Genome-wide identification and transcript analysis of TCP transcription factors in grapevine

Background The plant-specific TCP transcription factors play different functions in multiple processes of plant growth and development. TCP family genes have been identified in several plant species, but no comprehensive analysis of the TCP family in grapevine has been undertaken to date, especially their roles in fruit development. Results A total of 18 non-redundant grapevine TCP (VvTCP) genes distributing on 11 chromosomes were identified. Phylogenetic and structural analysis showed that VvTCP genes were divided into two main classes - class I and class II. The Class II genes were further classified into two subclasses, the CIN subclass and the CYC/TB1 subclass. Segmental duplication was a predominant duplication event which caused the expansion of VvTCP genes. The cis-acting elements analysis and tissue-specific expression patterns of VvTCP genes demonstrated that these VvTCP genes might play important roles in plant growth and development. Expression patterns of VvTCP genes during fruit development and ripening were analyzed by RNA-Seq and qRT-PCR. Among them, 11 VvTCP genes were down-regulated during different fruit developmental stages, while only one VvTCP genes were up-regulated, suggesting that most VvTCP genes were probably related to early development in grapevine fruit. Futhermore, the expression of most VvTCP genes can be inhibited by drought and waterlogging stresses. Conclusions Our study establishes the first genome-wide analysis of the grapevine TCP gene family and provides valuable information for understanding the classification and functions of the TCP genes in grapevine.

Background TCP proteins are a small family of plant-specific transcription factors and play important roles in multiple processes of plant growth and development by regulating cell growth and proliferation [1][2][3]. TCP transcription factors were named after four founding members: TEOSINTE BRANCHED1 (TB1) from Zea mays, CYCLOIDEA (CYC) from Antirrhinum majus, PROLIFERATING CELL NU-CLEAR ANTIGEN FACTOR 1 and 2 (PCF1 and PCF2) from Oryza sativa [4][5][6]. TCP proteins are featured by the TCP domain, a highly conserved 59-residue-long basic helix-loop-helix (bHLH) structure at the N-terminus, which is associated with DNA binding, protein-protein interaction and protein nuclear localization [7]. Based on the sequence features and homology of the TCP domains, TCP family members were classified into two subfamilies: Class I (represented by the PCF proteins) and class II (represented by CYC and TB1) [2,8]. The most noticeable difference between these two subfamilies is that class I members show a four-amino acids deletion in the basic region of the TCP domain [2]. The class II TCP members are further subdivided into two subclades (CIN and CYC/ TB1) based on the difference among their TCP domain. Furthermore, several class II members have an argininerich motif (R domain) with unknown functions, which is hypothesized to be involved in facilitation of proteinprotein interaction [1,2].
Increasing evidences show that TCP transcription factors play versatile functions in multiple physiological and biological processes during plant growth and development, such as branching [9,10], leaf morphogenesis [11,12], flower development [13,14], seed germination [15,16], hormone pathways [17,18] and response to environmental stress [19]. In Arabidopsis, AtTCP14 and AtTCP15 have been shown to regulate embryonic growth during seed germination by gibberellin signaling pathway [16]. They also could regulate leaf shape and internode length by promoting cell proliferation [12]. AtTCP16 is observably expressed in developing microspores, and its down-regulation generated 50% abnormal pollen in transgenic plants [20]. Recently, strong experimental evidence supports that class I members of TCP proteins could be implicated in fruit development and ripening [21,22]. Three tomato TCP genes (SlTCP12, SlTCP15 and SlTCP18) are preferentially expressed in the tomato fruit and their expressions are regulated by ripening-related transcription factor, such as RIPENING INHIBITOR (RIN) and COLORLESS NON-RIPENING (CNR) [21]. The strawberry FaTCP11 gene participates in ripening-related processes and regulates flavan-3-ols synthesis [23].
The functions of most class II members of TCP family have been elucidated. For example, the TB1 gene involves in the fate of maize axillary meristems [5] and the CYC gene affects the asymmetry, size and cell types of petals and stamens in Antirrhinum flower [4]. In Arabidopsis, AtTCP18 and AtTCP12, two homologs of TB1, are involved in suppressing bud outgrowth [9]. The tomato orthologs SlTCP9 (SlBRC1a) and SlTCP7 (SlBRC1b) also show similar functions in axillary bud initiation and outgrowth [24]. AtTCP1, the homolog of CYC, mediates plant growth and development by regulating the expression levels of brassinosteroid biosynthesis gene DWARF4 [25]. Five CIN-like genes including AtTCP2, AtTCP3, AtTCP4, AtTCP10 and AtTCP24 were targeted by miR319 and have been shown to be involved in regulating leaf and flower development [14,[26][27][28]. Moreover, AtTCP3 can increase flavonoid biosynthesis by interacting with R2R3-MYB proteins [29] and dominant-negative variant of AtTCP3 leads to shorter and crinkled siliques [30]. Transient over-expression of FvTCP9 in strawberry fruits dramatically promotes the expression of a series of genes involved in fruit color and aroma metabolism, suggesting that class II member of TCP family could be participated in fruit development and ripening processes [31].
To date, a number of TCP family members have been characterized in both dicots and monocots with the completion of entire genome, such as Arabidopsis [32], tomato [21], apple [33], strawberry [31], bamboo [34] and switchgrass [35]. However, little is known about the TCP family in grapevine [36], which is one of the most important fruit crop growing around the world with great nutritive and commercial value [37][38][39]. Due to the important roles of TCP transcription factors during plant growth and development, we performed for the comprehensive analysis of the VvTCP transcription factor family in grapevine. In the present study, 18 non-redundant TCP genes were identified from grapevine and were subsequently performed a systematic analysis including chromosome location, phylogenetic relationships, gene structure, conserved motif and cis-acting elements. We further analyzed the expression of VvTCP genes in diverse tissues, different stages of fruit development and ripening, as well as in response to hormones and stress treatment. This study provides reliable investigation of the VvTCP gene family and facilitates further functional characterization of TCP members in grapevine.

Identification of putative VvTCP in grapevine
Two different methods were peformed to identify and annotate TCP genes in grapevine genome. Firstly, the hidden Markov model (HMM) profile of the conserved TCP domain (PF03634) was downloaded from the Pfam database (http://pfam.janelia.org) and used to screen all grapevine proteins in the 12× coverage assembly of the V. vinifera PN40024 genome. Secondly, all Arabidopsis TCP protein sequences, which were downloaded from the Arabidopsis Information Resource (TAIR) database (http://www.arabidopsis.org), were used as queries to screen against grapevine genome database by using DNAtools software. Subsequently, all non-redundant VvTCP protein sequences were further verified for the presence of the TCP domain by screening against the Pfam (http://pfam.sanger.ac.uk/), InterProScan (http://www.ebi.ac.uk/Tools/pfa/iprscan/) and SMART (http://smart.embl-heidelberg.de/) database. The molecular weights (MW), isoelectric points (pI) and grand average of hydropathicity (GRAVY) of VvTCP proteins were calculated by the ExPasy website (https://web.expasy. org/protparam/). The subcellular location of VvTCP proteins was predicted by WoLF PSORT (http://www.genscript.com/psort/wolf_ psort.html).

Sequence alignment and phylogenetic analysis
Sequences of the 24 Arabidopsis and 22 rice TCP proteins were retrieved from TAIR (https://www.arabidopsis.org/) and rice genome database (http://rice.plantbiology.msu. edu/), respectively. The sequences of 30 tomato TCP family members were retrieved from the Solanaceae Genomics Network (https://solgenomics.net/). The sequences of 19 strawberry TCP family members were retrieved from PlantTFDB (http://planttfdb.cbi.pku.edu.cn/). The Antirrhinum CYC and maize TB1 were retrieved from NCBI database (https://www.ncbi.nlm.nih.gov/). ClustalX 2.0 software was used to perform the multiple sequence alignments of the amino acid sequences of the TCP proteins of grapevine, Arabidopsis, rice, tomato and strawberry. An unrooted phylogenetic tree based on the full length protein sequences sequence alignments was constructed using MEGA 7.0 software and the neighbor-joining method with the following parameters: pairwise alignment, 1000 bootstrap replicates, Poisson correction model, uniform substitution rates and complete deletion. Moreover, another phylogenetic tree was also constructed using all protein sequences of TCP domain in grapevine for further analysis. The motif logos of the VvTCPs were generated by submitting the sequences to the MEME website (http://meme.nbcr.net/ meme/cgi-bin/meme.cgi). Below are the parameters of MEME used: maximum number of motifs, 20; minimum motif width, 6; and maximum motif width, 50.
Chromosomal location, gene structure, and duplication analysis All VvTCP genes were mapped to grapevine chromosomes based on physical positions at the Grape Genome CRIBI website (http://genomes.cribi.unipd.it/) and the map was drawn using the MapInspect software. Accordingly, the cDNA sequences and their corresponding genomic DNA sequences of VvTCP members were obtained from the grapevine genome, then the exon-intron organization were identified by comparing the coding sequences with their corresponding genomic sequences using the GSDS software (http://gsds.cbi. pku.edu.cn) [40]. Tandem duplicated genes were defined by checking their physical locations on individual chromosomes and were identified as adjacent paralogous on a grape chromosome, with no more than one intervening gene [41]. For synteny analysis, the synteny blocks were detected by MCScanX software (http:// chibba.pgml.uga.edu/mcscan2/), with the E-value set below 1 × 10 − 5 taking reference from a previous study [42]. The diagrams were generated by the program Circos version 0.63 (http://circos.ca/) [43].

Expression profiles of VvTCPs in various organs and different berry developmental stages
The expression profiles of VvTCP genes were determined in a Vitis vinifera cv 'Corvina' (clone48) gene expression atlas of various organs at different developmental stages. Microarray data were obtained from the NCBI gene expression omnibus (GEO) datasets under the series entry GSE36128 (http://www.ncbi.nlm.nih.gov/geo/) [45]. The mean of expression value of each gene in all tissues/organs were analyzed and graphically represented using Multi Experiment Viewer (MeV) software [46]. The expression patterns of VvTCP genes in fruit developmental stages were acquired from gene expression omnibus (GEO) database of NCBI (GSE77218), which measured using RNAsequencing (RNA-Seq) data [47]. Berries from 3 year old grapevine trees 'Fujiminori' (V. vinifera× V. labrusca) were sampled in triplicate at the green fruit expanding (40DAF or DAF40), veraison (65DAF or DAF65), and ripe (90DAF or DAF90) stages throughout the growing season. Furthermore, expression analyses of VvTCP genes in 10 different grapevine (Vitis vinifera) varieties at four berry development stages were based on RNA-seq data (accession numbers GSE62744 and GSE62745) downloaded from the NCBI GEO datasets [48]. The 10 varieties contained five red-skinned (Sangiovese, Barbera, Negro amaro, Refosco and Primitivo) and five white-skinned berries (Vermentino, Garganega, Glera, Moscato bianco and Passerina). Berries were sampled in triplicate at four developmental stages, the pea-sized berry stage at 20d after flowering, the berries beginning to touch stage just prior to veraison (Pre_veraison), the berry-softening stage at the end of veraison (End_veraison), and the fully ripe berry stage at harvest.

The expression of VvTCP under stress condition
To investigate the expression profiles of TCPs in response to different stress treatment (Cu, salt, waterlogging and drought stress), grapevine RNA-seq data sets (SRA accession no. SRP070475 and SRP074162) were retrieved from NCBI GEO database (https://www.ncbi. nlm.nih.gov/geo/) or from published supplemental data sets [37,[49][50][51]. Two-year-old 'Summer Black' (hybrids of V. vinifera and V. labrusca) grapevine were used to investigate the expression of TCP genes in response to abiotic stresses. Cu stress of potted grapevine plants was simulated with 100 μM CuSO 4 and salt stress was treated with 0.8% NaCl [37,50]. The control plantlets were similarly treated with distilled water. Waterlogging treatment were performed by immersing the plants to water for 48 h [51] and drought treatment was performed by withholding water 20 days [49]. Grapevine plantlets grown in the standard conditions were used as a control. All types of samples were three replicates and the third and fourth unfolded leaves from the shoot apex was collected from treatment and control groups during deep sequencing. The analysis of RNA-seq data was according to previous method [37] and the RPKM (Reads Per Kilobase per Million mapped reads) values were used to estimate the gene expression level. The heatmap of TCP genes was exhibited using R software (http:// www.bioconductor.org/).

Plant growth condition and gene expression analysis using qRT-PCR
Four-years-old 'Fujiminori' grapevine trees, grown in the standard field conditions at the Qingdao Agricultural University fruit farm, Qingdao, China, were chosen as the experimental material. To investigate gene expression profiles of TCP genes during berry development and ripening, grapevine berry samples were also collected at three time points: the green fruit expanding stage (40 DAF), veraison (70 DAF) and ripe/harvest stages (90 DAF) throughout the growing season. All samples were collected in triplicate from each of the sampling points. The samples were immediately frozen in liquid nitrogen and stored at − 80°C until use.
A total of 200 mg of the grapevine tissues were used from above mentioned samples for total RNA isolation using the modified CTAB method [38], followed by DNaseI (Tiangen, Beijing, China) digestion to eliminate any contaminating DNA. For qRT-PCR analysis, the first-strand cDNAs was synthesized from the 1 μg RNA using a PrimeScriptTM RT Reagent Kit (TaKaRa, Dalian, China) according to the manufacturer's instructions. Expression pattern of various genes obtained from Microarray data was validated by qRT-PCR. The primers used for the qRT-PCR were designed using Primer 3.0 online and details of the primer sequences were presented in Additional file 2: Table S1. The grapevine housekeeping gene Actin (AB073011) was used as the internal control. The qRT-PCR was peformed using SYBR® Premixm Ex Taq™ (TaKaRa, Japan) with the Applied Biosystems 7500 Real-Time PCR System. All the experiments were carried out with three biological replicates. The 2 -ΔΔCT method was used to estimate the relative expression level [52].

Subcellular localization of grapevine TCP genes
Based on the grapevine genome and public NCBI database, the full coding sequences of three randomly selected VvTCP genes were PCR-amplified with highfidelity HS polymerase (TaKaRa Biotechnology, Dalian, China) using the primers listed in Additional file 2: Table S1. To construct green fluorescent protein (GFP)-tagged VvTCP, the three cloned VvTCP genes (35S, VvTCP2-GFP, 35S: VvTCP3-GFP and 35S: VvTCP18-GFP) were inserted into the pCAMBIA1300 vector, respectively. After electroporation of these construction into Agrobacterium tumefaciens EHA105, the transformed bacterial cells were activated and infected into the leaf tissue of Nicotiana benthamiana as previously described [53]. The transient expression of VvTCPs-GFP was observed 72 h later using a laser confocal microscope (Zeiss LSM700, Germany), the mCherry-labelled nuclear marker (NF-YA4-mCherry) was used to visualize the nucleus.

Identification of TCP gene family in grapevine
In order to identify and obtain the TCP genes in grapevine genome, the BLAST searches were performed at NCBI and other public databases. Subsequently, the HMM profile was employed to perform a global search of the grapevine genome (http://genomes.cribi.unipd.it/ grape/). After removing the redundant sequences, 18 non-redundant VvTCP genes were identified and mapped onto 11 out of 19 grapevine chromosomes (Additional file 1: Figure S1). Further, 18 VvTCP genes were annotated as VvTCP1 to VvTCP18 on the basis of their distributions in genome and relative linear orders among the respective chromosome.
Protparam tool was used to analyze the physical and chemical characterizations of the VvTCP proteins ( Table 1). The length of VvTCP proteins varied from 169 to (VvTCP14) 460 amino acid residues (VvTCP9). VvTCP14 showed the lowest value of the molecular weight (17.72 kDa), while the highest of the molecular weight (48.54 kDa) was observed in VvTCP6. The values of theoretical isoelectric point (pI) ranged from 6.09 to 9.71. The value of the aliphatic index ranged from 56.37 to 80.36, which suggested that the VvTCP proteins contained rich aliphatic amino acids. The GRAVY of all VvTCP proteins was less than zero, indicating that VvTCPs were hydrophilic. The majority of VvTCP proteins were predicted to be located on the nucleus by WoLF PSORT, but a few of them may be located in other subcellular compartments, such as chloroplast and cytoplasm (Table 1).

Phylogenetic analysis and classification of the VvTCP family
To explore the evolutionary and phylogenetic relationships between grapevine TCP proteins and other known TCPs, the full length of 115 TCP proteins from grapevine, Arabidopsis, rice, strawberry, tomato and two TCP genes (TB1 and CYC) with known function were used to construct a phylogenetic tree using Neiboring-Joining method (Fig. 1). Furthermore, in order to assess a better understanding of phylogenetic relationships of VvTCP members, multiplealignment of the core TCP domain of the all VvTCPs was also performed. Both the phylogenetic analysis and TCP domain alignment suggested that the grapevine TCP proteins were classified into two classes: class I (or PCF) contained 10 genes and class II contained 8 genes (Figs. 1a and 2a). Four-amino-acid fewer in the basic domain of class I than class II proteins was the most striking difference observed between these two classes (Fig. 2a). Additionally, the phylogenetic tree showed that class II could be further divided into two subclades, CYC/TB1 and CIN (Figs. 1a and 2a). Furthermore, all Arabidopsis, rice, strawberry and tomato TCPs existed the same class or clade as previous reports [21,31,54], confirming the reliability of our phylogenetic tree. According to the classification, the CYC/ TB1 subclade contained 3 VvTCP genes (VvTCP1, VvTCP10 and VvTCP15) and the CIN subclade included 5 VvTCP genes (VvTCP5, VvTCP6, VvTCP8, VvTCP11 and VvTCP18).
Expect for the TCP domain, several class II TCP members also share an R domain, which is an approximately 18-residues arginine-rich motif. As shown in Fig. 2b, four class II proteins, VvTCP1, VvTCP10 and VvTCP15 from grapevine class II CYC/TB1 as well as VvTCP6 from CIN, contained the R domain at the C-terminus of the TCP domain. The VvTCP6 in the CIN subclade was less conserved than CYC/TB1 subclade, in agreement with the previous in tomato and Phalaenopsis equestris [21,55]. Additionally, three CIN subclade genes (VvTCP5, VvTCP6 and VvTCP18) included the potential miR319 target site and displayed high sequence homology with the Arabidopsis and tomato miR319-targeted TCP genes (Figs. 1a and 2c).

Gene structure analysis and conserved motif identification
To further understand into the evolutionary relationships and structural features of the TCP protein in grapevine, the exon/intron structures and conserved motifs of VvTCPs were investigated. The conserved TCP domain sequences of VvTCP protein were used to construct a new phylogenetic tree, which also divided the VvTCP proteins into three subgroups (Fig. 3a). As shown in Fig. 3b, almost all VvTCP genes exhibited highly conserved exon-intron organization: 12 out of 18 VvTCP genes were no intron, four VvTCP genes had one intron, and two VvTCP genes had two introns. As expected, most of VvTCP genes within same subfamily exhibited similar distribution patterns of exon/intron in terms of exon length and intron number, which supported the classification of subclade and evolutionary relationship (Fig. 3b).
To get more insight into the diversity of motif compositions among VvTCPs, five conserved motifs were identified by MEME program. The results showed that the highly conserved TCP domain (motif 1) was existed in all VvTCP proteins ( Fig. 3c and Additional file 1: Figure  S2). The conserved R domain (motif 3) was hit in four class II VvTCP proteins. All class I members were characterized by motif 2 in C-terminal TCP domain. By comparison, the N-terminal TCP domain of motif 4 was detected in all class II proteins. Additionally, motif 5 were exclusively present in PCF, which was consistent with the previous report that some motifs existing in a particular subgroup may contribute to the specific function of those genes in the subgroup [31,56]. Togerher, VvTCP proteins clustered in same subgroup demonstrated similar motif composition, which was in agreement with the gene structure analysis.

Tandem duplication and synteny analysis of VvTCP genes
To reveal the mechanism for expansion and evolution of the VvTCP gene family, potential gene duplication events were investigated in the of grapevine genome. As illustrated in Fig. 4 and Additional file 3: Table S2, eight pairs of paralogous VvTCP genes were identified and distributed on different chromosomes in grapevine, whereas no tandem duplication events were observed, suggesting that segmental duplications were the main causes for the amplification of VvTCP gene family. In addition, six genes involved in two segmental duplication events (VvTCP1/ VvTCP10/VvTCP15 and VvTCP5/VvTCP8/VvTCP18).
Furthermore, a large-scale comparative synteny maps between grapevine and Arabidopsis, grapevine and tomato was analyzed at genome-wide levels with purpose to clarify the origin and function of TCP genes. A total of eight pairs of TCP genes were identified between grapevine and Arabidopsis (Additional file 1: Figure S3 and Additional file 3: Table S2), while 37 pairs of TCP   genes, including 12 VvTCP genes and 17 SlTCP genes, showed syntenic relationship (Additional file 1: Figure  S4 and Additional file 1: Table S2), suggesting that most TCPs had orthologous in Arabidopsis and tomato. Among the synteny events between grapevine and tomato, 8 VvTCP genes were found to be associated with at least three synteny events, such as VvTCP5-SlTCP1/SlTCP2/SlTCP3/ SlTCP10/SlTCP30 and VvTCP15-SlTCP7/SlTCP8/SlTCP9/ SlTCP22/SlTCP25 (Additional file 3: Table S2). Interestingly, six out of these eight genes were in CIN and CYC/TB1 subclade, indicating a higher conservation of CIN and CYC/TB1 than PIF subclade in TCP gene family.
Promoter Cis-regulatory elements analysis of grapevine VvTCP genes To further insight into the gene function and regulation mechanism of VvTCP genes, the cis-regulatory elements in promoter sequences were analyzed. The  Table S3). As show in Fig. 5, two cis-acting regulatory elements involved in endosperm expression (GCN4_motif and Skn-1_motif) were identified in promoter region of 6 and 17 VvTCP genes, respectively. Three cis-acting regulatory elements were related to meristem expression (CAT-box, CCGTCCbox and dOCT) in plant growth and development.
The shoot-specific expression element (as-2-box) and circadian control element (circadian) were found in 7 and 8 VvTCP genes, respectively. Additionally, the flavonoid biosynthetic (MBSI), zein metabolism regulation element (O2 site) and root specific (motif I) regulatory element were also found in the promoter region of the VvTCP genes ( Fig. 5; Additional file 4: Table S3).
Ethylene-responsive element (ERE) and auxin-responsive element (AuxRR-core and TGA-element were observed in 5 and 6 VvTCP genes respectively ( Fig. 5; Additional file 4: Table S3). Plenty of hormone-responsive elements were observed in the promoter region of VvTCP genes, revealing that hormones could play important functions in the regulation of plant growth and development (Fig. 5). In stress-related cis-acting elements, anaerobic induction (ARE), drought-inducibility (MBS), heat stress (HSE) and low-temperature (LTR) responsiveness element were also detected in the promoters of 14, 15, 12 and 6 VvTCP genes, respectively ( Fig. 5; Additional file 4: Table S3).

Tissue-specific expression patterns of VvTCP genes in grapevine
To gain more insights in potential roles of VvTCP genes during grapevine development, the organicspecific expression patterns of all the VvTCP genes were analysed using an expression atlas of V. vinifera cv. 'Corvina' from the GEO DataSets (GSE36128), which contained 42 various organs/tissues at different developmental stages obtained by microarray analysis [45]. Hierarchical clustering was used to present the relative expression levels of VvTCP genes in different tissues. As showed in Fig. 6, some VvTCP genes shared similar expression profiles in various tissues, while other VvTCP genes presented significant tissuespecific expression patterns, possibley suggesting the functional divergence of VvTCP genes in grapevine  Supplementary Table S2) organs/tissues during development. For example, VvTCP6 and VvTCP12 were constitutively high expressed in almost all tested issue, whereas VvTCP1, VvTCP5, VvTCP10 and VvTCP14 were expressed at extreme low levels in all tissues (Fig. 6, Additional file 5: Table S4).
In contrast, the expression levels of VvTCP2 and VvTCP16 were very high in young inflorescence, seedling and woody stem and they were relatively low expression in seed and pollen, implying that they might be involved in the development of inflorescence, seedling and woody stem (Fig. 6, Additional file 5: Table S4). VvTCP15 was only high relative expression level in latent bud, bud swell or bud burst and at almost undetectable levels in other tissues, suggesting that VvTCP15 might plant an important role in the development of grapevine buds (Fig. 6, Additional file 5: Table S4). VvTCP11 displayed high expression in petal, young inflorescence or well-developed inflorescence, indicating an involvement in flower development. Additionally, VvTCP8 showed relatively high expression in young leaves and inflorescence, VvTCP18 was extremely high transcript levels in young leaves, mature leaves, burst bud and bud after burst. Remarkably, some VvTCP genes (VvTCP2, 3, 6, 8, 9 and 11) were gradually decreased expression patterns from the green fruit stage to the veraison/ripe stage (Fig. 6, Additional file 5: Table  S4), which indicated that these genes might play important roles in fruit development. These results prompted us to investigate the transcript accumulation patterns of VvTCP genes during grapevine fruit development and ripening.

Expression patterns of VvTCP genes during different berry developmental stages
To understand the potential function of VvTCP genes in berry development and ripening, the transcript accumulation patterns of 18 VvTCP genes were investigated during three fruit developmental stages in grapevine using the expression profiles from the GEO DataSets (GSE77218) [47]. As shown in Fig. 7a, five VvTCP genes (VvTCP1, 5, 10, 14 and 15) were almost undetectable during the whole processes of berry development in grapevine (Fig. 7a, Additional file 6: Table S5). Eleven VvTCP genes (VvTCP2, 3, 4, 6, 7, 8, 9, 11, 12, 16 and 18) displayed the highest expression levels at green fruit stage, and then showed decreasing trend from veraison till to ripe stage, indicating potential roles during early berry development. On the contrary, the expression of VvTCP13 was increased gradually during three berry development stages. In order to validate the expression pattern of VvTCP genes in the various developmental stages of the berry by microarray data, qRT-PCR analysis of 12 detectable VvTCP genes was further performed at three berry development stages. As was expected, qRT-PCR results were highly consistent with the RNA-Seq data except for VvTCP17 (Fig. 7b). For example, VvTCP2, VvTCP3 and VvTCP7 showed a relatively high expression levels in green stage but decreased sharply in veraison stage, and then changed slightly from veraison to ripening stage (Fig. 7b). VvTCP13 was significantly higher expression in at ripe stage than that in green stage. However, the expression profiles of VvTCP17 did not correspond with RNA-Seq data. VvTCP17 was relatively high expression in veraison berry from RNA-Seq data, whereas the qRT-PCR analysis showed the highest expression in green berry stage (Fig. 7b). All these results implied that VvTCP genes might be involved in grapevine fruit development.
To provide more information on the berry developmental and ripening functions of VvTCP genes in grapevine, we investigated their transcript accumulation patterns among 10 different grapevine varieties by using microarray data (accession numbers GSE62744 and GSE62745), which consists of four different fruit developmental stages (the pea-sized berry stage at 20d after flowering, the berries beginning to touch stage just prior to veraison, the berry-softening stage at the end of veraison, and the fully ripe berry stage at harvest [48]. As shown in Fig. 8, four detected VvTCP genes (VvTCP6, 7, 9 and 11) were relatively higher expression in pea-sized berry and Pre_veraison stage and rapidly down-regulated during ripening, which were corresponded with the data from RNA-Seq and qRT-PCR analysis. Interestedly, VvTCP6 were intensely expressed in pea-sized berry, implying that VvTCP6 may play an important role during the early stages of grapevine berry development. Additionally, VvTCP13 was only detected at ripe stage (Fig. 8), which indicated that VvTCP13 might function in grapevine fruit ripening. All these results indicated that some VvTCP genes might play important roles in grapevine fruit development.

Transcript profiling of VvTCP genes under various abiotic stress treatments
Copper (Cu), salt, waterlogging and drought are common types of abiotic stresses in vineyards. To investigate the potential roles of the VvTCP genes in responses to Fig. 7 Expression profile of grapevine VvTCP genes during three fruit developmental stages. a Hierarchical clustering of the transcript accumulation profiles of 18 VvTCP genes during three berry developmental stages. b RT-qPCR transcript analysis of 12 selected VvTCP genes at three berry developmental stages. Berries from 3 year old 'Fujiminori' grapevine trees were sampled in triplicate at the fruit expanding (40DAF or DAF40), veraison (65DAF or DAF65), and ripe (90DAF or DAF90) stages throughout the growing season. The experiments were repeated three times and provided consistent results different environmental stresses, the RNA-seq datas were collected for the 18 VvTCP genes in the leaves of the grapevine exposed to Cu, NaCl, waterlogging and drought treatment (Fig. 9, Additional file 7: Table S6). Overall, the VvTCP genes responded to waterlogging and drought stress to a greater extent than to Cu and NaCl treatment. For example, eight VvTCP genes were regulated in response to waterlogging treatment and seven VvTCP genes responded to drought stress (Fig. 9, Additional file 7: Table S6). In contrast, only three (VvTCP8, 9 and 13) and one (VvTCP3) VvTCP genes were down-regulated expression in response to Cu and salinity stress, respectively, while the other VvTCP members were only slightly down-regulated or remained nearly unchanged (Fig. 9, Additional file 7: Table S6). Notably, three VvTCP genes (VvTCP8, 9 and 13) responded to at least three treatments, indicating that these genes might be involved in multiple stress response processes. Moreover, the expression difference of VvTCP depended on the type of stress. VvTCP13 was up-regulated in response to waterlogging stress, but was down-regulated in response to Cu and drought stress.

Subcellular localization of VvTCP proteins
It is well know that the nuclear localization of transcription factors is very important for regulate the transcription of target genes by binding to specific ciselements in their promoters. Previous studies have showed that TCP proteins were predominantly located in the nucleus, such as FvTCP8, FvTCP9 and FvTCP13 in strawberry. In this study, the majority of VvTCP proteins were predicted to be located on the nucleus by WoLF PSORT (Table 1). To characterize the subcellular localization of the VvTCP, three cloned VvTCP genes (VvTCP2-GFP, VvTCP3-GFP and VvTCP18-GFP) were introduced into the pCAM-BIA1300 vector by CaMV 35S promoter. The recombinant three fusion constructs was infiltrated into N. tabacum epidermal cells. As indicated in Fig. 10, green fluorescence signals from the expressed fusion VvTCP2-GFP, VvTCP3-GFP and VvTCP18-GFP were specifically distributed within the nuclei as confirmed by a mCherry-labelled nuclear marker (NF-YA4-mCherry). These results showed that VvTCP2, VvTCP3 and VvTCP18 were nuclear proteins, and   consistent with the prediction results and previous studies in strawberry [31].

Discussion
The plant-specific TCP transcription factors are known to play important roles in diverse aspects of physiological and biological processes during plant growth and development. To date, the TCP gene family have been investigated and characterized in various plant species such as Arabidopsis [32], tomato [21], apple [33], strawberry [31] and peach [22]. However, virtually no systematic and comprehensive informations of the TCP gene family in grapevine, a nutritious and economically important fruit crop all over the world, have been undertaken. In present study, 18 non-redundant VvTCP genes were identified and analyzed from grapevine genome. Furthermore, we peformed a multi-level analysis of the VvTCP genes in grapevine by investigating their evolutionary relationships, gene structure, protein motifs, duplication events, cis-acting elements, expression profiles in different tissues and developmental stages and under various stress treatment. The systematic characterization of VvTCP genes in grapevine will provide a better foundation for further functional studies of this gene family during grapevine growth and development.
Evolutionary conservation and divergence of the VvTCP gene family in grapevine Phylogenetic analysis and sequence alignment showed that all 18 VvTCP were classified into three major subgroups, which was consistent with the previous described in Arabidopsis, rice, tomato and strawberry [21,31,54]. Each subgroups contained TCP genes from Arabidopsis, rice, tomato, strawberry and grapevine (Fig. 1a). Furthermore, VvTCP members from the same group or subgroup shared a similar motif composition and intron/exon organization. For example, motif 2 and 4 were only present in class I and class II subgroup, respectively (Fig. 3). The consistency of the motif compositions and the exon/intron structures of VvTCP genes further supported the close evolutionary relationships.
In addition, the number of TCP genes was relatively conserved among Arabidopsis (24 members), rice (22 members) and strawberry (19 members) (Fig. 1b). However, it was significantly smaller than that present in tomato (30 members) and apple (52 members), which was consistent with the genome sizes of tomato (960 Mb) [57] and apple (742 Mb) [58], implying that TCP genes in various plants have expanded to different degrees. It is found that many TCP genes in tomato and Arabidopsis had three counterparts in grapevine (Additional file 1: Figure S3 and Additional file 1: Figure S4), indicating that the expansion of TCP family in grapevine may be caused by genome duplication events such as segmental duplication and whole-genome duplication. Our analysis showed that the number of paralogous TCP gene pairs accounted for over 50% of the entire TCP gene family in grapevine (Fig. 4). The fact supported the view that segmental duplication was a predominant duplication event for TCP genes and the major contributor to the expansion of TCP gene family in grapevine, as described previously, in Arabidopsis and cotton [59].

Functional divergence of VvTCP genes in plant growth and development
Increasing evidences suggest that TCP gene family involved in a wide range of functions during plant growth and development processes such as in shoot branching, leaf development, flower development and senescence [2]. The expression pattern of VvTCP genes in 42 different grapevine organs/tissues were investigated using an expression atlas of V.vinifera cv Corvina [45]. The expression analysis indicated that some VvTCP genes can be classified together according to their expression abundant in specific tissues of grapevine, probably reflecting their involved in a common metabolic and/or developmental process.
Previous studies have shown that TCP genes play important roles in leaf senescence and development [60]. In Arabidopsis, five CIN-like subfamily of TCP (AtTCP2, 3, 4, 10 and 24) were targeted by miRNA319 [61]. Ectopic expression of a miR319-insensitive TCP4 (mTCP4) gene led to the formation of miniature leaves during early stages of leaf development [62]. Proper regulation by miR319a of TCP4 was also important for petal and stamen development [14,63]. Furthermor, overexpression of miR319 or inhibition of multiple CIN-like TCP genes leat to delayed senescence, whereas overexpression of CIN-like TCP genes accelerates leaf senescence by activating biosynthesis of the hormone jasmonic acid [26,64]. In grapevine, VvTCP5, VvTCP6 and VvTCP18, the three closest homologs of these Arabidopsis CIN-like genes, had putative binding sites for VvmiR319b (Fig. 2c). VvTCP5 was only relatively high expression level in senescing leaf (Fig. 6), implying its potential function in leaf senescence. VvTCP6 and VvTCP18 were expressed at high levels in young leaf, mature leaf, young inflorescence, bud after burst and burst bud (Fig. 6), which were in agreement with the previous reports that CIN-like TCP genes and its post-transcriptional regulator miR319 play pivotal roles in leaf and flower development. These observations suggested that these miRNA-targeted VvTCP genes were likely to perform similar roles in leaf and flower development in grapevine to those of the Arabidopsis homologs.
The CYC/TB1 subgroup is mainly participated in the axillary meristems development [2]. In Arabidopsis, AtTCP1, the most closely homolog of CYC, was involved in the longitudinal elongation of leaves. AtTCP1 was strong expression in the petiole, lower portion of the inflorescence stem, and the midrib and distal region of expanding rosette leaves [65]. VvTCP10, which was closely homology with AtTCP1, was transcribed at relatively high levels in young inflorescence, green stem, bud after burst, burst bud or well-developed inflorescence and was almost undetectable in other tested tissues (Fig. 6). This result was partly consistent with the expression profile of AtTCP1 in Arabidopsis, indicating that VvTCP10 might play roles in flower and bud development in grapevine. AtTCP18 acted downstream of auxin and strigolactone to coordinate axillary bud outgrowth and up-regulation of AtTCP18 led to an inhibition of lateral branching. In contrast, mutation of AtTCP18 resulted in an increased number of rosette branches. AtTCP12 displayed a weaker or no mutant phenotype compared with AtTCP18 [9,66]. The VvTCP1, the homolog of the AtTCP12, was almost undetectable in all tissues, which implied that VvTCP1 was a potential functional redundant TCP member.
By contrast, most Class I genes, which usually play roles in cell growth and proliferation, exhibited more widespread and less tissue-specific expression patterns, such as in leaf, flower, stem and fruit (Fig. 6). These findings suggested that these Class I VvTCP genes might play various regulatory roles at multiple growth and development process. For example, AtTCP14 and AtTCP15 were involved in cell proliferation during seed, leaf and internode development [12,16]. AtTCP16 was proposed to modulate early pollen development [20]. AtTCP19 and AtTCP20 negatively regulated the onset of leaf senescence by jasmonate signaling pathway [7,17]. All of these AtTCP genes had at least one counterpart in grapevine, indicating that Class I TCP in grapevine might share similar functions with Arabidopsis homologs. Taken together, the above-mentioned results from model plants underlined that the TCP family members performed diverse biological functions in multiple plant growth and development processes.

Potential roles of VvTCP genes during berry development and ripening
Fruit development and ripening is a complex process and requires highly coordinated developmental events which were mainly controlling by a set of TFs regulatory networks [67]. In grapevine, all three genes from CYC/ TB1 subgroup, VvTCP1, VvTCP10, and VvTCP15 were not expressed in grapevine fruit (Fig. 7a), implying that these three genes in CYC/TB1 subgroup were rarely related to berry development and ripening. Similarly, some SlTCP genes in CYC/TB1 subgroup, such as SlTCP7, SlTCP8, SlTCP9 and SlTCP22, were almost undetectable in fruits [21], indicating that TCP genes in CYC/TB1 subgroup might not be associated with fruit development and ripening in tomato and grapevine. Interesting, the CYC/TB1 subgroup genes in strawberry, FvTCP6 and FvTCP14, showed the increased expression during the berry ripening process, and overexpression FvTCP9 transiently by agro-infiltration in strawberry fruits effectively increased the expression levels of ripening-related genes [31], which suggested that they might be involved in strawberry fruit ripening. Taken together, the TCP genes in CYC/TB1 might play variable roles in fruits development and ripening of different species.
Additionally, little was known about the functions of CIN clade members in fruit development. For example, PpTCP.C1 and PpTCP.D1.1 were high transcript accumulation levels in early fruits in peach, but were not associated with fruit ripening, suggesting that these two CIN clade members were likely to be involved in early peach fruit development. In this study, the CIN clade included five TCP members in grapevine, and of these genes, VvTCP5 was not expressed in grapevine fruits, indicating that VvTCP5 was irrelevant to grapevine fruit development and ripening. Moreover, four VvTCP genes in CIN clade was also relatively high expression in early grapevine fruit (Fig. 7), which were generally in agreement with the expression profile of PpTCP.C1 and PpTCP.D1.1 in peach. The similar expression patterns suggested that VvTCP genes in CIN clade was likely to perform roles similar in early fruit development in grapevine.
In tomato, three TCP genes in PCF (class I) subgroup, including SlTCP12, SlTCP15, and SlTCP18, were dominantly expressed in tomato fruits, implying that these TCPs were likely to play important roles during fruit development and ripening [21]. In present study, nine out of ten VvTCP genes in PCF subgroup were down-regulated expression during the berry ripening process, except for VvTCP13 (Fig. 7), indicating that these VvTCP genes might play a regulatory role during the early stages of berry development. Of these nine genes, VvTCP7, VvTCP9 and VvTCP12 which were individually homologous of SlTCP15, SlTCP12 and SlTCP18, might be involved in fruit development, due to the high levels of expression in the developing grapevine fruit (Figs. 7 and 8). In peach, the expression of PpTCP.A2 was negatively correlated to fruit ripening, and silencing of PpTCP.A2 enhanced the expression of PpACS1 and increased ethylene production, indicating that PpTCP.A2 was probably involved in fruit ripening by regulating ethylene biosynthesis [22]. Similarly, the expression of VvTCP2, which was a homologous gene of PpTCP.A2, was consistent with the expression profile of PpTCP.A2 and implied that the VvTCP2 gene was likely to play similar roles to PpTCP.A2 in grapevine fruit development and ripening.

Conclusions
In conclusion, 18 VvTCP genes were identified in the grapevine genome, which were distributed on 11 chromosomes. These VvTCP genes were divided into two classes based on the phylogenetic and structural feature. A lot of cis-acting elements were observed in the VvTCP promoter sequences, implying that VvTCP gene were controlled by a complex regulatory network. VvTCP genes might play important roles during grapevine growth and development as indicated by their spatial and temporal expression patterns. Notably, most VvTCP genes in grapevine were higher expressed in fruitlets than in other developmental and ripening fruits, indicating that these VvTCP genes were probably involved in early development in grapevine fruit. Taken together, all these findings will lay a solid foundation for for further unraveling the functions of VvTCP genes in grapevine growth and development.