Genome-Wide Identiﬁcation and Characterization of the CCT Gene Family in Foxtail Millet ( Setaria italica ) Response to Diurnal Rhythm and Abiotic Stress

: The CCT gene family plays important roles in diurnal rhythm and abiotic stress response, affecting crop growth and development, and thus yield. However, little information is available on the CCT family in foxtail millet ( Setaria italica ). In the present study, we identiﬁed 37 putative SiCCT genes from the foxtail millet genome. A phylogenetic tree was constructed from the predicted full-length SiCCT amino acid sequences, together with CCT proteins from rice and Arabidopsis as representatives of monocotyledonous and dicotyledonous plants, respectively. Based on the conserved structure and phylogenetic relationships, 13, 5, and 19 SiCCT proteins were classiﬁed in the COL , PRR , and CMF subfamilies, respectively. The gene structure and protein conserved motifs analysis exhibited highly similar compositions within the same subfamily. Whole-genome duplication analysis indicated that segmental duplication events played an important role in the expansion of the CCT gene family in foxtail millet. Analysis of transcriptome data showed that 16 SiCCT genes had signiﬁcant diurnal rhythm oscillations. Under abiotic stress and exogenous hormonal treatment, the expression of many CMF subfamily genes was signiﬁcantly changed. Especially after drought treatment, the expression of CMF subfamily genes except SiCCT32 was signiﬁcantly up-regulated. This work provides valuable information for further study of the molecular mechanism of diurnal rhythm regulation, abiotic stress responses, and the identiﬁcation of candidate genes for foxtail millet molecular breeding. study, we identiﬁed 37 putative SiCCT genes and classiﬁed the genes into three subfamilies. We conducted a comprehensive bioinformatic analysis of the gene location, chromosomal distribution, exon–intron structure, motif composition, regulatory sequences, phylogenetic relationships, and synteny. The expression of the CCT genes was analyzed in different tissues by semiquantitative RT-PCR; diurnal rhythm oscillations were investigated using transcriptome data, and expression levels under different treatments were assessed by quantitative real-time PCR (qRT-PCR) analysis. The results provide insights into the function and evolution of the SiCCT gene family and its potential roles in foxtail millet growth and stress responses and provide a foundation for molecular breeding to improve the stress tolerance of foxtail millet.


Introduction
Many environment factors, including abiotic stresses and photoperiodic rhythm, affect growth and development, and thus yield [1][2][3]. The molecular mechanisms by which plants adapt to environmental changes are very complex, and a number of gene families are involved in the regulatory pathways. The CCT gene family has been shown to play important roles in plant response to environmental changes, including circadian rhythm and abiotic stress [4][5][6]. Structural analysis and functional study of CCT family genes in plants will help us to explore the molecular mechanism of environment stress response, which is significant for improving the environmental adaption of plants.
The CCT family genes contain a conserved CCT domain and are primarily known in flowering plants [7]. According to the distribution of conserved domains, CCT family genes are classified into three subfamilies, comprising the CCT motif (CMF) subfamily that contains a single CCT motif [8], the CONSTANS-like (COL) subfamily that contains a CCT domain and one or two zinc-finger B-Box domains [9], and the PSEUDORESPONSE REGU-LATOR subfamily (PRR) that possesses a pseudo-receiver domain and a CCT motif [5]. The study of cormophyte and streptophyte lineages indicates that the CMF subfamily evolved from the COL subfamily through the loss of the B-Box domains [8]. These changes in the domains further enhanced the functional diversification of the CCT family.

Gene Identification, Chromosomal Location, and Phylogenetic Relationships of CCT Family Members in Foxtail Millet
The genome sequence of millet was obtained from the Ensembl Plant database (http://plants.ensembl.org/index.html, accessed on 30 September 2021). We downloaded the hidden Markov model of the CCT domain from the Pfam database (http://pfam. sanger.ac.uk/, accessed on 30 September 2021). The SiCCT family members were extracted from the genome database with HMMER 3.1 (http://hmmer.janelia.org, accessed on 30 September 2021). The candidate genes were screened against the National Center for Biotechnology databases (https://www.ncbi.nlm.nih.gov/cdd/, accessed on 12 October 2021) to determine the 37 final genes. Information on the CCT proteins were calculated using the ExPaSy ProtParam tool (http://web.expasy.org/protparam/, accessed on 12 October 2021), including length, molecular weight, and isoelectric point. We used the subcellular localization prediction tool Plant-mPLoc (http://www.csbio.sjtu.edu.cn/ bioinf/plant-multi/, accessed on 12 October 2021) [29] to predict the likely location of the protein. The MapChart software [30] was used to map the chromosomal locations of the genes.
The protein sequences of the CCT domain-containing genes of Arabidopsis and rice were downloaded from the Ensembl Plant database on the basis of relevant research reports [5,8,9,31,32]. The full-length CCT sequences of Arabidopsis, rice, and foxtail millet were selected for comparison with MEGA X version 7 software [33]. Subsequently, a multiple sequence alignment was used to construct a maximum likelihood phylogenetic tree with the MEGA software with the following parameters: Poisson model, pairwise deletion, and 1000 bootstrap replications.

Collinearity Analysis and Gene Duplication
To explore the synteny of orthologous CCT genes obtained from foxtail millet and other selected species, we used the MCScan (Python version) tool (https://github.com/ tanghaibao/jcvi/wiki/MCscan-(Python-version), accessed on 25 October 2021) to draw the synteny map. When visualizing the results, the gene filtering parameter in the small collinearity block was set to 30. Circos [34] was used to map all SiCCT genes to foxtail millet chromosomes based on physical location information from the millet genome database. A multiple collinearity scanning kit (MCScanX) [35] was used to analyze gene duplication events. KaKs_Calculator 2.0 [36] was used to calculate K a and K s to detect replication events.

Gene Structure and Conserved Sequence Analysis
The location of introns, exons, and untranslated regions in the genes was extracted from the gene finding format (GFF3) file. Conserved domains of millet CCT proteins were elucidated with the MEME Suite (http://meme-suite.org/, accessed on 1 November 2021) [37]. The optimized parameters were employed as follows: the maximum number of motifs was 10, and the optimum width was from 6 to 50. The map of the gene structure and motifs was drawn using TBtools [38].

Cis-Acting Element Analysis
The genomic DNA sequence 1000 bp upstream of the translation initiation codon for the SiCCT genes was used for cis-acting element analysis.

Plant Materials and Treatments
Millet 'Yugu1' was used in this study. Plants were grown in fields and in a greenhouse in Shenyang, Liaoning Province, China. The stress-treated plants were grown in a greenhouse in Hoagland's nutrient solution [40]. Abscisic acid, salt stress, and drought stress were applied separately by treating plants with 100 µM ABA, 200 mM NaCl solution, or 20% PEG6000, respectively. Plant leaves were collected after treatment for 0, 4, 8, and 12 h, with three biological replicates for each treatment. For low-temperature treatment, plants were placed in a 4 • C incubator for 12 h, and samples were collected at 0, 6, and 12 h. The leaf, immature seed, spikelet, root, leaf sheath, shoot apical meristem, stem, im-mature leaf, and seed were collected separately at three developmental stages from field-grown foxtail millet plants for semiquantitative RT-PCR.

Analysis of CCT Gene Expression in Millet and Quantitative Real-Time PCR
Total RNA was extracted from each tissue using a plant RNA extraction kit (Accurate Biotechnology Co., Ltd., Changsha, China) in accordance with the manufacturer's instructions. The cDNA was synthesized from 2 µg total RNA with M-MLV reverse transcriptase (Accurate Biotechnology Co., Ltd.). Semiquantitative PCR was performed using GoTaq ® Green Master Mix (Promega, Shanghai, China) with specific primers (synthesized by Genewiz, Suzhou, China). The GoTaq ® qPCR Master Mix kit (Promega) was used for qRT-PCR reactions using a C1000 Thermal Cycler and quantified using a CFX96 Real-Time System (Bio-Rad, Hercules, CA, USA). The relative transcript level was calculated using the 2 −∆∆Ct method [41]. Primers used in this experiment are enlisted in Table S1. Actin gene (SETIT_010361mg) was used as internal control for this experiment. Transcriptome data used in this study were downloaded from the NCBI SRA database under accession number PRJNA684771 [42]. We processed per kilobase per million (FPKM) data for diurnal rhythm analysis. Yi et al. [42] sampled the two groups of samples at the beginning of light and the beginning of darkness, respectively, and there were two groups of data at 16 h, 20 h, and 24 h. Therefore, we draw the line graph using the average of the FPKM values of these two sets of data.

Identification and Analysis of CCT Gene Family Members in Foxtail Millet
A total of 37 predicted CCT genes were extracted from the foxtail millet genome using a hidden Markov model (PF06203). The genes were designated as genes from SiCCT1 to SiCCT37 based on their chromosomal distribution ( Table 1). The gene identifier, genomic position, predicted molecular weight, protein length, and isoelectric point of each gene are listed in Table 1. The length of the proteins ranged from 160 aa (SiCCT27) to 760 aa (SiCCT35); the molecular weight varied from 17.6 kDa (SiCCT27) to 82.7 kDa (SiCCT35), and the isoelectric point ranged from 4.37 (SiCCT32) to 10.18 (SiCCT27). All of the genes were predicted to be localized in the nucleus except SiCCT4. SiCCT4 was predicted to be localized in both the mitochondria and the nucleus.
Thirty-six SiCCT genes were mapped to eight of the nine foxtail millet chromosomes. The number of SiCCT genes varied among the chromosomes, and no genes were detected on chromosome V ( Figure 1). Four chromosomes (IV, VI, VII, and VIII) contained three genes, and two chromosomes (II and III) carried four genes. Chromosome IX contained the largest number of nine genes. One gene was not mapped to a chromosome and was recategorized as unplaced ( Figure 1). SiCCT14 and SiCCT15 were located on chromosome 2 and identified as a pair of tandem repeats.

Evolutionary and Synteny Analysis of CCT Genes in Foxtail Millet and Other Species
To detect the evolutionary relationships among CCT family members, a maximum likelihood phylogenetic tree was constructed based on 99 full-length CCT protein sequences, comprising 32 from Arabidopsis, 30 from rice, and 37 from foxtail millet (Figure 2

Evolutionary and Synteny Analysis of CCT Genes in Foxtail Millet and Other Species
To detect the evolutionary relationships among CCT family members, a maximum likelihood phylogenetic tree was constructed based on 99 full-length CCT protein sequences, comprising 32 from Arabidopsis, 30 from rice, and 37 from foxtail millet ( Figure 2 and Table S2). The CCT proteins were resolved into three main clades consisting of eight groups, which were designated A to H. Group B represented the PRR subfamily with 15 members. The CMF subfamily comprised groups A (23 members), C (4 members), E (5 members), and G (9 members). The remaining groups (D with 16 members, F with 18 members, and H with 9 members) belonged to the COL subfamily. Figure 1. Chromosomal location of putative SiCCT genes. Thirty-six genes were mapped to the millet chromosomes, and one predicted gene was not mapped. The chromosome number is shown above each chromosome, and the length of each chromosome is indicated on the left in megabases (MB).

Evolutionary and Synteny Analysis of CCT Genes in Foxtail Millet and Other Species
To detect the evolutionary relationships among CCT family members, a maximum likelihood phylogenetic tree was constructed based on 99 full-length CCT protein sequences, comprising 32 from Arabidopsis, 30 from rice, and 37 from foxtail millet ( Figure 2 and Table S2). The CCT proteins were resolved into three main clades consisting of eight groups, which were designated A to H. Group B represented the PRR subfamily with 15 members. The CMF subfamily comprised groups A (23 members), C (4 members), E (5 members), and G (9 members). The remaining groups (D with 16 members, F with 18 members, and H with 9 members) belonged to the COL subfamily.  To enable further inference on the phylogenetic affinities of the foxtail millet CCT family, synteny analyses were performed between the CCT genes of foxtail millet and other species, comprising the dicotyledons Arabidopsis, tomato (Solanum lycopersicum), and alfalfa (Medicago truncatula) and the monocotyledons rice, sorghum (Sorghum bicolor), and maize ( Figure 3). Only one pair of homologous genes between foxtail millet and Arabidopsis, and two pairs between foxtail millet and each of alfalfa and tomato, were detected. A greater number of gene pairs were detected between foxtail millet and the other monocotyledonous species, namely rice (26 orthologous gene pairs), sorghum (28 orthologous gene pairs), and maize (31 orthologous gene pairs). Among these genes, several genes comprised more than three pairs in a collinear relationship between foxtail millet and the other species (Table S3), such as for SiCCT34, SiCCT35, and SiCCT36. SiCCT34 was detected in all studied species except Arabidopsis, suggesting that this gene may be evolutionarily conserved and may have played an important role in the evolution of CCT gene family. gene pairs), and maize (31 orthologous gene pairs). Among these genes, several genes comprised more than three pairs in a collinear relationship between foxtail millet and the other species (Table S3), such as for SiCCT34, SiCCT35, and SiCCT36. SiCCT34 was detected in all studied species except Arabidopsis, suggesting that this gene may be evolutionarily conserved and may have played an important role in the evolution of CCT gene family. Collinearity within the SiCCT family was analyzed to explore segmental duplication events within the SiCCT genes. Six pairs of segmental-duplicated genes were found on foxtail millet chromosomes ( Figure 4). We calculated the nonsynonymous substitution rate (Ka)/synonymous substitution rate (Ks) ratio for each gene pair to examine the evolutionary mechanism of the CCT genes ( Table 2). All Ka/Ks values were less than one, except for SiCCT32/SiCCT36. To further explore the evolution of CCT genes in other Poaceae species, we also selected the CCT genes of maize and sorghum to plot the respective intraspecific collinearity relationships ( Figures S1 and S2). Twenty-three and 10 pairs of segmental-duplicated genes were found in the maize and sorghum genomes, respectively. Collinearity within the SiCCT family was analyzed to explore segmental duplication events within the SiCCT genes. Six pairs of segmental-duplicated genes were found on foxtail millet chromosomes ( Figure 4). We calculated the nonsynonymous substitution rate (K a )/synonymous substitution rate (K s ) ratio for each gene pair to examine the evolutionary mechanism of the CCT genes ( Table 2). All K a /K s values were less than one, except for SiCCT32/SiCCT36. To further explore the evolution of CCT genes in other Poaceae species, we also selected the CCT genes of maize and sorghum to plot the respective intraspecific collinearity relationships ( Figures S1 and S2). Twenty-three and 10 pairs of segmentalduplicated genes were found in the maize and sorghum genomes, respectively.

Gene Structure and Conserved Motif Analysis
The phylogenetic tree of SiCCTs was constructed and divided into 10 groups (I to X; Figure 5A). To better understand the evolution of CCT genes in foxtail millet, we analyzed the intron-exon structure of the identified CCT genes ( Figure 5B). Most SiCCT genes contained one to three introns (eight genes with one intron, seven with two introns, and seven with three introns). A smaller proportion of the genes contained four to eight introns (two genes with four introns, four with five introns, three with six introns, four with seven introns, and one with eight introns). Only SiCCT20 lacked an intron.
The phylogenetic tree of SiCCTs was constructed and divided into 10 groups (I to X; Figure 5A). To better understand the evolution of CCT genes in foxtail millet, we analyzed the intron-exon structure of the identified CCT genes ( Figure 5B). Most SiCCT genes contained one to three introns (eight genes with one intron, seven with two introns, and seven with three introns). A smaller proportion of the genes contained four to eight introns (two genes with four introns, four with five introns, three with six introns, four with seven introns, and one with eight introns). Only SiCCT20 lacked an intron. We constructed domain diagrams of the 37 predicted SiCCT genes to study their conserved motifs using the MEME Suite of sequence analysis tools ( Figure 5C). The CCT domain was composed of motifs 1 and 3, and all SiCCT genes had complete CCT institutional domains, except for SiCCT14, SiCCT15, and SiCCT24. SiCCT14 only had motif 3, and SiCCT24 and SiCCT15 only had motif 1. Motif 2 was the B-Box domain, which was specific to the COL family. The PRR domain was composed of motif 5 and motif 10. In the PRR subfamily, SiCCT9, SiCCT4, SiCCT35, and SiCCT11 had intact PRR domains, whereas SiCCT26 only had motif 10. Furthermore, we found that SiCCT29, SiCCT31, SiCCT27, and SiCCT18 had independent evolutionary branches in the phylogenetic tree and their CCT domains (motif 1 and 3) were closer to the 5 end than other CCT genes. These results indicated that the domains of these proteins were relatively conserved and conformed with the subfamilial classification of the CCT family.

Analysis of Cis-Acting Elements in the SiCCT Gene Promoter Region
To predict the function and regulatory mechanism of the SiCCT genes, we analyzed the cis-acting elements in the promoter region of the genes. Based on the results of a PlantCARE analysis, the cis-acting elements were divided into three categories: plant growth and development, abiotic and biotic stresses, and phytohormone responsive (Table S4). We selected 27 cis-acting elements from the promoter of the SiCCT genes and divided them into the three categories ( Figure 6 and S3). In the growth and development group, the G-Box element involved in light response was most numerous and was detected in 24 of the 37 genes. In addition to light-response elements, the CAT-Box element involved in the regulation of meristem expression was present in 14 genes. An O 2 -site element involved in zein metabolism regulation was detected in nine genes. In the abiotic and biotic stresses group, multiple response elements were observed, such as oxidation, (ARE and GC-motif), wounding (box-S and WUN-motif), drought (MBS), high temperature (STRE), low temperature (LRE), and defense (CCAAT-box). The most numerous elements in the stress-response group were MYB and MYC binding sites, which are general stressresponsive elements. In the phytohormone-responsive group, we detected many elements involved in hormone response, including elements responsive to ABA (ABRE), salicylic acid (as-1), indoleacetic acid (AuxRR-core and TGA-element), gibberellins (GARE-motif and P-box), and methyl jasmonate (TGACG-motif). Among them, the number of ABRE elements was 92, which was the largest number of elements related to hormone regulation. In the cis-element analysis, a large number of elements related to light response and diurnal rhythm were detected in the promoter region of CCT genes. To further explore the CCT genes response to diurnal rhythm, an expression analysis of CCT genes under diurnal rhythm was performed. The transcriptome data (PRJNA684771) used for diurnal rhythm analysis were downloaded from NCBI database according to the research of Yi et al. [42]. The expression pattern of 16/37 CCT genes shoedn diurnal rhythm oscillations, which includes all PRR members, 6 COL members, and 5 CMF members (Figure 7 and Table S5). The expression of SiCCT4 and SiCCT9 reached the highest level at 12 h, and the expression of SiCCT11 and SiCCT35 reached the highest at 8 h, while SiCCT26 showed different rhythm oscillation than other PRR members. In addition to the PRR family genes, the COL subfamily members (SiCCT2, 5, 7, 8, 16, and 17) and CMF subfamily members (SiCCT19, 28, 31, 32, and 33) also showed diurnal rhythms, and the expression peak of these genes appeared in different times. Among them, the expression peak of SiCCT5, 8,16,19, and 31 reached the highest level at dark.

Expression Analysis of SiCCT Genes in Different Tissues and under Abiotic Stress and Exogenous Hormone Treatments
We used semiquantitative RT-PCR to investigate the expression of the SiCCT genes in various tissues, comprising the leaf, immature seed, spikelet, root, leaf sheath, shoot apical meristem, stem, immature leaf, and seed ( Figure S4). Four genes (SiCCT17, SiCCT22, SiCCT26, and SiCCT31) were highly expressed; three genes (SiCCT18, SiCCT24, and SiCCT37) were weakly expressed, and two genes (SiCCT14 and SiCCT16) were not expressed in all tissues. The RT-PCR results showed that the SiCCT genes were expressed in various tissues and differed in expression patterns.
In the cis-element analysis, a large number of elements related to light response and diurnal rhythm were detected in the promoter region of CCT genes. To further explore the CCT genes response to diurnal rhythm, an expression analysis of CCT genes under diurnal rhythm was performed. The transcriptome data (PRJNA684771) used for diurnal rhythm analysis were downloaded from NCBI database according to the research of Yi et al. [42]. The expression pattern of 16/37 CCT genes shoedn diurnal rhythm oscillations, which includes all PRR members, 6 COL members, and 5 CMF members (Figure 7 and Table S5). The expression of SiCCT4 and SiCCT9 reached the highest level at 12 h, and the expression of SiCCT11 and SiCCT35 reached the highest at 8 h, while SiCCT26 showed different rhythm oscillation than other PRR members. In addition to the PRR family genes, the COL subfamily members (SiCCT2, 5, 7, 8, 16, and 17) and CMF subfamily members (SiCCT19, 28, 31, 32, and 33) also showed diurnal rhythms, and the expression peak of these genes appeared in different times. Among them, the expression peak of SiCCT5, 8,16,19, and 31 reached the highest level at dark. Many genes in the CMF subfamily have been reported to be involved in abiotic stress and hormone regulation in plants [22,24]. To clarify whether foxtail millet CMF subfamily genes responded to abiotic stresses and ABA treatment, we analyzed their expression patterns under different treatments ( Figure 8). Overall, the transcript abundance of many genes was significantly increased, especially under ABA and drought treatment. After ABA treatment, all genes were up-regulated except SiCCT3 and SiCCT12. Among the upregulated genes, the expression of SiCCT13 was the highest, significantly up-regulated by more than 29 times, and its expression peaked at 8h. Under drought induction, the expression levels of all genes were induced except SiCCT22, SiCCT29, and SiCCT32. The expression level of SiCCT3 was significantly up-regulated by 40-fold at 12 h. The expression levels of half of the genes were induced by salt treatment, and the expression of most genes was induced by more than three times. Eight genes responded to low-temperature treatment, of which the expression levels of five genes were induced and three genes were inhibited. SiCCT31 was most significantly induced by low temperature, and the expression level of SiCCT31 was up-regulated by more than eight times at 8 h.
In general, most of the CMF subfamily genes showed a trend for up-regulated expression in response to abiotic stress and hormone treatment ( Figure S5A). After ABA treatment, 10 genes were up-regulated, followed by drought treatment. The expression of SiCCT37 was up-regulated under the four treatments. The expressions of SiCCT21, SiCCT28, and SiCCT18 was induced by exogenous ABA, salt, and drought. The number of genes down-regulated by salt treatment were the most among all treatments ( Figure  S5B). The expressions of SiCCT22 and SiCCT29 was inhibited by salt and drought. Many genes in the CMF subfamily have been reported to be involved in abiotic stress and hormone regulation in plants [22,24]. To clarify whether foxtail millet CMF subfamily genes responded to abiotic stresses and ABA treatment, we analyzed their expression patterns under different treatments ( Figure 8). Overall, the transcript abundance of many genes was significantly increased, especially under ABA and drought treatment. After ABA treatment, all genes were up-regulated except SiCCT3 and SiCCT12. Among the up-regulated genes, the expression of SiCCT13 was the highest, significantly up-regulated by more than 29 times, and its expression peaked at 8 h. Under drought induction, the expression levels of all genes were induced except SiCCT22, SiCCT29, and SiCCT32. The expression level of SiCCT3 was significantly up-regulated by 40-fold at 12 h. The expression levels of half of the genes were induced by salt treatment, and the expression of most genes was induced by more than three times. Eight genes responded to low-temperature treatment, of which the expression levels of five genes were induced and three genes were inhibited. SiCCT31 was most significantly induced by low temperature, and the expression level of SiCCT31 was up-regulated by more than eight times at 8 h.
In general, most of the CMF subfamily genes showed a trend for up-regulated expression in response to abiotic stress and hormone treatment ( Figure S5A). After ABA treatment, 10 genes were up-regulated, followed by drought treatment. The expression of SiCCT37 was up-regulated under the four treatments. The expressions of SiCCT21, SiCCT28, and SiCCT18 was induced by exogenous ABA, salt, and drought. The number of genes down-regulated by salt treatment were the most among all treatments ( Figure S5B). The expressions of SiCCT22 and SiCCT29 was inhibited by salt and drought.

Discussion
Stress tolerance and photoperiod regulation are important factors for plant adaption to growth in different regions. The CCT gene family has been reported to be involved in the regulation of photoperiod and stress responses in plants [13,43,44]. CCT genes have been identified in many crop species at the whole-genome level. For example, 41 and 53 CCT family genes have been identified in rice [45] and maize [44], respectively. Foxtail millet exhibits the characteristics of strong tolerance to abiotic stress and sensitivity to photoperiod. Therefore, we identified possible CCT genes from the foxtail millet genome. In this study, 37 putative CCT family genes were identified in foxtail millet (Table 1). A phylogenetic analysis of the CCT family genes of foxtail millet, Arabidopsis, and rice resolved the genes into eight groups within three clades (Figure 2). Several pairs of SiCCT genes homologous to the reported OsCCT and AtCCT genes were found in the phylogenetic tree. For example, SiCCT6 was homologous to HD1 and CO, SiCCT5 and SiCCT8 were homologous to DTH2, SiCCT35 was homologs to HD2, and SiCCT4 and SiCCT26 were homologous to APRR1. The high homologies among the genes may imply that the proteins perform similar functions.
Whole-genome duplication (WGD) and nested chromosome fusions (NCFs) are important driving forces in the evolution of flowering plants [46,47]. WGD can provide a plant with the opportunity for diversification in gene functions, and NCFs can change the number of chromosomes. These processes are important for speciation and the evolution of novel gene functions. Flowering plants have experienced many WGD events [48], of which three WGD events have occurred in the genomes of cereal grasses [49]. To explore the evolution of SiCCT genes in cereals, we constructed a collinearity map for maize and sorghum ( Figures S1 and S2) and calculated the Ka and Ks values for foxtail millet, sorghum, and maize (Table 2). We then estimated the timing of the doubling event of the SiCCT genes. The doubling time points for sorghum and maize were also calculated as

Discussion
Stress tolerance and photoperiod regulation are important factors for plant adaption to growth in different regions. The CCT gene family has been reported to be involved in the regulation of photoperiod and stress responses in plants [13,43,44]. CCT genes have been identified in many crop species at the whole-genome level. For example, 41 and 53 CCT family genes have been identified in rice [45] and maize [44], respectively. Foxtail millet exhibits the characteristics of strong tolerance to abiotic stress and sensitivity to photoperiod. Therefore, we identified possible CCT genes from the foxtail millet genome. In this study, 37 putative CCT family genes were identified in foxtail millet (Table 1). A phylogenetic analysis of the CCT family genes of foxtail millet, Arabidopsis, and rice resolved the genes into eight groups within three clades (Figure 2). Several pairs of SiCCT genes homologous to the reported OsCCT and AtCCT genes were found in the phylogenetic tree. For example, SiCCT6 was homologous to HD1 and CO, SiCCT5 and SiCCT8 were homologous to DTH2, SiCCT35 was homologs to HD2, and SiCCT4 and SiCCT26 were homologous to APRR1. The high homologies among the genes may imply that the proteins perform similar functions.
Whole-genome duplication (WGD) and nested chromosome fusions (NCFs) are important driving forces in the evolution of flowering plants [46,47]. WGD can provide a plant with the opportunity for diversification in gene functions, and NCFs can change the number of chromosomes. These processes are important for speciation and the evolution of novel gene functions. Flowering plants have experienced many WGD events [48], of which three WGD events have occurred in the genomes of cereal grasses [49]. To explore the evolution of SiCCT genes in cereals, we constructed a collinearity map for maize and sorghum ( Figures S1 and S2) and calculated the K a and K s values for foxtail millet, sorghum, and maize (Table 2). We then estimated the timing of the doubling event of the SiCCT genes. The doubling time points for sorghum and maize were also calculated as controls. Fox-tail millet separated from maize and sorghum approximately 27 million years ago (Mya), and maize and sorghum diverged approximately 13 Mya [25]. Based on the estimated dates for these two nodes and the respective estimated divergence dates for the Poaceae (98.2 Mya) [50] and the subfamily Panicoideae (48 Mya) [25], we divided the calculated WGD events into time periods comprising an intermediate ancestor period (~98.2 Mya), Poaceae ancestor period (~98.2 Mya to~48 Mya), Panicoideae ancestor period (foxtail millet:~48 Mya to~27 Mya; maize and sorghum:~48 Mya to~13 Mya), and post-speciation period (foxtail millet:~27 Mya; maize and sorghum: 13 Mya). Based on the calculations, the CCT genes of foxtail millet and sorghum diverged in relatively ancient times, with the diversification of most of the genes concentrated between the intermediate ancestor period and the Poaceae ancestor period. These findings suggest that SiCCT genes have hardly changed after speciation. From an evolutionary point of view, SiCCT genes have been essentially stable since entering the Panicoideae ancestor period. These results indicate that the functions of SiCCT genes were determined before the divergence of foxtail millet, which further implies that the functions of the genes are conserved.
Tandem repeats and fragment repeats gave rise to 14 genes, which were important for CCT family amplification in foxtail millet (Figures 1 and 4). We speculate that the generation of these new genes might be caused by changes in the domains. The generation of the CMF subfamily resulted from the loss of the B-Box domain of the COL subfamily members [5]. Therefore, we investigated the domains of SiCCT proteins and observed that most proteins had complete CCT domains except SiCCT14, SiCCT15, and SiCCT24 ( Figure 5). The CCT domain was composed of motifs 1 and 3. SiCCT15 and SiCCT14 had no other domain differences except in the CCT domain (SiCCT15 only had motif 1 and SiCCT14 only had motif 3). Both SiCCT14 and SiCCT15 were tandem repeats. Therefore, we consider that SiCCT14 and SiCCT15 were generated during the process of tandem duplication, owing to the separation of motif 1 and motif 3 in the CCT domain. The differences of SiCCT14 and SiCCT15 were not detected in SiCCT24. However, SiCCT24 formed a collinear pair with SiCCT25. The motif 3 in SiCCT24 may have been lost when segment duplication caused by the WGD occurred. In addition, the protein structure of six pairs of collinear genes were compared. Three pairs were observed to be structurally different (SiCCT24/SiCCT25, SiCCT32/SiCCT36, and SiCCT2/SiCCT17). In addition to the loss of motif 3 in SiCCT25 to form SiCCT24, SiCCT36 lost three domains (motifs 4, 7, and 9), and SiCCT17 lost the B-Box (motif 2) to form SiCCT32 and SiCCT2, respectively, during the WGD event. Previous studies have shown that the deletion of the B-Box domain affects grain vernalization and response to LD [8]. Therefore, after domain deletion, these proteins are likely to have assumed novel functions, which are worthy of further exploration. Furthermore, we found that SiCCT29, SiCCT31, SiCCT27, and SiCCT18 had independent evolutionary branches in the phylogenetic tree and their CCT domains were closer to the C terminus of protein than other CCT genes. This suggested that they may had different functions with other SiCCT genes.
It had been found that CCT genes responded to photoperiod signal and are involved in diurnal rhythm regulation, therefore affecting the flowering time of plants [44,51]. A number of elements related to light response and diurnal rhythm regulation were found on the promoter of SiCCT, including G-BOX, I-box, Sp1, and so on. Therefore, we further analyzed the expression rhythmicity of the CCT gene according to the transcriptome data. PRR subfamily genes are key factors in the regulation of plant diurnal rhythm. Figure 7 showed that the expression of SiCCT4, SiCCT9, SiCCT11, and SiCCT35 had strong diurnal rhythm. Among them, the expression of SiCCT11 and SiCCT35 reached the highest value at 8 h of light, and they had similar expression patterns in diurnal rhythm. In addition, the phylogenetic tree analysis showed that both SiCCT35 and SiCCT11 were closely related to Hd2. Hd2 is a gene related to rice heading date regulation, which can inhibit heading date under long-day condition [52]. Therefore, we speculated that SiCCT35 and SiCCT11 might play an important role in regulating the heading time of foxtail millet. In addition to the genes in the PRR subfamily, some genes in the COL and CMF subfamilies show strong diurnal rhythms. These results suggested that many genes in the SiCCT family contributed an important function in the regulation of foxtail millet diurnal rhythm.
In addition to photoperiod regulation, the function of some CCT genes is also associated with abiotic stresses. Among such genes, many members of the CMF subfamily have been reported, such as ZmCCT and GHD7 [22,24]. In the present study, many of the CMF subfamily genes showed sensitivity to ABA and stress treatment (Figure 8). In response to ABA treatment, all genes except SiCCT3 and SiCCT12 were up-regulated. Analysis of cis-acting elements ( Figure 6) revealed that no ABA-responsive element was detected in SiCCT3 and SiCTT12, which might be the reason why their expression was not induced by exogenous ABA. Most CMF family members were induced by drought stress except SiCCT32. They also had different degrees of response to low temperature and salt stress. Interestingly, SiCCT37 expression was up-regulated in response to all treatments. Thus, this gene may have multiple functions, and the further study of the expression pattern and functions of this gene may be of importance for elucidating abiotic stress responses.
The foxtail millet CCT gene family was analyzed in this study. Thirty-seven full-length SiCCT genes were characterized and classified into three subfamilies according to their domains. The collinearity analysis and phylogenetic relationships of SiCCT genes provide valuable clues for dissecting the evolutionary characteristics of the SiCCT genes. The expression analysis of SiCCT genes responding to diurnal rhythm and abiotic stresses provide insights into their potential functions in the growth and development of foxtail millet. The present research will be helpful for further functional research into SiCCT genes and, potentially, for the improvement of the environmental adaptability and stress tolerance of foxtail millet.
Interchromosomal relationships of maize CCT genes; Figure S3. Distribution map of cis-acting elements on each gene; Figure S4. Expression profiles of SiCCT genes in different tissues of millet. The site of expression is shown at the top, and the gene name is shown on the right; Figure S5. The detailed numbers of simultaneously up-regulated (A) and down-regulated (B) SiCCT genes by ABA (abscisic acid), NaCl, drought, and low temperature; Table S1: Primer name and sequence information of SiCCT; Table S2: List of CCT genes in Arabidopsis, rice, and foxtail millet; Table S3: List of homology genes of the CCT genes of foxtail millet with 6 representative plant species; Table S4: The number and function statistics of promoter elements predicted in this study; Table S5:   Institutional Review Board Statement: Not applicable.

Informed Consent Statement: Not applicable.
Data Availability Statement: The data and materials obtained in this study are available from the corresponding author on reasonable request.