Cell Structure and Function
Online ISSN : 1347-3700
Print ISSN : 0386-7196
ISSN-L : 0386-7196
Proteostasis stress alters mitochondrial levels and localization of SOD2 during cellular senescence
Yasuhiro Takenaka, Masataka Hirasaki, Yoshihiko Kakinuma
著者情報
ジャーナル オープンアクセス HTML
電子付録

2026 年 51 巻 2 号 p. 289-298

詳細
Abstract

Proteostasis decline is a hallmark of aging and contributes to stress-induced premature senescence (SIPS) through proteotoxic stress and mitochondrial dysfunction. Although mitochondria-derived reactive oxygen species (ROS) are critically involved in SIPS, the behavior of nuclear-encoded mitochondrial antioxidant proteins under proteostasis impairment remains poorly understood. Here, we investigated the dynamics of superoxide dismutase 2 (SOD2) during SIPS induced by the proteasome inhibitor MG132 or the lysosomal inhibitor bafilomycin A1 (BAFA1) in human fibroblasts. Both MG132 and BAFA1 induced cellular senescence and caused an early reduction of SOD2 protein levels in the mitochondrial fraction without a corresponding decrease in SOD2 mRNA or total cellular SOD2 protein levels. Immunofluorescence analysis revealed that SOD2 was partially dissociated from mitochondria under these conditions, indicating impaired mitochondrial localization rather than decreased protein abundance. Quantitative colocalization analysis confirmed a significant reduction in SOD2–mitochondria overlap, which was partially restored by short-term co-treatment with antimycin A (AA), a mitochondrial electron transport chain inhibitor that attenuates excessive ROS production. Furthermore, mislocalized SOD2 was frequently associated with aggregate-like structures, and biochemical analyses demonstrated an increase in high-molecular-weight SOD2-containing complexes under proteostasis stress. These findings suggest that proteostasis disruption is associated with early changes in the intracellular distribution and aggregate-associated accumulation of SOD2, leading to a transient reduction of mitochondrial SOD2 and potentially contributing to mitochondrial oxidative imbalance and the progression of cellular senescence.

Key words: cellular senescence, MnSOD, SOD2, mitochondria, oxidative stress, proteostasis, MG132, proteasome

Graphical Abstract

Introduction

Proteostasis is the dynamic process to maintain an optimal cellular proteome by modulating pathways of protein synthesis and protein clearance; the latter is mainly maintained by chaperones and two proteolytic systems, the ubiquitin-proteasome system (UPS) and the autophagy-lysosome system. Gradual decline of proteostasis is one of the hallmarks of aging in various organisms (Kikis, 2016; Sampaio-Marques and Ludovico, 2018; Yu and Hyun, 2021) including mammalians (Hwang et al., 2007; Meller and Shalgi, 2021), which causes a malfunction of transport, folding, and modification of de novo synthesized proteins and results in the widespread accumulation of protein aggregates and proteotoxicity.

Cellular senescence is implicated in organismal aging and age-related tissue dysfunction. In addition to replicative senescence, stress-induced premature senescence (SIPS) provides an experimental model to investigate how cellular stresses, such as oxidative stress, DNA damage, mitochondrial dysfunction, and impaired proteostasis, promote senescence-like phenotypes. Therefore, experimental perturbation of proteostasis provides a useful approach to examine the relationship between proteostasis decline and cellular senescence. Indeed, prolonged inhibition of UPS or the autophagy-lysosome system induces SIPS (Chondrogianni et al., 2003; Takenaka et al., 2022). Mitochondrial dysfunction is fundamentally involved in SIPS induced by impaired proteostasis (Takenaka et al., 2022), consistent with the role of mitochondria not only as bioenergetic factories but also as platforms for intracellular signaling (Bohovych and Khalimonchuk, 2016). Modest inhibition of mitochondrial function during SIPS progression leads to attenuation of excess ROS levels and senescent phenotypes (Takenaka et al., 2023), although how impaired proteostasis promotes mitochondrial ROS production remains to be elucidated. Also little understood is the potential role of nuclear-encoded mitochondrial antioxidant proteins such as superoxide dismutase 2 (SOD2) and glutathione peroxidase 4 (GPx4) in the progression of the SIPS by impaired proteostasis. SOD2 is a mitochondrial matrix enzyme that detoxifies superoxide and is closely related to aging and cellular senescence. Overexpression of SOD2 in Drosophila significantly extends lifespan (Sun et al., 2002), suggesting that SOD2 may play a critical role in aging. Additional evidence came from transgenic mouse study in which SOD2 overexpression decreased superoxide levels in aged mouse brain and slightly extended the lifespan (Hu et al., 2007). In contrast, mitochondrial oxidative stress caused by SOD2 deficiency promotes cellular senescence of mouse skin in vivo (Velarde et al., 2012). However, some studies failed to find a direct link between SOD2 and aging (Weyemi et al., 2012). Thus, the role of SOD2 in organismal aging and cellular senescence is complex and requires further investigation (Iakovou and Kourti, 2022). GPx4 reduces lipid hydroperoxides and protects cellular membranes, including mitochondrial membranes, from oxidative damage. However, it remains unclear how changes in localization and modification of these antioxidant proteins under stressed conditions contribute to mitochondrial ROS production and consequent cellular senescence.

To clarify these aspects, we studied protein localization and abundance of SOD2 and GPx4 in the course of cellular senescence induced by the proteasome inhibitor MG132 or the V-ATPase inhibitor bafilomycin A1 (BAFA1). Our findings indicate that both proteins are transiently depleted in mitochondria at early phase of SIPS progression, suggesting that impaired proteostasis affects the mitochondrial localization and intracellular distribution of antioxidant defense proteins. In particular, SOD2 showed an altered intracellular distribution and accumulation state under proteostasis stress. Partial inhibition of mitochondrial function by antimycin A (AA), a classical inhibitor of electron transport chain (ETC) complex III, restored the transient reduction of SOD2 protein in the mitochondrial fraction and attenuated the eventual senescent phenotype.

Materials and methods

Cell culture

MRC-5 cells at 30 population doubling levels (PDL) were obtained from the JCRB Cell Bank. The cells were maintained in EMEM (Fujifilm Wako Pure Chemical Corp, Osaka, Japan) supplemented with 10% fetal bovine serum (FBS) (Biological Industries, Cromwell, CT, USA) and penicillin-streptomycin (Nacalai Tesque, Kyoto, Japan) at 37°C in a humidified atmosphere of 5% CO2. The cells were passaged every 2–3 days. For MG132 or BAFA1 treatment, MRC-5 cells (38–40 PDL) were plated 6.4 × 105 cells per well in a 10-cm dish, allowed to grow for 24 h, and then changed the culture medium to EMEM containing 0.25 μM MG132 (ChemScene, Monmouth Junction, NJ, USA) or 13 nM BAFA1 (Sigma-Aldrich, St. Louis, MI, USA). On day 5 of drug treatment, the culture medium was replaced with regular EMEM containing no drug. Senescence-associated β-galactosidase (SA-β-gal) staining was performed as described previously (Dimri et al., 1995).

Preparation of mitochondrial fraction

The procedure for isolation of the mitochondrial fraction from cultured cells was as described by Clayton et al (Clayton and Shadel, 2014) with slight modifications. The buffer volume, number of strokes, homogenizer, and pestle were optimized for 0.5–1.0 × 106 cells. Frozen cell pellets were resuspended in hypotonic buffer and homogenized using a disposable plastic pestle (As One Corp., Osaka, Japan) with matched Safe-Lock tubes (Eppendorf, Hamburg, Germany). The detailed procedure is summarized in Fig. S1.

Immunoblot analysis

Protein concentration of mitochondrial and whole cell lysates was determined using Quick Start Bradford dye (Bio-Rad, Hercules, CA, USA). Mitochondrial (2–5 μg) and whole cell extracts (10–15 μg) were mixed with 5× SDS sample buffer containing 2-mercaptoethanol, heated at 95°C for 3 min, and then loaded onto an SDS-polyacrylamide gel (SDS-PAGE). For the immunoblot analysis of fixed SOD2 complex, cells were fixed in 4% paraformaldehyde (PFA) in phosphate-buffered saline (PBS) for 10 min at 25°C, washed with PBS, lysed in RIPA buffer, sonicated briefly, and centrifuged at 21,500 × g for 5 min. The supernatant was mixed with 5× SDS sample buffer, incubated at 95°C for 3 min or 37°C for 15 min, and loaded onto SDS-PAGE. Proteins were transferred to a PVDF membrane and incubated in 1% Western Blocking Reagent (Sigma-Aldrich) at 25°C for 1 h. The membrane was then incubated overnight with antibodies diluted in 0.5% Western Blocking Reagent. Primary antibodies used were as follows: SOD2 (sc-133134, Santa Cruz Biotechnology, Dallas, TX, USA), GPx4 (sc-166570, Santa Cruz Biotechnology), Tom40 (sc-365467, Santa Cruz Biotechnology), GAPDH (G8795, Sigma-Aldrich), ubiquitin (Ub; sc-8017, Santa Cruz Biotechnology), V5 (R960-25, Thermo Fisher Scientific, Waltham, MA, USA), p21 (sc-53870, Santa Cruz Biotechnology), p53 (sc-126, Santa Cruz Biotechnology), phosphorylated form of histone H2A.X (γH2AX) (sc-517348, Santa Cruz Biotechnology), phospho-S6 (p-S6; sc-293144, Santa Cruz Biotechnology), S6 (sc-74459, Santa Cruz Biotechnology), LC3 (PM036, MBL Life Science, Tokyo, Japan) and histone H3 (4499, Cell Signaling Technology, Danvers, MA, USA). The chemiluminescence signal was visualized with ImmunoStar LD (Fujifilm Wako Pure Chemical), and detected with LAS4000 (GE Healthcare, Chicago, IL, USA). The intensity of the obtained protein band was quantified using Fiji/ImageJ software ver. 2.0.0.

Quantitative PCR

Total RNA was isolated from MRC-5 cells using ISOGEN reagent (Nippon Gene, Tokyo, Japan) and reverse-transcribed using ReverTra Ace qPCR RT Master Mix with gDNA Remover (Toyobo, Osaka, Japan). cDNAs of human SOD2, GPx4, and 18S ribosomal RNA were amplified using THUNDERBIRD Next SYBR qPCR Mix (Toyobo) and the Thermal Cycler Dice Real Time System III (Takara Bio, Kusatsu, Japan). Relative expression levels of SOD2 and GPx4 were normalized to 18S ribosomal RNA and calculated using the comparative Ct method. Primer sequences are as follows: Hs_SOD2_RT-UP1, 5'-AACGGGGACACTTACAAATTGCT-3' and Hs_SOD2_RT-LP1, 5'-CCCAGTTGATTACATTCCAAATAGCTT-3' for SOD2; Hs_GPx4_RT-UP1, 5'-CGCGGGCTACAACGTCAAATTCGAT-3' and Hs_GPx4_RT-LP1, 5'-CCACGCAGCCGTTCTTGTCGAT-3' for GPx4; and LEM-m18S-F, 5'-CGGCTACCACATCCAAGGAA-3' and LEM-m18S-R, 5'-GCTGGAATTACCGCGGCT-3' for 18S ribosomal RNA.

Immunofluorescence

For immunofluorescence analyses, cells were fixed with either 4% PFA/PBS or methanol, permeabilized with 0.5% Triton X-100 in PBS for 10 min, and incubated with Blocking One Histo (Nacalai Tesque) for 1 h. Cells were then incubated with diluted primary antibody overnight. Primary antibodies used were as follows: SOD2 (sc-133134, Santa Cruz Biotechnology) and COX4 (PM063MS, MBL Life Science). After washing with TBS-T, cells were incubated with a secondary antibody and counterstained with Hoechst 33342. Images were acquired using an inverted fluorescence microscope IX83 or a laser scanning microscope FV1200 (Olympus, Tokyo, Japan). Fluorescence images were analyzed and quantified using Fiji/ImageJ ver. 2.0.0.

Staining of protein aggregates

Protein aggregation was visualized using ProteoStat Aggresome Detection Reagent (Enzo Life Science, Farmingdale, NY, USA), according to the manufacturer’s instructions. Cells were plated in μ-Slide 8 well chamber slide (ibidi GmbH, Gräfelfing, Germany), treated with MG132 for 24 h, fixed with 4% PFA/PBS, and permeabilized with 0.5% Triton X-100. Cells were stained with ProteoStat Reagent for 30 min at 25°C and observed under FV1200 (Olympus).

Immunoprecipitation

MRC-5 cells (~40 PDL) were treated with MG132 or BAFA1 for 24 h, and whole-cell lysates were prepared using IP lysis buffer (25 mM Tris-HCl, pH 7.4, 150 mM NaCl, 0.5% NP-40, 5% glycerol, 1 mM EDTA, and protease inhibitor cocktail [cOmplete, Roche]). Cell pellets were lysed by brief sonication and clarified by centrifugation. Lysates were pre-cleared by incubation with Dynabeads Protein G (Thermo Fisher Scientific) at 4°C for 40 min with rotation. After magnetic separation, the supernatants were transferred to new tubes. Pre-cleared lysates were incubated with anti-V5 antibody (R960-25, Thermo Fisher Scientific) for 45 min at 4°C with rotation, followed by incubation with Dynabeads Protein G overnight at 4°C. Beads were washed three times with TBS-T, and bound proteins were eluted by boiling in SDS sample buffer for subsequent immunoblot analysis.

FRAP assay

MRC-5 cells were transduced with lentiviral vectors expressing SOD2-EGFP or COX8-EGFP (pLJM1 backbone), in which full-length human SOD2 or COX8 cDNA was fused to the N-terminus of EGFP. Cells (40–45 PDL) were plated on glass-bottom dishes, treated with 0.25 μM MG132 for 24 h, and stained with 0.05 μM MitoBright LT Red (Dojindo, Kamimashiki, Japan) for 30 min. FRAP experiments were performed on live cells in HBSS at 37°C using FV1200 (Olympus) equipped with a UPlanSApo 60× oil immersion objective. A circular region of interest (ROI) was photobleached at 488 nm using full laser power, and fluorescence recovery was monitored for 90 s with images acquired every 3 s. Image sequences were processed using Fiji/ImageJ software ver. 2.0.0. Mean fluorescence intensity within the bleached ROI was measured over time and corrected for background. Data were normalized to pre-bleach intensity and plotted as relative fluorescence recovery curves. Correction for acquisition-related photobleaching, bleaching efficiency, and molecular mobility during the bleaching process was not applied; therefore, the FRAP data were not used to calculate absolute mobile or immobile fractions. Quantitative analysis and curve visualization were performed using GraphPad Prism ver. 5.0 (GraphPad Software, San Diego, CA, USA). Custom analysis of FRAP trajectories, including averaging and normalization, was performed using Fiji plugins developed by the Stowers Institute.

Statistical analysis

We conducted one-way ANOVA test followed by Tukey’s multiple comparison test. Two groups were compared using a non-parametric Mann-Whitney U test. Differences between groups with a p value of <0.05 were considered significant. All data were analyzed using the GraphPad Prism and presented as the mean (SD) of the obtained values.

Results

MG132 or BAFA1 treatment alters the intracellular distribution of SOD2

We previously established a cellular senescence model induced by MG132 or BAFA1 (Takenaka et al., 2022), in which normal human fibroblasts (MRC-5; 36–39 PDL) were treated with low concentrations of MG132 or BAFA1 for 5 consecutive days, followed by replacement with fresh medium without the inhibitors (Fig. 1A). By post-treatment day 5 (PD5), the cells exhibited a marked reduction in proliferation and gradually adopted an enlarged, flattened morphology. Induction of stress-induced premature senescence (SIPS) was confirmed by conventional SA-β-gal staining (Fig. 1B) and immunoblot analysis of senescence markers, including p21 and γH2AX (Fig. S2). Mitochondrial protein import is impaired by proteotoxic stress and is closely associated with mitochondrial dysfunction accompanied by oxidative stress (Pfanner et al., 2025; Sutandy et al., 2023). In our previous study, mitochondrial ROS levels were significantly increased as early as day 1 of MG132 or BAFA1 treatment (Takenaka et al., 2022), suggesting that the antioxidant response is insufficient to counteract drug-induced ROS. To examine this possibility, we analyzed the protein levels of two mitochondrial antioxidant enzymes, SOD2 and GPx4, in isolated mitochondrial fractions (Mt) from MG132- or BAFA1-treated cells (details are provided in Fig. S1). On day 1, SOD2 was significantly depleted in the mitochondrial fraction in MG132-treated cells (Fig. 1C and D), whereas total SOD2 protein (Fig. 1C and E) and SOD2 mRNA levels (Fig. 1F) remained unchanged compared with control cells. The levels of TOM40 and GAPDH in the mitochondrial fraction were largely unchanged upon drug treatment. In contrast, GPx4 protein levels were significantly reduced not only in the mitochondrial fraction (Fig. 1C and D) but also in the whole-cell lysate (Fig. 1C and E) on day 1. GPx4 mRNA levels were concomitantly decreased under MG132 treatment (Fig. 1F), suggesting that the reduction of GPx4 protein reflects decreased transcript levels. On day 5, protein levels of both SOD2 and GPx4 were increased in both fractions (Fig. 1D and E). Similar results were obtained in BAFA1-treated cells, in which mitochondrial SOD2 and GPx4 levels were reduced on day 1 (Fig. 1G and H), while whole-cell GPx4 protein levels were also decreased (Fig. 1I). In contrast, SOD2 and GPx4 mRNA levels were unchanged in BAFA1-treated cells on day 1 (Fig. 1J).

Fig. 1

Depletion of mitochondrial SOD2 and GPx4 during early stages of MG132- or BAFA1-treatment

(A) Experimental scheme for induction of SIPS in MRC-5 cells. Cells (38–40 PDL) were treated with 0.25 μM MG132 or 13 nM BAFA1 continuously for 5 days without medium refreshment. On day 5, the medium was replaced with inhibitor-free EMEM and cells were cultured for an additional 5 days. Day, days during drug treatment; PD, post-treatment days. (B) Representative images of untreated control and MG132- or BAFA1-treated cells stained for SA-β-gal activity (blue precipitate). Cells were stained on post-treatment day 5 (PD5). Cont, untreated young MRC-5 cells (39 PDL). Scale bar = 50 μm. (C) Immunoblot analysis of SOD2, GPx4, TOM40 (mitochondrial loading control), and GAPDH (cytosolic loading control) in MG132-treated cells. Mt, crude mitochondrial fraction; Whole, whole-cell lysate. Cont, control cells. Five μg (Mt) or 15 μg (Whole) of protein was loaded per lane. (D) Quantification of mitochondrial SOD2 and GPx4 levels in MG132-treated cells, normalized to TOM40. Values are mean ± SD (n = 4). *p<0.05. (E) Quantification of whole-cell SOD2 and GPx4 levels normalized to GAPDH. Values are mean ± SD (n = 4). ns, not significant. (F) Relative mRNA expression levels of SOD2 and GPx4 determined by quantitative RT-PCR. Four independent cDNA samples were analyzed. (G) Immunoblot analysis of BAFA1-treated cells as in (C). (H, I) Quantification of mitochondrial (H) and whole-cell (I) SOD2 and GPx4 levels in BAFA1-treated cells. Values are mean ± SD (n = 4). *p<0.05. (J) Relative mRNA expression levels of SOD2 and GPx4 in BAFA1-treated cells determined by quantitative RT-PCR.

Because GPx4 protein levels were decreased in both the mitochondrial fraction and whole-cell lysate on day 1, the reduction of mitochondrial GPx4 appeared to primarily reflect a decrease in total GPx4 abundance. In contrast, SOD2 was selectively reduced in the mitochondrial fraction without a corresponding decrease in total protein or mRNA levels; therefore, we focused on SOD2 in the subsequent analyses.

MG132 or BAFA1 treatment induces mislocalization of SOD2 from mitochondria

Mass spectrometry analyses have shown that aggregation of nuclear-encoded mitochondrial proteins, including respiratory chain subunits, occurs as an early event in MG132-treated cells (Rawat et al., 2019). To investigate the mechanism underlying the reduction of SOD2 in the mitochondrial fraction on day 1, we examined its intracellular localization. In control cells, SOD2 predominantly colocalized with the mitochondrial marker COX4. In contrast, in MG132- and BAFA1-treated cells, a fraction of SOD2 was detected in extra-mitochondrial (COX4-negative) regions on day 1 (Fig. 2A and S3), suggesting altered intracellular distribution of SOD2 under proteostasis stress. We next examined the effect of co-treatment with the mitochondrial complex III inhibitor antimycin A (AA), which has been reported to attenuate ROS levels and senescence phenotypes under these conditions (Takenaka et al., 2023). In MG132-treated cells, colocalization between SOD2 and mitochondria was significantly reduced, as quantified by Pearson’s correlation coefficient, whereas AA co-treatment partially restored this colocalization (Fig. 2B and C). Consistent with these observations, mitochondrial SOD2 levels were restored by AA co-treatment on day 1 (Fig. 2D), while total SOD2 protein levels remained unchanged (Fig. 2E). These results suggest that AA partially restores the mitochondrial localization of SOD2 under MG132-induced proteostasis stress.

Fig. 2

Mislocalization of SOD2 in MG132- or BAFA1-treated MRC-5 cells

(A) Representative immunofluorescence images of control and MG132-treated MRC-5 cells on day 1, stained for SOD2 (green), COX4 (mitochondrial marker, purple), and Hoechst 33342 (nuclei, blue). Arrowheads indicate SOD2 signals that do not colocalize with COX4. Scale bar = 100 μm. (B) Representative immunofluorescence images of MG132-treated cells with or without 1 μM antimycin A (AA) co-treatment, stained as in (A). (C) Quantification of colocalization between SOD2 and COX4 signals using Pearson’s correlation coefficient (Coloc2, Fiji/ImageJ). Data represent mean ± SD (n = 30 cells). *p<0.05. (D) Immunoblot analysis and quantification of mitochondrial SOD2 levels normalized to TOM40 in MG132-treated cells with or without AA co-treatment on day 1. Values are mean ± SD (n = 4). *p<0.05; ns, not significant. (E) Immunoblot analysis and quantification of whole-cellular SOD2 levels normalized to GAPDH in MG132-treated cells with or without AA co-treatment on day 1. Values are mean ± SD (n = 4).

MG132 or BAFA1 treatment promotes association of mislocalized SOD2 with aggregate-like structures

We next examined whether mislocalized SOD2 is associated with protein aggregate-like structures. In MG132- and BAFA1-treated cells, a fraction of extra-mitochondrial SOD2 signals colocalized with ProteoStat-positive aggregate-like structures on day 1, whereas little overlap was observed in control cells (Fig. 3A–C and S4A–C). These observations indicate that mislocalized SOD2 is at least partially associated with aggregate-like structures under proteostasis stress. To further characterize the biochemical accumulation state of SOD2, we analyzed supernatant and pellet fractions prepared under different detergent extraction conditions. However, no treatment-dependent increase in detergent-insoluble SOD2 was detected in MG132- or BAFA1-treated cells (Fig. S5A). These results suggest that MG132- or BAFA1-induced changes in SOD2 accumulation state are not accompanied by detectable accumulation of SOD2 in detergent-insoluble pellet fractions under the extraction conditions used here. We therefore examined SOD2-containing protein complexes using PFA-fixed cells. Detergent-soluble lysates prepared from PFA-fixed cells and incubated at 37°C for 15 min revealed a significant increase in high-molecular-weight SOD2 species in MG132- and BAFA1-treated cells compared with control cells (Fig. 3D). Upon heat denaturation at 95°C for 3 min, these high-molecular-weight species were largely resolved into monomeric and lower-molecular-weight forms, indicating that the observed SOD2 species represent crosslinked protein complexes. In contrast, GAPDH remained predominantly monomeric under these conditions.

Fig. 3

Mislocalized SOD2 associates with aggregate-like structures in MG132- or BAFA1-treated MRC-5 cells

(A, B) Representative images of MRC-5 cells treated with MG132 (A) or BAFA1 (B) for 1 day. Cells were immunostained for SOD2 (green) and COX4 (purple), and stained with ProteoStat (protein aggregation detection reagent, red). Nuclei were stained with Hoechst 33342 (blue). Arrowheads indicate extra-mitochondrial SOD2 signals colocalizing with ProteoStat-positive structures. (C) Representative images of untreated control cells. Scale bar = 20 μm. (D) Immunoblot analysis of SOD2-containing high-molecular-weight species in 4% PFA-fixed cells. Lysates were incubated at 37°C (15 min) or 95°C (3 min) in SDS sample buffer containing 2-mercaptoethanol, separated by SDS-PAGE, and immunoblotted (IB) with anti-SOD2 or GAPDH antibodies. C, M, and B denote control, MG132-, and BAFA1-treated cells, respectively. Right panel: Quantification of high-molecular-weight SOD2 species normalized to GAPDH. Values are mean ± SD (n = 4). *p<0.05.

We next examined potential modifications of SOD2. However, no detectable ubiquitination of SOD2 was observed on day 1 (Fig. S5B). Finally, a subset of mislocalized SOD2 signals partially colocalized with LAMP1-positive compartments in MG132-treated cells on day 1 (Fig. S6A), whereas little overlap was observed in untreated cells (Fig. S6B), suggesting a potential association of mislocalized SOD2 with lysosome-related compartments.

As a supplementary analysis, we also examined the apparent fluorescence recovery kinetics of SOD2-EGFP-containing structures using fluorescence recovery after photobleaching (FRAP). In MRC-5 cells expressing SOD2-EGFP, SOD2-EGFP signals were observed as both network-like and punctate structures (Fig. S7A). The network-like structures largely overlapped with the mitochondrial marker, whereas the punctate structures showed weaker colocalization, suggesting the presence of distinct SOD2-EGFP signal patterns. We first analyzed mitochondria-targeted COX8-EGFP as a reference marker. FRAP analysis of COX8-EGFP showed no significant difference in apparent fluorescence recovery between control and MG132-treated cells on day 1 (Fig. S7B–D). Although COX8-EGFP is not a freely diffusible mitochondrial matrix protein, this result suggests that MG132 treatment did not uniformly alter the apparent fluorescence recovery of the EGFP-fused mitochondrial proteins examined in our experimental system. We next examined SOD2-EGFP. In punctate SOD2-EGFP structures, apparent fluorescence recovery was comparable between control and MG132-treated cells on day 1 (Fig. S7E–G). In contrast, FRAP analysis of network-associated SOD2-EGFP showed reduced apparent fluorescence recovery in MG132-treated cells on day 1 compared with control cells (Fig. S7H–J). This reduction was no longer observed at post-treatment day 4 (PD4) (Fig. S7J). Together, these FRAP observations suggest altered dynamic behavior of network-associated SOD2-EGFP under MG132-induced proteostasis stress, although further analysis would be required to define its relationship to mitochondrial matrix dynamics.

Discussion

Proteostasis impairment has been shown to disrupt mitochondrial protein homeostasis, including the cytosolic aggregation of nuclear-encoded mitochondrial proteins (Rawat et al., 2019). Consistent with this notion, we observed an early reduction of mitochondrial SOD2 and GPx4 levels upon MG132 or BAFA1 treatment (Fig. 1), accompanied by increased mitochondrial ROS as previously reported (Takenaka et al., 2022). Notably, SOD2 depletion in the mitochondrial fraction occurred without a reduction in total protein or mRNA levels, indicating that post-transcriptional mechanisms underlie this phenomenon. In contrast to SOD2, GPx4 levels were reduced in both mitochondrial and whole-cell fractions, accompanied by a decrease in mRNA levels (Fig. 1), suggesting transcriptional regulation as a primary mechanism. These results highlight distinct regulatory mechanisms for mitochondrial antioxidant enzymes under proteostasis stress. Our imaging analyses further revealed that a fraction of SOD2 was redistributed to extra-mitochondrial regions under proteostasis stress (Fig. 2). Mislocalized SOD2 partially colocalized with aggregate-like structures and, to a lesser extent, with LAMP1-positive compartments (Fig. 3 and S6), suggesting that cytosolic SOD2 may be retained outside mitochondria under these conditions. However, under detergent-containing extraction conditions, SOD2 did not show detergent-insoluble pellet fractions (Fig. S5A), but instead formed high-molecular-weight species in PFA-crosslinked soluble lysates (Fig. 3D), indicating that SOD2 is associated with soluble protein complexes rather than classical insoluble aggregates, although the precise proportion of SOD2 distributed among detergent-soluble and detergent-insoluble fractions remains to be determined.

The mechanism underlying SOD2 mislocalization remains unclear. One possibility is that proteostasis stress interferes with mitochondrial protein import, resulting in cytosolic retention of SOD2. Alternatively, changes in protein conformation or interactions may alter the intracellular distribution of SOD2. Recent studies have shown that a cytosolic pool of SOD2 can interact with E3 ubiquitin ligase UBR2 under starved conditions and regulate proteasomal degradation (Ibrahim et al., 2025). In parallel, proteasome inhibition may also impair amino acid recycling, creating a pseudo-starvation-like condition. In this context, extra-mitochondrial SOD2 observed in our study may localize to such proteostasis-related compartment. In addition, co-treatment with antimycin A restored mitochondrial SOD2 levels without affecting total SOD2 expression (Fig. 2D and E), suggesting that modulation of mitochondrial or cellular redox states may influence SOD2 localization under proteostasis stress. However, the precise mechanism by which AA rescues SOD2 localization requires further investigation.

As a supplementary analysis, we performed FRAP using SOD2-EGFP-expressing cells to examine the apparent fluorescence recovery kinetics of SOD2-EGFP-containing structures. Unexpectedly, among the SOD2-EGFP signal patterns examined, network-associated SOD2-EGFP exhibited reduced apparent fluorescence recovery under proteostasis stress, whereas punctate SOD2-EGFP structures showed no significant change (Fig. S7). Moreover, this reduction in apparent fluorescence recovery was transient and recovered after removal of the stress (Fig. S7J). These findings suggest that SOD2 mislocalization and altered fluorescence recovery represent distinct but related responses to proteostasis stress, likely reflecting subtle changes in the interaction network or physicochemical state of mitochondria-associated SOD2.

This study has several limitations. The molecular mechanisms underlying SOD2 mislocalization and altered mobility remain unresolved, and potential post-translational modifications or interacting partners were not fully characterized. Further studies will be required to clarify how proteostasis stress affects mitochondrial protein homeostasis at the molecular level. In conclusion, our findings demonstrate that proteostasis stress alters the intracellular distribution and accumulation state of SOD2 during the early phase of stress-induced premature senescence. These changes may represent an early response linking impaired proteostasis to mitochondrial oxidative imbalance and cellular senescence.

Author Declaration Statements

Funding

The research was supported by JSPS KAKENHI Grant Numbers JP21K11605 and JP24K14708.

Conflict of Interest Statement

The authors declare no conflicts of interest.

Data Availability Statement

All data that support the findings of this study are available upon request from the corresponding author.

Author Contribution Statement

YT and YK designed the research and wrote the manuscript; YT performed all experiments and analyzed the data; MH assisted with the analysis of microscopic images.

Ethics Approval and Consent to Participate

Ethics approval is not required in this study.

Patient Consent for Publication

Not applicable.

Acknowledgments

We are grateful to Ms. Miyuki Komura (Evident/Olympus) for her technical assistance with the FRAP experiments. We are also grateful to Santa Cruz Biotechnology, Inc. (Dallas, TX, USA) and Cosmo Bio Co., Ltd. (Tokyo, Japan) for providing sample antibodies.

References
Abbreviations

AA

antimycin A

BAFA1

bafilomycin A1

COX4

cytochrome c oxidase subunit 4

COX8

cytochrome c oxidase subunit 8

Ct

threshold cycle

EGFP

enhanced green fluorescent protein

EMEM

Eagle’s minimum essential medium

ETC

electron transport chain

FBS

fetal bovine serum

FRAP

fluorescence recovery after photobleaching

GAPDH

glyceraldehyde-3-phosphate dehydrogenase

GPx4

glutathione peroxidase 4

HMW

high-molecular-weight

IB

immunoblotting

LAMP1

lysosomal-associated membrane protein 1

PBS

phosphate-buffered saline

PD

post-treatment day

PDL

population doubling level

PFA

paraformaldehyde

PVDF

polyvinylidene difluoride

qPCR

quantitative PCR

ROI

region of interest

ROS

reactive oxygen species

RT-PCR

reverse transcription PCR

SA-β-gal

senescence-associated β-galactosidase

SIPS

stress-induced premature senescence

SOD2

superoxide dismutase

TBS-T

Tris-buffered saline containing Tween 20

TOM40

translocase of the outer mitochondrial membrane 40

UPS

ubiquitin-proteasome system

 
© 2026 The Author(s)

This is an open access article distributed under the terms of the Creative Commons BY (Attribution) License (https://creativecommons.org/licenses/by/4.0/legalcode), which permits the unrestricted distribution, reproduction and use of the article provided the original source and authors are credited.
https://creativecommons.org/licenses/by/4.0/legalcode
feedback
Top