Blog

By Huan Yang, Hongtao Shen, Jing Li, and Lian-Wang Guo 

Excerpt from the article published in Autophagy, 15:9, 1539-1557, Published online: 14 Mar 2019 DOI: 10.1080/15548627.2019.1586248

Editor’s Highlights

• Autophagy lysosome-mediated pathway (ALP) degrades and recycles cellular wastes including proteins, lipids, and dysfunctional organelles, and is a key element of the Proteostasis Network.
• SIGMAR1 is a putative chaperone protein primarily residing in the membranes of endoplasmic reticulum (ER) and nuclear envelope, that not share sequence homology with any other mammalian proteins.
• SIGMAR1 modulates a variety of signaling pathways including ion channels, GPCRs, lipid rafts, ER stress response, as well as chromatin remodeling. 
• A SIGMAR1 association with fusion machinery proteins ATG14, STX17, and VAMP8. Organelle fusion is a critical step for autophagic waste clearance.
• SIGMAR1 modulates autophagosome-lysosome fusion, a converging point of different types of macroautophagy including mitophagy.
• The SIGMAR1 C-terminal half molecule alone is reported to be sufficient for its chaperone activity 
• SIGMAR1 N-terminal half (SIGMAR1N80) alone produced positive BiFC and co-IP results, suggesting its association with (or proximity to) the lysosomal fusion factor complex.

ABSTRACT

Autophagosome-lysosome fusion is a common critical step in various forms of macroautophagy/autophagy including mitophagy, the selective degradation of mitochondria. Regulations of this fusion process remain poorly defined. Here we have determined the role of SIGMAR1, a unique endoplasmic reticulum membrane protein. Knockout of Sigmar1impaired mitochondrial clearance without altering the PINK1-PRKN/Parkin signaling, in mouse retinal explants and cultured cells treated with carbonyl cyanide m-chlorophenyl hydrazone (CCCP) for induction of mitophagy. SIGMAR1 depletion also caused accumulation of autophagosome markers LC3-II and SQSTM1, but did not change the levels of BECN1 and ATG7, proteins associated with autophagosome biogenesis. Lysosomal pH and protease activities were not negatively affected. However, sigmar1 knockout partially compromised autophagosome-lysosome fusion in CCCP-treated NSC34 cells, as revealed by reduced GFP fluorescence quenching of GFP-RFP-LC3-II puncta and co-localization of lysosomes with mitochondria. Furthermore, SIGMAR1 co-immunoprecipitated with ATG14, STX17, and VAMP8 (but not SNAP29), proteins key to autophagosome-lysosome membrane fusion. Re-expressing SIGMAR1 in the null background rescued clearance of mitochondria and autophagosomes. In summary, we started out finding that sigmar1 knockout impaired the clearance of mitochondria and autophagosomes, and then narrowed down the SIGMAR1 modulation to the autophagosome-lysosome fusion step. This study may shed new light on understanding autophagy-associated cyto-protection and disease mechanisms.

Abbreviations: APEX2, a genetically engineered peroxidase; BiFC, bimolecule fluorescence complementation; CCCP, a mitophagy inducing compound; CRISPR, clustered regularly interspaced short palindromic repeats; EM, electron microscopy; ER, endoplasmic reticulum; MAP1LC3/LC3, microtubule-associated protein 1 light chain 3; SIGMAR1, sigma non-opioid intracellular receptor 1.

Introduction

Autophagy is a lysosome-mediated mechanism that degrades and recycles cellular wastes including proteins, lipids, and dysfunctional organelles [1–3]. Mitophagy is the selective autophagic degradation of mitochondria [4–6]. These highly regulated processes play important roles in aging and diseases [7,8]. Autophagy is characterized by the formation (biogenesis) of phagophores that engulf cellular wastes (cargo), mature into autophagosomes and fuse with lysosomes wherein cargo is degraded. Thus, autophagosome-lysosome fusion is a critical step for final clearance of autophagic cargo such as damaged mitochondria that are otherwise harmful to the cell. While the mechanisms of autophagosome biogenesis have been intensively studied, regulations of autophasome-lysosome fusion remain poorly understood [9].

SIGMAR1 is a putative chaperone protein [10] primarily residing in the membranes of endoplasmic reticulum (ER) and nuclear envelope [11,12]. This receptor is unique in that it does not share sequence homology with any other mammalian proteins [13]. SIGMAR1 modulates a variety of signaling pathways including ion channels, GPCRs, lipid rafts, ER stress response [14,15], as well as chromatin remodeling [16]. Studies including our own [17,18] support a protective role of SIGMAR1 in neurodegenerative diseases [19–24] and SIGMAR1 mutations are linked to familial amyotrophic lateral sclerosis (ALS) [21,25]. The crystal structure of human SIGMAR1 has been recently solved [26]. In spite of these advances, the function of SIGMAR1 is not well understood at the molecular level [15], especially in autophagy.

Knockdown of SIGMAR1 with siRNAs led to accumulation of autophagosomes and enhanced ER stress in NSC34 neuronal cells [27,28]; SIGMAR1 antagonists elevated autophagosome marker LC3-II in cancer cells [29] and in astrocytes [30,31]. In addition, targeted siRNA screens in the RPE1 cell line also predicted SIGMAR1 as one of the autophagy regulators [32]. However, important mechanistic questions remain regarding the role of SIGMAR1 in autophagy. Given the critical importance of mitophagy/autophagy in cellular functions and human health, it is imperative to investigate the molecular role of SIGMAR1 in these processes.

Since it has never been clearly addressed whether SIGMAR1 regulates mitophagy [33], the mitochondria-specific autophagy, we first set out to determine the role of SIGMAR1 in mitochondrial clearance. To confirm the results, we performed experiments in parallel using sigmar1 knockout retinal tissues and CRISPR-mediated knockout and knockdown cell lines as well as primary cells isolated from sigmar1 knockout mice. We found that SIGMAR1 deficiency impaired the clearance of mitochondria under the treatment with carbonyl cyanide m-chlorophenyl hydrazone (CCCP), a mitophagy inducer, and accelerated apoptosis. We then narrowed down the underlying mechanism to partial blockage of the autophagosome-lysosome fusion step. Our results identify SIGMAR1 as a novel modulator of this autophagic organelle fusion event.

Results

Knockout and inducible knockdown of SIGMAR1 using the CRISPR-Cas9 technology

To unambiguously define the role of SIGMAR1 in mitophagy, we used several tissue and cell model systems. 1) Retinal explants or mouse embryonic fibroblasts (MEFs) were isolated from wild type (WT) and sigmar1 knockout (KO) mice. 2) SIGMAR1 KO HEK293 (Figure 1(a,b)) and NSC34 [34] cell lines were generated using a CRISPR-Cas9 genome-editing approach via lentiviral expression of nuclease-active Cas9 and SIGMAR1/Sigmar1 gRNAs. SIGMAR1/Sigmar1KO single clones were selected. 3) To induce SIGMAR1 knockdown, we used a lenti-vector to express SIGMAR1 gRNA and nuclease-deficient Cas9 (dCas9) fused to a transcription repressor (KRAB), which suppresses SIGMAR1 expression in a targeted manner (Figure 1(c)). We achieved ~90% SIGMAR1 knockdown in SH-SY5Y cells 3 d after doxycycline induction (Figure 1(d)), which is highly efficient especially considering slow turnover of the SIGMAR1 protein [35].

Mitochondria clearance is impaired in sigmar1 KO mouse retinal explants and sigmar1 KO cells

In recent studies [36–38], mitophagy has been generally induced by treating cells with low-dose carbonyl cyanide m-chlorophenyl hydrazone (CCCP), a mitochondrial uncoupler, followed by western blot assay to measure expression of mitochondrial marker proteins VDAC1 and TIMM23. These two markers decrease upon lysosomal degradation of mitochondria (i.e., mitophagy), thus their accumulation serves as readouts of impaired mitophagy. As shown in Figure 2, A and B, in WT retinal explants, VDAC1 and TIMM23 were both substantially reduced in similar proportions 24h after CCCP induction of mitophagy. However, in KO retinal explants these two markers did not decrease, but rather increased 2–3 fold after CCCP treatment. We confirmed this finding in the NSC34 neuronal cell line (Figure 2(c,d)). Consistently, while VDAC1 and TIMM23 continuously decreased in WT cells during the time course of CCCP treatment, they accumulated in KO cells during the 1–12h treatment. This accumulation was also observed via fluorescent imaging of DsRed-Mito labeled mitochondria in KO (vs WT) NSC34 cells following CCCP treatment (Figure 2(e,f)). These opposite results (KO vs WT) suggest impaired mitophagy in KO cells. To further confirm a specific role of SIGMAR1 in mitophagy, we SIGMAR1 Sig1R knockdown in the SH-SY5Y neuronal cell line (Figure 2(g)). We found that mitophagy was impaired also by SIGMAR1 knockdown (Figure 2(g-i)), as indicated by substantially reduced relative degradation of TIMM23 (difference between CCCP and DMSO conditions) in knockdown cells (vs WT) (Figure 2(i)). While CCCP caused TIMM23 level changes in opposite directions in SIGMAR1 KO and WT cells (Figure 2(b,d)), the TIMM23 changes were not in opposite directions in SIGMAR1 knockdown and WT cells (Figure 2(h)). This likely reflects a lesser impact of induced knockdown compared to permanent KO. The combined results from three experimental models demonstrate that SIGMAR1 modulates outcomes of mitophagy.

The PINK1-PRKN/Parkin pathway in mitophagy is not affected by sigmar1 KO

Thus far, the PINK1-PRKN/Parkin pathway is the best characterized early event of mitophagy [39]. Membrane depolarization in damaged or dysfunctional mitochondria leads to PINK1 stabilization on the outer membrane. PINK1 then activates the E3 ubiquitin ligase activity of PRKN and its recruitment from the cytosol to mitochondria. Activated PRKN ubiquitinates mitochondrial outer membrane proteins, which trigger phagophore engulfment and lysosomal degradation of mitochondria [40].

To determine whether PINK1-PRKN signaling was affected by sigmar1 KO, we transfected NSC34 cells to express both EGFP-PRKN and DsRed-Mito. Fluorescence microscopy illustrated homogeneous distribution of EGFP-PRKN in the cytosol in the absence of the mitophagy inducer CCCP (DMSO condition; Figure 3(a))(Figure S1). In contrast, 4h after CCCP treatment, EGFP-PRKN staining patterns became punctate and mostly overlapping with mitochondrial marker DsRed-Mito, indicating PRKN recruitment to mitochondria. We counted the number of cells that showed EGFP-PRKN/DsRed-Mito overlapping (indicative of PRKN recruitment), and did not see a difference between WT and KO (Figure 3(b)). Thus, the PINK1-PRKN pathway appeared unaffected by the absence of SIGMAR1. Reinforcing this conclusion, levels of stabilized PINK1 protein (the long variant) were similar in WT and KO cells during the 12-h CCCP treatment to induce mitophagy (Figure 3(c)).

We then evaluated whether the next stage of mitophagy, the engulfment of mitochondria by autophagosomes, was disturbed by Sigmar1 KO. We co-transfected EGFP-LC3 and DsRed-Mito into NSC34 cells and induced mitophagy with a 6-h CCCP treatment. Upon autophagosome maturation, EGFP-LC3 becomes lipidated and associates with phagophore membranes. It appears punctate [41], and yellow if co-localizing with DsRed-Mito (engulfed mitochondria) (Figure 3(d)). Quantification of mitochondria-phagophore/autophagosome co-localization using Pearson’s correlation coefficient did not show a significant difference between WT and KO cells (Figure 3(e)). Thus, Figure 3 shows that neither PINK1-PRKN signaling nor mitochondria engulfment by autophagosome was disrupted by Sigmar1 KO.

Autophagosome clearance is impaired in sigmar1 KO mouse retinal explants and KO cells

After mitochondria engulfment by phagophores, mitophagy merges into autophagic flux, which features constant generation (biogenesis) and lysosomal degradation of autophagosomes and cargos therein [2]. As such, mitophagy is an integral process of autophagy and disturbed autophagy alters outcomes of mitophagy [42]. To further dissect the causal mechanism of mitophagy impairment due to SIGMAR1 loss, we evaluated the impact of Sigmar1 KO on autophagy. We measured (Figure 4(a)) protein levels of LC3-II, a commonly used autophagosome marker, and also SQSTM1 which ‘tags’ waste proteins for autophagic degradation [41]. To determine whether SIGMAR1 plays a general role in autophagy, the experiments were first carried out in the basal condition without CCCP treatment. In KO retinal explants, these two proteins were at least 2-fold higher than WT controls. This result was replicated in non-neuronal (HEK293) as well as neuronal (NSC34) Sig1R KO cells vs WT cells (Figure 4(b)). A similar result was obtained by calculating the LC3-II:LC3-I ratio. While the LC3-II:LC3-I ratio is often used to monitor the conversion of cytosolic LC3-I to the lipidated form (LC3-II, localized on autophagosomes), impaired autophagosome-lysosome fusion, and hence impaired degradation of LC3-II, also contributes to an increase in this ratio. We also observed significantly more accumulation of LC3-II in SIGMAR1 knockdown SH-SY5Y cells compared to control cells (DOX+ vs DOX-, Figure 4(c,d)) (Figure S2). Together, these results consistently demonstrate accumulation of autophagosomes due to loss of the SIGMAR1 protein, reminiscent of the mitochondria accumulation observed in SIGMAR1/sigmar1 KO tissues and cells (Figure 2).

Accumulation of autophagosomes could result from increased biogenesis or reduced lysosomal degradation or both [41,43]. We assessed the effect of SIGMAR1 depletion in NSC34 cells on BECN1 and ATG7, proteins key to autophagosome biogenesis (Figure 4(e)). The data show that levels of both proteins in KO and WT cells were essentially the same, although CCCP induced a slight increase in WT cells (Figure 4(f)). To further confirm unaltered biogenesis in KO vs WT cells, we pharmacologically blocked autophagic flux using bafilomycin A1 (Figure 4(g)). This inhibitor blocks the fusion of autophagosomes with lysosomes and hence the autophagosome degradation; thus increased LC3-II accumulation would reflect enhanced autophagosome biogenesis [44]. However, we did not see a difference of LC3-II levels between KO and WT in the presence of bafilomycin, although LC3-II accumulated in KO (vs WT) cells in the absence of bafilomycin, either with or without CCCP treatment (Figure 4(h))(Figures S3 and S4). This result is similar to that of BECN1 and ATG7, suggesting undisturbed autophagosome biogenesis in sigmar1 KO cells. Significantly, in the basal condition without CCCP, sigmar1 KO (vs. WT) caused 2-fold more LC3-II accumulation (Figure 4(h), DMSO). Conversely, bafilomycin A1, which fully blocks autophagosome degradation, resulted in further accumulation of LC3-II and also mitochondrial markers (Figure S5) in sigmar1 KO cells. That the absence of the SIGMAR1 did not result in complete blockage of the clearance of autophagosomes and mitochondria suggests the existence of a yet unknown compensatory mechanism or factor.

Autophagosome closure is not disrupted in sigmar1 KO cells

An important step in the formation of autophagosome is that a phagophore extends and encloses cytoplasmic contents (cargo). We therefore determined whether depletion of SIGMAR1 affects autophagosome membrane closure by using a highly cited method [45]. The basic principle is that once sealed in the autophagosome, SQSTM1 which ‘labels’ cargo, should be protected from degradation by proteinase K; conversely, if the autophagosome is not completely closed, SQSTM1 will be exposed to proteinase K and disappear. We isolated autophagosomes and analyzed the ratio of SQSTM1 levels (with vs without proteinase K), which reflects protection of SQSTM1 from proteolysis and hence the level of autophagosome closure (Figure 5(a)). Our data indicated that this ratio was not significantly sigmar1 in Sig1R KO cells compared to WT cells, either in the basal, CCCP, or starvation condition (Figure 5(b)). SQSTM1 was completely degraded after Triton X-100 solubilization, verifying active proteinase K. Moreover, since autophagosomes are known to be sensitive to redox [2], we also compared the levels of reactive oxygen species (ROS) in WT and KO cells. However, only a slight increase of ROS was detected in KO vs WT cells (Figure 5(c)). This result is consistent with the lack of difference in autophagosome closure between WT and KO cells. In addition, we did not observe a difference in the autophagosome subcellular positioning pattern between WT and KO cells, as shown by mCherry-LC3 fluorescence (Figure S6).

Taken together, the multiple lines of experimental evidence presented above suggest that the cause of impaired mitochodria/autophagosome clearance in KO cells was likely at the lysosomal degradation step. Indeed, autophagosome ‘turnover’ or LC3-II degradation rate, calculated as a difference of LC3-II levels with vs without Baf.A1 blocking its lysosomal degradation (Figure 4(i)), was substantially reduced in KO vs WT cells.

Lysosomal functional integrity is not compromised in sigmar1 KO mouse retinal explants and sigmar1 KO cells

Retarded autophagosome degradation could result either from dysfunction of lysosomes or impairment of autophagosome-lysosome fusion [46]. Since the lysosome is the final common destination of autophagic cargos including mitochondria [37], we first assessed the influence of sigmar1 KO on lysosomal activities. CTSB and CTSD are the main proteases that execute autophagosme cargo protein degradation, therefore their activities serve as sensitive indicators of lysosomal health [47]. In addition, an acidic pH is critical for normal proteolytic activities of lysosomal proteases. Our data from retinal explants (Figure 6(a,b)) indicate that protein levels of mature CTSB were not significantly changed in KO vs WT tissues; mature CTSD was increased by ~2 fold in the KO; LAMP2, a marker protein residing in the lysosome membrane, was slightly increased. Confirming this result, very similar outcomes occurred in sigmar1 KO vs WT NSC34 cells (Figure 6(c,d)). Consistent with an increase of mature CTSD detected by western blot, its proteolytic activity was also increased in KO vs WT NSC34 cells (Figure 6(e)). This difference did not occur in response to CCCP treatment, although CCCP elevated CTSD activity in WT cells. In agreement, lysosomal pH remained unaltered in KO compared to WT cells (Figure 6(f)). These data show that the lysosomal capacity for digesting autophagic cargo was not compromised by sigmar1 KO.

Autophagosome-lysosome organelle fusion is partially compromised in sigmar1 KO cells

Since Sigmar1 KO did not negatively affect lysosomes, we inferred that inefficient autophagosome-lysosome fusion might be the underlying cause of mitophagy impairment. To monitor the organelle fusion, we transfected NSC34 cells to express LC3 fused with tandem GFP and mRFP. As observed in WT cells, autophagosomes (LC3-II puncta) appeared red as GFP fluorescence was quenched by the acidic environment of lysosomes, indicating normal fusion of these two organelles (Figure 7(a)). Interestingly, more LC3-II puncta appeared yellow in KO cells. This contrast between WT and KO cells revealed that the fusion of autophagosomes to lysosomes was partially disrupted in the absence of SIGMAR1. We predicted that the mitochondria engulfed with in autophagosomes would not be able to enter lysosomes and this was confirmed by a pronounced decrease of co-localization of mEmerald-lysosome with DsRed-Mito in KO cells compared to WT cells (Figure 7(b)).

Consistent with the foregoing results, we observed mCherry-SIGMAR1 at the edge of EGFP-LC3-labled autophagosome puncta (Figure 8(a)), and SIGMAR1-mCherry juxtaposed with mEmerald-labeled lysosomes (Figure 8(b)), suggesting possible SIGMAR1 proximity to these two organelles. We re-examined this using EM with resolutions enhanced by a SIGMAR1-GFP-APEX2 construct expressed on a sigmar1-null background. As shown in Figure 8(c), SIGMAR1 mostly localized in the ER membrane, including the nuclear envelope, as we previously reported (11). SIGMAR1 was also found at the ER-lysosome contacts, consistent with the fluorescence microscopy data. Together, the different lines of evidence suggest a role of SIGMAR1 in autophagosome-lysosome fusion, possibly involving the membrane fusion machinery.

A SIGMAR1 association with the SNARE complex in autophagosome-lysosome fusion

While multiple proteins are involved in autophagosome-lysosome fusion, the SNARE complex composed of STX17-SNAP29-VAMP8 is directly responsible for the final membrane fusion [48]. Most recently, ATG14 was found to regulate the fusion process by directly interacting with STX17 and SNAP29 [9]. Thus a protein complex facilitating final membrane fusion between autophagosomes and lysosomes is constituted at least by these four proteins. Moreover, STX17 and ATG14 both have been shown at ER-mitochondria contact sites [43], where SIGMAR1 primarily resides in the ER [10]. We therefore surveyed which of these proteins is in close proximity to SIGMAR1 in living cells, using a bimolecular fluorescence complementation (BiFC) method [49]. Interaction or close proximity of two proteins brings together two halves (VN and VC) of a ‘split’ Venus (YFP variant) allowing fluorescent emmissions. Based on the prevailingly used model of SIGMAR1 topology [15], fusing VN to the C-terminus of the full-length SIGMAR1 molecule should localize VN in the ER lumen, whereas its fusion to the C-terminus of the truncated version, SIGMAR1N80 (amino acids 1–80)(Figure S7), should position it to the cytosolic side. We found that three of the four fusion machinery proteins, i.e., ATG14, STX17, and VAMP8, induced Venus fluorescence after co-transfection with SIGMAR1N80-VN (amino acids 1–80) (Figure 9(a-d)). The pair of SNAP29-SIGMAR1N80 did not produce fluorescence, serving as a negative control. Further demonstrating the specificity of the BiFC method, pairing SIGMAR1N80 with CANX (calnexin), an irrelevant ER residing protein with its C-terminus localized in the cytosol, did not lead to positive signal (Figure S8). This result addresses a potential concern that aggregation of overexpressed ER proteins may cause false-positive BiFC. The full-length SIGMAR1 (SIGMAR1-VN) did not induce Venus fluorescence after co-expression with any of the four proteins (Figure 9), possibly because of a ‘pointing-away’ spatial orientation of the VN moity at the C-terminus of the full-length SIGMAR1 molecule.

We then re-examined the BiFC results by co-IP assay using NSC34 cells co-expressing FLAG-tagged full-length (or N80) SIGMAR1 and one of the four fusion machinery proteins tagged with a V5 peptide (Figure 9(e-h)). Replicating the BiFC result, ATG14, STX17, and VAMP8, but not SNAP29, each co-precipitated with either SIGMAR1N80 or full-length SIGMAR1. Taken together, our results obtained using two different methods suggest that SIGMAR1 is in close proximity of the autophagosome-lysosome fusion machinery protein complex.

We further examined the observed Sig1R association with the autophagosome-lysosome fusion machinery. Western blot assay indicated that the protein levels of either VAMP8 or STX17 were not significantly different between WT and KO (Figure 10(a))(Figure S9). We then determined the impact of SIGMAR1 KO on the interaction between VAMP8 and STX17, which reside on lysosomes and autophagosomes, respectively upon the organelle fusion. Compared to the basal condition (DMSO control), the co-IP of endogenous VAMP8 with V5-tagged STX17 drastically increased after CCCP treatment, indicating the formation of the VAMP8-STX17 protein complex. The comparison between KO and WT NSC34 cells under CCCP treatment showed a slight decrease (~33%) of VAMP8 co-IP (densitometry normalized to V5-STX17) in the KO (Figure 10(b)). We next determined whether sigmar1 KO affects the lysosomal localization of VAMP8, which is known to be synthesized in the ER, processed in the Golgi, and then transported to lysosomes to assume its function in autophagosome-lysosome fusion [9]. As indicated in Figure 11, the colocalization of endogenous VAMP8 with the mEmerald label of lysosomes did not show a significant difference between WT and KO, either in the basal (DMSO control) or CCCP condition.

Combined, the results presented above suggest that SIGMAR1 engages in autophagosome-lysosome fusion, not by indirectly affecting protein levels of VAMP8 and STX17 and the distribution of VAMP8 to lysosomes.

Re-expressing SIGMAR1 in a null background rescues mitophagy and autophagy

To further examine a specific role of SIGMAR1 in the autophagic process, we performed rescue experiments by re-introducing SIGMAR1 to the KO background. As either the full-length or SIGMAR1N80 co-precipitated with fusion machinery proteins (Figure 9), we expressed these two constructs and assessed CCCP-induced mitophagy. As shown in Figure 12(a,b), whereas VDAC1 and TIMM23 accumulated in KO NSC34 cells (as opposed to WT) after 12-h CCCP treatment, expression of the full-length SIGMAR1 restored KO cells to a WT-like phenotype. In contrast, SIGMAR1N80 was not able to produce this effect. Interestingly, SIGMAR1N80 expressed in WT cells prevented mitophagy induction by CCCP treatment while it slightly promoted basal mitochondria clearance (without CCCP) compared to empty vector control (Figure 12(c,d)).

Given that the C-terminal half molecule has been reported to account for the SIGMAR1 chaperone function [10], we tested its potential in the rescue of impaired mitochondrial clearance in sigmar1 KO NSC34 cells. However, the KO cells without (i.e., EV control) or with expression of the FLAG-tagged C-terminal half of Sig1R molecule (C120-FLAG) showed similar patterns of change in VDAC1 and TIMM23 protein levels during CCCP treatment (Figure 12(e,f)), indicative of a lack of rescue. This result together with that of SIGMAR1N80 (Figure 12(b)) suggest that it would require the full-length SIGMAR1 to rescue impaired mitophagy in KO cells.

We therefore re-examined the rescue effect of the full-length SIGMAR1 using MEFs, which were isolated from WT and KO mice to confirm the mechanisms in primary cells without CRISPR genetic manipulations. While TIMM23, VDAC1, and LC3-II accumulated by 2 fold in KO vs WT cells, after re-introducing the full-length SIGMAR1 to the KO background these KO cells reverted to a WT-like phenotype (Figure 12(g,h)). It is also important to note that the experiments with MEF cells were conducted in the basal condition without CCCP treatment. Thus, the KO-exacerbated accumulation of autophagosomes in MEF cells is consonant with the result obtained with NSC34 cells in the basal condition (Figure 4(h), DMSO). These experiments without CCCP demonstrate a sigmar1 KO-specific impairment of autophagy, minimizing a concern of possible non-specific effects arising from CCCP [50].

Collectively, the results from the rescue experiments demonstrate that the full-length SIGMAR1 was capable of rescuing the clearance of mitochondria and autophagosomes that was partially blocked due to SIGMAR1 depletion.

Apoptosis is exacerbated in sigmar1 KO cells and sigmar1 KO mouse retinal explants under the mitophagy-inducing condition

Finally, we evaluated the effect of Sig1R KO-induced impairment of mitophagy on apoptosis. We first transfected NSC34 cells with the Apoliner construct for apoptosis assay, in which GFP is cleaved off if caspases are activated. As indicated in Figure 13(a,b), while cleaved GFP slowly increased in WT cells during mitophagy induction by 12-h CCCP treatment, this increase was significantly accelerated in KO cells. To re-examine this observation, we used a different method by immunoblotting cleaved CASP3, which showed a very similar result (Figure 13(a,c)). Furthermore, we verified exacerbated apoptosis in KO (vs WT) cells using a third method, CASP3-CASP7 activity assay (Figure 13(d)). Importantly, CCCP-treated KO retinal explants also showed elevated CASP3 compared to WT control (Figure 13(e)).

In summary, in the current study we first observed that SIGMAR1 deficiency caused accumulation of mitochondria under the treatment with CCCP for induction of mitophagy. The extended determination of various factors involved in mitophagy/autophagy then led us to an interpretation that the autophagosome-lysosome fusion process was partially compromised in the absence of SIGMAR1 (Figure 14). Furthermore, the impairment of autophagosome clearance due to SIGMAR1/sigmar1 KO also occurred in the basal condition without mitophagy induction.

Discussion

In this study we report that SIGMAR1 plays a role in autophagosome-lysosome fusion. We first established that mitophagy was impaired by either permanent knockout or induced knockdown of SIGMAR1. We then narrowed down the underlying mechanism to partially disrupted autophagosome-lysosome fusion, and further obserevd a SIGMAR1 association with fusion machinery proteins ATG14, STX17, and VAMP8. Organelle fusion is a critical step for autophagic waste clearance yet its regulation is poorly defined [51]; our finding of SIGMAR1 as a modulator of this process may have broad implications in disease mechanisms. Recent reports support a protective role of SIGMAR1 in neurodegenerative diseases [18,24,52]. As SIGMAR1 is a unique ER resident protein with no homologs in mammalian genomes, investigation of SIGMAR1 as a new intervention target has been widely advocated [52,53]. However, advancing SIGMAR1-targeted interventions into clinical applications has been hampered by poor understanding of SIGMAR1 molecular actions [54], and whether SIGMAR1 knockout affects mitophagy remained underexplored.

Several lines of evidence in the current study support an important role of SIGMAR1 in mitophagy. First, impaired mitophagy due to SIGMAR1/sigmar1 knockout or knockdown was reproducibly observed in retinal tissues and four different cell types. A lesser impairment by knockdown vs knockout suggests that residual SIGMAR1 in knockdown cells is able to facilitate mitophagy to some degree. Consistent with the mitophagy relevance of these observations, no significant changes (KO vs WT) in VDAC1 and TIMM23 levels occurred without the CCCP induction of mitophagy, as indicated in Figure 2(b,d)(0h), 2F(0h), 2H and Figure 12(b) (0h). Thus, this result confirms the specificity of VDAC1 and TIMM23 as mitochondrial markers [37]. Interestingly, we observed accumulation of these two mitochondrial markers in sigmar1 KO (vs WT) MEFs without CCCP treatment. Different from the cell lines used in other experiments, MEFs are primary cells that were isolated from mice. We therefore gather that the manipulations during MEF isolation and purification might have caused cellular stress and triggered mitochondrial damage even without adding CCCP. Second, manifesting the SIGMAR1 specificity, re-introducing SIGMAR1 to a null background rescued mitophagy. Third, Sigmar1 knockout in retinas and cells exacerbated apoptosis confirming a protective effect of SIGMAR1-assisted mitophagy. A SIGMAR1 cyto-protective function was previously attributed to a variety of mechanisms, including antioxidation, ER stress response, and Ca²⁺ homeostasis, among others [10]; here our data suggest that SIGMAR1 preserved cell survival also via mitophagy. Fourth, consistent with mitophagy impairment, autophagosome accumulation occurred in sigmar1 knockout retinas and cells that were induced for mitophagy. In support of this result, accumulation of LC3-II was previously observed using siRNA for SIGMAR1 knockdown [27,28] or antagonists to block the SIGMAR1 binding site [55], although the underlying mechanism remained unclear.

Another important finding in this study is that SIGMAR1 modulates autophagosome-lysosome fusion, a converging point of different types of macroautophagy including mitophagy [39]. Thus far, regulatory mechanisms of this fusion event remain poorly understood [33] and our study suggests a novel mode of SIGMAR1 molecular function involved in this event. Impairment of the lysosomal organelle fusion due to Sigmar1knockout was revealed by compromised lysosomal quenching of GFP fluorescence, and also by mitigated co-localization of lysosomes with autophagosomes or mitochondria. Hence, inefficient autophagosome-lysosome fusion and cargo degradation may explain the accumulation of autophagosome and mitochondrial markers caused by SIGMAR1 deficiency observed here or by SIGMAR1 inactivation in other studies [56]. Alternatively, considering a possible role of SIGMAR1 in mitochondrial homeostasis [10], the observed accumulation of mitochondria may be explained by mitochondrial damage caused by Sigmar1 KO per sewithout involving impairment of autophagosome-lysosome fusion. However, in our experimental setting, Sigmar1 KO did not affect the PINK1-PRKN signaling of early-phase mitophagy, which would otherwise have been potentiated if Sigmar1 KO had caused mitochondrial damage. Sigmar1 KO did not appear to affect autophagosome formation either, as no changes were observed in BECN1 and ATG7 protein levels and autophagosome closure. Moreover, we did not observe a negative effect of SIGMAR1 ablation on events later than the lysosomal fusion step, i.e., lysosome proteolytic activity. Of note, another study (reported during the submission of this manuscript) provides evidence of fusion impairment due to the aggregation of mutated SIGMAR1 [28]. Though consistent with our data, aggregation of overexpressed SIGMAR1 per se causes cellular stress and may affect autophagic flux by different mechanisms. Nevertheless, these detailed analyses allowed us to pinpoint reduced autophagosome-lysosome fusion as the most likely mechanism underlying the observed mitophagy impairment. Consonantly, we did not observe an obvious Sigmar1 KO-caused alteration of VAMP8 and STX17 protein levels (Figure 10 and S9), lysosomal localization of VAMP8 (Figure 11), or subcellular distribution of autophagosomes (Figure S6) or lysosomes (Figure S10). A change in either of these would have otherwise secondarily affected autophagosome-lysosome fusion. Rather, we observed a SIGMAR1 proximity to the lysosomal membrane fusion protein complex using two experimental approaches, BiFC and co-IP.

Though intensively studied for pharmacology and psychoactive effects, SIGMAR1 remains under-investigated for its molecular functions [54]. The C-terminal half of the molecule is thought to be responsible for the SIGMAR1 chaperone activity and the known SIGMAR1 molecular interactions [10]. To date, known SIGMAR1 molecular actions, mainly modulations of ER stress response, redox, channel activity, GPCRs, and Ca²⁺ homeostasis, are almost all attributed to its role as a molecular chaperone that assists protein folding [54], and the SIGMAR1 C-terminal half molecule alone is reported to be sufficient for its chaperone activity [10,57]. Interestingly, our data showed that expressing the SIGMAR1 C-terminal half in sigmar1 KO NSC34 cells could not rescue mitophagy that was impaired due to KO, thus differentiating SIGMAR1’s chaperone function from its role in mitophagy/autophagy observed here. Conversely, we observed that the SIGMAR1 N-terminal half (SIGMAR1N80) alone produced positive BiFC and co-IP results, suggesting its association with (or proximity to) the lysosomal fusion factor complex. It is noteworthy that the Co-IP approach is not subject to an alternative interpretation of ‘false-positive’ interactions made possible by a long arm formed with VN and VC proteins in the case of BiFC. Of note, in BiFC experiments the full-length SIGMAR1 did not restore Venus fluorescence. There are two possible scenarios. First, if the SIGMAR1 C-terminus localizes in the ER lumen – the prevailing view in the literature [54,58], VN fused to this end must be separated by the ER membrane from VC that was attached to a cytosol-exposed fusion factor protein, e.g., STX17. Second, based on the SIGMAR1 crystal structure indicating the C-terminal domain embedded in the ER membrane on the cytosolic side [26], VN at the SIGMAR1 C-terminus should be ‘pinned’ onto the ER membrane and thus inaccessible to VC. Either scenario reasonably explains the lack of Venus fluorescence restoration by the full-length SIGMAR1. As conflicting reports exist on the subcellular localization of the SIGMAR1 C-terminus, distinguishing the two scenarios awaits more information. Nonetheless, a possible SIGMAR1N80 association with the fusion machinery complex departs from the known chaperone-based SIGMAR1 interactions, e.g., with GPCRs and ion channels [10,55].

However, when re-expressed in KO cells, it was the full-length SIGMAR1 rather than SIGMAR1N80 that rescued autophagic cargo degradation. On the other hand, SIGMAR1N80 prevented mitochondria clearance in WT cells. To reconcile those results, we propose that SIGMAR1N80 prevents mitophagy in WT cells probably by acting as a dominant-negative variant against the endogenous full-length SIGMAR1. Thus, a reasonable interpretation of our results is that while the N80 portion of the SIGMAR1 protein may associate with fusion machinery proteins, a full function of SIGMAR1 in autophagosome-lysosome fusion likely requires the full-length molecule. Along this line, an alternative interpretation of our data could be that the ablation of SIGMAR1 may lead to misfolding or mislocalization of the fusion machinery proteins, hence indirectly mitigating autophagosome-lysosome fusion. However, this interpretation appears not to be supported by our results that neither protein levels of VAMP8 and STX17 nor the lysosomal distribution of VAMP8 were altered in sigmar1 KO cells in comparison to WT cells. Aside from a possible SIGMAR1 engagement with the the autophagosome-lysosome fusion machinery, our data cannot rule out another scenario whereby SIGMAR1 may organize ER microdomains that serve as ‘docking’ sites for these two organelles to constellate and then fuse. Consistent evidence includes deformation of ER microdomains due to SIGMAR1 knockdown [27] and binding of SIGMAR1 with various lipids such as cholesterol, sphingosine, and ceramide which are important components of the ER membrane [59]. While intriguing, a role for SIGMAR1 in the autophagosome-lysosome fusion event requires further investigations to elucidate; e.g. co-IP experiments with multiple cell types to detect endogenous protein-protein interactions involving SIGMAR1 and the fusion factors, and assays to monitor organelle fusion using in vitro reconstituted vesicles [9].

In summary, we started out finding that SIGMAR1 depletion impaired mitophagy, and finally narrowed down the SIGMAR1’s role to the autophagosome-lysosome fusion step. While mitophagy features specialized signaling in its initiation and progression, it merges with general macroautophagy upon lysosomal organelle fusion as damaged mitochondria ‘hitchhike’ on autophagosomes for disposal to lysosomes [6]. In other words, mitophagy and general macroautophagy share the autophagosome-lysosome fusion mechanism (Figure 14). Consistently, in addition to the accumulation of mitochondrial markers, autophagosome (LC3-II) accumulation was also exacerbated due to SIGMAR1 ablation when mitophagy was induced with CCCP (Figure 4(c/d,g/h)). It is worth noting that exacerbated autophagosome accumulation in sigmar1 KO (vs WT) also occurred upon serum starvation (Figure S3), a condition commonly used to induce general macroautophagy [2,60]. Moreover, KO-caused impairment of autophagy was robustly observed even in the basal condition without CCCP treatment (Figures 4(a/b,g/h), and 12(e/f)). This is a significant observation that further confirms an important role of SIGMAR1 in autophy as it is well known that basal autophagy occurs ubiquitously in normal conditions without cellular stress [2]. These results collectively suggest a general SIGMAR1-medated role in macroautophagy pathways that involve autophagosome-lysosome fusion. However, with regard to autophagosome accumulation, Sigmar1 KO only accounted for 55% of the effect of Baf.A1 (Figure 4(h)) which fully blocks autophagosome-lysosome fusion [44,60]. This contrast implicates that a currently unknown mechanism(s) may compensate for the loss of the SIGMAR1 function. Indeed, sigmar1 KO mice do not exhibit overt phenotypes in normal conditions, suggesting the presence of compensatory modulators [15]. Furthermore, it has been widely observed that SIGMAR1’s protective functions may not obviously manifest until challenged with various stressors [53,54]. These interesting contexts underscore the importance of better understanding the molecular underpinning of SIGMAR1 functions, in particular, its regulations, either direct or indirect, in autophagosome-lysosome fusion.

Conclusions

Neither regulations of autophagosome-lysosome fusion nor molecular functions of SIGMAR1 are well understood. Our results reported herein suggest a SIGMAR1 molecular function in modulating the fusion process. Autophagosome-lysosome fusion is a key step in autophagic processes including mitophagy, and their dysregulation is associated with a broad range of disease conditions. Therefore, further research to elucidate SIGMAR1-associated molecular mechanisms in autophagosome-lysosome fusion will contribute to the understanding of autophagy-related disease mechanisms.