The Calpain-7 protease functions together with the ESCRT-III protein IST1 within the midbody to regulate the timing and completion of abscission

  1. Elliott L Paine
  2. Jack J Skalicky
  3. Frank G Whitby
  4. Douglas R Mackay
  5. Katharine S Ullman
  6. Christopher P Hill
  7. Wesley I Sundquist

  • Curated by eLife

    eLife logo

    eLife assessment

    The manuscript by Paine and coworkers provides a fundamental improvement on how the enzymatic activity of CALPAIN7 (a Cys protease) influences cytokinesis mediated by the ESCRT (endosomal sorting complexes required for transport) pathway. The authors provide a convincing molecular and cellular basis for one of the several key steps involved in membrane fission during the separation of dividing eukaryotic cells. These findings should be of interest to a wide scientific audience including biochemists, structural biologists, and cell biologists.

Reviewed by

Read the full article

Listed in

Log in to save this article

Abstract

The Endosomal Sorting Complexes Required for Transport (ESCRT) machinery mediates the membrane fission step that completes cytokinetic abscission and separates dividing cells. Filaments composed of ESCRT-III subunits constrict membranes of the intercellular bridge midbody to the abscission point. These filaments also bind and recruit cofactors whose activities help execute abscission and/or delay abscission timing in response to mitotic errors via the NoCut/Abscission checkpoint. We previously showed that the ESCRT-III subunit IST1 binds the cysteine protease Calpain-7 (CAPN7) and that CAPN7 is required for both efficient abscission and NoCut checkpoint maintenance (Wenzel et al., 2022). Here, we report biochemical and crystallographic studies showing that the tandem microtubule-interacting and trafficking (MIT) domains of CAPN7 bind simultaneously to two distinct IST1 MIT interaction motifs. Structure-guided point mutations in either CAPN7 MIT domain disrupted IST1 binding in vitro and in cells, and depletion/rescue experiments showed that the CAPN7-IST1 interaction is required for (1) CAPN7 recruitment to midbodies, (2) efficient abscission, and (3) NoCut checkpoint arrest. CAPN7 proteolytic activity is also required for abscission and checkpoint maintenance. Hence, IST1 recruits CAPN7 to midbodies, where its proteolytic activity is required to regulate and complete abscission.

Article activity feed

  1. Version published to 10.7554/elife.84515 on eLife
  2. eLife

    Author Response

    Reviewer #2 (Public Review):

    Please note that I am not a structural biologist and cannot critically evaluate the details of figures 1 to 3; my review focuses on the cell biology experiments in figures 4 and 5.

    Paine and colleagues investigated structural requirements for the interaction between the ESCRT-III subunit IST1 and the protease CAPN7. This is a continuation of previous work by the same group (Wenzel et al., eLife 2022), which showed that Capn7 is recruited to the midbody by Ist1 and that Capn7 promotes both normal abscission and NoCut abscission checkpoint function. In this article, the structural determinants of the Ist1-Capn7 interaction are characterised in more detail, focusing on the structure of Capn7 MIT domains and their binding to Ist1. Notably, point mutations in Capn7 MIT domains known to mediate binding to Ist1 and midbody recruitment are shown here to be required for abscission functions, as expected from the authors' previous paper. Furthermore, the report shows that a Capn7 point mutant lacking proteolytic activity behaves as a loss-of-function in abscission assays, despite showing normal midbody localisation. These are important results that will help in future studies to understand how the Capn7 protease regulates abscission mechanistically.

    The report is clearly written and the results support the main conclusions. Some technical limitations and alternative interpretations of the data should be discussed in the text, as outlined below.

    1. It is not always clearly stated how the results presented in this report relate to those in the Wenzel paper. For example, the finding that Ist1 recruits Capn7 to midbodies (p. 6 and figure 4) was first shown in the Wenzel paper. The novelty here is not that Capn7 MIT mutants fail to localise to midbodies, but that they phenocopy the previously described knockdown of Capn7, failing to support normal abscission and NoCut function (fig. 5). This supports and extends the findings of Wenzel et al. It is important to make this explicit and explain the conceptual advances shown here more clearly.

    We take the reviewer’s point and we have now clarified this issue in the text (e.g., page 7, lines 4-5).

    1. The NoCut checkpoint can be triggered by chromatin bridges, DNA replication stress, and nuclear basket defects, but only basket defects are tested here. Therefore, it is not clear if NoCut is still functional in Capn7-defective cells after replication stress and/or with chromatin bridges. Ideally, this should be tested experimentally, or alternatively discussed in the text, especially since the molecular details of how NoCut is engaged under different conditions remain unclear. For example, "abscission checkpoint bodies" proposed to control abscission timing form in response to nuclear basket defects and aphidicolin treatment, but not in the presence of chromatin bridges (Strohacker et al., eLife 2021).

    We appreciate the reviewer’s excellent suggestion. We have now performed the requested experiments and added a new figure showing that CAPN7 is also required to maintain the NoCut checkpoint when it is triggered by DNA bridges (new Figure 6A) or by replication stress (new Figure 6B).

    1. The current data suggest that Capn7 is a regulator of abscission timing, but in my opinion do not quite establish this, for two main reasons. First, abscission timing is not directly measured in this study. Time-lapse imaging would be required to rule out alternative interpretations of the data in figure 5. For example, a delay in an earlier cell cycle stage could in principle lead to a decrease in the overall fraction of midbody-stage cells. Second, the absence of the midbody is not necessarily a marker of complete abscission. Indeed, midbody disassembly is associated with the completion of abscission in unchallenged HeLa cells, but not in cells with chromatin bridges (Steigemann et al, Cell 2009). Midbodies remain a useful marker for pre-abscission cells, but the absence of midbodies should not be immediately interpreted as completion of abscission without further assays. Formally, a direct measurement of abscission timing would require imaging of the plasma membrane, for example using time-lapse phase-contrast microscopy (Fremont et al., 2016 Nat Comm). These limitations should be mentioned in the text.

    We note that midbody numbers are not our only measure of abscission delay/failure - we also measure the numbers of multinucleate cells and sum the two. Nevertheless, we understand the reviewer’s point and have therefore noted that we are using increased frequencies of cells with midbody connections and multiple nuclei as surrogate markers for abscission defects and NoCut-induced abscission delays (page 7, lines 13-14 and line 17).

    1. IST1 plays a role in nuclear envelope sealing by recruiting the co-factor Spastin (Vietri et al., Nature 2015), a known IST1 co-factor also confirmed in the previous interactome screen (Wenzel et al. 2022). CAPN7 could have a role in maintaining nuclear integrity upon the KD of Nup153 and Nup50 (Mackay et al. 2010) instead of/in addition to its proposed role in delaying abscission as part of the NoCut checkpoint at the midbody. I don't think the authors can differentiate between these two possibilities, and it would be interesting to consider their possible implications on how the "NoCut" checkpoint is triggered.

    The reviewer again makes good points, and we agree that in addition to participating in abscission, CAPN7 may be involved in closure of the nuclear envelope and that nuclear envelope closure may, in turn, be linked to satisfaction of the NoCut checkpoint. This involvement would nicely explain our observations that both SPAST and CAPN7 participate in both NoCut and abscission. We are in an unusual situation, however, because other colleagues in our field have told us in private communications that they observe that CAPN7 does, in fact, participate in nuclear envelope closure. We find that observation interesting and exciting but it is their discovery, not ours, and we have therefore refrained from doing analogous experiments ourselves. As a compromise, we have added the following text to the penultimate section of our paper (page 8, lines 34-35 through page 9, lines 1-11):

    “Our discovery that both CAPN7 and SPAST participate in the competing processes of cytokinetic abscission and NoCut delay of abscission may appear counterintuitive, but we envision that the MIT proteins could participate in both processes if they change substrate specificities or activities when participating in NoCut vs. abscission; for example, via different sites of action, post-translational modifications, and/or binding partners. We note that, in addition to its well documented function in clearing spindle microtubules to allow efficient abscission (Yang et al., 2008), SPAST is also required for ESCRT-dependent closure of the nuclear envelope (NE) (Vietri et al., 2015). The relationship between NE closure and NoCut signaling is not yet well understood, and it is therefore conceivable that nuclear membrane integrity is required to allow mitotic errors to sustain NoCut signaling. It will therefore be of interest to determine whether or not CAPN7, in addition to its midbody abscission functions, also participates in nuclear envelope closure and, if so, whether that activity is connected to its NoCut functions.”

    We think that this additional text explains what we (and the reviewer) consider to be an attractive model, but leaves open the question of CAPN7 involvement in nuclear envelope closure to be resolved by our colleagues.

    1. Figure 5 should include images of representative cells, highlighting midbody-positive and multinucleated cells. Without images, it is not possible to evaluate the quality of these data.

    We appreciate this suggestion and have now added images showing midbody-positive and multinucleated cells from the quantified datasets to allow assessment of our data quality (new Figures 5B and 5D).

  3. eLife

    eLife assessment

    The manuscript by Paine and coworkers provides a fundamental improvement on how the enzymatic activity of CALPAIN7 (a Cys protease) influences cytokinesis mediated by the ESCRT (endosomal sorting complexes required for transport) pathway. The authors provide a convincing molecular and cellular basis for one of the several key steps involved in membrane fission during the separation of dividing eukaryotic cells. These findings should be of interest to a wide scientific audience including biochemists, structural biologists, and cell biologists.

  4. eLife

    Reviewer #1 (Public Review):

    The current manuscript builds on a previous publication from the same group(s) (https://doi.org/10.7554/eLife.77779) that identified several new interacting partners of the ESCRT pathway. The authors show via fluorescence polarization anisotropy (FPA) and NMR spectroscopy that the microtubule-interacting and trafficking (MIT) domains of CALPAIN7 bind to the IST1 subunit of the ESCRT-III complex. The authors used a powerful combination of biochemical, structural, and cell biological tools. The experiments are designed well and are performed to a high standard. The vast majority of the conclusions are supported by the data.

    The authors report the X-ray crystal structure of the MIT-IST complex show the exact residues involved in the interaction and the mode of binding. They validate their findings and put them in a biological context by introducing mutations into the key residues in the MIT domain of CALPAIN7 and IST domain of ESCRT-III that can disrupt and restore the molecular interactions using in vitro biochemical assays and in vivo immunofluorescence microscopy imaging.

    The writing is exceptional. I could not spot a single typographical or grammatical error. The manuscript is easy to read. It is concise and is nicely supported by high-quality figures. It was a pleasure to read.

    The authors provide a great amount of detailed information about the experiments performed and have deposited the majority (if not all) of the plasmids used in the study in a public depository. They should be commended for this decision.

  5. eLife

    Reviewer #2 (Public Review):

    Please note that I am not a structural biologist and cannot critically evaluate the details of figures 1 to 3; my review focuses on the cell biology experiments in figures 4 and 5.

    Paine and colleagues investigated structural requirements for the interaction between the ESCRT-III subunit IST1 and the protease CAPN7. This is a continuation of previous work by the same group (Wenzel et al., eLife 2022), which showed that Capn7 is recruited to the midbody by Ist1 and that Capn7 promotes both normal abscission and NoCut abscission checkpoint function. In this article, the structural determinants of the Ist1-Capn7 interaction are characterised in more detail, focusing on the structure of Capn7 MIT domains and their binding to Ist1. Notably, point mutations in Capn7 MIT domains known to mediate binding to Ist1 and midbody recruitment are shown here to be required for abscission functions, as expected from the authors' previous paper. Furthermore, the report shows that a Capn7 point mutant lacking proteolytic activity behaves as a loss-of-function in abscission assays, despite showing normal midbody localisation. These are important results that will help in future studies to understand how the Capn7 protease regulates abscission mechanistically.

    The report is clearly written and the results support the main conclusions. Some technical limitations and alternative interpretations of the data should be discussed in the text, as outlined below.

    1. It is not always clearly stated how the results presented in this report relate to those in the Wenzel paper. For example, the finding that Ist1 recruits Capn7 to midbodies (p. 6 and figure 4) was first shown in the Wenzel paper. The novelty here is not that Capn7 MIT mutants fail to localise to midbodies, but that they phenocopy the previously described knockdown of Capn7, failing to support normal abscission and NoCut function (fig. 5). This supports and extends the findings of Wenzel et al. It is important to make this explicit and explain the conceptual advances shown here more clearly.
    2. The NoCut checkpoint can be triggered by chromatin bridges, DNA replication stress, and nuclear basket defects, but only basket defects are tested here. Therefore, it is not clear if NoCut is still functional in Capn7-defective cells after replication stress and/or with chromatin bridges. Ideally, this should be tested experimentally, or alternatively discussed in the text, especially since the molecular details of how NoCut is engaged under different conditions remain unclear. For example, "abscission checkpoint bodies" proposed to control abscission timing form in response to nuclear basket defects and aphidicolin treatment, but not in the presence of chromatin bridges (Strohacker et al., eLife 2021).
    3. The current data suggest that Capn7 is a regulator of abscission timing, but in my opinion do not quite establish this, for two main reasons. First, abscission timing is not directly measured in this study. Time-lapse imaging would be required to rule out alternative interpretations of the data in figure 5. For example, a delay in an earlier cell cycle stage could in principle lead to a decrease in the overall fraction of midbody-stage cells. Second, the absence of the midbody is not necessarily a marker of complete abscission. Indeed, midbody disassembly is associated with the completion of abscission in unchallenged HeLa cells, but not in cells with chromatin bridges (Steigemann et al, Cell 2009). Midbodies remain a useful marker for pre-abscission cells, but the absence of midbodies should not be immediately interpreted as completion of abscission without further assays. Formally, a direct measurement of abscission timing would require imaging of the plasma membrane, for example using time-lapse phase-contrast microscopy (Fremont et al., 2016 Nat Comm). These limitations should be mentioned in the text.
    4. IST1 plays a role in nuclear envelope sealing by recruiting the co-factor Spastin (Vietri et al., Nature 2015), a known IST1 co-factor also confirmed in the previous interactome screen (Wenzel et al. 2022). CAPN7 could have a role in maintaining nuclear integrity upon the KD of Nup153 and Nup50 (Mackay et al. 2010) instead of/in addition to its proposed role in delaying abscission as part of the NoCut checkpoint at the midbody. I don't think the authors can differentiate between these two possibilities, and it would be interesting to consider their possible implications on how the "NoCut" checkpoint is triggered.
    5. Figure 5 should include images of representative cells, highlighting midbody-positive and multinucleated cells. Without images, it is not possible to evaluate the quality of these data.

  6. eLife

    Reviewer #3 (Public Review):

    A central step in cell division is the formation of midbody abscission that separates two daughter cells at the end of cytokinesis. The ESCRT, endosomal sorting complexes required for transport, plays a critical role in this process. Specifically, the ESCRT-III proteins are actively recruited throughout the cell at the membrane fission sites, and their oligomerization into filaments is necessary to constrict the cell membranes to the fission point. Fundamental structural elements in ESCRT-III interactome are the so-called MIT-interacting motifs (MIMs) located at the protein's C terminal portion. Recently, Sundquist and co-workers (eLife 2022) identified several cofactors interacting with ESCRT-III subunits directly implicated in abscission. Among those cofactors, they identified Calpain-7, a cysteine protease whose function is still unclear. Calpain-7 comprises two MIT domains that target ESCRT-II subunit IST1. Here, the authors use structural methods and cell assays to characterize the interactions between Calpain-7 and IST1. For the structural studies, they constructed a minimalistic system in which MT1 and MT2 domains of Calpain-7 interact with the two MIMs localized in the IST1 construct. The truncated constructs interact with high affinity, recapitulating the strength of interaction expected for the full-length constructs in the cell. Using fluorescence polarization anisotropy binding isotherms, these researchers obtained solid binding data, showing a dissociation constant of 0.09 uM for the construct containing both MIMs, ~2 uM for the second MIM domain, and 100 uM for the first MIM. These data suggest a synergistic binding mechanism between the two MIM domains. The authors expressed and purified these constructs in recombinant systems and obtained purified isotopically labeled proteins to study by NMR. To characterize the binding by NMR, the authors studied the IST1 constructs with the two MIMs in the absence and presence of Calpain-7. The IST1 construct displays a well-resolved NMR fingerprint, with most resonances assigned to specific residues. Upon addition of the Calpain-7 construct, the resonances of the residues involved in the binding either broaden beyond detection or shift significantly, which supports the fluorescence binding studies. Given the high affinity, these authors were able to crystallize these complexes and identify the binding interfaces that parallel the solution NMR studies. Mutational studies confirm the hot spots for the interactions, and the authors concluded that the MIT:MIM binding interface is responsible for the association of the full-length constructs of Calpain-7 and IST1 in the cell. Using localization experiments, the authors concluded that IST1 is responsible for recruiting Calpain-7 to midbodies, and the presence of both MIT domains of Calpain-7 and MIM domains is required for localization. Taken together, the biophysical characterization of these complexes and the cell assays led the authors to conclude that IST1 binding to Calpain-7 is necessary for its role in abscission and Nocut checkpoint maintenance.
    In my opinion, the research is well executed and also supported by their previous finding (see Sundquist 2022 eLife). The paper is succinct and well-written.

  7. Version published to 10.1101/2022.10.18.512775v2 on bioRxiv
  8. Version published to 10.1101/2022.10.18.512775v1 on bioRxiv