Genome-wide identification and expression analysis of the B-box transcription factor gene family in grapevine (Vitis vinifera L.)

B-box (BBX) zinc-finger transcription factors play important roles in plant growth, development, and stress response. Although these proteins have been studied in model plants such as Arabidopsis thaliana or Oryza sativa, little is known about the evolutionary history or expression patterns of BBX proteins in grapevine (Vitis vinifera L.). We identified a total of 25 VviBBX genes in the grapevine genome and named them according to the homology with Arabidopsis. These proteins were classified into five groups on the basis of their phylogenetic relationships, number of B-box domains, and presence or absence of a CCT domain or VP motif. BBX proteins within the same group showed similar exon-intron structures and were unevenly distributed in grapevine chromosomes. Synteny analyses suggested that only segmental duplication events contributed to the expansion of the VviBBX gene family in grapevine. The observed syntenic relationships between some BBX genes from grapevine and Arabidopsis suggest that they evolved from a common ancestor. Transcriptional analyses showed that the grapevine BBX genes were regulated distinctly in response to powdery mildew infection and various phytohormones. Moreover, the expression levels of a subset of BBX genes in ovules were much higher in seedless grapevine cultivars compared with seeded cultivars during ovule development, implying a potential role in seed abortion. Additionally, VviBBX8, VquBBX15a and VquBBX29b were all located in the nucleus and had transcriptional activity except for VquBBX29b. The results of this study establish the genome-wide analysis of the grapevine BBX family and provide a framework for understanding the biological roles of BBX genes in grapevine.


Background
Transcription factors (TFs) play varied and important roles in plant growth, development and biological responses [1]. Zinc-finger TFs are one of the most populous classes of TFs in plants, and can be classified into several families based on the number and location of characteristic amino acid sequence motifs [2]. The B-box (BBX) zinc-finger TFs contain one or two conserved domains of approximately 40 amino acids near the amino terminus, and may also contain a CCT (CONSTANS, CO-like and TOC1) domain and/or a valine-proline (VP) motif at the carboxyl terminus [3]. In Arabidopsis thaliana (Arabidopsis), BBX genes have been identified and classified into five subfamilies based on presence of these domains [4]. Additionally, it has been reported that the conserved B-box domain mediates protein-protein interactions, while the CCT domain functions in transcriptional regulation [5,6].
BBX proteins participate in myriad biological processes in plants. In Arabidopsis, AtBBX1, AtBBX4, AtBBX7 and AtBBX32 regulate photoperiodic flowering [7][8][9][10]. Meanwhile, similar flowering roles have been reported in rice [11], barley [12], sorghum [13] and Chinese Cabbage [14]. Other studies have found that several AtBBX proteins participate in seedling photomorphogenesis through the HY5-COP1 regulatory module [15][16][17]. For instance, the HY5 transcription factor directly binds to a G-box cis-element present in the promoters of AtBBX30 and AtBBX31 and represses their expression, thus negatively regulating photomorphogenesis [18]. In rice, OsBBX14 promotes photomorphogenesis by directly binding the T/G-box cis-element of the OsHY5L1 promoter under blue light conditions [19]. In pear, PpBBX16 (the homolog of AtBBX22) and PpHY5 jointly activate the expression of PpMYB10 and other structural genes to positively regulate light-induced anthocyanin accumulation [20]. In apple, MdBBX20 integrates the influence of ultraviolet radiation and low temperature to promote the accumulation of anthocyanin [21]. In addition, BBX proteins have also been found to participate in response to environmental stress. For example, AtBBX31 promotes tolerance to UV-B radiation in Arabidopsis [22], and CmBBX22 regulates leaf senescence in chrysanthemum [23]. Heterologous expression of apple MdBBX10 in Arabidopsis enhances tolerance to salt [24]. In grapevine, VvCOL and VvCOL1 (VviBBX2 and VviBBX5, respectively) participate in flowering and bud dormancy [25].
Grapevine (Vitis vinifera L.) is one of the most economically important perennial fruit crops throughout the world. Grapes can be consumed fresh or dried, or can be processed into juice, wine, and jam. Seedless grapevine cultivars are particularly important, especially for fresh and dried fruit. However, worldwide production of both seeded and non-seeded grapes is increasingly limited by biotic and abiotic stresses. Cultivar improvement through traditional breeding and biotechnology is an exciting prospect, but options have been limited by the general lack of knowledge about key genes that mediate stress responses. Fortunately, the publication of a draft grapevine genome [26] has facilitated the identification of transcription factors. In this study, we identified members of the BBX family in grapevine from the draft genome sequence and gained insight into their potential function based on gene and protein structure, phylogeny, synteny, subcellular localization and transcriptional activity, as well as expression during ovule development, in response to pathogen challenge and various phytohormones. Taken together, this work will be helpful for future studies of BBX gene functions in grapevine.

Identification and characterization of grapevine BBX genes
To identify BBX genes in the grapevine genome, we employed a Hidden Markov Model (HMM)-based approach and the amino acid sequence profile of the Bbox-type zinc-finger domain (Pfam; PF00643). The resulting protein sequences were assessed for the presence of a B-box domain as defined by the Simple Modular Architecture Research Tool (SMART; http:// smart.embl-heidelberg.de/) and the Conserved Domain Database (CDD; https://www.ncbi.nlm.nih.gov/ Structure/cdd/cdd.shtml). This resulted in the identification of 25 putative BBX genes. For the sake of nomenclature and consistency, these were designated as VviBBXs (Table 1, Additional file 1: Text S1), based on the recently proposed grapevine nomenclature system [27]. The length of the encoded proteins ranged from 127 to 469 amino acids, and their predicted molecular mass ranged from~14.3 to 50.9 kDa. The isoelectric points of the predicted proteins ranged from~4.1 to 8.7 (Table 1).

Phylogeny and conserved domains of the grapevine BBX proteins
To analyze the evolutionary relationship and potential functional divergence of the VviBBX gene family, a total of 205 BBX proteins, including 32 from Arabidopsis, 29 from tomato, 30 from rice, 64 from apple and 25 from pear, were used to construct a phylogenetic tree (Fig. 1, Additional file 2: Text S2). This resolved the grapevine BBX proteins into five clades which mostly corresponded to their assigned structural groups (Table 1, Fig. 3a). However, an exception was found in the above clades. The VviBBX27 protein was presumptively phylogenetically in clade IV based on the structure group, but it was located in phylogenetic clade V (Fig. 1). As shown in the phylogeny tree, it is evident that BBX genes of the woody plants (grapevine, apple and pear) clustered together. And most of the grapevine BBXs also clustered together with proteins from Arabidopsis and tomato, instead of rice, consistent with the closer relationship of grapevine to the two eudicots.
The conserved sequences of the B-box1 and B-box2 zinc finger domains were C-X 2 -C-X 7-8 -C-X 2 -D-X-A-X-L-C-X 2 -C-D-X 3 -H-X 2 -N-X 4 -H and C-X 2 -C-X 8 -C-X 7 -C-X 2 -C-X 4 -H(N)-X 6-8 -H, respectively. In addition, the CCT domain of twelve of the grapevine proteins with the form of R-X 5 -R-Y-X 2 -K-X 3 -R-X 3 -K-X 2 -R-Y-X 2 -R-K-X 2 -A-X 2 -R-X-R-X 2 -G-R-F-X-K was highly conserved. A graphical representation of amino acid conservation with these motifs is shown in Fig. 2. Alignment of the protein sequences revealed that the B-box1 domain was more conserved than the B-box2 as a result of five absolutely  Fig. S1). Based on amino acid sequence conservation, number of B-box domains, and the presence or absence of the CCT domain, the 25 grapevine BBX proteins fell into five distinct structural classes (Table 1), which is consistent with previous results in Arabidopsis [4]. Group I, comprising three of the proteins, contained two B-box domains and one CCT domain. The seven representatives of Group II also contained two B-box domains and a CCT domain, but were distinguished from Group I based on the absence of the highly conserved amino acid sequence (SANPLARR) in the B-box2 domain and VP motif amino-terminal to the CCT domain seen in Group I proteins (Additional file 4: Fig. S2). Group III, comprising two proteins, contained one B-box domain and one CCT domain. The eight members of Group IV contained two B-box domains, while Group V proteins (five members) had only one B-box domain.

Analysis of conserved protein motifs and exon-intron structure of VviBBX genes
To gain additional insight into the conservation and diversification of the grapevine BBX gene family, we analyzed the conserved protein motifs encoded by the genes, as well as exon-intron structures (Fig. 3). Sixteen conserved motifs were identified (Fig. 3b), with four corresponding to B-box1 (Motifs 1/5), B-box2 (Motif 3), and CCT (Motif 2). Interestingly, we observed that Motifs 8, 13, and 14 were present only in Group III, which might contribute to the functional divergence of BBX genes. Motifs 6 and 7 were seen in all members of Group II, but also in VviBBX27 in Group IV, suggesting that VviBBX27 may have evolved from a Group II gene. The motif sequences and logos are listed in Additional file 5: Table S1. Additionally, VviBBX22b was the longest BBX gene (14.3 Kb). We also found that three genes in Group V (VviBBX28, VviBBX30 and VviBBX32) had a single exon, while all others carried between two

Chromosomal distribution and synteny analysis among VviBBX genes
Based on their annotated genomic locations, the 25 VviBBX genes were found to be widely distributed among the grapevine chromosomes ( Fig. 4). Chromosome 12 contained the most VviBBX genes (four), whereas Chromosomes 1 and 19 both possessed three genes, Chromosomes 3, 4, 14, and 18 had two VviBBX genes, and Chromosomes 5, 7, 9, 10, 11, and 17 had only one gene. The chromosomal location of VviBBX28 was on the chromosome Unknown. Segmental duplications and tandem duplications contribute to the evolution of gene families [28]. According to Fig. 4 and Additional file 6: Table S2, four segmental duplication events have occurred: VviBBX9 to VviBBX10, VviBBX12a to VviBBX12b, VviBBX19a to VviBBX19b, and VviBBX21a to VviBBX21b. However, no tandem duplication was observed according to the   Table S1. c Exon-intron structure of grapevine BBX genes. Exons are represented by pink boxes and black lines connecting two exons represent an intron. The Roman numerals (I-V) indicate the five structural groups foregoing descriptions of Holub [29], and thus only segmental duplication seems to have taken part in the evolution of the grapevine BBX gene family. To gain insight into the evolutionary relationship between VviBBX and AtBBX genes, we analyzed genomic synteny. A total of 26 gene pairs, comprising 17 VviBBXs and 23 AtBBXs, were identified (Fig. 5, Additional file 7: Table S3). Among those, we found nine orthologous pairs, and also identified eight orthologous gene pairs with one grapevine gene corresponding to multiple Arabidopsis genes. We noted that AT2G32310 was not included in the Arabidopsis BBX family, but contained a CCT domain which was also found in VviBBX10. Finally, three orthologous gene pairs where multiple grapevine genes corresponded to a single Arabidopsis gene were found (Additional file 7: Table S3). In brief, these syntenic relationships suggest that about two-thirds of the BBX genes appeared before the divergence of grapevine and Arabidopsis. To investigate potential selective pressure for VviBBX gene duplication events, we calculated the nonsynonymous (Ka) and synonymous (Ks) substitution rates. Between grapevine and Arabidopsis, or grapevine alone, all segmentally duplicated gene pairs showed Ka/Ks ratios of < 1, suggesting that they had evolved primarily under purifying selection. The divergence time of the segmental duplication event was calculated as between~77 and 110 million years ago (Mya) in grapevine alone (Additional file 6: Table S2), and between~102 and 349 Mya, with an average of 178.8 Mya, in grapevine and Arabidopsis (Additional file 7: Table S3).

VviBBXs gene expression profiles in response to E. necator inoculation and hormone treatments
To help identify a possible function of the VviBBX genes in response to powdery mildew, we inoculated healthy plants of the powdery-mildew resistant genotype 'Shang-24' with Erysiphe necator, the causative agent of grapevine powdery mildew, and monitored the expression of the 25 VviBBX genes by semi-quantitative RT-PCR (Real-time polymerase chain reaction). Within 12 h after inoculation, the expression levels of ten genes (VviBBX2, VviBBX8, VviBBX11, VviBBX12b, VviBBX21a, VviBBX22a, VviBBX22b, VviBBX28, VviBBX29a and VviBBX29b) were up-regulated, while those of seven others (VviBBX7, VviBBX9, VviBBX10, VviBBX12a, VviBBX15a, VviBBX15b and VviBBX30) were downregulated (Fig. 6a). Expression of VviBBX6, VviBBX19b, VviBBX21a and VviBBX25 peaked at 12 h postinoculation, whereas VviBBX6 and VviBBX19b decreased sharply at 24 h and remained relatively stabile for the remainder of the measurement period.
Plant hormones, such as abscisic acid (ABA), ethylene (Eth), methyl jasmonate (MeJA), and salicylic acid (SA), play important roles in regulating developmental processes and signaling networks involved in plant responses to biotic and abiotic stresses [30]. In this study, we evaluated the transcriptional response of the 25 BBX genes in plants exposed to these hormones (Fig. 6b). In plants treated with ABA, VviBBX15b, VviBBX21b, VviBBX22a, VviBBX28 and VviBBX29b were upregulated to various degrees, while VviBBX29a was down-regulated. VviBBX22b transcript levels decreased slightly at the first three sampling times after ABA treatment, but then increased. In plants treated with ethylene, most VviBBX genes were down-regulated, while VviBBX19b and VviBBX32 showed decreased expression at least at 12 and 24 h after treatment. After treatment with MeJA, three genes (VviBBX27, VviBBX29a and VviBBX30) were up-regulated, while four genes (VviBBX12a, VviBBX15b, VviBBX21b and VviBBX22b) were down-regulated. Interestingly, two genes, VviBBX2 and VviBBX15a, showed an obvious decrease in expression at the early stages and increase in expression at the later stages. After treatment with SA, eight genes (VviBBX2, VviBBX12b, VviBBX15a, VviBBX15b, VviBBX19a, VviBBX19b, VviBBX21b and VviBBX22b) were up-regulated, and two genes (VviBBX28 and VviBBX30) were down-regulated. Expression of VviBBX21b peaked 1 h after treatment, whereas that of both VviBBX19a and VviBBX19b peaked 3 h after treatment. These transcriptional responses show that the grapevine BBX genes are regulated by multiple phytohormones. To support the results of the semiquantitative RT-PCR analyses, expression of six, randomly-selected VviBBX genes was determined using quantitative RT-PCR (Fig. 6c, d), and results of both analysis approaches were generally consistent.

Expression analysis of VviBBX genes during ovule development
To insight into the potential functions of VviBBX genes during grapevine development, the gene expression atlas contained 54 various organs and tissues at different developmental stages was performed (Additional file 8: Fig.  S3, Additional file 9: Table S4) based on the GEO Data-Sets (GSE36128) [31]. We noted most of the VviBBX genes showed different levels of expression in all organs and tissues. For example, VviBBX6 and VviBBX11 were high expressed relatively ubiquitously. VviBBX10 showed higher expression level in stamen and pollen than other tissues. VviBBX22b exhibited lowest levels of expression in senescencing leaf (Additional file 8: Fig. S3). These results indicated the functional diversification of VviBBX genes during grapevine growth and development. Additionally, previous transcriptome analyses of seed development in grapevine hybrids also suggested that VviBBX transcription factors might be involved in seed size [32]. Hence, we analyzed the expression of the genes in two seeded cultivars ('Red Globe' and 'Kyoho') and two seedless cultivars ('Thompson Seedless' and 'Flame Seedless'), at 27, 30, 33, 36 and 39 days after flowering (DAF) (Fig. 7a) to investigate a potential function for grapevine BBX genes in ovule development. We noticed that most VviBBX genes were expressed differentially according to cultivar, suggesting a potential role in seed development or abortion. Four genes (VviBBX8, VviBBX28, VviBBX29b and VviBBX30) exhibited much higher expression levels in seedless cultivars than seeded cultivars. In particular, VviBBX8 showed a hundred-fold higher expression in seedless cultivars. In contrast, higher expression levels in seeded cultivars relative to seedless cultivars were observed for eight genes (VviBBX5, VviBBX6 VviBBX10, VviBBX11, VviBBX12a, VviBBX15a, VviBBX19b and VviBBX21a), suggesting that they might function in normal development of the ovule. Expression of four randomly selected genes was assessed by quantitative RT-PCR (Fig. 7b), and results were generally consistent with those obtained from semi-quantitative RT-PCR.

Subcellular localization and transcriptional activity of the three BBX proteins
As shown in Table 1, subcellular localization software predicted that approximately 80% BBX proteins were located at the nucleus. To verify the predicted localization pattern of BBX proteins in cells, three genes with one in V. vinifera cultivars 'Thompson Seedless' (VviBBX8) and two from Chinese wild V. quinquangularis accession 'Shang-24' (VquBBX15a and VquBBX29b), which strongly responded to powdery mildew, hormones and/ or ovule development, were cloned for subcellular localization in tobacco leaves (Additional file 10: Fig.  S4). The green fluorescence signals from VviBBX8-GFP, VquBBX15a-GFP and VquBBX29b-GFP fusion proteins were all observed specifically in the nucleus of tobacco leaves, suggesting that the three fusion proteins were localized in the nucleus (Fig. 8a), and these were consistent with the prediction results (Table 1). Using the Yeast Two-Hybrid System (Y2H), the transcriptional activities of BBX proteins were also analysed. Yeast cells transformed with the positive control and negative control all grew well on SD/−Trp (Lacking tryptophan) and SD/ −Trp/X-α-Gal (5-Bromo-4-chloro-3-indolyl-α-Dgalactopyranoside). Positive blue colonies of yeast cells transformed with pGBKT7-VviBBX8 and pGBKT7- VquBBX15a were observed on SD/−Trp/X-α-Gal/AbA (Aureobasidin A), while yeast cells transformed with pGBKT7-VquBBX29b did not survive, suggesting that the VviBBX8 and VquBBX15a had transcriptional activity and VquBBX29b possessed no activation ability in yeast (Fig. 8).

Discussion
In this investigation, 25 grapevine BBX genes were systematically identified and characterized using bioinformatic approaches. Detailed information about these genes, including gene ID, accession numbers, structural group classification, and physiological and biochemical properties of the encoded protein are given in Table 1. The number of BBX family genes was the same as found in pear [33] and was less than in Arabidopsis [4], rice [34], tomato [35], potato [36], apple [37] and maize [38], but higher than eight other plant species [38,39]. We also found that the largest number of BBX members in a plant was 64 from apple, while the least was 19 in millet (Additional file 11: Fig. S5). The remarkable variation in gene number among plants, although potentially attributable to unfinished genome sequence, could reflect species-specific duplications or deletions during evolution. The genes were named VviBBXs (Table 1) based on the nomenclature rules and mapped to specific chromosomes according to their annotated genomic location (Fig. 4).
We constructed logos alignment for the two conserved B-box domains and CCT domain as shown in Fig. 2. The distribution of conserved amino acid residues between the B-box1 domain and B-box2 domain was similar, but not identical. The five Cys residues and two His residues were conserved to a greater extent than the remainder of the amino acid residues in the B-box1 domain. However, the first His residue following the second Cys-X-X-Cys was replaced by Asn in the B-box2 domain (VviBBX9 and VviBBX10). This variation is not seen in Arabidopsis [4], rice [34], tomato [35] or apple [37] and might bring about a new function for these two genes. The 32 Arabidopsis BBX genes were unambiguously classified into five structural groups [4], and BBX members from other plants, including grapevine in this study, can be assigned into five groups (  [37], and 10, 2, 8 and 5 in grapevine (Additional file 11: Fig. S5). We carried out a phylogenetic analysis of BBX protein sequences from grapevine and five other plants. The 25 BBXs in grapevine were classified into five clades, mostly corresponding with the structural groups except for VviBBX27 (Table 1, Fig. 1, Fig. 3a). And some exceptions were also seen in tomato [35], potato [36] and apple [37]. For example, apple MdBBX7 and MdBBX59 with only one B-box domain were presumptively classified to Clade V, but they were phylogenetically in Clade IV, which contained two B-box domains [37].
We analyzed the conserved motifs in the grapevine BBX family genes by MEME, and only 16 motifs were authenticated with E-value < 0.05. The majority of the VviBBX proteins in the same group showed similar motif distribution, but exceptions were found (Fig. 3b). For example, Motif 8, Motif 13 and Motif 14 were found only in Group III, whereas Motif 4, Motif 6 and Motif 7 were mostly specific to Group II. Motif 10 was found only in VviBBX12a and VviBBX12b, the two genes of Group II. These results suggested that BBX proteins might have some unique functions. A previous study reported that exon-intron structure could be used to support phylogenic relationships in a gene family [40]. We found that the number of exons varied from 1 to 5, and most exons showed conserved positions among the 25 VviBBX genes. We also found that three genes (VviBBX28, VviBBX30 and VviBBX32) in Group V, accounting for 12% of the entire family, had no introns, which was the same percentage observed in the pear BBX family [33]. As shown in Fig. 3, VviBBX genes containing a similar exon-intron structure clustered together in the phylogenetic tree. Indeed, we observed four gene pairs (VviBBX9/VviBBX10, VviBBX12a/VviBBX12b, VviBBX19a/VviBBX19b and VviBBX21a/VviBBX21b) possessing the same number of exons with nearly the same exon lengths (Fig. 3c), suggesting that they might have been generated from segmental or tandem duplication, which is supported by our synteny analysis (Fig. 4). However, exon-intron gain or loss, which contributes to expansion and diversification of gene families [41], was observed within the same VviBBX gene group. For example, VviBBX22b had three exons, while the paralogous gene VviBBX22a contained only two exons (Fig.  3c), suggesting that VviBBX22a might have lost one exon during evolution. A similar observation was reported within the five members of the grapevine bHLH IX subfamily, VvbHLH103-VvbHLH107. VvbHLH103 contained five exons, while the other four members contained six, suggesting VvbHLH103 might also lost one of its exons in the course of evolution [42].
Segmental and tandem genomic duplication are important driving forces in gene family expansion [28]. Duplicated genes generally undergo selection, including purifying selection, positive selection or neutral selection, to adapt to the unstable environmental conditions during the long period of evolution. In the present study, four segmentally duplicated gene pairs were found on grapevine Chromosomes 1, 3, 4, 12, 14, 18 and 19, whereas no tandemly duplicated genes were identified (Fig. 4). This result is similar to those from studies of BBX gene family evolution in pear [33], rice [34], tomato [35], maize, sorghum, stiff brome and millet [38]. Moreover, the segmentally duplicated gene pairs, (e.g., VviBBX21a/VviBBX21b) comprised the same group and exhibited similar exon-intron structures and motifs (Fig.  3b, c). A previous study in rice also showed that OsBBX genes likely resulting from segmental duplication were part of the same group [34]. These observations demonstrate that segmental duplications contributed to expansion of the BBX gene family in grapevine. We also identified 26 orthologous BBX gene pairs resulting from segmental duplications between grapevine and Arabidopsis, suggesting that they might have a common ancestor, and therefore, similar functions. For example, expression of AtBBX5 was shown to be strongly induced by ABA, leading to enhanced abiotic stress tolerance [43]. In this study, we found that the orthologous VviBBX5 is expressed to relatively high levels in ABAtreated plants (Fig. 6b). This implicates VviBBX5 in abiotic stress responses in grapevine.
To date, the functions of BBXs have been widely reported, including photoperiodic regulation of flowering, shade avoidance, seedling photomorphogenesis, anthocyanin accumulation and abiotic stress. For example, in chrysanthemum, CmBBX24 was shown to enhance drought and low temperature tolerance by modulating gibberellin biosynthesis [44]. A Group IV BBX protein in Arabidopsis, AtBBX21, functions in ABA signaling [45], and is a positive regulator of photomorphogenesis [17]. Our analyses of grapevine BBX gene expression in response to powdery mildew infection and phytohormone treatments (Fig. 6a, b) showed that these genes are diversely regulated. For instance, VviBBX29b was strongly induced after upon powdery mildew infection, while VviBBX15a was strongly induced and VviBBX15b was slightly repressed in response to MeJA treatment (Fig.  6c, d). These results provide evidence that BBX proteins may participate in response to biotic stress and hormonal signal transduction pathways.
In the present study, the expression levels of most VviBBX genes peaked 27-33 days after full bloom in seedless grapevine varieties (Fig. 7a). Similarly, some previous studies had reported that ovule abortion in stenopermocarpic seedless cultivars takes place at about this time [32,46]. The expression patterns of VviBBX8 and VviBBX29b were significantly up-regulated in nearly all ovule developmental stages in seedless grapevine cultivars, relative to seeded varieties. However, VviBBX6 and VviBBX15a were significantly up-regulated in seeded grapevine cultivars (Fig. 7b). Taken together, the difference in expression levels of VviBBX genes between the seeded and seedless cultivars suggests that they might be involved in ovule abortion, and also have a function in normal seed development.
AtBBX14 and AtBBX15 has been shown to participate in various signaling pathways to uniformize plant growth and development in Arabidopsis [15]. In this study, VquBBX15a, the same clade of AtBBX14 and AtBBX15, has the representative features of the BBX family, which was located in the nucleus (Fig. 8a) and had transcriptional activity (Fig. 8b). These results indicate that VquBBX15a functions as a transcriptional activator of diverse downstream genes. Furthermore, we found that VquBBX29b possessed no activation ability in yeast, maybe it needs some required modification or assistance of other proteins. Definitely, more experiments are needed to confirm this suppose in the future.

Conclusions
In this study, 25 VviBBX genes were identified based on the latest version of grapevine genome annotation. And we also carried out a comprehensive analysis of the BBX gene family, including phylogeny, conserved domain, motif compositions, exon-intron configurations, chromosomal distributions, genomic synteny and expression pattern analysis. The transcription of VviBBX in response to powdery mildew infection, various hormonal treatments and during seed development indicated that VviBBX genes might participate in the corresponding signal transduction pathways and seed abortion. Additionally, the subcellular location and transcriptional activity assays of three BBX members were verified, suggesting that BBX proteins might activate the expression of various downstream genes in nucleus. Taken together, genome-wide analysis of the VviBBX family will provide a fundamental basis for further research on the functions of BBX genes in grapevine.

Analysis of exon-intron structure and conserved motifs
Exon-intron structures of the confirmed BBX genes were determined according to the alignments of their coding sequences and genomic full-length sequences in the Grapevine Genome (12X) database (https://www. genoscope.cns.fr/vitis/) [26]. The diagrams of exon-intron structures were generated using the online program Gene Structure Display Server 2.0 (http://gsds.cbi.pku.edu.cn) [56]. Conserved motifs of all BBX proteins were identified using the online MEME analysis tool (http://meme-suite. org/tools/meme) [57] with the maximum number of motifs being set at 16, and other default parameters. Only motifs with E-value < 0.05 were present. TBtools software was used to draw the map of the conserved motif [58].

Chromosomal localization and synteny analysis
The chromosomal locations of each VviBBX gene were identified according to physical location information from the latest version of grapevine genome annotation. Syntenic blocks for the grapevine BBX genes, as well as between grapevine and Arabidopsis, were identified and analyzed using the MCScanX software [59]. Synteny analysis and chromosomal location diagrams were generated in a globe plot using the program Circos-0.69-6 (http://circos.ca) [60]. The nonsynonymous (Ka) and synonymous (Ks) substitution rates of each gene pairs were calculated using the TBtools software [58]. The Ks values were used to calculate the divergence time with the following formula: T = Ks/2λ (λ = 6.5 × 10 − 9 for Grapevine) [61].

RNA isolation and expression profiling
Total RNA was isolated from grapevine tissues using an EZNA Plant RNA Kit (R6827-01, Omega Bio-tek, USA). RNA quality was assessed by the ratio of A 260 /A 280 and quantity was determined using a NanoDrop 2000 Spectrophotometer (Thermo Fisher Scientific, Wilmington, DE, USA). First-strand cDNAs were synthesized by reverse transcription of 1000 ng total DNA-free RNA using the Prime Script RT reagent Kit (TaKaRa Biotechnology, Dalian, China) following the manufacturer's instructions. The resulting cDNA was diluted six fold for use in semi-quantitative RT-PCR and quantitative RT-PCR experiments.
Oligonucleotide primers for each VviBBX gene were designed using Primer Premier 5.0 software (PREMIER Biosoft International, Palo Alto, CA, USA), and assessed for potential alternative target sequences utilizing the Primer-BLAST online program in the NCBI database (https://www.ncbi.nlm.nih.gov/tools/primer-blast/index. cgi?LINK_LOC=BlastHome) (Additional file 12: Table  S5). Expression profiles of VviBBXs in various organs and tissues based on the microarray data obtained from the NCBI gene expression omnibus (GEO) datasets under the series entry (GSE36128) (https://www.ncbi. nlm.nih.gov/geo/query/acc.cgi?acc=GSE36128) [31]. The grapevine ACTIN1 gene (GenBank accession no. AY680701) and EF1-α gene (GenBank accession no. EC931777) were used as the endogenous control. Semiquantitative RT-PCR and quantitative RT-PCR were carried out as previously described [65]. The results of semi-quantitative RT-PCR were visualized using the software GeneSnap (Version 7.08; SynGene, Cambridge, England) and heat maps were generated using TBtools software [58]. The 2 −ΔΔCT method was used to calculate gene relative expression levels from quantitative RT-PCR amplification [66], and StepOne software (Version 2.3; Applied BioSystems, USA) was used to analyze the relative expression.

Subcellular localization and transcriptional activity of BBX proteins
The subcellular location of BBX proteins was predicted by CELLO (http://cello.life.nctu.edu.tw) [67]. The full-length of coding sequences of three BBX genes were amplified with high fidelity PrimeSTAR ® Max DNA Polymerase (TaKaRa Biotechnology, Dalian, China) from 'Thompson Seedless' (VviBBX8) and 'Shang-24' (VquBBX15a and VquBBX29b). The three grapevine BBX coding regions without the termination codon were inserted into a pCambia2300-GFP vector driven by the CaMV 35S promoter. The resulting constructs were transformed into Agrobacterium tumefaciens GV3101 and infected into the leaves of Nicotiana benthamiana as previously described [68]. The transient expression of the fused proteins was observed by a laser scanning confocal microscope (Olympus FV3000, Japan) after 3 d at room temperature. The green fluorescence and DAPI (4,6-diamidino-2-phenylindole dihydrochloride) were excited with a 488-nm and 405-nm laser line, respectively.
Transcriptional activity assays were performed using the Matchmaker ® Gold Yeast Two-Hybrid System (Clontech, Mountain View, CA, USA) as described in the manual. Full-length BBXs were fused in the pGBKT7 vector containing the DNA-binding region of GAL4, and transferred into the yeast strain Y2H. Meanwhile, pGBKT7-53 co-transformed with pGADT7-T was used as positive control, and pGBKT7-Lam co-transformed with pGADT7-T was used as negative control. The transformants were grown at 30°C for 3-5 days. Transcriptional activation activity was determined by positive blue colonies on the selective solid medium plate SD/ −Trp, supplemented with 40 μg/mL X-α-Gal and 200 ng/mL AbA. Primers used for gene clone, subcellular localization and transcriptional activity were listed in Additional file 13: Table S6.