The conserved domain of COL proteins was downloaded from the Pfam database (http://pfam.xfam.org/family/PF06203 and PF00643) and was used as the template region to search StCOL sequences in the potato genome database (http://spuddb.uga.edu/). Hidden Markov model (HMM) software was used to select candidate StCOL genes, which were then submitted to NCBI (https://www.ncbi.nlm.nih.gov/) and UniProt (http://www.uniprot.org) for confirming the final StCOL gene family members.

The online website cluster (https://pir.georgetown.edu/pirwww/search/multialn.shtml) was used to perform multiple alignments of the predicted protein sequences of the 15 StCOLs in potato, 17 AtCOLs in Arabidopsis, and 13 SlCOLs in tomato (Zobell et al., 2005). The unrooted phylogenetic tree was constructed using the MEGA 7 software with 1000 bootstrap replicates (Hall, 2013). The StCOLs were generated according to the phylogenetic relationships of 17 AtCOL genes with the StCOL gene and a specific gene nomenclature system.

The MEME (http://meme.sdsc.edu/meme/cgi-bin/meme.cgi) online software was used to assess the conservative area of the StCOL gene family protein sequence. The maximum ordinal number was set to 10, and the optimum motif width was 6–50 amino acid residues (Bailey et al., 2009). The gene structures of StCOLs were determined to obtain gene exons and containing substructures by Gene Structure Display Server (http://gsds.cbi.pku.edu.cn/). In addition, the cis-acting elements of StCOL promoter regions in potato were predicted by PlantCARE (PlantCARE/HTML/http://bioinformatics.psb.ugent.be/webtools/).

Collinearity analysis was conducted through PlantDGD software (http://pdgd.njau.edu.cn:8080/) (Qiao et al., 2019). The Multiple Collinearity Scan toolkit (MCScanX) was used to identify StCOL gene sequence repetition events. BLASTP was used to identify the species, symbiotic homologous pairs of the protein sequence. The protein sequence parameter was set to 1) alignment significance: e-value (default: 1 × 10⁻⁵), 2) MATCH_SCORE: final score (default: 50) (Wang et al., 2012). The obtained data were plotted by Circos online software (http://circos.ca/). Then, MapInspect software was used to locate COL genes whose relative positions have been identified in the potato genome database on potato chromosomes.

ExPASy online software (http://web.expasy.org/protparam/) was used to predict and analyze the amino acid number, theoretical molecular number, isoelectric point, and other physicochemical properties of the potato COL protein.

The publicly available RNA-seq data in the potato genome (http://spuddb.uga.edu/) were downloaded. The RNA-seq data of the root, tuber pith, tuber cortex, young tuber, mature tuber, shoot apex, tuber sprout, tuber peel, stamen, flower, petiole, leaf, stem, and stolon were selected to analyze the expression of StCOL genes. The “Normalized” function was used to normalize gene expression and construct the heat maps of StCOL gene expression using TBtools (Chen et al., 2020). The expression data are presented in Supplementary Table S1.

The leaves and tuber of the early maturing potato variety “Favorita” at three different growth stages (20, 42, and 65 days) were collected. The low-temperature sensitive variety “Desiree” was subjected to 2°C for 0 h, 1 h, 2 h, 24 h, and 48 h, and the leaves of all treatments were collected. The collected samples were stored in a −80°C refrigerator. The total RNA of all collected samples was isolated by a TaKaRa MiNiBEST Universal RNA Extraction Kit (TaKaRa, Beijing, China). Reverse transcription was performed with StarScript II RT Mix with gDNA Remover (GeneStar, Beijing, China). The PCR amplification was conducted on a CFX96 touch real-time PCR detection system (Bio-Rad, Hercules, CA, United States). β-actin was used as an internal reference gene. The Ct (2^−ΔΔCT) method was used to calculate the relative expression of the StCOL genes (Livak and Schmittgen, 2001). The experiment was performed with three biological replicates. The primer sequences for qRT-PCR are listed in Supplementary Table S2. Duncan’s test with p < 0.05 was used to indicate significant differences.

A total of 15 potato COL genes were identified in the potato genome database. The genes were named StCOL1–StCOL15. The amino acid length of StCOL gene family members ranged from 347 to 453 aa, the molecular weight was 38.65–49.92 kD, and the isoelectric point was 5.13–6.09 (Supplementary Table S3). The isoelectric point of all 15 StCOL proteins was less than 7, the instability index was 6.94–59.96, and the aliphatic index ranged from 57.18 to 71.48. The grand average of hydropathicity ranged from −0.867 to −0.331, all of which were less than 0, indicating that StCOL proteins were hydrophilic.

In order to understand the evolutionary relationships among the StCOL gene family, Arabidopsis, and tomato, 15 StCOLs, 17 AtCOLs, and 13 SlCOLs were constructed by the phylogenetic tree (Figure 1). The StCOL protein family can be divided into three subfamilies: Group Ⅰ, Group Ⅱ, and Group Ⅲ. Group Ⅰ contained five StCOLs, six AtCOLs, and four SlCOLs. Group Ⅱ consists of three StCOLs, four AtCOLs, and three SlCOLs. Group III consists of seven StCOLs, seven AtCOLs, and five SlCOLs.

A total of 10 conserved motifs were obtained by predicting the StCOL protein conserved motifs through the MEME website (Figure 2A). All 15 StCOL proteins contain motif 1 (CCT domain) and motif 2 (zinc finger B-box domain). The five StCOLs in group I all contain motif 1, motif 2, and motif 3. There is a motif 9 in StCOL9 and StCOL11, but the location is slightly different. In group Ⅱ, StCOL12, StCOL14, and StCOL15 all contain motif 1, motif 2, and motif 7. In group ⅠⅡ, all seven StCOL proteins contained motif 1, motif 2, motif 3, and motif 6, of which four StCOL proteins contained all eight motifs. In order to further understand the structural and conserved characteristics of the StCOL gene family, exons and introns of the StCOL gene family were analyzed. The 15 StCOL genes all contain between two and four exons and introns. The number of introns and exons of genes in group III and group Ⅱ was the same. The number of exons and introns varies in the different groups, which indicates that the functions of the different genes may be specific.

Seventeen cis-acting elements related to stress response, growth and development, and hormone regulation were predicted in the potato COL gene family (Figure 3). Among them were five cis-acting elements related to growth and development, six cis-acting elements related to stress response, and six cis-acting elements related to hormone regulation response. Twelve genes contained response elements necessary for anerobic induction, 10 genes contained abscisic acid response elements, 11 genes contained gibberellin response elements, four genes contained cis-acting elements related to low-temperature stress response, and 15 StCOL contained light response elements. Most of the growth-related response elements are photoregulatory response elements. These results indicate that the COL gene may play an important role in anerobic induction, abscisic acid response, gibberellin response, low-temperature stress response, and light response conditions.

In order to understand the physicochemical properties of COL proteins in potato, the post-translational modifications were performed at ScanProsite (http://ca.expasy.org/tools/scanprosite/). It mainly includes six kinds of post-translational modifications in the 15 StCOL proteins, including phosphorylation sites protein kinase C (PKC), casein kinase II (CK II), CAMP-CGMP-kinase (cAMP-cGMP), tyrosine kinase (Tyr), N-myristoylation (N-Myr), and N-glycosylation (N-Glyc) (Supplementary Table S4). Each StCOL protein contains at least three and up to six post-translational modifications. The phosphorylation sites protein kinase C, casein kinase II, and N-myristoylation existed in 15 StCOL proteins. A total of eight StCOL proteins had cAMP-cGMP sites, five StCOL proteins with Tyr sites, and 13 StCOL proteins with N-Glyc sites were observed. The results suggested that StCOL proteins may exert their function through phosphorylation sites protein kinase C, casein kinase II, and N-myristoylation.

Fifteen COL genes were found to be unequally distributed on eight potato chromosomes through chromosome localization (Figure 4A). Chr5 has the largest distribution with four COL genes. Chr3, Chr4, Chr8, and Chr9 contain only one COL gene. StCOLs were mainly distributed at the top or end of chromosomes. MCScanX was used to analyze the evolution and function of 15 StCOLs in the potato chromosome genome. The results showed that four pairs of the StCOL gene family had gene duplication events (Figure 5). There were no tandem repeats and large fragment replication, including StCOL14 on Chr03, StCOL12 on Chr12, StCOL6 and StCOL1 on Chr07, and StCOL8 gene clusters on Chr12. There were gene duplication events in StCOL5 and StCOL8 on Chr12. Interspecies collinearity analysis of COL genes of tomato, potato, and Arabidopsis identified 17 homologous genes between tomato and potato and 21 homologous genes between potato and Arabidopsis, indicating that there were more direct homologous genes, close relatives, and similar gene functions among these species (Figure 4B). A total of 13 StCOL genes are homologous in tomato and Arabidopsis (Supplementary Table S5), indicating that these StCOL genes may have a common ancestor in different species that evolved from the same ancestor.

The expression patterns of StCOL family genes in different potato tissues were analyzed using potato transcriptome data. The results showed that 10 genes were up-expressed to a certain extent in all tissues of potato, while the remaining five genes were down-expressed in potato (Figure 6). StCOL2 is highly expressed in root, tuber pith, stem, and other tissues, indicating that these genes play an important role in potato growth and development. High expression levels of StCOL1, StCOL2, StCOL6, StCOL7, and StCOL10 were found in flowers and may play a role in regulating the flowering process in potato. The StCOL2, StCOL4, StCOL7, StCOL10, StCOL11, and StCOL13 were more highly expressed in root, tuber pith, tuber cortex, and young tuber tissues than other genes. It may be that these genes play a major role in the formation of potato tubers.

The relative expression levels of StCOL genes were measured at different maturities of potato leaves and tubers. The qRT-PCR results showed that the StCOL2, StCOL5, StCOL6, StCOL8, and StCOL9 were highly expressed in the leaves of potato growing to day 42 (Figure 7). The expression levels of StCOL4 were significantly higher in potato tubers than in leaves. With increased growth time, the expression level of StCOL7 and StCOL15 decreased in leaves and tubers. The expression levels of StCOL2, StCOL6, and StCOL9 presented opposite patterns in leaves and tuber during the same period. These results indicated that the StCOL genes may play an important role in the formation of potato tubers.

The relative expression levels of StCOL genes were measured after different cold stress treatment times by qRT-PCR. The results showed that the StCOL1, StCOL6, StCOL12, and StCOL14 were highly expressed after treatment with 2°C for 1 h (Figure 8). Compared with 0 h, the expression levels of StCOL5, StCOL7, and StCOL9 were significantly down-expressed after different lengths of cold stress treatment times. These results indicated that the StCOL genes may play an important role in potato response to cold stress.