TAX1BP1 recruits ATG9 vesicles through SCAMP3 binding

  1. Yutaro Hama
  2. Yoshitaka Kurikawa
  3. Takahide Matsui
  4. Noboru Mizushima
  5. Hayashi Yamamoto

Reviewed by

Read the full article

Listed in

Log in to save this article

Abstract

Macroautophagy is a cellular process that delivers cytoplasmic material to lysosomes for degradation via autophagosomes. It often involves the selective degradation of ubiquitinated proteins. During selective macroautophagy, five ubiquitin-binding adaptors, p62, NBR1, OPTN, NDP52, and TAX1BP1, form biomolecular condensates with ubiquitinated proteins and recruit ATG9 vesicles, which serve as the initial membrane source required for autophagosome formation. However, the molecular details underlying the cargo/adaptor-dependent recruitment of ATG9 vesicles remain unclear. Here, we show that ATG9 vesicles are recruited by three cargo adaptors: TAX1BP1, NBR1, and OPTN. We also find that ATG9A itself is not the determinant for recruitment by these cargo adaptors, and that TAX1BP1-dependent ATG9 vesicle recruitment is mediated by SCAMP3, a transmembrane protein on the ATG9 vesicles, through binding to the coiled-coil 1 domain of TAX1BP1. These findings provide mechanistic insights into the cargo/adaptor-dependent assembly of ATG9 vesicles in mammals.

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

    The authors do not wish to provide a response at this time.

  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

    Summary:

    In this paper, Hama et al look to address ongoing questions regarding the recruitment of core autophagy factors to protein aggregates during aggrephagy. Using cells that lack all known aggrephagy receptors (PentaKO HeLa cells) provides the authors with a 'blank canvas' with which to cleanly dissect a process otherwise fraught with mechanistic redundancy. By this approach, the authors isolate a previously unidentified mechanism by which TAX1BP1 recruits ATG9A vesicles to ubiquitin-positive aggregates. Using mass spectrometry, the authors identify SCAMP3 as a component of ATG9 vesicles that is responsible for recruiting the vesicles to cargo. Moreover, they provide additional mechanistic insight through the subsequent identification and mutagenesis of a putative interaction interface between TAX1BP1 and SCAMP3.

    Major comments:

    1. Statistics:
      • a) The description of the statistical methods is sparse. In the methods section, the authors' state that statistical methods are described in each figure legend. However, for most figures they were excluded.
      • b) Where statistics are discussed, Mann-Whitney was used. However, this appears to be the incorrect test in many cases. Per GraphPad (the author's preferred statistical package) "Use Mann-Whitney test only to compare two groups. To compare three or more groups, use the Kruskal-Wallis test followed by post tests. It is not appropriate to perform several Mann-Whitney (or t) tests, comparing two groups at a time."
    2. Rigor and reproducibility - The authors report n=~30 cells in most experiments. However, it's unclear if these 30 cells represent more than a single experimental replicate. While the trends in the data are quite convincing, this a significant limitation of this study.
    3. Puncta identification - it's unclear how the authors called puncta. The quantification of these images (i.e. the box and whiskers plots) makes compelling points that support the authors' interpretation. However, in many cases the authors are calling puncta amidst a fairly speckled image. How were puncta distinguished from speckles? When did something rise to the level of a puntum? If it was computationally called, please provide the methods and pipeline. If data were manually scored, then the lack of replicates rises to a more significant concern - would other investigators have scored the data similarly? Is it possible that the data were scored with implicit bias based on expected outcomes of the investigator? Where data scored in a blinded fashion?
    4. Do the findings presented here have functional implications? While the data on recruitment of ATG9A to TAX1BP1 is clear, it's unclear whether the SCAMP3/TAX1BP1 interaction is functionally important for lysosomal delivery of cargo. In part, this is because the authors must work in autophagy-deficient cells (ATG9AKO or FIP200KO) to observed many of their effects. One way to address function might be to ask whether lysosomal delivery of TAX1BP1 is affected upon SCAMP3KO (e.g. in PentaKO cells to remove the effects of other receptors).

    Minor comments:

    1. The authors IP TAX1BP1CC1 and SCAMP3, but an IP between full length TAX1BP1 (WT and K248E) and SCAMP3 would more fully demonstrate the sufficiency of this binding site.
    2. Optional: Is the TAX1BP1 and SCAMP3 interaction direct?. The data are suggestive, but this question is not fully resolved due to the in vivo nature of the assays. Formally, there could be an adapter that facilitates TAX1BP1/SCAMP3 interaction. This could be formalized by testing binding between purified soluble domains of SCAMP3 and TAX1BP1CC1.
    3. Figure 4F - the y axis has a typographical error

    Significance

    This is a crisply written manuscript with a generally clean experimental approach. While there are many directions the authors could take this work in the future, the data largely stand on their own as a short, concise advance. The work adds to our current understanding of the regulation of ATG9A recruitment during selective autophagy, which is under-explored in comparison to starvation-induced autophagy. Mechanistic insights are provided in the form of a new interacting protein SCAMP3, which is present in ATG9A vesicles and is required for the recruitment of ATG9A vesicles to TAX1BP1. The data are predominantly convincing, albeit with significant caveats. Limited sample sizes and lack of functional effects (see 'major concerns') limit the impact of this work. However, the data have merit on their own. This is a specialized study that will be appreciated by those interested in selective autophagy and the mechanism of autophagosome formation.

    Reviewer expertise: selective autophagy and autophagosome formation

  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

    Ubiquitin-binding adaptors are a group of autophagy adaptor proteins that facilitates the formation of isolating membrane around a specific substrate during selective autophagy. At least two ubiquitin-binding adaptors, NBR1 and OPTN, have previously been demonstrated to recruit ATG9-positive vesicles to ubiquitin-positive biomolecular condensates. It is postulated that this recruitment mediates the formation of the isolating membrane around the ubiquitinated biomolecular condensates. In this manuscript, the authors utilized the HeLa cells deficient in the expression of five ubiquitin-binding adaptors, p62, NBR1, OPTN, NDP52 and TAX1BP1, to show that expressing TAX1BP1 colocalizes with ATG9A. The authors interpreted this observation to suggest that TAX1BP1 can recruit ATG9A vesicles independent of other adaptors. Interestingly, the TAX1BP1 positive structures differ from that of OPTN and NBR1 because they do not contain ubiquitin, suggesting a difference in substrate specificity. They further show that TAX1BP1 differs from OPTN and NBR1 because ATG9 recruitment required SCAMP3 to recruit ATG9A. Based on these observations, the authors propose that TAX1BP1 recruits ATG9A via SCAMP3.

    Major comments

    • The work presented in this manuscript is of high quality, however, the reliance on a single experimental assay (colocalization microscopy studies) limits the relevance of its findings. The conclusions are based on colocalization studies in wild-type and CRISPR/cas9 knockout HeLa cell lines. Therefore, it is not certain that their findings are restricted to these PentaKO cell lines. Given that these adaptors are required for selective autophagy, the pentaKO cells may have adapted. Unfortunately, the current study lacks additional systems physiologically relevant models or systems. Such experiments are needed to validate their findings in the HeLa pentaKO cell. (Optional) Alternatively, identify the substrate(s) specific to the TAX1BP1-SCAMP3-ATG9A mediated autophagosome vs. that of NBR1 and/or OPTN in these cell lines.
    • There are several issues with the immunofluorescence data throughout the manuscript. For example, in Figure 1C, the authors quantify 30 cells per condition and perform a Mann-Whitney test, presumably using each cell as a data point, as this is not specified in the legend. There are two problems with this approach, the first being there is only one indicated independent trial performed, and the second that treating each cell as a data point for statistical analysis overinflates the power. These experiments should instead be performed across at least 3 independent trials (especially given the wide range of values seen in some conditions such as Fig. 2b) with 30 cells counted per trial, and statistical analysis should be performed using the means from each trial. See Lord, S.J. et al. (JCB 2020, PMID: 32346721) for a detailed explanation. Further, how each statistical test is performed (using means vs. all points) and the number of independent trials conducted should be communicated in the figure captions. Finally, beyond Fig. 1, the statistical test used is not reported. If a Mann-Whitney test was used for all statistical analyses, this should be revisited as with two or more variables (cell type, treatment, etc.), this is not the appropriate test.

    Minor Points

    • The experimental design rationale is not clear throughout the manuscript. For example, why FIP200 KO cells are used in Fig. 3c and Fig. 4 is not apparent, and only in Fig. 6 (Lines 211-213) is it clearly stated. The authors should revisit the results section and ensure the experimental rationale is clearly explained.
    • It is curious that condensates formed in muGFP-NDP52 and muGFP-TAX1BP1 cells lack ubiquitin (Fig. 3c). Is this image representative of most cells? if so it should be discussed.
    • The authors are missing some relevant citations:
      • Lines 57-58, the authors may consider citing more recent literature (Olivas, T.J. et al., JCB 2023; Broadbent, D.G. et al., JCB 2023; Nguyen, A. et al., Mol Cell 2023) investigating ATG9 vesicles in autophagosome biogenesis.
      • In lines 58-66, the authors do not mention that NBR1 was also shown to bind FIP200 (Turco, E. et al., Nat Communications 2021, PMID: 34471133).
    • Some minor changes are recommended for clarity:
      • In lines 106-110, the authors may consider explaining that "growing" conditions are cells grown in regular culture conditions, as it is a bit unclear whether these cells are treated with a drug, etc., to induce p62-condensate formation. Something as simple as "...p62-double-positive structures under normal culture conditions, hereafter referred to as growing conditions," would help the reader.
      • The general axis labelling of "Ubiquitin-positive rate of FIP200 puncta (%)" (Fig 1D) is confusing wording. The authors may consider changing to "% of ubiquitin-positive FIP200 puncta" for readability.
      • The axis title in Fig. 4F has a typo.
    • Several figure legends include data interpretation, which should generally be excluded from figure legends. For example, the first line of Fig. 4B should be removed.
    • The authors may consider how ATG9A is recruited to ubiquitin-independent selective autophagy cargoes that are degraded independently of NBR1, p62, OPTN, NDP52, and TAX1BP1 in their discussion.
    • Supplementary figure S2-right side blot. For best practice in publications, I encourage the authors to provide a single blot of FIP200, ATG13, and ATG9A immunoblot for Penta KO and its KO derivatives. Presently, the blot appears to have been spliced together

    Significance

    Although different ubiquitin-binding adaptor proteins have been shown to have some substrate specificity, how they mediate the selective degradation of a substrate remains unclear. This manuscript provides evidence that TAX1BP1, like NBR1 and OPTN, can also recruit ATG9A positive structures independent of the autophagosome initiation complex ULK1-complex. However, they do show that the mechanism of ATG9A vesicles differs from the other two adaptors in that it requires SCAMP3, a membrane protein found in endosomes. The data provided by the authors are of high quality. However, the study relies on only one experimental system/assay, which limits the relevance of its findings and could benefit from testing these findings using additional systems and/or physiologically relevant models, as well as increasing the rigor of the quantitative analysis.

    Besides these two major shortcomings, the finding is novel and adds to our current understanding of autophagy adaptor proteins. The difference in the mechanism of ATG9 vesicles compared to NBR1 and OPTN may contribute to the selective nature of these autophagy adaptors.

    This manuscript is most appropriate for specialized basic researchers in the field of selective autophagy.

    Our expertise is in selective autophagy regulation and substrate selectivity of selective autophagy.

  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

    Macroautophagy is a catabolic pathway for various intracellular components mediated by the formation of autophagosomes followed by lysosomal degradation. ATG9 vesicles provide the initial autophagosomal membrane source and thereby recruitment of ATG9 vesicles to the autophagosome formation sites serves as a critical step for autophagy induction. However, the precise molecular mechanism of the ATG9A recruitment is not fully understood. In this study, Hama et al. report two distinct pathways; ULK complex-dependent ATG9 vesicle recruitment during starvation-induced autophagy, and selective autophagy receptor TAX1BP1-dependent ATG9 vesicle trafficking through the binding to SCAMP3, which was identified as an ATG9 vesicle component by the authors. Unfortunately, the authors were unable to demonstrate the impact of the TAX1BP1-SCAMP3-ATG9 vesicle axis on cellular physiology, presumably due to the existence of compensatory mechanisms mediated by other selective autophagy receptors and ULK complex, which limits the impact of the findings presented. Having said that, the study is technically well executed and provides a new insight into the regulation of cargo/receptor-mediated ATG9 vesicle recruitment. This reviewer has a few comments that should be addressed to strengthen the authors' conclusions.

    1. In the Co-IP experiments (Fig 5A-C), binding of TAX1BP1 to SCAMP3 is assessed by using the CC1 domain fraction of TAX1BP1, which may yield an artificial binding to SCAMP3. Could the authors confirm binding of full length TAX1BP1 wild-type and K248E mutant to SCAMP3?
    2. In Fig 5D, the authors showed that SCAMP3 localises to immuno-isolated ATG9A-positive vesicles. Is it a direct interaction between the two proteins? Could the authors provide the evidence that the interaction is retained in the presence of a detergent by immunoprecipitation? If the interaction is indirect, can the authors discuss candidate proteins that mediate binding of SCAMP3 to ATG9 vesicles?
    3. Related to the comment #2, it is interesting that the knockout of ATG9A does not affect SCAMP3-positive "ATG9 vesicle" formation. What is the nature of "ATG9 vesicles" lacking ATG9A?
    4. Could the authors confirm that K284E mutation in TAX1BP1 abrogates the localisation of SCAMP3 to the TAX1BP1 condensates as in Fig 6E? This will reinforce the claim that TAX1BP1 binding to SCAMP3 facilitates ATG9 vesicle recruitment.
    5. Could the authors discuss the potential reasons differentiating TAX1BP1 from other CC-domain containing autophagy receptor proteins (NDP52, OPTN and NBR1), which enables it to bind to SCAMP3. For instance, does TAX1BP1 have charged residues facing outwards in its CC domain that could be responsible for this specificity?
    6. In Fig 3C and 6E, no colocalisation of TAX1BP1 and ubiquitin was observed in TAX1BP1 condensates. In the context of "cargo-driven recruitment" of ATG9 vesicles, what cellular component(s) could trigger TAX1BP1-mediated SCAMP3/ATG9 vesicle recruitment? In the Discussion, authors mentioned that ferritin-NCOA4 was not the target of the TAX1BP1-SCAMP3 axis. Could the authors test if any of the other known TAX1BP1 cargo proteins localise to TAX1BP1 condensates in Penta KO/FIP200 KO/muGFP-TAX1BP1 cells?

    Minor:

    1. Fig 4F: Typo in y axis.

    Significance

    General assessment: provide a summary of the strengths and limitations of the study. What are the strongest and most important aspects? What aspects of the study should be improved or could be developed?

    The main finding of the study is a new pathway of ATG9 vesicle recruitment through the interaction of TAX1BP1 with SCAMP3, which provides a novel insight into molecular mechanisms of autophagosome biogenesis. However, the axis is implied to be redundant for functional autophagy in wild-type cells, and lack of data providing a biological function of the axis in cellular physiology will limit impact attracting broader readers outside of molecular mechanism of autophagy.

    Advance: compare the study to the closest related results in the literature or highlight results reported for the first time to your knowledge; does the study extend the knowledge in the field and in which way? Describe the nature of the advance and the resulting insights (for example: conceptual, technical, clinical, mechanistic, functional,...).

    Interaction between ATG9A and selective autophagy receptors OPTN and NBR1 has been reported (doi: 10.1083/jcb.201912144; 10.15252/embr.201948902). This study provides an additional mechanistic insight into the regulation of ATG9 vesicle recruitment through another autophagy receptor TAX1BP1 interacting with SCAMP3 which was newly identified as an ATG9 vesicle component in this study. Given the predominant functions of ATG9A in TNF cytotoxicity and plasma membrane integrity as well as TAX1BP1 in neuronal proteostasis and iron homeostasis (doi: 10.1126/science.add6967; 10.1038/s41556-021-00706-w; 10.1016/j.molcel.2020.10.041; 10.15252/embr.202154278), the interaction between TAX1BP1 and ATG9A would potentially have uncovered but important role in mammals. Autophagy-independent lysosomal degradation regulated via ULK component, ATG9 and TAX1BP1 might be related in this context (10.1016/j.celrep.2017.08.034).

    Audience: describe the type of audience ("specialized", "broad", "basic research", "translational/clinical", etc...) that will be interested or influenced by this research; how will this research be used by others; will it be of interest beyond the specific field?

    This study will be of interest to the basic researchers working on molecular and structural mechanisms of bulk and selective macroautophagy. Unfortunately, the lack of data demonstrating the relevance of the findings for cellular physiology will limit the impact on researchers in broader fields such as pathology and drug discovery.

    Please define your field of expertise with a few keywords to help the authors contextualize your point of view. Indicate if there are any parts of the paper that you do not have sufficient expertise to evaluate.

    Molecular cell biology of autophagy; neurodegenerative and lysosomal storage disorders.

  5. Version published to 10.1101/2023.08.18.553817v1 on bioRxiv