Dynein and dynactin move long-range but are delivered separately to the axon tip

  1. Alexander D. Fellows
  2. Michaela Bruntraeger
  3. Thomas Burgold
  4. Andrew R. Bassett
  5. Andrew P. Carter

Reviewed by

Read the full article

Listed in

Log in to save this article

Abstract

Axonal transport is essential for neuronal survival. This is driven by microtubule motors including dynein, which transports cargo from the axon tip back to the cell body. This function requires its cofactor dynactin and regulators LIS1 and NDEL1. Due to difficulties imaging dynein at a single-molecule level, it is unclear how this motor and its regulators coordinate transport along the length of the axon. Here, we use a neuron-inducible human stem cell line (NGN2-OPTi-OX) to endogenously tag dynein components and visualize them at a near-single molecule regime. In the retrograde direction, we find that dynein and dynactin can move the entire length of the axon (>500 µm). Furthermore, LIS1 and NDEL1 also undergo long-distance movement, despite being mainly implicated with the initiation of dynein transport. Intriguingly, in the anterograde direction, dynein/LIS1 moves faster than dynactin/NDEL1, consistent with transport on different cargos. Therefore, neurons ensure efficient transport by holding dynein/dynactin on cargos over long distances but keeping them separate until required.

Article activity feed

  5. Version published to 10.1083/jcb.202309084
  6. 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

    Reviewer #1:

    1. If doable, image dynein and dynactin simultaneously in the Halo-DYNC1H1/DCTN4-SNAP iNeurons. Co-movement of dynein and dynactin towards the somatodendritic compartment and their separate movement in the anterograde direction along the axon would provide the most convincing evidence for the key claims of the manuscript.

    Please see the planned revision section for our response

    Reviewer #2:

    Major comment (requires additional experimentation)

    1. While the data presented do certainly suggest that dynein and Lis1 are transported anterogradely on separate vesicular cargoes from dynactin and Ndel1, the study would be much stronger if supported by dual imaging of dynein and dynactin to prove that these proteins do indeed move in association with separate vesicular populations. I would like to see dual-color kymograph traces showing that the proteins move independently. The authors should be able to accomplish this using their dual Halo-DYNC1H1/DCTN4-SNAP hESC line. To acquire and analyze this data might take several months, but it would greatly strengthen this paper. If the authors do this experiment, they may also be able to address the mechanism of reversal of anterograde cargoes which they speculate about in the Discussion, which would add even more interest and insight.

    Please see the planned revision section for our response

    Minor comments (addressable without additional experimentation)

    1. The authors deduce that 1-4 Halo fluorochromes corresponds to 1-2 dynein molecules. This implies that the cells are homozygous for the Halo tag, but I do not see this addressed explicitly. The authors should state explicitly whether the lines generated for their study are heterozygous or homozygous for the tag. If the cells are heterozygous, which would seem most likely, then they may be underestimating the number of dyneins per spot and should take this into account.

    We have added whether lines are homozygous or heterozygous to the manuscript. We also include a new Supplementary Figure panel (Fig S6) showing the genotyping data. In summary, all lines are homozygous except for PAFAH1B1-Halo (hESCs) which is heterozygous.

    1. Why are the moving spots lower in intensity than the NEM-treated static spots. It appears to suggest that they may be associated with different structures. This should be clarified and discussed.

    Our data suggest that the fast-moving spots have fewer dyneins than NEM treated static spots. We suggest this is because the fast-moving cargos are smaller than the average cargo and therefore have fewer dyneins on them. This is also supported by the smaller number of dyneins reported previously on endosomes as compared to the large lysosomes. We have clarified this in the discussion (page 7-8).

    1. The authors state in the Results that most of the dynein spots were diffusing, often along microtubules, but they do not visualize microtubules so how do they know this? They may need to remove the phrase "often along microtubules".

    This has been removed.

    1. At the end of the Introduction the authors state that their data "allow us to understand how the dynein machinery drives long-range transport in the axon". This is an overstatement. The "how" in this sentence is not addressed in this study.

    We have softened the sentence by adding the phrase “better understand”.

    1. The conclusion that dynein binds to cargos stably throughout their transport along the axon is based on measurements of the fastest moving cargoes but the authors do not provide data on the distribution of velocities for the entire population of retrograde cargoes. It is not valid to extrapolate the behavior of a small number of cargoes to the entire population. The average may be much slower than the fastest cargoes. Moreover, even for the fastest organelles the authors cannot say that the dynein is stably bound because they did not track single cargoes and thus do not know that the cargoes moved continuously in one single bout of movement for 500 µm; it is possible that the cargoes moved in multiple consecutive bouts interrupted by brief pauses and dynein motors may have exchanged between bouts.

    We have added a section to the discussion to highlight that other cargos may behave differently from the fastest ones (page 7). We have also clarified the assumptions that lead us to expect a slower arrival time of the first signal (page 5).

    1. The authors say that "it is clear that at least some dyneins remain on cargoes throughout their transport along the axon". As explained above, the data do not prove this so this statement should be removed.

    We have softened this sentence from “it is clear” to “our results suggest” and explained in more detail why we make this conclusion

    1. The authors note that most of the dynein spots were not moving processively and state that this is consistent with prior studies showing that only a subset of dynein is actively involved in transport. However, as they note elsewhere, dynein is both motor and cargo and most axonal dynein is transported at slow average velocities so maybe they should be more explicit about what they mean by "involved in transport".

    We have clarified we mean fast axonal transport and thank the reviewer for highlighting this point.

    1. When the authors note that most of the dynein in axons is transported in the slow component of axonal transport, they should also cite the work of Pfister and colleagues who were the first to show this (PMID 8824315 and 8552592).

    This was an omission on our part. The references have now been added.

    1. The authors propose that dynein and Lis1 are transported together but there were significantly fewer anterogradely transported Lis1 particles than dynein particles. This should be discussed.

    We have added more information to the discussion. Although we cannot rule out this effect being due to the heterozygous tagging of our LIS1 cell line, we do not witness the same decrease in events in the retrograde direction. Therefore, we believe there is a subset of anterogradely moving dynein lacking LIS1. As discussed in the manuscript, this subset may already be bound to dynactin and therefore not require LIS1.

    1. For the statistical analysis, the authors should provide p values in the legends for the comparisons that are judged to be "not significant". The authors should also be consistent in how they label differences that are not significant - they mark them as "ns" in Fig. 1, but in the other figures they do not, leaving some ambiguity about whether particular comparisons were not tested or were found to be not significant. For example, in Fig. 4C the average speed of the dynactin is about 0.5 µm/s greater than for the other proteins and the spread in the data suggest that this could be significant, but no significance is indicated on the plot, implying p>0.05. It is not clear how confident we can be that there is no difference.

    We have now included all p values in the figure legends and have removed the “ns” in Fig 1D. In our revised manuscript, only significant differences are highlighted in the figures.

    Reviewer #3:

    • if I look at the kymographs, trajectories appear rather complex, pausing, standing still, moving and everything mixed. The explanation of how actual trajectories are extracted and on what basis is very short, too short for me. I think the authors should expand this. Furthermore, I think it would be good if the authors would present, in their kymographs examples of the tracked (and also the not included) tracks. Maybe in supplementary info.

    The analysis of this data used the Trackmate Fiji plugin. This tracks spots frame to frame in a movie and then outputs the data of the tracks. No data was extracted from kymographs but they were used as a graphical illustration of the moving spots. To better explain our analysis pipeline, we have expanded our methods section and have added an example of a tracked movie (Video 15) as well as highlighted the tracked spots in one kymograph example (Figure 7S).

    • I found 'velocity' ill defined. I get the impression, judging from the number of points (compared to the other parameters) that the authors determine the average velocity of each individual trajectory. That is an important parameter (but should indeed be called 'trajectory averaged' velocity), but might not be the only one useful to learn from the data, where trajectories do not always appear to have constant speeds (pausing, etc.). Why do the authors not determine point-to-point velocities and plot histograms of those for all the trajectories (simply plot histograms of all the displacements between subsequent data points in trajectories)? This might provide great insight into the actual maximum velocity and the fraction of pausing or moving in opposite direction etc., providing much more molecular detail than currently extracted from the data.

    The reviewer is correct. We have measured the average velocity of the spots from the beginning of the track to the end. We have clarified this in the text. Furthermore, as stated above in the revision plan, we are currently doing the additional analysis and will include it in the final revision

    • I was a bit surprised to read that the authors have gone to the effort to create a dual-color labeled cell line, but did not do actual correlative two-color measurements (or at least show them). It would be so insightful to see dynein and dynactin move separately in the anterograde direction.

    Please see the planned revision section for our response.

  7. 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

    Fellows and coauthors present a signle-molecule study toward dynein regulation in axons. They observe that dynein in vivo makes very long runs and that regulators LIS1 and NDEL1 cotransport with dynein all the (retrograde way). Remarkably, different components of the dynein complex appear to be transported in different ways/velocities in the antorograde direction. Overall experiments are well conducted, I only have a couple of important questions regarding data analysis. Some aspects should be explained better, more steps should be shown and here and there I think the authors could, with minimal effort, obtain much more out of their data (see below). Nevertheless, I think this is an important study, on of the first single-molecule efforts to understand axonal transport in the cell (see below). Key findings are important for our understanding of dynein regulation.

    My concerns:

    • if I look at the kymographs, trajectories appear rather complex, pausing, standing still, moving and everything mixed. The explanation of how actual trajectories are extracted and on what basis is very short, too short for me. I think the authors should expand this. Furthermore, I think it would be good if the authors would present, in their kymographs examples of the tracked (and also the not included) tracks. Maybe in supplementary info.
    • I found 'velocity' ill defined. I get the impression, judging from the number of points (compared to the other parameters) that the authors determine the average velocity of each individual trajectory. That is an important parameter (but should indeed be called 'trajectory averaged' velocity), but might not be the only one useful to learn from the data, where trajectories do not always appear to have constant speeds (pausing, etc.). Why do the authors not determine point-to-point velocities and plot histograms of those for all the trajectories (simply plot histograms of all the displacements between subsequent data points in trajectories)? This might provide great insight into the actual maximum velocity and the fraction of pausing or moving in opposite direction etc., providing much more molecular detail than currently extracted from the data.
    • I was a bit surprised to read that the authors have gone to the effort to create a dual-color labeled cell line, but did not do actual correlative two-color measurements (or at least show them). It would be so insightful to see dynein and dynactin move separately in the anterograde direction.

    Referee Cross-Commenting

    I think we agree on the key points:

    • in principle, great study
    • quantification / tracking could go a bit further and should be explained better
    • manuscript / conclusions would be strengthened substantially if the authors could do some 2-color experiments to correlated dynein / dynactin movements in anterograde vs retrograde direction.

    Significance

    I think this is an important and exciting manuscript. As an in vivo single-molecule biophysicist with great interest in intracellular transport, I have been astonished in the lack of people trying to take single-molecule data on the motor involved, in particular neurons. I believe this is the only way to find out how transport actually works and what role motors play. Mutants is not enough, bulk data is not enough, in vitro is not enough. This is what the field needs (and many in the field do not seem to be aware of this...). Great that Fellows and coauthors took on this task and show some really exciting data. I am not an expert on their stem-cell labeling approach so cannot judge on that. The imaging seems to be done well. As discussed above, I think there might be much more in the data than the authors now get out, so I would encourage them to do some additional analysis. But overall, this effort is important and I think the conclusions will stand and provide important new insights in dynein regulation in the cell.

  8. 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

    Summary - The authors use a CRISPR knock-in gene editing strategy to label endogenous dynein, dynactin (p62 or Arp11) and dynein regulators (Ndel1 and Lis1) with Halo or SNAP tags. They do this in human iPSC and ESC cell lines engineered to express doxycycline-inducible NGN2 cloned into a "safe harbor" site of the genome. They induce the cells to differentiate into iNeurons using doxycycline and image the tagged proteins in axons with single molecule sensitivity using HILO illumination. The paper is clearly written, the description of the methods is thorough, and the data and figures (including the videos) are of good quality. The use of gene editing to knock the tags into the endogenous gene loci is a superior strategy to classic overexpression strategies. The authors also make effective use of microfluidic chambers to ensure the axons are uniformly orientated and coaligned over a distance of 500µm.

    Major comment (requires additional experimentation)

    1. While the data presented do certainly suggest that dynein and Lis1 are transported anterogradely on separate vesicular cargoes from dynactin and Ndel1, the study would be much stronger if supported by dual imaging of dynein and dynactin to prove that these proteins do indeed move in association with separate vesicular populations. I would like to see dual-color kymograph traces showing that the proteins move independently. The authors should be able to accomplish this using their dual Halo-DYNC1H1/DCTN4-SNAP hESC line. To acquire and analyze this data might take several months, but it would greatly strengthen this paper. If the authors do this experiment, they may also be able to address the mechanism of reversal of anterograde cargoes which they speculate about in the Discussion, which would add even more interest and insight.

    Minor comments (addressable without additional experimentation)

    1. The authors deduce that 1-4 Halo fluorochromes corresponds to 1-2 dynein molecules. This implies that the cells are homozygous for the Halo tag, but I do not see this addressed explicitly. The authors should state explicitly whether the lines generated for their study are heterozygous or homozygous for the tag. If the cells are heterozygous, which would seem most likely, then they may be underestimating the number of dyneins per spot and should take this into account.
    2. Why are the moving spots lower in intensity than the NEM-treated static spots. It appears to suggest that they may be associated with different structures. This should be clarified and discussed.
    3. The authors state in the Results that most of the dynein spots were diffusing, often along microtubules, but they do not visualize microtubules so how do they know this? They may need to remove the phrase "often along microtubules".
    4. At the end of the Introduction the authors state that their data "allow us to understand how the dynein machinery drives long-range transport in the axon". This is an overstatement. The "how" in this sentence is not addressed in this study.
    5. The conclusion that dynein binds to cargos stably throughout their transport along the axon is based on measurements of the fastest moving cargoes but the authors do not provide data on the distribution of velocities for the entire population of retrograde cargoes. It is not valid to extrapolate the behavior of a small number of cargoes to the entire population. The average may be much slower than the fastest cargoes. Moreover, even for the fastest organelles the authors cannot say that the dynein is stably bound because they did not track single cargoes and thus do not know that the cargoes moved continuously in one single bout of movement for 500 µm; it is possible that the cargoes moved in multiple consecutive bouts interrupted by brief pauses and dynein motors may have exchanged between bouts.
    6. The authors say that "it is clear that at least some dyneins remain on cargoes throughout their transport along the axon". As explained above, the data do not prove this so this statement should be removed.
    7. The authors note that most of the dynein spots were not moving processively and state that this is consistent with prior studies showing that only a subset of dynein is actively involved in transport. However, as they note elsewhere, dynein is both motor and cargo and most axonal dynein is transported at slow average velocities so maybe they should be more explicit about what they mean by "involved in transport".
    8. When the authors note that most of the dynein in axons is transported in the slow component of axonal transport, they should also cite the work of Pfister and colleagues who were the first to show this (PMID 8824315 and 8552592).
    9. The authors propose that dynein and Lis1 are transported together but there were significantly fewer anterogradely transported Lis1 particles than dynein particles. This should be discussed.
    10. For the statistical analysis, the authors should provide p values in the legends for the comparisons that are judged to be "not significant". The authors should also be consistent in how they label differences that are not significant - they mark them as "ns" in Fig. 1, but in the other figures they do not, leaving some ambiguity about whether particular comparisons were not tested or were found to be not significant. For example, in Fig. 4C the average speed of the dynactin is about 0.5 µm/s greater than for the other proteins and the spread in the data suggest that this could be significant, but no significance is indicated on the plot, implying p>0.05. It is not clear how confident we can be that there is no difference.

    Referee Cross-Commenting

    There seems to be agreement among all three reviewers that the authors should perform dual imaging of dynein and dynactin to prove that these proteins do indeed move together in the retrograde direction but separately in the anterograde direction. This would strengthen the study greatly.

    Significance

    General assessment - There are now multiple papers that have analyzed axonal transport of cargoes in iPSC-derived neurons, but this one appears to be the first to do it by tagging dynein motors and with single-molecule sensitivity. The principal conclusions are (1) that dynein is capable of long-range movement in axons and (2) that dynein moves dynein/Lis1 complexes are transported anterogradely in association with distinct cargoes from dynactin/Ndel1 complexes. The former is a modest conclusion and is entirely expected so not very impactful, but the latter is interesting and novel. The difference between the average velocities for the four proteins in the anterograde and retrograde directions is striking. All four move at similar velocities in the retrograde direction but in the anterograde direction, dynein and Lis1 move significantly faster than dynactin and Ndel1. Given these data, it is reasonable to infer that these proteins are being transported in two separate sets of cargoes. As the authors note in their Discussion, this is important because it could provide a mechanism for transporting dynein components anterogradely in a less active form that could then be activated when the components come together in the distal axon. However, I feel that one critical experiment is missing, which is to perform dual labeling of anterogradely transported dynein and dynactin in the same cells (see major comment). Without this experiment, the evidence is indirect.

    Audience - If confirmed by the dual labeling experiment, the authors' conclusions would represent a conceptual and mechanistic insight into the mechanism of bidirectional transport in axons that would be of broad interest to neuronal cell biologists studying neuronal trafficking.

    Expertise - This reviewer has expertise in the neuronal cytoskeleton, live imaging and axonal transport and has some experience working with iPSC-derived neurons.

  9. 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

    To image dynein in the axon at a single-molecule level, Fellows et al. used neuron-inducible human stem cell lines to Halo/SNAP tag endogenous dynein components by gene editing, and visualized fluorescently labeled protein molecules in differentiated neurons in microfluidic chambers by HILO microscopy-based live imaging. Using those cutting edge technologies, the authors demonstrate that in the axon, not only dynein and dynactin but also the dynein regulators LIS1 and NDEL1 can move long distance retrogradely towards the somatodendritic compartment. They also show that dynein /LIS1 move faster than dynactin/NDEL1 in the anterograde direction, suggesting that they are delivered separately to the distal end of the axon. The approach to study subcellular motility of endogenous dynein/dynactin is creative, the data are solid. I would like to suggest one experiment to support more strongly the authors' conclusions:
    If doable, image dynein and dynactin simultaneously in the Halo-DYNC1H1/DCTN4-SNAP iNeurons. Comovement of dynein and dynactin towards the somatodendritic compartment and their separate movement in the anterograde direction along the axon would provide the most convincing evidence for the key claims of the manuscript.

    Referee Cross-Commenting

    I agree with Reviewer 2 that the authors should clarify whether the knockin lines for dynein are homozygous. I also agree with both Reviewers 2 and 3 that the authors should do more analysis of the kymographs to obtain more information.

    Significance

    This is an elegant study on dynein motility and transport in vivo. The experimental approaches and findings presented in this manuscript are very valuable contributions to the field of dynein/dynactin and axonal transport. The results showing that dynein/dynactin can move long-range retrogradely in the axon are in good agreement with previous findings that dynein-driven cargo transport is highly processive, and the data suggesting that dynein and dynactin/NDEL1 are trafficked separately to the distal tip of the axon provide new insights into the regulatory mechanisms for the subcellular distribution and activity of molecular motors. Together these findings provide conceptual advances for understanding axonal transport. They will be of great interest to not only scientists in the field of intracellular transport but also those in cellular neurobiology.

  10. PREreview

    This Zenodo record is a permanently preserved version of a PREreview. You can view the complete PREreview at https://prereview.org/reviews/8362691.

    This review reflects comments and contributions from Kamaldeep Singh, Prithviraj Rajebhosale, Luciana Gallo, Ryan Cubero & Femi Arogundade. Review synthesized by Kamaldeep Singh.

    In this study, Fellows et al. investigated the motility of endogenously-tagged dynein motors and its regulators along the length of the axons using live imaging of neuron-inducible human-stem cell lines (iNeurons) as a model system. Using highly inclined and laminated optical sheet (HILO) imaging of iNeurons, they show that dynein and dynactin are transported at different speeds to the distal tip of the axon. Further, use of SNAP and Halo-tag conjugated with highly stable fluorophores also allowed them to show that single molecules of dynein and dynactin can traverse the entire length of the axon (>500 um). In summary, this study has contributed in advancing the cell biological understanding of dynein and its regulators in mammalian axons.

    We believe that the following positive aspects make the findings of the study strong and convincing:

    1. Use of neuron-inducible human stem cell-derived iNeurons as a model system is a significant advantage as it provides a better understanding of mammalian cell biology for studying axonal transport compared to traditional cancer cell lines.

    2. Use of microfluidic devices to separate axons from the somatodendritic compartments allows a convincing and clear demonstration of retrograde and anterograde transport in the axons and dynamics thereof.

    3. Use of CRISPR to endogenously tag the dynein heavy chain and the ARP11 subunit of dynactin with a SNAP and HaloTag allowed the authors to selectively label the motos with high spatio-temporal resolution. This not only aided in imaging dynein/dynactin at a near single-molecule level but also allowed them to figure that single molecules of dynein and dynactin can traverse the entire length of the axon (>500um).

    4. The photobleaching analysis on dynein spots in iNeurons provides valuable insights into the possible number of dynein molecules per cargo (which also agrees well with recent measurements of dynein number on endosomes in literature) and supports the claim of detecting single molecules under the given experimental conditions.

    We also noted several points regarding the study and the manuscript (as major or minor comments) which if addressed, could possibly make the study better:

    Major comments:

    • The authors observed mostly 1 and 2 step bleaching for dynein (Figure 2). Here, while they commented on the two step events being predominant and representative of the cytoplasmic dynein dimer, an explanation for the single bleaching events is lacking. Is it possible that these could be Halo-tagged dynein molecules dimerized with endogenous, untagged dynein from monoallelic targeting by CRISPR? It would therefore be nice to clarify and show the data that validates their CRISPR knock-in efficiency.  

    • The authors mention: "We saw many distinct dynein spots in the axonal compartment, most of which were diffusing, often along microtubules (Fig S2A, Video 4)." How did the authors determine that these events were diffusive and that they occured along microtubules?

    Minor comments: 

    • Figure 3 is incorrectly labeled as Figure 2 in the legend. Kindly correct this.

    • "Thought" is misspelled as "thught" in the last paragraph of Introudction section.

    • Fig 3C is incorrectly addressed as "Fig 3B" under the results section "Dynein moves long range".

    • In the videos attached along with the manuscript: It would be nice if there could be additional markers - for instance arrows tracking the particles (just like box on the spot where photobleaching was performed) -  to help readers focus on the main point the authors are trying to make with respect to a given video.

    • While the microfluidic devices utilized in the study might be standard in the trafficking field, it would be nice if the authors could provide a detailed description such as the devices' exact dimensions and manufacturer/supplier details.

    Authors have already hinted towards many unanswered questions, possible experiments and also listed anticipated outcomes for many such questions. Here are a few suggestions for lines of investigation that one could undertake in the future research:

    • As it has already been emphasized by authors, LIS1 is an important regulator of dynein activation. Therefore, simultaneous imaging of LIS1 and dynein might allow authors to identify whether the pausing of dynein is due to exchange of LIS1 within the complexes. Similar studies conducted with dynein together with other dynein cofactors, activators, or even kinesins will be useful to address many unanswered questions in the field.

    • Given that the authors can compartmentalize the somatodendrites and axons, it would also be interesting to know whether supply of dynein at the distal tip is regulated in response to different stimuli for e.g. in response to perturbations in neuronal activity.

    We wish the authors best of luck for all of their future research endeavours!

    Competing interests

    The author declares that they have no competing interests.

  11. Version published to 10.1101/2023.07.03.547521v2 on bioRxiv
  12. Version published to 10.1101/2023.07.03.547521v1 on bioRxiv