OUP user menu

Dependence of Saccharomyces cerevisiae Golgi functions on V-ATPase activity

Isaac Corbacho, Francisco Teixidó, Isabel Olivero, Luis M. Hernández
DOI: http://dx.doi.org/10.1111/j.1567-1364.2011.00784.x 341-350 First published online: 1 May 2012


The V-ATPase of Saccharomyces cerevisiae is an ATP-dependent proton pump responsible for acidification of the vacuole and other internal compartments including the whole secretory pathway. We have studied the behavior of several glycoprotein processing reactions occurring in different Golgi compartments of representative vmaΔ mutants. We found that outer chain initiation is not altered in the mutants while mannosylphosphate transfer, α(1, 3)-linked mannoses addition, and α factor maturation seem to be affected. The results suggest a gradation in the dependence of Golgi functions on V-ATPase activity, from early Golgi (unaffected) to late Golgi (significantly reduced). These findings are in agreement with the internal pH of Golgi cisternae measured in mammalian cells, which is more acidic in the late region. The mutant defects can be partially restored by buffering the external medium to pH 6.0, which supports the existence of a mechanism that, in the absence of a functional V-ATPase, could contribute to pH regulation at least in the late Golgi.

  • Saccharomyces cerevisiae
  • V-ATPase
  • VPH1
  • STV1
  • mannosylphosphorylation
  • Golgi pH regulation


Most intracellular processes of eukaryotic cells are dependent on the pH. Enzymes working inside the cell have a relatively narrow pH range, which implies that the pH maintenance in the cytosol and organelles is crucial for a variety of cell functions, including membrane trafficking, receptor-mediated endocytosis, and protein synthesis, processing, transport, and degradation (Kane, 2006; Forgac, 2007; Martinez-Munoz & Kane, 2008). V-ATPases play a central role in this maintenance. Saccharomyces cerevisiae has a multisubunit membrane-bound V-ATPase composed of two domains: V0 and V1. V0 is embedded in the membrane and comprises six different subunits named a, c, c′, c′′, d, and e, while V1 is water soluble and comprises eight more protein subunits named A, B, (C), D, E, F, G, and H (Zhang et al., 2008). V-ATPase was initially described in the vacuolar membrane of S. cerevisiae (Kakinuma et al., 1981; Klionsky et al., 1990), but was later shown to be present in Golgi membranes. The location in vacuole or Golgi membranes was related to the existence of two isoforms of subunit ‘a’ encoded by STV1 and VPH1. The Vph1p-containing complexes were located in the vacuole, while the ones functioning in the Golgi had Stv1p instead (Manolson et al., 1994; Kawasaki-Nishi et al., 2001b). A similar subunit-dependent location has also been described in mammalian cells (Forgac, 2007) and in plant cells (Seidel et al., 2008). However, it has been shown as well that the distribution is not permanent; rather, it can change in some cells under particular conditions (Toyomura et al., 2003). Actually, the fact that deletion of either STV1 or VPH1 resulted in a wild-type phenotype in some cases, suggested that both gene products could be present in any of the locations. In a recent study (Samarao et al., 2009), it has been reported that purified membranes of ER, Golgi, and vacuole of S. cerevisiae endow V-ATPase activities with different properties depending on the location. In a previous work, we found that transfer of mannosylphosphate groups to N-linked oligosaccharides was dependent on the V-ATPase activity, because single-deletion mutants on V-ATPase subunits showed a low dye-binding phenotype (Corbacho et al., 2005). This process is assigned to a late Golgi region. In this work, we checked the dependence of several processes related to protein secretion/glycosylation, which have been assigned to distinct Golgi regions, on the V-ATPase activity. The study adds one more phenotype to the long list of defects previously described for vma mutants (Kane, 2007). In addition, we studied in representative vmaΔ mutants the influence of changes in extracellular pH on several Golgi-located processes, as well as on detoxification by methylene blue.

Materials and methods

Strains and growth conditions

Saccharomyces cerevisiae: wild-type strains BY4741 (MATa, his3Δ1, leuΔ0, met15Δ0, and ura3Δ0) and BY4742 (MATα; his3Δ1; leuΔ0; lys2Δ0; ura3Δ0), and single deletants (using kanMX module) in both backgrounds for genes VMA1 (YDL185W), VMA4 (YOR332W), VMA5 (YKL080W), VMA6 (YLR447C), VMA7 (YGR020C), VMA8 (YEL051W), VMA9 (YCL005W-A), VMA10 (YHR039C-A), VMA11 (YPL234C), VMA12 (YKL119C), VMA13 (YPR036W), VMA16 (YHR026W), VMA21 (YGR105W), VMA22 (YGR155W), STV1 (YMR054W), and VPH1 (YOR270C) were obtained from EUROSCARF (European Saccharomyces cerevisiae Archive for Functional Analysis).

Saccharomyces cerevisiae bar1Δ MATa was also from EUROSCARF.

The double deletant stv1Δ vph1Δ, as well as vma2Δ (YBR127C) and vma3Δ (YEL027W), was constructed for this study. The reason for construction of the single deletants vma2Δ and vma3Δ was to guarantee that the mutations are in a clean genetic background because in a previous study (Corbacho et al., 2005), we found additional defects in a commercial single deletion strain. For stv1Δ vph1Δ, in a stv1Δ strain (BY4742 MATα his3Δ1; leu2Δ0 lys2Δ0 ura3Δ0 YMR054W::kanMX4), a cassette replacement of VPH1 gene with hphMX was made (Goldstein & McCusker, 1999). hphMX cassette was amplified from plasmid pAG32 (EUROSCARF) using the forward primer 5′TGAATCAAAAAAAAAACATTTAAAGGT TACACAAGGAAAATAATGCGTACGCTG CAGGTCGAC and the reverse primer 5′GAAGTACTTAAATGT TTCGCT TTTTTTAAAAGT CCTCAAAAT TTAATCGATGAATTCGAGCTCG (sequences for VPH1 targeting are underlined).

hphMX cassette was PCR-amplified in a 50-μL reaction using PCR Extender System (5 PRIME), forward and reverse primers, 100 ng of plasmid pAG32 as template, 0.2 mM dNTPs, and 5 U of polymerase. Amplification of the cassettes was initiated with a 2-min 94 °C denaturation, followed by 30 amplification cycles: 94 °C for 30 s, 65 °C for 30 s, and 72 °C for 6 min.

About 2 μg of PCR product was used to transform the S. cerevisiae stv1Δ strain using the lithium acetate method (Gietz & Woods, 2003). Cells were grown overnight in YPD at 30 °C on a rotator before plating transformants onto selective media (YPD supplemented with 200 μg mL−1 G418 and 300 μg mL−1 hygromycin B) (Goldstein & McCusker, 1999). Colonies were re-grown twice more in selective media. Homologous integration of dominant drug resistance cassettes was verified by PCR from genomic DNA. For these PCR reactions, a direct primer was designed upstream of VPH1, and reverse primers were inside either the VPH1 or the hphMX sequence.

For construction of vma2Δ and vma3Δ, a similar protocol was followed, using the wild-type BY4741 as parental strain. In all cases, transformation of deleted strains with the corresponding wild-type gene resulted in a wild-type phenotype.

Unless other conditions are specified, cells were grown at 30 °C in solid or liquid YPD (1% yeast extract, 2% peptone, and 2% d-glucose). When needed, YPD was buffered at different pH values, using either 200 mM citrate-phosphate buffer (for the pH range 5–5.5) or 200 mM phosphate buffer (for the pH range 6–7). For the methylene blue tests, 0.003% of methylene blue was added to standard solid YPD, and the plates are inoculated with 10 μL of a culture previously grown on standard liquid YPD to an OD of 2.

Glycosylation analysis methods

The staining of the yeast cells with alcian blue (Sigma) was performed essentially by the method described by Ballou (Ballou, 1990) and later optimized (Mañas et al., 1997; Corbacho et al., 2005). In brief, cells were grown in 200 μL of YPD in 96-well microtitre plates. After inoculation, the plates were incubated for 48 h at 30 °C, which will favor that all samples will end up with a similar cell mass (Corbacho et al., 2005), and then kept at 4 °C for 48 h more. For staining, the plates were centrifuged for 5 min at 1000 g in an A-2-MTP rotor (Eppendorf), and the culture medium was discarded. The cells in the pellet were washed with 200 μL of 50 mM acetic acid. After a further centrifugation, the pellets were resuspended in 200 μL of a solution of 0.1% alcian blue in 50 mM acetic acid, centrifuged again, and washed twice with 50 mM acetic acid without removing the supernatant the second time. The blue color on the pellet of each well was quantitated as described previously (Corbacho et al., 2005). The plates were scanned from the bottom, the images were stored, and the intensity was quantitated on the computer screen visually and, when necessary, with the aid of Image J, a public domain image analysis software package downloaded from the NIH website (Rasband, W.S., ImageJ, US. National Institutes of Health, Bethesda, Maryland, USA, http://rsb.info.nih.gov/ij/, 1997–2005). The low dye-binding phenotype (ldb) shown by some S. cerevisiae strains reflects a decrease in affinity by the alcian blue dye because of defects in N-linked oligosaccharides phosphorylation (Mañas et al., 1997).

Native gel electrophoresis of the external invertase migration pattern was carried out as described previously (Ballou, 1990; Mañas et al., 1997). As previously indicated for alcian blue staining, the growth of the strains until stationary phase in the same volume of medium will also favor that all samples will end up with a similar cell mass. Specific antibodies against α(1, 3)-linked mannoses were obtained in our laboratory for a previous study (Olivero et al., 2001).

Vacuole vital staining

For quinacrine staining, exponentially growing cells (5 × 106) were harvested and resuspended in phosphate buffer 50 mM pH 7.6 containing 2% glucose, 100 mM quinacrine (Sigma), and supplemented with 20 mM bafilomycin (Sigma) when indicated (Umemoto et al., 1990). Cells were incubated for 15 min at 30 °C, resuspended in phosphate buffer pH 7.6, and examined with a Nikon Eclipse E600 microscope, equipped with fluorescence and a Zeiss AxioCam digital camera.

For the neutral red staining, exponential phase cells harvested as earlier and resuspended in 25 μL of YPD pH 7 were stained with 5 μL of neutral red (Sigma) (2 mg mL). Cells were incubated for 5 min at room temperature and observed by phase contrast microscopy.

α; factor production assay

For α factor production assays, a modification of a previous method (Sprague, 1991) was followed. The MATα strains to be tested were grown in buffered and nonbuffered liquid YPD. Then, 108 cells of each strain were harvested, washed with 1 mL of the same fresh medium, and resuspended again in 10 μL of the same medium. Finally, the cell suspensions were spotted onto buffered and nonbuffered YPD plates and incubated for 48 h at 20 °C. The plates were previously inoculated with a lawn of 5 × 105 cells of the α factor-sensitive strain S. cerevisiae bar1Δ MATa. In addition, a ‘quantitative mating assay’ was also used (Sprague, 1997).

Statistical analysis

Data of blue color intensity (Figs 5–7) were summarized in mean ± SD. Comparisons of the variables between pH levels were performed by using Kruskal–Wallis rank-sum test with the Dunnet's method for multiple comparisons with pH 6 group. The analysis was performed using the IBM-SPSS Statistics 19 (SPSS Inc. IBM Company). The level of statistical significance was set at P-value ≤ 0.05 (two tailed).

Results and discussion

Effect of V-ATPase activity on late Golgi processes: mannosylphosphorylation of N-linked oligosaccharides and transfer of α;(1, 3)-linked terminal mannoses are dependent on V-ATPase activity

In a previous study (Corbacho et al., 2005), we found that single-deletion mutants of S. cerevisiae, in genes encoding subunits of V-ATPase, showed ldb phenotype when growing in standard YPD medium, with the exception of stv1Δ and vph1Δ which behaved very close to the wild type. The ldb phenotype is a consequence of a reduction in the amount of mannosylphosphate groups attached to N-linked oligosaccharides (Mañas et al., 1997). According to these results, we suggested that mannosylphosphate transfer to N-linked oligosaccharides, a process that occurs in a late Golgi region (Jungmann et al., 1999; Munro, 2001), might be dependent on V-ATPase activity.

To further study this finding, we used the macrolide antibiotic bafilomycin, a potent specific inhibitor of V-ATPases (Bowman et al., 1988; Dröse & Altendorf, 1997). Wild-type cells treated with the antibiotic were checked for ldb phenotype, and the results were compared to representative vma mutants: vma2Δ and vma3Δ. VMA2 encodes subunit B of the peripheral V1 domain, while VMA3 encodes subunit c of the transmembrane V0 domain. As seen in Fig. 1a, the effect of bafilomycin on wild type results in an ldb phenotype very close to that shown by both vmaΔ mutants. Wild type and mnn6Δ have been included as controls of known behavior for dye binding (Corbacho et al., 2005). These results confirm that the transfer of mannosylphosphate groups to glycoprotein-linked oligosaccharides is dependent on the activity of the V-ATPase in S. cerevisiae. In addition, Fig. 1 also shows that vacuole acidification in the wild type was blocked by bafilomycin in the same way as in the vmaΔ mutants grown without the antibiotic, as demonstrated by the lack of staining with quinacrine and neutral red (d–g). Untreated wild type shows intense staining with both dyes (b and c). Although similar results to these shown in b–g have been reported previously, we considered interesting to include them in the figure as controls for known behavior of vmaΔ strains.


Alcian blue staining of vma2Δ, vma3Δ, and wild type (BY4741) + 20 mM bafilomycin. Controls of known alcian blue affinity: wild type, mnn9Δ, mnn6Δ (a). Quinacrine staining of wild type (b); wild type + 20 mM bafilomycin (d) and vma3Δ (f). Neutral red staining of wild type (c); wild type + 20 mM bafilomycin (e) and vma3Δ (g).

To verify whether the effect of V-ATPase activity on mannosylphosphorylation was specific for the transfer of mannosylphosphate groups, or just a consequence of a general malfunction of the Golgi apparatus, we also checked the transfer of terminal α(1, 3)-linked mannoses to the glycoprotein-linked oligosaccharides. This reaction is catalyzed by Mnn1p (Ballou, 1990), and it also occurs in a middle-late Golgi region (Yip et al., 1994; Munro, 2001). The presence of terminal α(1, 3)-linked mannoses on oligosaccharides of the different strains was examined by checking the agglutination with specific antibodies against these residues. The results are shown in Fig. 2. All wild-type cells were precipitated with the antibodies in an agglutination assay, while the control strain mnn1Δ, which lacks α(1, 3)-linked mannoses, did not show any agglutination, and the cells remained in the supernatant. The vma2Δ and vma3Δ mutants showed an intermediate behavior because around 25% of the cells remained in the supernatant in the same agglutination test. The results suggest that inactivation of V-ATPase causes a pH deregulation in the Golgi, affecting glycosylating reactions that occur in that location.


Agglutination assay of indicated strains with anti-α(1, 3) antibodies.

Dependence of outer chain elongation on V-ATPase activity

Outer chain building in S. cerevisiae starts by the addition of a single α (1, 6)-linked mannose to the ‘inner core’ which was built in the ER. The transfer is catalyzed by an enzyme encoded by OCH1 and localized in cis-Golgi (Nakayama et al., 1992). None of the vmaΔ strains showed an invertase size in agreement with a complete lack of the outer chain in the N-linked oligosaccharides (Corbacho et al., 2005) (see och1Δ in Fig. 3). The results suggest that the activity of Och1p is not significantly dependent on V-ATPase activity.


Migration pattern in native gel electrophoresis of external invertase produced by representative vmaΔ strains. Wild type (BY4741) and och1Δ were used as controls.

Elongation of the outer chain occurs in two sequential steps catalyzed by the protein complexes mannan polymerases I and II (M-Pol I and M-Pol II) (Jungmann & Munro, 1998; Munro, 2001). They add the α (1, 6)-linked backbone to the first mannose transferred by Och1p, so they are localized on a cis-middle region of the Golgi apparatus. Because the product of M-Pol I is the substrate for M-Pol II, probably M-Pol I is located mainly in the cis region, while M-Pol II acts more displaced to the middle Golgi. As stated earlier, Fig. 3 shows that none of the invertase molecules synthesized by vmaΔ strains exhibit a migration pattern similar to och1Δ, which suggests that activity of M-Pol I is not affected in the vmaΔ mutants. We chose strains deleted in genes that encode subunits of the V0 or V1 domains. Some strains show a slightly but detectable reduction in invertase size or increase in heterogeneity, when compared to the wild type, probably due to a partial affectation of M-Pol II activity. The minor differences observed between some vma strains are difficult to understand because all of them are deletion strains; however, they were found in several repetitions of the experiment. The results suggest that V-ATPase dependence of Golgi functions is not evenly distributed through all the Golgi stacks, but rather shows a gradation from early (unaffected) to late (affected) Golgi regions. It has been described that the steady-state pH of Golgi in mammalian cells varies from 6.7 in early regions to 6.0 in late regions [reviewed in (Paroutis et al., 2004)]. These differences could explain why the late regions are more dependent on the acidification by V-ATPase activity than the early ones in S. cerevisiae and point to a similar pH gradation in the yeast's Golgi.

α; Factor maturation

The Golgi functions checked above all correspond to activity of transferases that participate in the building of the outer chain of N-linked oligosaccharides. To discard a specific effect of V-ATPase on the glycosyl transferases involved in the glycosylation pathway, we also tested the α factor maturation, another Golgi-located activity. The α factor in S. cerevisiae is encoded by MF(α)1 and MF(α)2 genes. It is synthesized as an inactive pre-pro-protein that is glycosylated and processed through the secretory pathway (Julius et al., 1983, 1984). The last steps in processing are proteolytic cleavages by the sequential action of Kex2p, Kex1p, and Ste13p proteases (Julius et al., 1983, 1984; Dmochowska et al., 1987; Manolson et al., 1994; Perzov et al., 2002), which occurs late in the secretory pathway. To check for the synthesis and secretion of active α factor, we used a method described previously (Sprague, 1991). The assay is based on a halo test. A drop of a cell suspension of the tested strains was deposited on a plate inoculated with a tester strain especially sensitive to the pheromone (bar1Δ MATa; see ), and the halos were observed.

Figure 4 shows that wild-type cells produce active α factor that is responsible for the clear halo of inhibition of bar1Δ growth. In the kex2Δ mutant, the halo is not observed at all, as expected for a strain that does not secrete active α factor. The vma2Δ and vma3Δ strains show an intermediate behavior. The halo produced by these strains is smaller and more diffuse than the one produced by wild type. The experiment was repeated four times with similar results. The results confirm that pH maintenance by V-ATPase activity is required for a general Golgi function. However, as is the case with the above shown glycosylating enzymes, their effect is not very marked probably due to an alternative mechanism for pH maintenance in Golgi.


α Factor secretion test. The patches of the indicated α strains were inoculated on a YPD plate previously spread with a sensitive strain bar1Δ MATa.

To further support these results, the mating efficiency of vma2Δ was compared to that of wild type by using a previously described quantitative mating assay (Sprague, 1991). We found a mating efficiency of 0.50 for wild type and 0.35 for vma2Δ, which is also in agreement with the previous suggestion that the malfunction of Golgi apparatus in the mutant affects α factor processing. Values are mean of five different experiments (standard deviation of 0.055 and 0.057, respectively).

Effect of extracellular pH on Golgi function

Several previous studies have provided evidence for the influence of extracellular pH on the growth rate of S. cerevisiae mutants defective in V-ATPase activity (Nelson & Nelson, 1990; Manolson et al., 1994; Perzov et al., 2002). Nelson and Nelson (Nelson & Nelson, 1990) showed that double disruptant strains on genes encoding subunits of V-ATPase were unable to grow at pH higher than 6.5. They suggested that a lower pH in the culture medium is an alternative mechanism for the acidification of internal organelles by fluid phase endocytosis, which allows survival of the cells. Manolson and co-workers (Manolson et al., 1994) found that single-deletion mutants in several VMA genes were unable to grow in media with 100 mM CaCl2 or 4 mM ZnCl2, media with glycerol as carbon source, or media buffered to pH 7.5. In the case of VPH1 and STV1, single-deletion mutants behaved as the wild type, while the double deletant showed the mutant phenotype. Similar conclusions have been reached by others (Perzov et al., 2002). In this work, we have investigated the behavior of representative vmaΔ mutants growing in media buffered at different pHs. We checked the influence of extracellular pH on several phenotype characteristics: mannosylphosphate transfer, detoxification of methylene blue, α(1, 3)-linked mannoses addition, and α factor maturation.

Figure 5 shows that mannosylphosphorylation of wild type and mutants mnn9Δ and mnn6Δ is independent of external pH, because no appreciable changes in alcian blue affinity are observed. The strains vma2Δ and vma3Δ show a different behavior. In nonbuffered YPD, they show a weak ldb phenotype that is roughly maintained at other pHs, with the exception of pH 6. At this pH, the affinity of the mutant cells for alcian blue increases (see Fig 5b), which suggests that external pH has some influence on mannosylphosphate transfer. Data of blue color intensity from strains vma2Δ and vma3Δ were analyzed by using the Kruskal–Wallis rank-sum test (with the Dunnet's method for multiple comparisons with pH 6 group). In both cases, the results were statistically significant with P-value < 0.0001.


(a) Alcian blue staining of the indicated strains grown in YPD media buffered at different pHs. (b) Quantitation of blue color intensity. Values represented in (b) are obtained from eight different measurements (mean ± SD). The scales indicate relative intensity.

As stated before, the steady-state pH of Golgi varies from cis to trans regions, with values ranging from 6.7 to 6.0 in mammalian cells (Paroutis et al., 2004). In S. cerevisiae, the values must be very close to these because an external pH of 6.0 could contribute to having a pH in Golgi very close to the steady state in the trans region, even in the presence of a mutant V-ATPase. Martinez-Munoz & Kane (2008) carried out a complete study of internal pH changes in wild-type S. cerevisiae and some vma mutants, in response to glucose and KCl additions. They found that glucose addition results in an increase in vacuolar pH in vma2 and vma3 mutants and wild-type cells treated with the V-ATPase inhibitor concanamycin A. This response was diminished by fixing the external pH to 5, which suggests that, at least in the absence of an active V-ATPase, external pH clearly contributes to internal pH homeostasis. As mannosylphosphorylation occurs in a middle-late Golgi region, at an external pH of 6, the transferase would find a working pH quite close to the wild type. In addition, we found that lowering the external pH to 5, the late Golgi functions related to N-glycosylation are not restored in the mutants, probably because the resulting Golgi pH was slightly different (probably lower) than in the wild type. These findings point to a different Golgi behavior when compared to previous data for cytoplasm or vacuole (Padilla-Lopez & Pearce, 2006; Martinez-Munoz & Kane, 2008).

Methylene blue is a dye that has been used to distinguish dead and living yeast cells. It is assumed that, at certain concentrations, the dye is excluded by living cell membranes, while it penetrates and stains dead cells (Kucsera et al., 2000). We found that addition of methylene blue to YPD at a concentration of 0.003% allowed us to stain colonies from vma2Δ mutants in a pH-dependent manner, while wild-type colonies remained unstained independently of the pH. As in the previously described alcian blue staining, we found that at pH 6, the behavior of vma mutants was closer to the wild type than at any other pH (not shown). This experiment also suggests that external pH contributes to pH homeostasis inside the cell and that pH 6 is the most appropriate to favor the detoxification mechanism. At other pHs, the detoxification mechanism does not work properly, and the cells finally dye and become stained.

The same experiment on external pH dependence was also carried out by checking α(1, 3)-linked mannoses addition and α factor maturation. We were able to detect significant differences in the phenotype of vma2Δ and vma3Δ grown on YPD buffered to pH 6. As is shown above for alcian blue staining and methylene blue detoxification, in both characteristics tested, the phenotype was closer to the wild type (see Figs 2 and 4) at pH 6 (results not shown).

Behavior of vph1Δ, stv1Δ, and the double mutant vph1Δ stv1Δ

It is generally assumed that subunit ‘a’ of V-ATPase is involved in targeting the enzyme to different subcellular locations in different organisms (Forgac, 2007; Seidel et al., 2008). In S. cerevisiae, VPH1 and STV1 encode two isoforms of the ‘a’ subunit, which are responsible for the location in the vacuolar or Golgi membranes, respectively (Manolson et al., 1994; Kawasaki-Nishi et al., 2001a). It has been shown that deletion of either VPH1 or STV1 results in a wild-type phenotype for some cell functions (Perzov et al., 2002; Corbacho et al., 2005), suggesting that in the absence of one of the isoforms of subunit ‘a’, the other can replace it. However, it has also been reported that VPH1 is required for vacuolar acidification (Manolson et al., 1992) and that quinacrine uptake into the vacuole is only partially restored when STV1 is overexpressed on a vph1Δ mutant (Manolson et al., 1994).

Here, we have analyzed the behavior of vph1Δ, stv1Δ, and the double mutant vph1Δ stv1Δ regarding the same Golgi functions stated earlier: mannosylphosphate transfer, α(1, 3)-linked mannoses addition, and α factor maturation as well as methylene blue detoxification, at different pHs, to determine the contribution of both isoforms to pH maintenance in the vacuole and in several Golgi regions. In addition, we used neutral red staining to determine the vacuole acidification on the mentioned strains.

Figure 6 shows the results obtained for methylene blue detoxification (a and b) or alcian blue staining (c and d). Results were similar in both tests. The single mutants stv1Δ and vph1Δ behave as the wild type independently of the pH, while the double mutant stv1Δ vph1Δ exhibits a mutant phenotype, with blue colonies in methylene blue-containing YPD (a) or a moderated ldb phenotype (c). In both cases, a certain complementation of the mutant phenotype can be seen in the double deletant when the media were buffered to pH 6. As in previous Fig. 5, the different blue color intensity at pH 6 in the double mutant stv1Δ vph1Δ was tested by the Kruskal–Wallis rank-sum test with the Dunnet's method. In both cases (b and d), the results were statistically significant with P-value < 0.0001. The same behavior was observed when α(1, 3)-linked mannose addition and α factor maturation were checked (results not shown). These results definitely demonstrate that acidification of Golgi in stv1Δ strains can be achieved by Vph1p-containing V-ATPase, which is usually located in the vacuolar membrane. Also, they support the previous observation that an external pH 6 partially compensates the mutant phenotypes caused by a malfunction of the Golgi apparatus.


pH dependence of methylene blue detoxification (a and b) and alcian blue staining (c and d) of the indicated strains. (b and d) show the quantitation of blue color intensity. Values represented in (b) and (d) are obtained from eight different measurements (mean ± SD). The scales indicate relative intensity.

The function of the vacuolar-located V-ATPase can be checked by monitoring quinacrine uptake into the vacuole or by neutral red staining of the vacuole (see Fig. 1). In a previous study (Manolson et al., 1994), it was shown that vph1Δ was unable to transport quinacrine into the vacuole and that this function was only partially restored when STV1 was overexpressed. Our results with neutral red staining corroborate that vph1Δ and the double deletant stv1Δ vph1Δ do not acidify the vacuole while stv1Δ does (Fig. 7).


Vacuolar staining of indicated strains with neutral red. (a) Wild type; (b) vph1Δ stv1Δ; (c) vph1Δ; (d) stv1Δ.

The results in this section confirm that the two isoforms of subunit ‘a’ of the V-ATPase are not totally interchangeable. While Vph1p seems to replace Stv1p in stv1Δ, giving rise to a functional Golgi-located V-ATPase, this does not happen in the contrary situation. Stv1p does not make a functional vacuolar-located V-ATPase in vph1Δ, because the lack of staining by neutral red demonstrates that the vacuole is not acidified in this mutant (Fig. 7b). Only when STV1 is overexpressed do some of the Stv1p-containing V-ATPase complexes localize in the vacuolar membrane. The reason for this behavior is unknown, but one can speculate about a possible explanation. It is generally accepted that V-ATPase is assembled in the ER and then travels to its final destinations which depend on the ‘a’ subunit isoform. The contribution of each isoform-containing V-ATPase to acidification of a particular compartment is determined by its time of residence (reviewed in Kane, 2006 and Forgac, 2007). Bearing this in mind, the Vph1p-containing V-ATPase complex must travel from the ER to the vacuolar membrane through the secretory pathway. This implies that it must go through the whole Golgi apparatus and that, during the transit, it will contribute to Golgi acidification, although to a lesser extent than Stv1p-containing complexes. In stv1Δ mutants, the contribution of the Vph1p-containing ATPase to Golgi acidification would probably be enough to maintain the physiological pH, resulting in a wild-type phenotype. In contrast, the Stv1p-containing V-ATPase travels from ER to Golgi membranes, but the vacuole is not included in the pathway. That would explain the mutant phenotype in the vph1Δ mutants. They are not able to acidify the vacuole because they do not have any V-ATPase in the vacuolar membrane. Overexpression of STV1 could result in a saturation of the regular pathway and thus in a relocation of some of the Stv1p-containing V-ATPase complexes to the vacuolar membrane. Those relocated V-ATPase complexes might be responsible for the partial restoration of the vacuolar pH.

The findings presented in this work complement and extend the study of Samarao and co-workers (Samarao et al., 2009) on the function of V-ATPase along the secretory pathway. We found that dependence of Golgi functions on V-ATPase activity increases from early (unaffected) to late (affected) regions and that the external pH can partially compensate the lack of an active V-ATPase in Golgi.


This work was supported by Grant PRI07A087 co-financed by the Junta de Extremadura and FEDER. We thank Dr M. González Velasco for his help with statistical analysis. I.C. and F.T. are recipients of post- and predoctoral studentships from the Junta de Extremadura. We thank Maria Rico for technical assistance.


  • Editor: Guenther Daum


View Abstract