A live-cell marker to visualize the dynamics of stable microtubules throughout the cell cycle

Jansen et al. introduce StableMARK (Stable Microtubule-Associated Rigor-Kinesin), a live-cell marker to visualize stable microtubules. This live-cell marker enables the exploration of different MT subsets throughout the cell cycle to understand how they contribute to cellular organization and transport.

I am sorry that our answer on this occasion is not more positive, and I hope that this outcome will not dissuade you from submitting other manuscripts to us in the future. In this study the authors describe the development and initial characterization of a probe for live imaging of stable microtubules based on a rigor mutant of kinesin-1. Using this sensor they image stable microtubules in U2OS cells and report that while the microtubules themselves are stable they display dynamic behaviors such as undulation, looping, and sliding. The manuscript is well written and the work is of high quality. Similar probes, such as those based on the N-terminal domain of ensconsin, that image stable microtubules in live cells have been reported before, and these are not discussed here nor is there a comparison of the kinesin-1 sensor to these probes. Prior studies have also revealed that stable microtubules exhibit looping and sliding behaviors and the present study also does not provide a thorough and quantitative exploration of the dynamic behavior of stable microtubules using the new kinesin-1 based probe. Thus, regrettably, in its present form this work does not establish that this new sensor allows for interrogation of cell biological problems in ways previously impossible nor provide novel cell biological insights as a proof of principle, which are required for consideration as a JCB Tools paper. We would, however, be interested in seeing a more complete study in which these issues were addressed, if the authors wish to do so.
1. In Figure 1 the authors focus a lot of their attention on acetylation. They show that this marker is marking stable microtubules but not necessarily the acetylated ones. They compare the colocalization of Kin-1 and acetylated MTs on taxol-stabilized and hyper-acetylated microtubules. In the former their marker binds all MTs whereas without the taxol stabilization it binds a subset of MTs. Could this be due to their Kinesin rigor construct binding differently to taxol and non taxol stabilized MTs? Their interpretation is supported by live cell nocodazole treatment, but it would be reassuring to see a quantification of the amount of stable MTs before and after washout with and without their marker. A small point but the dashed lines in fig 1b and S1b showing where the line profiles in s1c were taken from are black and very difficult to see, perhaps make them more obvious to the eye? 2. In Figure 2, the high curvature of stabilized microtubules has been observed by other groups before. Citation of this earlier work is encouraged (for example, see Friedman et al JCB 2010).
3. One major issue with the argument that the authors are trying to make is that it is not clear whether their rigor kinesin is actually stabilizing microtubules. Microtubule dynamics parameters need to be determined in control cells and cells expressing their rigor kinesin. Just reporting EB1 comet numbers is not enough -what are the microtubule dynamic parameters? These are highly feasible experiments, especially for this lab. I am sure that at higher concentration levels the rigor kinesin will affect microtubule dynamics. What is the expression range for which this is a passive marker? 4. In vitro work showed that de-tyrosination affects the binding of kinesin-1 more so than acetylation ( Kaul et al 2014 -also not cited) yet the authors seem to have focused on acetylation. Is it possible that this is a marker for de-tyrosinated microtubules? They show that at low levels the amount of acetylated tubulin is about the same with and without the rigor mutant. ( Figure 3A). It would be reassuring to see the same for de-tyr tubulin, and to see a quantification for this. If the marker responds more to tyrosination than acetylation they also may be missing a fraction of the stabilized microtubules.
5. The Rab6A transport data needs to be quantified to see whether transport speeds are changed by the presence of more possible roadblocks (ie rigor kinesin) on the MTs.( Figure 3G). The transport kinetics of other cargos should also be analyzed that use other motors than the two used for Rab6A transport (for example, a connection between lysosome transport and acetylated MTs has been previously established). The serum starvation experiments are nice, but all they show is that the marked MTs can be depolymerized by this assay (already known). In general, the comment that "our data demonstrate that in the subpopulation of low rigor-expressing cells, artifacts of the MT cytoskeleton are minimal" is not backed up by sufficient data. It might very well be the case, but they need to do more to prove this so that this can be a marker that can be used with confidence in the field.
6. Their investigation of the mechanism of stabilization is technically well-performed but I have reservations with their interpretation. Upon ablation they see ~50% of the MTs depolymerize from one end which they interpret as an aging effect, where stabilization starts at the +tip and is followed by stabilization along the lattice. There have been several papers on motor induced damage and long-range lattice effects, including the stabilization of the lattice by even substiochiometric amounts of kinesin (Andreu-Carbo et al, 2022, Triclin et al 2021Peet et al 2018). These need to be discussed here as they point to the possible perturbing role that their marker can actually have. If these microtubules also are healed more, they will also be more stable when ablated because of the incorporation of GTP tubulin (Aumeier et al. 2016;Vemu et al 2018).
7. When discussing laser ablation experiments and the asymmetry in depolymerization between the plus and minus ends, the authors strangely cite a paper from 2014. This is textbook work from Salmon and colleagues for example and should be cited.
8. The authors make the interesting observation that the rigor kinesin coated MTs do not have dynamic (EB3) tips. Based on this they speculate that therefore there is a "special tip structure" on stabilized microtubules that prevents both growth and shrinkage. Could it be that their rigor mutant interferes with normal EB3 binding? 9. What is the effect of this rigor mutant on microtubule dynamics in vitro? These are highly feasible experiments that should be performed. And does it affect EB binding? Mechanistic data on what the effect of this mutant is on microtubule dynamics and binding of some ubiquitous effectors is needed. This paper describes the use of a rigor mutant of kinesin-1 to identify stable microtubules (MTs) in living cells. The authors show convincing data that a 2X-mNeon-tagged version of the rigor kinein-1 marks stable, acetylated MTs in cells. They then use the rigor kinesin to follow stable MT dynamics in interphase and mitotic cells and show a number of behaviors of these MTs, many of which have been reported previously. Overall, the development of a live cell probe for stable MTs is an unmet need for cell biologists and has the potential to reveal novel behaviors of these MTs. However, the paper has some major issues (see below) that need to be addressed before the rigor kinesin-1 is characterized well enough to be a reliable live cell marker for stable MTs. Key issues: 1. The authors need to establish the basis for recognition of the stable MTs by the rigor knesin-2 sensor. The rigor mutant labels stable, acetylated MTs, yet they also show by manipulating acetylation that it is not the basis for the detection of the stable MTs. Without understanding the basis for the recognition of the stable MTs by the rigor mutant, MTs detected by the rigor kinesin are only operationally defined as stable. This feature of the paper has great potential to create confusion/controversy when drawing conclusions from results with the rigor kinesin obtained in different systems and by different investigators. There is abundant evidence that kinesin-1 exhibits elevated binding to MTs with detyrosinated tubulin (e.g., Liao G and Gundersen GG, JBC, 1996;Dunn S et al, JCS 2008;Cai D et al, PLOS Biol, 2009;Sirajuddin M. et al, NCB, 2014), and the rigor kinesin should be tested against this posttranslational modification. 2. The relationship between the level of rigor kinesin-1 expression and artifacts caused by its expression are incompletely documented. The authors conclude (and I am not sure I agree, for example EB1 comets and hence dynamic MTs are clearly affected in Fig. 3C), that at "low" expression there is little effect on MTs, whereas at "high" expression, stable MTs increase, dynamic MTs are reduced, and overall MT distribution is disrupted. "Low" and "high" expression can mean different things to different people. Consequently, for the method to be reproducible and widely useful, it is critical that the authors define the level of rigor kinesin-1 expression that begins to show the artifacts they observe. For example, a comparison of the level of the rigor kinesin-1 to the endogenous level of kinesin-1 (or tubulin) would be helpful. Given the cell-cell variability with transient expression, it might be easier to make this comparison with stable cell lines with low and high expression. 3. There is very little new about stable MTs that is reported with the rigor kinesin leading me to wonder whether it will be all that useful. Part of this reflects the fact that the authors repeat previous findings that they are apparently unaware of. For example, previous studies have shown that stable MTs are curly (Piperno G et al, JCB, 1987 and others), do not grow for extended periods (Schulze E and Kirschner M, JCB, 1986;Webster D et al., PNAS, 1987;Infante AS et al, JCS, 2000;Palazzo AF et al, NCB, 2001), and depolymerize slowly when severed (Walker RA et al, JCB, 1989;Infante AS et al, JCS, 2000). The authors also seem unaware of all the molecular studies implicating numerous factors (Rho, CLASPs, ELKs, mDia, INF2) in stable MT formation. These studies should be acknowledged and additional studies should be considered to show the utility of the rigor kinesin. Additional comments: 4. The experiments showing that acetylated tubulin accumulates in MTs resistant to nocodazole and increases in taxol treated MTs shown in Fig 1F-G are not new or news. The data should be moved to the supplement. 5. The description of the behaviors of rigor kinesin decorated MTs are imprecise and in some cases over interpreted. For example, it is unclear how the authors draw the conclusion that the MT shown in Fig. 2C is sliding as it could also be treadmilling or trapped in actin flow. The MT in Suppl Fig2C might be moving rather than depolymerizing. And, I do not see active looping in fig. 2E as described. 6. I observe a substantial change in the overall distribution of MTs cause by "high" rigor kinesin expression in Fig. 3A. This should be mentioned along with the other alterations cause by expression of the rigor kinesin. 7. The comparison in Fig3B between rigor kinesin expression and acetylated tubulin levels needs to include the acetylation levels in non-expressing cells. Otherwise, their claim that expression of the rigor kinesin does not affect acetylated tubulin levels is not valid. Additionally, normalizing the rigor kinesin levels to the total tubulin levels makes no sense to me -the graph should simply show rigor kinesin expression. 8. The authors should plot EB1 comets per cell rather than EB1 comets/um to capture the clear reduction in the numbers of comets in both "low" and "high" expressing cells shown in Fig. 3C. 9. The data that Rab6A transport is unaffected needs quantification beyond the anecdotal example shown in Fig. 3G-I. 10. Many have noted that the turnover of spindle MTs (excluding mid-zone MTs, which are well known to be more stable), is an order of magnitude faster than interphase MTs. The rigor kinesin seems to decorate a subset of spindle MTs raising the question whether it is really detecting stable MTs in all cases. This result also illustrates the uncertainty with what the rigor kinesin is actually detecting.
Reviewer #3 (Comments to the Authors (Required)): Jansen et al. describe a new live-cell sensor that specifically detects stable microtubules. This is an important advance for cell biology, as so far, labelling of distinct microtubule species in cells was restricted to the use of antibodies, i.e. fixed cells. Being able to detect specific microtubule subtypes in living cells is a huge advance that will help the community to substantially advance the understanding of roles of microtubule specialization in living cells, and to image how those microtubules behave. The authors already provide an exciting glimpse on those opportunities by showing how they new live-cell marker allows them to follow the faith of stable microtubules throughout cell division. The presence of stable microtubules within a dynamic microtubule network as it is found in undifferentiated interphase cells has for many decades intrigued the scientific community. Because these microtubules are the only ones in cells to accumulate posttranslational modifications such as acetylation and detyrosination, it has for long been disputed whether these modifications are the cause or the consequence of stabilization of those microtubules. Recent work showing that acetylation decreases the flexural rigidity of microtubules, thereby making them more resilient to repeated mechanical stress suggested that acetylation could cause at least longevity of microtubules in cells. However, whether the observed stable microtubules in cells are solely stable because of their posttranslational modifications has remained an open question. The current work now provides an innovative approach to study stable microtubules in cells. The authors found that a rigor mutant of the motor protein Kif5b, when expressed at low levels in cells, specifically binds to stable microtubules, which are typically acetylated (tubulin acetylation at K40) in cells. To test whether the species of microtubules that their sensor detects could simply be the acetylated microtubules, the authors performed a series of experiments with drugs that either stabilize microtubules, or prevent their polymerization, or boost tubulin acetylation. Strikingly, they find that boosting acetylation does not increase microtubule stability in cells, and accordingly their sensor does not label more microtubules in cells with boosted acetylation. However, it does so in cells treated with the microtubule stabilizer taxol, indicating that the nature of stable microtubules -detected with the new sensor -is not restricted to their posttranslational acetylation. After showing that their new sensor can be used as a live-cell marker for stable microtubules, the authors demonstrate by different complementary experiments that the binding of (low doses) of their new sensor does not perturb overall microtubule behavior in cells, including posttranslational modifications, microtubule dynamics, motility of kinesin motors, or intracellular architecture. They go on demonstrating that the microtubules that are specifically labelled with their sensor are indeed stable: they demonstrate that they do not depolymerize after laser cutting, which typically would happen for dynamic microtubules. They further demonstrate that the new sensor is not exclusively binding stable microtubules, but binds other microtubules in cells much more transiently, suggesting a mechanism that increases its binding affinity specifically to the stable microtubules. Finally, the authors demonstrate how the new sensor can be used to monitor the faith and dynamics of stable microtubules in dividing cells. These impressive movies illustrate the potential of their new tool to obtain unpreceded insight into the behavior of specific microtubule subtypes in cells. The manuscript is carefully written, figures are of high quality and most experiments have been performed thoroughly with appropriate controls. Nonetheless, several points need to be address before the paper can be considered for publication.
Major points: 1) In Fig. 3A the authors show that low expression levels of their sensor do not increase tubulin acetylation in cells. However, as a control example, they show a cell with highly bundled microtubules around the nucleus (row 1), which is not the case of the cell expressing the sensor (row 2). One could speculate that the acetylation levels in the control are higher due to the microtubule bundling, which is absent from the cell in row 2 -and thus conclude that low expression levels of the sensor lead to levels of acetylation that can only be attained in control cells upon microtubule bundling. To show that this is not the case, the authors should choose cells that are more similar in terms of microtubule bundling for this figure, i.e. a control without microtubule bundles. 3) In Fig. 5D, the authors show how stable microtubules accumulate in the spindle midzone over time using their new sensor. However, they do not measure the total microtubule load of the midzone, thus it is not clear whether what they measure is simply an accumulation of microtubules. 4) One of the major drawbacks of their tool is that they need to select cells with "low expression levels". What this actually means remained unclear throughout the manuscript, nor did they mention how many cells in a population of transfected cells would fulfill this criterium and could thus be used for analysis. The authors should thus show a large field of view of a typical cell culture transfected with their sensor, and label the cells in this field that they consider "low expression" and "high expression". They might also want to give numbers for this, which should be easy by re-analyzing their experiments. Did the authors consider the use of lentivirus to transduce cells with their sensor? This would most likely allow a better tuning of the expression levels.
Minor points: 1) The authors have developed a great tool that will certainly be used a lot in the future. However, the name they have chosen for their binder, rigor-2xmNeongreen, is hard to memorize and even hard to pronounce. They should think of a more simple, intuitive name. Also, they should use the same name throughout the manuscript -currently they use different nomenclatures in the figures and the text. Along the same lines, the name for the stable cell line -U2OS-FlpIN;Kif5b-rigor-2xmNeongreen -is unpronounceable.
2) Introduction line 34: the authors chose a difficult-to-understand sentence to say that microtubules carrying low levels of tubulin PTMs, which is often visualized by the presence of the C-terminal tyrosine on alpha-tubulin, are also dynamic. They should clarify this for readers not familiar with the field of tubulin PTMs. It is also important to mention that microtubules are not strictly partitioned in tyrosinated and detyrosinated microtubules in cells, both forms do co-exist in the same microtubules and some microtubules have more detyr-tubulin, while others are more labelled with tyr-tubulin.
3) Along the lines of the point mentioned above, in line 76 the authors mention that their new sensor labels a subset of detyrmicrotubules. Obviously, this reflects that the stable microtubules the sensor detects are not necessarily identical with the detyrpool of microtubules, but it might also reflect that being detyr-positive can represent a range of detyr-levels on these microtubules. If the authors want to make a strong point about they, they should co-stain their cells with the tyr-tubulin antibody to check whether the detyr-microtubules that are not bound by the sensor are perhaps more tyrosinated? Alternatively, they should at least point out this possibility, and given that the figure is in the supplement, this could be done in the figure legend. 4) In the laser cutting experiments (Fig. 4) the authors show that microtubules labelled with their sensor do not depolymerize when cut. While they have shown in previous experiments that the new sensor does not appear to change microtubule dynamics, they have never formally shown that it does not prevent depolymerization, thus this possibility does still remain. The authors should mention this. They might consider showing that another recently published live-cell sensor of microtubules that is specific to tyrosinated microtubules (thus labelling the more dynamic microtubules) does not prevent depolymerization under the conditions the authors use in their experiments. Fig. 4L that their sensor does also label dynamic microtubules, but much more transiently, is highly intriguing. Why does the sensor fall off the microtubules so quickly? In the light of the current literature, did the authors consider the possibility that the rigor mutant of kinesin actually could extrude single tubulin molecules from the microtubule lattice?

A live-cell marker to visualize the dynamics of stable microtubules throughout the cell cycle
Klara I. Jansen et al.

Reviewer #1
Cells contain a mixture of stable microtubules with slow turnover and dynamic microtubules that turn over rapidly. These different microtubule populations frequently are post-translationally modified, mainly acetylation, detyrosination and glutamylation. However, while antibodies for these modifications have been in use for a long time and have served as proxy markers for stable microtubules, there are no live markers for stable microtubules. The lack of such markers has been an impediment into deeper cell biological investigations. The authors explore in this study the use of a rigor form of Kinesin-1 to preferentially label and image stable microtubules in live cells. They show they can image these stable microtubules over long times and observe their fate during cell division.

While the idea is interesting and the imaging is well-done (not surprising from the Kapitein group), I have strong reservations about the rigor and significance of this work as it currently stands. This field has been muddled by a lot of phenomenology and a lack of mechanism. The work described here, while potentially interesting and useful, will only further confuse the field in its current state because it is not clear (i) what this rigor kinesin construct recognizes and (ii) what its effect is on microtubule dynamics. These points would have to be addressed in order for this marker to be reliable. In addition, given that this is a resource paper, the authors should show that their observations hold in at least one other cell line. There is also a lack of scholarship, with both old and new work not being cited. The authors fail to cite and discuss relevant studies that have direct impact on the interpretation of their experiments.
My specific comments are below. I hope the authors will consider them.
 We are grateful to the reviewer for the thoughtful feedback on our manuscript. The reviewer raises two key questions: what does the marker recognize and what is the effect on microtubule dynamics?
In our manuscript, we describe a multitude of experiments that, in our opinion, convincingly demonstrate that our marker labels a subset of stable, long-lived microtubules. The existence of these microtubules is well-known and earlier correlative work has revealed that these stable microtubules are nocodazole-resistant, enriched in post-translational modifications, sensitive to serum starvation, and vary in abundance during the cell cycle. We directly confirm these findings in live cells using our new marker, which provides new opportunities to monitor the dynamics of stable microtubules during the cell cycle, as demonstrated in our mitosis experiments. In our opinion, this wide array of experiments validates our marker as a live-cell marker for the subset of long-lived, stable microtubules, even when it remains unknown what exactly this motor recognizes. In a similar fashion, EB proteins were used as markers for dynamic microtubules many years before it was understood what exactly they recognize at the microtubule plus-end.
Nonetheless, we were also very keen on understanding the selectivity of our marker and therefore started a collaboration with structural biologists when we first established our marker. Based on the emerging evidence in the literature that microtubules can exist in different lattice states and that these can be both recognized and reinforced by different microtubule-associated proteins, we hypothesized that stable microtubules inside cells feature an expanded microtubule lattice and that our marker recognizes this expanded lattice state. Our collaborators therefore analyzed microtubule lattice spacing inside cells using cryo-focused ion beam milling and cryo-electron microscopy, which revealed that a fraction of microtubules in wildtype cells indeed have an expanded lattice. Moreover, correlative cryo-light and electron microscopy revealed that our marker specifically decorates microtubules that have an expanded lattice. Together, these results demonstrate that stable microtubules inside cells display an expanded lattice and that this is what our marker recognizes. Given the enormous amount of highly-specialized work performed by our collaborators and following the editorial guidance that we received, these results will be presented in a separate (follow-up) manuscript with different first and last authors.
With respect to the second point, the effect on microtubule properties, we realize that there is always an interplay between recognizing a conformation and inducing a conformation. Therefore, if we overexpress our marker at high levels it could alter the lattice of non-stable microtubules and thereby promote stabilization by promoting the binding of additional stabilizing proteins. Our manuscript has carefully addressed this by examining the effect of different levels of overexpression on acetylation, microtubule organization, and intracellular transport. Upon request of the reviewer, we performed additional controls and we have measured the concentration range in which our marker is a passive marker that does not alter microtubule physiology. These results are discussed in point 3 below.
In addition to these two key points, we now show the use of our marker in a panel of commonly used cell lines. We also went carefully over our references and added more citations to earlier work.  Our data is consistent with the notion that acetylation is a modification that accumulates on stable microtubules, but does not directly confer significant stability to microtubules (i.e. protection to nocodazole-induced depolymerization). Thus, in control situations acetylation is found on stable, long-lived microtubules, while chemically promoting acetylation leads to the acetylation of dynamic microtubules without stabilizing them, as confirmed by nocodazole treatment. Our finding that our marker (now termed StableMARK) binds to Taxol-stabilized microtubules is consistent with the established result that Taxol induces an expanded lattice (shown in vitro, as well as in the new manuscript of our collaborators mentioned above).
For the revised manuscript, we quantified the data from the serum starvation assay and demonstrate that StableMARK-decorated MTs are not over-stabilized, as they behave similarly to the control condition upon prolonged serum starvation. In addition, reappearance of stable MTs upon release of prolonged serum starvation is similar between StableMARK-expressing and nontransfected cells. This data is presented in Fig.4G-I. Finally, we have also changed the dashed line in Fig.1B from black to red to increase visibility. Figure 2, the high curvature of stabilized microtubules has been observed by other groups before.

Citation of this earlier work is encouraged (for example, see Friedman et al JCB 2010).
We agree that the high curvature of acetylated microtubules is a well-known feature that has been reported by many groups. We did not intend to suggest that these were new findings, but rather wanted to demonstrate that our marker specifically recognizes these stable microtubules and that we can now observe the spatiotemporal dynamics of this specific subset in live cells. We clarified this in the text and included more citations to earlier work.

One major issue with the argument that the authors are trying to make is that it is not clear whether their rigor kinesin is actually stabilizing microtubules. Microtubule dynamics parameters need to be determined in control cells and cells expressing their rigor kinesin. Just reporting EB1 comet numbers
is not enough -what are the microtubule dynamic parameters? These are highly feasible experiments, especially for this lab. I am sure that at higher concentration levels the rigor kinesin will affect microtubule dynamics. What is the expression range for which this is a passive marker?  Indeed, many microtubule-associated proteins will stabilize microtubules when expressed at high enough levels. Furthermore, live-cell markers are generally known to alter the system of interest when expressed at too high levels. This is well-known for LifeAct, EB1/3, and MT markers such as GFP-tubulin or the MT-binding domain of MAP7 (see Supporting Figure 1, below). Despite the artifacts that all of these widely used markers can induce, they remain well-accepted tools for cell biology because experimenters can be trained to recognize normal versus abnormal cells.
For our new marker, we have extensively analyzed how the level of stable and dynamic microtubules changes as a function of expression level. This revealed that the level of stable microtubules in cells (as identified by staining for acetylation) does not change at the expression levels that we use in our imaging. As expected, in higher expressing cells, the level of acetylation does increase, which is most likely the consequence of microtubule overstabilization leading to more long-lived microtubules that accumulate more modifications. Importantly, our serum starvation experiments reveal that stable microtubules can still disappear when decorated with our marker, demonstrating that our marker does not over-stabilize this specific subset. Now the question emerges whether experimenters that can be trained to identify normal cells while using LifeAct, EB1/3 or other MT markers can also be trained to identify cells with the proper level of our marker for stable microtubules. We think this is possible for a number of reasons. As shown in our manuscript, our marker overlaps with acetylated microtubules, which have a very stereotyped, perinuclear distribution and are often highly curved (see point 2). In our experience, it is quite straightforward to identify the population of cells whose expression levels result in a distribution of the marker that is consistent with stainings for acetylation in control cells. In addition, for the revised manuscript we have performed a series of fluorescence correlation spectroscopy experiments in order to measure the concentration of our marker in cells that we classify as low, medium or high expressing. This revealed that our marker can be found in concentrations ranging around 0.02-72 M, where concentrations below 2 M correspond to expression levels where we find no over-acetylation. This data is present in Sup. Fig.3D Therefore, we believe that the general behavior of our marker (mimicking wildtype cells at levels below 2 M and gradually inducing overstabilization in a concentration-dependent manner above 3 M) is very similar to many other, widely used live cell markers and that the various control experiments presented in our manuscript should be sufficient to guide new users towards the proper use of this marker in their model systems.
In our revised manuscript we also present further evidence that many stable microtubules do not have a dynamic end. Figure 6D-F shows that EB3 comets only rarely emerge from StableMARK-decorated microtubules, whereas Sup. Fig.4 B-D shows that also in wildtype cells most acetylated microtubules do not have a non-modified extension. Thus, in cells with expression levels between 0.02-2 M, dynamic microtubules are not decorated with our marker. Consistently, we do not see a difference in the growth of microtubule plus ends between cells with and without our marker (Fig. 4F).

4.
In vitro work showed that de-tyrosination affects the binding of kinesin-1 more so than acetylation ( Kaul et al 2014 -also not cited) yet the authors seem to have focused on acetylation. Is it possible that this is a marker for de-tyrosinated microtubules? They show that at low levels the amount of acetylated tubulin is about the same with and without the rigor mutant. (Figure 3A). It would be reassuring to see the same for de-tyr tubulin, and to see a quantification for this. If the marker responds more to tyrosination than acetylation they also may be missing a fraction of the stabilized microtubules.
 For the revised manuscript, we have now also tested the overlap between our marker and detyrosinated microtubules. We find that in control U2OS cells, detyrosinated microtubules are very sparse and that our marker clearly marks more microtubules. When we increase detyrosination by overexpression of VSH1 and SVBP most microtubules become detyrosinated (but not stabilized), while our marker still labels a subset of microtubules. This data is presented in Sup. Fig.1A-G. Based on this new data and the earlier data on acetylation, we conclude that our marker does not recognize either of these modifications directly, but does localize to microtubules that are acetylated. As discussed above, we have now found that the stable microtubules recognized by kinesin-1 have an expanded lattice and we think that this is what determines the selective binding of our marker, either directly or indirectly.

The
Rab6A transport data needs to be quantified to see whether transport speeds are changed by the presence of more possible roadblocks (ie rigor kinesin) on the MTs.( Figure 3G). The transport kinetics of other cargos should also be analyzed that use other motors than the two used for Rab6A transport (for example, a connection between lysosome transport and acetylated MTs has been previously established).
 We have now quantified the speeds of lysosomes and Rab6a vesicles in control cells and low StableMARK expressing cells. We found that the distributions of vesicle speeds in control cells and rigor-expressing cells are very similar. This data is presented in Fig.5F-I.
The serum starvation experiments are nice, but all they show is that the marked MTs can be depolymerized by this assay (already known). In general, the comment that "our data demonstrate that in the subpopulation of low rigor-expressing cells, artifacts of the MT cytoskeleton are minimal" is not backed up by sufficient data. It might very well be the case, but they need to do more to prove this so that this can be a marker that can be used with confidence in the field.
 As argued in our response to point 3, we believe that our data demonstrates that, when used at the proper expression level, our marker labels the existing subset of stable microtubules and does not lead to stabilization of additional microtubules. Nonetheless, our marker could still lead to additional stabilization of the stable microtubules. Earlier work has shown that upon serum starvation in 3T3 cells, stable microtubules disappear. We reasoned that if our marker would confer a strong additional stabilization, these microtubules might no longer disappear upon serum starvation. We therefore repeated this classic experiment and found that decoration with our marker did not prevent the depolymerization of stable microtubules, indicating that our marker does not overstabilize stable microtubules. In the revised manuscript, we have further quantified these experiments by comparing the level of acetylation with and without serum starvation in cells with and without expression of our marker and found no difference between non-transfected control cells and low StableMARK-expressing cells (Fig.4G-I).
6. Their investigation of the mechanism of stabilization is technically well-performed but I have reservations with their interpretation. Upon ablation they see ~50% of the MTs depolymerize from one end which they interpret as an aging effect, where stabilization starts at the +tip and is followed by stabilization along the lattice. There have been several papers on motor induced damage and longrange lattice effects, including the stabilization of the lattice by even substiochiometric amounts of kinesin (Andreu-Carbo et al, 2022, Triclin et al 2021Peet et al 2018). These need to be discussed here as they point to the possible perturbing role that their marker can actually have. If these microtubules also are healed more, they will also be more stable when ablated because of the incorporation of GTP tubulin (Aumeier et al. 2016;Vemu et al 2018).

Other in vitro experiments (Peet et al 2018, Shima et al 2018)
have revealed that kinesins can induce other changes in the microtubule lattice that are more subtle, but perhaps also more consequential. Purified kinesin motors can cause small extensions of in vitro polymerized microtubules, most likely by inducing a different microtubule lattice spacing (Peet et al 2018, Shima et al 2018). Thus, even without inducing the exchange of tubulin subunits, kinesins can induce a GTP-like (expanded) tubulin conformation within a GDP-tubulin lattice. As explained above, our collaborators have now found that a fraction of microtubules in wildtype cells indeed features an expanded lattice. Moreover, correlative light and electron microscopy revealed that our marker specifically decorates microtubules with an expanded lattice. These results support a model in which stable microtubules inside cells feature a specific lattice, which might be recognized and reinforced by a large number of MAPs, including kinesin-1. Since lattice expansion appears to be a cooperative effect, this also explains why our marker (and many other MAPs) can induce microtubule stabilization when expressed too highly, but is mostly reading out the lattice state at low concentrations. Following the editorial guidance that we received, these results will be presented in a separate manuscript by our collaborators.

When discussing laser ablation experiments and the asymmetry in depolymerization between the plus and minus ends, the authors strangely cite a paper from 2014. This is textbook work from Salmon and colleagues for example and should be cited.
 We cited the 2014 paper because it had used the exact same experimental configuration. In our revised manuscript, we now also cite the original work from the Salmon lab.

The authors make the interesting observation that the rigor kinesin coated MTs do not have dynamic (EB3) tips. Based on this they speculate that therefore there is a "special tip structure" on stabilized microtubules that prevents both growth and shrinkage. Could it be that their rigor mutant interferes with normal EB3 binding?
 To address this question we reasoned that if the lack of dynamic tips on stable microtubules would be induced by our marker, we should be able to find dynamic ends on many acetylated microtubules in non-transfected cells. We performed super-resolution microscopy (STED) on U2OS cells stained for acetylated tubulin and total tubulin and quantified the percentage of acetylated MT tips that had a dynamic end (as manifested by a stretch of total tubulin extending from the acetylated MT tip). We found that 89% of the acetylated MT tips did not have a dynamic end, which is very similar to what we found in our live-cell imaging experiments using StableMARK and EB3. We therefore conclude that the limited polymerization at the tip of stable MTs is not caused by the presence of low levels of StableMARK. This new data is presented in Sup. Fig.4B-D.

What is the effect of this rigor mutant on microtubule dynamics in vitro? These are highly feasible experiments that should be performed. And does it affect EB binding? Mechanistic data on what the effect of this mutant is on microtubule dynamics and binding of some ubiquitous effectors is needed.
 As discussed in the previous comment, we now provide additional support for our finding that most of the stable microtubules recognized by our marker in U2OS cells do not have dynamic end. Therefore, in vitro assays for microtubule dynamics that use seeds and free tubulin do not really recapitulate the properties of this specific subset of microtubules. Nonetheless, we initiated the purification of the rigor mutant to test its effect on in vitro microtubules. Unfortunately, we experienced some unexpected difficulties and delays in setting up these experiments. Because, as explained, there is no straightforward connection between such reconstitution assays and our cellular assays, we decided to focus on addressing the other comments.

This paper describes the use of a rigor mutant of kinesin-1 to identify stable microtubules (MTs) in living cells. authors show convincing data that a 2X-mNeon-tagged version of the rigor kinein-1 marks stable, acetylated MTs in cells. They then use the rigor kinesin to follow stable MT dynamics in interphase and mitotic cells and show a number of behaviors of these MTs, many of which have been
reported previously. Overall, the development of a live cell probe for stable MTs is an unmet need for cell biologists and has the potential to reveal novel behaviors of these MTs. However, the paper has some major issues (see below) that need to be addressed before the rigor kinesin-1 is characterized well enough to be a reliable live cell marker for stable MTs.
 We thank the reviewer for the thoughtful comments on our manuscript. We believe that the additional data presented in our revised manuscript addresses the key concerns of the reviewer.  (e.g., Liao G and Gundersen GG, JBC, 1996;Dunn S et al, JCS 2008;Cai D et al, PLOS Biol, 2009;Sirajuddin M. et al, NCB, 2014), and the rigor kinesin should be tested against this posttranslational modification.
 In our opinion, an operational definition of microtubule stability is the most correct one. Microtubules are stable if they resist cold-treatment or nocodazole treatment, or if they remain intact after being mechanically challenged (e.g. by laser ablation). This is also the reason why we validated our marker in these different conditions. The exact mechanism of stabilization might differ from cell to cell, because many different microtubule-associated proteins can contribute to stability and their expression differs from cell to cell.
Nonetheless, we were also very keen on understanding the selectivity of our marker and therefore started a collaboration with structural biologists when we first established our marker. Based on the emerging evidence in the literature that microtubules can exist in different lattice states and that these can be both recognized and reinforced by different microtubule-associated proteins, we hypothesized that stable microtubules inside cells feature an expanded microtubule lattice and that our marker recognizes this expanded lattice state. Our collaborators therefore analyzed microtubule lattice spacing inside cells using cryo-focused ion beam milling and cryo-electron microscopy, which revealed that a fraction of microtubules in wildtype cells indeed have an expanded lattice. Moreover, correlative cryo-light and electron microscopy revealed that our marker specifically decorates microtubules that have an expanded lattice. Together, these results demonstrate that stable microtubules inside cells display an expanded lattice and that this is what our marker recognizes. Given the enormous amount of highly-specialized work performed by our collaborators and following the editorial guidance that we received, we decided to present these results in a separate (follow-up) manuscript with different first and last authors.
For the revised manuscript, we have now also tested the overlap between our marker and detyrosinated microtubules. We find that in control U2OS cells, detyrosinated microtubules are very sparse and that our marker clearly marks more microtubules. When we increase detyrosination by overexpressing VSH1 and SVBP, most microtubules are detyrosinated (but not stabilized), while our marker still labels a subset of microtubules. This data is presented in Sup. Fig.1A-G. Based on this new data and the earlier data on acetylation, we conclude that our marker does not recognize either of these modifications, but does localize to microtubules that are acetylated. As discussed above, we have now found that the stable microtubules recognized by kinesin-1 have an expanded lattice and we think that this is what determines the selective binding of our marker, either directly or indirectly. Fig. 3C), that at "low" expression there is little effect on MTs, whereas at "high" expression, stable MTs increase, dynamic MTs are reduced, and overall MT distribution is disrupted. "Low" and "high" expression can mean different things to different people. Consequently, for the method to be reproducible and widely useful, it is critical that the authors define the level of rigor kinesin-1 expression that begins to show the artifacts they observe. For example, a comparison of the level of the rigor kinesin-1 to the endogenous level of kinesin-1 (or tubulin) would be helpful. Given the cell-cell variability with transient expression, it might be easier to make this comparison with stable cell lines with low and high expression.

The relationship between the level of rigor kinesin-1 expression and artifacts caused by its expression are incompletely documented. The authors conclude (and I am not sure I agree, for example EB1 comets and hence dynamic MTs are clearly affected in
 Indeed, many microtubule-associated proteins will stabilize microtubules when expressed at high enough levels. Furthermore, live-cell markers are generally known to alter the system of interest when expressed at too high levels. This is well-known for LifeAct, EB1/3, and MT markers such as mCherry-tubulin or the MT-binding domain of MAP7 (see Supporting Figure 1, above). Despite the artifacts that all of these widely used markers can induce, they remain well-accepted tools for cell biology because experimenters can be trained to recognize normal versus abnormal cells.
For our new marker, we have extensively analyzed how the level of stable and dynamic microtubules changes as a function of expression level. This revealed that the level of stable microtubules in cells (as identified using staining for acetylation) does not change at the levels that we use in our imaging. As expected, in higher expressing cells, the level of acetylation does increase, which is most likely the consequence of microtubule overstabilization leading to more long-lived microtubules that accumulate more modifications. Furthermore, our serum starvation experiments reveal that stable microtubules can still disappear when decorated with our marker, demonstrating that our marker also does not overstabilize this specific subset. Now the question emerges whether experimenters that can be trained to identify normal cells while using LifeAct, EB1/3 or other MT markers can also be trained to identify cells with the proper level of our marker for stable microtubules. We think this is possible for a number of reasons. As shown in our manuscript, our marker overlaps with acetylated microtubules, which have a very stereotyped, perinuclear distribution and are often highly curved. In our experience, it is quite straightforward to identify the population of cells whose expression levels result in a distribution of the marker that is consistent with stainings for acetylation in control cells. In addition, for the revised manuscript we have performed a series of fluorescence correlation spectroscopy experiments in order to measure the concentration of our marker in cells that we classify as low, medium or high expressing. This revealed that our marker can be found in concentrations ranging around 0.02-72 M, where concentrations below 2 M corresponds to expression levels where we find no over-acetylation. This data is present in Sup. Fig.3D.
Therefore, we believe that the general behavior of our marker (mimicking wildtype cells at levels between 0.02-2 M and gradually inducing overstabilization in a concentration-dependent manner above 3M) is very similar to many other, widely used live cell markers and that the various control experiments presented in our manuscript should be sufficient to guide new users towards the proper use of this marker in their model systems.

There is very little new about stable
MTs that is reported with the rigor kinesin leading me to wonder whether it will be all that useful. Part of this reflects the fact that the authors repeat previous findings that they are apparently unaware of. For example, previous studies have shown that stable MTs are curly (Piperno G et al, JCB, 1987 and others), do not grow for extended periods (Schulze E and Kirschner M, JCB, 1986;Webster D et al., PNAS, 1987;Infante AS et al, JCS, 2000;Palazzo AF et al, NCB, 2001), and depolymerize slowly when severed (Walker RA et al, JCB, 1989;Infante AS et al, JCS, 2000). The authors also seem unaware of all the molecular studies implicating numerous factors (Rho,CLASPs,ELKs,mDia,INF2) in stable MT formation. These studies should be acknowledged and additional studies should be considered to show the utility of the rigor kinesin.

 As recognized by all three reviewers, 'the development of a live cell probe for stable MTs is an unmet need for cell biologists and has the potential to reveal novel behaviors of these MTs.'
The goal of this manuscript was to carefully validate this new marker and to demonstrate that we can reproduce earlier findings on stable microtubules obtained more indirectly. As such our demonstration of curly microtubules, lack of growth, slow depolymerization upon severing etc. etc. was not intended to make new claims, but to validate our marker. We clarified this in the text and included more references to earlier work. In addition, we demonstrate new insights into the (re)organization of stable microtubules throughout the cell cycle. However, we never intended to extensively study the multiple mechanisms of stabilization by the well-known factors listed by the reviewer. After depositing our preprint, we have already shipped out our plasmid to many labs that have specific research questions and are excited to finally have a validated marker for live-cell imaging of stable microtubules, for example in specific cell types of various model organisms.
As outlined in point 1, we have now also used our marker in a collaborative study with cryo-EM experts to study the lattice spacing of stable microtubules inside cells. Cryo-EM is an extremely powerful technique for structural studies inside cells, because it preserves cellular ultrastructure so well. Nonetheless, it is impossible to identify different microtubule subset, because the fixation procedures are incompatible with antibody-based labeling procedures. So this is an example where our live-cell marker is extremely useful. Because it already decorates a subset of microtubules in live cells, it can also be used to identify these microtubules in cryo-fixed cells and to compare the microtubule lattice of stable and dynamic microtubules.
Additional comments: Fig 1F-G are not new or news. The data should be moved to the supplement.  As pointed out by the reviewers, there are widely varying ideas about the causes and consequences of microtubule stabilization. These experiments are connected to the subsequent experiments in control cells and StableMARK-expressing cells, which together convey an important point and we prefer to present them together. Based on the feedback we received when presenting this work, it also seems that, despite earlier key work from the Gundersen lab that we cite, the result that hyperacetylation does not lead to stabilization of the newly acetylated microtubules is not considered general knowledge. Fig. 2C is sliding as it could also be treadmilling or trapped in actin flow. The MT in Suppl Fig2C might be moving rather than depolymerizing. And, I do not see active looping in fig. 2E as described.

The description of the behaviors of rigor kinesin decorated MTs are imprecise and in some cases over interpreted. For example, it is unclear how the authors draw the conclusion that the MT shown in
 The speckled labeling by our marker enabled us to discriminate between sliding and treadmilling. The kymograph shown in Fig. 2D (now Fig.3D) shows that the speckles are moving together, which indicates that the microtubule moves as a whole and rules out treadmilling. Because the velocity of such motile events was in the order of the velocity of microtubule-based motors (see original Fig.  2F (now Fig.3F) and much faster than the typical speed of actin flow, we reasoned that sliding was the most logical interpretation. In Suppl Fig. 2C, our interpretation that this is a partial depolymerization event was based on the observation that the speckled pattern on the remaining microtubule does not change over time, as also shown in the kymograph in Suppl Fig. 2D. For Fig.  2E (now Fig.3E), we changed the wording to active curvature induction.
6. I observe a substantial change in the overall distribution of MTs cause by "high" rigor kinesin expression in Fig. 3A. This should be mentioned along with the other alterations cause by expression of the rigor kinesin.
 This was already mentioned in the text, but we will better highlight this.

The comparison in Fig3B between rigor kinesin expression and acetylated tubulin levels needs to include the acetylation levels in non-expressing cells. Otherwise, their claim that expression of the rigor kinesin does not affect acetylated tubulin levels is not valid. Additionally, normalizing the rigor kinesin levels to the total tubulin levels makes no sense to me -the graph should simply show rigor kinesin expression.
 Fig.3B the mean + SD for non-expressing cells is shown by the green lines. We normalized the rigor kinesin to the total tubulin levels in order to robustly compare images taken on different days, with perhaps slightly different illumination conditions.
8. The authors should plot EB1 comets per cell rather than EB1 comets/um to capture the clear reduction in the numbers of comets in both "low" and "high" expressing cells shown in Fig. 3C.
 We chose to quantify per area, because we noted that larger (untransfected) cells typically have more comets.
9. The data that Rab6A transport is unaffected needs quantification beyond the anecdotal example shown in Fig. 3G-I.
 For the revised version, we quantified the speeds of lysosomes and Rab6a vesicles in control cells and low StableMARK expressing cells. We found that the distributions of vesicle speeds in control cells and rigor-expressing cells are very similar. This data is presented in Fig.5F-I.  We thank the reviewer for the positive assessment of our manuscript and the helpful suggestions that have helped us to improve our manuscript.

Major points:
1) In Fig. 3A   We changed the example shown for the control condition. In addition, we now added Sup. Fig.6A, which shows for six different cell types a large field of view with multiple rigor-transfected and non-transfected cells. Here, the rigor-transfected cells are indiscernible from the non-transfected cells in the acetylated and total tubulin channel.
2) Line 175, Fig 3G-  In our revised manuscript, we quantified the speeds of lysosomes and Rab6a vesicles in control cells and in low expressing StableMARK cells. We found that the distribution of vesicle speeds in control cells and rigor-expressing cells are very similar. This data is presented in Fig.5F-I.
3) In Fig. 5D, the authors show how stable microtubules accumulate in the spindle midzone over time using their new sensor. However, they do not measure the total microtubule load of the midzone, thus it is not clear whether what they measure is simply an accumulation of microtubules.
 Because mitotic cells are very sensitive to photo-toxicity, we could not follow the accumulation of StableMARK and a total tubulin marker over time in live cells. Therefore, we quantified the accumulation of acetylated tubulin and total tubulin during different stages of cytokinetic bridge formation in U2OS WT cells. This clearly showed the relative increase in acetylated tubulin (and thus stabilization of MTs) during the formation of the cytokinetic bridge. This data is presented in Sup. Fig.5B,C.

4) One of the major drawbacks of their tool is that they need to select cells with "low expression levels". What this actually means remained unclear throughout the manuscript, nor did they mention how many cells in a population of transfected cells would fulfill this criterium and could thus be used for analysis.
The authors should thus show a large field of view of a typical cell culture transfected with their sensor, and label the cells in this field that they consider "low expression" and "high expression". They might also want to give numbers for this, which should be easy by re-analyzing their experiments. Did the authors consider the use of lentivirus to transduce cells with their sensor? This would most likely allow a better tuning of the expression levels.
 We agree that the need to select cells with the proper expression level could be perceived as a drawback. However, in our experience the situation is not different for many other widely used live-cell markers. For example LifeAct, EB1/3, and MT markers such as mCherry-tubulin or the MT-binding domain of MAP7 are generally known to alter the system of interest when expressed at too high levels. Despite the artifacts that all of these widely used markers can induce, they remain well-accepted tools for cell biology because experimenters can be trained to recognize normal versus abnormal cells.
Now the question emerges whether experimenters that can be trained to identify normal cells while using LifeAct, EB1/3 or other MT markers can also be trained to identify cells with the proper level of our marker for stable microtubules. We think this is possible for a number of reasons. As shown in our manuscript, our marker overlaps with acetylated microtubules, which have a very stereotyped, perinuclear distribution and are often highly curved. In our experience, it is quite straightforward to identify the population of cells whose expression levels result in a distribution of the marker that is consistent with stainings for acetylation in control cells. In addition, for the revised manuscript we have performed a series of fluorescence correlation spectroscopy experiments in order to measure the concentration of our marker in cells that we classify as low, medium or high expressing. This revealed that our marker can be found in concentrations ranging from 0.02 µM -72 µM , where concentrations below 2 M corresponds to expression levels where we find no over-acetylation. This data is presented in Sup. Fig.3D.
Therefore, we believe that the general behavior of our marker (mimicking wildtype cells at levels below 2 M and inducing overstabilization above 3 M) is very similar to many other, widely used live cell markers and that the various control experiments presented in our manuscript should be sufficient to guide new users towards the proper use of this marker in their model systems.

Minor points:
1) The authors have developed a great tool that will certainly be used a lot in the future. However, the name they have chosen for their binder, rigor-2xmNeongreen, is hard to memorize and even hard to pronounce. They should think of a more simple, intuitive name. Also, they should use the same name throughout the manuscript -currently they use different nomenclatures in the figures and the text. Along the same lines, the name for the stable cell line -U2OS-FlpIN;Kif5b-rigor-2xmNeongreen -is unpronounceable.
 Following the reviewer's suggestion, we have changed the name of our marker to StableMARK (Stable Microtubule-Associated Rigor-Kinesin) and applied it uniformly in text and figures.
2) Introduction line 34: the authors chose a difficult-to-understand sentence to say that microtubules carrying low levels of tubulin PTMs, which is often visualized by the presence of the C-terminal tyrosine on alpha-tubulin, are also dynamic. They should clarify this for readers not familiar with the field of tubulin PTMs. It is also important to mention that microtubules are not strictly partitioned in tyrosinated and detyrosinated microtubules in cells, both forms do co-exist in the same microtubules and some microtubules have more detyr-tubulin, while others are more labelled with tyr-tubulin.
 We clarified this in our revised manuscript.
3) Along the lines of the point mentioned above, in line 76 the authors mention that their new sensor labels a subset of detyr-microtubules. Obviously, this reflects that the stable microtubules the sensor detects are not necessarily identical with the detyr-pool of microtubules, but it might also reflect that being detyr-positive can represent a range of detyr-levels on these microtubules. If the authors want to make a strong point about they, they should co-stain their cells with the tyr-tubulin antibody to check whether the detyr-microtubules that are not bound by the sensor are perhaps more tyrosinated? Alternatively, they should at least point out this possibility, and given that the figure is in the supplement, this could be done in the figure legend.
 For the revised manuscript, we have now also tested the overlap between our marker and detyrosinated microtubules. We find that in control U2OS cells detyrosinated microtubules are very sparse and that our marker clearly marks more microtubules than strictly the set of detyrosinated microtubules. When we increase detyrosination by overexpressing VSH1 and SVBP most microtubules are detyrosinated (but not stabilized), while our marker still labels a subset of microtubules. This data is presented in Sup. Fig.1A-G. Based on this new data and the earlier data on acetylation, we conclude that our marker does not recognize either of these modifications, but does localize to microtubules that are acetylated. As discussed in point 5 below, we now also have structural data revealing that the stable microtubules recognized by kinesin-1 have an expanded lattice and we think that this is what determines the selective binding of our marker, either directly or indirectly. (Fig. 4)   Earlier laser ablation experiments in cells without our marker revealed that, even in the absence of minus-end stabilizing proteins, a subset of microtubules would not depolymerize upon laserinduced severing (Jiang et al., Dev Cell 2014). This indicates that slow depolymerization of a subset of microtubules is not necessarily induced by the expression of our marker. To examine whether our marker could potentially overstabilize stable microtubules and prevent their depolymerization, we have performed serum starvation experiments. Previous work has shown that stable microtubules disappear upon starvation. We found that in cells expressing our marker, starvationinduced depolymerization was not impaired, indicating that our marker does not prevent depolymerization. We have further quantified these findings in the revised manuscript (see Fig.3I). Fig. 4L that their sensor does also label dynamic microtubules, but much more transiently, is highly intriguing. Why does the sensor fall off the microtubules so quickly? In the light of the current literature, did the authors consider the possibility that the rigor mutant of kinesin actually could extrude single tubulin molecules from the microtubule lattice?

5) The observation in
 Recent studies have indeed reported that motor proteins can promote tubulin turnover in reconstitution assays with purified proteins (Triclin et al., 2021, Andreu-Carbo et al., 2022. This process is generally believed to depend on the energy derived from ATP hydrolysis. Since our marker does not hydrolyze ATP, it is unlikely that it will cause the type of microtubule damage that has been reported in the above mentioned in vitro assays. . Thus, even without inducing the exchange of tubulin subunits, kinesins can induce a GTP-like (expanded) tubulin conformation within a GDP-tubulin lattice. Because all these results were from in vitro experiments with purified components, we were curious to understand how this related to cells and thus started a collaboration with structural biologists to study microtubule lattice spacing inside cells. This revealed that a fraction of microtubules in wildtype cells indeed has an expanded lattice. Moreover, correlative light and electron microscopy revealed that our marker specifically decorated microtubules with an expanded lattice. These results support a model in which stable microtubules inside cells feature a specific lattice, which might be recognized and reinforced by a large number of microtubuleassociated proteins, including kinesin-1. We therefore think that the transient binding to dynamic microtubules reflects a much lower affinity for compacted lattices and the inability of our marker to adopt the two-head bound rigor-conformation on these microtubules. More specifically, we hypothesize that the inter-dimer spacing of the expanded lattice promotes the rigor to adopt a twohead-bound state, while for the compacted lattice, this is not be the case. These structural studies of microtubule lattice spacings in cells were spearheaded by our collaborators and are documented in a complementary manuscript, following the editorial guidance that we received from Journal of Cell Biology.