Identification and expression analysis of xyloglucan endotransglucosylase/hydrolase (XTH) family in grapevine (Vitis vinifera L.)

Identification and biochemical analysis of XTHs in grapevine

The sequence annotations of the whole genome and the gene GFF3 file were downloaded by using CRIBI v2.1 (https://urgi.versailles.inra.fr/Species/Vitis/Annotations) (Canaguier et al., 2017). We also downloaded hidden Markov models (PF00722 and PF06958) of the XTH domain from the Pfam database (http://pfam.xfam.org) and obtained the candidate gene sequence numbers of the grapevine XTH family with HMMer software (Potter et al., 2018). To avoid duplication and the inclusion of sequences without XTH family domain characteristics, sequences without the XTH domain and sequences showing alternative splicing were removed. The EMBL-EBI online tool (http://pfam.xfam.org/search/sequence) (Gaia et al., 2021) was used to further analyze secondary structure domains, and the sequences without typical XTH domains were removed.

The relative molecular weight (MW), hydrophilicity (GRAVY), and isoelectric point (pI) of these VvXTHs were predicted and analyzed using ExPASy (https://www.expasy.org/) (Duvaud et al., 2021). Single peptide (SP) prediction was performed on the SignalP v4.1 server (https://services.healthtech.dtu.dk/service.php?SignalP-5.0).

Gene structures were analyzed with Gene Structure Display Server software (http://gsds.gao-lab.org/) (Hu et al., 2015). Conserved motifs in VvXTHs were statistically identified with the online Multiple EM for Motif Elicitation (MEME) software (https://meme-suite.org/meme/tools/meme) (Bailey et al., 2009), and TBtools was then used for the clustering and visualization of VvXTHs in grapevine. Multiple protein sequence alignments were performed with ClustalX software and the Espript 3.0 online program (https://espript.ibcp.fr/ESPript/ESPript/) (Larkin et al., 2007).

Phylogenetic analysis of VvXTHs in grapevine

To investigate the phylogenetic relationship of VvXTHs, the 34 VvXTH protein sequences from grapevine and the 33 AtXTH protein sequences from Arabidopsis were used for multiple sequence alignment by using the Clustal W program within MEGA 11.0 software (Sudhir et al., 2018). The phylogenetic tree was built using the neighbor-joining (NJ) method with 1,000 bootstrap replications and the p-distance model and was then validated by the maximum likelihood method. To better visualize the phylogenetic tree, the final tree diagram file (*.nwk) was uploaded from MEGA to Figtree and EVOLVIEW online software (http://www.evolgenius.info/evolview/) (Balakrishnan et al., 2019).

The Grape Genome Browser (12X) (http://www.genoscope.cns.fr/externe/GenomeBrowser/Vitis/) provided chromosomal location data for all VvXTHs. We used TBtools to identify and illustrate the distribution of genes on chromosomes. MCScanX with the default parameters was applied to identify gene duplication events. The CIRCOS program (https://github.com/CJ-Chen/TBtools) was used to analyze syntenic relationships among VvXTHs. VvXTHs falling within the identified collinear blocks were regarded as segmental events, and any two genes separated by a distance of less than 100 KB whose similarity exceeded 75% were considered tandem duplications. To visualize the synteny relationships of orthologous XTHs derived from grapes and Arabidopsis, Dual Synteny Plotter software (https://github.com/CJ-Chen/TBtools) was applied to construct a syntenic analysis map (Xie et al., 2018). The Arabidopsis sequences were obtained from The Arabidopsis Information Resource (TAIR) database (https://www.arabidopsis.org/) (Han et al., 2013). TBtools software was used to calculate the nonsynonymous (Ka) and synonymous (Ks) substitution rates and Ka/Ks ratio of each gene pair. Divergence times were calculated as follows: T = Ks/2 λ (λ = 6.5 × 10⁻⁹ for grapevine) (Li et al., 2019a).

Cis-element analysis of XTHs in grapevine

The sequences within 1,500 base pairs (bp) upstream of the starting codon of the VvXTHs were obtained from Ensembl Plants (http://plants.ensembl.org/index.html) as the promoter regions (Dan et al., 2017). The cis-elements were predicted with PlantCARE Web Tools (http://bioinformatics.psb.ugent.be/webtools/plantcare/html/) (Magali et al., 2002) and New PLACE Web Tools (https://www.dna.affrc.go.jp/PLACE/?action=newplace). TBtools was used to draw heatmaps and build clustering trees.

Gene expression analysis of XTHs in different grapevine organs and tissues

To understand the spatial and temporal expression patterns of VvXTHs during development, a high-throughput microarray data, from a gene expression atlas generated from different organs/tissues at different developmental stages (Marianna et al., 2012), was employed for further analysis. According to the gene ID, the expression profiles of VvXTHs was extracted from the GSE36128 data set, and we then normalized the average expression value of each gene in 54 samples (including green and woody tissues and organs at different developmental stages as well as specialized tissues such as pollen and senescent leaves). TBtools was used to draw heatmaps and build clustering trees.

To verify the reliability of the results obtained from the GSE36128 data set, the organ-specific expression patterns were examined with quantitative real-time RCR (qRT-PCR) using the five different organs (tendril, root, stem, leaf and flower) from 5-year-old trees of grapevine “Crimson” growing at the experiment station of Shandong Agricultural University (Taian, Shandong, China).

RNA extraction and expression analysis of VvXTHs

The tissue culture seedlings of Vitis vinifera cv “Crimson” seedless were grown on 1/2 Murashige and Skoog (MS) solid medium with 0.2 mM indole-3-butytric acid (IBA) under a 16-h-light/8-h dark cycle at 24 ± 1 °C for 6 weeks. Six-week-old seedlings, which transcription level changes more pronounced, were transferred to liquid medium containing 200 mM NaCl or 200 mM mannitol for salt and drought stress treatments, respectively. The treated seedlings were extracted and separated into leaves and roots for 0, 3, 6, 9, 12, and 24 h upon treatment, immediately frozen in liquid nitrogen and stored at −80 °C for RNA extraction. For each sample, three biological replicates were collected.

Total RNA was extracted from the samples treated with NaCl and mannitol using a HiPure HP Plant RNA Mini Kit (Magen, Guangzhou, China) based on the supplier’s instructions. Subsequently, first-strand cDNA was synthesized from total RNA with the PrimeScript™ RT reagent kit with gDNA Eraser (Vazyme Biotech Co., Nanjing, China). qRT-PCR was performed using a SYBR^® PrimeScript™ RT-PCR Kit (TaKaRa, Dalian, China) according to the supplier’s instructions with a CFX96TM Real-Time PCR Detection System. Gene expression levels were normalized against the average expression of the internal reference gene, and the baseline and Ct (threshold cycles) value were automatically determined by the CFX Manager software program. The relative expression levels of VvXTHs were calculated using the 2^−ΔΔCt comparative Ct method. The internal reference gene used in this study was Vvβ-actin7 (XM_034827164), which has been proved to be a most stable gene for normalization by comparison with other reference gene (Vvβ-actin101: XM_002265440) (Fig. S1). All experiments were performed with three biological replicates, and all the primers used in this study are listed in Table S1. To visualizing the relative difference, the expression level of 0 h treatments for salt and drought stresses and tendril for plant tissue specificity was set as 1, respectively. Then, TBtools was used to draw a heatmap for visualization.

Identification and analysis of VvXTHs in grapevine

Forty-two sequences were identified by searching for two domains (Pfam: PF00722 and PF06955) with the HMMer program. We deleted six alternative splicing sequences and two sequences without typical XTH domains. As a result, we finally identified 34 VvXTHs. Described by previous studies (Cao et al., 2016; Fu, Liu & Wu, 2019; Li et al., 2018; Wan et al., 2014), we named these genes according to their chromosomal locations and named them VvXTH1-VvXTH34.

The analysis of the physical and biochemical data of the 34 VvXTHs, including their amino acids (AAs), MWs, SPs, pIs, total average hydrophilicity (GRAVY) and subcellular localization, revealed that they contained 251~369 AAs. The MW ranged from 28.5 to 41.7 kDa, while the pI ranged from 4.61 to 9.45. All XTHs exhibited hydrophilicity. Subcellular location prediction results showed that most of the genes are localized in the plasma membrane (29), while a few were localized extracellularly (5), including VvXTH10 and VvXTH12 in group IIIA and VvXTH2, VvXTH32, and VvXTH33. The majority of the proteins (80%) contained signal peptides, which were approximately 25-AA long (Table 1).

Phylogenetic analysis and classification of VvXTHs

To investigate the evolutionary relationships and functional associations of VvXTHs with AtXTHs, we built a phylogenetic tree utilizing the protein sequences of XTHs from Vitis vinifera and Arabidopsis (Fig. 1). The VvXTHs were grouped according to the previous grouping method applied for AtXTHs and the evolutionary relationship between grapes and Arabidopsis. The results of the phylogenetic analysis indicated that the 34 VvXTHs could be divided into three groups, including 27 VvXTHs in group I/II, two in group IIIA, and five in group IIIB. In addition, one XTH protein (VvXTH11) was classified into the original ancestral group. Group I/II contained most of the members, and substantial similarity could be observed between some members of the group. The termini of the phylogenetic tree branch showed a total of 22 sister pairs, eight of which were orthologous pairs between Arabidopsis and grapevine, and six were grape homolog gene pairs. This analysis revealed that the number of VvXTHs was slightly expanded in comparison to the number of XTHs in Arabidopsis.

Thirty-four VvXTHs were unevenly distributed on 13 chromosomes. In particular, Chr.11 contained the largest number of VvXTHs (15), whereas other chromosomes contained considerably fewer genes. For example, a total of 4, 3, and 2 genes were located on Chr.5, Chr.10 and Chr.1, respectively. In addition, Chr.2, Chr.3, Chr.6, Chr.7, Chr.8, Chr.12, Chr.15, Chr.16, and Chr.17 each contained only one gene (Fig. 2A). Therefore, it can be inferred that there should be no observable association between the number of XTHs and the length of chromosomes. Furthermore, the genes located on Chr.11 and Chr.5 were closely clustered together. According to the chromosome location and genome annotation information, a total of 81 tandem duplicate gene pairs were obtained (Fig. 2A). To determine the relationships among VvXTH members, we performed a collinearity analysis and found no VvXTH within the identified collinear blocks, which indicated that segmental duplication is not involved in VvXTH expansion (Fig. 2B). These results prove that the expansion of VvXTHs, especially group I/II gene members, was driven by tandem duplication. We also traced the duplication time of VvXTHs by analyzing their Ka, Ks and Ka/Ks ratio. The Ka/Ks ratios of all VvXTHs were less than 1, ranging from 0.07 to 0.28. The duplication times of all VvXTHs were also calculated. The duplication times ranged from 2.91–78.39 Mya (million years ago) (Table S2). To further evaluate the evolution and development of the VvXTH family, we constructed a comparison diagram of grapes and Arabidopsis. Eight VvXTHs were shown to be synonymous with XTHs of Arabidopsis. Among these genes, VvXTH10 is collinear with AtXTH31 and AtXTH32 in Arabidopsis, while VvXTH1 is collinear with AtXTH27 and AtXTH28 in Arabidopsis (Fig. 2C). It is speculated that there may be functional redundancy among these genes, which implies that they have important roles in evolutionary progress.

Gene structural and multiple sequence alignment analysis of VvXTHs

Gene structural analysis revealed that the closely related genes within this subfamily are characterized by a similar structure, which can be further verified by the results of phylogenetic analysis (Fig. 3A). With the exception of VvXTH32, which lacks any introns, all other VvXTH members contain 2~4 different introns. In particular, group I/II contains a large number of members, and most members have two introns. The gene sister pairs, including VvXTH23/26, VvXTH25/28, VvXTH14/15, and VvXTH8/9 at the terminal branch of the evolutionary tree, have highly similar exon/intron structures. In addition, compared with the adjacent gene VvXTH27, VvXTH24 has lost an exon and exhibits different intron and exon lengths. The members of group IIIA all have three introns and four exons and present high structural similarity. The members of group IIIB have developed different numbers and lengths of intron/exon structures during the long evolutionary process (Fig. 3C). In general, most VvXTHs present the same intron/exon structure pattern and remain conserved during evolution, which is consistent with the results obtained in other plants.

Based on the results of MEME motif analysis, motif three and motif four are highly conserved in all VvXTHs. Motif three is a characteristic domain that catalyzes enzymatic contact reactions, which denoted as (D/N)E(I/L/A/V/F)(D/T)(F/I)E(F/I/L)LG (Figs. 3B and 3D). Among these, the first glutamate residue (E) indicates the catalytic nucleophile that initiates the enzymatic reaction, and the second E residue functions represents a base to activate the entrant substrate. In addition, all members except for VvXTH30 contain motif one; moreover, members of the same group share a similar motif composition (Fig. 3B). For instance, motif two only exists in group I/II, and motif eight only exists in group IIIA and IIIB. Genes in the same clade, especially those that are closely related, such as (1) VvXTH23, VvXTH26, VvXTH25, and VvXTH28; (2) VvXTH6, VvXTH7, VvXTH8, and VvXTH9; and (3) VvXTH10 and VvXTH12, can share much more similar motif structures (Fig. 3B). In addition, motif seven, motif nine, and motif 10 only exist in group I/II, and most of the group members contain the motifs mentioned above (Fig. 3B). The members of group IIIA contain five motifs with the same distribution. Group IIIB members contain 4–6 motifs; while they all share four identical motifs, only VvXTH4 and VvXTH32 have motif six, and only VvXTH33 does not have motif five. The results of multiple sequence alignment also confirmed that the conserved domain active site is present in all VvXTHs. Moreover, with the exception of VvXTH2 (IIIB), VvXTH10 (IIIA), and VvXTH12 (IIIA), potential N-glycosylation residues are located near the active site in the 31 other VvXTHs (Fig. 3E). Conserved domain predictions suggested that members of the same subfamily may have similar structures and may be involved in similar functions. Thus, we need to focus on distinctive members that may present surprising functions that remain to be discovered.

Organ-specific expression pattern analysis of VvXTHs

Through the expression profile (GSE36128) analysis of the GEO data set, we obtained the specific expression patterns of VvXTH members in different organs and developmental periods of grapevine to predict the functions of VvXTHs in growth and development (Fig. 4). According to the results of cluster analysis, the VvXTH families were classified into four groups (A–D): group A contains seven genes with high expression levels in berry peels, skins, shafts, and tendrils; group B includes four members with high expression levels only in stems and tendrils but low expression in other organs; group C includes eight genes with very low expression levels in all organs; and group D is the largest (15 members) subfamily and shows the highest expression in berries, shafts, and tendrils. In addition, the VvXTHs had higher expression levels in the pulp, peel, and stem during the V (veraison), MR (mid-ripening), and R (ripening) periods, indicating that VvXTHs may be related to fruit ripening. In short, the four groups of VvXTHs present specific expression profiles depending on the organ and developmental stage. This interesting phenomenon may be due to the specific functions of these specific genes in related tissues.

To verify the reliability of the organ-specific expression profiles, qRT-PCR analysis was conducted on five different tissues (tendril, root, stem, leaf and flower) of grapevine “Crimson” for VvXTHs, then the qRT-PCR results were compared with the data obtained from the GSE36128 data set of the cultivar “Corvina” (Marianna et al., 2012) with the same tissues at the corresponding developmental stages. It was found that the expression patterns of VvXTHs were generally consistent with the data obtained from the GSE36128 data set (Fig. S2), which suggests that temporal and spatial expression of VvXTHs is generally similar in different cultivars, even grown in different conditions.

Transcriptional profiles of VvXTHs under abiotic stress

The PlantCARE database and New PLACE database were utilized to identify cis-elements in the DNA sequences 1.5 KB upstream of the VvXTH start codons. The results showed that all 34 VvXTHs contained a variety of abiotic and biotic stress response elements, phytohormone response elements, and growth and development-related response elements (Fig. 5A). Similarly, in the New PLACE database, all 34 VvXTHs were predicted to contain phytohormone response elements and elements involved in the responses to a variety of abiotic stresses, including cold and heat, ABA, dehydration and salinity (osmotic) stress (Fig. 5B). As shown in Fig. S3, two drought stress response elements (MYB and MYC) exist in almost all members, and 80% of the gene members contain defense and stress response elements (STREs), which indicates that the VvXTH family probably shows important functions when plants are subjected to abiotic or biotic stress. Abscisic acid response elements (ABREs) and salicylic acid response elements (TCA elements) are abundantly present in VvXTH family members, which indicates that the VvXTH family may also be involved in hormone regulation during plant growth and the response to stress. The presence of a meristem development control element (CAT-box) in most VvXTH members suggests that the VvXTH family may have significant effects on the regulation of plant growth and development. In particular, the 34 VvXTHs were predicted to contain a large number of light response elements in the New PLACE database, and light-responsive elements (GATABOX and Box-4) were present in most VvXTH members, suggesting that the VvXTH family plays important roles in photosynthesis and photomorphogenesis.

Promoter analysis demonstrated the widespread presence of cis-elements associated with abiotic stress in the promoter regions of VvXTHs, revealing the possible induction of VvXTH expression by abiotic stress. To further investigate the potential roles of VvXTHs in response to abiotic stress, especially drought stress and salt stress, we selected 14 VvXTH members harboring abiotic response cis-elements for further study. Six-week-old grape seedlings were exposed to 200 mM NaCl or 200 mM mannitol, and the expression of 14 VvXTHs was examined in separated leaves and roots. In roots, the expression levels of 11 genes were upregulated under salt stress, among which four members were significantly upregulated. The expression of VvXTH5, VvXTH20, and VvXTH34 was increased by more than two-fold, and that of VvXTH4 was increased by more than four-fold. Interestingly, the expression of VvXTH4 peaked after 9 h of treatment, presenting an obviously different pattern from the other genes. This shows that these genes may respond to salt stress in different ways. Under drought stress, most of the genes whose expression was upregulated reached a peak after 3 h of treatment, and some genes were upregulated by more than four-fold (VvXTH3 and VvXTH20). VvXTH10 expression reached a peak after 12 h of treatment, indicating that this gene may be expressed at a later time. The number of upregulated VvXTHs in leaves relative to roots decreased after stress treatment, but these genes were more highly upregulated. Among these genes, VvXTH3, VvXTH10, and VvXTH31 were upregulated by approximately ten-fold. The above genes might play particularly crucial roles in the leaf response to salinity stress. Taken together, our findings indicate that the expression of VvXTHs could be altered by salt and drought stress, suggesting that VvXTHs may participate in reactions to abiotic damage, especially under salt and drought stress.