A force-sensitive mutation reveals a spindle assembly checkpoint-independent role for dynein in anaphase progression

  1. David Salvador-Garcia
  2. Li Jin
  3. Andrew Hensley
  4. Mert Gölcük
  5. Emmanuel Gallaud
  6. Sami Chaaban
  7. Fillip Port
  8. Alessio Vagnoni
  9. Vicente José Planelles-Herrero
  10. Mark A. McClintock
  11. Emmanuel Derivery
  12. Andrew P. Carter
  13. Régis Giet
  14. Mert Gür
  15. Ahmet Yildiz
  16. Simon L. Bullock

Reviewed by

Read the full article

Listed in

Log in to save this article

Abstract

The cytoplasmic dynein-1 (dynein) motor organizes cells by shaping microtubule networks and moving a large variety of cargoes along them. However, dynein’s diverse roles complicate in vivo studies of its functions significantly. To address this issue, we have used gene editing to generate a series of missense mutations in Drosophila Dynein heavy chain (Dhc). We find that mutations associated with human neurological disease cause a range of defects in larval and adult flies, including impaired cargo trafficking in neurons. We also describe a novel mutation in the microtubule-binding domain (MTBD) of Dhc that, remarkably, causes metaphase arrest of mitotic spindles in the embryo but does not impair other dynein-dependent processes. We demonstrate that the mitotic arrest is independent of dynein’s well-established roles in silencing the spindle assembly checkpoint. In vitro reconstitution and optical trapping assays reveal that the mutation only impairs the performance of dynein under load. In silico all-atom molecular dynamics simulations show that this effect correlates with increased flexibility of the MTBD, as well as an altered orientation of the stalk domain, with respect to the microtubule. Collectively, our data point to a novel role of dynein in anaphase progression that depends on the motor operating in a specific load regime. More broadly, our work illustrates how cytoskeletal transport processes can be dissected in vivo by manipulating mechanical properties of motors.

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 manuscript Salvador-Garcia et al. examine how several mutations in the Dynein heavy chain (DHC) influence Dynein function in an in vitro reconstituted system and in Drosophila embryos. Most importantly, they identify a novel substitution mutation (S3372C, generated as a by-product of a targeted CRISPR mutagenesis screen) that leads to a novel phenotype specifically in early Drosophila embryos (metaphase arrest) and that only impairs DHC function under load in vitro. Most surprisingly, the metaphase arrest in embryos does not appear to be due to a failure to inactive the spindle-assembly-checkpoint (SAC), a known DHC function. This suggests that DHC has a hitherto unappreciated function in allowing spindles in the Drosophila embryo to progress from metaphase-to-anaphase.

    The manuscript is generally well written and conveys the major conclusions clearly and concisely. The data is generally of high quality, and largely supports the main conclusions, although there is one set of relatively straight-forward experiments that I think would be an important addition (see major comment #1, below).

    Major Comments

    1. The observation that the S3372C mutation causes a mitotic arrest that is not SAC dependent (i.e. it still largely occurs even in a Mad2 mutant background) is very surprising, and is the basis of the authors claim of a new, DHC-dependent, mechanism that allows embryo spindles to progress into anaphase. I think it would be important to assess whether the SAC components are still localising to the kinetochores in these S3372C, Mad2 double mutants (e.g. is Rod still recruited to high-levels in the double mutant?). If the SAC components are still being recruited to the spindle (suggesting that they are still detecting that the spindle is not ready to go into anaphase), is it worth considering that Mad2 may not be essential for SAC function in these embryos? I say this because I find it hard to imagine how any, presumably mechanical, failure at the kinetochore that leads to the improper metaphase/anaphase transition in the S3372C mutants, would signal to the rest of the spindle to not transition to anaphase if the SAC is truly inactivated. Do the authors think these embryos have a completely unrelated surveillance system that detects the S3372C-dependent error (whatever that is) and arrests the spindles specifically in embryos? Or is the error itself sufficient to cause a spindle-wide arrest, which seems improbable?
    2. I was surprised the authors made no attempt to quantify the level of over-accumulation of Dlic (Figure 6) or Rod (Figure 7) (and the lack of over-accumulation in other regions of the spindle). The images are convincing, so I don't doubt that this is the case, but I think some sort of quantification would be useful and I don't think it would be hard to come up with a way to do this (even just drawing a ROI around the approximate areas of interest). It would also be interesting to know whether other proteins like Spc25 (Figure 6) and Cdc20 (Figure S6) are recruited to normal levels at kinetochores.

    Minor Comments

    1. In the Discussion the authors state: "Our discovery of a missense mutation that strongly affects nuclear divisions in the embryo without disrupting other dynein functions offers a unique tool to study the mitotic roles of the motor". This should be reworded, as it suggests that the mutation effects all mitotic functions of DHC, which is clearly not the case (and also applies only to the embryo).
    2. I think it worth more explicitly stating that there is no evidence that the defects the S3372C mutation lead to in the in vitro reconstituted system are the cause of the in vivo defects observed in the embryo. The authors are careful not to directly claim this, and I agree with their assertion that this is the most "parsimonious explanation" for their data, but I'm sure they would agree that this is far from proven, and it might be worth emphasising this point a little more.

    Significance

    This is a well conducted study that significantly extends the author's previous work on how mutations in DHC (initially indentified in human patients) effect DHC function (Hoang et al., PNAS, 2017). The paper reports the striking central finding that the S3372C mutation produces a very unusual mitotic arrest phenotype specifically in Drosophila embryos, and the authors link this to the also striking finding that this mutation only disrupts DHC function in vitro when DHC is working under load. As mentioned above, this link is not proven here, but this is a solid working hypothesis that is potentially of significant interest to those working on molecular motors and their role in fundamental cell biology and human disease.

    I am a cell biologist with expertise in the cytoskeleton, particularly during early Drosophila embryogenesis.

  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

    In the manuscript by Salvador-Garcia et al., the authors assess the physiological consequences of dynein mutations in flies and in vitro. In addition to characterizing the manner by which disease causing mutations affect fly development and some aspects of their cell physiology, the authors focus on sporadic mutations that arose during the course of generating their mutant fly lines. Of particular interest was a mutation in the dynein MTBD: S3722C. This mutation caused very interesting phenotypes in flies (e.g., infertility in females likely due to mitotic arrest) as well as in reconstituted motility assays (e.g., reduced stall force). The authors posit that the mitotic arrest phenotype is a consequence of a dynein's role in initiating anaphase onset, and that this role is distinct from its well established role in silencing the spindle assembly checkpoint.

    The paper is very well written, and the data are of high quality. Most of the claims - with one major exception (described below) - are well supported by the data. I have a few comments that might help to strengthen the conclusions.

    Major comment:

    1. In brief, I'm not convinced the mitotic arrest phenotype is not a consequence of impaired SAC silencing by the mutant dynein. The main tool the authors use to support their claim is a Mad2 mutant. They use this to determine if preventing SAC function (with the mutant Mad2) can override the ability of S3722C cells to progress into anaphase. The Mad2 mutant does in fact increase the proportion of cells exiting mitosis (from 0.8 to 14% of cells); however, the low number (14%) suggests that an inability to silence the SAC is not the reason the cells are not entering anaphase (i.e., it is SAC silencing-independent). My question is how penetrant the Mad2 mutant is? For example, how many cells with this Mad2 mutant would exit mitosis if the authors perturbed mitosis some other way (e.g., treatment with high concentrations of nocodazole)? If the number is still low (~14%), then this might be why the mutant can't rescue the mitotic exit phenotype for S3722C cells, and would challenge the following statements: "...it suggests that this can make, at best, a minor contribution to the mitotic arrest phenotype"; and "Remarkably, the MTBD mutation does not appear to block anaphase progression in embryos by preventing the well-characterized role of kinetochore-associated dynein in silencing the SAC, as the defect persists when the checkpoint is inactivated by mutation of Mad2. Collectively, these observations indicate that kinetochore dynein has a novel role in licensing the transition from metaphase to anaphase." Although previous studies might have assessed the penetrance of this mutant in other cell types, given the cell specificity of the mitotic arrest phenotype for S3722C (in embryonic cells, but not L3 neuroblasts), it will be important to provide such additional evidence in embryonic cells (e.g., nocodazole treatment of embryonic cells) to support these statements, especially in light of the bold conclusions and hypotheses they are making (e.g., "For example, the apparent variability in tension between sister kinetochores in S3372C embryos, which could reflect abnormal force generation by the mutant motor complex, might prevent APC/C activation through the complex series of signaling events that respond to chromosome biorientation."). Although it would be fascinating if the authors are correct that dynein provides another role in licensing anaphase onset, the well-established role for dynein in checkpoint silencing currently seems like the most parsimonious explanation.

    Minor comments:

    1. The S3722C mutant appears to accumulate to higher-than-normal levels at KTs and to some extent along the spindle MTs. In addition to the representative images, it would be helpful to see a quantitation of this phenomenon for WT, S3722C, and S3722C/+ along with statistics.
    2. Although the mean inter-KT distance was unchanged between WT and S3722C cells, the authors note that the deviation from the mean was higher. Could this simply be a consequence of more highly dynamic oscillations of KT pairs (similar to that seen with Hec1-S69D in DeLuca et al., JCB 2018)? More dynamic oscillations could potentially lead to more variable distances between KT pairs.
    3. It is interesting that the S3722C/Mad2-mutant cells are enriched in telophase (Fig. 7G). Does this not suggest another arrest point for these cells?
    4. The authors state: "Immunostaining revealed that whereas α1-tubulin was present throughout the spindle apparatus, α4-tubulin was enriched at the spindle poles (Figure S7A)." Although I agree the a4-tubulin appears somewhat enriched at the poles with respect to a1-tubulin, a quantitation (with statistics) would be needed to support this claim. That being said, I agree the isotype is unlikely to account for the S3722C phenotype.
    5. Trapping data show reduced stall force, yet increased stall time at low resistive forces for the mutant. This finding could potentially account for the reduced velocity of GFP-Rod noted in cells; however, I wonder if the authors noted altered velocity for dynein-driven bead movement under load in their trapping assays? This information would be useful to include in their manuscript.
    6. Is there a defocused spindle pole phenotype in the mutant cells? The cells in Video 1 and Fig. S6c appear to show as much, although other cells do not.
    7. The authors state: "This may reflect the relatively short length of Drosophila neurons making them less sensitive to partially impaired cargo transport." Could the extent of the phenotypes also be related to the lifespan of the flies? Do any of the diseases caused by these mutations have late-onset in patients? I wonder if a subtle defect in dynein behavior might not manifest for numerous years due to only minor changes in motility?

    Referees cross-commenting

    I wanted to reiterate my skepticism regarding the possibility that their data strongly support a SAC-independent role for dynein in the metaphase-to-anaphase transition (it seems Reviewer #3 might agree with me). I don't think it's impossible, but I'm not convinced they've made a very strong case for this model, which is noted in the title. The fact that proteins are accumulating at aligned kinetochores in the mutant cells (e.g., Rod and DLIC) in fact are consistent with a SAC silencing defect. Along these lines, I think reviewer #1's point regarding RZZ is a good one (that the mutant dynein is incapable of evicting RZZ specifically from aligned KTs), and should be tested prior to publication.

    My primary concern is that their conclusion is based entirely on the fact that the Mad2 mutant does not fully restore mitotic exit to the dynein mutant cells. Given the cell-specificity of their dynein phenotype (in embryonic cells, but not L3 neuroblasts), I think testing the penetrance of their Mad2 mutant in the embryonic cells would need to be assessed. In my review I suggested nocadazole, when I realized I meant to say reversine (oops!).

    Significance

    The manuscript by Salvador-Garcia et al. is a very interesting study dissecting the physiological consequences of dynein mutations in flies and in vitro. This study will be of high interest to those in the dynein/molecular motor field, as well as those that study mitosis and kinetochore function. One of the most interesting findings in the study is the identification of a mutation in dynein that specifically impacts its motility in conditions of high load. This mutant provides a novel tool to dissect load-dependent transport for dynein in other systems. Moreover, the study suggests a novel role for dynein in promoting anaphase onset; if the authors can provide additional support for this claim, the impact of this study would be greater.

    Field of expertise: molecular motors, kinetochore function

  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

    The authors start out by examining the cellular and organismal effects of 6 human disease-linked mutations, as well as the mouse legs-at-odd-angles (Loa) mutation, after introducing the mutations into the D. melanogaster dynein heavy chain (Dhc) by genome editing. This reveals an overall correlation between the severity of the effect on dynein motor activity in vitro (determined in a previous study) and the penetrance of the corresponding mutant phenotype in the fly, with a couple of interesting exceptions that illustrate the value of performing structure-function analysis of dynein in animal models. The authors then focus on an additional missense mutation in the Dhc microtubule binding domain, fortuitously generated by imprecise editing, that results in a striking phenotype. The S3372C mutation supports normal development, including normal axonal transport of mitochondria and asymmetric mitosis of larval neuroblasts, but female flies are infertile. Through elegant genetics, ectopic disulfide bond formation with a nearby residue is ruled out as the cause of the maternal effect. S3372C results in metaphase arrest of early embryonic divisions, characterized by over-accumulation of dynein light intermediate chain (Dlic) and the dynein recruitment factor Rod at kinetochores, as well as by reduced poleward streaming of Rod along spindle microtubules. Surprisingly, the S3372C-induced metaphase arrest cannot by bypassed by inhibiting the spindle assembly checkpoint, implying that dynein promotes the metaphase-to-anaphase transition not solely through its known function in spindle assembly checkpoint silencing. In vitro motility and optical trapping experiments show that the mutant motor performs normally in a load-free regime but exhibits reduced peak force production and excessive pausing under load. Furthermore, molecular dynamics simulations reveal how the mutation affects dynein's interaction with microtubules, including a change in the positioning of the stalk. The authors conclude that the S3372C mutation specifically perturbs high-load functions of dynein, explaining the selective phenotype observed in vivo.

    The experiments are technically on a very high level, the results are presented in a clear manner, and the conclusions are fully supported by the data.

    Minor suggestions (optional):

    • In the first part of the paper, where Dhc mutations associated with neurological disease are examined, the H3808P and F579/Loa mutations are shown to cause mis-accumulation of synaptic vesicles in axons. The authors may want to perform this assay for the K129I, R1557Q, and K3226T mutations, as this would strengthen the comparative analysis of in vitro versus in vivo effects, summarized in Figure 1C. For example, K129I has a more severe effect in vitro than the Loa mutation, but the Loa mutation has a more pronounced phenotype on the organismal level. Would this also be the case in a cell-based assay?
    • The observation that the metaphase arrest of S3372C mutant embryos cannot be alleviated by the checkpoint-defective Mad2 mutant is very intriguing, as is the observation that Dlic and the RZZ subunit Rod over-accumulate at/near kinetochores. As discussed by the authors, one possibility is that the arrest is a consequence of dynein's failure to disassemble the corona by stripping, but, surprisingly, in a manner unrelated to dynein's role in SAC silencing. In this regard, it is interesting to note that fly RZZ mutants do not undergo metaphase arrest in the early embryo (Williams and Goldberg, 1994; Défachelles et al., 2015), whereas knockdown of Spindly, which functions in dynein recruitment downstream of RZZ, does lead to arrest (see Figure 2 in Clemente et al., 2018; PMID 29615558). Taken together, this raises the possibility that it is the failure to remove RZZ (and other associated corona components) from kinetochores that inhibits anaphase onset in S3372C embryos. It would therefore be interesting to test whether the metaphase arrest in S3372C embryos is alleviated in RZZ mutants.

    Referees cross-commenting

    The Mad2 mutant the authors use is a P-element insertion that was described by Buffin et al 2007 as a null mutant with regards to SAC signaling (it also does not produce any detectable protein by Western blot; Figure 1b). Nevertheless, since the analysis in Buffin et al was restricted to larval brains, I agree with reviewer #2 that it remains to be formally demonstrated that this Mad2 mutant fully abolishes the SAC in the early embryo. Unfortunately, as far as I am aware, reversine does not work well in Drosophila. An alternative would be to combine Dhc(S3372C) with the other Mad2 mutant used by Buffin et al, which (besides not producing detectable protein) lacks the Mad1 binding domain and can therefore be expected to be a definitive checkpoint null in all tissues.

    Significance

    The cytoplasmic dynein 1 motor complex participates in a multitude of cellular processes that require microtubule minus-end-directed motility. Whereas in vitro reconstitution efforts have led to a detailed understanding of the motor's motile properties, the essential requirement of dynein for development has made it challenging to dissect how the motor contributes to specific aspects of intracellular organization and cell division in vivo. The need for a mechanistic understanding of how dynein motility is used for diverse functions in vivo is underscored by missense mutations in the motor subunit that cause human neurological disease. In this interesting and insightful study, Salvador-Garcia and colleagues characterize several missense mutations in dynein heavy chain (Dhc) using biochemical assays and genetic approaches is the fly, which reveals the distinct effects of disease-causing mutations in vivo and uncovers an unanticipated novel function of dynein in regulating mitotic progession.

    This beautifully executed study has important implications for dynein's role in mitosis, in particular its role at the kinetochore, and is of broad interest to cell biologists studying the cytoskeleton, as it demonstrates that examining motor mutants with altered mechanical properties in vivo can reveal specific motor functions.

  5. Version published to 10.1101/2023.08.03.551815v1 on bioRxiv