The ER membrane protein complex governs lysosomal turnover of a mitochondrial tail-anchored protein, BNIP3, to restrict mitophagy

  1. Jose M Delgado
  2. Logan Wallace Shepard
  3. Sarah W Lamson
  4. Samantha L Liu
  5. Christopher J Shoemaker

Reviewed by

Read the full article

Listed in

Log in to save this article

Abstract

Lysosomal degradation of autophagy receptors is a common proxy for selective autophagy. However, we find that two established mitophagy receptors, BNIP3 and BNIP3L/NIX, violate this assumption. Rather, BNIP3 and NIX are constitutively delivered to lysosomes in an autophagy-independent manner. This alternative lysosomal delivery of BNIP3 accounts for nearly all of its lysosome-mediated degradation, even upon mitophagy induction. To identify how BNIP3, a tail-anchored protein in the outer mitochondrial membrane, is delivered to lysosomes, we performed a genome-wide CRISPR screen for factors influencing BNIP3 flux. By this approach, we revealed both known modifiers of BNIP3 stability as well as a pronounced reliance on endolysosomal components, including the ER membrane protein complex (EMC). Importantly, the endolysosomal system regulates BNIP3 alongside, but independent of, the ubiquitin-proteosome system (UPS). Perturbation of either mechanism is sufficient to modulate BNIP3-associated mitophagy and affect underlying cellular physiology. In short, while BNIP3 can be cleared by parallel and partially compensatory quality control pathways, non-autophagic lysosomal degradation of BNIP3 is a strong post-translational modifier of BNIP3 function. More broadly, these data reveal an unanticipated connection between mitophagy and TA protein quality control, wherein the endolysosomal system provides a critical axis for regulating cellular metabolism. Moreover, these findings extend recent models for tail-anchored protein quality control and install endosomal trafficking and lysosomal degradation in the canon of pathways that ensure tight regulation of endogenous TA protein localization.

Article activity feed

  1. Review Commons

    Note: This rebuttal was posted by the corresponding author to Review Commons. Content has not been altered except for formatting.

    Learn more at Review Commons

    Reply to the reviewers

    1. General Statements

    __Response: __Thank you to all the reviewers for their helpful efforts on behalf of our manuscript. At current, we have addressed most of the reviewers’ major comments, including providing additional replicates for many experiments and clarifying ambiguous points in the text. Related data, figures and text have been adjusted accordingly. We believe that these changes have improved our manuscript, both strengthening our main conclusions and clarifying ambiguous text.

    Several still-ongoing experiments are elaborated below. These experiments are well within the abilities of our lab and can be completed in short order.

    Specific responses to the individual concerns addressed by the reviewers are outlined below.

    Please feel free to contact me if I can be of any help in the decision process.

    2. Description of the planned revisions

    Insert here a point-by-point reply that explains what revisions, additional experimentations and analyses are planned to address the points raised by the referees.

    [Reviewer 1]

    Comment: Across the manuscript, NIX levels appear to be unresponsive to most treatments in the MDA-MB-231 line, including hypoxia treatment. This is an unusual result and raises questions about the role of NIX in MDA-MB-231 line, mainly that BNIP3 is the primary driver of mitophagy in this system. Indeed, Figure 7D indicates that there is very little mitophagy contribution by NIX since knockout of BNIP3 is sufficient to abolish mitophagy almost completely. Therefore, the effects seen on mitophagy following EMC3 knockout in Figure 7 might be smaller in a line that is responsive to NIX mitophagy. It would be beneficial to analyse basal mitophagy flux in an additional cell line, for example U2OS (FigS1E) in which NIX is responsive to hypoxia.

    Response: Thank you for bringing this intriguing insight to our attention. We have seen that EMC3 knockout prevents lysosomal delivery of BNIP3 in U2OS cells (Fig S2D). However, we don’t know what the effects on mitophagy are in U2OS, or the extent to which mitophagy is dependent on BNIP3 and/or NIX. To test this, we will perform the suggested experiment, taking mt-Keima expressing U2OS cells testing the role of NIX and/or BNIP3 in mitophagy.

    Comment: Following on from comment 1 above, Figure 7 would benefit with an analysis of hypoxia (or DFP, or cobalt chloride) stimulation of mitophagy to assess whether mitophagy levels are higher in EMC3 KOs. The authors argue that BNIP3 is trafficked to the ER during mitophagy and is not turned over by mitophagy itself, it would therefore be interesting to test if BNIP3 is prevented from being removed from mitochondria whether this would affect the rate or levels of mitophagy under stimulating conditions.

    __Response: __To address this question, we will perform mitoflux analysis on EMC3 KO cells +/- hypoxia.

    Comment: Figure 4B: The localisation of tf-BNIP3 is reminiscent of ER in BTZ treated samples. How much of the protein is on mitochondria in the presence of BTZ? Does MLN4924 cause a similar issue?

    __Response: __To address this question, we will perform fluorescence microscopy of tf-BNIP3 cells co-expressing mito-BFP under these treatments and utilize our Coloc2 plugin pipeline to monitor correlation.

    Comment: Can the authors assess whether BNIP3 that is on mitochondria is transferred to the ER (perhaps through photoswitchable GFP-BNIP, activated on mitos and then observe its transfer to ER)? This seems important in order to address the possibility that BNIP3 that is being turned over by the endolysosome is being delivered directly to the ER.

    __Response: __This is an interesting question and a curiosity also shared by Reviewer #2. To test this hypothesis, we will utilize a photo-switchable Dendra2 fluorophore to track BNIP3 in the cell via microscopy.

    [Reviewer #2]

    Comment: How is BNIP3 inserted into the outer membrane? A previous study from the Weissman lab proposed that MTCH2 serves as insertase. The authors did not mention MTCH1 and MTCH2 in context of Fig. 2B. Were these proteins not found? Did the authors test the relevance of MTCH2 in their assay? This aspect should be addressed and mentioned.

    __Response: __Thank you for the insight and suggestion. We were intrigued when the Weissman/Voorhees paper characterizing MTCH1/2 was published. Consistent with their findings, MTCH2 was found in the “suppressor” population of our tf-BNIP3 CRISPR screen, but given our 0.5-fold change threshold, the gene was not validated (fold change value = 0.46, Table S1). We suspect the lack of significance stems from the redundancy with MTCH1. Consequently, we would hypothesize that MTCH1/2 are the responsible insertases. To formally address this suggestion, we plan to genetically perturb MTCH1/2 and look at BNIP3 localization and mitophagy.

    Comment: The authors generated an interesting BNIP3 mutant with a C-terminal Fis1 anchor. This variant is constantly located in the outer membrane (which is shown here). The physiological consequence of the constitutive distribution on mitochondria is however only superficially studied. The authors should characterize this interesting mutant in some more depth.

    __Response: __In the original manuscript, we characterized BNIP3(Fis1TMD) for lysosomal delivery and mitophagy. Going forward, we will perform Seahorse oxygen consumption experiments and mitochondrial network analysis to view the physiological consequences of constitutive expression of BNIP3(Fis1TMD) on the outer membrane.

    3. Description of the revisions that have already been incorporated in the transferred manuscript

    Please insert a point-by-point reply describing the revisions that were already carried out and included in the transferred manuscript. If no revisions have been carried out yet, please leave this section empty.

    [Reviewer #1]

    Comment: Continuing from comment 2, given that the authors conclude that BNIP3 is not turned over by mitophagy, can they examine whether BNIP3 is excluded from sealed mitophagosomes?

    __Response: __We have softened the wording of our conclusions to reflect that the vast majority of BNIP3 lysosomal degradation is by this alternative pathway and not mitophagy. However, we do not wish to completely dismiss that BNIP3 is present on mitophagosomes. Rather, if mitophagosomes contain BNIP3, they seemingly account for only a very small portion of BNIP3 degradation in the cell, to the extent that it is not easily detectable by our assays (Lines 414-419). Definitively identifying whether BNIP3 is in sealed mitophagosomes will be part of future studies using CLEM or FIB-SEM techniques.

    Comment: Is the BNIP3(FisTMD) expressed to equivalent levels to WT BFP-BNIP3? Given that theFis1 form of BNIP3 cannot traffic to endolysosomes, its levels might be higher. In addition, overexpression of the BNIP3-Fis construct was used to make the argument that dimerization is not important for mitophagy. But the authors should also take into account the possibility that with overexpression, the potential efficiency afforded to mitophagy via dimerization of endogenous proteins may be negated, and therefore hidden. Given this, I don’t think that the authors can confidently conclude that dimerization does not contribute to mitophagy, and that instead its main role is ER-endolysosomal turnover of BNIP3.

    __Response: __We thank the reviewer for pointing out the possible over-interpretation of our data. Overexpression is an important caveat to consider. We would expect the Fis1 form of BNIP3 to be higher in protein levels given its deficiency in endolysosomal trafficking. Still, as the reviewer points out, over-expression could be mitigating the effect of our dimerization mutants. This caveat is now discussed in the manuscript and our interpretations regarding this fact have been greatly softened (Lines 373-376, Lines 449-462).

    Comment: Please include molecular weight markers for all western blots.

    __Response: __All western blots have now been labeled with molecular weight markers.

    Comment: Figure 5A-G: These data do not make a convincing case for the role of dimerization and are very difficult to follow. Only the mislocalized S172A mutant was responsive to Baf treatment, while the LG swap mutant which is mitochondrial and cannot dimerize is unaffected by Baf treatment. Figure 5H-I utilize a construct of BNIP3 that is missing most of the protein and which has very low turnover (Figure 5B). Unfortunately these results don’t make a highly convincing case about the biology of native, full length, mitochondrial BNIP3. The authors are advised to either strengthen the dimerization argument, or perhaps lighten the language around the main conclusions from these data.

    __Response: __Thank you for bringing the lack of clarity to our attention. Both dimer mutants of BNIP3 (S172A and LG swap) are insensitive to Baf-A1 treatment. These results hold for full-length BNIP3 using either the tf (Fig 5D) or IRES (Fig 5I) reporter. To demonstrate that defects in lysosomal transport were due to dimerization defects (and not other, unanticipated effects of the mutations), we looked at whether chemically induced dimerization could reverse the trafficking defects. Indeed, forced dimerization of the ER-restricted variant rescued ER-to-lysosome trafficking. From this, we conclude that that dimerization is a critical facet of BNIP3 trafficking to the lysosome.

    We have re-worked the relevant text (both in results and discussion) to clarify major points and lighten the language around the conclusions from these data (described below).

    First, as mentioned above, we have added a significant discussion about the limitations of our assay and of possible interpretations. (Lines 300-303, Lines 323-326, Lines 483-489).

    Second, with regards to the specific construct used in this experiment, we have expanded the results section to better describe our rationale and approach (Lines 304-308). In short, because dimerization of native BNIP3 occurs within the membrane, we aimed to place the DmrB domain as close to the TM segment as possible. Due to the topology of TA proteins, a C-terminal tag isn’t possible. Therefore, we used the shortest truncation version of BNIP3 (117-end) that undergoes measurable lysosomal delivery. This was an important experimental consideration, and one we did not sufficiently rationalize in the original manuscript. We now include this point in the text.

    [Reviewer #2]

    Comment: The authors show that BNIP3 on the ER is not stable but degraded by the proteasome. Does this require ERAD factors? Is the mitochondrial BNIP3 protein likewise degraded by proteasomal degradation? It is not clear whether both BNIP3 pools are constantly turned over or whether degradation exclusively/predominantly occurs on the ER surface.

    __Response: __These are fascinating mechanistic questions. We hope to thoroughly address these questions in a subsequent study. However, as a teaser, we have included the basic answer to these questions in Fig 5I.

    To preliminarily characterize the proteasomal degradation of ER- and mitochondrial-BNIP3, we utilized our IRES reporter system - adapted from Steve Elledge’s system for degron monitoring (Fig 5I). Strikingly, our ER-restricted BNIP3 mutation (S172A) is sensitive to inhibition of both the proteasome and the AAA-ATPase p97/VCP, a key extractase for ERAD substrates. These data tentatively suggest an ERAD-dependent degradation mechanism (although many follow-up studies will be needed to confirm the mechanistic details). In sharp contrast, our mitochondrial-restricted mutant (LG Swap) is sensitive to proteasome inhibition by Bortezomib, but it is insensitive to VCP inhibition. The differential requirement for VCP suggests that proteasomal degradation occurs on both cellular pools of BNIP3 albeit through different mechanisms.

    Comment: The results of the screen shown in Fig. 2B are particularly interesting for readers. The glutathione peroxidase GPX4 was found as a top hit among the EMC components. GPX4 protects membranes (including those of mitochondria) against oxidative damage, is a major component of ferroptosis and linked to mitochondrial dysfunction and mitophagy. The authors should mention this interesting hit in the context of their discussion of the lipid-sensing properties of the dimerizing TM domains of BNIP3.

    __Response: __Thank you to Reviewer #2 for bringing this to our attention. The relationship between GPX4 and BNIP3 flux is very interesting. We have incorporated GPX4 into the discussion section (Lines 457-459).

    [Reviewer #3]

    Comment: For all of the tf-BNIP3 FACS data (all violin plots), it is unclear how many biological replicates were performed. The author only stated that at least 10,000 cells were analyzed per sample, but I believe this is for each biological replicate. To better demonstrate the biological replicates, the authors should consider using bar graphs of the medians(triplicates) with error bars.

    Response: We have included biological replicates of FACS data in all primary figures (except for Fig.1C). Biological replicates, represented as medians (in triplicate), are indicated in figure legends.

    Comment: In Fig 3D, it is unclear as to why there is no basal state accumulation of BNIP3 protein levels compared to Baf1A treated condition especially with USO1 and SAR1A KO samples. Is this because BNIP3 are targeted for proteasomal degradation? I think Fig 3D should include a BTZ treatment next to Baf1A to account for the lack of basal state accumulation of BNIP3.

    Response: We apologize for the lack of clarity on this point. Yes, the reviewer’s interpretation of the data is correct. This point is more clearly elaborated in the text of our revised manuscript (Lines 219-223). Our results indicate that when lysosomal degradation is diminished, the expected increase in total BNIP3 protein levels is attenuated by proteasomal degradation (as evidenced by the hyperstability of BNIP3 upon Bortezomib treatment in mutant backgrounds). As requested, we have included the same knockout panel, now treated with BTZ (Fig S2E). These genetic data are further supported by Fig 3E, where a small molecule inhibitor of vesicle trafficking, Brefeldin-A, ameliorates the effect of lysosomal inhibition (BafA1) but exacerbates the effect of proteasome inhibition.

    Comment: Truncation of proteins could affect their protein stability even during their synthesis. For Fig 5B and 6B, the authors should show the blots for the expression of the different truncated mutants to prove that the change in BNIP3 stability and their effect of mitoflux (or lack thereof), is not due to poor expression of these mutants.

    Response: These were important potential caveats to document, and we thank the reviewer for their comment.

    We note that, due to differences in transduction efficiency, western blot data is an incomplete measure for relative expression levels – it cannot distinguish between fraction of cells transduced and expression level per cell. However, RFP fluorescence (Fig 5B) and BFP fluorescence (Fig 6B) are fluorescent internal controls allowing us to assess expression levels with single cell resolution. We have provided histograms of RFP and/or BFP intensity (new Fig S4A, Fig S5B), which provides support that overall expression levels of these constructs are similar. Critically, any variation we observe does not correlate with any of the effects we report.

    In addition, we have clarified the figure axis in Fig 5B to indicate that the value we are reporting is the “fold-stabilization upon BafA1 treatment”. The original figure legend wasn’t clear. Our metric (fold-stabilization) is internally normalized to compensate for differences in expression level. This is an important clarification.

    Comment: For the data in Fig 7, the authors demonstrated that treating cells with proteasomal inhibitor increases mitoflux. Since the proteasome targets monomeric BNIP3 for degradation, the logical assumption is that BTZ drives dimerization of BNIP3. Can the authors demonstrate this in an approach similar to Fig 5C? This simple experiment will add significant insight into the study.

    __Response: __Thank you for the suggestion. As Fig 5C relied on BNIP3 over-expression, we thought it even more informative to assess the effects of BTZ on dimerization of endogenous BNIP3. Indeed, we see accumulation of an SDS-resistant BNIP3 dimer in cells treated with BTZ (new Fig S2E, line 221). We hypothesize that BTZ indirectly drives dimerization of BNIP3 by accumulating the total levels of the protein, potentiating monomers to form additional stable dimers.

    Comment: In line 168-169, "In addition, multiple suppressor genes identified from our screen had previously been reported including TMEM11..." -- Unclear what biology they are reported to be involved in

    __Response: __We have clarified this line to read: "In addition, we recovered multiple known suppressors of BNIP3 flux, including outer membrane protein spatial restrictor TMEM11, mitochondrial protein import factors DNAJA3 and DNAJA11, and mitochondrial chaperone HSPA9"

    Comment: Along the line with Major comment 2, the explanation for Fig 3D needs to be better elaborated, perhaps to include the role of proteasome already at this point (if the authors think this is the reason why basal BNIP3 levels remains low with USO1 and SAR1A KO).

    __Response: __We have included a discussion about compensation by the proteasome in these genetic backgrounds (lines 219-226) and have referred to the newly incorporated western blot (new Fig S2E).

    Comment: Line 302-304, I believe that statement only refers to Fig S4C and the statement for Fig5G is in the next sentence. Please remove Fig5G from line 304. It was confusing to read.

    __Response: __The reference of Fig 5G has been removed.

    Comment: Line 367, there is a reference for Fig S5C but that figure is missing.

    __Response: __The spurious reference has been removed.

    Comment: Line 410-411, are there any reported clinical cases of EMC mutations with phenotypes that could be explained by elevated mitophagy?

    __Response: __Thank you for the suggestion. There are clinical presentations of EMC mutations and splice variants in diseases and conditions related to the central nervous system (PMID: 23105016, PMID: 26942288, PMID: 29271071). However, all characterization has been done in the clinical setting looking at clinical presentations/symptoms and not molecular or cellular characterization. We have added a line to the discussion about this speculative correlation between EMC deficiency and mitophagy (lines 516-519).

    4. Description of analyses that authors prefer not to carry out

    Please include a point-by-point response explaining why some of the requested data or additional analyses might not be necessary or cannot be provided within the scope of a revision. This can be due to time or resource limitations or in case of disagreement about the necessity of such additional data given the scope of the study. Please leave empty if not applicable.

    [Reviewer #1]

    Comment: Figure 3B: Are the red puncta observed in USO1 and SAR1A cells a product of higher levels of ER-phagy owing to BNIP3's high presence on the ER membrane?

    __Response: __This is an intriguing hypothesis. We will test whether this is true using a USO1/ATG9A dual KO. However, we don’t think this result is critical to the overall arc of the manuscript and we will not include these data if they indicate otherwise.

  2. Review Commons

    Note: This preprint has been reviewed by subject experts for Review Commons. Content has not been altered except for formatting.

    Learn more at Review Commons

    Referee #3

    Evidence, reproducibility and clarity

    In this manuscript, the authors provided a comprehensive study on the regulation of BNIP3 protein levels by both the non-autophagic lysosomal and proteasomal degradation pathways. Using various nifty tools, the authors demonstrated that ER-localised BNIP3 dimers are rerouted via the ER-lysosomal pathway for non-autophagic lysosomal degradation whereas BNIP3 monomers on both the ER and mitochondria are targeted for proteasomal degradation. Together, these pathways help to repress hyperactivation of basal mitophagy.

    Overall, I find this project very well executed and the manuscript is also very clear and concise. The key conclusions of this work are supported by orthogonal approaches, thus making the data highly convincing. I only have a few major and several minor comments for this manuscript:

    Major comments:

    1. For all of the tf-BNIP3 FACS data (all violin plots), it is unclear how many biological replicates were performed. The author only stated that at least 10,000 cells were analysed per sample but I believe this is for each biological replicate. To better demonstrate the biological replicates, the authors should consider using bar graphs of the medians (triplicates) with error bars.
    2. In Fig 3D, it is unclear as to why there is no basal state accumulation of BNIP3 protein levels compared to Baf1A treated condition especially with USO1 and SAR1A KO samples. Is this because BNIP3 are targeted for proteasomal degradation? I think Fig 3D should include a BTZ treatment next to Baf1A to account for the lack of basal state accumulation of BNIP3.
    3. Truncation of proteins could affect their protein stability even during their synthesis. For Fig 5B and 6B, the authors should show the blots for the expression of the different truncated mutants to prove that the change in BNIP3 stability and their effect of mitoflux (or lack thereof), is not due to poor expression of these mutants.
    4. For the data in Fig 7, the authors demonstrated that treating cells with proteasomal inhibitor increases mitoflux. Since the proteasome targets monomeric BNIP3 for degradation, the logical assumption is that BTZ drives dimerization of BNIP3. Can the authors demonstrate this in an approach similar to Fig 5C? This simple experiment will add significant insight into the study.

    Minor comments:

    1. In line 168-169, "In addition, multiple suppressor genes identified from our screen had previously been reported including TMEM11..." -- Unclear what biology they are reported to be involved in
    2. Along the line with Major comment 2, the explanation for Fig 3D needs to be better elaborated, perhaps to include the role of proteasome already at this point (if the authors think this is the reason why basal BNIP3 levels remains lowl with USO1 and SAR1A KO).
    3. Line 302-304, I believe that statement only refers to Fig S4C and the statement for Fig 5G is in the next sentence. Please remove Fig5G from line 304. It was confusing to read.
    4. Line 367, there is a reference for Fig S5C but that figure is missing.
    5. Line 410-411, are there any reported clinical cases of EMC mutations with phenotypes that could be explained by elevated mitophagy?

    Significance

    My expertise lies in organelle-selective autophagy and protein homeostasis. Overall, I think this is a very strong manuscript and the data are very solid. The work adds to our current understanding of the basal regulation of BNIP3 which was not previously explored. The novelty of this work lies in the unexpected regulation of BNIP3 via an autophagy-independent, lysosomal pathway and the observation has the potential to be extended to the regulation of the stability of other tail-anchored proteins. This is a very specialised study and will be of interest to the mitophagy and transmembrane protein regulation community.

  3. Review Commons

    Note: This preprint has been reviewed by subject experts for Review Commons. Content has not been altered except for formatting.

    Learn more at Review Commons

    Referee #2

    Evidence, reproducibility and clarity

    Mitochondria are degraded by autophagy in a reaction that depends on mitophagy receptors, such as BNIP3 and BNIP3L/NIX. These receptors are regulated on the level of gene expression, but also on a post-translational level. This study elucidates the processes which control the levels of BNIP3 in a process that relies on the alternative distribution of the protein to different cellular compartments. BNIP3 is a tail-anchored protein that is inserted into the mitochondrial outer membrane and, alternatively, into the ER membrane. The authors show that insertion into the ER is dependent on the EMC complex and, to a lesser extent, on the GET complex. After insertion into the ER membrane, BNIP3 can be trafficked to lysosomes for degradation or, alternatively, be degraded by proteasomal proteolysis. This study provided evidence that the conditional distribution of BNIP3 to these different intracellular locations is used to control mitophagy even though the specific conditions which determine the alternative destinations remain largely unexplored. The study is of high technical quality, all important controls are shown, and the text is well written.

    Specific points

    1. The mechanism of the alternative distribution is not addressed here. Is the location of BNIP3 dependent on where the newly synthesized protein is initially targeted to (such as in the case of Pink1) or is there a constant redistribution and flux of the protein between the two membranes? This is an important aspect which should be experimentally addressed and some data to this should be already published as part of this study since this aspect is important for the final model proposed.
    2. How is BNIP3 inserted into the outer membrane? A previous study from the Weissman lab proposed that MTCH2 serves as insertase. The authors did not mention MTCH1 and MTCH2 in context of Fig. 2B. Were these proteins not found? Did the authors test the relevance of MTCH2 in their assay? This aspect should be addressed and mentioned.
    3. The authors show that BNIP3 on the ER is not stable but degraded by the proteasome. Does this require ERAD factors? Is the mitochondrial BNIP3 protein likewise degraded by proteasomal degradation? It is not clear whether both BNIP3 pools are constantly turned over or whether degradation exclusively/predominantly occurs on the ER surface.
    4. The authors generated an interesting BNIP3 mutant with a C-terminal Fis1 anchor. This variant is constantly located in the outer membrane (which is shown here). The physiological consequence of the constitutive distribution on mitochondria is however only superficially studied. The authors should characterize this interesting mutant in some more depth.
    5. The results of the screen shown in Fig . 2B are particularly interesting for readers. The glutathione peroxidase GPX4 was found as a top hit among the EMC components.GPX4 protects membranes (including those of mitochondria) against oxidative damage, is a major component of ferroptosis and linked to mitochondrial dysfunction and mitophagy. The authors should mention this interesting hit in the context of their discussion of the lipid-sensing properties of the dimerizing TM domains of BNIP3.

    Significance

    Many studies in the last years focused on the roles of Pink and Parkin in the context of mitophagy, a system that also relies on alternative protein targeting (in that case between the inner and outer membrane of mitochondria). The study here shows that BNIP3, another highly important mitophagy receptor, uses in principle a similar strategy, however, here the alternative targeting occurs between the mitochondrial outer membrane and the ER membrane. Mechanistic insights are provided, for example also into the different domains of BNIP3 and their relevance for targeting and mitophagy. The study therefore addresses an important aspect, is of excellent quality and will be of interest for a broad readership.

  4. Review Commons

    Note: This preprint has been reviewed by subject experts for Review Commons. Content has not been altered except for formatting.

    Learn more at Review Commons

    Referee #1

    Evidence, reproducibility and clarity

    Recent work by several groups has revealed that NIX and BNIP3 levels can be regulated through ubiquitination, mediated by FBXL4, to restrict mitophagy. In this study, the authors identify an additional avenue for the regulation of BNIP3 levels involving the transfer BNIP3 from mitochondria to the ER and eventually into the endolysosomal system for degradation. In addition, the authors argue that most of BNIP3 turnover during mitophagy is through the newly identified ER pathway and not through mitophagy. There was little to no endolysosomal turnover observed for NIX, and therefore ethe authors predominantly focused on BNIP3. Key ER transfer factors required for BNIP3 endolysosomal turnover were identified through whole genome CRISPR/Cas screening, and include EMC3. Knockout of EMC3 results in slightly higher levels of mitophagy under basal conditions, and higher levels of mitophagy following proteasome inhibition with BTZ, supporting the overall conclusion that BNIP3 levels are regulated by lysosomal turnover.

    Major Comments:

    1. Across the manuscript, NIX levels appear to be unresponsive to most treatments in the MDA-MB-231 line, including hypoxia treatment. This is an unusual result and raises questions about the role of NIX in MDA-MB-231 line, mainly that BNIP3 is the primary driver of mitophagy in this system. Indeed, Figure 7D indicates that there is very little mitophagy contribution by NIX since knockout of BNIP3 is sufficient to abolish mitophagy almost completely. Therefore, the effects seen on mitophagy following EMC3 knockout in Figure 7 might be smaller in a line that is responsive to NIX mitophagy. It would be beneficial to analyse basal mitophagy flux in an additional cell line, for example U2OS (Fig S1E) in which NIX is responsive to hypoxia.
    2. Following on from comment 1 above, Figure 7 would benefit with an analysis of hypoxia (or DFP, or cobalt chloride) stimulation of mitophagy to assess whether mitophagy levels are higher in EMC3 KOs. The authors argue that BNIP3 is trafficked to the ER during mitophagy and is not turned over by mitophagy itself, it would therefore be interesting to test if BNIP3 is prevented from being removed from mitochondria whether this would affect the rate or levels of mitophagy under stimulating conditions.
    3. Continuing from comment 2, given that the authors conclude that BNIP3 is not turned over by mitophagy, can they examine whether BNIP3 is excluded from sealed mitophagosomes?
    4. Figure 4B: The localisation of tf-BNIP3 is reminiscent of ER in BTZ treated samples. How much of the protein is on mitochondria in the presence of BTZ? Does MLN4924 cause a similar issue?
    5. Is the BNIP3(FisTMD) expressed to equivalent levels to WT BFP-BNIP3? Given that the Fis1 form of BNIP3 cannot traffic to endolysosomes, its levels might be higher. In addition, overexpression of the BNIP3-Fis construct was used to make the argument that dimerization is not important for mitophagy. But the authors should also take into account the possibility that with overexpression, the potential efficiency afforded to mitophagy via dimerization of endogenous proteins may be negated, and therefore hidden. Given this, I don't think that the authors can confidently conclude that dimerization does not contribute to mitophagy, and that instead its main role is ER-endolysosomal turnover of BNIP3.
    6. Can the authors assess whether BNIP3 that is on mitochondria is transferred to the ER (perhaps through photoswitchable GFP-BNIP, activated on mitos and then observe its transfer to ER)? This seems important in order to address the possibility that BNIP3 that is being turned over by the endolysosome is being delivered directly to the ER.

    Minor comments:

    1. Figure 3B: Are the red puncta observed in USO1 and SAR1A cells a product of higher levels of ER-phagy owing to BNIP3's high presence on the ER membrane?
    2. Please include molecular weight markers for all western blots.
    3. Figure 5A-G: These data do not make a convincing case for the role of dimerization and are very difficult to follow. Only the mislocalized S172A mutant was responsive to Baf treatment, while the LG swap mutant which is mitochondrial and cannot dimerize is unaffected by Baf treatment. Figure 5H-I utilise a construct of BNIP3 that is missing most of the protein and which has very low turnover (Figure 5B). Unfortunately these results don't make a highly convincing case about the biology of native, full length, mitochondrial BNIP3. The authors are advised to either strengthen the dimerization argument, or perhaps lighten the language around the main conclusions from these data.

    Significance

    Overall, this is a valuable and important study that provides an important new advance into how mitophagy is regulated by mitophagy receptors. It adds another layer of regulation in addition to the ubiquitin-proteasome mediated restriction of mitophagy reported by others. The data are predominantly convincing and make a strong argument for endolysomal turnover of BNIP3 to regulate its levels. This study will be of high interest to the field of mitophagy. There is also general interest to the field of mitochondrial biology that a TA mitochondrial (and peroxisomal) protein can be extracted from mitochondria, transferred to the ER, and eventually to the endolysosomal system.

    Reviewer expertise: mitophagy mechanisms, autophagosome formation

  5. Version published to 10.1101/2023.03.22.533681v1 on bioRxiv