2.1 Deletion of Ubp2 restores respiratory growth defects in the absence of Hsp31 paralogs

Mitochondrial structure and function are essential indices of cellular health and have been studied extensively owing to the dynamicity of the organelle. The Hsp31 paralogs in Saccharomyces cerevisiae, which belong to the DJ-1 superfamily, have been identified as crucial modulators of mitochondrial health [14,15]. It has been reported that a combinatorial deletion of Hsp31 and Hsp34 results in a growth defect in glycerol at non-permissive temperatures, suggesting the role of Hsp31 paralogs in maintaining mitochondrial integrity [15]. Alteration in the expression of Fzo1 due to the absence of Hsp31 paralogs manifests as an increase in hyperfused mitochondrial structures. However, the mechanism by which these paralogs perturb the mitochondrial dynamics remains elusive.

The process of fusion is a finely orchestrated mechanism mediated by the action of a conserved class of cysteine-dependent deubiquitinases. Recent evidence suggests that Ubp2 is a crucial modulator of mitochondrial dynamics and essential for maintaining the fusogenic status of Fzo1 [38]. Since Hsp31 paralogs regulate Fzo1 expression and mitochondrial dynamics, we investigated a possible functional cross-talk between Ubp2 and Hsp31 paralogs. Using yeast genetic manipulation, we created ∆ubp2 in the BY4741 strain background. The deletion of Ubp2 (∆ubp2) leads to an impairment in the growth in both complete (S.C. Dextrose) and non-fermentable media (S.C. Glycerol) at non-permissive temperature of 37°C (Fig 1A). At the same time, among the four Hsp31 paralogs, only the cells lacking Hsp31(∆31) exhibit a slight reduction in growth in glycerol at a non-permissive temperature of 37 °C (Fig 1A) [15]. The cells lacking Hsp34 (∆34) have growth comparable to WT cells. However, together deletion with Hsp31 (∆31∆34) leads to severely compromised growth in glycerol in conditions of heat stress at 37°C (Fig 1A). To test a genetic interaction between Hsp31 paralogs (Hsp31 and Hsp34) and Ubp2, we constructed combination of deletion strains lacking Ubp2 in the background of ∆31(∆31∆ubp2), ∆34(∆34∆ubp2) and ∆31∆34 (∆31∆34∆ubp2). Interestingly, the strains ∆31∆ubp2 and ∆34∆ubp2 alone had a more compromised growth phenotype in glycerol at 37°C, comparable to ∆ubp2 (Fig 1A). Intriguingly, on the other hand, the deletion of Ubp2 in the absence of 31 paralogs (∆31∆34∆ubp2)showed rescue in the growth phenotype comparable to WT in a non-fermentable carbon source and non-permissive temperature of 37°C (Fig 1A).

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 1. Ubp2 deletion restores the respiratory growth defect of Δ31Δ34 strain.

(A) Growth phenotype assessment. The indicated yeast strains were allowed to grow up to mid-log phase in SC dextrose broth at 30°C. Ten-fold serially diluted cells were spotted on the indicated media and incubated at permissive temperature (30°C) and non permissive temperature (24°C and 37°C). Images were captured at 36h and 72h for dextrose and glycerol, respectively. (B) Assessment of the Ubp2-HA levels in the absence of Hsp31 paralogs. Whole-cell lysate of the indicated strains was subjected to immunoblotting and probed for Ubp2-HA levels in late log phase(12h). (C) Densitometric analysis for probing the difference in expression of Ubp2-HA levels. (D) Growth phenotype assessment upon complementation with Ubp2. The indicated strains were allowed to grow up to the mid-log phase in SC Leu- media and subjected to grow at permissive and non-permissive temperatures in dextrose and glycerol. Images were captured for dextrose and glycerol after 36h and 72h, respectively. Unpaired Student t-test was performed for statistical analysis. Error bars represent the standard deviation in median values from 3 biological replicates. Asterisks indicate the p-value, *, p < 0.05; **, p < 0.01; ***, p < 0.001; ****, p < 0.0001.

https://doi.org/10.1371/journal.pgen.1011353.g001

To determine whether the growth rescue by the Ubp2 in the absence of Hsp31 paralogs is due to the consequence of its regulated expression, we first analyzed the levels of Ubp2 in WT and ∆31∆34 strains. To test this, Ubp2 was genomically tagged with hemagglutinin (HA) at the C-terminus and expressed in WT and ∆31∆34. The whole cell lysates were subjected to immunoblotting and probed for the Ubp2 levels. The immunoblot analysis revealed that ∆31∆34 strains showed up to a 2.5-fold increment in the expression of Ubp2 levels in cells from late log phase (Fig 1B and 1C). While, WT cells exhibit a significant reduction in Ubp2 expression in the late log phase compared to the cells collected from the early-log phase (S1B Fig, compare lanes 1 and 3),Ubp2 expression remained unaltered in ∆31∆34 across mid-log (6h) and late-log (12h) phases (S1B Fig, compare lanes 2 and 4), suggesting Hsp31-dependent regulated expression of Ubp2. To further ascertain this observation, the Ubp2 as the molecular player in regulating the growth defect in ∆31∆34, we overexpressed Ubp2 (under TEF promoter) in the background of ∆ubp2, ∆31∆34∆ubp2, and subjected the strains to growth analysis. The overexpression of Ubp2-HA was ascertained by immunoblotting (S2 Fig). The phenotypic assessment in glycerol indicated that Ubp2 overexpression in ∆ubp2 shows a phenotype comparable to WT. In contrast, Ubp2 overexpression in ∆31∆34 leads to growth sensitivity in glycerol at 37°C, equivalent to ∆31∆34 alone (Figs 1D and S16). These experimental evidence suggest a strong functional cross-talk between Hsp31 paralogs and Ubp2 in terms of regulating respiratory growth in conditions of heat stress.

2.2 Ubp2 regulates mitochondrial integrity and turnover in cells lacking Hsp31 paralogs

The phenotypic growth defects in respiratory deficient conditions are closely correlated to perturbance of mitochondrial health. Therefore, we assessed mitochondrial-specific functional parameters across all the strains. In particular, the changes in the mitochondrial morphology are assessed to ascertain the overall cellular health. To microscopically visualize mitochondrial structures, the cells were transformed with MTS-mCherry constructs that specifically translocate to the organelle due to targeting sequence and decorate the mitochondria [14]. In agreement with the previous findings, the ∆ubp2 strain exhibited fragmented mitochondria [39] (Fig 2A, compare panels 2 with panel 1). In contrast, the ∆31∆34 strain showed hyperfused mitochondrial morphology [15] (Fig 2A, compare panels 3 with 1). Most interestingly, upon deletion of Ubp2 in ∆31∆34 background, mitochondrial morphology was reversed, which now closely resembles the WT cells (Fig 2A, compare panels 4 with panel 1). The fragmented, intermediate, and tubular states of the mitochondria were subjected to quantification for each respective strain. The results highlight that the cells exhibiting intermediate mitochondrial morphology in ∆31∆34∆ubp2 are comparable to that of the WT in correlation with the reversal of growth phenotype (Fig 2B) compared to the extensive tubular network observed in ∆31∆34 strain. On the other hand, Ubp2 deletion in the background of single deletion strains of Hsp31 paralogs (∆31∆ubp2, ∆34∆ubp2) exhibited fragmented mitochondria, comparable to the ∆ubp2 strain (S3 Fig). Further, to validate the genetic interaction between Ubp2 and Hsp31 paralogs, Ubp2 was overexpressed in ∆31∆34∆ubp2, and the mitochondrial network was assessed.Corroborating with the phenotype, Ubp2 overexpression in ∆31∆34∆ubp2 restores the hyperfused mitochondrial structures compared to WT (S12 Fig, compare panels 1 and 5). Intriguingly, the catalytically inactive form of Ubp2 (ubp2_C745S) could not restore the hyperfused network of mitochondria (S12 Fig, compare panels 1,5 and 6), highlighting the role of catalytic cysteine in the maintenance of mitochondrial dynamics.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 2. Ubp2 deletion reverts the defect in mitochondrial integrity in the absence of Hsp31 paralogs.

(A) Assessment of mitochondrial integrity. Yeast strains expressing MTS-mCherry grown in SC Leu^- dextrose till the mid-log phase were subjected to microscopic analysis to visualize mitochondria. Scale bar (10 μm). The three different mitochondrial morphologies scored have been represented (Fragmented, Intermediate, Tubular cells). (B) Quantification of cells exhibiting specific mitochondrial morphology. (C) Quantification of total mitochondrial mass by flow cytometric analysis. Cells grown until the mid-log phase were subjected to NAO staining and quantified using the BD FACS Verse instrument. (D) Evaluation of the functional mitochondrial mass by flow cytometry. Cells grown until the mid-log phase were subjected to JC-1 staining, a potentiometric dye, and acquired using BD FACS Verse. (E) Measurement of ATP levels. Mitochondria were isolated from the indicated strains, and ATP was measured using a fluorescence-based assay. One-way ANOVA with Tukey’s multiple comparison test was used for significance analysis. Error bars represent the standard deviation in median values from 3 biological replicates. Asterisks indicate the p-value, *, p < 0.05; **, p < 0.01; ***, p < 0.001; ****, p < 0.0001.

https://doi.org/10.1371/journal.pgen.1011353.g002

As we observed varied mitochondrial structures across the different strains, we probed for the total and functional mitochondrial mass change using flow cytometric analysis. NAO (10-N-nonyl-acridine orange) is used specifically for staining cardiolipin, an indicator of the total mitochondrial mass in cells. Upon flow cytometry, deletion of Ubp2 did not show a significant difference in the total mitochondrial mass compared to WT (Fig 2C). However, the ∆31∆34 strain exhibited a substantial increase in the total mitochondrial mass, as reported previously [15]. Interestingly, there was no difference in the total mitochondrial mass in ∆31∆34∆ubp2, and it was comparable to ∆31∆34 (Fig 2B). This indicates that deletion of Ubp2 in the background of WT and ∆31∆34 does not alter the total mitochondrial mass. The functionality of mitochondria was assessed by staining the cells with a potentiometric dye, JC-1, and subjecting them to flow cytometric analysis. A significant decrease in the functional mitochondrial content in ∆ubp2 was observed compared to WT (Fig 2D). Contrastingly, the ∆31∆34 strain exhibited a significant increase in the functional mitochondrial mass [15], as reported previously. Strikingly, the deletion of Ubp2 in the background of ∆31∆34 restored the functional mitochondrial mass comparable to WT (Fig 2D). In order to validate the result, Ubp2 was overexpressed in the background of ∆31∆34∆ubp2 and the total and functional mitochondrial mass was measured (S13A and S13B Fig). The total mitochondrial mass upon overexpression of Ubp2 in ∆31∆34∆ubp2 is comparable to ∆31∆34, consistent with the previous results(S13A Fig). Moreover, the functional mitochondrial mass increases upon Ubp2 overexpression in ∆31∆34∆ubp2 and is comparable to ∆31∆34 (S13B Fig).

Furthermore, as mitochondrial function is closely associated with respiration status, leading to ATP generation, we measured the mitochondrial ATP levels across all the strains. Corroborating the previous data, we found a significant decrease in ATP levels in cells lacking Ubp2, while ∆31∆34 exhibit higher ATP levels. At the same time, the ATP levels in ∆31∆34∆ubp2 have restored to WT, consistent with the reversal of mitochondrial-specific phenotypes correlating to the suppression of growth defects (Fig 2E). These results suggest that Ubp2 plays a crucial role in maintaining functional mitochondrial mass in conjunction with Hsp31 paralogs.

Mitochondrial homeostasis is regulated by mitophagy, which clears the damaged and aged mitochondria [40]. We observed that loss of Ubp2 does not affect the total mitochondrial mass, albeit involving the functional mitochondrial pool. We assayed the strains for organellar turnover to investigate whether Ubp2 has a role in modulating mitophagy. Mitophagy is monitored by tagging the outer mitochondrial membrane protein (OM45) with GFP at the C-terminus and subjecting the strains to microscopic and immunoblotting analysis [41]. The cells were grown in dextrose until the mid-log phase and then in glycerol for the subsequent time points as indicated. Samples were collected, and cells were lysed for analysis using immunoblotting. The results suggest that the process of mitophagy is initiated in the WT strain after 36h of mitophagy induction (Fig 3A). On the other hand, the ∆ubp2 exhibits mitophagy initiation at 24h of induction, probably due to the presence of fragmented smaller mitochondrial structures. However, the intensity of the processed GFP in ∆ubp2 is less than the WT at the same time points. This indicates that even though mitophagy is initiated at an earlier time in ∆ubp2, clearance of damaged mitochondria is slower than the WT, accumulating fragmented and nonfunctional mitochondrial mass. ∆31∆34 exhibits comparable mitophagy rate to WT, albeit at a lower processing efficiency. However, ∆31∆34 showed slower mitophagy induction possibly due to the presence of hyperfused mitochondrial structures (Fig 3A and 3B). Interestingly, ∆31∆34∆ubp2, which has restored mitochondrial integrity exhibits kinetically faster mitochondrial clearance by induction observed at 36h as compared to ∆31∆34 alone (Fig 3A and 3B). However, the intensity of processed GFP is less than ∆31∆34 and not comparable to WT, suggesting that Ubp2 deletion partially suppresses the process of mitophagic flux. The mitophagy deficient ∆atg32 strain served as a negative control for GFP processing (Fig 3A, last panels).

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 3. Role of Ubp2 in regulating mitophagy.

(A, B) Mitophagy induction by Western analysis. Cells expressing OM45-GFP from the indicated strains were cultured in glycerol for the mentioned time points, and the lysates were subjected to Western blotting and quantified using Two-Way ANOVA for significance analysis.Error bars represent the standard deviation in median values from 3 biological replicates. Asterisks indicate the p-value, *, p < 0.05; **, p < 0.01; ***, p < 0.001; ****, p < 0.0001 (B). ∆atg32 was used as a positive control to compare the defects in mitophagy. (C) Microscopic analysis of mitophagy induction. Cells expressing OM45-GFP were grown in YP-glycerol for 48h and subjected to fluorescence microscopy. The vacuole is stained with FM4-64 dye. Scale bars (10 µm). Images were zoomed up to 2X and represented.

https://doi.org/10.1371/journal.pgen.1011353.g003

To validate the results further, we also subjected the strains to microscopy to observe the vacuolar internalization of processed GFP, after staining with FM4–64 to visualise the vacuole. In correlation with the immunoblotting results, ∆ubp2 exhibits processed GFP comparable to WT. Still, the fluorescence intensity was higher than WT after 48h of mitophagy induction indicating faster turnover (Fig 3C, compare panels 1 and 2). Similarly, OM45-GFP is internalized to a lesser extent in ∆31∆34 compared to ∆31∆34∆ubp2 (Fig 3C, compare panels 3 and 4). However, the level of processed OM-45 GFP fluorescence intensity in ∆31∆34∆ubp2 is lesser than WT, indicating that Ubp2 partially restores the mitophagic turnover. The cells are deficient in mitophagy (∆atg32), did not show any processed OM-45 GFP fluorescence in the vacuole, and were used as a negative control (Fig 3C, compare panels 1–4 with 5).

2.3 Fzo1 levels and turnover are restored upon deletion of Ubp2 in the absence of Hsp31 paralogs

The expression and stability of proteins like Fzo1 and Dnm1 in Saccharomyces cerevisiae determine the dynamicity of the mitochondria [42]. Our results exclusively highlight variations in mitochondrial health, and we analyzed for changes in the expression of mitochondrial dynamic proteins. In agreement with previous findings, we observed a significant decrease in Fzo1 levels in ∆ubp2 compared to WT (Fig 4A and 4B). In contrast, a substantial increment in the expression of Fzo1 was observed in ∆31∆34 strains [15] (Fig 4A and 4B). Interestingly, ∆31∆34∆ubp2 exhibited a restoration in levels of Fzo1, comparable to WT (Fig 4A and 4B). To ascertain the role of Ubp2 in stabilizing Fzo1 and promoting fusion, Ubp2 overexpressed in the WT background, and phenotypic analysis was performed. In agreement with the previous results, WT/Ubp2(↑) exhibits a compromised phenotype in glycerol at 37°C (S5A Fig). At the same time, mitochondria exhibited hyperfused morphology upon Ubp2 overexpression (S5B Fig). The hyperfusion of mitochondria results from enhanced expression of Fzo1 (S5C Fig).

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 4. Fzo1 levels are restored upon deletion of Ubp2 in the absence of Hsp31 paralogs.

(A, B) Evaluation of the relative steady-state levels of Fzo1 by western blotting. The cells from indicated strains lysed, and Fzo1 levels were probed in the whole cell lysate (A). The blots were subjected to quantification by densitometry, and the fold change in Fzo1 expression was represented. (C, D) Assessment of the turnover kinetics of Fzo1 using cycloheximide assay. The cells were subjected to treatment with cycloheximide for the indicated time points, and the lysates were subjected to western blotting (C). The blots were subjected to quantification densitometrically and represented the rate at which Fzo1 is degraded across the strains (D). Two-way ANOVA was performed using three biological replicates. (E, F) Evaluation of ubiquitination status of Fzo1. Genomically tagged Fzo1 at the C-terminus with HA-tag was subjected to pull-down analysis using HA-conjugated beads and probed with anti-HA and anti-ubiquitin specific antibodies (E). The Western blots were subjected to quantitation by densitometry, and fold change in the ubiquitination levels is represented (F). One-way ANOVA with Tukey’s multiple comparison tests was used for significance analysis. Error bars represent the standard deviation in median values from 3 biological replicates. Asterisks indicate the p-value, *, p < 0.05; **, p < 0.01; ***, p < 0.001; ****, p < 0.0001.

https://doi.org/10.1371/journal.pgen.1011353.g004

To validate the genetic interaction between Ubp2 and Hsp31 paralogs, Ubp2 overexpressed in ∆31∆34∆ubp2, and the steady-state levels of Fzo1 were analyzed by immunoblotting. The results suggest a significant increment in Fzo1 expression upon overexpression of Ubp2 in ∆31∆34∆ubp2 as compared to WT (S6A Fig). Furthermore, there were no significant differences in the expression of fission protein Dnm1 across the strains (S15 Fig). The results further suggest that the absence of both Hsp31 paralogs and Ubp2 keeps the levels of Dnm1 unaltered in the cells, which is in agreement with the previous findings [15,39].

As there was a restoration of Fzo1 steady-state levels in ∆31∆34∆ubp2, we tested for the turnover rate of Fzo1 across the four strains by utilizing the cycloheximide chase assay (CHX). The turnover kinetics revealed that Fzo1 has a shorter half-life in ∆ubp2 compared to WT (Fig 4C and 4D). As opposed to ∆ubp2, the ∆31∆34 strain exhibited stable expression of Fzo1, correlating with the increased steady-state expression (Fig 4C and 4D). However, ∆31∆34∆ubp2 exhibits restoration of Fzo1 dynamics, comparable to WT (Fig 4C and 4D). Furthermore, upon overexpression of Ubp2 in ∆31∆34∆ubp2, the turnover kinetics of Fzo1 was found comparable to WT (S6B Fig).

Next, we determined if the altered Fzo1 levels are ubiquitinated as ubiquitin conjugation on Fzo1 can contribute to fusion or direct it towards proteolysis. The fusogenic ability depends on the ubiquitination status of the Fzo1 protein, dictated by the activity of Ubp2 [32]. To further investigate the ubiquitination status of the strains, we tagged Fzo1 C-terminally using hemagglutinin (HA) tag. The strains were then subjected to immunoprecipitation and pulled down using HA-specific antibodies. The pull-down fraction was subjected to immunoblotting and probed with an anti-ubiquitin antibody. We observed that the absence of Ubp2 leads to increased ubiquitination of Fzo1 as compared to WT (Fig 4E). However, ∆31∆34 exhibits a marked decrease in the ubiquitination status of Fzo1, corroborating with its increased stability in the absence of Hsp31 paralogs. Interestingly, there was a significant restoration in the levels of ubiquitinated Fzo1 in ∆31∆34∆ubp2 as it is comparable to WT (Fig 4E and 4F). These findings highlight that the ubiquitination status of Fzo1 dictates its stability and turnover kinetics across the strains.

2.4 Ubp2 deletion restores cell cycle progression in cells lacking Hsp31 paralogs

Mitochondrial structure transition is integral for the cells to advance into subsequent cell cycle phases [43–45]. Previous studies indicate that cells lacking Hsp31 paralogs switch to hyperfused mitochondrial structures as an adaptive strategy to ameliorate the increased ROS levels. At the same time, this adaptive strategy is detrimental to cell cycle progression, as observed in the ∆31∆34 strain [15]. Therefore, we hypothesized that remission of hyperfused mitochondrial structures in ∆31∆34∆ubp2 would restore the cell cycle progression.

To test this hypothesis, the strains were arrested at the G1 phase using alpha factor and synchronously released in dextrose media. Subsequently, the cells were fixed and subjected to flow cytometric analysis. Upon deletion of Ubp2, the cells exhibited no delay in cell cycle progression and resembled WT (Fig 5). The loss of Ubp2 does not cause perturbation in a cell cycle, as the mitochondrial distribution remains unaffected. However, the ∆31∆34 strain exhibited slower progression as more than 50% of the cells continued to be at the G2/M phase (from t = 80 min to t = 100 min), per the presence of hyperfused mitochondrial structures [15] (Figs 5A and S14). Interestingly, restoration of mitochondrial integrity in ∆31∆34∆ubp2 led to the usual progression in cell cycle and is comparable to the WT. Taken together, the results indicate that remission of mitochondrial structures upon deletion of Ubp2 in the absence of Hsp31 paralogs enables the cells to undergo unimpeded progression through the cell cycle.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 5. Cell cycle is restored upon deletion of Ubp2 cells lacking Hsp31 paralogs.

(A) Cell cycle analysis. The restoration of cell cycle was analysed using PI staining after synchronizing the strains using alpha factor. Release after G1-arrest was compared at the indicated time points. (B,C) Assessment of cell morphology and size. The cell morphology and size in the indicated strains were analysed by microscopy (B). Scale bar (15 µm). The relative differences in cell size in the yeast strains (WT, Δ31Δ34, and Δ31Δ34∆ubp2 cells) were quantified using ImageJ (C). The values were plotted using GraphPad Prism 5.0. One-way ANOVA with Tukey’s multiple comparison test was used for significance analysis. Error bars represent the standard deviation in median values from 3 biological replicates. Asterisks indicate the p-value, *, p < 0.05; **, p < 0.01; ***, p < 0.001; ****, p < 0.0001.

https://doi.org/10.1371/journal.pgen.1011353.g005

The progression of the normal cell cycle is well correlated with cell size and growth. Hence, we determined the change in cell size. Microscopic analysis revealed that ∆31∆34 has cells with an average cell size of ~9 µm exhibiting increased cell size than WT (average cell size of 4–6 µm) and, in the stationary phase, exhibit pseudohyphal characteristics (Fig 5B and 5C) [15]. On the other hand, ∆ubp2 cells exhibit normal cell size compared to WT. Interestingly, ∆31∆34∆ubp2 has reversed the cell population with a size similar to the WT (Fig 5B and 5C). Thus, balanced mitochondrial dynamics are integral in maintaining cell cycle and cell size.

2.5 Ubp2 regulates basal ROS levels in the absence of Hsp31 paralogs

Active mitochondrial respiration contributes significantly to the pool of cellular basal ROS levels. Hence, we probed for the basal levels of ROS across the strains. The yeast cells grown in dextrose-containing media were treated with ROS sensing dye, H₂DCFDA (2’,7’,-dichlorodihydrofluorescein diacetate), and subjected to flow cytometric and microscopic analysis to measure the total cellular ROS levels. The flow cytometry data reveals that ∆ubp2 has basal ROS levels comparable to WT. A significant increase in the basal ROS levels was observed in ∆31∆34, consistent with the previous findings (Fig 6A) [15]. Interestingly, the basal ROS levels were significantly restored in ∆31∆34∆ubp2 and the levels comparable to WT (Fig 6A). These findings were further confirmed by microscopic analysis with H₂DCFDA dye staining. Like flow cytometric data, a significant increment in green fluorescence intensity was observed for ∆31∆34 strain. On the other hand, the green fluorescence was significantly restored in ∆31∆34∆ubp2 strains, similar to WT and ∆ubp2 strains (Fig 6B). These results were further validated by overexpressing Ubp2 in the background of ∆31∆34∆ubp2 and consistent with the findings, we found a significant increment in the basal cellular ROS levels indicated by increased staining, as compared to WT (S7B and S7C Fig). Interestingly, overexpression of the catalytically inactive mutant of Ubp2 did not increase the basal ROS levels, which was equivalent to WT and ∆31∆34∆ubp2 (S9 Fig);suggesting the role of cysteine dependent function of Ubp2 in regulating the redox status.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 6. Basal ROS levels are restored upon restoration of mitochondrial dynamics.

(A,B) Evaluation of basal ROS levels by H₂DCFDA. Total ROS levels were measured in indicated strains by staining the cells with H₂DCFDA and subjected to flow cytometric analysis (A). Similarly, the cells from indicated strains were stained with H₂DCFDA and subjected to microscopic analysis (B). Scale bar (15 µm). (C,D) Evaluation of basal mitochondrial ROS. The mitochondrial ROS lelevls in indicated strains were determined by MitoSOX dye using flow cytometric analysis (C). Similarly, the mitochondrial ROS lelevls were probed by MitoSOX dye using microscopic analysis (D). Scale bar (10 µm). One-way ANOVA with Tukey’s multiple comparison test was used for significance analysis. Error bars represent the standard deviation in median values from 3 biological replicates. Asterisks indicate the p-value, *, p < 0.05; **, p < 0.01; ***, p < 0.001; ****, p < 0.0001.

https://doi.org/10.1371/journal.pgen.1011353.g006

The basal mitochondrial specific ROS was probed using mitochondrial specific dye, MitoSOX. The cells grown in dextrose-containing media were subjected to flow cytometry and microscopic analysis. The flow cytometric data revealed a non-significant difference in basal mitochondrial ROS in ∆ubp2 strain, which is comparable to WT. However, a two-fold increment in mitochondrial ROS was observed in ∆31∆34 (Fig 6C). Interestingly, the mitochondrial ROS levels of ∆31∆34∆ubp2 were restored to WT (Fig 6C). To substantiate the flow cytometry results, the microscopic analysis was performed using MitoSOX dye. The total mitochondrial ROS levels in WT and ∆ubp2 were comparable (Fig 6D, compare panels 1 and 2). A significant increment in basal mitochondrial ROS in ∆31∆34 was evident upon the microscopic analysis (Fig 6D, compare panels 1 and 3). Consistent with the above findings, ∆31∆34∆ubp2 exhibits a significant restoration of mitochondrial ROS distribution compared to WT (Fig 6D, compare panels 1 and 4), which correlates well with the reversal of mitochondrial integrity and functions. Overexpression of Ubp2 in the background of ∆31∆34∆ubp2 leads to increment of the basal mitochondrial ROS levels as indicated by MitoSOX staining (S8 Fig). However, the cysteine mutant of Ubp2 was ineffective in increasing mitochondrial ROS upon its overexpression in ∆31∆34∆ubp2 (S8 Fig).

2.6 Ubp2 modulates oxidative stress sensitivity in GSH dependent manner

Our experiments provide compelling evidence highlighting the significance of balanced mitochondrial dynamics in maintaining normal cellular growth. The findings also suggest that the mitochondrial integrity is restored in ∆31∆34∆ubp2 strain due to the suppression of basal ROS levels. As mitochondrial health is directly correlated to restoration in the basal and mitochondrial ROS, we further probed for the response of the strains in the presence of extraneous oxidative stress. To test, the yeast cells from the mid-log phase were subjected to 1 mM H₂O₂ treatment, followed by growth analysis by spot test. The results reveal that ∆ubp2 is not sensitive to extraneous oxidative stress as the growth is comparable to WT. On the other hand, ∆31∆34 showed maximum sensitivity to oxidative stress, consistent with previous findings [15]. Strikingly, the ∆31∆34∆ubp2 strain showed partial restoration of sensitivity to extraneous stress (Fig 7A). In addition, Ubp2 was overexpressed in ∆31∆34∆ubp2, and the strains were subjected to phenotypic analysis upon H₂O₂ treatment. The growth assay suggests that upon complementation, the phenotype is reversed, whereby ∆31∆34∆ubp2/Ubp2(↑) is more sensitive to stress (S7A Fig). Interestingly, overexpression of the cysteine mutant of Ubp2 does not sensitise ∆31∆34∆ubp2 cells to H₂O₂ stress and phenocopies ∆31∆34∆ubp2 (S10 Fig).

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 7. Loss of Ubp2 provides partial oxidative stress resistance in the absence of Hsp31 paralogs.

(A) Growth phenotype assessment. Yeast strains were grown till mid-log phase and were subjected to H₂O₂ stress. (B,C) Evaluation of Glutathione levels using flow by microscopy and cytometry. The GSH levels were determined by monochlorobiamine staining and subjected to microscopy (B), Scale bar (25 µm). The normalised GSH levels were probed using flow cytometry (C) by monochlorobiamine staining. (D, E) Assessment of relative GSH/GSSG ratio in the mitochondrial fraction (D) and cytoplasmic fraction (E) of the indicated strains, using Glutathione assay kit. One-way ANOVA with Tukey’s multiple comparison test was used for significance analysis. Error bars represent the standard deviation in median values from 3 biological replicates. Asterisks indicate the p-value, *, p < 0.05; **, p < 0.01; ***, p < 0.001; ****, p < 0.0001.

https://doi.org/10.1371/journal.pgen.1011353.g007

Our results indicate that cross-talk between Ubp2 and Hsp31 paralogs regulates mitochondrial dynamics via Fzo1 and contributes to oxidative stress response. The glutathione pool is a crucial indicator of oxidative stress, wherein the reduced form of glutathione (GSH) acts as a ROS scavenger and mitigates oxidative stress. GSSG is the oxidized form of glutathione, which increases significantly in conditions of oxidative stress to cells. Interestingly, the GSH and GSSG ratio is also a crucial indicator of mitochondrial fusion, whereby GSSG mediates the stability of mitofusins and promotes mitochondrial fusion in conditions of oxidative stress [46]. As our results indicate that mitochondrial remodeling via Fzo1 is a consequence of oxidative stress response, we further probed for the GSH levels using monochlorobimaine (MCB). Monochlorobimine is a cell-permeable nonfluorescent probe that shows fluorescence only when it reacts with the reduced glutathione pool in cells. Cells in the mid-log phase were treated with MCB and subjected to microscopy and flow cytometric analysis. It is evident from microscopic analysis that ∆ubp2 does not show any difference in MCB staining as the fluorescence intensity is comparable to WT (Fig 7B). This corroborates the fact that Fzo1 stability is altered without affecting the redox parameters in the absence of Ubp2. Interestingly, ∆31∆34 exhibits a significant decrease in GSH levels, as indicated by the minimal MCB fluorescence in cells lacking Hsp31 paralogs (Fig 7B). Most interestingly, ∆31∆34∆ubp2 exhibit restored levels of GSH, thereby restoring redox parameters (Fig 7B).

Furthermore, the flow cytometric analysis indicated a similar pattern of MCB uptake across the indicated strains. GSH levels remain unaltered in ∆ubp2 and are comparable to WT (Fig 7C). In line with the microscopic results, flow cytometric analysis showed a significant decrement in GSH levels in ∆31∆34 strain. At the same time, ∆31∆34∆ubp2 exhibits restored GSH levels comparable to WT (Fig 7B).The complementation study further validated these results, wherein ∆31∆34∆ubp2/Ubp2↑ exhibits lesser GSH level and is comparable to ∆31∆34 (S11 Fig). Additionally, the GSH and GSSG levels were measured in the mitochondrial and cytoplasmic compartments. The GSH/GSSG ratio decreased significantly in ∆31∆34 in the mitochondria (Fig 7D) and cytoplasm (Fig 7E), indicating an increment in the oxidized glutathione levels. Opposingly, ∆31∆34∆ubp2 has a GSH/GSSG ratio comparable to WT, indicating restored redox status (Fig 7D and 7E). As the GSH and GSSG levels were perturbed in ∆31∆34, we reinstated the redox status by treating the ∆31∆34 cells with N-acetyl cysteine (NAC). We found that the mitochondrial morphology and total mass were reinstated and resembled WT (S17A and S17C Fig). Moreover, the reversal of mitochondrial phenotype in NAC-treated ∆31∆34 cells is due to a decrease in Fzo1 steady state levels (S17B Fig), thereby highlighting the role of Hsp31 paralogs in regulating mitochondrial dynamics in a redox-dependent manner.

In summary, our observations firmly establish a cross-talk between Ubp2 and Hsp31 paralogs in maintaining mitochondrial health in normal physiological conditions and under oxidative stress.

4.1 Generation of yeast strains and plasmid construction

The yeast strains utilized in this study are mentioned in S1 Table. Deletions were done using homologous recombination in the background of BY4741 (MATα his3Δ1 leu2Δ0 met15Δ0 ura3Δ0). Strains with deletions of Hsp31 paralogs (∆31, ∆34, and ∆31∆34) were generated previously in the lab [14]. Ubp2 was knocked out by generating an appropriate DNA cassette comprising regions homologous to 5’ and 3’ UTR of the gene UBP2 using primer sets P5/P6, with nat1 as the selection marker. The cassette was transformed in WT, ∆31, ∆34, and ∆31∆34 and consequently, ∆ubp2, ∆31∆ubp2, ∆34∆ubp2, ∆31∆34∆ubp2 strains were generated. Subsequently, the deletion strains were confirmed using primers P7 and P8.

For cell cycle analysis, ∆bar1 and ∆31∆34∆bar1 were generated previously in the lab [15], and ∆31∆34∆ubp2∆bar1 was generated using primers P5 and P6. All the deletions were confirmed using the primers P7/P8. Ubp2, Fzo1, and Dnm1 were HA-tagged genomically at the C-terminus by transforming the indicated strains with DNA cassettes encoding for hphNT1resistance gene with P9/P10, P1/P2, and P12/P13, respectively.

UBP2 with HA-tag at C-terminus was cloned into pRS415 plasmid with TEF promoter for complementation studies. The constructed plasmid was transformed into ∆ubp2 and ∆31∆34∆ubp2. For mitochondria visualization with a microscope, the strains were transformed with a previously constructed plasmid expressing MTS-mCherry under the influence of TEF promoter [14].

4.2 Western blot analysis

The indicated strains were subjected to growth until the log phase (A₆₀₀= 0.6-0.8 OD) or until the stationary phase (A₆₀₀= 1–2 O.D.) in necessary experiments. An equal O.D. (A₆₀₀ = 2.5 OD) was pelleted for the indicated strains and subjected to incubation with 10% TCA (trichloroacetic acid) at 4°C. It was followed by acetone washes and subjected to lysis by bead beating method. Acid-washed glass beads were added to the dried pellets along with 1X SDS dye (50 mM Tris-HCl, pH 6.8, 2% SDS, 0.1% bromophenol blue, 10% glycerol, and 100 mM β-mercaptoethanol). Consequently, the samples were heated at 92°C for 15 min, followed by a spin at 14,000 rpm at room temperature (23–25°C). The supernatant was loaded onto different percentages of SDS-polyacrylamide gels depending on the molecular weights of the protein to be separated and probed. The total protein content in each strain was quantified by staining the polyacrylamide gel and normalized for equal loading. The samples were transferred to an activated PVDF membrane for immunoblotting analysis. The blots were first blocked in 0.1% TBST skim milk, followed by multiple washes with 0.1% TBST. The blots were incubated with the respective primary and secondary antibodies (mentioned individually and in the S2 Table), followed by probing with luminol solution (BioRad).

4.3 Microscopic analysis

The mitochondrial morphology was analyzed for indicated strains by transforming the mTS-mCherry construct. Due to the mitochondrial targeting signal sequence, the mitochondria were decorated explicitly with mCherry and visualized using a Delta Vision Elite microscope at 100X magnification. The transformed cells were collected from the mid-log phase and spread on agar padding. These non-fixed cells were subjected to microscopic analysis.

To measure the ROS with specific-dyes (H₂DCFDA and MitoSox), cells from the mid-log phase were subjected to staining with the respective dyes in the dark. After incubation of the cells with dye for 30 min, the cells were washed with sterile PBS. The dye-treated cells were spread on agarose padding and subjected to microscopy analysis at 100x magnification.

4.5 Quantification of mitochondrial network morphology

ImageJ was used to measure different mitochondrial morphologies across the strains. Maximum intensity projections of all the fields were created, following which ‘thresholding’ was done to remove the background. Particles were measured by setting a scale based on the pixels by using the ‘Measure particle’ function. The particles were then analyzed based on a range using the ‘analyze particle’ function. Cells were counted per field, and the mitochondrial morphologies were scored based on the dimensions measured.

4.6 Measurement of total and functional mitochondrial mass

Total mitochondrial mass was measured by staining the cells with cardiolipin-specific dye called nonyl acridine orange (NAO). Cells from the mid-log phase (A₆₀₀= 0.6 OD) were treated with NAO for 30 min (in the dark) at 30°C at 900 rpm. To understand the role of redox status in regulating mitochondrial mass, cells were grown overnight with 5mM NAC and subjected to flow cytometry after staining with NAO. Similarly, cells from the mid-log phase were treated with a potentiometric sensitive dye called JC-1 to measure the functional mitochondrial mass. Subsequently, the cells were washed in sterile PBS and subjected to flow cytometric analysis (BD FACS Verse).

4.7 Pull down analysis

The ubiquitination status of Fzo1 was determined by the pull-down assay performed according to a previously established protocol [57]. Briefly, 200 mL of culture from the mid-log phase (A₆₀₀= 0.6 OD) was pelleted down and suspended in ice-cold immunoprecipitation (IP) buffer (1 M Tris pH = 7.5, 5 M NaCl, 0.5 M EDTA). The slurry was subjected to bead beating (1 min, four cycles) with intermittent incubation on ice. The lysed cells were centrifuged at 16000 rpm for 5 min at 4°C. The supernatant (containing the cytosolic fraction) was discarded, and the pellet (containing all the membrane fractions) was subjected to solubilization using IP buffer containing 0.5% Nonidet P-40 (NP-40). The membrane fraction was solubilized using a nutator at 4°C for 2h. Meanwhile, Protein G beads (Protein G Sepharose 4 Fast Flow beads, Catalogue no: 17061801, Cytiva) were equilibrated with IP buffer and were incubated with anti-HA antibody (Monoclonal anti-HA tag antibody, Catalogue no: SAB2702217, Sigma Aldrich) at 10 rpm, 4°C for 2h. Consequently, the antibody equilibrated agarose beads were incubated with the solubilized membrane fraction for binding (10 rpm, 4°C, 12h). The protein bound beads was eluted with loading dye and separated on 8% SDS-PAGE gel and subjected to blotting on PVDF-memebrane (37.5:1 ratio of acrylamide: polyacrylamide) [57]. The blots were incubated with either with anti-HA (Monoclonal anti-HA tag antibody, Catalogue no: SAB2702217, Sigma Aldrich) and anti-ubiquitin antibody (Ubiquitin antibody, Catalogue no: NB300–130, Novus Biologicals). Anti-mouse secondary antibody (Amersham ECL Mouse IgG, Catalogue no: NA931V, Cytiva) was used for probing blots with anti-HA antibody and anti-ubiquitin antibody. The blots were consequently probed using luminol solution (BioRad).

4.8 Estimation of ROS levels

To probe for the basal cellular and mitochondria specific ROS, cells from mid-log phase were treated with H₂DCFDA and MitoSox respectively. The dye treated cells were subjected to flow cytometry and microscopic analysis. To measure the ROS levels after extraneous stress, cells from the mid-log phase were treated with 1 mM H₂O₂. The treated cells were washed with 1X PBS and stained with the respective dyes.

4.9 Mitophagy induction

All the indicated strains were tagged with GFP at the C-terminus of mitochondrial OM45. The strains were then subjected to growth in YP-dextrose media till the mid-log phase. These cells were shifted to grow in YP-glycerol media for the indicated time points at the permissive temperature of 30°C, thereby inducing mitophagy. Lysates were collected and subjected to Western blotting. The blots were probed using anti-GFP antibody (Anti-Green Fluorescent Protein (GFP) antibody, Catalogue no: G1546, Sigma-Aldrich), followed by probing with anti-mouse secondary antibody(Amersham ECL Mouse IgG, Catalogue no: NA931V, Cytiva). The densitometric intensities of processed GFP and OM45-GFP were quantified and normalized to the corresponding loading control. Consequently, the ratio of processed GFP to total GFP (processed GFP + OM45-GFP) was calculated and plotted. Furthermore, cells undergoing mitophagic induction were subjected to microscopic analysis, and the vacuolar localization of processed OM45-GFP was detected.

5.1 Estimation of glutathione

Mitochondrial and cytoplasmic fractions were isolated and seperated from the indicated strains,as per the protocol previously published [58]. The purity of the mitochondrial fractions was checked by probing the mitochondrial extracts with anti-Tim44(expressed in mitochondrial inner membrane) antibody (Gifted, Elizabeth A Craig lab,University of Wisconsin-Madison)along with anti-Pgk1 antibody (Gifted, Pundi N Rangarajan lab, Indian Institute of Science) [59,60]. The total glutathione and GSSG levels were measured using the glutathione GSH/GSSG assay kit (Sigma, Catalogue no. MAK440) as per the indicated protocols. Mitochondrial isolates (100 µg protein) and cytoplasm (100 µg protein) were incubated with the reagents, and absorbance was recorded at the wavelength of 460nm. All the values were then calculated based on the standard curve, and the ratio was plotted.

5.2 Statistical analysis

All statistical analysis was performed using GraphPad Prism 6.0 software. Error bars represent Standard deviation (S.D) derived from at least three biological replicates.The respective statistical test performed has been performed for the experiments represented Asterisks used in the figures represent the following significance values: *, p ≤ 0.05; **, p ≤ 0.01; ***, p ≤ 0.001; ****, p ≤ 0.0001; ns, not significant.