Hyperoxidation of mitochondrial peroxiredoxin limits H2O2‐induced cell death in yeast

Abstract Hydrogen peroxide (H2O2) plays important roles in cellular signaling, yet nonetheless is toxic at higher concentrations. Surprisingly, the mechanism(s) of cellular H2O2 toxicity remain poorly understood. Here, we reveal an important role for mitochondrial 1‐Cys peroxiredoxin from budding yeast, Prx1, in regulating H2O2‐induced cell death. We show that Prx1 efficiently transfers oxidative equivalents from H2O2 to the mitochondrial glutathione pool. Deletion of PRX1 abrogates glutathione oxidation and leads to a cytosolic adaptive response involving upregulation of the catalase, Ctt1. Both of these effects contribute to improved cell viability following an acute H2O2 challenge. By replacing PRX1 with natural and engineered peroxiredoxin variants, we could predictably induce widely differing matrix glutathione responses to H2O2. Therefore, we demonstrated a key role for matrix glutathione oxidation in driving H2O2‐induced cell death. Finally, we reveal that hyperoxidation of Prx1 serves as a switch‐off mechanism to limit oxidation of matrix glutathione at high H2O2 concentrations. This enables yeast cells to strike a fine balance between H2O2 removal and limitation of matrix glutathione oxidation.


Appendix
. (corresponds to Figure 2) (A) The response of mitochondrial matrix-localized roGFP2-Tsa2ΔCR probe, in wild-type and Δprx1 cells to the addition of 10 µM antimycin A. Cells were grown in SGal (-Leu) medium and harvested at early exponential phase. Lighter colored curves are controls showing the probe response upon the addition of 0.1% (v/v) ethanol. (B) Volcano plot of the differential mRNA expression in Δprx1+empty vector cells, compared to Δprx1+Prx1-WT cells. P-values are plotted against enrichment (log2(fold change (FC))). Cutoff was set at ±0.32 log2(fold change (FC)). Cells were grown in SGal (-Ura) medium and harvested at early exponential phase. (C) The cellular process GO terms analysis of the differential expression output data represented in (B) performed using the online tool GOrilla. (D) Growth curve of wild-type, Δprx1, Δctt1 and Δctt1Δprx1 cells in SD medium complemented with all amino acids and 1 mM H2O2 or water as a control. (n = 3 biological replicates with cells obtained from independent cultures for each replicate). Significance for the difference in the time the cultures reach 50% of their maximal OD600 was assessed with a Student's, 2-tailed, unpaired t-test. **p < 0.01. Compared to the other strains, Δctt1Δprx1 cells have an extended lag phase. (E) The response of cytosolic roGFP2-Tsa2ΔCR probe, in Δtsa1Δtsa2 and Δtsa1Δtsa2Δprx1 cells grown in SGal (-Leu) medium, to the addition of 0.1 mM exogenous H2O2. Both strains exhibit very similar responses with the Δtsa1Δtsa2Δprx1 cells allowing for slightly faster recovery of the cytosolic roGFP2-Tsa2ΔCR probe after bolus H2O2 incubation.
OxD refers to the degree of sensor oxidation. Error bars represent the standard deviation (n = 3 biological replicates, with cells taken from independent cultures for each individual biological replicate). Significance was assessed with the t-test. **p < 0.01. Figure 3) (A) The response of cytosolic and mitochondrial matrix-localized Grx1-roGFP2 probes, expressed in BY4742 wild-type yeast cells with an empty vector or in Δglr1 cells transformed either with an empty vector or a vector encoding wild-type Glr1 grown in SGal (-Leu,-Ura), to a bolus of exogenous 2.5 mM H2O2. Expression of Glr1 from a strong promotor improves maintenance of EGSH in the matrix compared to wild-type cells indicating the presence of only limiting amounts of Glr1 in the matrix of wild-type cells.

Appendix FigureS3. (corresponds to
(B) Cartoon model representing that Glr1 activity in the matrix is limiting for reduction of glutathione disulfide.
OxD refers to the degree of sensor oxidation. Error bars represent the standard deviation (n = 3 biological replicates, the mean of 3 technical replicates in each case, in which the probe response was measured 3 times, with cells obtained from 3 independent cultures, for every strain and probe combination). Figure S4. (corresponds to Figure 4) (A) 'Acute stress-washout' assay. BY4742 wild-type yeast cells expressing a mitochondrial matrix-localized Grx1-roGFP2 probe were pre-treated for the indicated times with the indicated amounts of H2O2. Afterwards, cells were washed and the response of matrixtargeted Grx1-roGFP2 to the addition of 1 mM H2O2 was measured. For this experiment cells were grown in SGal (-Leu) medium and harvested at early exponential phase. Preincubation with H2O2 attenuates the subsequent response of the Grx1-roGFP2 probe towards a 1 mM bolus application of H2O2. (B) ∆prx1 cells transformed with a plasmid encoding wild-type Prx1, a plasmid encoding the matrix-targeted D-amino acid oxidase (Su9-DAO) and expressing a mitochondrial matrixlocalized Grx1-roGFP2 were grown in SGal (-Leu -Ura -His) to early exponential phase before the addition of 0.15 M L-alanine for the indicated times. Subsequently, cells were washed and the response of Grx1-roGFP2 to the addition of 1 mM H2O2 was measured. Incubation with Lalanine (not a substrate of DAO) does not affect the Grx1-roGFP2 response. (C) The response of mitochondrial matrix Grx1-roGFP2, expressed in wild-type and Δtsa1Δtsa2 cells, to the addition of exogenous H2O2. Cells were grown in in SGal (-Leu). Δtsa1Δtsa2 cells exhibit at higher H2O2 concentrations a lower response of the matrix-targeted Grx1-roGFP2 probe. In these cells higher amounts of H2O2 can reach the matrix and we propose can there hyperoxidize Prx1 impairing oxidation of GSH during bolus H2O2 treatment. Error bars represent the standard deviation (n = 3 biological replicates, the mean of 3 technical replicates in each case, in which the probe response was measured 3 times, with cells obtained from 3 independent cultures). (D) Redox shift assay to establish the redox state of Prx1 in the samples of (B). The experiment was performed as described in Figure 4F. The redox state of Prx1 does not change upon addition of L-alanine (E) Redox shift assay to establish the redox state of Prx1 in the samples of Figure 4D. The cysteine C91 in Prx1 becomes partially inaccessible upon incubation of DAO-expressing cells with D-alanine. This is in line with the decreased Grx1-roGFP2 response in Figure 4D and argues for hyperoxidation of Prx1 under these conditions. (F) The graph shows steady state Grx1-roGFP2 oxidation following H2O2 pre-treatment and cytosolic translation inhibition as well as the maximum probe oxidation in response to the subsequent second H2O2 treatment. Data refers to the experiment performed in Figure 4J. (G) Graphic representation of the 'Halo assay' for growth sensitivity upon chronic exposure to H2O2. The zone of growth inhibition is reported as the percentage of the radius of the circle where no colonies grow in relation to the maximal radius of the plate. The example plate is representative for one of the three biological replicates of the experiment depicted in (H), specific to BY4742 + empty plasmid. (H) H2O2 sensitivity of yeast cells expressing Zea mays aquaporins in their plasma membrane. Cultures of wild-type S. cerevisiae cells (BY4742) co-transformed with empty plasmid, or an plasmid encoding active aquaporin PIP25 or the inactive H199K mutant, were diluted in sterile distilled water to an OD600 of 0.1 and dispersed on plates SD (-Leu) agar plates. A filter disk placed in the middle of the plate was infused with 1 M H2O2 and the zone of growth inhibition "Halo" was recorded after 2 days growth at 30°C. Cells expressing a functional aquaporin in the plasma membrane are more sensitive to chronic H2O2 stress. (n = 3 biological replicates, with cells taken from independent cultures for each individual biological replicate). (I) H2O2 and diamide sensitivity of yeast cells lacking the mitochondrial peroxiredoxin Prx1. Cultures of wild-type and ∆prx1 cells were diluted in sterile distilled water to an OD600/mL of 0.1 and dispersed on plates containing the different carbon sources glucose (YPD, fermentation), galactose (YPGal, fermentation and respiration), and glycerol (YPG, respiration is enforced). A filter disk placed in the middle of the plate was infused with 1 M H2O2 or 1 M diamide and the zone of growth inhibition was recorded after 2 days growth at 30°C. (n = 3 biological replicates, with cells taken from independent cultures for each individual biological replicate). (J) Layout of the 'acute stress assay' experiment preformed in (K, L, M, N) and in Figures 5E,  6A, 7C-D -refer to the "Acute stress assay" section of materials and methods for a detailed explanation. (K) H2O2 'acute stress' assay for yeast cells expressing Zea mays aquaporins in their plasma membrane. Cultures of wild-type S. cerevisiae cells (BY4742) co-transformed with an empty plasmid or a plasmid encoding, active aquaporin PIP25 or the inactive H199K mutant were grown in SD (-Leu) medium and harvested at early exponential phase. Cells were pre-treated with the indicated amounts of H2O2 for 30 mins. Afterwards, the cells were diluted and a fixed volume plated on YPD plates. The number of viable colonies was counted after 2 days growth at 30°C, here represented as a percentage with respect to the 0 mM pre-treatment. Cells expressing a functional aquaporin in the plasma membrane are more sensitive to acute H2O2 stress. Error bars represent standard deviation (n = 3 biological replicates, with cells taken from independent cultures for each individual biological replicate). (L) H2O2 'acute stress' assay. Wild-type and Δprx1 cells pre-grown in YPD to early exponential phase were treated with the indicated amounts of H2O2 for 30 mins. Afterwards, the cells were diluted and a fixed volume plated on YPD plates. The number of viable colonies was counted after 2 days growth at 30°C, here represented as a percentage in respect of the 0 mM pretreatment. Cells lacking Prx1 are more resistant to acute H2O2 treatment. Error bars represent standard deviation (n = 5 biological replicates, with cells taken from independent cultures for each individual biological replicate). (M) H2O2 'acute stress' assay. Δprx1 cells co-transformed with either an empty vector or a plasmid encoding wild-type Prx1 were grown in SGal (-Ura) to early exponential phase. Cells were subsequently treated with the indicated amounts of H2O2 for 30 mins. Afterwards, the cells were diluted and a fixed volume plated on YPD plates. The number of viable colonies was counted after 2 days grown at 30°C, here represented as a percentage in respect of the 0 mM pre-treatment. Δprx1 cells with reintroduced Prx1 are more sensitive to acute H2O2 stress. Error bars represent standard deviation (n = 3 biological replicates, with cells taken from independent cultures for each individual biological replicate). (N) H2O2 'acute stress' assay. Δprx1Δctt1 cells co-transformed with an empty plasmid or with a plasmid encoding wild-type Prx1 were grown in SGal (-Ura) and harvested at early exponential phase. Cells were subsequently treated with the indicated amounts of H2O2 for 30 mins. Afterwards, the cells were diluted and a fixed volume plated on YPD plates. The number of viable colonies was counted after 2 days grown at 30°C, here represented as a percentage in respect of the 0 mM pre-treatment. Cells lacking Prx1 are more resistant to acute H2O2 stress even in the absence of the CTT1 gene. This indicates that a main contribution to acute H2O2 resistance stems from properties of Prx1 and not the upregulation of Ctt1 during an adaptive response. Error bars represent standard deviation (n = 3 biological replicates, with cells taken from independent cultures for each individual biological replicate).

Appendix
In all graphs, significance was assessed with a Student's, 2-tailed, unpaired t-test. *p < 0.05; **p < 0.01; ***p<0.001;****p < 0.0001. Figure S5. (corresponds to Figure 5) (A) Scheme illustrating the rationale behind the design of Prx1-P233stop. Alignment of amino acids in the C-terminal region of S. cerevisiae Prx1, S. cerevisiae Tsa1 and S. pombe Tpx1. A truncated SpTpx1 variant (truncation site indicated by the arrow) is more resistant to hyperoxidation by H2O2. We tested whether truncation of Prx1 at an analogous position would render Prx1 more resistant to H2O2-induced hyperoxidation. Collectively, the data in Figures 5 and 6 indicate that this is indeed the case. At present, the molecular basis of this increased resistance remains unclear. (B) Redox shift assays of Prx1 variants in Δprx1 transformed with a plasmid encoding the Prx1-P233stop variant or wild-type Prx1 to assess Prx1-P233stop migration behavior on SDS-PAGE and assess accessibility to mmPEG in shift experiments. Prx1-P233stop can be shifted by mmPEG and at steady state (lane 1) is fully reduced. At the resolution of this gel, differences in full-length Prx1 migration behavior cannot be deduced. (C) Tables summarizing the significance assessed with a Student's, 2-tailed, unpaired t-test for all the combination of the strains used in Figure 5E. *p < 0.05; **p < 0.01; ***p < 0.001;****p < 0.0001. The response of a mitochondrial matrix-localized Grx1-roGFP2 probe, expressed in Δprx1 cells containing an empty plasmid or a plasmid encoding either Su9-sa1, PRDX3, PRDX5 or Su9-PRDX6, Su9-PfAOP or Su9-PfAOP-L109M, to the addition of 1 mM exogenous H2O2 after a preceding pretreatment with either 10 mM H2O2 or water as a control (acute stress assay). Some peroxiredoxins mediate efficient GSH oxidation (e.g. HsPRDX6, AOP L109M), others are less so (e.g. HsPRDX5). After preincubation with H2O2, transfer of oxidation from H2O2 onto EGSH is prevented for some peroxiredoxins (e.g. AOP L109M), for others not (e.g. HsPRDX6). Error bars represent the standard deviation (n = 3 biological replicates, the mean of 3 technical replicates in each case, in which the probe response was measured 3 times, with cells obtained from 3 independent cultures, for every strain). OxD represents the degree of sensor oxidation. Cells were grown to an early exponential phase in SGal medium lacking the appropriate amino acids for plasmid selection. (C) Tables summarizing the significance assessed with a Student's, 2-tailed, unpaired t-test for all the combination of the strains used in Figure 6A. *p < 0.05; **p < 0.01; ***p < 0.001;****p < 0.0001. (D) Effect of acute H2O2 exposure on Prx1 redox state. Δprx1 cells co-transformed with a plasmid encoding wild-type Prx1 and grown in SGal (-Ura) medium to early exponential phase were treated with the indicated amounts of H2O2 for 10 mins. Afterwards, the cells were either directly modified with the alkylating agent mmPEG24 or first reduced with TCEP (reduction of disulfide bonds and sulfenylated cysteines, but not hyperoxidized sulfinylated and sulfonylated cysteines) and then modified with mmPEG24. The lower band present in the TCEP-treated samples indicates and irreversible hyperoxidation of the cysteine. At around 0.5 mM exogenous H2O2 (in some assays at 1 mM) Prx1 becomes hyperoxidized (observe TCEP treatment does not reduce the blocked cysteine residue). (E) Effect of acute H2O2 exposure on cells lacking PRX1. Δprx1 cells co-transformed with an empty plasmid and grown in SGal (-Ura) medium to early exponential phase were treated with the indicated amounts of H2O2 for 10 mins. Afterwards, the cells were either directly modified with the alkylating agent mmPEG24 or first reduced with TCEP and then modified with mmPEG24. The controls for Δprx1 cells expressing wild-typePrx1 from a plasmid are provided. In Δprx1 cells, no band can be observed indicating specificity of the antibody. (F) Effect of acute H2O2 exposure on Δprx1 cells complemented with a plasmid encoding Su9-PRDX6. Cells were grown in SGal (-His) and treated with the indicated amounts of H2O2 for 10 mins. Afterwards, the cells were either directly modified with the alkylating agent mmPEG24 or first reduced with TCEP and then modified with mmPEG24. At around 0.5 mM external H2O2 PRDX6 becomes inaccessible to mmPEG modification. This is however not hyperoxidation as it can be reverted by treatment with TCEP. The PRDX6 cysteine remains (at least partially) in a non-hyperoxidized state upon treatment with up to 10 mM H2O2. Tables   Appendix Table S1. Yeast strains used in this study.