known CIPK members in plants and provide an experimental basis for further elucidating the functions and mechanisms of abiotic stress responses mediated by CIPKs in foxtail millet.