Functional Characterization of the Grapevine γ-Glutamyl Transferase/Transpeptidase (E.C. 2.3.2.2) Gene Family Reveals a Single Functional Gene Whose Encoded Protein Product Is Not Located in Either the Vacuole or Apoplast

Abstract

γ-glutamyl transferases/transpeptidases (E.C. 2.3.2.2, GGTs) are involved in the catabolism of many compounds that are conjugated to glutathione (GSH), which have a variety of roles. GSH can act as storage and transport vehicle for reduced sulfur; it is involved in the detoxification of xenobiotics and also acts as a redox buffer by utilizing its thiol residue to protect against reactive oxygen species, which accumulate in response to biotic and abiotic stress. Furthermore, many distinctive flavor and aroma compounds in Sauvignon blanc wines originate from odorless C5- and C6-GSH conjugates or their GGT catabolized derivatives. These precursors are then processed into their volatile forms by yeast during fermentation. In many plant species, two or more isoforms of GGTs exist that target GSH-conjugates to either the apoplast or the vacuole. A bioinformatics approach identified multiple GGT candidates in grapevine (Vitis vinifera). However, only a single candidate, VvGGT3, has all the conserved residues needed for GGT activity. This is intriguing given the variety of roles of GSH and GGTs in plant cells. Characterization of VvGGT3 from cv. Sauvignon blanc was then undertaken. The VvGGT3 transcript is present in roots, leaves, inflorescences, and tendril and at equal abundance in the skin, pulp, and seed of mature berries and shows steady accumulation over the course of whole berry development. In addition, the VvGGT3 transcript in whole berries is upregulated upon Botrytis cinerea infection as well as mechanical damage to leaf tissue. VvGGT3-GFP fusion proteins transiently over-expressed in onion cells were used to study subcellular localization. To confirm VvGGT3 activity and localization in vivo, the fluorescent γ-glutamyl-7-amido-4-methylcoumarin substrate was added to Nicotiana benthamiana leaves transiently over-expressing VvGGT3. In combination, these results suggest that the functional VvGGT3 is associated with membrane-like structures. This is not consistent with its closely related functionally characterized GGTs from Arabidopsis, radish and garlic.

Keywords: GGT; New Zealand Sauvignon blanc; glutathione; grape berry development; volatile thiol; γ-glutamyl transpeptidase.

Figure 1

Phylogeny of selected plant γ-glutamyl transferases/transpeptidases (GGTs). The accession numbers of these GGTs are presented in Table 1 . GGTs group into two distinctive subgroups, then further into the monocot and dicot clades. Subgroup 1 has GGTs, which are thought to prevent oxidative stress by degrading the oxidized form of glutathione (GSSG) and the breakdown of extracellular GSH in the apoplast. Subgroup 2 has GGTs which are thought to degrade GSH-conjugates in the vacuole. The GGT from grape, VvGGT3, highlighted in red, falls under subgroup 2. Multiple sequence alignments of deduced protein sequence that were used in this analysis can be found in Supplementary Data Sheet 2 .

Figure 2

Fluorescent images of Nicotiana benthamiana leaf discs infiltrated with the over-expressing VvGGT3 construct five days prior to the addition of the γ-glutamyl-7-amido-4-methylcoumarin substrate. Images are representative of the result obtained from at least 10 independent infiltrations. (A) Leaves transiently transformed with the VvGGT3 over-expressing construct and subsequently infiltrated with the substrate. (B) Untransformed leaves at the same age infiltrated with the substrate. (C) Leaves transiently transformed with the VvGGT3 over-expressing construct and subsequently infiltrated with molecular grade water. (D) Leaves transiently transformed with Agrobacterium tumefaciens GV3101 (MP90) without a transgene and infiltrated with the substrate. Scale bars = 100 μm.

Figure 3

Confocal images of onion cells transiently transformed with either: the VvGGT3-GFP fusion protein (top row), VvGGT3_N75-GFP fusion protein (middle row), or GFP alone (bottom row). The merged images demonstrate that neither the full length (VvGGT3) nor the first 75 AA of the protein (VvGGT3_N75) localize to the vacuole as predicted by phylogenetic analysis. The free GFP localizes to the nucleus and cell membranes as described in Ohkama-Ohtsu et al. (2007b). The images presented are representative of cells obtained from epidermal peels from three independent biolistic events on three individual bulb leaf sections bulb for each construct. Each peel was inspected to compare cells from at least three different fields of view. Two independent experiments carried out. Scale bars = 100 µm.

Figure 4

The mean relative abundance of VvGGT3 in a range of Sauvignon blanc tissues and stress treatments. All data is derived from biological replicates and technical triplicates were performed for each RT-qPCR gene targets. (A) The mean relative abundance in root (root tip, ∼15 mm), leaf (∼225 mm²), E-L 15 inflorescence (Coombe, 1995), and tendril (∼20 mm), and the skin, pulp, and seed fraction of mature berries [21°Brix, mostly E-L 38 berries (Coombe, 1995)], the expression level of tendril was set to 1. (B) The mean relative abundance in grape berries throughout development was studied in the 2006, 2007, 2008, and 2009 growing seasons. For each time point, the phenological stages mostly represented E-L 30, 31, 32, 35, 36, 36, 37, and 38, respectively (Coombe, 1995). Veraison in each season took place approximately 50 to 60 days after anthesis (daa). The level at 20 daa was set to 1 (n = 4 biological replicates, n = 3 biological replicates for 20 and 30 daa). (C) The mean relative abundance in whole grape berries from bunches prior to harvest [21 °Brix, mostly E-L 38 berries (Coombe, 1995)] infected with B. cinerea. Berries that showed no sign of infection were considered the control and set to 1, berries of healthy appearance adjacent to infected berries were considered not infected (no infection), and grape berries that showed signs of infection but were not fully infected (infected) were analyzed ( Supplementary Figure 6 ). (D) The mean relative abundance over 24 h in mechanically damaged young grape leaves (∼2.5–3 cm wide). The level of unwounded leaves was set to 1. For all assays (A–D), the geometrical means of VvActin and VvEF1α as reference genes were used as a normalization factors. Different letters indicate statistically significant differences, P < 0.05, ANOVA Tukey-Kramer, n = 3 biological replicates for graphs in (A), (C), and (D), means ± SEM.