In Saccharomyces cerevisiae, the cell integrity pathway plays a role in the oxidative stress response. In this study, we show that the Pkc1 protein mediates oxidative signalling by helping to downregulate ribosomal gene expression when cells are exposed to hydrogen peroxide. An active actin cytoskeleton is required for this function, because the cells blocked in actin polymerisation were unable to repress ribosomal gene transcription. Following the invertase secretion pattern, we hypothesize that oxidative stress induced by hydrogen peroxide could have affected the latter steps of secretion. This would explain why the Pkc1 function was required to repress ribosomal biogenesis.
We have recently demonstrated that oxidative stress induces transient and dosage-dependent ribosomal gene repression upon treatment with hydrogen peroxide (H2O2) (Petkova et al., 2010). Other types of environmental stress are also known to regulate ribosomal gene transcription (Mizuta & Warner, 1994). The cells affected in the secretory pathway require the Pkc1 function to repress ribosomal gene expression (Nierras & Warner, 1999). Some cell wall receptors are believed to sense the increase in turgor pressure caused by secretion impairment and to transmit the signal to Pkc1 (Li et al., 2000). In response to oxidative stress, Mtl1, a cell wall receptor and component of the CWI (Cell Wall Integrity) or Pkc1-MAPk pathway, signals ribosomal gene repression through TOR1 and RAS2 inhibition (Petkova et al., 2010). Mtl1 is also required for cell survival in response to oxidative stress. According to Vilella et al. (2005), of all the elements integrating the Pkc1-MAPk pathway, only Rom2 and Pkc1 are essential for cell viability following treatment with different oxidants. Mtl1 also activates the Pkc1-MAPk pathway when cells are exposed to oxidants (Petkova et al., 2010). Pkc1 is a protein kinase C and the key element in the CWI pathway. Given the close connection between Mtl1 and Pkc1 in the oxidative stress response, we wondered whether the Pkc1 protein also had a role in the ribosomal gene repression that occurs in response to oxidative stress. To check this, we grew wild-type and pkc1 cells to log phase and treated both cultures with H2O2. As pkc1 is lethal in the absence of the cell wall stabilizer sorbitol, we added this reagent to the culture media to a final concentration of 0.8 M. The presence of sorbitol in the culture media conferred wild-type cells with more resistance to ribosomal gene repression in response to 1 mM H2O2 treatment. This is probably due to the fact that, to some extent, H2O2 affects the CWI (not shown), even though the main target for H2O2 is not necessarily the cell wall (Vilella et al., 2005; Petkova et al., 2010). We therefore increased the H2O2 concentration to 10 mM in SD medium containing sorbitol and the required amino acids, without affecting wild-type cell viability (not shown). We took samples at several times (as shown in Fig. 1) and processed them for subsequent northern blot analysis. We checked the RPSA, RPS28A and RPL30 ribosomal genes and also U1 as a loading control. In the presence of 0.8 M sorbitol and 10 mM H2O2, wild type did not exhibit any substantial loss of viability (not shown, cell viability was checked in plates by doing serial dilutions from exponentially growing cultures), whereas ribosomal gene transcription was dramatically repressed (Fig. 1a) similarly to that described in Petkova et al. (2010). However, in the case of pkc1 strain, repression was almost absent (Fig. 1), while cell viability was seriously affected (not shown); this effect was similarly to one previously reported (Vilella et al., 2005). These results demonstrate that Pkc1 is required to transmit the oxidative signal to ribosomal biogenesis in response to oxidative stress. Another of the functions also regulated by Pkc1 and affected by oxidative stress is the organisation of the actin cytoskeleton. Oxidative stress induces actin depolarisation and depolymerisation (Vilella et al., 2005; Pujol et al., 2009) upon H2O2 treatment. The actin cytoskeleton transiently depolarizes upon H2O2 treatment. We observed that under the experimental conditions applied in this study, actin depolarisation was both qualitatively and quantitatively identical to that previously described (not shown, and Vilella et al., 2005). This suggested that actin depolarisation and ribosomal gene repression upon oxidative treatment are two functions that occur simultaneously. To ascertain whether actin polymerisation activity influences ribosomal gene expression in the oxidative stress response, wild-type cells were treated with 150 μM latrunculin for 1 h to completely depolarize the actin cytoskeleton. H2O2 was then added to the cultures. Samples were taken and processed for Northern analyses. Figure 1a shows that latrunculin A did not affect the ribosomal gene expression compared to wild-type cells. However, wild-type cells treated with 150 μM latrunculin A and subsequently with 10 mM H2O2 were unable to repress the transcription of the ribosomal genes that were tested. In conclusion, actin polymerisation activity is required to repress ribosomal biogenesis under oxidative conditions. Three plausible possibilities could explain our findings: (i) Pkc1 signals ribosomal gene repression and actin polymerisation divergently, in response to oxidative stress; (ii) Pkc1 linearly signals actin polymerisation and then ribosomal gene expression and (iii) actin is an oxidative stress sensor (as proposed by Gourlay & Ayscough, 2005), which acts upstream of Pkc1 in the signalling pathway. This point deserves further research in future studies.
It is well known that arresting secretion causes ribosomal gene repression. We wondered whether oxidative stress could also provoke secretion impairment. To answer this question, we decided to check both the early and late stages of the secretory pathway. Carboxypeptidase Y is a vacuole protein that is transported from ER to the Golgi apparatus, and then via late endosomes to the vacuole. The vacuolar form of CPY is the mature form (M) that displays the greatest electrophoretic mobility, whereas CPY localisation to the Golgi apparatus renders an immature or unprocessed form (P) that displays less electrophoretic mobility. SEC18 has been reported to be required for the transport of carboxypeptidase Y through the yeast Golgi complex but not for the final delivery of CPY to the vacuole (Graham & Emr, 1991). We therefore used the sec18-1 mutant to test whether H2O2 could have affected secretion from ER to the Golgi and analysed the CPY protein. A blockade in the first steps of secretion affects CPY transport (Graham & Emr, 1991). We grew cultures of the sec18-1 thermosensitive mutant to exponential phase at 25 °C and then shifted them from 25 to 37 °C for 2 h to block secretion from ER to the Golgi. In Fig. 1b, we observe that in a sec18-1 mutant growing at 25 °C, anti-CPY detected a single band corresponding to the mature processed form of CPY. When the cultures were shifted from 25 to 38 °C for 2 h, anti-CPY detected two forms: a more abundant band, corresponding to the mature or processed form (M), and a less abundant band, corresponding to the unprocessed form (P). H2O2 treatment did not cause any defects in CPY processing, as can be observed in Fig. 1b. (We used increasing concentrations of H2O2, and the result did not change, data not shown.) We therefore concluded that H2O2 did not affect the first stages of secretion in Saccharomyces cerevisiae.
FM4-64 lipophilic styryl dye is a vital stain, which is used to follow bulk membrane internalisation and transport to the vacuole in yeast (Vida & Emr, 1995). These authors have described that whereas wild type showed significant vacuolar membrane staining, the sec1-1 thermosensitive mutant was impaired in vesicle to plasma membrane secretion at 37 °C, but not at 25 °C. We followed FM4-64 internalisation in exponentially growing wild-type cells that were both treated with H2O2 or untreated (Fig. 1c). Whereas in wild-type cells FM4-64 correctly stained vacuole membranes in exponential cultures, upon H2O2 treatment, this dye suffered impairment in both its capture and internalisation (Fig. 1c). These results further strengthen the hypothesis that oxidative stress affects the secretory pathway.
To check the late stages of secretion, we used anti-invertase to detect the glycosylated forms of the enzyme invertase. Invertase is an enzyme that catalyses the hydrolysis of sucrose into fructose and glucose. As depicted in Fig. 1d, blocking secretion, using the thermosensitive mutant sec1-1, induced a marked reduction in invertase secretion. When we treated cells with increasing concentrations of H2O2, we observed an equivalent reduction in invertase secretion. Budding yeast expresses two different forms of invertase (Gascon & Ottolenghi, 1967), both being encoded by SUC2. The biggest form that is of glycoprotein nature is located in the cell wall and is known as heavy invertase. The second type is a free carbohydrate form that is located and accumulates intracellularly in an unglycosylated form. This form is the small or light invertase, which is constitutively synthesized. However, the heavy chain is synthesized under conditions of catabolic derepression; it is highly glycosylated and secreted on the cell periphery. We hypothesized that if H2O2 affected secretion, invertase would not be correctly glycosylated and as a consequence would not be properly secreted. We used an anti-invertase antibody to detect the unglycosylated form of invertase in wild-type cells growing in rich media. Upon shifting to a low glucose concentration (0.3%), an elevated band was revealed, which corresponded to different hyperglycosylated forms at 25 °C in a sec1-1 mutant. When the cells were shifted to 38 °C for 2 h, the wide band corresponding to the glycosylated forms almost disappeared, indicating a blockage in the later steps of secretion. We investigated the response to increasing concentrations of H2O2 and observed that the multiple bands of glycosylation gradually disappeared with increasing concentrations of H2O2. These results suggested that oxidative stress caused by H2O2 impaired invertase secretion to a level equivalent to that observed in the sec1-1 mutant. Taking all the results shown in Fig. 1 together, we can conclude that oxidative stress affects the later steps of secretion and consequently causes ribosomal gene repression mediated by the Pkc1 function. This process also requires polymerisation of the actin cytoskeleton.
We are very grateful to Dr Luis Rodriguez Dominguez from the Universidad de la Laguna (Spain) for providing with anti-invertase polyclonal antibody. M.I.P. was funded by the Generalitat de Catalunya. This work was supported by the Spanish Ministry Education and Science (Spanish Government), through Grant BFU2009-11215.