ATP-hydrolysis and proton pumping by the V-ATPase (vacuolar proton-translocating ATPase) are subject to redox regulation in mammals, yeast and plants. Oxidative inhibition of the V-ATPase is ascribed to disulfide-bond formation between conserved cysteine residues at the catalytic site of subunit A. Subunits containing amino acid substitutions of one of three conserved cysteine residues of VHA-A were expressed in a vha-A null mutant background in Arabidopsis. In vitro activity measurements revealed a complete absence of oxidative inhibition in the transgenic line expressing VHA-A C256S, confirming that Cys256 is necessary for redox regulation. In contrast, oxidative inhibition was unaffected in plants expressing VHA-A C279S and VHA-A C535S, indicating that disulfide bridges involving these cysteine residues are not essential for oxidative inhibition. In vivo data suggest that oxidative inhibition might not represent a general regulatory mechanism in plants.
The V-ATPase (vacuolar proton-translocating ATPase) is a nanometre-scale molecular machine that performs rotational catalysis to drive protons across membranes. The ATP-hydrolysis motor V1 and the proton turbine Vo are functionally connected by a central shaft. ATP hydrolysis at the catalytic nucleotide-binding sites found at the interfaces of the hexameric arrangement of V1 subunits A and B induces rotation of the central shaft (subunits d, D and F), which is conveyed to the proteolipid ring (subunits c, c′ and c″). Rotation of the c-ring against subunit a, that provides the second part of the proton channel, drives protons across the membrane. Co-rotation of the two domains is prevented by four connections between V1 and V0 (subunits a, E, G, C and H).
The V-ATPase is found in the endomembrane system of all eukaryotes as well as at the plasma membrane of specialized animal cells. V-ATPase activity controls cytosolic and lumenal pH as well as the membrane potential and the resulting electric and chemical gradients create the protonmotive force, which drives secondary active transport. Furthermore, organellar acidification is essential for vesicular trafficking along both the exocytotic and endocytotic pathways [1,2]. Given that V-ATPase function is required for various physiological processes in many compartments, regulation of V-ATPase activity is assumed to be of crucial importance . Plant V-ATPase activity is controlled by a number of regulatory mechanisms, including differential localization of VHA-a isoforms [4,5], transcriptional regulation  and biochemical modulation by phosphorylation or directly by adenylates [7–9]. However, the physiological relevance of many of these regulatory mechanisms has not been demonstrated in vivo.
On the basis of in vitro studies, V-ATPases from mammals, fungi and plants are subject to oxidative inactivation and activity can be recovered by reducing agents [10–12]. Three cysteine residues are conserved in subunit A (VHA-A in Arabidopsis) of all eukaryotes and formation of an intramolecular disulfide-bond between Cys261, that resides within the nucleotide-binding P-loop motif, and Cys539 of subunit A was suggested to be responsible for oxidative inhibition, whereas disulfide formation between Cys284 and Cys539 is assumed to recover the activity and to represent the active state of the V-ATPase [11,13–15]. However, as Cys284 and Cys539 are essential for correct folding or stability of the A subunit in yeast [16,17], it has not been possible to test this model. Evidence for the in vivo importance of redox regulation of the V-ATPase is thus circumstantial. The yeast CYS4 mutant is deficient in cysteine and hence in glutathione biosynthesis. The less reducing cytosolic environment in this mutant results in partial loss of V-ATPase-activity that can be restored by addition of reduced glutathione or by replacing Cys261 by a serine .
Similar to other V-ATPases, the plant enzyme is reversibly inactivated by H2O2, GSSG and GSNO (nitrosoglutathione), and activity can be recovered by reducing agents such as DTT (dithiothreitol). Interestingly, redox-dependent activity changes are not restricted to the VHA-A subunit, but also involve intramolecular disulfide bridge formation in the stalk subunit VHA-E .
To elucidate the in vivo relevance of redox regulation of the plant V-ATPase, we individually addressed the importance of the conserved cysteine residues in the VHA-A subunit by site-directed mutagenesis and complementation of a null allele of Arabidopsis VHA-A . The impact of physiological oxidizing agents such as GSSG, H2O2 or GSNO on the V-ATPase variants was analysed by recording their ATP hydrolysis activity. The results of the present study demonstrate the differential importance of Cys256, Cys279 and Cys535 for the plant V-ATPase and allow us to draw conclusions on the mechanism of redox regulation.
Plant materials and growth conditions
Arabidopsis thaliana (Col-0) was grown in soil-culture in a growth chamber with 12 h of light (240 μmol quanta·m−2·s−1, 19°C) and 12 h dark (18°C) with 60% relative humidity.
For the patch-clamp experiments, A. thaliana plants were cultivated in climate chambers under conditions described previously . According to Beyhl et al. , mesophyll protoplasts enzymatically isolated from the leaves of 5–7-week old plants were subjected to osmotically induced release of the vacuoles.
For vacuolar pH measurements, seeds of A. thaliana ecotype Col-0 seeds and the respective mutant lines were surface-sterilized with ethanol and sown on to plates containing 0.5× MS salt (Duchefa) solidified with 0.5% phyto agar (Duchefa). The pH of the medium was adjusted to 5.8 using KOH. After stratification for 48 h at 4°C, seedlings were grown at 22°C with a photoperiod of 16 h.
Plasmid constructs and plant transformation
Site-directed mutagenesis was performed by PCR using a hybrid BamHI fragment including the VHA-A promoter and the first four exons fused to a cDNA fragment of the remaining exons  in pBluescript as template. The presence of mutations was confirmed by sequencing and the resulting mutant BamHI fragments were cloned into the binary vector pPZP312 in which the VHA-A terminator was inserted as a SalI–HindIII fragment.
Binary plasmids were introduced into Agrobacterium tumefaciens strain GV3101:pMP90 and selected on 5 μg/ml rifampicin, 10 μg/ml gentamycin and 100 μg/ml spectinomycin. vha-A/+ plants were transformed using standard procedures, and transgenic plants were selected on soil by spraying with 0.1% Basta®. Presence of the vha-A allele in the T2 generation was assured by selection on 50 μg/ml kanamycin.
Isolation of tonoplasts and V-ATPase activity measurements
Tonoplasts were isolated as described previously . The tonoplast suspension was supplemented with Brij58 buffer [50 mM Tricine, pH 8.0, 20% (v/v) glycerol and 2% (w/v) Brij58], resulting in a V-ATPase/detergent ratio of 1:10 and incubated for 30 min at 4°C. The samples were centrifuged at 98000 g and 4°C. The supernatant contained the solubilized V-ATPase. The activity was measured by recording the released phosphate according to the method of Benzini with modifications as described by Dietz et al. . Bafilomycin-sensitive ATP hydrolysis was related to V-ATPase activity. GSNO (50–750 μM), GSSG (1–10 mM) and H2O2 (50–2000 μM) were applied as oxidizing agents and tonoplast membranes were incubated for 30 min at 4°C before measuring V-ATPase activity.
Patch-clamp experiments on vacuoles were performed in the whole-vacuole configuration as described previously [23,24], following the convention for electrical measurements on endomembranes . An EPC10 patch-clamp amplifier (HEKA) was used for macroscopic current recordings from mesophyll cell vacuoles at a data acquisition rate of 10 ms. After macroscopic currents were low-pass filtered at 100 Hz, data were digitized by the integral LIH8+8 of the EPC10 (HEKA). The data were stored on a windows-environment computer. The clamped voltages were corrected for the liquid junction potentials . Different software such as Patchmaster (HEKA) and IGOR PRO (WaveMetrics) were used for data acquisition and off-line analysis. Macroscopic currents were normalized to membrane capacitance (Cm) of the respective vacuole to allow quantitative comparison of the proton transport capacity among different vacuoles. The holding potential for V-ATPase measurements was set to 0 mV. Symmetrical solute conditions (except of pH and substrate) were chosen with bath and pipette solutions both composed of 100 mM KCl, 5 mM MgCl2 and 1 mM CaCl2. The solutions were adjusted to an osmolality of 300, 350 or 400 mOsmol/kg−1 with D-sorbitol. The pH values of the solutions were adjusted to pH 7.5 (bath media) with 10 mM Hepes/Tris and with 10 mM Mes/Tris to pH 5.5 (pipette solution). Bath solution additionally containing 5 mM Mg-ATP and 2 mM H2O2 were applied to the cytosolic side of the vacuolar membrane via an application pipette in front of the respective vacuoles.
Vacuolar pH measurements
Vacuolar pH measurements in Arabidopsis roots were essentially performed as described previously  with some minor modifications. Seedlings (6-day-old) were incubated in liquid medium (0.5× MS, pH 5.8) containing 10 μM of the membrane-permeant pH sensitive dye BCECF [2′,7′-bis-(2-carboxyethyl)-5(6)-carboxyfluorescein]-AM (acetoxymethyl ester) (Molecular Probes, Invitrogen) and 0.02% Pluronic F-127 (Molecular Probes, Invitrogen). Dye loading was conducted for 1 h at 22°C in the dark followed by two washing steps, 5 min each, in liquid medium. For the H2O2 treatment, the dye-loaded seedlings were incubated for 15 min in liquid medium (0.5× MS, pH 5.8) containing 2 mM H2O2. Fluorescence microscopy was performed on a Leica SP5II confocal laser-scanning microscope equipped with an inverted DMI6000 microscope stand using a Leica HCX PL APO CS 20.0×0.70 IMM UV objective. BCECF was sequentially excited using the 458 and 488 nm laser line of the argon laser. Fluorescence emission was detected between 510 and 550 nm. The average fluorescence intensities were measured in defined regions of interest using ImageJ v1.43 (NIH) followed by calculating the ratio between the intensities of the 488 nm and the 458 nm excited images. In situ calibrations were performed by treating BCECF-loaded seedlings for 20 min with pH equilibration buffers containing 50 mM Mes-BTP (bis-Tris propane) (pH 5.2–6.4) or 50 mM Hepes-BTP (pH 6.8–8.0) and 50 mM ammonium acetate. Calibrations were performed in the presence and absence of 2 mM H2O2. Ratio values were plotted against the pH and the calibration curves were generated using a sigmoidal Boltzmann fit (see Supplementary Figure S1 at http://www.BiochemJ.org/bj/448/bj4480243add.htm).
SDS/PAGE and Western blot analysis
Tonoplasts were isolated and proteins were separated by non-reducing SDS/PAGE (12.5% gels). The samples were solubilized in SDS/PAGE buffer containing 125 mM Tris/HCl, pH 6.8, 2.5% (w/v) SDS, 10% (v/v) glycerol and 0.02% Bromophenol Blue. Western blot analysis was performed with a rabbit antibodies directed against the holoenzyme of Kalanchoe daigremontiana, and against the VHA-E or VHA-A subunits [27–29]. The secondary antibody was conjugated to alkaline phosphatase, and bands were detected after colour development in carbonate buffer (100 mM NaHCO3, pH 9.8, and 1 mM MgCl2) with 5-bromo-4-chloro-3-indolylphosphate and p-Nitro Blue Tetrazolium chloride as substrates. For redox titration, the redox potentials were calculated on the basis of the Nernst equation for a two-electron reaction and realized by corresponding ratios of reduced and oxidized DTT. The total concentration of DTT was 50 mM.
Chemical cysteine cross-linking
Cross-linking of V-ATPase was performed using p-PDM (N,N′-p-phenylbismaleimide) as a cross-linker. Approximately 0.2 μg/μl protein (50 mM Hepes, pH 7.5) was mixed with p-PDM in a final molar ratio of 1:100. The reaction was carried out for 45 min at room temperature (23°C). Cross-linking was stopped by addition of 20 μl Stop buffer (125 mM Tris, pH 6.8, 20% glycerol, 5% 2-mercaptoethanol, 4% SDS and 0.003% Coomassie Brilliant Blue). The samples were immediately subjected to SDS/PAGE.
Amino acid sequences were obtained from the NCBI database and crystal structures were from the RCSB PDB database (VHA-A: PDB code 1VDZ, ; and hexameric head: PDB code 3GQB ). Alignments were performed using ClustalW and intramolecular mapping of cysteine residues was performed with SwissProt PDB Viewer. The theoretical probability of disulfide formation was predicted using DISULFIND .
Mapping of conserved cysteine residues in VHA-A
The catalytic V-ATPase subunit A is highly conserved across all kingdoms. The amino acid sequence of Arabidopsis VHA-A shares 48–50% identity with archaebacterial and 59–69% with eukaryotic subunit A sequences. The residues Cys256, Cys279 and Cys535 of VHA-A are conserved among all eukaryotic sequences. Cys256 is located within the nucleotide-binding P-loop domain. Interestingly, Cys256 is replaced by a serine residue in thermophilic and anaerobic archaebacteria such as Pyrococcus horikishii (Figure 1A) and Thermus thermophilus. Cys256, Cys279 and Cys535 correspond to Ser238, Cys261 and Cys514 of the P. horikoshii VHA-A subunit. The structure of P. horikoshii VHA-A has been resolved previously . In P. horikoshii, the distances between these amino acid residues range between 11 and 29 Å (1 Å=0.1 nm) (Figure 1B) and flexible regions, that may have an impact on the distances between these cysteine residues, were not reported by Maegawa et al. . Although the nucleotide-binding state of the crystallized VHA-A was not specified in that report, the distances exceed 7 Å, which is the maximum spacing between cysteine residues that allows disulfide formation . The sequence-based tool DISULFIND also predicted the absence of intramolecular disulfide bonds with a high level of confidence . The fact that Cys256 is conserved in aerobic, but not in anaerobic, organisms points to its function in redox regulation of V-ATPase activity and indicates that intramolecular disulfide formation involving Cys256, if enabled by structural flexibility, is not essential for V-ATPase activity. The data obtained from the available crystal structure represents a first indication for the absence of intramolecular disulfide formation involving Cys256.
Alignment and distance between conserved cysteine residues in VHA-A
In vivo analysis of conserved cysteine residues
The Arabidopsis genome contains a single gene encoding the catalytic subunit VHA-A , making it an ideal model system to study the structure and function of the catalytic subunit in vivo. To individually address the function of the conserved cysteine residues in vivo, residues 256, 279 and 535 of VHA-A were mutated to serine. Expression of the resulting proteins was placed under the control of the endogenous VHA-A promoter and terminator sequences. Loss of function alleles of VHA-A cause gametophytic lethality  and thus heterozygous (vha-A/+) plants were transformed with the respective mutant constructs as well as with a wild-type control. Homozygous vha-A/vha-A plants were identified in the T3 generation for all transgenes. The presence of the introduced mutation was verified for two independent lines each by PCR and sequence analysis. With the exception of Cys279, all transgenic lines complemented with the VHA-A mutants were found to be indistinguishable from the wild-type (Figure 2A). Plants expressing C279S-substituted VHA-A showed reduced growth as indicated, e.g. by a decrease in rosette diameter (Figure 2C). Differences in C279S and wild-type became apparent at an age of four weeks (Figure 2). Western blot analysis showed that all mutant proteins were expressed at levels similar to endogenous VHA-A (Figure 2C). On the basis of these results none of the three conserved cysteine residues is essential for function or stability of VHA-A and thus this enabled us to study the mechanism underlying oxidative inhibition and its biological relevance both in vitro and in vivo.
Growth phenotype of the VHA-A mutants C256S, C279S and C535S compared with the wild-type
Biochemical characterization of V-ATPase complexes lacking conserved cysteine residues
To determine the biochemical properties of the VHA-A variants, tonoplast-enriched membranes were isolated from the leaves of wild-type and the complemented transgenic lines and incubated with either the reducing agent DTT or the oxidizing agent H2O2. The proteins were separated by non-reducing SDS/PAGE, immobilized on a membrane and immunodetected using antibodies against VHA-A. Mobility shifts indicative of disulfide formation were not observed for VHA-A, neither in the wild-type nor in the variants C256S, C279S and C535S (Figure 3A). More specifically, the electrophoretic mobility of wild-type VHA-A was unaffected by redox conditions in the range of −410 mV to −250 mV, thus indicating that intramolecular disulfide formation does not occur in this range (Figure 3B). The conclusion is supported by the absence of characteristic doublet bands that indicate transition between the fully reduced and fully oxidized state. In Figure 3, the 55 kDa band probably corresponds to VHA-B, which is also recognized by the applied antibody . As disulfide bridge formation in the stalk subunit VHA-E has been reported , tonoplast-enriched membranes of the transgenic lines C256S, C279S and C535S were also subjected to Western blotting with an antibody directed against VHA-E. The Western blots revealed a preference for the dimeric and monomeric form under reducing conditions, whereas high-molecular-mass aggregates on expense of the monomeric form appeared under oxidizing conditions. The three transgenic lines were further characterized by distinct patterns of the high-molecular-mass aggregates (Figure 3C), so that under oxidizing conditions the conformation of VHA-E seems coupled to the presence of individual cysteine residues within VHA-A. Thus we isolated V-ATPases from wild-type plants and the three transgenic lines C256S, C279S and C535S. The samples were chemically cross-linked using the cysteine cross-linker p-PDM and proteins were separated by SDS/PAGE. Products were visualized with an antibody directed against the VHA-A subunit. In the Western blot analysis of the cross-linked wild-type V-ATPase, a band was visible with an apparent mass of approximately 100 kDa. This band can also be seen in the C279S and C535S lines, but it is absent if Cys256 is substituted with a serine residue (Figure 4). The same was observed if an antibody against VHA-E was applied (Supplementary Figure S2 at http://www.BiochemJ.org/bj/448/bj4480243add.htm). These results might point to an intermolecular disulfide bond formation involving Cys256. Although the most promising and reported candidate for such a disulfide bridge with VHA-A might be VHA-B, the size of the observed cross-linking product falls below the expected molecular mass of 123 kDa for VHA-A and VHA-B.
Western blot analysis of isolated V-ATPase from the transgenic plants compared with the wild-type
Cysteine cross-linking of V-ATPase subunits
In vitro and in vivo activity measurements of V-ATPase complexes lacking conserved cysteine residues
As we failed to find evidence in support of intramolecular disulfide formation in VHA-A, we next determined the effect of C256S, C279S and C535S on V-ATPase activity and oxidative inhibition. Whereas bafilomycin-sensitive ATP hydrolysis of tonoplast-enriched membranes from wild-type, C256S and C535S plants proceeded at rates of 2.5–3.1 μmol phosphate·mg of protein−1·s−1 under control conditions, C279S plants were characterized by reduced ATP-hydrolysis activity (1.2 μmol phosphate·mg of protein−1·s−1). To compare the effects of the thiol-modifying reagents H2O2, GSSG and GSNO on ATP hydrolysis, the results are therefore given as relative activity after setting the initial ATPase activity to 100%. The transgenic lines C279S and C535S behaved similarly to the wild-type, for which H2O2, GSSG and GSNO were potent inhibitors of ATP hydrolysis. In contrast, C256S was largely unaffected by H2O2, GSSG and GSNO. Only very high concentrations of GSNO exceeding the physiological range had a significant effect on ATP hydrolysis (Figure 5).
V-ATPase activity in response to oxidizing agents
To determine the effect of the C256S and C535S mutations on proton transport, the patch-clamp technique was applied and ATP-stimulated proton currents were recorded in the whole-vacuole configuration (Figure 6A). Interestingly, whereas ATP hydrolysis was unaffected under control conditions for C256S and C535S, the proton currents were reduced in both mutants, indicating that their coupling rate (nH+/ATP) was reduced. However, in the presence of 2 mM H2O2, proton currents in C535S, but not in C256S, were significantly inhibited, supporting the notion that oxidative inhibition is not based on intramolecular disulfide-bond formation.
Effect of H2O2 on V-ATPase-mediated H+-currents and vacuolar pH
We have shown previously that the vha-a2 vha-a3 mutant lacking the tonoplast V-ATPase showed a significant increase in vacuolar pH compared with the wild-type , thus we analysed the effects of the C256S, C279S and C535S mutations on vacuolar pH under control conditions and in the presence of 2 mM H2O2. Vacuolar pH was measured in epidermal and cortex cells of 6-day-old Arabidopsis roots using the cell-permeant pH-sensitive dye BCECF-AM. Previous studies have demonstrated that BCECF specifically accumulates in the central vacuole of Arabidopsis root cells and is therefore suitable to measure vacuolar pH [5,35]. To rule out any interference of the fluorescent properties of the pH sensor with the oxidizing environment of H2O2, pH calibrations were performed either under standard conditions or in the presence of 2 mM H2O2 (Supplementary Figure S1). No differences in vacuolar pH between the wild-type and the VHA-A mutants were observed, neither under control conditions nor after 15 min of 2 mM H2O2 treatment (Figure 6B). However, as demonstrated previously , the vha-a2 vha-a3 (pH 6.28) mutant, lacking the tonoplast V-ATPase, showed a significant increase in vacuolar pH compared with the wild-type (pH 5.77) (Figure 6B).
Reactive oxygen species such as H2O2 function as signalling molecules in the plant cell. Their implication on e.g. development and response to abiotic stresses might be partially mediated by modulating the energization of endomembranes in plants. The V-ATPase represents a highly abundant proton pump at the plant endomembranes and adjusting V-ATPase activity to the metabolic needs in dependence on intracellular conditions and environmental stimuli is a central requirement for cellular homoeostasis. Therefore the capability of the plant V-ATPase to be regulated by post-translational modulation of its redox state has been analysed. Oxidative inhibition of the V-ATPase is a common feature of all eukaryotic V-ATPases and is assumed to involve three conserved cysteine residues: Cys261, Cys284 and Cys539, that were initially identified in bovine brain and yeast VHA-A. Interestingly, Cys261 is replaced by a serine residue in the anaerobic Archaebacteria P. horikoshii and T. thermophilus (Figure 1). This might point to its importance under aerobic conditions. In Methanosarcina mazei Gö1, even Cys539 is replaced by a serine residue, so that only Cys284 is conserved in all kingdoms. This is in agreement with our finding that expression of functional subunit A is enabled in the absence of Cys256 and Cys535. Even Cys279 is not strictly required for VHA-A, however, the reduced growth and V-ATPase activity of plants expressing VHA-A C279S indicates that it is important for full functionality. Its counterpart in yeast has been shown to be important for correct folding or for the stability of VHA-A , partial unfolding or reduced stability might explain the general reduction of V-ATPase activity in the C279S mutant line in A. thaliana. This would explain why Cys279 is highly conserved in all kingdoms, whereas Cys256 and Cys535 can be substituted by a serine residue without an effect on activity.
On the basis of biochemical analysis, including labelling of cysteine-reactive cysteine residues of subunit A with fluorescein-maleimide, redox regulation was suggested to occur by intramolecular disulfide formation between Cys261 and Cys539, whereas the active state was proposed to involve a disulfide bond between Cys284 and Cys539 [11,14]. The analysis of the substitutions C284S and C539S of the VHA-A subunit VMA1p in yeast failed to support this hypothesis, as both proteins do not complement the Δvma1 phenotype. Both cysteine residues are required for proteolytic processing of the Vma1p-precursor in yeast and hence are of structural importance for VMA1p, and mutation leads to destabilization of VMA1p [16,36].
The distance between Ser232 (corresponds to Cys256 in A. thaliana) and Cys539 (corresponds to Cys535 in A. thaliana) of subunit A of the T. thermophilus A-ATPase is 26 Å , but the distance between Cys254 and Cys532 was postulated to be 5–6 Å for the bovine enyzme . The β-subunit of the related F-ATPase unoccupied by nucleotide shows a 20 Å movement of the C-terminal domain away from the glycine-rich loop relative to the ATP and ADP bound forms, and this change has been suggested to be part of the catalytic cycle. If the V-ATPase A subunit undergoes a similar change, disulfide bond formation would lock the A subunit in a closed conformation, which is typical in the absence of nucleotides and blocks nucleotide binding, thus inhibiting activity . However, inhibitory intramolecular disulfide formation within VHA-A was not observed in A. thaliana in the present study. The electrophoretic mobility of VHA-A from A. thaliana was not altered in the reduced state, so that a conformational alteration of VHA-A as caused by disulfide bridge formation appears unlikely. This finding is in contrast with the observed electrophoretic mobility of the insect VHA-A and small angle X-ray scattering data with isolated insect V-ATPase that indeed indicated structural alterations of the head domain under non-reducing conditions. Also in the present study, tryptic digests of the reduced and oxidized state revealed alterations of e.g. VHA-B, but gave no hints for redox-dependent alterations of VHA-A . It is noticeable that ATP hydrolysis of isolated V-ATPases and proton transport of isolated vacuoles is sensitive to oxidizing conditions, whereas the vacuolar pH of intact root cells was unaffected. As it has been shown that the glutathione pool is fully oxidized under these conditions , we conclude that, at least in Arabidopsis roots, V-ATPase activity is not directly coupled to the cytosolic redox potential. This is in contrast with the yeast V-ATPase, where vacuolar acidification is impaired in the glutathione-deficient mutant vma41-1 . Dietz et al.  discussed the role of thioredoxins rather than glutathione in maintaining the V-ATPase in the reduced state in plants. On the other hand, the peroxidase activity of plasma membrane-associated annexins has been suggested to be involved in membrane protection against oxidative stress in maize . Since annexins bind to phosphatidylserine and phosphatidylcholine and both phospholipids are also abundant at the tonoplast, such membrane-associated peroxidases might be candidates for H2O2 detoxification at the tonoplast as well [40,41]. However, the V-ATPase is efficiently protected against oxidation in the plant cell and oxidative inhibition of the V-ATPase might not generally contribute to V-ATPase regulation in vivo. Furthermore, H2O2 is known as trigger for programmed cell death , and the destructive vacuole-mediated programmed cell death relies on acidification of the vacuole. This process is of importance for the formation of tracheary elements or for virus defence . Therefore, H2O2 sensitivity of the V-ATPase would have a severe impact on plant development and virus defence. Thus H2O2-mediated activation of V-ATPase might occur in order to support programmed cell death. H2O2-induced activation of ion transport that involves a single cysteine residue has been reported for the K+ channel SKOR . However, neither enhancement nor inhibition of V-ATPase by H2O2 has been observed in vivo and the in vitro data revealed basal sensitivity of V-ATPase towards oxidative conditions.
In summary: (i) the oxidative inhibition of ATP hydrolysis and proton transport was mostly prevented by the C256S substitution in VHA-A, demonstrating that intramolecular disulfide formation is not involved in redox regulation of the V-ATPase; (ii) Cys279 is required for activity, most likely by its involvement in folding or stability; and (iii) oxidative conditions did not affect the vacuolar acidification in root cells, indicating efficient protection of V-ATPase in vivo. Taken together, these results suggest that oxidative inhibition of the plant V-ATPase involves Cys256 of VHA-A exclusively, and seems to have less importance for regulation of the plant V-ATPase.
We have shown in the present study that oxidative inhibition of the plant V-ATPase does not involve intramolecular disulfide-bond formation within VHA-A as has been proposed for both the yeast and mammalian enzymes. Instead, our data revealed on the one hand that, in plants, modulation of VHA-A Cys256 is sufficient for inhibition and on the other hand that oxidative inhibition of the V-ATPase does not occur in Arabidopsis roots. Physiological analysis of plants expressing the C256S mutation will now allow us to determine the functional relevance of this biochemical mechanism.
Thorsten Seidel performed cross-linking, bioinformatical analyses and co-wrote the paper. Stefan Scholl and Melanie Krebs established and characterized all transgenic lines, and performed the pH measurements. Miriam Hanitzsch measured the ATP hydrolysis. Florian Rienmüller, Irene Marten and Rainer Hedrich did the patch-clamp experiments. Patricia Janetzki performed the redox shift assays of proteins and participated in cross-linking. Karl-Josef Dietz co-wrote the paper. Karin Schumacher conceived and designed the project and co-wrote the paper.
We thank members of the Seidel and Schumacher groups for technical support, helpful discussions and critical reading of the paper prior to submission.
The work was supported by the Deutsche Forschungsgemeinschaft [grant numbers SFB 613, TP A5 (to T.S. and K.-J.D.) and SPP1108 (to K.S.)].