Background

The glyoxalase system is an almost ubiquitous pathway for the detoxification of methylglyoxal (MG) in living systems. The typical glyoxalase system can detoxify the highly cytotoxic compound, MG to the nontoxic D-lactate through the sequential activity of two enzymes - glyoxalase І (GLYI; lactoylglutathione lyase) and glyoxalase II (GLYII; hydroxyacylglutathione hydrolase). First, MG is converted into hemithioacetal (HTA) in the presence of reduced glutathione (GSH), then GLYI catalyses the HTA isomerisation into S-D-lactoyl-glutathione (SLG). Subsequently, SLG is hydrolysed to D-lactate by GLYII, recycling GSH to the system [1, 2]. Recently, a one-step detoxification of MG has been proposed for a few organisms. This is mediated by glyoxalase III (GLYIII) enzyme without any cofactor [3,4,5,6,7].

Glyoxalase family identification has so far been carried out in four plant genomes Arabidopsis, Oryza sativa, Glycine max and Medicago truncatula [8,9,10]. These glyoxalase gene identifications have been based on sequence analysis. In the genome wide analysis, eleven GLYI genes have been described in Arabidopsis thaliana and in rice [8]. However, further study revealed that only three Arabidopsis thaliana genes (AtGLYI2, AtGLYI3 and AtGLYI6) and two rice genes (OsGLYI8 and OsGLYI11.2) were functional and contained all the binding sites required for glyoxalase activity. In addition, the activity of AtGLYI2 was dependent on Zn2+, while the activities of AtGLYI3, AtGLYI6 and OsGLYI11.2 were dependent on Ni2+ [11,12,13]. Specifically, OsGLYI-8 did not show any strict requirement for a metal ion for its activity [14]. Similarly, two out of five GLYII proteins from Arabidopsis thaliana (AtGLYII2 and AtGLYII5) and two out of three GLYII proteins from rice (OsGLYII2 and OsGLYII3) showed GLYII enzyme activity [15,16,17,18,19,20]. Glyoxalase II of all species, including human, yeast and Arabidopsis, contains a highly conserved metal binding motif (THXHXDH) that is also present in the family of metallo-β-lactamases, which are known to require Zn(II) [21,22,23]. However, the glyoxalase III enzyme does not require any cofactor and is reported to be a member of the DJ-1/Pfp1 (PARK7/Pyrococcus furiosus protease I) superfamily with a conserved catalytic triad Glu-Cys-His in its active site. Mutagenesis studies of Hsp31 revealed that the Cys-185 and Glu-77 were essential for catalysis, whereas His-186 was less crucial for enzymatic function, although it participates in the catalytic process [4, 24]. Site-directed mutagenesis of the conserved cysteine in the N-terminal domain of OsDJ-1C, which can utilise MG as substrate to produce D-lactate in a glutathione-independent manner in rice, resulted in the loss of GLYIII activity. This confirms it is a functional enzyme and that cysteine is necessary for its activity [7].

The glyoxalase system has been well studied in the animal kingdom and shown to play numerous roles in cell division and proliferation, microtubule assembly, embryogenesis, maturation and cell death [2]. An active plant glyoxalase enzyme was first reported in Douglas fir needles [25]. Thereafter, several studies have reported the presence of glyoxalase activity in both monocotyledons and dicotyledons [26,27,28,29,30,31]. It was found that over-expressed GLYI and/or GLYII in plants significantly increased their tolerance to various abiotic stresses such as salinity, drought, extreme temperature and heavy-metal toxicity [32,33,34,35,36]. Thus, glyoxalases have been suggested as biomarkers for plant stress resistance [31, 37]. In addition to abiotic stress responses, glyoxalase genes have also been reported to be regulated by biotic stresses. A rice GLYI gene was found to be down-regulated after infection by Xanthomonas oryzae pv. oryzae or by Pyricularia grisea [38]. Proteomic comparison analysis of maize kernel embryos showed significant up-regulation of GLYI protein in the resistant embryos in response to Aspergillus flavus infection [39]. Similarly, induction of GLYI activity was observed in rice in response to attack by brown planthopper (Nilaparvata lugens) and in Brassica after infection by Sclerotinia sclerotiorum. After fungal infection, intercellular MG levels were significantly increased in susceptible genotypes, so high GLYI activity was crucial for the detoxification of MG [39, 40].

The grapevine (Vitis vinifera L.) is a worldwide cultivated crop of high economic value. Sequencing of the highly homozygous grapevine PN40024 genome provided the opportunity to analyse the grapevine genome and to identify gene families [41]. Grape downy mildew, caused by the oomycete Plasmopara viticola, occurs in most parts of the world, especially in those where it is wet during the vegetative growth period. A major outbreak of the disease can cause severe losses in both yield and berry quality [42]. Genetic and gene expression profiling analyses show that Rpv1, NPR1 homologs, and PR protein encoding genes contribute to the function of downy mildew resistance in grapevine [43,44,45]. Moreover, overexpression of VpPR10.1 can enhance the transgenic grape plants resistance to downy mildew [46, 47]. It is important to find more crucial genes for grape downy mildew tolerance. In this study, we identify the gene families of the glyoxalase system in V. vinifera and discuss their roles in response to downy mildew. The existence of glyoxalases as multigene families in grape suggests there could be a number of undiscovered roles for these genes.

Results

Identification and detailed analysis of glyoxalase families in grape

To identify all the glyoxalase proteins in grape, lactoylglutathione lyase domain (PF00903), metallo-beta-lactamase domain (PF00753) and DJ-1/PfpI (PF01965) were downloaded from Pfam and used as queries to search the Grape Genome Database using HMMER with default E-values (< 1.0) [48]. The proteins were then manually confirmed using Pfam and BLASTP along with the previously reported GLYI proteins of Arabidopsis, rice, soybean and M. truncatula [8,9,10]. Proteins were identified that contained the glyoxalase domain (PF00903) and that contained all four conserved binding sites required for glyoxalase activity- i.e. (a) the active site, (b) the metal binding site, (c) the GSH binding site and (d) the dimer interface. These proteins were classified as VvGLYI-like (Table 1 and Additional file 1: Table S1) and those without all the four conserved binding sites were identified as non-glyoxalases (Additional file 2: Table S2) [8, 11]. Proteins that contained the metallo-beta-lactamase domain (PF00753) along with conserved metal binding sites (THXHXDH/H/D/H), an active site (C/GHT) and seven conserved GSH binding sites (C/K(R)/F(Y)/Y/N/R/K) and that had a putative hydroxyacyl glutathione hydrolase function were classified as VvGLYII-like (Table 1, Additional file 3: Table S3 and Additional file 4: Table S4) [8,9,10, 49,50,51]. Likewise, proteins that contained the DJ-1/PfpI (PARK7/Pyrococcus furiosus protease I) domain (PF01965) and Glu-Cys-His/Tyr catalytic triad were classified as VvGLYIII-like [6, 10]. A total of four GLYI-like, two GLYII-like and three GLYIII-like proteins were identified in the grape genome. These were named VvGLYI-like1 to VvGLYI-like4, VvGLYII-like1 to VvGLYII-like2 and VvGLYIII-like1 to VvGLYIII-like3 according to their location order on the grape chromosomes [8]. Meanwhile, all the grape glyoxalase-like genes identified were analysed in detail (Table 1). They were found to be located on eight different chromosomes. The CDS length of the VvGLY-like members varied from 708 bp to 1344 bp. The largest protein was VvGLYIII-like1 with a length of 447 aa, while the smallest was VvGLYI-like2 with a length of 235 aa (Table 1). Detailed information on the glyoxalase genes in grape are listed in Table 1, including their accession numbers, protein lengths, locations and similarities to the Arabidopsis’s orthologues.

Table 1 Detailed information for the GLYI, GLYII and GLYIII genes identified in Vitis vinifera L
Full size table

Protein motif identification and gene structure analysis of glyoxalase-like families of grape

The exon-intron structure of the VvGLYI-like, VvGLYII-like and VvGLYIII-like genes was varied (Fig. 1). There were eight exons in VvGLYI-like3 and VvGLYI-like2 genes (Fig. 1), while VvGLYI-like1 and VvGLYI-like4 both had nine exons. The numbers of exons in VvGLYII-like1 and VvGLYII-like2 were eight and seven respectively (Fig. 1). VvGLYIII-like1 and VvGLYIII-like3 each had eight exons, while VvGLYIII-like2 had five exons (Fig. 1). Genes with longer coding sequences are significantly more important and evolve more slowly than genes with shorter CDSs and they contain more functional domains within the gene [52].

Fig. 1
figure 1

Gene structures of grape glyoxalase genes and domain architectures of grape glyoxalase proteins. The exons are represented by black boxes while the introns are represented by black lines. Only the exons are drawn to scale because some introns are too long. The glyoxalase domain (PF00903) of VvGLYI-like is represented by an orange box. For VvGLYII-like proteins, β-lactamase domain is represented by a blue box and the hydroxyacylglutathione hydrolase C-terminus (HAGH-C) domain is represented by an orange hexagon. The DJ/Pfpl domain (PF01965) of VvGLYIII-like protein is represented by a dark green box. The domain position is indicated by the number inside the box or hexagon. The full length of protein is indicated by the numbers at both ends

Full size image

All predicted VvGLYI-like proteins were analysed using Pfam to identify the glyoxalase domain (PF00903). Results indicate that VvGLYI-like1 and VvGLYI-like4 contain two glyoxalase domains while the other proteins had only a single PF00903 domain (Fig. 1). Analysis of the VvGLYII-like proteins showed that they both had a metallo-beta-lactamase domain along with a hydroxyacylglutathione hydrolase C-terminus domain (HAGH-C) (PF16123) (Fig. 1). The HAGH-C domain was a substrate binding site usually found in the C-terminus of GLYII proteins along with the catalytic domain (PF00753) [53]. The previously reported DJ-1 proteins of rice and Arabidopsis all contain two DJ-1/PfpI domains, while the DJ-1 proteins of E. coli, Drosophila, Caenorhabditis elegans and human contain only one [5, 7]. Structural analysis of the three grape GLYIII-like proteins show they all have two DJ-1/PfpI domains (Fig. 1). Unlike human and other homologs, the plant glyoxalases seem to have higher efficiency or better regulatory properties. Thus, plants seem to have greater adaptability to their ever-changing environments from which they are unable to escape.

Conserved binding sites and phylogenetic analysis of operative GLYI proteins

GLYI is a metalloenzyme with conserved metal binding sites of H/QEH/QE, requiring divalent metal ions (Ni2+/Co2+ or Zn2+) for activation. It has been reported that GLYIs dependent on Zn2+ have longer domains (more than 140 aa) than Ni2+/Co2+ dependent ones (around 120 aa) and have unique regions in their sequences [11, 12, 54]. To determine metal ion dependency, the N-terminal lactoylglutathione lyasedomains of the operative GLYI proteins were aligned using ClustalW along with the human GlyI domain [55] (Additional file 5: Figure S1). Among the four VvGLYI proteins, VvGLYI-like1, VvGLYI-like3 and VvGLYI-like4 were predicted to be Ni2+-dependent because their domains were around 120 aa and specific Zn2+-dependent regions were absent. Meanwhile, VvGLYI-like2 was predicted to be Zn2+-dependent because of its domain length of 151 aa and because it had conserved regions for Zn2+-dependence (Additional file 1: Table S1 and Additional file 5: Figure S1).

To gain a broader understanding of the phylogeny of GLYIs, an unrooted phylogenetic tree was constructed using the Maximum Likelihood (ML) method, and based on the multiple sequence alignment of the lactoylglutathione lyase domains of the proteins listed in Additional file 1: Table S1 (Fig. 2a). According to the phylogenetic tree, the putative GLYI proteins can be divided into two subfamilies (I and II) (Fig. 2a). Subfamily I contains all the putative Ni2+-dependent GLYI proteins while the putative Zn2+-dependent GLYI proteins were in subfamily II. This provides a basis for our future research on gene functions.

Fig. 2
figure 2

Phylogenetic relationships of GLY-like proteins from various plant species. Unrooted trees were produced using MEGA5.1 software with the Maximum Likelihood method and 1000 bootstrap replicates using the N-term domain of GLYIs (a), full length amino acid sequences of GLYIIs (b) and the two domains of GLYIIIs respectively. Glyoxalase protein sequences are provided as Additional files 13, 14 and 15. Members of the glyoxalase gene families from grape, Arabidopsis, rice, soybean and Medicago truncatula are marked by red, green, yellow, purple and blue respectively. The domains of human, mouse, Drosophila, Caenorhabditis elegans and Escherichia coli are marked with a white rhombus. The bootstrap values are shown near the nodes. Scale bar represents 0.1 or 0.2 amino acid substitution per site

Full size image

Conserved sites and phylogenetic analysis of putative GLYII proteins

The full length of the VvGLYII-like proteins and of the previously reported GLYII proteins of Arabidopsis, rice, soybean and M. truncatula [8,9,10] were manually analyzed using BLASTP and multiple sequence alignment to analyse for the presence of conserved binding sites required for GLYII activity (Additional file 3: Table S3, Additional file 4: Table S4 and Additional file 6: Figure S2). The putative GLYII-like proteins in which all binding sites are present are listed in Additional file 3: Table S3. We speculate that these proteins should be studied further with regard to their roles as active GLYII proteins.

We constructed another tree using the ML method, based on multiple sequence alignment of the full length of the putative GLYII proteins listed in Additional file 3: Table S3 from grape, Arabidopsis, rice, soybean and M. truncatula [8,9,10] (Fig. 2b). This tree was divided into two subgroups. VvGLYII-like1 is in the branch I, along with AtGLYII-5 which has been reported to be an active GLYII protein located in the mitochondria and to have a predominant metal center of Fe(III)Zn(II) but does not seem specifically to bind manganese [18]. VvGLYII-like2 is in branch II, along with AtGLYII-2 which is reported to be an active GLYII protein located in the cytosol and that has been shown to bind a mixture of Zn, Fe or Mn [21]. We speculate that these proteins in the two branches may be located in different places in the cell and have different metal dependences (Fig. 2b).

Phylogenetic analysis of putative GLYIII proteins

It has been reported that both DJ-1and Hsp31 proteins contain the Glu-Cys-His catalytic triad. Among these, the glutamate and cysteine residues in the same locations in the two enzymes directly involved in catalysis are essential for GLYIII enzymatic activity. In contrast, histidine residues at different sequence and structural positions, are less important to enzyme activity [4, 5]. Previously reported DJ-1 protein sequences from Arabidopsis, rice, soybean and M. truncatula along with VvGLYIII-like proteins were aligned using ClustalW [9, 10, 56]. Incomplete sequences and proteins not containing the highly conserved glutamate and cysteine residues either in the N-terminal or C-terminal were discarded (Additional file 7: Figure S3). The remaining sequences were used for phylogenetic analysis. Using the MEGA 5.1 tool, a ML tree of the N-terminal and C-terminal DJ-1/PfpI domains from grape, Arabidopsis, rice soybean and M. truncatula along with the well characterised DJ-1/Hsp31 proteins from human, mouse, Drosophila, C. elegans and E. coli (Fig. 2c). This tree showed three major clades, I to III. Clade II contained all the DJ-1 proteins from the animal species. In Clade III, heat-inducible molecular chaperones (Hsp31) from E. coli, AtDJ-1D and OsDJ-1C have been shown in previous studies to be a primary GLYIII enzyme [4, 6, 7]. Thus VvGLYIII-like2 in this clade is speculated to be an important GLYIII enzyme worthy for further study. In Clade I, AtDJ-1a and AtDJ-1b exhibited weak specificity towards MG but higher specificity towards glyoxal and AtDJ-1a was involved in the oxidative stress response [23, 57]. Thus, we speculate that VvGLYIII-lik1 and VvGLYIII-lik3 in Clade I may exhibit weak glyoxalase activities or have different functions from glyoxalase.

Expression profiles of grape glyoxalase genes in different tissues and under downy mildew stress

To investigate the expression of glyoxalase genes in the different tissues of grape, we analysed the expressions of glyoxalase genes in shoots, stems, leaves, flowers, tendrils and ovules at 20/30 and 40 days after flowering (DAF) in V. vinifera L. cv. ‘Thompson seedless’ by qRT-PCR. VvMADS9 which is specifically expressed in flowers, served as the positive control [58, 59]. Based on the expression profiles, all the glyoxalase genes were expressed in ovules and most had higher expressions in stems, leaves, tendrils, but lower expressions in flowers. However, some differences were noted. For instance, VvGLYI-4 had lower expressions in roots, stems, leaves than in tendrils and ovules (Fig. 3). With confirmation of the expressions of glyoxalase-like genes in grape, PR1 and NPR1 (involved in the defence response) and VvGAPHD (a housekeeping gene) were evaluated by qRT-PCR following infection with P. viticola (Fig. 4). There were no significant differences in the expressions of VvGAPHD after infection. This indicates our methods are reliable. The expressions of PR1 and NPR1 were found to increase 24–48 h after inoculation. However, VvGLYI-like2, VvGLYI-like3, VvGLYI-like4, VvGLYII-like2 and all three VvGLYYIII-like genes were first expressed at 6–12 h, expression then decreased but increased again at 96–120 h, while VvGLYII-like1 was down-regulated. Meanwhile, VvGLYI-like1 was expressed most strongly at 48 h but like PR1 and NPR1, expression then decreased. It is speculated that VvGLYI-like1could play an important role in grape’s defence response to P. viticola (Fig. 4). The expression patterns of the various glyoxalase members are different. This indicates the various glyoxalase members play different roles in grape downy mildew stress modulation pathway.

Fig. 3
figure 3

Expression analysis of grape glyoxalase genes in different tissues. All qRT-PCR experiments were carried out with three biological replications, and three technical repeats. The highest expression level was set at 1. Duncan’s multiple range test (P ≤ 0.05) is used to calculate the significant differences showing by different letters. Standard deviations are shown by error bars

Full size image
Fig. 4
figure 4

Expression analysis of grape glyoxalase genes under downy mildew stress. Eight time points after downy mildew inoculation are used in the analysis. The qRT-PCR experiments are conducted with three biological replicates and three technical repeats. The highest expression level was set at 1. Standard deviations are shown by error bars and Duncan’s multiple range test (P ≤ 0.05) was used to calculate the significant differences indicated by different letters

Full size image

Discussion

Identification of glyoxalase genes in plants

Genome-wide analysis of the glyoxalase enzymes have been carried out previously in rice, Arabidopsis, soybean and M. truncatula [8,9,10]. However, these families have not been studied in grape. Previous studies on GLYIs and GlYIIs in soybean and M. truncatula based on the presence of conserved metal ion binding sites and substrate binding sites to determine whether a protein sequence had enzyme activity [9, 10]. However, the screening criterias used in these studies were not comprehensive. Jain et al. screened GLYI-like gene sequences in Arabidopsis and determined they contained not only metal binding sites, but also GSH binding sites, active sites and dimer interfaces [11]. Limphong et al. characterised GlYIIs in Arabidopsis and identified 14 amino acids that are important for substrate binding, metal binding and catalytic activity [49]. Using the new screening criteria, the present study identifies four GLYI-like proteins and two GLYII-like proteins in grape, 10 GLYI-like proteins and three GLYII-like proteins in soybean and five GLYI-like proteins and two GLYII-like proteins in M. truncatula. The number of previously reported proteins in full families and putative active members of all five plant species are listed in Additional file 8: Table S5, including the length of their genomes [60,61,62,63,64]. Reports on the structural properties of plant GLYIII proteins are few, with structure reported only for Arabidopsis AtDJ-1d protein, the conserved binding site was Glu-Cys-His triad [6]. Similarly, OsDJ-1C, a rice DJ-1 protein, generated through homology modelling studies using AtDJ-1d as a template, also had the conserved three-dimensional structure [7]. These glyoxalase families display wide differences in their sequences which might lead to diversity in their biochemical functions. However, at this stage, little is known about plant DJ-1 proteins.

Predicted metal ion dependence properties of glyoxalases

There are two types of metal dependency of GLYI proteins, Zn2+ and non-Zn2+ (mainly Ni2+). Regardless of the metal activation, both types have four conserved metal-binding residues (H/QEH/QE) [13, 57, 65]. Interestingly, metal dependence specificity of predicted GLYI proteins can easily be identified based on the length and sequence of proteins. The amino acid length of Zn2+-activated GLYI proteins is greater than of Ni2+-activated GLYI proteins and there is also a unique region in the sequence of Zn2+-activated proteins [11, 13, 56]. According to these criteria, in this study, three out of the four VvGLYI-like proteins are likely to be Ni2+-dependent and VvGLYI-like2 was speculated to be Zn2+-dependent. The same pattern of Ni2+-dependence dominance has been found in other plant species, two out of three Arabidopsis GLYI proteins are Ni2+-activated [11]. In addition, according to this study seven out of 10 soybean GLYI-like proteins and four out of five M. truncatula GLYI proteins are predicted to be Ni2+-activated (Additional file 1: Table S1). Moreover, previous reports have detected the presence of two GLYI domains in a single protein from Saccharomyces cerevisiae, Plasmodium falciparum and O. sativa [13, 66, 67]. Two domains could form two active sites in a single protein and both active sites were found to play roles in P. falciparum but one of the active sites was found to have no function in O. sativa [13, 67]. Here, VvGLYI-like1 and VvGLYI-like4 had two conserved glyoxalase domains (PF00903) (Fig. 1). They also possess four conserved metal binding residues in the N- terminal and three in the C- terminal (Additional file 5: Figure S1 and Additional file 9: Figure S4). That is the C- terminal domain of VvGLYIlike-1 and VvGLYI-like4 may have no function, as in rice. Similarly, GLYII enzymes also require divalent cations for their activity. GLYII proteins of plants, usually possess a binuclear metal binding center [16, 18, 20, 23]. AtGLYII-5 is reported to have a predominant metal center of Fe(III)Zn(II) but does not appear to bind Mn specifically [18]. However, AtGLYII-2 has been shown to bind a mixture of Zn, Fe or Mn and it also exhibits positive cooperativity in metal binding [21]. The gene dependence on metal ions still needs to be proved by experiments. We just predicted the metal dependence, providing some ideas for future research.

Previous characterisations of GLYIII activity in all species display a metal ion-independent nature [5, 6]. However, Hsp31 and human DJ-1 as glyoxalases reveal the existence of the essential catalytic residues, glutamic acid and cysteine, together with a histidine residue that assists catalysis [4, 5]. However, the histidine positioning in DJ-1 and Hsp31 are different. In the present study, both domains in the N-terminal and C-terminal of VvGLYIII-like2 resembled Hsp31, AtDJ-1d and OsDJ-1C in that the cysteine and histidine were contiguous (Additional file 7: Figure S3). These proteins are reported to be important detoxification GLYIII enzymes in Arabidopsis, rice and E. coli [6, 7]. Thus we predict that VvGLYIII-like2 could have the same functions as these and might play important roles as a GLYIII in grape. In contrast, the N-terminal of VvGLYIIIlike-3 resembled that of Drosophila in which histidine is replaced by tyrosine but the histidine in the C-terminal of VvGLYIII-like3 was replaced by phenylalanine [68]. However, the N-terminal and C-terminal of VvGLYIII-like1 are in the same as human DJ-1protein (Additional file 7: Figure S3). Although the catalytic core is similar, catalytic activity was quite variable among the members of the GLYIII family and this suggests a variety of function for the members in this family.

Different expression patterns of glyoxalase genes in response to stress

The role of the glyoxalase system in resistance to multiple biotic and abiotic stresses has been well characterised in some studies. Recently, it has been reported that the glyoxalase pathway of rice not only increases tolerance to salinity, drought and extreme temperatures but also reduces damage from sheath blight fungus (Rhizoctonia solani) [69]. Moreover, some glyoxalase members of soybean such as GmGLYI-6/9/20 and GmGLYII-6/10 show special expression patterns in response to various pathogenic infection [9]. However, whether glyoxalase proteins participate in downy mildew response is unknown. In this study, we found that after downy mildew inoculation, most glyoxalase genes had two periods of high expression at 6–12 h and at 96–120 h. In particular, VvGLYII-like1 had a down-regulated expression pattern after downy mildew inoculation. After fungal infection, intercellular MG levels are significantly enhanced in susceptible genotypes, so high GLYI activity is needed [5]. Nevertheless, a rice GLYI gene was found to be down-regulated after infection with Xanthomonas oryzae pv. oryzae or with Pyricularia grisea [38]. Like PR1 and NPR1, VvGLYI-like1 was highly expressed 48 h after inoculation. Thus, VvGLYI-like1 may play an important role in the defence response to downy mildew. Whether the other glyoxalase-like genes are involved in defence responses still needs to be clarified. The infection process of P. viticola is complex and takes a long time, the expression levels of genes during this process are very variable. Further studies are required to determine if over expression of glyoxalase genes could offer direct protection to plants from pathogens.

Moreover, the high identity of glyoxalase genes between grape and Arabidopsis at the protein level indicates that glyoxalase genes may have similar conserved functions to their Arabidopsis orthologue genes (Additional file 10: Table S6). As with the grape glyoxalase-like genes, AtGLYI-2, AtGLYI-3, AtGLYI-6, AtGLYII-2 and AtGLYII-5 are highly expressed at all the stages of development [8]. The frequency of ESTs or cDNAs available in different databases has been considered as a useful tool for preliminary analysis of gene expression [70]. The search for ESTs was performed by comparing the grape glyoxalase cDNA sequences against EST sequences available at NCBI in BLASTN searches with the search parameters: maximum identity, > 95%; length, > 300 bp; and E value, < 10− 10 [71]. The ESTs were identified for all of the GLY genes, indicating they are expressed. However, the frequency of ESTs was low, indicating that these are expressed at very low levels. The matched EST sequences were derived from various grape tissues or libraries such as root, leaf, berry, seed, bud, flower, pericarp, shoot, clusters and stem, indicating the differential expression of VvGLY-likes in grape tissues (Additional file 11: Table S7).

Conclusions

This study identifies four GLYI-like, two GLYII-like and three GLYIII-like glyoxalase proteins in grape. All these newly-identified members were further analysed for their homology, essential binding sites for catalytic activity, structural properties, expressions in different tissues and expressions under stress from downy mildew infection. This study lays a foundation for researching the roles of GLY genes of grape in response to downy mildew stress. Most glyoxalase genes were highly expressed at 6–12 h and 96–120 h post inoculation but VvGLYII-like1 expression was down-regulated. In particular, VvGLYI-like1 was highly expressed 48 h after inoculation and, like PR1 and NPR1, might be involved in defence response to downy mildew. Despite many recent advances in functional studies of glyoxalase genes, the biological functions of most glyoxalase genes in physiological and developmental processes or plant defence still require elucidation. The bioinformatic analysis and expression patterns of the grape glyoxalase gene families conducted in our study provide an overview of the composition and expression of glyoxalase genes in grapevine that will facilitate selection of candidate genes for cloning and for further functional characterisation.

Methods

Identification of glyoxalase genes in grape

To identify the GLYI, GLYII and GLYIII genes in grape, the Hidden Markov Model (HMM) profiles of the glyoxalase domain (PF00903), metallo-beta-lactamase domain (PF00753) and DJ-1/PfpI (PF01965) were downloaded from the Pfam database (http://pfam.xfam.org/). They were then used as queries to search the Grape Genome Database (12X) (http://www.genoscope.cns.fr/externe/GenomeBrowser/Vitis/) using HMMER3.1 with default E-values (< 1.0) [48]. All putative GLYI, GLYII and GLYIII genes were manually confirmed by searching the NCBI (https://www.ncbi.nlm.nih.gov/) and the Pfam program again with score value ≥100 and e-value less than 10− 30. To find the possibly active GLYI proteins, each of the identified VvGLYI-like protein sequences along with putative GLYI proteins in rice, Arabidopsis, soybean and M. truncatula was analysed using BLASTP at the score value ≥100 and E value ≤0. Proteins with conserved binding sites, active sites, metal binding sites, GSH binging sites and dimer interfaces were assumed to be GLYI-like. The predicted proteins are named: prefix “Vv” for Vitis vinifera, followed by GLYI-like, GLYII-like or GLYIII-like, and lastly the Arabic numbers (e.g. 1, 2, 3) based on their locations on the chromosome.

Exon-intron and protein structure analysis

The exon and intron of the glyoxalase-like genes were identified according to the Grape Genome Browser. Diagrams were drawn using the online tool GSDS 2.0 (http://gsds.cbi.pku.edu.cn) [72]. Coding sequences (CDs) and untranslated region (UTR) sequences of glyoxalase genes are provided in Additional files 16. All the newly predicted grape glyoxalase-like proteins were analysed using Pfam (http://pfam.xfam.org) to reveal the position of domains.

Multiple sequence alignment and phylogenetic analyses

Two unrooted phylogenetic trees were constructed respectively with amino acid sequences of N-terminal lactoylglutathione lyase domain (PF00903) in GLYI-like and full length of GLYII-like proteins from grape Arabidopsis, rice, soybean, M. truncatula using MEGA 5.1 with the Maximum Likelihood method and 1000 bootstrap replicates [73, 74]. In addition, N-terminal and C-terminal DJ-1/PfpI (PF01965) domain from grape Arabidopsis, rice, M. truncatula along with DJ-1/PfpI domain from human, mouse, Drosophila, C. elegans and Hsp31 from E. coli were used to construct another phylogenetic tree using the same method and the sequences are shown in Additional files 13, 14 and 15 respectively. Multiple sequence alignment was carried out using ClustalW [75] and visualized with Jalview [76].

Expression of grape glyoxalase genes in different tissues and under downy mildew stress

‘Thompson Seedless’ (Vitis vinifera L.) was grown in the vineyard of Northwest A&F University, Yangling, Shaanxi, China and grapevines were managed according to local standards. Samples of roots, stems, leaves, flowers, tendrils and ovules were collected. Ovules were collected 20 DAF (small globular embryo), 30 DAF (globular embryo) and 40 DAF (aborted embryo). For the downy mildew treatment, downy mildew (Plasmopara viticola) infected leaves were collected from the vineyard. Sporangia were isolated using vacuum aspiration [77]. Detached healthy leaves of ‘Thomson seedless’ were obtained from the same position on the shoot (fifth to sixth unfolded leaf). These were surface sterilized with 70% (vol/vol) ethanol and then rinsed in deionised water. The abaxial leaf surfaces were inoculated with 100 μl droplets of an aqueous suspension of sporangia (2 × 106 sporangia ml− 1). The control plants were neither inoculated nor elicitor-treated. The petioles were then covered with a piece of wet cotton, incubated in the dark at 25 °C and > 95% relative humidity (RH) [78]. Inoculated leaves were then collected after 0, 6, 12, 24, 48, 72, 96 and 120 h, these were immediately immersed in liquid nitrogen and stored at − 80 °C pending analysis. All the plant materials were replicated three times.

RNA isolation, qRT-PCR and statistical analysis

The EZNA Plant RNA Kit (Omega, Guangzhou, China) was used to extract the total RNA following the manufacturer’s instructions. For qRT-PCR (Quantitative Real-time PCR) analysis, the first-strand cDNA was synthesised using PrimeScript RTase (Takara, Dalian, China) according to the manufacturer’s instructions. Primers were designed by Primer Premier 5.0 (Premier, Canada) and these are listed in Additional file 12: Table S8. The qRT-PCR was carried out using SYBR green (Takara, Dalian, China) with the IQ5 real time PCR system (Bio-Rad, Hercules, CA, USA). Grape Actin 7 (Accession no. XM_002282480) was used as an internal control. VvMADS9 (Accession no. NM_001280946), VvPR1 (Accession no. XM_002273752) and VvNPR1 (Accession no. XM_002281439) were used as a positive control. Three biological repeats were used and each reaction was carried out three times. The normalised expression method (2-△△C(t) method) was used to analyse the relative expression levels with the highest level of expression set as 1. Statistical analyses were conducted using SPSS 18.0 Software (Chicago, IL) and Excel. Statistically significant analyses were based on Duncan’s biological repeats and each reaction is reported as a mean with standard deviations (SDs) presented as error bars.