Genome-Wide Characterization of B-BOX Gene Family and Their Responses to Light Quality and Cold Stress in Tomato

Background: Perceiving incoming environmental information is critical for optimizing plant growth and development. Multiple B-box proteins (BBXs) play essential roles in light-dependent developmental processes in plants. However, whether BBXs function as a signal integrator between light and temperature in tomato plants remains elusive. Results: In this study, 31 SlBBX genes were identied from the newly released tomato (Solanum lycopersicum) genome sequences, and were clustered into ve subgroups with phylogenetic analysis. Gene structure and protein motif analyses showed relatively high conservation of closely clustered SlBBX genes within each subgroup; however, genome mapping analysis indicated the uneven distribution of the SlBBX genes on tomato chromosomes. Synteny analysis indicated that segmental duplication events happened in the expansion of the SlBBX genes in tomato. Promoter cis-regulatory elements prediction indicated that SlBBX genes were highly responsive to light, hormones, and stress conditions. Furthermore, the transcript analysis revealed that various SlBBX genes differed signicantly in expression after exposure to different light quality and low temperatures, while 61.3% of SlBBX genes were responsive to both light and low temperatures. Conclusions: Our study presented a genome-wide survey of SlBBX gene family in tomato, and emphasized their functions in perceiving light quality and low temperature, which may improve the current understanding of SlBBX gene functions in integrating light and temperature signals for plant adaptation to adverse environments.

Upon light irradiation, BBX21 directly binds to BBX22, HY5 and its own promoter regions and activates their transcription [11][12][13]. Moreover, both BBX21 and HY5 can associates with the BBX11 promoter to promote its transcription, while BBX11 binds to the HY5 promoter to activate its transcription [14]. Thus, these three TFs (BBX21, HY5 and BBX11), regulate the transcription, forming a positive feedback loop that precisely regulates plants photomorphogenesis. Moreover, BBX20, BBX21, BBX22 and BBX23 interact with HY5 to increase its transcriptional activity toward the target genes [15][16][17], whereas BBX24, BBX25, BBX28 and BBX29 suppress photomorphogenesis by reducing the activity of HY5 [17][18][19][20]. The HY5 positively controls BBX22 at the transcriptional level [21], whereas it represses the transcription of BBX30 and BBX31 by binding to the promoters of these two genes [22,23]. Meanwhile, BBX30 and BBX31 promote the expression of BBX28 and BBX29 by directly binding to the promoter regions of these genes [20]. In addition, direct interactions between BBX32 and BBX21 lead to inhibition in the BBX21-HY5 [24]. Interestingly, BBX4, accumulated after exposure to red light, directly interacts with phyB to promote photomorphogenesis in Arabidopsis [25]. These mechanisms are also found in other plant species. For example, OsBBX14 induces OsHY5L1 gene expression to stimulate photomorphogenesis in rice [26]. MdBBX37 associates with MdHY5 promoter to inhibit its expression in apple [27]. Additionally, MdBBX22 and MdBBX25/MdCOL4 bind to the MdHY5 promoter to increase and decrease the transcriptional activation of MdHY5, respectively [28]. Both PpBBX16 and PpBBX18 interact with PpHY5 to increase the biochemical activity of PpHY5, while PpHY5 binds to the promoter region of PpBBX18 to promote the transcription of PpBBX18 in pear [29,30]. Furthermore, the interaction of PpBBX21 with PpHY5 and PpBBX18 affects the bioactive heterodimers formation of PpHY5-PpBBX18 [30]. Therefore, speci c BBXs and HY5 constitute an important regulatory network to precisely control normal plant growth and development.
BBX proteins also act vital roles in regulatory networks that control plant adaption to abiotic stress.
Previous studies show that both BBX5 and BBX21 positively regulate plants tolerance to drought and salt stress in Arabidopsis [34,35]. BBX24/STO directly interacts with H-protein promoter binding factor 1 (HPPBF-1), which is a salt-responsive MYB TF, to enhance the root growth and salt tolerance in Arabidopsis [34]. CmBBX22 also positively regulates the plant drought tolerance [36]. In addition, MdBBX10 enhances tolerance to salt and drought by modulating ABA signaling and ROS accumulation [37]. In Arabidopsis, BBX18 and BBX23 control thermomorphogenesis [38]. Both MdBBX20 and MaCOL1 are responsive to low temperatures in apple and banana, respectively [39,40]. Furthermore, ZFPL, a homologous gene of AtBBX32, enhances cold tolerance in the grapevine [41]. CmBBX24 also increases plant cold tolerance in Chrysanthemum [42]. However, whether SlBBXs are involved in light and cold response in tomato remains to be explored.
In the present study, 31 SlBBX genes were identi ed and characterized in tomato. Gene distribution, synteny analyses, the architecture of exon-intron and motifs differences were investigated. Furthermore, the three-dimensional structure and evolution of SlBBX proteins were performed. Promoter analysis showed that the cis-elements in light signaling, hormones and stress response were the main elements of SlBBXs promoters. Meanwhile, we found that multiple SlBBX genes were either up-regulated or downregulated in response to different light quality and low temperatures, and 61.3% of SlBBX genes were responsive to both light and low temperatures. Therefore, our results suggest that SlBBXs are an essential component of light and temperature cues, which function to integrate environmental stimuli and plant hormones to coordinate plant growth under low temperatures.

Identi cation and characterization of SlBBX Genes in tomato
Based on the gene annotation as well as the conserved B-box motif characteristic of the BBX members, a total of 31 SlBBX genes were identi ed. The detailed information [gene name, gene identi ers, chromosome location, theoretical isoelectric point (pI), molecular weight (MW), genomic, coding sequences (CDS), peptide length, number of exon and intron, instability and aliphatic index, the grand average of hydropathicity (GRAVY) values and subcellular localization) of each SlBBX was presented in Table 1. The lengths of CDS and amino acids (AA) of 31 SlBBXs range from 267 bp and 88 aa (SlBBX18) to 1428 bp and 475 aa (SlBBX27), respectively. Thus, varied MW and pI were observed among SlBBX proteins. The MW of SlBBXs varies from 9.57 (SlBBX18) to 53.14 kDa (SlBBX27). The pI ranged from 4.25 (SlBBX5 and SlBBX7) to 9.28 (SlBBX26), with 74.2% SlBBXs with a pI lower than seven, which indicated that most of the SlBBX proteins were acidic in nature. The pI ranged from 4 to 9 in SlBBX proteins contained one (single) or two (double) B-box domains, while it decreased when plus a CCT domain, especially in the SlBBX proteins with a B-BOX domain plus a CCT domain (Additional le 1: Figure S1), which suggested that the CCT domain in SlBBX proteins may decrease their pI. Majority of SlBBXs were grouped into unstable proteins because their instability index was greater than 40, except for SlBBX6 in this family ( Table 1). The predicted aliphatic index ranged from 50.05 to 97.43 in SlBBX proteins. All SlBBX proteins, with the exception of SlBBX18, were predicted to be hydrophilic due to the GRAVY value (< 0). Subcellular localization predicted that most SlBBXs (23 of 31) were localized in the nuclear region, ve of them, including SlBBX5, SlBBX 6, SlBBX17, SlBBX25 and SlBBX31 in the chloroplast, while SlBBX16 and SlBBX18 in the cytoplasm, SlBBX19 in the peroxisome (Table 1). In addition, none of the 31 BBX proteins have a transmembrane domain, which indicated that these SlBBX proteins were not located on the cell membrane (Additional le 1: Figure S2).
Protein sequences, phylogenetic analysis and threedimensional structure of SlBBXs The domains logos and the sequences of the B-box1, B-box2 and CCT domain of the SlBBX proteins are shown in Fig. 1. Eight members out of the 31 SlBBXs, were characterized by the occurrence of two B-box domains and also a conserved CCT domain, whereas four members of them had a valine-proline (VP) motif ( Table 2). Only two B-BOX domains were found in ten SlBBXs, whereas ve members had one B-box domain and also a CCT domain, and eight members had only one B-box domain ( Table 2). Among the three domains, we found that each tomato B-box motif contained approximately 40 residues with the consensus sequence C-X2-C-X8-C-X2-D-X4-C-X2-C-D-X3-H-X8-H (Fig. 1). The conserved C, C, D and H residues ligated two zinc ions [2]. Additionally, the consensus sequence of the conserved CCT domain To better reveal the evolutionary relationships, we generated a phylogenetic tree based on the 32 AtBBXs and 31 SlBBXs (Fig. 2). All sequences were clustered into ve subfamilies according to the phylogenetic analysis and previous article [2]. The BBX genes in clade I (group 1) had two concatenated B-box domains, a CCT domain and a VP motif except for SlBBX1 and SlBBX2, which did not have a VP motif and a CCT domain. The members of group II were characterized by two B-box domains and also a CCT domain with the only exception for SlBBX7, which contained two B-box domains only, and SlBBX8 and SlBBX10, which only had one B-box domain and a CCT domain. In the group III, all the members contained one B-box domain as well as a CCT domain. The group IV and V possessed two and one B-box domain, respectively; nonetheless, SlBBX27 that contained two B-box was also grouped into V. Additionally, BBX proteins from two species showed scattered distribution across the branches of the evolutionary tree, which implies that the duplication events occurred after the lineages diverged.
Protein structural features are crucial for understanding the biological properties as well as the evolutionary origins of proteins. Here, we found that most members of SlBBX proteins in a subfamily had a similar three-dimensional structure (Fig. 3). In addition, we found physical connections in each protein sector in the tertiary structure. Moreover, a distinct functional role, and an independent mode of sequence divergence in the protein family, re ected the evolutionary histories of the conserved biological properties of BBX proteins.
Gene structure, conserved motifs, chromosomal localization and synteny analysis of SlBBXs The evolution of multigene families can be driven by gene structural diversity. Examination of the genomic DNA sequences revealed that most SlBBXs contained one to ve introns, while SlBBX16, SlBBX17 and SlBBX30 had no introns ( Fig. 4b and Table 1; Additional le 1: Figure S3). Among them, nine SlBBXs had one intron, followed by ten SlBBXs with two introns, ve SlBBXs with three exons, four SlBBXs with four exons, and one SlBBXs with ve introns. Generally, members of each subclass, which are most closely related, exhibited analogous exon-intron structures. For instance, the members in group I and V had one to two, and zero to one intron, respectively ( Fig. 4a and 4b; Additional le 1: Figure S3). However, a few SlBBX genes showed dissimilar exon-intron arrangements. For instance, SlBBX18 and SlBBX19 had high sequence similarity, but SlBBX18 and SlBBX19 contained two and ve introns, respectively ( Fig. 4a and 4b; Additional le 1: Figure S3). These divergences indicated that both the gain and loss of introns during evolution, may better explain the functional diversity of SlBBX homologous genes.
To further examine the structural features of SlBBXs, the conserved motif distributions were analyzed. Twenty conserved motifs were predicted (Fig. 4c), while multilevel consensus sequences and the E-value of them were shown in Additional le 2: Table S1. The results showed that motif 17 was the largest motif depending on the width, followed by motif 8 and motif 2 (Additional le 2: Table S1). Motif 1 was found in all the SlBBXs (Fig. 4c). Notably, 74.2% and 70.1% SlBBXs contained motif 4 and motif 3, respectively.
Motif 2 was unique to the group I, II and III of SlBBXs, while motif 5 was unique to group II except for the SlBBX27. Motif 10 was found only in group III of SlBBXs. Our results showed that members that are most closely related in the phylogenetic tree contained common motifs on the basis of alignment and position, which indicated that they may have a similar biological function.
Chromosomal locations showed that 31 SlBBX genes were unevenly distributed on the 12 chromosomes (Fig. 5a). A maximum number of SlBBX genes were found on chromosome 12 (Chr12), comprising of six genes. Five genes were located on Chr2 and Chr7. Four and three SlBBX genes were located on Chr5 and Chr4, respectively. Both Chr1 and Chr6 contained two members of SlBBX genes, whereas only one gene was detected on Chr3, 8, 9 and 10. Additionally, no SlBBX genes were found on Chr11.
To examine the duplication of SlBBX genes, the MCScanX program was used. Thirty-six pairs of SlBBXs were identi ed as segmental duplication in the tomato genome (Fig. 5b). Chr2, 7 and 12 had more duplication regions, which partially explain the greater numbers of SlBBX genes that were located on these three chromosomes. Although SlBBX1 and SlBBX3 were located on the same chromosome (Fig. 5a), and their sequence identity was 83% (Additional le 1: Figure S4), they were not tandem duplication. To further examine the evolutionary relationships between SlBBXs and AtBBXs, a synteny analysis was performed with MCScanX software. A total of 16 of SlBBX-AtBBX orthologous pairs were identi ed (Fig. 5c), which indicated the existence of numerous SlBBX genes prior to the divergence of Arabidopsis and tomato. Some members of SlBBXs were not localized in the syntenic block, suggesting that these genes might have certain speci city due to their evolution time.
Analysis of cis-elements in the promoter region of SlBBXs Transcription factors directly bind the cis-elements in regulatory networks controlling gene expression; therefore, analysis of the putative cis-elements is critical to examine the expression of SlBBX genes. A total of 61 major cis-elements were predicted from the PlantCARE database (Fig. 6a), including 22 light responsive, 12 hormone responsive, 11 stress responsive and 16 development. The number of light responsive cis-elements was the largest in the promoters of 31 SlBBX genes (Fig. 6b). The number of ciselements in the promoters of SlBBX17 and SlBBX2 was the largest and least, respectively. The major light responsive elements contained box4 (21%), G-box (17.9%) and CMA1a/2a/2b (14.3%), which were located on 87.1% (27/31), 83.9% (26/31) and 96.8% (30/31) of SlBBXs promoters, respectively (Fig. 6c). The most common motif were the JA-responsive elements (MYC), abscisic acid (ABA)-responsive element (ABRE), and ethylene-responsive element (ERE), accounting for 24.8%, 21.5% and 17.2% of the scanned hormone responsive motifs, respectively. The stress responsive elements MYB, STRE (stressrelated elements) and WUN were located on 96.8% (30/31), 90.3% (28/31) and 77.4% (24/31) of 31 SlBBX genes promoters, respectively. In the development category, various growth and development related elements, such as AT-rich element (19.2%), O 2 -site for zein metabolism regulation (13.7%), CATbox for meristem expression (12.3%), GCN4_motif required for endosperm expression (9.6%), were found. Our ndings suggest that the promoter regions of SlBBX genes that contained the cis-elements played a critical role in the light and stress responses.

SlBBX genes expression in response to different light quality
To assess whether light signaling regulates SlBBXs, we investigated the gene expression of SlBBXs in tomato plants grown at dark (D), white (W) and different light quality [purple (P), blue (B), green (G), yellow (Y), red (R), and far-red (FR)] conditions. In comparison with D, light decreased the transcripts of SlBBX1, SlBBX8, SlBBX10 and SlBBX12, while it increased the transcripts of SlBBX7, SlBBX13 and SlBBX15 (Fig. 7). Plants grown at R light conditions showed higher expression of SlBBX4, SlBBX14, SlBBX23, SlBBX24 and SlBBX29 than those grown at other light qualities. Whereas FR light signi cantly up-regulated the transcripts of SlBBX7, SlBBX13, SlBBX15, SlBBX21, SlBBX25, SlBBX26 and SlBBX27, it obviously down-regulated the transcripts of SlBBX14, SlBBX16, SlBBX18, SlBBX24, SlBBX28, SlBBX30 and SlBBX31 (Fig. 7). Transcripts of SlBBX16, SlBBX17, SlBBX18, SlBBX30 and SlBBX31 were induced, while transcripts of SlBBX5, SlBBX6, SlBBX19, and SlBBX20 were inhibited in plants when grown at B light conditions. SlBBX15 was induced by G light irradiation at 6 h, whereas SlBBX9 and SlBBX28 were repressed (Fig. 7). Y light led to an obvious reduction in expression of SlBBX9 and SlBBX31. Obviously, the P light increased the expression of SlBBX3, SlBBX5, SlBBX6, SlBBX15, SlBBX19, SlBBX20, SlBBX21, SlBBX26 and SlBBX27, but decreased the expression of SlBBX10 and SlBBX16. Interestingly, SlBBX4, SlBBX23 and SlBBX29 were only responsive to R light, while SlBBX7, SlBBX13 and SlBBX25 were induced just in response to FR light. Meanwhile, R light induced the expression of SlBBX14 and SlBBX24, but FR light inhibited their expression (Fig. 7). In general, the results showed that SlBBX genes might act a critical role in response to light quality signaling. Expression pattern of the SlBBX genes in response to chilling stress To investigate whether SlBBX genes participated in chilling stress, we analyzed the transcriptome data of tomato plants after chilling stress [43]. Results revealed that the expression levels of ten members of SlBBX family genes, including SlBBX3, SlBBX16, SlBBX17, SlBBX19, SlBBX24, SlBBX26, SlBBX28, SlBBX29, SlBBX30 and SlBBX31, were higher in tomato plants after chilling stress than those grown at optimal temperature conditions (Fig. 8). Furthermore, we found the transcripts of SlBBX1, SlBBX7, SlBBX9, SlBBX12, SlBBX13, SlBBX15, SlBBX18, SlBBX21, and SlBBX27 were signi cantly decreased after chilling stress. These ndings suggest that SlBBX genes might have an important role in response to chilling stress, whereas further studies are essencial to investigate the mechanism.

Discussion
Here, we identi ed and characterized 31 SlBBX genes in tomato ( Fig. 1; Tables 1 and 2), which contained two new members (SlBBX30 and SlBBX31) in comparison with the previous studies [44]. BBX proteins are characterized by one or two B-box domains at the N-terminal and, in some cases, a CCT domain at the Cterminal [1]. Here, we found both the newly retrieved SlBBX proteins (SlBBX30 and SlBBX31) contain a Bbox domain at the N-terminal ( Fig. 1 and Table 2), and they were also clustered in group V (Fig. 2). In addition, as shown in Fig. 3, the three-dimensional structures of SlBBX30 and SlBBX31 were similar to the other members of SlBBXs, which further indicated these two proteins were new SlBBX proteins. There were ve subfamilies in the 32 members of Arabidopsis BBXs according to the combination of different conserved domains [2]. However, the conserved domain-based classi cation of tomato BBX proteins was rather complex. As shown in Fig. 2, SlBBX1-6 were classi ed into group I, which had two B-boxes and a CCT plus a VP domains, whereas SlBBX1 lacked a VP domain, and SlBBX2 only contained two B-boxes (Table 2). Meanwhile, SlBBX7-12 were clustered into group II, which possessed two B-boxes and a CCT domains; however, SlBBX8 and SlBBX10 had one B-box and a CCT domains, while SlBBX7 contained two B-boxes. Group V contained only one B-box, except for SlBBX27, which contained two B-boxes. We investigated the detail of sequence alignment in SlBBXs (Fig. 1), and found a high degree of conservation of the B-box1 domain among SlBBX7-12, thus the clustering results of these proteins were similar to that based on B-box1. These results indicated that during the process of evolution, some SlBBX proteins lost the B-box2 domain. Since gene duplications play a vital role in genomic rapid expansions during evolution [45], we speculated the new two genes (SlBBX30 and SlBBX31) were retrieved because of genes duplication events. The identi ed SlBBX genes were distributed unevenly (Fig. 5a). There were no SlBBXs located in Chr11, while SlBBX30 and SlBBX31 were located in the Chr6 and Chr7, respectively. As shown in Fig. 5b, almost SlBBX genes were located within syntenic blocks. Among them, SlBBX30 on Chr6 and SlBBX31 on Chr7 had similarities with SlBBX28 on Chr12. These results revealed that segmental duplication events happened in the expansion of the SlBBX genes family in tomato.
Accumulating evidence showed that some BBX proteins act as central players in a variety of lightregulated physiological processes in plants. Here, we found that the number of light responsive ciselements was the largest in the promoters of 31 SlBBX genes (Fig. 6), which indicated that SlBBX genes were regulated by light signaling. Thus, we examined the gene expression of all the SlBBXs in response to different light quality. Results showed that light decreased the transcripts of SlBBX1, SlBBX8, SlBBX10 and SlBBX12, while increased the transcripts of SlBBX7, SlBBX13 and SlBBX15 compared with dark (Fig. 7). Previous studies had demonstrated that COP1, which is degraded after illumination, works as an E3 ubiquitin ligase that targets a variety of light signaling factors for ubiquitination and degradation in darkness [9,46]. For example, COP1 interacts with multiple BBXs, such as CO/BBX1 and BBX10, and subsequently degrades them by the 26S proteasome system [33,47]. Nevertheless, COP1 stabilizes BBX11 rather than degradating it [14], which suggests that COP1 likely degrades a yet unidenti ed component(s) targeting BBX11. Thus, COP1 may also control the stability of SlBBX proteins, including SlBBX1, SlBBX7, SlBBX8, SlBBX10, SlBBX12, SlBBX13 and SlBBX15, in the transition from dark to light. Interestingly, we found that SlBBX4, SlBBX23 and SlBBX29 were only in response to R light, while SlBBX7, SlBBX13 and SlBBX25 were just in response to FR light (Fig. 7). These results indicated that these SlBBX proteins might directly interact with the photoreceptors, which sense R and FR light signals. Similarly, recent work has revealed that phyB directly interacts with BBX4 and positively regulates its accumulation in red light in Arabidopsis [25], which demonstrates that photoreceptors may directly control some BBX proteins. In addition, the results showed that R light induced the expression of SlBBX14 and SlBBX24, whereas FR light inhibited their expression (Fig. 7), which implied that these two SlBBX proteins might function antagonistically to regulate some plant physiological processes, such as shade avoidance and the elongated of hypocotyls. Here, we observed that B light induced the gene expression of SlBBX16, SlBBX17, SlBBX18, SlBBX30 and SlBBX31, whereas inhibited the transcripts of SlBBX5, SlBBX6, SlBBX19 and SlBBX20 (Fig. 7). Previous work demonstrated that BBX28/BBX29 and BBX30/BBX31 could precisely control each other by forming a feedback loop in Arabidopsis [19,20,22,23]. Thus, these SlBBX proteins may work in concert with each other and some unidenti ed factors to regulate the plant growth in response to light signaling.
Light and temperature are more or less inter-related during plant growth and stress response [48]. Studies previously showed that BBX18 and BBX23 are involved in the thermomorphogenesis in Arabidopsis [38]. Both MdBBX20 and MaCOL1 are responsive to low temperatures in apple and banana, respectively [39,40]. ZFPL, a homologous gene of AtBBX32, enhances cold tolerance in grapevine [41]. Furthermore, CmBBX24 also increases plant cold tolerance in Chrysanthemum [42]. However, whether SlBBXs regulate plant cold response in tomato remains elusive. Here, we observed that there were numerous hormones and stress responsive cis-elements in the promoters of SlBBX genes (Fig. 7). Furthermore, the results showed that low temperatures induced the expression of SlBBX3, SlBBX16, SlBBX17, SlBBX19, SlBBX24, SlBBX26, SlBBX28, SlBBX29, SlBBX30 and SlBBX31, whereas inhibited the transcripts of SlBBX1, SlBBX7, SlBBX9, SlBBX12, SlBBX13, SlBBX15, SlBBX18, SlBBX21 and SlBBX27 (Fig. 8). We have demonstrated that SlHY5 positively regulates plant cold tolerance [49,50], and a variety of BBX proteins interact with HY5 [19,20,22,23]; thus, it indicates that SlBBXs may play critical roles in plant cold tolerance in tomato.

Conclusions
In conclusion, this investigation found two new members (SlBBX30 and SlBBX31) of the SlBBX gene family. We made a systematic analysis of the 31 SlBBX genes, including their conserved domain, phylogenetic relationship, three-dimensional structure, gene structure, chromosome location, gene duplication and cis-elements analysis. The results suggested that 31 members of SlBBXs were distributed unevenly in the whole genome, and no SlBBXs were located on chromosome 11. Gene duplication analysis indicated that segmental duplications had driven the expansion of the tomato BBX genes. Gene expression analysis revealed that multiple SlBBX genes highly responsive to light quality and low temperatures, which lay a foundation for understanding their biological functions in response to the crosstalk between light and temperature.

Plant material and growth conditions
Seeds of wild-type tomato (Solanum lycopersicum) in the cv 'Ailsa Craig' (Accession: LA2838A) background were obtained from the Tomato Genetics Resource Center (http: //tgrc.ucdavis.edu) as previously reported [48]. Seedlings, which grown in pots with a mixture of one part vermiculite to three parts peat , receive Hoagland nutrient solution. The growth conditions for tomato seedlings were 25/20 o C (day/night) temperature with a 12 h photoperiod, the light intensity of 600 µmol m -2 s -1 , and 65% relative humidity.

Light and cold treatments
The six-leaf stage plants were used for all experiments. Plants were grown at white light conditions with an aerial temperature of 25°C or 4°C for the cold treatment in controlled environment growth chambers (Ningbo Jiangnan instrument factory, Ningbo, China). For light quality treatments, plants were exposed to dark (D), white light (W) or different wavelength [purple (P), 394 nm; blue (B), 450 nm; green (G), 522 nm; yellow (Y), 594 nm; red (R), 660 nm and far-red (FR), 735 nm, Philips] light from 6:00 AM to 6:00 PM. The light intensity was 100 µmol m -2 s -1 . The Lighting Passport (Asensetek, Model No. ALP-01, Taiwan) was used to measure light intensity and light quality as a previous study [51].

Genome-wide identi cation of SlBBX genes in tomato
The protein sequence of Arabidopsis BBXs were downloaded from the TAIR database (https:// www.arabidopsis.org/). Tomato BBX proteins were searched and downloaded from three public databases, including the NCBI database (http:// www.ncbi.nlm.nih.gov/), the Phytozome 11.0 database (https://phytozome.jgi.doe.gov/pz/portal.html) and the Sol Genomics Network tomato database (https://solgenomics.net/). We chose the candidate BBX accroding to the E-value (1e -5 ) and the highest similarity scores. All the putative BBX genes were submitting to the InterProScan database (http://www.ebi.ac.uk/interpro/), SMART (http://smart.embl-heidelberg.de/) and Conserved Domains Database (http://www.ncbi.nlm.nih.gov/cdd/) to further con rm their completeness and existence of the core domains. The proteins without the B-Box domain and duplicate proteins were removed.
Chromosomal location, gene structure, tandem duplication and synteny analysis SlBBX genes were mapped to tomato chromosomes according to the Phytozome 11.0 database with the MapChart software. The chromosome distribution diagram was drawn by the online website MG2C (http://mg2c.iask.in/mg2c_v2.1/) with the information from Sol Genomics Network (http://www.solgenomics.net).
Exon and intron structures of the SlBBXs were determined accroding to their corresponding full-length gene sequences in Phytozome11.0 database. We performed gene structure analysis of the SlBBX genes by using the gene structure display server (GSDS, http://gsds.cbi.pku.edu.cn/) [59].

Cis-elements of promoters analysis
To identify potential light-, stress-, hormone-and development-related cis-elements, the 2000-bp genomic DNA sequence upstream of the start codon (ATG) of SlBBX genes were obtained from the tomato genome database. The cis-elements in these SlBBX genes promoter were identi ed by using the Plant Cis-Acting Regulatory Element (PlantCARE; http://bioinformatics.psb.ugent.be/webtools/plantcare/html/) [61].

Gene expression analysis
Total RNA was extracted from tomato leaves as previously [62,63]. RNA-seq data of the tomato plants after exposure to 25°C or 4°C for 6 h were obtained and performed as described previously [43,49]. Three biological replicates for various samples were prepared. For the gene transcript analysis, the cDNA template for real-time RT-PCR was synthesized using a Rever-Tra Ace qPCR RT Kit with a genomic DNAremoving enzyme (Toyobo). qRT-PCR experiments were carried out with an SYBR Green PCR Master Mix Kit (TaKaRa) using an Applied Biosystems 7500 Real-Time PCR System (qTOWER 3 G, Germany). The PCR was run at 95°C for 3 min, followed by 40 cycles of 30 s at 95°C, 30 s at 58°C, and 1 min at 72°C. The Tomato ACTIN2 gene was used as an internal control to normalize small differences in template amounts. The relative gene expression was calculated as described previously [64]. The primers sequence was in Additional le 2: Table S2.