Efficient PHB2 (prohibitin 2) exposure during mitophagy depends on VDAC1 (voltage dependent anion channel 1)

ABSTRACT

Exposure of inner mitochondrial membrane resident protein PHB2 (prohibitin 2) during autophagic removal of depolarized mitochondria (mitophagy) depends on the ubiquitin-proteasome system. This uncovering facilitates the PHB2 interaction with phagophore membrane-associated protein MAP1LC3/LC3. It is unclear whether PHB2 is exposed randomly at mitochondrial rupture sites. Prior knowledge and initial screening indicated that VDAC1 (voltage dependent anion channel 1) might play a role in this phenomenon. Through in vitro biochemical assays and imaging, we have found that VDAC1-PHB2 interaction increases during mitochondrial depolarization. Subsequently, this interaction enhances the efficiency of PHB2 exposure and mitophagy. To investigate the relevance in vivo, we utilized porin (equivalent to VDAC1) knockout Drosophila line. Our findings demonstrate that during mitochondrial stress, porin is essential for Phb2 exposure, Phb2-Atg8 interaction and mitophagy. This study highlights that VDAC1 predominantly synchronizes efficient PHB2 exposure through mitochondrial rupture sites during mitophagy. These findings may provide insights to understand progressive neurodegeneration.

Introduction

Autophagic degradation of mitochondria (mitophagy) is one of the ways by which defective mitochondria are eliminated [1–3]. This is an essential process that upholds the quality of the mitochondrial population within the cellular environment [4–6]. The presence of various routes to support mitophagy underscores the significance of this process in ensuring the survival of cells [7]. In this context, one of the extensively investigated pathways is PINK1 (PTEN induced kinase 1)-PRKN (parkin RBR E3 ubiquitin protein ligase)-mediated mitophagy [1,2]. Mutations in these proteins are closely associated with neurodegenerative conditions, notably Parkinson disease (PD) [8,9]. It is unequivocally acknowledged that following mitochondrial depolarization, PINK1 recruits PRKN to the outer mitochondrial membrane (OMM), where PRKN subsequently ubiquitinates various target proteins. PINK1 can phosphorylate multiple OMM proteins, with some of them serving as docking sites for PRKN, yet only a limited subset of these proteins have been identified as fulfilling this particular function [10,11]. From an alternate standpoint, it is proposed that PINK1 might phosphorylate a preexisting ubiquitin chain on OMM, subsequently initiating the recruitment of PRKN. As a result, the substrate undergoes further ubiquitination, thereby amplifying the process. Mitochondrial E3 ligases (MARCHF5/MITOL [membrane associated ring-CH-type finger 5], for example) are essential for such initial seeding effects, but in the absence of PRKN, they alone are insufficient to induce mitophagy [12,13]. Up to this stage, the process can be partially reversed, as the ubiquitination mediated by PRKN is known to be counteracted by various deubiquitinase enzymes [14]. If ubiquitinated proteins on OMM are not subjected to deubiquitination, they can either be degraded by the Ubiquitin-proteasome complex or associate with SQSTM1/p62 (sequestosome 1) [15,16]. SQSTM1, serving as a receptor protein, enables the linkage between ubiquitinated proteins and MAP1LC3/LC3 (microtubule associated protein 1 light chain 3). Earlier research has indicated that the ubiquitin-proteasome system is equally vital in depolarization-induced mitophagy, although the precise rationale remained unclear [17,18]. Recently, it has come to light that mitophagy requires proteasome activity-dependent rupture of OMM and exposure of inner mitochondrial membrane (IMM) protein PHB2 (prohibitin 2) [19]. PHB2 can function as an LC3 receptor. Unlike cardiolipin, which translocate to OMM during depolarization [20], OMM rupture site formation is important for such PHB2 binding with LC3 [19].

PHB1 and PHB2 are membrane proteins that are highly conserved and exist as multimeric ring complexes within mitochondria. They are involved in several crucial processes, such as mitochondrial shaping, biogenesis and maintenance of mitochondrial DNA [21–23]. Notably, PHB2 has the intriguing capacity to control the release of CYCS (cytochrome c, somatic) by safeguarding OPA1 (OPA1 mitochondrial dynamin like GTPase) long forms from mitochondrial proteases. Although cells lacking PHB2 can survive, they are highly sensitized to apoptotic stimuli [23]. Recently it was revealed that the PHB complex’s scaffolding effect on PARL (presenilin associated rhomboid like)-PGAM5 (PGAM family member 5, mitochondrial serine/threonine protein phosphatase) can also regulate PINK1 processing [24]. The significance of PHB2 in mitophagy is further emphasized by another study, which showed that PRKN can ubiquitinate PHB2 when OMM is disrupted [25].

While several studies have investigated the PINK1-PRKN pathway up to the stage of OMM rupture and subsequent PHB2 exposure, none elucidated how PHB2 remains in close proximity to these sites of rupture. Here, we investigated two possible explanations: i) PHB2 exposure after mitochondrial depolarization is a random event or ii) PHB2 interacts with an OMM PRKN target, thus degradation of such a protein may increase the efficiency of PHB2 exposure. If this interacting protein on OMM could form oligomers (or form a multimer with other OMM PRKN targets), it may further enhance the efficiency.

We utilized an unbiased approach to identify a potential partner of PHB2 that could be targeted for ubiquitination by PRKN. Our study reveals that during stress-induced mitophagy, VDAC1 (voltage dependent anion channel 1) can facilitate the exposure of PHB2 to LC3. This study enhances our understanding of the underlying mechanisms in mitophagy, and the implications are broad. For example, our findings have the capacity to explain issues related to mitochondrial quality control in progressive neurodegenerative disorders.

Results and discussion

Association of VDAC1 with PHB2 increases following mitochondrial depolarization

The primary aim of our study was to identify possible candidates capable of interacting with both PHB2 and PRKN. We reasoned that proteasome mediated degradation of such a protein on the depolarized mitochondrial surface could potentially result in the exposure of PHB2. Fig. S1A depicts different possible scenarios illustrating variations in PHB2 exposure following mitochondrial depolarization and the degree of interaction with LC3. The scenarios depicted are as follows: (1) PHB2 is randomly exposed after the degradation of the PRKN substrate, (2) the PRKN target can oligomerize and PHB2 is available near the OMM rupture site, (3) the PRKN target can oligomerize, but PHB2 is not abundantly located near the rupture site, (4) the substrate oligomerizes and engages with PHB2, (5) the PRKN substrate does not undergo oligomerization but interacts with PHB2. Our hypothesis aligns with the notion that scenarios 2, 4, and 5 could potentially result in significantly greater PHB2 exposure. We utilized BioGrid open database to identify a total of 233 proteins that could interact with both PRKN and PHB2 (Fig. S1B). Out of these, 88 are mitochondrial proteins and 18 of them localize on OMM (Mitocarta 2.0 database). Importantly, we conducted GEO analysis (from GEO data set GSE20333 and GSE20186, Fig. S1C and D respectively) on these common interactors and found significantly low expression of VDAC1 in substantia nigra (SN) region of PD brain. This is noteworthy since compromised mitophagy is strongly linked with progressive neurodegeneration in PD. Next, mitochondria were isolated from rat brain and divided in two equal portions. In one portion mitochondria were depolarized with carbonyl cyanide m-chlorophenylhydrazone (CCCP; 10 µM, 20 min). The other portion received equal amount of vehicle. We used protein concentrators with 50- or 100-kDa cut-off polyethersulfone membranes to segregate the mitochondrial proteins from the homogenate. VDAC1 has a molecular mass of ~ 34 kDa. Hence, if VDAC1 forms complexes with proteins of a molecular mass ranging from 15–20 kDa or higher, these membranes will obstruct their passage. We observed a significantly higher retention of VDAC1 and VDAC2 (voltage dependent anion channel 2) by the 50 and 100 kDa cut-off membranes when mitochondria were depolarized, suggesting possible oligomer or complex formation (Fig. S1E, Data S5). We selected VDAC1 for further investigation due to its high abundance on OMM and its high affinity towards PRKN [14]. CCCP treatment (10 µM) did not alter total level of VDAC1 in isolated rat brain mitochondria (Fig. S1E, Data S5).

Next, we wanted to reconfirm that VDAC1 can interact with PHB2. As VDAC1 and PHB2 share similar molecular mass (~34 kDa), to avoid false positive signals we utilized WT or VDAC1 KO HEK 293T cells. We transfected VDAC1 KO cells with VDAC1-3×FLAG and immunoprecipitated FLAG from the samples depicted in Fig. S1F. After immunoblotting, a shift of VDAC1 signal was noted in VDAC1-3×FLAG transfected samples when compared to the WT untransfected cells (Fig. S1F, Data S5). This is due to the additional ~ 3 kDa 3XFLAG attachment with VDAC1. The signal for VDAC1 also overlapped with that of FLAG. In the VDAC1-3XFLAG transfected samples we detected PHB2 signal at similar positions when compared to the WT untransfected cell samples. Protein samples from untransfected VDAC1 KO cells did not show a VDAC1-3XFLAG-PHB2 signal (Fig. S1F, Data S5).

Previously it was noted that VDAC1 exhibits weak associations with one another, making it challenging to detect the oligomeric forms (or complexes with PHB2, in this instance) by SDS-PAGE unless chemically crosslinked [26,27]. At first, we treated isolated rat brain mitochondria with CCCP (10 µM, 20 min) and prepared EGS crosslinked protein samples for SDS-PAGE. Immunoblot analysis showed prominent overlap between VDAC1 and PHB2 signals at ~ 34, ~72, ~130 and ≥ 180 kDa (Figure 1A, Data S1). Following CCCP treatment, we consistently observed an increase in VDAC1-PHB2 signal at ~ 130 kDa and ≥ 180 kDa (Figure 1A). To confirm that the expression levels of two proteins are similar, non-crosslinked samples were incubated with β-mercaptoethanol and separated by SDS-PAGE. Immunoblots could not detect any alteration in the total protein levels (Figure 1A, Data S1). To investigate this phenomenon under endogenous conditions we utilized SH-SY5Y cell line. We found that VDAC1 levels or mitochondrial mass did not reduce within 2 h of CCCP (10 µM) treatment (Fig. S2A, and B, Data S6). We did not find any notable OMM rupture within this time period of CCCP treatment. Interestingly, increased contact points between OMM and IMM were evident (Fig. S2C). We observed an increase in high molecular mass VDAC1 and PHB2 signals (at ~ 72, ~ 130 and ≥ 180 kDa) after 2 h CCCP treatment in SH-SY5Y cells (Figure 1B). To confirm whether the heightened colocalization signals of VDAC1 and PHB2 in the immunoblots result from their interaction, we subjected isolated rat brain mitochondria to CCCP treatment (for 20 min) and subsequently conducted immunoprecipitation for either VDAC1 or PHB2. Likewise, SH-SY5Y cells were exposed to CCCP for 2 h, mitochondria were isolated and the experiment was repeated. In both scenarios, we observed an increased interaction between VDAC1 and PHB2 upon CCCP treatment (Figure 1C, Data S1). For further confirmation, we designed a split GFP based VDAC1-PHB2 interaction sensor, where GFP moiety 1–10 was attached to VDAC1 C-terminal and GFP β strand 11 was attached to N-terminal end of PHB2. These two parts of GFP remained non-fluorescent (Fig. S2D). However, GFP can self-assemble and demonstrate prominent fluorescence when VDAC1 and PHB2 come in close proximity. VDAC1-GFP (1–10) and GFP (11)-PHB2 were expressed in VDAC1 KO HEK 293T cells. Control HEK 293T cells (without CCCP) demonstrated puncta like GFP signals inside mitochondria (Figure 1D). 20 min CCCP treatment led to heighted GFP signal, signifying increased self-assembly of GFP and thus higher interaction between VDAC1 and PHB2 (Figure 1D).

Figure 1.

VDAC1-PHB2 complex formation during CCCP induced depolarization. (A) isolated rat brain mitochondria are treated with CCCP (10 µm) for 20 min, crosslinked (CS) with EGS and subjected to immunoblotting. Non-crosslinked samples from the same mitochondrial pool were incubated with β mercaptoethanol (βME) and immunoblotted after SDS PAGE. ~72 (appeared between 72 and 55 kDa), ~130 and ≥ 180 kDa band intensity from the crosslinked samples were normalized by ~ 34 kDa band intensity and represented as bar graphs. Arrowheads demarcate ~ 72, ~130 kDa bands and (}) indicate ≥ 180 kDa bands. The experiment was repeated for six times. Experiments with non-crosslinked samples were repeated three times. (B) SH-SY5Y cells are treated with CCCP (10 µm) for 2 h and isolated EGS CS mitochondria are processed for immunoblotting. ~72, ~130 and ≥ 180 kDa band intensity were normalized by ~ 34 kDa band intensity and represented as bar graphs. The experiment was repeated for 4 times. (C) isolated rat brain mitochondria are treated with CCCP (20 min), or isolated mitochondria from 2-h CCCP treated SH-SY5Y cells are immunoprecipitated for either VDAC1 or PHB2. After SDS-PAGE, immunoblots are probed with the indicated antibodies. The experiment was repeated 3 times. (D) VDAC1 KO HEK 293T cells expressing VDAC1-GFP (1-10), GFP (11)-PHB2 and mitoBFP are treated with vehicle or CCCP (10 µm, 20 min). 3D rendered z stack images are represented to demonstrate increased GFP signal after CCCP treatment. The experiment was repeated for 3 times and at least 45 cells were considered. Scale bar: 10 µm. GFP intensity was normalized by MitoBFP intensity and mean values are represented as the bar graph. (E) rat brain mitochondria are treated as mentioned in a and crosslinked with DSP or BMH. Immunoblots represent VDAC1 and PHB2 localization after SDS-PAGE. (*) indicate the missing signal of VDAC1 in the BMH crosslinked samples. (F) mitochondria isolated from control or VDAC1 knockout (KO) HEK 293T cells are treated as mentioned in C and crosslinked as mentioned in figure. Immunoblots are probed with VDAC1 and PHB2 antibodies. Arrowhead demarcates ~ 130 kDa band. Bar graphs represent mean ± SEM. Student’s t test. *p ≤ 0.05, **p ≤ 0.01, ***p ≤ 0.001 when compared to control group.

VDAC1-PHB2 complex constitutes the central element within the larger molecular assemblies

Next, we wanted to determine whether this increased signal of VDAC1 at ~ 130 and ≥ 180 kDa after CCCP treatment is dependent on the presence of PHB2 or not. Knockdown of PHB2 can alter mitochondrial morphology, polarization and membrane architecture [22,23,28]. Given that mitochondria are already compromised, the impact of CCCP will not be decipherable in a PHB2 null background. So, we utilized the fact that rat PHB2 does not have cysteine residue and thus should remain insensitive to BMH crosslinking. Although CCCP treated samples exhibited increased VDAC1 or PHB2 signal at ~ 130 and ≥ 180 kDa after DSP crosslinking, no such effect was found after BMH crosslinked samples (Figure 1E). Rat VDAC1 contains two cysteines, thus VDAC1 signal at ~ 72 kDa in BMH crosslinked samples represent VDAC1 dimer lacking any PHB2. Interestingly, VDAC1 signal at ~ 130 kDa was not visible in BMH crosslinked samples, indicating the necessity of PHB2 to form the complex at ~ 130 kDa (Figure 1E). However, this could also potentially be due to an unknown interactor of VDAC1 that lacks cysteine residues. So, we addressed this issue from the reverse angle. To determine whether increase in PHB2 signal at ~ 130 and ≥ 180 kDa after CCCP treatment is VDAC1 dependent or not, we used VDAC1 KO HEK 293T cells. The total amount of PHB2 does not change in absence of VDAC1 (Fig. S3A-C, Data S7). We isolated mitochondria from VDAC1 KO HEK 293T cell lines and treated with CCCP (10 µM, 20 min). The inter-reliant alteration of PHB2 and VDAC1 signals (at ~ 130 and ≥ 180 kDa) due to depolarization was visible in mitochondrial samples which were isolated from VDAC1 WT HEK 293T cells (Figure 1F, Data S1). Although DSP crosslinked protein samples from WT cells showed increased VDAC1 signal at ~ 130 kDa after CCCP treatment, BMH crosslinking demonstrated no such effects. On the other hand, increased PHB2 signals were detected (at ~ 130 kDa) when protein samples from WT cells were crosslinked with DSP. As anticipated, PHB2 signal was not detectable when BMH was used for crosslinking. Interestingly, PHB2 signals did appear after DSP crosslinking at ~ 130 and ≥ 180 kDa even when VDAC1 was absent. However, the CCCP induced increased signal at ~ 130 kDa were not observed. The PHB2 signals observed at the same locations in the absence of VDAC1 might be a result of PHB2 binding to other proteins and these complexes remained unaffected by mitochondrial depolarization.

In order to ascertain if VDAC1 oligomerization is necessary for the formation of the ~ 130 kDa complex, we pre-treated rat brain mitochondria with VDAC1 oligomerization inhibitor 4′-diisothiocyanostilbene-2,2′-disulfonic acid (DIDS) [26,27,29]. DIDS effectively inhibited the complex formation of VDAC1 at ~ 130 kDa in response to CCCP treatment (Fig. S3D, Data S7). If PHB2-VDAC1 (with a combined molecular mass of ~ 68 kDa) serves as the fundamental component of ~ 130 kDa or larger complexes, the breakdown of these complexes should result in signals corresponding to their individual molecular mass (~34 kDa). DSP crosslinked samples showed signals for VDAC1 and PHB2 at ~ 34, ~72 and ≥ 130 kDa in 2D immunoblots when the SDS PAGE was done in non-reducing conditions (Fig. S3E). Complete disintegration of DSP crosslinked samples by β mercaptoethanol exhibited VDAC1 and PHB2 signals only at ~ 34 kDa (Fig. S3E) in immunoblots after 2D SDS-PAGE. After first dimension SDS-PAGE, partial reduction of EGS crosslinked samples demonstrated clear signals of VDAC1 and PHB2 at ~ 34 and ~ 72 kDa (Fig. S3F). Smear like signals above 72 kDa represent intact or partially disintegrated complexes. Further reduction of ~ 72 kDa complexes (indicated by rectangle, Fig. S3F) yielded signals of VDAC1 and PHB2 at ~ 34 kDa (Fig. S3F).

VDAC1 is required for PHB2 exposure and efficient mitophagy after CCCP treatment

Next, we investigated whether VDAC1 KD, KO or inhibiting oligomerization could impede PHB2 exposure and subsequently halt mitophagy. As previously shown, absence of VDAC1 does not interfere with depolarization induced PRKN localization to mitochondria (Fig. S3G) [16]. We monitored mitophagy at four different stages as mentioned in Fig. S4A. We employed protease (trypsin) protection/digestion assay to assess the integrity of OMM. Fig. S4B demonstrates three different scenarios where trypsin has differential access to TOMM20 (translocase of outer mitochondrial membrane 20; OMM), PHB2 (IMM), ATP5F1A/ATP5A (ATP synthase F1 subunit alpha; IMM-cristae) and HSPD1/HSP60 (heat shock protein family D (Hsp60) member 1; matrix), when incubated for a specific time period and at a specific temperature. Trypsin digestion assay in SH-SY5Y cells revealed that ATP5F1A, is partially digested by trypsin after 4 h of CCCP treatment. This might indicate the beginning of robust OMM breakdown (Fig. S4C and S4B- second scenario, Data S8) [19]. Co-treatment of proteasome inhibitor Mg-132 protected against CCCP-induced PHB2 exposure at 4 h (Fig. S4D, Data S8). VDAC1 oligomers can form pores big enough to release mtDNA (macro-pore) [26]. These macro-pores might uncover PHB2 situated near VDAC1 oligomers, without requiring OMM rupture by proteasome (Fig. S4B- third scenario). Treating isolated mitochondria with CCCP did not increase the sensitivity of PHB2 or other inner components like ATP5F1A and HSPD1 toward trypsin (Fig. S4E, Data S8). Thus, PHB2 exposure through VDAC1-macro pore formation was not perceived. After 4 h CCCP treatment, PHB2 sensitivity toward trypsin demonstrated that OMM is disintegrated in SH-SY5Y or HEK 293T cells (Figure 2A, Data S2). Interestingly, PHB2 is not efficiently exposed to trypsin after CCCP treatment in VDAC1 KD or KO cells when compared to the control (Figure 2A, Data S2). DIDS treatment also inhibited CCCP treatment induced PHB2 exposure in SH-SY5Y cells (Figure 2A, Data S2).

Figure 2.

VDAC1 is required for depolarization induced PHB2 exposure and mitophagy. (A) immunoblots represent protease protection assay for PHB2, mitochondrial inner membrane (ATP5F1A) and matrix (HSPD1) markers in the cell lines after the mentioned treatments. Cells were incubated with CCCP (10 µm)/dids (100 µm) for 4 h. Bar graphs represent mean fold change (± SEM). *p ≤ 0.05, ***p ≤ 0.001 when compared to control group. N = 3, Student’s t test. (B) Representative images of duo link proximity ligation assay for PHB2 and LC3 in SH-SY5Y cells after the mentioned treatments. Cells are treated with DIDS (100 µm) and/or CCCP (10 µm) for 4 h. Scale bar: 10 µm. Bar graphs represent mean number of PLA puncta (± SEM)/cell. ***p ≤ 0.001 when compared to control group. ^##p ≤0.01/^###p ≤0.001 when compared to the only CCCP treated group. The experiment is repeated three times. One-way ANOVA followed by Tukey’s multiple comparisons test. (C) SH-SY5Y or HEK 293T cells are treated with CCCP for 8 h. DIDS is co-treated with CCCP in the mentioned group. Immunoblots for the mentioned proteins are representative of at least three different experiments. Bar graphs represent mean fold change ± SEM. n = 3, *p ≤ 0.05, ***p ≤ 0.001, Student’s t test. (D) quantification of mitophagic cells after transfecting mitoKeima is performed by FACS analysis. Bar graphs represent mean fold change (± SEM) in number of high mitophagic cells after 8 h of CCCP (10 µm) treatment in the mentioned cell lines. The experiment is repeated 3 times. **p ≤ 0.01, ***p ≤ 0.001, Student’s t test.

PHB2 regulates mitophagy mostly by interacting with LC3. To determine whether the compromised unmasking of PHB2, as indicated by the protease protection assay, can affect its interaction with LC3 we performed a proximity ligation assay (PLA). PLA showed that the interaction between PHB2 and LC3 was dependent on the presence and oligomerization of VDAC1 (Figure 2B). Our findings propose a model wherein LC3 interact with PHB2 at the sites of OMM rupture formed due to the degradation of VDAC1. To characterize this further, we used structured illumination microscopy (SIM). GFP-LC3 transfected SH-SY5Y cells were treated with CCCP and we selected two time points for imaging– 2 h and 4 h. After 2 h of CCCP treatment, alterations in VDAC1 levels were minimal, whereas at 4 h, VDAC1 was significantly reduced (Fig. S4F). We did not detect GFP-LC3 puncta near mitochondria in control cells (Fig. S4F). However, at 2 h of treatment, there was increased GFP-LC3 near mitochondria, although the signal did not overlap with PHB2. Interestingly, after 4 h of CCCP treatment, the VDAC1 signal was reduced and GFP-LC3 overlapped with PHB2. When cells were simultaneously treated with CCCP and Mg-132 for 4 h, although GFP-LC3 was present near mitochondria, it failed to overlap with the PHB2 signal (Fig. S4F). We also determined the level of ATP5F1A and HSPD1 (which reflect mitochondrial mass) in different experimental conditions. Clear decrease in ATP5F1A or HSPD1 was noticed after CCCP treatment, in control SH-SY5Y or HEK 293T cells. However, VDAC1 KD/KO or DIDS treatment inhibited this decrease in mitochondrial mass (Figure 2C, Data S2). Furthermore, the number of cells with high level of mitophagy was quantified after transfecting the mentioned cell lines with mitoKeima (Figure 2D). We found that VDAC1 was necessary for the increase in high mitophagic cells in response to CCCP treatment (Figure 2D).

Notably, we discovered that VDAC1 oligomerization is also crucial for mitophagy, as the mere presence of VDAC1. However, it is not clear why a single molecular interaction between VDAC1 and PHB2 is not sufficient in this scenario. It is possible that an elevated quantity of VDAC1 in the complex and its subsequent degradation by the proteasome complex could lead to larger rupture sites on the OMM (macro-slit). This might potentially increase the effectiveness of PHB2-LIR exposure (as suggested in Fig. S1A, 4th column).

Porin-PHB2 complex formation increases during rotenone induced mitophagy, in vivo

To investigate the role of VDAC1 in mitophagy in vivo, we employed Drosophila as a model organism. Drosophila has been extensively used to delineate mitochondrial elimination pathways and as CCCP cannot be fed to the flies, many of these studies administered rotenone to induce mitophagy [1,30,31]. Similar to previous studies [30], we also found that 1 mm rotenone exposure can induce mitophagy, as assessed in live wings or thoracic muscle of mitoKeima expressing flies (Figure 3A). Results obtained from fixed thoracic muscles in mCherry-Atg8a/GABARAP (GABA type A receptor-associated protein) expressing flies confirmed our findings, where rotenone treatment led to an increase in mCherry, Atg8a and blw (Drosophila equivalent to ATP5F1A) overlapping signals (Figure 3B). Transmission electron microscopy (TEM) images also confirmed the existence of mitophagic structures in fly thorax following 2 d of rotenone treatment (Figure 3C). To determine whether rotenone induced mitochondrial stress can also lead to increased porin (equivalent to mammalian VDAC)-Phb2 complex formation, we isolated mitochondria from control and rotenone-treated Drosophila and crosslinked with EGS for immunoblotting. As seen in our in vitro experiments, rotenone treatment increased porin and Phb2 signals at ~ 72, ~130 and ≥ 180 kDa (Figure 3D). Immunoprecipitation of porin or Phb2 from isolated Drosophila mitochondria likewise revealed increased interaction between these proteins when flies were exposed to rotenone (Figure 3E and Fig. S4G, Data S3).

Figure 3.

Rotenone treatment induces mitophagy and increases porin-Phb2 complex formation in Drosophila. (A) images showing mitoKeima signal from live fly wing (Act5C-GAL4:UAS mitoKeima) after 1 d or 2 d rotenone (1 mm). For wing, images are taken from the indicated region in bright field (scale bar: 10 µm). Images for mitoKeima were also taken from live thoracic muscle after 2 d of rotenone treatment. For thoracic muscle mitochondria, we dissected the thorax and isolated muscles are imaged after submerging in phosphate buffer saline, pH 7.0 (scale bar: 20 µm). White arrowheads indicated mitophagic structures. (B) Drosophila expressing Atg8a-mCherry (Act5C-GAL4:UAS-mCherry-Atg8a) are anaesthetized after 1 d or 2 d of rotenone treatment. Fixed fly thorax is immunostained for blw, mCherry and Atg8a. Images are captured from the indicated portion of thorax as mentioned in bright field or fluorescent image of the whole body (scale: 100 µm) and the representative merged images are provided. White arrowheads indicate colocalization of mCherry, blw and Atg8a signals, which appeared as bright white puncta (scale bar: 20 µm). (C) fixed unstained thorax samples are processed for transmission electron microscope imaging and the selected region from the rotenone treated group demonstrate presence of clear mitophagic structures after 2 d of rotenone treatment (1 mm). Scale bar is as indicated in the images. (D) isolated mitochondria from control and rotenone treated Drosophila are crosslinked with EGS and immunoblotted for porin and Phb2 as mentioned previously. Arrowheads indicate the ~ 70 and ~ 130 kDa signals. (}) indicate signals which appeared above 180 kDa molecular mass marker. Immunoblots are representative of three different experiments. (E) Drosophila treated with rotenone (1 d) are anaesthetized and mitochondrial protein samples are processed for immunoprecipitation. At least 20 flies are processed for individual experiment. Representative blots demonstrate signal for the mentioned proteins after co- immunoprecipitation (IP) for either porin or Phb2. Immunoblots are representative of three different experiments.

Porin is required for rotenone induced PHB2 exposure and mitophagy in Drosophila

Drosophila has four different isoforms of VDAC, out of which porin shows the highest similarity with mammalian VDAC1. Although porin and porin2 have similar functions, porin is expressed ubiquitously, whereas porin2 is restricted mostly to sperm cells [32,33]. To determine whether porin is required for rotenone induced OMM rupture, Phb2 exposure and mitochondrial elimination, in vivo, we utilized WT, porin KO (A2/A2) and porin reintroduced/revertant (RV) Drosophila lines (Fig. S4H, Data S8) [34]. Thoracic mitochondria of A2/A2 flies appeared different from WT or RV lines (Fig. S4I). We treated these flies with rotenone for 1 d and mitochondria were isolated. When exposed to trypsin, Phb2 displayed sensitivity in WT or RV fly lines (Figure 4A, Data S4). In A2/A2 flies, Phb2 was not degraded by trypsin (Figure 4A, Data S4). Proximity ligation assay for Atg8a/LC3 and Phb2 showed increased proximity after rotenone treatment in the thoraces of WT and RV Drosophila. However, there were no significant changes detected in the thorax of A2/A2 flies (Figure 4B). Confocal microscopic images also revealed the presence of increased mitophagic structures in the thorax of WT and RV flies following 2 d of rotenone treatment (Figure 4C and Fig. S4J). TEM images confirmed the presence of mitophagic structures in these flies (Figure 4C). No clear alterations are observed in A2/A2 fly thorax (Figure 4C and Fig. S4J). Immunoblot analysis showed a decrease in blw and Hsp60 in WT or RV flies after a two-day rotenone treatment. There were no significant changes observed in A2/A2 flies (Figure 4D, Data S4).

Figure 4.

Vdac/porin is necessary for rotenone induced mitophagy in Drosophila. (A) blots demonstrate protease protection assay of Phb2 in mitochondria isolated from WT, porin KO (porin A2/A2) and porin revertant (porin RV) Drosophila lines, after 1 d of rotenone treatment. Bar graphs represent mean ± SEM. *p ≤ 0.05. n = 3-4, Student’s t test. (B) Representative images of duo link proximity ligation assay (PLA) for Phb2 and Atg8a (red) in control or rotenone (2 d) treated fly thorax of the mentioned lines are provided. Phase contrast and DAPI images are merged with the PLA signals. Bar graphs represent mean number of PLA puncta (± SEM). *p ≤ 0.05, **p ≤ 0.01. n = 4-6. Student’s t test. (C) fixed thoracic samples of mentioned fly lines after 2 d of rotenone treatment are processed for immunostaining or for transmission electron microscopy. Confocal microscope images represent colocalization of blw (red) and Atg8a (green). Scale bar: 10 µm. Arrowheads in transmission electron microscopic images indicate mitophagic structures. Scale bar as indicated in the image. At least 5 fly thoraces are imaged. (D) mitochondrial marker proteins (blw and Hsp60) from Drosophila lines are quantified by immunoblotting after 1 d and 2 d of rotenone treatment. Bar graphs represent normalized mean intensity (± SEM) for blw and Hsp60. *p ≤ 0.05, **p ≤ 0.01. n = 3-4. One-way ANOVA followed by Dunnett’s multiple comparisons test.

Multiple research groups have reported that the activity of proteasome complex is required for breakdown of OMM at discrete sites during mitophagy. Discovery of the LC3 receptor – PHB2 exposure through these sites is relatively new and thus the efficiency of this unmasking process requires further characterizations. In this study, we report that the interaction between VDAC1 and PHB2 is crucial for effectively exposing PHB2 during depolarization or stress-induced mitophagy. The impact of this complex on mitophagy is also VDAC1 oligomerization dependent. Therefore, we propose that the degradation of oligomeric state of VDAC1 leads to the breakage of the OMM, creating a sufficient opening to expose PHB2. Our immunoprecipitation and protein crosslinking-based immunoblot assays suggest that VDAC1-PHB2 (~64 kDa) may serve as the fundamental unit, while a ~ 130 kDa component represents the simplest form of this complex which is relevant to PHB2 exposure and mitophagy. Importantly, our findings are applicable in an in vivo system where rotenone treatment leads to a similar increased interaction between Phb2 and porin. Additionally, the presence of porin is necessary for unmasking of Pbh2 and mitophagy in Drosophila, confirming the reproducibility of our in vitro findings. It is important to note that we employed CCCP or rotenone to induce stress in the system and investigated the influence of VDAC1-PHB2 complex during enforced mitophagy. Whether this interaction remains significant in the context of normal physiological mitophagy in different cell lineages during aging will necessitate further investigation.

An area that remained open for further study is whether the formation of this complex is specific to mitophagy, or it serves as a broader response to various forms of mitochondrial insults. VDAC1 oligomerization has been frequently observed in the context of apoptosis induction [27,29,35,36]. The idea projects multimeric VDAC1 complex as a gateway to release mitochondrial inner compartment components, especially CYCS. Conversely, VDAC1 is dispensable for the formation of mitochondrial permeability transition pore and VDAC1 or VDAC3 knockout does not reduce apoptosis [37]. This highlights that VDAC1 oligomerization and apoptosis might be stimulus and cell type specific [26]. However, VDAC1 oligomeric states on apoptotic OMM are not well characterized. Little is known about other potential interacting partners of VDAC1 that may contribute to the formation of these complexes in diverse ways, particularly in response to various mitochondrial stressors.

A prior study suggests that VDAC1 plays a pivotal role at the intersection of mitophagy and apoptosis. The nature of VDAC1 ubiquitination is one of the determining factors, where mono-ubiquitination inhibits apoptosis and poly-ubiquitination leads to mitophagy [38]. This study also points out the significance of the PINK1-PRKN pathway in cell survival. During mitochondrial stress, there exists a continual competition between mitophagy and the possible release of CYCS. Effective elimination of damaged mitochondria through mitophagy plays a crucial role in determining the outcome. A recent study has indeed shown that PINK1-PRKN-induced proteasome-dependent rupture of the OMM can lead to the release of CYCS when mitochondria are not adequately cleared through autophagic processes [39]. From our present findings, it appears that in situations where cellular autophagy mechanisms are active, the establishment of the VDAC1-PHB2 complex and the efficient exposure of PHB2 May have the potential to promote cell survival by facilitating mitophagy.

The question of whether the increased PHB2-VDAC1 interaction can offer protection against the impediments to mitophagy and the subsequent development of progressive neurodegenerative conditions (particularly in PD) warrants further examination. However, given its role in cell line and stimulus specific apoptosis, increased VDAC1 oligomerization may also lead to pathogenic conditions. Systemic lupus erythematosus (SLE) pathogenesis has recently been linked to elevated VDAC1 oligomerization [26]. Although movement disorders are commonly associated with SLE, the occurrence of PD in such patients is relatively rare. Indeed, a sizable cohort study that assessed the likelihood of PD development in SLE patients discovered that these individuals have a reduced risk of PD occurrence [40]. Further investigation is needed to determine whether the elevated VDAC1 oligomers associated with SLE include PHB2 as part of the complex, potentially promoting mitophagy for protection against PD development. It is worth noting that previous evidence has also shown reduced VDAC1 levels in the SN of PD brain [41] (GEO database NCBI; accession number: GSE20333 and GSE20186).

Further investigations and clarifications are required to identify other OMM proteins that are PRKN substrates and can produce similar effects. Among the potential candidates, MFN1 (mitofusin 1) or MFN2 and elements of the TOMM complex stand out. TOMM20, a component of the TOMM complex, is known to form high molecular weight assemblies during pyroptosis [42]. Nonetheless, the study provided evidence that mitochondrial depolarization induced by CCCP does not have the same impact. MFN2 has not yet been established as an interactor with PHB2, particularly in the context of CCCP treatment. However, the elimination of MFN2 does result in impaired mitophagy and retrograde neurodegeneration, possibly owing to its involvement in recruiting PRKN to the OMM [43,44]. These findings do not exclude the possibility that other affected roles of MFN2 May play a significant role in the development of PD-like symptoms, such as the tethering between the endoplasmic reticulum and mitochondria [45].

Even though VDAC1 is a prominent substrate for PRKN, there is a scarcity of research dedicated to identifying factors that might influence VDAC1’s involvement in mitophagy [14,15,17]. To summarize, our study demonstrated how VDAC1 can interfere with mitophagy by regulating PHB2 exposure toward LC3. While it is possible to bypass the need to degrade ubiquitinated proteins on OMM for mitophagy, exposure of PHB2 can increase the efficiency of the process. Thus, it can act as a critical factor in the development of neurodegenerative disorders. Further investigations are needed to delve into the connection between neurodegeneration and elements that impede the interaction between VDAC1 and PHB2.

Materials and methods

Reagents

All the reagents are purchased from Sigma-Aldrich, unless mentioned otherwise.

Animal ethics

Animal experimentation was done following the national guidelines (Care and Use of Animals in Scientific Research) formed by Committee for the Purpose of Control and Supervision of Experiments on Animals (CPCSEA), Animal Welfare Division, Ministry of Environment and Forests, Govt. of India. The protocol was evaluated and accepted by animal ethics committee of CSIR-Indian Institute of Chemical Biology, Kolkata, India (IICB/AEC/Meeting/May/2019/3, dated 31/05/2019).

Antibodies

The following primary antibodies are used: VDAC1 (Millipore, MABN504 and Abclonal, A19707), VDAC (Abcam, ab14734), VDAC2 (Abcam, ab37985), VDAC3 (Abcam, ab130561), TOMM20 (Abcam, ab186735), ATP5F1A/ATP5A (Abcam, ab14748), anti-mCherry (Abcam, ab183628), PHB2 (Sigma-Aldrich, HPA039874 and Santa Cruz Biotechnology, sc -133,094), anti-Flag (Sigma-Aldrich, F7425), HSPD1/HSP60 (Biobharati, BB-AB0209), actin (Santa Cruz Biotechnology, SC-58673), LC3 (Abcam, ab128025 and Abclonal, A17424), HRP-anti mouse (Biobharati, BB-SAB02C), HRP-anti rabbit (Biobharati, BB-SAB01C), HRP-anti goat (Abcam, ab6885), Alexa Fluor 568 anti-mouse (Invitrogen, A11004), Alexa Fluor 488 anti-mouse (Invitrogen, A11001), Alexa Fluor 568 anti-rabbit (Invitrogen, A11011), Alexa Fluor 488 anti-rabbit (Invitrogen, A11008), Alexa Fluor 647 anti-rabbit (Invitrogen, A21245), Alexa Fluor 647 anti ATP5F1A/ATP5A (Abcam, ab196198).

Plasmids

The following plasmids are used for the study: PRKN-YFP, GFP-LC3 and MitoKeima [1,46,47]. Plasmids for VDAC1 shRNA-puromycin, VDAC1-3XFLAG, VDAC1-GFP (1–10) and GFP (11)-PHB2 were designed in the laboratory and procured from Vector builder. MitoBFP is procured from Dr. Oishee Chakrabarti’s laboratory (Saha Institute of Nuclear Physics, Kolkata, India).

Drosophila lines

Act5C-GAL4/CyO (Bloomington Stock No. 25374) is gifted by Dr. Pralay Majumder (Presidency University, Kolkata, India), UASp-mCherry-Atg8a (Bloomington Stock No. 37750) is gifted by Prof. SC Lakhotia (Banaras Hindu University, Varanasi, India). UAS mitoKeima, porin A2/CyO and porin RV are procured from Korea Drosophila Resource Center (KDRC ID number 10,327, 10332 and 10,116, respectively) [34]. These lines are maintained in standard cornmeal food. Rotenone or vehicle administration is done by mixing the proper concentration with the food [30].

Cell culture and treatments

Human mid-brain derived dopaminergic neuroblastoma cell line SH-SY5Y or human embryonic kidney (HEK 293T) cells are cultured in DMEM (Thermo Fisher Scientific 12,800,017) with sodium bicarbonate (3.75 g/l; Sigma-Aldrich, S5761), fetal bovine serum (10%; Thermo Fisher Scientific 10,270–106), and penicillin-streptomycin (1%; Thermo Fisher Scientific 15,070,063) in a humidified incubator at 37°C at 5% CO₂. Control and VDAC1 knockout HEK 293T cells were procured from Abcam (ab255449 and ab255444). Cell transfections are done using Transfectin TM (Bio-Rad 1,703,351). Stable VDAC1 KD SH-SY5Y cell is obtained by transfecting VDAC1 shRNA and selected against puromycin (Sigma-Aldrich, P8833) treatment following standard protocol. MitoBFP, VDAC1-GFP (1–10) and GFP (11)-PHB2 were expressed in the mentioned cell lines for 36 h. For measuring the raw IntDen we utilized freely available ImageJ software. For 3D rendering we used volume j plugin.

CCCP (10 µM; Sigma-Aldrich, C2759) treatment, in presence or absence of MG-132 (50 µM; Sigma-Aldrich, SML1135) is given for different time points (1 to 8 h), as mentioned in the relevant experimental procedure.

Mitochondria isolation and treatments

Mitochondria were isolated from rat brain/cells/Drosophila by differential centrifugation following the protocol as mentioned previously [47]. In brief, Cell/tissue is homogenized in mannitol-sucrose buffer (225 mm mannitol [Sisco Research Laboratories 79,887], 75 mm sucrose [Sisco Research Laboratories 27,580], 5 mm HEPES [Sisco Research Laboratories 16,826], pH 7.4, 0.1 mm EGTA [Sigma-Aldrich, E4378]; supplemented with 2% BSA [Sisco Research Laboratories 85,171]) and then centrifuged at 1,500 × g (at 4°C for 6 min). The supernatant is again centrifuged at 7,000 × g (6 min). The pellet is washed with mannitol-sucrose buffer and re-suspended in appropriate amount of mannitol-sucrose buffer.

CCCP (10 µM) is used for 20 min to induce depolarization, then samples are crosslinked for another 30 min. Samples are immediately prepared in Laemmli buffer (1 M Tris [Sisco Research Laboratories 37,969], pH 6.8, 2% SDS [Sisco Research Laboratories 32,096], 40% glycerol [Sisco Research Laboratories 62,417], 0.4% bromophenol blue [Sisco Research Laboratories 93,676]) and processed for immunoblotting in mentioned conditions. The concentration of the crosslinkers are: bismaleimidohexane (BMH [Thermo Fisher Scientific 22,330]) 5 mm; dithiobis-succinimidyl propionate (DSP [Sigma-Aldrich, D3669]) 2 mm; ethylene glycol bis-succinimidyl succinate (EGS [Sigma-Aldrich, E3257]) 50 µM for rat brain mitochondria and 1 mm for isolated cell line mitochondria. In relevant experiments, 10 min prior treatment with DIDS (100 µM; Sigma-Aldrich, D3514) is given.

Protease (trypsin) protection/digestion assay

For isolated mitochondria (from rat brain and Drosophila), around 50 µg is treated with either buffer or 200 μg/ml trypsin (TPCK treated [Thermo Fisher Scientific 20,233]) for 10 min at 4°C in trypsin digestion buffer (10 mm sucrose, 0.1 mm EGTA-Tris [Sigma-Aldrich, E4378 and Sisco Research Laboratories 37,969], and 10 mm Tris-HCl, pH 7.4 [Sisco Research Laboratories,17560]); then, the samples are mixed with Laemmli buffer + β-mercaptoethanol (Sigma-Aldrich, M6250) and heated at 95°C.

For cells, we followed the protocol as described previously [47]. Briefly, cells from different genetic background are harvested after treatment (CCCP-10 µM, DIDS pretreatment-100 µM, 4 h) and permeabilized (0.015% digitonin [Sigma-Aldrich 300,410], 2 min), followed by washing with pH 7.4 PBS (0.137 M NaCl [Sisco Research Laboratories 33,205], 0.0027 M KCl [Sisco Research Laboratories 50,016], 0.01 M Na₂HPO₄ [Sisco Research Laboratories 97,768] 0.0018 M KH₂PO₄ [Sisco Research Laboratories, 5045]; pH 7.4). Cell pellet is resuspended in trypsin digestion buffer and divided equally into two parts; out of which one received trypsin (200 µg/ml). Both of the tubes are kept on ice for 5 min and then 2X Laemmli buffer + β-mercaptoethanol is added. Samples are separated in by SDS PAGE and processed for immunoblotting.

Immunoblot and immunoprecipitation

Crosslinked samples are resolved using SDS PAGE in non-reducing condition followed by Immunoblotting. For SDS-PAGE which is done under reducing condition, protein lysates are prepared in radioimmunoprecipitation assay buffer (20 mm Tris, pH 7.4, 130 mm NaCl, 1 mm Na₂EDTA [Sisco Research Laboratories 43,272], 1 mm EGTA [Sigma-Aldrich, E4378], 1% NP40 [Sigma-Aldrich 74,385] 1% sodium deoxycholate [Sisco Research Laboratories 96,876]) supplemented with protease inhibitor cocktail (Sigma-Aldrich, P8340). Immunoblotting is carried out following the standard procedure. Secondary antibodies conjugated to HRP were used (1:3000) and visualized with ECL chemiluminescence (Invitrogen 32,132 or Sigma-Aldrich, WBLUF0500).

For second dimension SDS-PAGE, gel pieces are incubated with either β-mercaptoethanol or 0.5 N NH₄OH (as mentioned, Sigma-Aldrich 221,228) for 1 h, 37°C with continuous shaking, followed by 20-min incubation with 2X Laemmli buffer. The gel pieces are kept on top of a resolving gel and sealed with agarose.

For third dimension SDS-PAGE, isolated gel pieces are further incubated with 0.5 N NH₄OH for 1 h, 37°C and resolved again.

For immunoprecipitation, approximately 1 mg protein sample kept in lysis buffer (25 mm Tris HCl, pH 7.4, 150 mm NaCl, 1% NP40, 1 mm EDTA, 5% glycerol, 0.1% Tween 20 [Sisco Research Laboratories 28,599]) and washed with equilibrated Protein A/G agarose beads (Santa Cruz biotechnology, sc-2003). Equal amount of cleared lysate is incubated with respective antibodies (1:100), 4°C overnight, and on the next day Protein A/G agarose beads are co-incubated for 2 h at room temperature. Afterwards beads are washed repeatedly and protein samples are eluted using 2X Laemmli buffer (+ β mercaptoethanol).

Evaluation of mitoKeima in cells and drosophila wing/muscle

Cells are transfected with mitoKeima and after 48 h, CCCP treatment is given to induce mitophagy (8 h). Signal intensity from single cell suspension is determined by FACS. High mitophagic cells [48] are quantified by Flowjo software.

For Drosophila, mitoKeima-expressing live fly wings are visualized under confocal microscope (Zeiss LSM980, Germany). For all the experiments Ex: 440/568 and Em: 610 nm is used. For thorax: thoracic muscles are dissected out on coverslip and kept in PBS (pH 7.4) while imaging.

Immunofluorescence and confocal imaging

Fixed cells (4% PFA [Sigma-Aldrich, P6148]) on coverslips are permeabilized with 0.1% triton X-100 (Sigma-Aldrich, T8787) and blocked with 4% BSA. For Drosophila thoracic muscle, samples are fixed in 4% PFA for 3 d and transferred to 30% sucrose. Cryosections on slides are permeabilized with 0.2% Triton X-100. BSA (8%; 1 h) is used for blocking. Samples are incubated with specific primary antibodies at 4^◦C overnight in a humid chamber and tagged with Alexa Fluor conjugated secondary antibody.

For Duo link-proximity ligation assay, 4% PFA fixed samples are blocked and incubated with primary antibodies (rabbit-PHB2 and mouse-LC3), as described above. A kit is used to develop the signal following manufacturers protocol (Sigma-Aldrich, DUO92101). We visualized the signal by using a high-resolution confocal microscope (Zeiss LSM980, Germany).

Transmission electron microscopy

To acquire TEM images, we followed the protocol as stated previously [47].

Statistical analyses

Bar graphs represented in the figures are mean ± SEM. Two-tailed Student’s t-test (or mentioned otherwise) or one-way ANOVA is used to decide the level of significance. Relevant post hoc tests are mentioned along with the figure legend. p ≤ 0.05 was considered as significant difference.

Funding Statement

The work was supported by CSIR - Indian Institute of Chemical Biology, Kolkata, India; Department of Biotechnology (DBT), Ministry of Science and Technology, India; Science and Engineering Research Board (SERB), Ministry of Science and Technology, India; International Brain Research Organization (IBRO) - Young IBRO Regions Connecting Awards.

Disclosure statement

No potential conflict of interest was reported by the author(s).

Data availability statement

Data that support the findings of this study are available from the corresponding author upon reasonable request.

Abbreviations

A2/A2

porin knockout;

BFP

blue fluorescent protein;

BMH

bismaleimidohexane;

CCCP

carbonyl cyanide m-chlorophenylhydrazone;

DSP

dithiobis-succinimidyl propionate;

DIDS

4′-diisothiocyanostilbene-2,2′-disulfonic acid;

EGS

ethylene glycol bis-succinimidyl succinate;

GEO

gene expression omnibus;

GFP

green fluorescent protein;

HEK 293T

human embryonic kidney 293T cells;

IMM

inner mitochondrial membrane;

kDa

kilodaltons;

KD

knockdown;

KO

knockout;

MAP1LC3/LC3

microtubule associated protein 1 light chain 3;

OMM

outer mitochondrial membrane;

PHB2

prohibitin 2;

PINK1

PTEN induced kinase 1;

PLA

proximity ligation assay;

PRKN

parkin RBR E3 ubiquitin protein ligase;

PD

Parkinson disease;

RV

porin revertant;

SDS-PAGE

sodium dodecyl sulfate-polyacrylamide gel electrophoresis;

SLE

systemic lupus erythematosus;

TEM

transmission electron microscopy;

VDAC1

voltage dependent anion channel 1;

WT

wild type.

Supplementary material

Supplemental data for this article can be accessed online at https://doi.org/10.1080/15548627.2024.2426116