Genome-Wide Identification and Expression Analysis of GA2ox, GA3ox, and GA20ox Are Related to Gibberellin Oxidase Genes in Grape (Vitis vinifera L.)

Gibberellin (GAs) plays the important role in the regulation of grape developmental and growth processes. The bioinformatics analysis confirmed the differential expression of GA2, GA3, and GA20 gibberellin oxidase genes (VvGA2oxs, VvGA3oxs, and VvGA20oxs) in the grape genome, and laid a theoretical basis for exploring its role in grape. Based on the Arabidopsis GA2oxs, GA3oxs, and GA20oxs genes already reported, the VvGA2oxs, VvGA3oxs, and VvGA20oxs genes in the grape genome were identified using the BLAST software in the grape genome database. Bioinformatics analysis was performed using software such as DNAMAN v.5.0, Clustalx, MapGene2Chrom, MEME, GSDS v.2.0, ExPASy, DNAsp v.5.0, and MEGA v.7.0. Chip expression profiles were generated using grape Affymetrix GeneChip 16K and Grape eFP Browser gene chip data in PLEXdb. The expression of VvGA2oxs, VvGA3oxs, and VvGA20oxs gene families in stress was examined by qRT-PCR (Quantitative real-time-PCR). There are 24 GAoxs genes identified with the grape genome that can be classified into seven subgroups based on a phylogenetic tree, gene structures, and conserved Motifs in our research. The gene family has higher codon preference, while selectivity is negative selection of codon bias and selective stress was analyzed. The expression profiles indicated that the most of VvGAox genes were highly expressed under different time lengths of ABA (Abscisic Acid) treatment, NaCl, PEG and 5 °C. Tissue expression analysis showed that the expression levels of VvGA2oxs and VvGA20oxs in different tissues at different developmental stages of grapes were relatively higher than that of VvGA3oxs. Last but not least, qRT-PCR (Real-time fluorescent quantitative PCR) was used to determine the relative expression of the GAoxs gene family under the treatment of GA3 (gibberellin 3) and uniconazole, which can find that some VvGA2oxs was upregulated under GA3 treatment. Simultaneously, some VvGA3oxs and VvGA20oxs were upregulated under uniconazole treatment. In a nutshell, the GA2ox gene mainly functions to inactivate biologically active GAs, while GA20ox mainly degrades C20 gibberellins, and GA3ox is mainly composed of biologically active GAs. The comprehensive analysis of the three classes of VvGAoxs would provide a basis for understanding the evolution and function of the VvGAox gene family in a grape plant.


Introduction
Gibberellins (GAs) form a group of tetracyclic diterpenes, and some of them are biologically active. The GA biosynthetic pathway has been elucidated, and its key components have been identified [1]. It serves as hormones in higher plants by controlling diverse growth and developmental processes, such Grapevine (Vitis vinifera L.) is part of the broadest cultivated and economically significant crops all over the world. The exogenous pre-bloom application of gibberellin 3 (GA3) to grapevine, which is an economically important crop that has long been an important component of the human diet, and is commonly used to induce seedlessness [31,32], establish early ripening [39], and enhance berry size in seedless cultivars [40][41][42]. With the recent report that grape genetic basis for bunch and berry traits could be linked with gibberellin activity related VvGAI1 gene, which could have a pleiotropic effect and be involved in their genetic regulation, since it is a negative regulator of GAs response [43]. Single nucleotide polymorphisms (SNPs) in the grape genome have been extensively studied, and these findings have led to a better understanding of the genetic diversity of cultivated grapevine [44][45][46]. Although gibberellin oxidase gene has been widely explored [16,30,47,48], a tiny part of GA2ox, GA3ox, and GA20ox gene in grape have been investigated [5,49]. Considering the roles of GA2ox, GA3ox, and GA20ox in response to GA3 during grapevine growth and development, we sought to identify GA2ox, GA3ox, and GA20ox family in grape. In this study, gibberellin oxidase genes were identified and analyzed by bioinformatics method. Using grape abiotic stress and tissue expression data in the GEO (Gene Expression Omnibus) database and screening of the highest expression gene by qRT-PCR (Quantitative real-time-PCR), to explore the function of three gibberellin oxidase genes in the synthesis and degradation of gibberellin, the expression of different tissues in different growth and development stages of grapes, and their response to abiotic stress. This provides a new theoretical basis for grape breeding.

Plant Materials and Treatments
'Pinot Noir' tube seedling was used as material for qRT-PCR. The single shoot stem of the seedlings was attached to a solid GS (modified B5 solid medium) and cultured under Light Emitting Diode (LED) white light (light for 16 h and dark for 8 h) in the incubator for 35 days. Subsequently, the seedings were treated by 5, 10, 15, and 20 mg·L −1 gibberellin (GA3) and 10 mg·L −1 GA3 inhibitor (Uniconazole) for 24 h after cultured for 35 days, and the distilled water was used as a control. All of the materials were collected from grape leaves and frozen in liquid nitrogen and stored at −80 • C for RNA extraction and gene expression analysis.

Identification of Grape Gibberellin Oxidase Genes GA2ox, GA3ox, and GA20ox
All of the sequences were downloaded from four databases: TAIR (Arabidopsis Information Resource) [37], GDR (Apple Information Resource) [50], Rice Genome Annotation Project Database, and Grape Genome Database. Then, a gene of GA oxidase family in Arabidopsis was selected as a seed to blast the three databases. SMART [51] and previous genome annotations are employed to confirm the sequence accuracy and remove genes without the 2OG-Fe (II) oxygenase domain. All of the 24 GA oxidase genes are shown in Table 1, and the information about Arabidopsis, rice, apple, and grape is listed in Table S1.
The gene structures (Exon and intron) were analyzed with GSDS v.2.0. The conserved domain of the protein was examined using the MEME online software. In addition, the maximum number of motifs in the conserved domain was placed at 10.

Phylogenetic Analysis of Systems
The accession numbers of GA2ox, GA3ox, and GA20ox of Arabidopsis thaliana, rice, apple, and grape were aligned using ClustalX v.2.0 (Conway Institute, University College Dublin, Dublin, UK) [50]. A phylogenetic tree was constructed via Molecular Evolutionary Genetics Analysis MEGA 7.0 (Pennsylvania State University, State College, PA, USA) [53] by using the neighbor-joining (NJ) method and adopts the following parameters: the mode adopted "Poisson model," the gap was set to "Complete deletion," and the check parameter was bootstrap = 1000 times, random seed.
The non-synonymous/synonymous (Ka/Ks; ω) value of duplicate gene pairs or triplicate gene groups (between any two genes in one triplicate gene groups) were calculated using DNAsp v.5.0 (University of Barcelona, Barcelona, Spain).

Analysis of the Cis-Acting Element, Subcellular Localization, and Secondary Structure
The promoter sequences of GA2ox, GA3ox, and GA20ox in grape were downloaded from the grape genome database (2000 bp), and the gene promoter elements were predicted and analyzed using the PlantCARE online site [59,60]. The subcellular localization of GA oxidases was analyzed by WoLF PSORT (https://wolfpsort.hgc.jp/). The secondary structure was identified through NPS@:SOPMA secondary structure prediction.

Acquisition and Analysis of Chip Expression Data in Grape
Expression data were retrieved from microarray platforms (Affymetrix GeneChip 16K Vitis vinifera Genome Array, Affymetrix®Inc., Santa Clara, CA, USA) [61], and the selected data were about 'Cabernet Sauvignon' grape under different exogenous ABA treatment conditions (accession numbers: GSE31662 and GSE31664). Abiotic stress data (accession number: GSE31594) were downloaded from the GEO database, and the expression data of GA2ox-, GA3ox-, and GA20ox-related genes were extracted from grapes. Heat maps were drawn using R language. Besides, tissue expression data was obtained from Platforms (GPL13936 NimbleGen 090918 Vitus vinifera exp HX12 [090918_Vitus_exp]) (accession numbers: GSE36128) [62], and the selected data were come from the Grape organs at different stages and the heat map was drawn using TBtools.

RNA Extraction, qRT-PCR and Statistical Analysis
RNA was extracted using a Spectrum Plant Total RNA kit (Sigma St. Louis, MO, USA). Using the extracted total RNA of grape leaves as a template, the reverse strand of complementary DNA (cDNA) was synthesized using the Reverse Transcriptase M-MLV (RNase H-) kit (TaKaRa Biotechnology. Lanzhou, China), and 0.5-2 µg of purified total RNA was reverse transcribed into the first strand cDNA that was used to qRT-PCR. Subsequently, qRT-PCR instrument (Light Cycler 96 Real-Time PCR System, Roche, Basel, Switzerland) was used for the test, and qRT-PCR was performed using TaKaRa SYBR Premix Ex Taq™ II (TaKaRa Biotechnology. Lanzhou, China). Cycling parameters were 95 • C for 30 s, 40 cycles of 95 • C for 5 s, and 60 • C for 30 s. For melting curve analysis, a program including 95 • C for 15 s, followed by a constant increase from 60 • C to 95 • C, was included following the PCR cycles.
GAPDH gene (GenBank accession no. CB973647) was used as the internal reference gene, and the primer sequence was shown in Table S2. The relative expression of the gene was calculated using 2 −∆∆CT method [63]. Samples, which served as cDNA stocks for PCR analysis were stored at −80 • C.
Data quantified from the qRT-PCR of three biological replicates were subjected to two-way ANOVA analysis, followed by Bonferroni's post-test for data comparison. A p-value of less than 0.01 was deemed to represent a significant difference and significant difference analysis using Duncan method.

Results
3.1. Identification of GA2ox, GA3ox, and GA20ox Genes in Grape A total of 24 candidate genes were obtained in the Grape Genome (12X), which contain 11 GA2ox, 6 GA3ox, and 7 GA20ox, respectively. VvGA2ox1-VvGA2ox11, VvGA3ox1-VvGA3ox6, and VvGA20ox1-VvGA20ox7 were appointed on the basis of the order of their gene identification number (ID) (Table 1, Figure S1). All these genes were widely distributed on 13 chromosomes. The largest distribution was located on the 3rd chromosome, and only one gene distributed on the 1st, 2nd, 5th, 7th, 12th, and 16th chromosomes, respectively. It distributes in three genes for the 4th and 19th chromosomes, respectively. Two gene distributions were detected on the 15th and 18th chromosomes, respectively. The CDS coding sequences of GA2ox, GA3ox, and GA20ox in grape ranged from 825 bp (VvGA3ox1) to 2037 bp (VvGA3ox4), amino acid sequence lengths ranged from 275-678. The molecular weight of these three types of gibberellin oxidase with large differences that between 31.27-75.98 kD. The largest difference in molecular weight is the GA3ox class with a difference of 44.71 kD. The molecular weight difference of the GA20ox class is 13.58 kD, and the smallest difference in molecular weight of GA2ox class with a difference of 12.48 kD. Three types of gibberellin oxidase proteins had hydrophilic values ranging from −0.05 to −0.54 and all of them were hydrophilic proteins. Predicted values of the isoelectric points of these three types were among 5.14-8.22, with only one basic protein (GA2ox2), one neutral (GA2ox8), and the remaining proteins were acidic. Furthermore, 72.73% (8) and 66.67% (4) of GA2ox and GA3ox proteins had an instability index of >40, respectively. However, 42.86% (3) of the GA20ox proteins had an instability index of >40, indicating that GA2ox and GA3ox oxidases were more stable than GA20ox oxidase.

Structural Analysis of GA2ox, GA3ox, and GA20ox Genes
The phylogenetic tree was built using GA2ox, GA3ox, and GA20ox protein sequences in grape, and gene structure analysis was performed ( Figure 1). The amino acid sequence comparisons revealed that grape GAoxs had a conserved Fe 2+ 2-oxoglutarate-dependent dioxygenase (2-ODD) domain in each ODD (Supplementary Figure S2). The number of exons of the 24 gibberellin oxidase genes, including VvGA3ox2 and VvGA3ox3, was between 2 and 6. VvGA3ox5 and VvGA3ox6 had only 2 exons, and VvGA3ox4 had 6 exons. The closer the genetic relationship was, the more similar the genetic structure between them would be, such as those between VvGA20ox2 and VvGA20ox6 and between VvGA2ox9 and VvGA2ox10, which contained the same number of exons, and have the same structure of genes. Three types of gibberellin oxidase genes were mostly within 3 kb, and the individual genes VvGA2ox3 and VvGA3ox4 reached about 6 kb, which might be related to their special features.
The phylogenetic tree was constructed using the GA2ox, GA3ox, and GA20ox proteins of Arabidopsis thaliana (At), Oryza sativa (Os), Malus domestica (Md), and Vitis vinifera (Vv) (Figure 2). In line with the evolutionary relationship and according to the Huang et al., 2015 research [50], they were divided into seven subgroups, namely GAox-A, GAox-B, GAox-C, GA20ox, GA3ox, C19 GA2ox and C20 GA2ox subgroups. The evolutionary relationship showed that the distribution of GA2ox, GA3ox, and GA20ox in each species was similar and evenly distributed in the seven subfamilies. We can classify VvGA2ox3, 4, 5 to C20 GA2ox subgroup, according to C20 GA2ox subfamily members of Arabidopsis that containing AtGA2ox7, 8 and OsaGA2ox5. VvGA3ox2, 3 and 6 can be classified in the GA3ox subfamily. VvGA2ox1, 2, 7, can be classified in the C19 GA2ox subfamily. And VvGA20ox1, 2, 4, 5 and 6 were classified in the GA20ox subfamily. In addition, VvGA20ox7, VvGA2ox9 and VvGA2ox10 can be classified to GAox-A. VvGA3ox1, VvGA2ox8 and VvGA2ox11 can be classified to GAox-B. VvGA20ox3, VvGA3ox5, VvGA3ox4 and VvGA2ox6 can be classified to GAox-C. Overall, the number of GA20ox and GA2ox was greater than that of GA3ox, indicating that GA20ox and GA2ox also experienced a dynamic evolutionary route and leading to functional redundancy. The phylogenetic tree was constructed using the GA2ox, GA3ox, and GA20ox proteins of Arabidopsis thaliana (At), Oryza sativa (Os), Malus domestica (Md), and Vitis vinifera (Vv) ( Figure  2). In line with the evolutionary relationship and according to the Huang et al., 2015 research [50], they were divided into seven subgroups, namely GAox-A, GAox-B, GAox-C, GA20ox, GA3ox, C19 GA2ox and C20 GA2ox subgroups. The evolutionary relationship showed that the distribution of GA2ox, GA3ox, and GA20ox in each species was similar and evenly distributed in the seven subfamilies. We can classify VvGA2ox3, 4, 5 to C20 GA2ox subgroup, according to C20 GA2ox subfamily members of Arabidopsis that containing AtGA2ox7, 8 and OsaGA2ox5. VvGA3ox2, 3 and 6 can be classified in the GA3ox subfamily. VvGA2ox1, 2, 7, can be classified in the C19 GA2ox subfamily. And VvGA20ox1, 2, 4, 5 and 6 were classified in the GA20ox subfamily. In addition, VvGA20ox7, VvGA2ox9 and VvGA2ox10 can be classified to GAox-A. VvGA3ox1, VvGA2ox8 and VvGA2ox11 can be classified to GAox-B. VvGA20ox3, VvGA3ox5, VvGA3ox4 and VvGA2ox6 can be classified to GAox-C. Overall, the number of GA20ox and GA2ox was greater than that of GA3ox, indicating that GA20ox and GA2ox also experienced a dynamic evolutionary route and leading to functional redundancy.
The conservative motif analysis of different subgroups of protein sequences was performed using MEME to obtain 10 conserved motifs ( Figure 3) and named motif 1 to motif 10 (Table S3). Motifs 1, 2, and 4 were conserved sequences shared by all of the genes ( Figure S3). Among them, motif 3 is unique to GA2ox and GA3ox. motifs 5, 6, and 8 are unique to GA2ox and GA20ox. motifs 7 and 9 are in GA2ox and GA3ox. motif 10 was only present in GA2ox9, 10, and GA20ox7. Six pairs of genes with homology greater than 90 were VvGA20ox2/VvGA20ox6, VvGA2ox3/VvGA2ox5, VvGA2ox9/VvGA2ox10, VvGA3ox2/VvGA3ox3, VvGA2ox8/VvGA2ox11, and VvGA2ox1/VvGA2ox2. The conserved sequence and protein lengths of each pair of proteins were basic and identical. The conservative motif analysis of different subgroups of protein sequences was performed using MEME to obtain 10 conserved motifs ( Figure 3) and named motif 1 to motif 10 (Table S3). Motifs 1, 2, and 4 were conserved sequences shared by all of the genes ( Figure S3). Among them, motif 3 is unique to GA2ox and GA3ox. motifs 5, 6, and 8 are unique to GA2ox and GA20ox. motifs 7 and 9 are in GA2ox and GA3ox. motif 10 was only present in GA2ox9, 10, and GA20ox7. Six pairs of genes with homology greater than 90 were VvGA20ox2/VvGA20ox6, VvGA2ox3/VvGA2ox5, VvGA2ox9/VvGA2ox10, VvGA3ox2/VvGA3ox3, VvGA2ox8/VvGA2ox11, and VvGA2ox1/VvGA2ox2. The conserved sequence and protein lengths of each pair of proteins were basic and identical.
We identified ten triplicated gene groups in the grape VvGA2ox, VvGA3ox, and VvGA20ox gene family, such as the VvGA3ox2-VvGA3ox3-VvGA3ox6 group. In addition, VvGA3ox2, VvGA3ox3 and VvGA3ox6 were collinear, showing that the triplication may derive from chromosome segmental triplication or a large-scale triplication event. All the triplicated gene groups in the grape VvGA2ox, VvGA3ox, and VvGA20ox family underwent purifying selection (Table S4). Based upon our chromosomal location, gene structure and motif analysis, we found that the genes may be functionally conserved within a triplicated gene group such as the VvGA3ox2-VvGA3ox3-VvGA3ox6, VvGA2ox3-VvGA2ox5-VvGA2ox4 and VvGA20ox4-VvGA20ox5-VvGA20ox1 groups. The motifs of one VvGA2ox, VvGA3ox, and VvGA20ox family member are also similar to the other members in the triplicated same gene group (Figures 1-3).

Codon Usage Bias Analysis
Analysis of codon usage parameters of VvGA2ox, VvGA3ox, and VvGA20ox gene families was listed in Table S5. Among these parameters, CAI, CBI, and NC are usually predicting gene expression levels. Generally, CAI and CBI are positively correlated, with values ranging from 0 to 1 and the closer a value is to 1; the stronger the codon preference, the higher the gene expression level, and the negative correlation to the number of effect codons (NC). NC value is generally from 20 to 61. 20 indicates that the synonymous codon bias is larger, and the closer to 61, the smaller the synonym codon bias. Highly expressed genes have a large degree of codon preference, with large CAI and CBI values and small NC values, low expression genes contain more types of rare codons, so the preference is low, CAI and CBI values are smaller and NC values are higher. Its value will be influenced by the amino acid composition of the gene and the length of the gene. From the data point of view, NC values of the VvGA2ox, VvGA3ox, and VvGA20ox gene families ranged from 45.88-57.27, and the expression level was average. We can find that the CAI and CBI values of VvGA2ox7 are quite high, while the NC values are relatively low, and suspected the expression level is relatively high. It is generally accepted that the two factors of GC amount and GC3s have an important influence on the codon usage of the gene. In these three gibberellin oxidase gene sequences, the amount of 21 gibberellin oxidase gene GC was less than 0.5, indicated that the grape gibberellin oxidase gene had no discernible preference for GC. The GC3s value was greater than 0.5, accounting for 41.7%, indicated that most of the grape gibberellin oxidase genes prefer codons ending in A/T.
A correlation analysis was conducted to further understand the influences of codon bias on gene properties (Table S6). On the basis of the correlation analysis, we can find that FOP and CBI, CAI is an extremely significant correlation, simultaneously, the content of GC and GC3s is a significant correlation, through these consequences, we can infer the gibberellin oxidase gene is mainly derived from the stress of mutation. Gravy values were negative and they were significantly negatively correlated with GC3s (p < 0.01), so gibberellin oxidase proteins were all hydrophobic proteins. The GC (G + C content) in the grape was significantly positively correlated with G3s (the frequency of occurrence of the corresponding base in the third position of the synonymous codon) and GC3s (the G + C content in the third position of the codon) (p < 0.01), which is significantly negatively correlated with T3s and A3s (p < 0.01), FOP (optimal codon usage frequency) and CBI (codon preference index), CAI (codon adaptation).
The indices were all significantly positively correlated (p < 0.01). GC and GC3s were significantly negatively correlated with T3s and A3s (p < 0.01), and GC, GC3s and NC (codon effective numbers) were substantially negative. Correlation (p < 0.05) indicates that base composition has an important influence on codon preference and gene expression level. The level of gene expression and codon usage preference is affected by the synonym codon the effect of the base in the third position. The more codon-preferred (lower NC values) genes, the more preferred they are to use optimal codons (higher FOP values) and higher G + C content, especially at G/C (larger GC3s).
RSCU is the relationship between the actual number of codons corresponding to an amino acid and the number of theoretical applications. It can intuitively reflect the degree of preference of codon usage, regardless of gene length and amino acid abundance. When the actual number of applications is the same as the number of theoretical applications, RSCU = 1, there is no preference for codons; when RSCU > 1, it indicates that the codon appears more frequently than other synonymous codons. Preferred codons; when RSCU < 1, the relative frequency of codons is low, which is a codon that is less frequently used by genes. By comparison (Table S7), in VvGA2ox, there are 23 the codon of RSCU > 1, 20 VvGA3ox, RSCU > 1, 21 VvGA20ox, RSCU > 1, Among them, the RSCU values of 17 codons in the grape GA2ox, GA3ox, and GA20ox were all greater than 1, indicating that these codons are preferred codons for the gene family. These codons are the most commonly used codons for the gibberellin oxidase gene, the optimal codon.

Subcellular Localization and Secondary Structure Analysis
The results of subcellular localization (Table S8) showed that the gibberellin oxidase gene family was mainly expressed in the chloroplast, the cytoplasm, and the nucleus; indicating that the gene family mainly existed in organs with strong photosynthesis and respiratory metabolism. Among them, VvGA2ox1, 4, VvGA20ox1, and 5 were not noted in the cytoplasm. VvGA2ox3, 9, VvGA20ox3, and 7 were not expressed in the chloroplasts, and the remaining genes were written in the cytoplasm. VvGA2ox5 and 7 were not expressed in the nucleus, but the other genes were written in the nucleus. Only four genes VvGA3ox2, 4, VvGA20ox5, 6 expressing in the mitochondria, and other genes were not noted. Only three genes, namely, VvGA2ox10, VvGA3ox4, and VvGA20ox5, were associated with the peroxisome. VvGA20ox1 and 2 were written in vacuoles. VvGA2ox1 and 2 were expressed in the Golgi. VvGA2ox10, 11, VvGA3ox1, 5, VvGA20ox2, 3, 4, and 6 were noted in the cytoskeleton, and these genes might be associated with cell wall formation.
The secondary structure of the grape gibberellin oxidase gene family encodes proteins mainly consisting of α-helix, β-turn, and irregular curl. Secondary structure analysis indicated that the 24 proteins encoded by the Gibberellin oxidase gene family were mainly α-helix and irregular coils (Table S9).

Gene Chip Expression Profile Analysis
These three kinds of genes were removed by downloading the gene expression data of ABA and abiotic stress from the GEO database to predict the expression of gibberellin oxidase gene in response to ABA and abiotic stress. The data corresponding to the grapes were tested by heat map (Figures 4 and 5). The expression of gibberellin oxidase gene in response to different periods of ABA stress significantly differed (Figure 4). With the prolongation of stress time, the genes with continuously upregulated expression levels were VvGA2ox9, 10, and VvGA20ox7. VvGA2ox1, 8,9,10, and 11 were continuously upregulated under ABA stress for 3, 10, and 14 d. Only the expression levels of VvGA2ox9, 10, and VvGA20ox7 were upregulated under ABA stress for 28 d, indicating that ABA was closely related to VvGA2ox9, 10, and VvGA20ox7.  In the graph, 3d-Un represent untreated the grape fruit just as control, 3d-ABA represent have been treated grape fruit with ABA, Simultaneously, 10d-Un, 14d-Un, 28d-Un, representing untreated grape fruit at 10, 14, and 28 days, otherwise, 10d-ABA, 14d-ABA, 28d-ABA, representing have been treated the grape fruit at 10, 14, and 28 days. Red or green shading represented the upregulated or downregulated expression level, respectively. The scale denoted the relative expression level. VvGAox genes subjected to ABA treatments for 3, 10, 14, and 28 days in grape. In the graph, 3d-Un represent untreated the grape fruit just as control, 3d-ABA represent have been treated grape fruit with ABA, Simultaneously, 10d-Un, 14d-Un, 28d-Un, representing untreated grape fruit at 10, 14, and 28 days, otherwise, 10d-ABA, 14d-ABA, 28d-ABA, representing have been treated the grape fruit at 10, 14, and 28 days. Red or green shading represented the upregulated or downregulated expression level, respectively. The scale denoted the relative expression level.
In response to different abiotic stresses, the expression of gibberellin oxidase gene was further significantly different ( Figure 5). VvGA2ox2, 3, 5, 6, VvGA20ox2, 4, 6, 7, VvGA3ox1, 4, and 5 were upregulated in response to salt stress. The expression of VvGA2ox7, VvGA20ox2, and VvGA3ox4 was stable. VvGA20ox1 also maintained a low expression level. The expression levels of VvGA2ox2, 3, 4, 5, 6, 7, 9, 10, VvGA20ox2, 3, 5, 6, 7, VvGA3ox1, 4, and 5 were significantly upregulated under PEG stress for 24 h. The expression levels of VvGA2ox2, 3, 4, 5, VvGA20ox3, and VvGA3ox5 were upregulated. The expression levels of VvGA2ox1, 8, 11, VvGA20ox1, 6, 7, and VvGA3ox6 were downregulated under 24 h of PEG stress. The expression levels of VvGA2ox1, 4, 6, 8, 11, VvGA20ox1, 3, 5, VvGA3ox1, 4, and 6 were upregulated in response to low-temperature stress at 5 • C. The expression levels of VvGA2ox1, 8, 11, VvGA20ox1, and 3 under low-temperature stress at 5 • C for 8 h were upregulated. These findings suggested that these upregulated genes had a certain role in the cold resistance of grape. Figure 4. Hierarchical clustering of the expression profiles of 23 VvGAoxgenes subjected to ABA treatments for 3, 10, 14, and 28 days in grape. In the graph, 3d-Un represent untreated the grape fruit just as control, 3d-ABA represent have been treated grape fruit with ABA, Simultaneously, 10d-Un, 14d-Un, 28d-Un, representing untreated grape fruit at 10, 14, and 28 days, otherwise, 10d-ABA, 14d-ABA, 28d-ABA, representing have been treated the grape fruit at 10, 14, and 28 days. Red or green shading represented the upregulated or downregulated expression level, respectively. The scale denoted the relative expression level.  Fourteen grape gibberellin oxidase genes were found in the Grape eFP Browser database, and the expression of these genes in different tissues of different growth stages of wine grapes was visualized using the TBtool ( Figure 6). Overall, the expression levels of VvGA2ox in different tissues were significantly higher than those of VvGA3ox and VvGA20ox. From a local perspective, VvGA2ox1 was expressed in other tissues except for the low expression levels in roots and seeds.

Gene Cis-Element Analysis of GA2ox, GA3ox, and GA20ox
PlantCARE was used to predict the upstream promoter elements of the gene and analyze the number of cis-acting elements related to hormones, meristems, adversity, and circadian rhythm to further explore the functions of GA2ox, GA3ox, and GA20ox (Figure 7), and the main cis-element selected in the Table S10. On the promoters of these three types of gibberellin oxidase, stress-related action elements were ubiquitous and relatively abundant. Hormone-related action elements, such as methyl jasmonate, gibberellin, and abscisic acid, were compared. In general, the amounts of auxin and salicylic acid were relatively small, and only six genes, namely, VvGA2ox2, VvGA20ox1, 4, 5, VvGA3ox1, and 6, were associated with meristematic tissues, and four genes, namely, VvGA2ox2, 6, VvGA3ox6, and VvGA20ox1, were related to circadian rhythm. Although all of the three genes are owned by the gibberellin oxidase gene, some of the gibberellin oxidase gene promoters contain gibberellin-related elements, such as VvGA2ox2, 3, 5, 9, VvGA3ox6, VvGA20ox1, and 7. On the promoters of GA20ox genes, only VvGA20ox1 and 7 did not contain gibberellin-related action elements, and others contained gibberellin-related action elements. Many genes contain abiotic stress elements, such as low temperature, drought, saline and other stress response components, while six genes do not get this cis-acting element. Specifically, the promoters of different genes contained the same species or even a number of similar or identical elements, and the similarity of their functions should need to be further verified.
Genes 2019, 10, 680 13 of 20 action elements were ubiquitous and relatively abundant. Hormone-related action elements, such as methyl jasmonate, gibberellin, and abscisic acid, were compared. In general, the amounts of auxin and salicylic acid were relatively small, and only six genes, namely, VvGA2ox2, VvGA20ox1, 4, 5, VvGA3ox1, and 6, were associated with meristematic tissues, and four genes, namely, VvGA2ox2, 6, VvGA3ox6, and VvGA20ox1, were related to circadian rhythm. Although all of the three genes are owned by the gibberellin oxidase gene, some of the gibberellin oxidase gene promoters contain gibberellin-related elements, such as VvGA2ox2, 3,5,9,VvGA3ox6, VvGA20ox1, and 7. On the promoters of GA20ox genes, only VvGA20ox1 and 7 did not contain gibberellin-related action elements, and others contained gibberellin-related action elements. Many genes contain abiotic stress elements, such as low temperature, drought, saline and other stress response components, while six genes do not get this cis-acting element. Specifically, the promoters of different genes contained the same species or even a number of similar or identical elements, and the similarity of their functions should need to be further verified.
Under different concentrations of gibberellin treatment, the expression levels of GA2ox were significantly upregulated, and the expression levels of GA3ox and GA20ox were downregulated. The expression levels of GA3ox and GA20ox were obviously upregulated when the samples were treated

Discussion
Gibberellins are associated with several aspects of growth and development, including seed maturation, stem elongation, floral induction, germination, pollen germination, and pollen tube growth [2,12,13,33,64]. Nevertheless, few studies have been conducted on GA2ox, GA3ox, and GA20ox in response to hormonal and abiotic stresses [34,38]. Therefore, we sought to study the characteristics of this gene family and the expression of gibberellin oxidase genes under gibberellin and uniconazole treatment was quantitatively analyzed in qRT-PCR to verify the mechanism of Under different concentrations of gibberellin treatment, the expression levels of GA2ox were significantly upregulated, and the expression levels of GA3ox and GA20ox were downregulated. The expression levels of GA3ox and GA20ox were obviously upregulated when the samples were treated with 10 mg·L −1 uniconazole. After 5 mg·L −1 gibberellin 3 (GA3) treatment, the upregulated genes were VvGA2ox1, 3,6,7,9,10,11,VvG3ox4, and VvGA20ox7. Among them, VvGA2ox7 was the highest, which relative expression was reached to 46.66 folds to compare with the control. Followed by the expression of VvGA2ox1 and 11, whose expression levels were 4.34 and 4.69 time that of the control, respectively. Under the treatments of the 10 mg·L −1 GA3, VvGA2ox5, 8, VvGA3ox1, VvGA20ox3, 5, and 6 were upregulated. The highest expression was observed in VvGA20ox5, which was 3.45 times more than control. Under the treatments of the 15 mg·L −1 GA3 was administered, the expression levels of VvGA2ox1, 11, VvGA3ox4, and 5 were upregulated. The gene with the highest expression was VvGA2ox11, which was 4.28 time of the control. After treatment with 20 mg·L −1 GA3 was administered, VvGA2ox1, 7, 11, VvGA3ox5, and VvGA20ox5 were 0.90, 1.78, 1.69, 1.13, and 1.79 time of the control, respectively. After treatment with 10 mg·L −1 the gibberellin inhibitor uniconazole was given, the upregulated genes were VvGA2ox3, 6, 10, VvGA3ox2, 4, 5, VvGA20ox1, 2, 3, 4, 6, and 7. Among them, VvGA20ox7 expression level was 2.58 times of the control.

Discussion
Gibberellins are associated with several aspects of growth and development, including seed maturation, stem elongation, floral induction, germination, pollen germination, and pollen tube growth [2,12,13,33,64]. Nevertheless, few studies have been conducted on GA2ox, GA3ox, and GA20ox in response to hormonal and abiotic stresses [34,38]. Therefore, we sought to study the characteristics of this gene family and the expression of gibberellin oxidase genes under gibberellin and uniconazole treatment was quantitatively analyzed in qRT-PCR to verify the mechanism of these genes in grape and analysis the expression of gene chip expression profile is related to ABA, abiotic stress and different tissues of grape.
In this study, we have identified and analyzed of GA2ox, GA3ox, and GA20ox gene. The effects of these three gibberellin oxidase genes on gibberellin synthesis has been explored and the role of the gene family in grapes has been examined. In this experiment, 24 gibberellin oxidase genes from the grape genome have been identified by bioinformatics analysis, but the number is different from other plants, such as 16 members in Arabidopsis thaliana [38], 21 members in rice [65,66], 24 members in soybean [37], and 18 members in cucumber [67]. The gibberellin oxidase family is large, which may be linked to genome size. VvGA2ox, VvGA3ox, and VvGA20ox were widely distributed on 13 chromosomes except chromosomes 6,8,11,13,14, and 17. All of them were allocated, and each chromosome had one to four genes possibly because GA was involved in various life activities of plants (seed germination, stem elongation, fruit development, etc.). Specific motifs in amino acid sequences are vital regions related to function. As subfamilies from 2OG-Fe (II) oxygenase superfamily, GA oxidases include the C-terminal of prolyl 4-hydroxylase α subunit, and the parts of the active site comprise an alpha2 beta2 complex with an α subunit [49,68,69]. The analysis of gene structure and conserved motifs revealed similar genetic structures with comparable genetic evolutionary relationships. The conservative motif analysis of grapes found that motif 3 was unique to GA2ox and GA3ox. Motif 6 and motif 8 were unique to GA2ox and GA20ox. Similar reports have also described the conserved motifs in Arabidopsis and cucumber species [37]. Similarly, different types of genes contained specific motifs possibly because of different types of gibberellin oxidases. The precise relationship should be further studied.
The results of the promoter cis-acting element analysis showed that GA2ox, GA3ox, and GA20ox contained some cis-acting elements related to hormones, meristems, stress, and circadian rhythm regulation. These genes were also hit by GA regulation. Jasmonic acid, IAA, SA, ABA, and stress regulation are consistent with previous studies [72]. These three genes belonged to the gibberellin oxidase gene, but some genes, such as VvGA2ox2, 3, 5, 9, VvGA3ox6, VvGA20ox1, and 7, and other gene promoters do not contain gibberellin-related elements, possibly regulation of endogenous GA content changes by other factors on the GA signaling pathway to regulate GA levels in plants. In total, 24 genes had hormone-, meristem-, stress-, and circadian rhythm-related elements on the upstream promoter. The expression data of these three genes under ABA and abiotic stress were downloaded and analyzed by the GEO database. The results showed that these three genes could respond to different duration of ABA and various abiotic stress conditions [36]. Besides, the expression analysis of different tissues in different developmental stages of grape showed that GA2ox7 was mainly expressed in seeds and fruits in C19 GA2ox subfamily. The expression of GA2ox1, 2 was basically the same, mainly expressed in fruits, buds, stamens, flowers, core skin and other tissues. GA20ox7 was highly expressed in roots, stems, leaves, flowers, buds, fruits and other tissues in the GA20ox subfamily. Compared with GA2ox and GA20ox, GA3ox is mainly expressed in stems, buds, inflorescences and pollen. We combined the codon preference analysis, grape gene chip expression and fluorescence quantitative analysis found that the three analysis results have similarity, that is, the expression level of VvGA2ox7 is relatively high.
The genes of the gibberellin oxidase gene in grape were treated with different concentrations of GA and uniconazole by qRT-PCR to investigate the relationship between GA2ox, GA3ox, and GA20ox in grapes, GA, and uniconazole. The results showed that the expression of GA2ox was upregulated in these genes after GA3 treatment, because the GA2-oxidase is a key enzyme in the degradation process of GA, which can inactivate the biologically active GAs and their precursors and other intermediates in plants, thus maintaining biologically active GAs in plants and balance between intermediates [15]. A previous analysis indicated that feedback regulation controls the concentrations of active GAs in higher plants [13]. In most plants, GA20ox and GA3ox, whose products oxidize the penultimate and final steps, respectively, and the production of bioactive GAs (GA1 and GA4) are downregulated by applied exogenous GA [13]. In contrast, the genes encoding GA2ox, which convert active GAs to inactive catabolites, are upregulated by GA treatment [16]. Our findings are similar to previous studies, our results in accordance with others consequences that indicated the GA2ox gene mainly functions to inactivate biologically active GAs, while GA20ox mainly degrades C20 gibberellins, and GA3ox is mainly composed of biologically active GAs [5,13]. In a nutshell, our study provides new insights into the functional divergence of the GAox family in grape and implicates the potential roles of positive selection in the evolution of GAox genes, which deserves further investigation.

Conclusions
Twenty-four gibberellin oxidase genes, including GA2ox, GA3ox, and GA20ox, were found in grape, which can be subdivided into seven subfamilies. GA2ox, GA3ox, and GA20ox were associated with abiotic stress expression and response to the expression of various exogenous hormones, they have different expressions in different organizations. The GA2ox gene mainly functions to inactivate biologically active GAs, while GA20ox mainly degrades C20 gibberellins, and GA3ox is mainly composed of biologically active GAs. Our findings provided a basis for conducting in-depth mechanistic studies on the distinct biological characteristics and adaptability of the grape in harsh environments.
Supplementary Materials: The following are available online at http://www.mdpi.com/2073-4425/10/9/680/s1, Figure S1. A map of the distribution of genes on chromosomes of GA oxidases in grape. Figure S2. Alignment of GA oxidases in grape. Figure S3. Motif sequences of GA oxidase in grape. Table S1. GA oxidase genes in Arabidopsis, Rice, Apples and Grapes. Table S2. Primers used for qRT-PCR of GA2ox, GA3ox and GA20ox gene family in grape. Table S3. Motif sequences of GA oxidase in grape. Table S4. Selection pressure of each grape VvGA2ox, VvGA3ox, and VvGA20ox duplication gene pairs in triplicate gene-group. Table S5. Codon usage characteristics of three gibberellin oxidase gene families in grape. Table S6. Correlation coefficient between the parameters of codon usage of three gibberellin oxidase gene families in grape. Table S7. Frequency of use of three grape gibberellin oxidase gene synonymous codons. Table S8. Subcellular location prediction of VvGAoxs gene in Vitis vinifera. Table S9. The secondary structure of VvGAoxs protein sequence in Vitis vinifera. Table S10. Number of cis-acting elements in GA2ox, GA3ox and GA20ox promoter of grape. Table S11. qRT-PCR data analysis.

Conflicts of Interest:
The authors declare no conflict of interest.