Author response:
The following is the authors’ response to the original reviews.
We have carefully addressed the insightful comments provided by the reviewers which thoroughly increased our comprehension of the dynamics of centriole amplification. The manuscript has been revised accordingly and put in the context of the two papers we published since our last submission, showing that MCC differentiation is a genuine cell cycle variant. A point by point answer to all reviewer comments is provided below.
Briefly:
We have streamlined terminology and nomenclature in text and figures / better define experimental conditions with nocodazole
We have tested the role of dyneins in the dynamics of centriole amplification
We have done correlative light and electron microscopy on the early stages of centriole amplification
We have analyzed a new single cell RNA seq dataset comparing canonical and MCC cell cycle variants in mouse brain progenitors
Collectively, this allowed us to make a clearer parallel with what occurs during centriole duplication and to demonstrate that centriole biogenesis in the MCC cell cycle is marked by the superimposition of 2 canonical centriole cycles.
We believe the manuscript will interest a broader readership since it now provides more fundamental insights on the mechanism of centriole biogenesis.
Public Reviews:
Reviewer #1 (Public Review):
The manuscript by Boudjema et al. describes the cellular events underlying centriole amplification and apical migration to allow the assembly of hundreds of motile cilia in multi-ciliated cells. For this, they use cell culture models in combination with fixed and live cell imaging using antibody staining and fluorescence from endogenously tagged centriole and deuterostome markers, respectively. The work is largely descriptive and functional analyses are restricted to treatment with the microtubule depolymerizing drug nocodazole. The imaging is state-of-the-art including confocal microscopy, live imaging with optical sectioning and high optical and temporal resolution, as well as super-resolution imaging by ultra-expansion microscopy.
The study does a good job of providing a very detailed description of the dynamics of centrioles and deuterostomes that lead to centriole amplification and apical migration in multiciliated cells. This detailed view was missing in previous work. It also reveals the involvement of microtubules at multiple steps: the formation of a cloud of deuterostome precursors, the nuclear envelope tethering of newly formed centrioles, their separation, and their migration to the apical surface.
It would have been useful to expand the analysis of the role of microtubules by including analyses of the requirement for specific microtubule motors, for a better understanding and additional evidence that microtubule-based transport is involved. A weak point is that there is no visualization of microtubules together with deuterosomes and centrioles at the different steps of centriole amplification and migration, to directly address how these structures may interact with and move along microtubules.
Overall, apart from experimental aspects and since this is largely a descriptive study, the manuscript would benefit from more precise language and a better description of the complex events underlying centriole amplification and movements.
We have streamlined terminology and nomenclature, clarified the description of the complex events, and test the role of dyneins in centriole amplification. Microtubules density in MCC does not allow to extract information from imaging. In addition, we have done correlative light and electron microscopy on the early stages of centriole amplification and analyzed a new single cell RNA seq dataset comparing canonical and MCC cell cycle variants in mouse brain progenitors. We also replied points by points to the reviewer specific comments.
Altogether, our new data allowed to demonstrate that centriole biogenesis in the MCC cell cycle is marked by the superimposition of 2 canonical centriole cycles. We believe the manuscript will interest a broader readership since it now provides more fundamental insights on the mechanism of centriole biogenesis.
Reviewer #2 (Public Review):
This important work will be of interest to centriole and cilia cell biologists. It describes in detail how microtubules control multiple aspects of centriole amplification in brain multiciliated cells. This study provides a greater time-resolved and molecular proteomic mapping of the different steps involved, with or without microtubule disruption. Boudjema et al. show that microtubules are important throughout the centriole amplification process, from the early stages, where the procentrioles emerge from a pericentriolar "nest", through the growth stage where microtubules maintain the perinuclear localisation, to the detachment stage, where microtubules assist in perinuclear disengagement and apical migration. The results are generally well supported by the evidence, but the manuscript would benefit significantly from some heavy editing to introduce more niche terms, standardize abbreviations in text, and labels on figures to help bring the readers, especially non-specialists, along with them - increasing the accessibility of their work.
We thank the reviewer for his/her enthusiasm. We have streamlined terminology and nomenclature and clarified the description of the complex events to increase the accessibility of our work. We also replied points by points to his/her specific comments.
Reviewer #3 (Public Review):
Summary:
In this manuscript, Boudjerna and Balagé et al. aim to elucidate the spatial origin of centriole amplification and the mechanisms behind the formation of an apical-basal body patch in multiciliated cells (MCCs). To this end, they focused on the role of microtubules and developed new tools for spatiotemporal and high-resolution analysis of different stages of centriole amplification, including the centrosome stages, A-stage, G-stage, and MCC-stage. Among these tools, the MEF-MCC cells grown on micropatterns stands out for its versatility as it is not tissue-specific and does not require epithelial cell-to-cell contact for differentiation. Additionally, the CEN2-GFP; mRuby-DEUP1 knock-in mouse model was used to study different stages of centriole amplification in physiological brain MCCs. This model offers an advantage over the previously described CEN2-GFP model by enabling the resolution of early events in centriole amplification through the visualization of DEUP1-positive structures and their dynamics. Finally, the authors leveraged powerful imaging techniques, including super-resolution microscopy, the U-ExM, and high-resolution live cell imaging in order to detect and track centriole amplification, elongation, disengagement, and migration.
By combining the MEF-MCC and knock-in mouse model with spatiotemporal imaging in control and nocodazole-treated cells (treated acutely or chronically), the authors define the sequence of events during centriole amplification, revealing the critical roles of microtubules for the first time. Initially, the centrosome-mediated microtubule network forms, organizing a pericentrosomal nest from which procentrioles and deuterosomes emerge. Their findings indicate the importance of microtubules in recruiting and maintaining pericentriolar material clouds that contain DEUP1, PCNT, SAS6, PLK1, PLK4, and tubulins. Following the amplification stage, the procentrioles mature, leading to cells displaying numerous MTOCs, as demonstrated by regrowth experiments. Mature centrioles then disengage from deuterosomes, attach to the nuclear envelope, and migrate to the apical surface facilitated by microtubules.
Strengths:
The manuscript provides new insights into the regulatory function of microtubules in centriole amplification. Addressing the role of microtubules during different stages of centriole amplification required the development of new tools to study brain MCCs, which will be useful in future studies of MCCs. A notable strength of this manuscript is the authors' thorough and quantitative analysis of highly dynamic processes in MCCs. The precision and detail in describing these dynamic events are impressive. This comprehensive analysis advances our understanding of MCC biology.
Weaknesses:
The role of microtubules and other molecular players during different stages of centriole amplification in brain MCCs can be further studied and strengthened using the tools developed in the manuscript. A more quantitative description of some of the analysis performed in the manuscript is required to strengthen the conclusions.
We thank the reviewer for his/her enthusiasm. We have tested the role of dyneins in the dynamics of centriole amplification, done correlative light and electron microscopy on the early stages of centriole amplification and analyzed a new single cell RNA seq dataset comparing canonical and MCC cell cycle variants in mouse brain progenitors. We also replied points by points to the reviewer specific comments.
Recommendations for the authors:
As you will see, all reviewers felt that the analyses of the involvement of microtubules should be strengthened by including controls and additional experiments. Also, they agree that significant text editing would help to improve the manuscript's accessibility and readability.
Specifically, they would suggest (1) streamline terminology and nomenclature in text and figures; (2) better define experimental conditions with nocodazole (concentrations used, effect on microtubules, effect on canonical centriole duplication); and (3), in the absence of other complementary genetic perturbation experiments, add a limitations paragraph in the discussion about conclusions drawn from nocodazole treatment alone.
Reviewer #1 (Recommendations For The Authors):
Main issues:
(1) The authors use variable terminology to describe the same or similar events/structures. For example, in Figure 1 they refer to "centrosome stage" where they observe a pericentrin "cloud", which they later refer to as a "nest". In all other figures the first stage is not referred to as the "centrosome stage" but as the "cloud stage". Again, they also describe the "cloud" as a "nest" occasionally, but not always. In the cartoon, the nest is termed "centrosome cradle". The variable and inconsistent use of terms is confusing and the authors do not provide any explanation for the use of one vs. another.
The text is now corrected. The centrosome stage corresponds to the stage preceding the beginning of centriole amplification in MCC progenitor. The pericentrosomal cloud of centriole and deuterosome elements forms later on, during the amplification A-stage. The formation of this cloud marks the beginning of A-stage, and persists up to G-stage where it dissolves. When we show that the cloud hosts the first stages of centriole biogenesis, we defined it as a “nest”. We do not use anymore the term craddle.
(2) What prompted the authors to use the term "nest"? It gives the impression that they describe aspecific physical entity/structure (also depicted in this way in Figure 3P, with microtubules outside of this structure), but what is the evidence for this?
The cloud is the spatial entity and the term “nest” is used to define a function of this transient compartment. We decided to keep the term “nest” as we now identified it with correlative light and electron microscopy, in addition to U-ExM, and show that the accumulation of centriole and deuterosome elements is accompanied by the formation of immature procentrioles, deprived of MT walls, as well as immature and empty deuterosomes. The scheme with MT outside the cloud/nest is misleading as we see MT organized by the mother centriole. We have now changed this.
(3) The "nest" may simply be a dynamic accumulation of precursor particles around the centrosome, similar to what has been described for centriolar satellites. Rather than proposing a new entity, I suggest testing whether the "nest" particles may colocalize with PCM1 and thus may be related to centriolar satellites. Based on the data, the nest would simply be the centrosomal MTOC that organizes a radial microtubule array on which particles move around its center. In the absence of other evidence, I am not convinced that a new term is needed.
We totally agree with the reviewer: the centrosome, as MTOC, concentrates centriolar and deuterosome components. This cloud is consistently dissolved when MT are depolymerized or dyneins inhibited. So, the physical entity is a “cloud”. We used the term “nest” to propose one function for this cloud which is to form deuterosomes and centrioles, before they move away for maturation. In fact, deuterosome and centriole formation are hindered when the cloud is dissolved. We have tried to edit the text all over the manuscript to make it clearer.
(4) Role of MTs: are microtubules required or do they just facilitate some of the investigated events?
The reason why the role of MT has not been tested yet during centriole amplification is probably because MT not only constitute the cell cytoskeleton on which molecular motors ride to transport cargos or distribute forces, they are also the core component of the structures we are studying. This is why we have tested a range of nocodazole concentrations and used concentrations where MT are perturbed but not entirely depolymerized, allowing centrioles to be produced (Fig. 4 Supplementary 1A-B). This may lead to an underestimation of the role of MT but we cannot study the role of MT on centriole amplification if centrioles cannot be formed.
Does multi-ciliation in these models eventually occur normally under the concentrations and treatment conditions used here? This should be tested and discussed in the context of whether microtubules are indeed required and at what step of the entire process (amplification, migration, ciliogenesis) they may be critical.
We did both chronic and acute treatments.
Chronic treatments were done to test the overall efficiency of centriole amplification when MT (or dyneins) are perturbed. Chronic treatments were used to assess the role of MT (or dyneins) on the global efficiency of centriole and deuterosome formation (number of cells able to amplify, number/size/loading of deuterosomes, final number of centrioles (Fig. 4H-I, Fig. 4 Supplementary 2 B-D). In these chronic treatment, we focused on centriole amplification and not ciliation since it was the scope of this study. Also, we did not take ciliation as a readout of amplification because ciliation is relying on MT polymerization.
Then, we also did acute treatments to test the role of MT (or dyneins) at each stage of amplification (A-amplification, G-growth, D-disengagement, M-migration; Fig. 4, 5, 7, 8 and associated supplementary figures). Since one stage is dependent on the precedent one, this enabled us to decipher the direct role of MT (or dyneins) on each single stage. We have now edited text, methods, legends and pictograms to be clear on whether acute or chronic treatment was done.
(5) Can the authors include control (non-amplifying) progenitors in their analyses? It would be useful to know what the signal and distribution of each specific marker are before differentiation begins (before the cloud stage).
Non amplifying progenitors are analyzed and constitute the so-called “centrosome stage”. We have now precised it and called it the “progenitor stage”.
(6) Figure 2: Again, the terminology is confusing, since the authors describe that DEUP1 forms a "cloud" with centrin during the A stage.
Corrections have been done as explained in point 1.
(7) Description Figure 3: the authors introduce yet another term: "halo" A-stage. Is this the early A stage? Again, this is not explained and confusing. More systematic and consistent description is needed.
Corrections have been done as explained in point 1. The term halos is used un the lab as it was the first term we used in our Nature paper in 2014 in reference to the halo described by Erich Nigg when they overexpressed Plk4. It was an error to use it in the manuscript.
(8) Nocodazole treatments: the used concentrations are quite high.
MCC develop a very dense and stable MT network that is not comparable to cycling cells. MT are very difficult to depolymerize entirely (Fig. 4 Supplementary 1A-B).
(a) To avoid non-specific effects the authors should test what the minimal concentration is that completely depolymerizes microtubules in their cell model and perform analyses at this concentration.
We have of course tested a range of nocodazole concentrations at the beginning of the study (Fig. 4 supplementary 1A-B), and used concentrations where MT are perturbed but not entirely depolymerized, allowing centrioles to be produced (see answer to point 4). In case it was not clear, we refer to this now several time and more clearly in the text and methods.
(b) They should demonstrate depolymerization of microtubules by microtubule staining in the acute and chronic noc treatments and at the different noc concentrations used.
This is, and was, in supplementary material (same, Fig. 4 supplementary 1A).
(c) The authors should demonstrate that the used nocodazole concentrations do not impair normal centriole biogenesis during the cell cycle in these cells; if so, impaired assembly of centriole wall MTs may contribute to the observed effects in Figure 4.
As mentioned in point 8b, we have of course tested a range of nocodazole concentrations at the beginning of the study (Fig. 4 supplementary 1A), and used concentrations where MT are perturbed but not entirely depolymerized, allowing centrioles to be produced (see answer to point 4). The ability of the cells to form centrioles during chronic treatments were always assessed using immunostainings of SAS6 and/or CEN2-GFP signals (now exemplified in Fig. 4 Supplementary 1B). We also did EM analysis on cells treated with the highest doses of nocodazole (Nocodazole 10 uM for 24h) and this showed that centrioles can form with, what seems to be MT walls, in cells totally deprived of cytoplasmic MT fibers (Fig. 4 Supplementary 3-4). However, this does not show that all the cells can, because the number of cells that can be analyzed by EM are not sufficient to conclude. Also, one cannot assess whether MT walls are properly polymerized. However, the absence of MT walls should not change the results of the Figure 4, which are based on DEUP1, SAS6 or CEN2-GFP signals for deuterosomes and centrioles. Also MT depolymerization affects the formation of deuterosomes, which should not be altered by MT wall defects as it is not affected, even when centriole formation is blocked (LoMastro et al., 2024). Last but not least, we now show that blocking dyneins, as a comparable and even greater effect, on the formation of the cloud, deuterosomes and centrioles (Fig. 4C-I and Supplementary Fig. 4), which confirms that MTOC function, rather that MT wall formation, explain the centriole biogenesis alteration shown in Figure 4.
(9) The authors repeatedly refer to the centriole-to-centrosome conversion of amplified centrioles and how this resembles centriole-to-centrosome conversion during the cell cycle. However, they incorrectly claim that this occurs at the G2/M transition. PLK1-dependent modification occurs at this stage, but conversion and PCM recruitment only occur after mitosis (see original work by the Tsou lab, which needs to be cited here).
We agree with the reviewer. We have now added additional data to show clearly that centriole biogenesis, which requires two cell cycles to proceed in cycling cells, is accelerated during the MCC cell cycle variant where the elongation and maturation cycles are superimposed. This is now clearly shown in Fig. 3, 5, 9 and discussed.
(10) Figure 6H-J: the authors claim that at low noc concentration, more D-stage cells showed incomplete disengagement than in controls, but the effect is shown only for the highest 10 µM concentration. Do any eof the phenotypes in Figure 6 also occur at the lowest noc concentration (assuming it depolymerizes MTs)? Again, it is crucial to demonstrate this, to exclude unspecific effects not linked to MT depolymerization.
An error was made on the figure (but not in the legend). In Figure 6, chronic treatments are at 1 or 5 µM. Only acute treatments were done using 10 µM. In both cases, MT are not entirely depolymerized in these experiments (Fig. 4 supplementary 1A).
(11) Disengagement, Figure 7: The authors describe that DEUP1 signal spreads all over the cytoplasm and becomes diffuse during this process, but one cannot see a diffusive signal throughout cells in the figures.
We pushed the contrast to make it clearer but the deuterosomes are still bright at this stage and it is difficult to have both signal clear (now in Fig. 6B). We have also changed the example in video (now video 19) to show it more clearly with DEUP1 channel alone.
(12) Figure 7: localization of disengaged centrioles at microtubule "nodes" is not clear from the images. There are many centrioles and random colocalization may be expected simply based on the high number. Higher resolution and/or magnification and quantification would be needed.
We have edited and now say that centrioles “colocalize” with MT which, since centrioles nucleate MT, seems normal. We agree that it could be random, but given the density of MT, and the number of centrioles, it does not seem opportune to us to quantify. We can just say that we never see centrioles is regions that are deprived of MT.
(13) The term "diffusive" to describe slow centriole movements in Figure 8 suggests that it is not motor or force-dependent, but there is no evidence for that. Movement based on opposing forces could produce a similar result, but would not be considered diffusive.
We agree. We have changed “diffusive” by “diffusive-like”.
(14) The manuscript would greatly benefit from the analysis of some candidate motor activities that may drive the movement and migrations of centrioles in this system. This would support the importance of the microtubule network for the specific steps in these processes, and better define its role beyond "being required". Dynein may be a candidate or minus end-directed kinesins. Since chemical inhibitors are available, these types of experiments would be straightforward.
We formerly tested ciliobrevin but had hard time because of the small stability of the drug. Since our submission to eLife, we tested dynapyrazol and dynarestin and found dynapyrazol very efficient in dissolving the Golgi, a good readout of dynein inhibition. We sought to test the role of dyneins, using dynapyrazol, on (i) the formation of the pericentrosomal cloud in A-stage, (ii) the oscillation of DEUP1+ structures during A-stage, (iii) the number, size, loading of deuterosome, (iv) the final number of centrioles, (v) the migration to the nuclear membrane and (vi) the final apical migration of centrioles. The results are now inserted in main and associated Fig. 4, 5, 7, 8, 9.
(15) Discussion:
"the role microtubules" lacks "of"
This is now edited.
"This lack is..." Lack of what?
This is now edited.
"reflexive link" - meaning of "reflexive" is not clear in this context
We have removed it.
In my opinion, the study does not identify a nest composed of DEUP1, PCNT, and Centrin2; it only shows that these components accumulate as particles around the centrosome, which functions as MTOC. Consequently, it seems that the "nest" does not exist when MT is depolymerized. One could consider the center of the centrosomal MT array as a nest in this context, but there is no evidence of a specific new structure as suggested by the way the term is used in the manuscript.
This is what we want to say: the center of the MT array become a nest in this context. We do not state that there is a specific new structure. We just say that MT and dynein dependent concentration of centriole and deuterosome components exists and that this region nests the birth of centrioles and deuterosomes. Also, this compartment is restricted in time and space, which justifies to use a specific term. The MTOC exists in the progenitor cell, while this compartment, marked by DEUP1, Centrin, PCNT accumulation, appears at the beginning of amplification and grows during A-stage to be dissolved at G-stage when all the deuterosomes and centrioles have moved away.
What is the evidence that "DEUP1 is a centrosomal protein before building deuterosome structures"? It would be good to refer to the specific experiment. Does DEUP1 localize at centrioles also in the absence of microtubules? If not, I would not consider it a centrosomal protein.
We have removed this statement to avoid misinterpretation.
"This reminds the centriole-to-centrosome conversion..." the sentence is missing an "of"; also, again the authors confuse the order of events during the cell cycle, where centrosome conversion occurs after completion of mitosis, not at G2/M transition.
We have removed this statement to avoid misinterpretation. Also, see Point 9.
"microtubule dependent nuclear migration" should be rephrased; it sounds as if the nucleus migrates.
This has been changed
The following discussion of disengagement being linked to association with the nuclear envelope and resembling the process in cycling cells is misleading. In cycling cells movement of centrioles along the nuclear envelope occurs at G2/M and drives centrosome separation (separation of centriole pairs) in preparation for mitosis, not centriole disengagement.
We are now clearer. We compare centriole-loaded deuterosome organization around the nuclear membrane to the migration of new centrosomes during early prophase (Fig. 5F-H, Fig. 5 Supplementary 2G-K).
Regarding the possibility that forces by microtubules generated by the daughter centriole drive disengagement also in cycling cells, I would argue that this is unlikely since the daughter centriole can only nucleate microtubules after disengagement has occurred (and conversion to centrosome/PCM recruitment). Once this happens, it may physically separate the disengaged centrioles, which is a different type of activity. Indeed, originally the term "disengagement" was coined to specifically describe the loss of the perpendicular engagement of daughter centrioles with their mothers (Tsou and Stearns, Nature, 2006).
We have removed this statement to avoid misinterpretation. The perpendicular engagement is difficult to assess on deuterosomes but we do see by live imaging, that attachment changes during D-stage, before centrioles detach clearly from deuterosomes.
"high resolutive" should be "high resolution"
Edit done.
"splitted" should be "split"
Edit done.
"Consistently, when the mitotic oscillator is dis-inhibited and cells enter pseudo-mitotic events, centrioles show clear and rapid cell-cycle like clustering" This sentence is not understandable without further explanation; what does mitotic oscillator refer to? What are pseudo-mitotic events? What is cell cycle-like clustering?
We have removed this statement.
Minor:
(1) Abstract: "Centriole number must be restricted to two..." Since cells are born with two centrioles and have 4 centrioles (2 pairs) when they enter mitosis, this sentence is inaccurate.
The sentence has changed.
(2) Abstract: "reflexive link"; I am not sure what the term "reflexive" refers to?
We have removed this statement to avoid misinterpretation.
(3) Figure 1C, D: it should be described better that the larger magnification panels represent overlays of many cells and what marker they show. This is not obvious since the smaller single-cell panels always show two different markers. Also, it would be more useful to show also single cells in the magnified view. The overlay does not allow us to see if a marker forms a cloud or a single dot, which is as important as the cell-to-cell variation in distribution.
We have clarified this in the text and the legend. The cell-to-cell variation cannot be estimated with the overlay, but the projection from several cells (number precised) allows to see that the signal is confined in a restricted region. Or not. Which is what we wanted to analyze.
Related to the above, the authors say that pericentrin forms a cloud at the top left in panel D, but there is only one confined centrosomal dot in the single-cell panel.
The sentence has changed.
(4) Results, Figure 2F; video 4: The authors claim connection and disconnection of DEUP1 aggregates with centrosomal centrioles; can the authors comment on the spatial resolution including in z in this movie to support this claim? Can they exclude that the structures are in proximity of each other rather than "connected"?
This is a single z-section of 500nm. The resolution in xy is 128nm/pixel. Given the sizes of deuterosomes and a mature centriole, and given the fact that we observed this dynamics in several cells in live, we can state that the structures are connected. This is consistent with deuterosomes frequently observed “kissing” the daughter centriole by EM in the present manuscript (Fig. 2D, Fig. 2 supplementary 3 and 4 and Fig. 4 Supplementary 3-4). One has to look carefully at the daughter centriole (marked “dc”) and span in on the serial sections to see the connected deuterosome (marked by a star): this is at very early stage and therefore it is small. We have not zoomed in since previous manuscript have already described this at later stages with bigger deuterosomes. You can refer to main or supplementary figures in previous manuscripts (Al Jord 2014, Khoury Damaa 2024) where serial sections span the entire deuterosomes and daughter centrioles and show, with nanometric resolution, that both structures are frequently sticked to each others on tens of nanometers.
(5) The term "dynamics" as used in the manuscript should be plural.
It has been used plural, except when for “dynamic microtubules” and “dynamic attachment to the nucleus”, which we think is ok? We have not found any other singular uses in our manuscript.
(6) Figure 5: what does "YL1/2 procentriole intensity" refer to in panel F? This should be the intensity of microtubule asters.
This has been modified.
(7) Figure 6 - supplement 1B: contrary to the claim in the text, one cannot see tight colocalization with the nuclear pore marker. This seems to be a very small subset of particles and even in those cases colocalization is not tight. Also, what is the relevance of nuclear pore colocalization?
We edit and change the phrasing as ‘colocalization with NPC’ is not the good term. What we want to say is that there is a tight connection with the nuclear envelope as shown by the localization of NPC on the same z-section as centrioles. This is why we present a single z, to show that centrioles and NPC are on the same z-plane of 500nm. NPC are stained to outline the nuclear membrane. This is also clearly visible for G-stage centrioles in the XY plane. We have now added an entire z-stack on video 18.
Reviewer #2 (Recommendations For The Authors):
To improve accessibility of their manuscript, we would suggest making the following edits:
(1) Define 'specialist' or 'niche' terms each time you introduce them, such as 'pericentrosomal nest', or 'flower-like structures'.
This has been clarified.
(2) Have a think about abbreviations, again ones that work for people outside the project- this paper uses 'PC' for 'procentriole' but for many 'PC' is 'Parental centriole' or Figure 6J talks about 'D total' or 'D partial', leaves readers confused.
This has been clarified.
(3) Standardize your abbreviations throughout particularly for your treatments- sometimes Noco sometimes, NOCO, or your imaging experiments sometimes Cen-GFP, sometime CEN2-GFP (Figure 7A, D vs. Figure 6) or DEUP1- mRuby, DEUP1-mRuby3 or mRuby3-DEUP1?
We now use Nocodazole or Noco in the text and the figure respectively, CEN2-GFP and mRubyDEUP1.
(4) About 10% of the population, including several key figures in this field, are red-green color blind. Although 4 colour fluorescence is difficult to get right for everyone, choosing palettes (especially for two colour panels) is inclusive. More so, greyscale or inverted monochrome images make it easier for everyone to visualize changes in localization, size, and intensity. Red on black small foci is particularly difficult to discern. For example, Figure 3 - more individual channels in grayscale with arrows to mc, dc, and cilia would be helpful - difficult to distinguish stainings.
We thank the reviewer for this comment and for this recommendation of being more inclusive. We have done the changes.
To improve the conclusions drawn, we suggest some revisions below:
(1) Since the paper really hangs on it, a clearer description of the rationale for when, how long and how much nocodazole treatment was done is needed. The logic currently is difficult to follow seemingly random jumps 10x concentration are used. Microtubules control many aspects of cell biology and could be impacted. For example, I particularly found Figures 6D and H difficult to follow i.e. the timing for 6H seems off.
MCC develop a very dense and stable MT network that is not comparable to cycling cells. MT are very difficult to depolymerize entirely. We have of course tested a range of nocodazole concentrations at the beginning of the study and shown the extent of MT depolymerization under each treatment. We used concentrations where MT are perturbed but not entirely depolymerized, allowing centrioles to be produced (see answer to point 4 reviewer 1). The level of perturbation of MT and consequences on centriole formation at the different timings and doses were done for each experiment and are exemplified in Fig. 4 supplementary 1A-B. This figure was already present in the first version of the manuscript but we have now edited text, methods and pictograms to clarify this.
(2) Perhaps an extension of this point- in general how interdependent are the processes? If there is a defect at the nest stage, how much are the later defects secondary to this, or do MTs genuinely play direct roles at all stages or are these knock-on effects? How do the authors rule this out? Defects in the nest, lead to smaller and more DEUP1+ foci, with defects in concentrating procentriole factors and centrin, which lead to... For example, Figure 4B looks like centrin is reduced upon noco treatment? Does noco treatment affect Cetn2GFP levels globally? Individual channels grayscale would help visualise this better.
See also our answer to reviewer 1 point 8c.
The stages are indeed interdependent. This is why we did both chronic and acute treatments. Chronic treatments were done to test the overall efficiency of centriole amplification when MT are perturbed. We typically used low dose of 1µM because nocodazole remains 48h in the culture medium. Acute treatments were done to test the role of MT at each stage of amplification (A-amplification, G-growth, D-disengagement, M-migration). Most of the acute treatments were done live and nocodazole was applied after the first time point of live monitoring. We used 10µM to have a rapid effect, and because nocodazole remains only several hours in the culture medium. This allowed to monitor the stage “n”, in cells where the stage “n-1” was completed without any drug which allowed to analyze a stage without having perturbed the precedent one.
We now also test the consequences of dynein inhibition using both acute and chronic dynapyrazole treatments. We show that except for centriole migration, dynein inhibition phenocopies MT depolymerization (centriole number, perinuclear organization and disengagement as well as deuterosome number/loading/size).
Nocodazole chronic treatments do affect intensity of CEN2-GFP at G-stage centrioles suggesting an altered A-to-G transition. In D-stage, CEN2-GFP signal seems normal. We now mention this in the text and in the Fig. 4 Supplementary 1B.
(3) The authors nicely show the importance of MTs in the structure of the nest from which procentrioles and DEUP1 positive structures emerge. They suggest this nest may be what supports procentriole generation in the absence of DEUP1 and parental centrioles. Firstly how does this nest look in the absence of DEUP1 and/or parental centrioles (centrinone treatment)? This may be what they are trying to show in Figure 5 Supplement 1 but it currently is very difficult to digest what it is showing relative to controls and whether this is significant in the way it is plotted.
The nest is conserved in the DEUP1KO with or without centrosomal centrioles, as shown by accumulation of Centrin and PCNT at the center of the self-organised MT network (Mercey et al., 2019). This is in fact what motivated our study on the role of MT in centriole amplification. We have edited the legend to precise the quantification done, which is not related to this question. In this quantification, we show that the increased propensity to accumulate PCNT by centriole-loaded deuterosomes between A and G-stage is maintained in the absence of deuterosomes, indicating that centrioles themselves accumulate/recruit PCNT.
(4) Can you do CLEM on DEUP1-Ruby and these early foci at the cloud stage to see if they are visible at the ultrastructural level, relative to procentrioles, microtubules, and other electron-dense structures?
We thank the reviewer for this question. We have done CLEM on the pericentrosomal cloud during very early steps of centriole amplification. This showed that DEUP1 early accumulation at the centrosome corresponds to a region rich in fibro granular aggregates, suggesting that DEUP1 may be translated here, through locally concentrated centriolar sattelites, known to be involved in local translation. Then, small deuterosomes and immature centrioles are formed, within this cloud of sattelites, confirming that the pericentrosomal cloud is a nest for centriole biogenesis (Fig. 2C-D + Fig. 2 Supplementary 2-6 for control and Fig. 4 Supplementary 3-4 for nocodazole treated cells). This also shows that immature deuterosomes are not necessarily round shaped, and can be deprived of centriole loading.
(5) Check the scale bars- see Fig 4E. Check throughout.
Done.
(6) Figure 3 Supplement 1 and 2 don't match the legend and are likely reversed - which one is right?
Done.
(7) Technical issue - I couldn't play videos 6 or 16? Check these work.
Done.
(8) Nomenclature mammalian proteins- mouse or human- should be all caps DEUP1, PLK4, SAS6,etc. Watch your units- space between number and unit.
This has been done.
(9) Many of the graphs involve three biological replicates but why not plot the mean of each of the three experiments and do stats? The number of events measured may conflate the significance. Try using Superplots.
Here is how we proceed: we count the number of occurrence of the phenotype we monitor, and the total number of cells. We apply a X<sup>2</sup> to test whether there is a significative difference between our replicates in each condition. If not, we pool the number of occurrence of the phenotype we monitor and the total number of cells for the 3 replicates, and for each condition. Finally we apply a X<sup>2</sup> between the different conditions. This is how we usually proceed to avoid comparing a mean of percentages. This is now explained in the methods.
Minor points:
(1) "DEUP1 is a centrosomal protein and assembles deuterosomes in the pericentrosomal region in brain MCC". I am not sure you have evidence that DEUP1 is a centrosomal protein. You don't seem to study the relationship between centrosomes and DEUP1? Rewrite this title and tone down this claim.
This has been modified.
(2) Why the crossbow micropattern (versus some other shape) - seems very specific but not discussed?
We wanted a shape where centrosome is not localized at the center of mass of the nucleus. Among the corresponding patterns, the crossbow was the one where differentiating cells had less propensity to detach.
(3) Figure 2 - are the foci of DEUP1 at the cloud stage smaller than at A stage? How do they grow? Measure the diameter at cloud stage, just after they leave the cloud and then once they move away from centrosomal cloud and each other. If so, and they do indeed grow in size from the cloud stage to the growth stage which I think your images suggest - do you envision this happening with the gradual addition of DEUP1 rather than fusion?
Early deuterosomes are not easy to detect by light microscopy, because of accumulation of DEUP1 in the cloud. We did CLEM on the cloud of early A-stage cells to resolve the earliest deuterosomes which are often very small (see Fig. 2D, Fig. 2 Supplementary 2-6) suggesting that they grow, either by fusion, which we never observe in our movies at later A-stage, or by accretion of DEUP1. However, by light microscopy, we can detect very early but big deuterosomes, which we see splitting later on into smaller ones. So, we cannot conclude on the mechanism that regulate deuterosome size. This is now discussed in the discussion of the manuscript.
You say in the discussion:
"Consistently, we never observed fusion events of DEUP1 condensates in our time-lapse experiments. More importantly, we did FRAP experiments on endogenously tagged mRuby-DEUP1 in cells at the different stages of centriole amplification, and did not find significant recovery, supporting that centrosomal DEUP1+ foci and deuterosomes are not liquid-like structures (Figure 8 Supplementary 2)." How do you prove there is no fusion of deuterosomes?
It is always difficult to prove the absence of something, we agree! But we did tens of movies with high temporal resolution and never observed fusion events. But, as we say in the previous question, the very early deuterosomes can be very small and we do not distinguish them from the DEUP1+ cloud by live imaging. So at this stage, we cannot say. But later on, during A- or G-stage and when deuterosomes are outside the cloud to be easily observed, we very often observe deuterosomes bumping into each others and stay in close contact for minutes, but then moving away. This, for us, supports the lack of fusion properties. But the question remains open. We now explain this in the manuscript and have added an example in video 28.
If they are getting bigger as I think your imaging suggests from cloud to growth stage, then how is this happening?
MT depolymerisation and dynein inhibition leads to the formation of very small deuterosomes. Dynein inhibition can even lead to a block in the formation of new deuterosomes suggesting that DEUP1 concentration is a crucial parameter for condensation into deuterosomes. Deuterosome growth may happen through oligomerization of DEUP1 molecules allowed by their dyne-independent concentration. Sorokin in 1968 proposed that a supersaturation of deuterosome components may lead to their solid crystallization into deuterosomes. Deuterosome size can also be regulated by a more complex molecular cascade, involving post-translational modifications of DEUP1 or PCM, such as phosphorylations driven by the cell cycle machinery. This would be consistent with the fact that deuterosomes are very big in the absence of CCNO, a cyclin required for entering the MCC cell cycle variant. This will need further investigations.
I'm not sure FRAP actually proves fusion doesn't happen.
Agreed, this is not what we wanted to say, we clarified. The FRAP experiment just suggests that it is not liquid-like.
It is technically difficult to laser ablate individual or only subsets of deuterosomes...
This is what was done but anyway, FRAP does not firmly show that deuterosome compartments are not liquid-like as we now precise.
(4) How do you fix your cells for expansion as you have no preservation of cytoplasmic microtubules? You are saying that there is a "nest" of MTs but beta tubulin ONLY stains the cilia and centriole - why is this? Tyrosinated tubulin on regular confocal shows strong cytoplasmic staining. See Figure 3.
Cytoplasmic microtubules do not preserve well through the expansion process. We did try a few different fixations and pre-extraction methods but they come at a trade-off to preserving centrioles. i.e. we could either preserve cytoplasmic tubes or centrioles but not both with the same processing method.
(5) "PCNT puncta partially overlap with centrin (Figure 3 Supplementary 2C). At this stage, PLK4, the master regulatory kinase, and SAS6, one of the first centriolar components are either absent or present as small foci within the cloud, often on the wall of the parent centrioles (Figure 3B-C)." some arrows to highlight this would be useful - difficult to see?
We have tried to make arrows on what is now Fig. 3 Supplementary 1 G, but there is to many CENTRIN colocalizing with PCNT. We have enhanced the contrast of the merge to make it more visible.
(6) Figure 3I legend - what are the arrows pointing at? Yellow and white on inserts? ". Around the same time as tubulin, centrin is also recruited to procentrioles (Figure 3I). This stage is probably the stage that we previously documented as A"
However you see centrin at DEUP1 foci in D, and you don't show any eg. SAS6 or PLK4 positive DEUP1+ structures lacking centrin specifically, centrin seems to be present on all the procentrioles in Figure 3I. Did I miss it where you show centrin negative procentrioles in the cloud?
Fig. 3I (now Fig. Supplementary 1J), yellow arrows are pointing at centrioles with non-acetylated MT while white arrows point at acetylated MT. This is now indicated in the legend.
Regarding CENTRIN, it is present as a diffuse staining around the centrosome since the very beginning of amplification (now in Fig. 3 Supplementary 1A with different contrasts), in addition to compose the parental centrioles. This staining can therefore overlap with DEUP1 staining when DEUP1 appears (Fig. 3 Supplementary 1B, E) but not necessarily. In live we observe that CENTRIN and DEUP1 foci can move independently at early stages (Fig. 2 Supplementary 1B, video 2). This is later on, as shown now in Fig. 3 Supplementary 1J (previously Fig. 3I), that procentrioles are all strongly positive for CENTRIN.
A new paper (Laporte et al., Cell 2024) recently showed that the recruitment of CENTRIN on duplicating procentrioles first occurs at the distal end, visible by a small dot, and then appears gradually at the level of the inner scaffold when procentriole reach 160nm, the stage where POC5 appears, which corresponds to the A-to-G transition in our MCC progenitors (Al Jord et al., 2014). One can therefore consider that the same is happening in our cells, and that, with the CENTRIN cloud, we have difficulties to detect the distal CENTRIN dot. We have changed the text to add this reference and discuss CENTRIN apparition in MCC procentrioles.
(7) " The DEUP1 asymmetry previously described at the centrosomal daughter centriole (Al Jord etal., 2014) becomes visible in some cells during the cloud stage (Figure 3B, N; Figure 3 Supplementary 2B) and in a majority of cells" difficult to see - maybe enlarge and single channel from Figure 3F-H in the supplemental Figure 3 to emphasise this?
We have either changed the pictures or the contrast to be more representative with the quantifications. This is visible in Fig. 3A, D, E, G; Fig3. Supplementary 1E and now using correlative light and EM in Fig. 2 Supplementary 2, 3, 4 and Fig. 4 Supplementary 3-4. One has to look carefully at the daughter centriole (marked “dc”). We have not zoomed in since previous manuscript have already described this at later stages with bigger deuterosomes. You can refer to main or supplementary figures in previous manuscripts (Al Jord 2014, Khoury Damaa 2024) where serial sections span the entire deuterosomes and daughter centrioles and show, with nanometric resolution, that both strutures are frequently sticked to each others on tens of nanometers.
(8) Do you have videos of DEUP1 oscillations with nocodazole to show a lack of oscillations?
We have now added videos of DEUP1 oscillations under nocodazole and dynapyrazole treatments.
(9) "In addition, co-staining of centrioles and nuclear pore proteins show a tight colocalization(Figure 6 Supplementary 1B)." I see the colocalisation in panel 1 but less obvious with panel 2 maybe have some more zoomed in panels and some quantification of the colocalization? Is it more striking at the G stage than the D stage?
We edit and change the phrasing as ‘colocalization with NPC’ is not the good term. There is too many centrioles and NPC, they cannot do otherwise than colocalize… What we want to say is that there is a tight connexion with the nuclear envelope. This is why we present a single z, to show that centrioles and NPC are on the same z-plane. This is also clearly visible for centrioles that are loaded on deuterosomes that are around the nuclear membrane in the XY plane. We also added a video to show an entire z-stack of this kind of staining.
(10) "Indeed, SAS6 normally disappears from procentrioles when centrioles are docked, just beforeciliation (Al Jord et al., 2014). This suggests that centrioles were able to degrade SAS6, a process also dependent on APC/C (Strnad et al., 2007), but failed to disengage from deuterosomes." Figure 6 Supplement 1E-F - are you sure it wasn't that Sas6 wasn't loaded correctly at the earlier stage and so is reduced recruitment rather than premature disengagement of Sas6? If it is indeed premature disengagement of Sas-6 - what about CP110 - does the CP110 get loaded and is it still present in noco treated cells arrested in the D phase?
We do not observe SAS6-negative procentrioles on deuterosomes at G-stage but only on deuterosomes in D-stage cells (cells with partly disengaged procentrioles). This is why we hypothesize that, because of the long duration of D-stage and knowing that SAS6 is finally degraded at the end of amplification (Al Jord et al., 2014), we are in the presence of cells where SAS6 has been degraded but where centrioles did not manage to disengage. This is now clarified in the text.
(11) Can you track deuterostome splitting live? Maybe not enough spatial or time resolution?
One has to monitor in 3D (multiple z because deuterosomes move a lot), 2 colors, high temporal resolution (dt=2-5’; to be able to track a single deuterosome), and long duration (deuterosomes are sometimes touching each other and then moving away, giving the impression that they split). This eventually leads to the bleaching of the mRuby fusion protein… We have put an example of what we think is a deuterosome splitting in Fig. 6E (former Fig. 7D). But we decided to finally monitor with low temporal resolution (dt=40’) to avoid photobleaching, and analyze numerous deuterosomes and cells to quantify the number and size of deuterosomes over time in single cells.
(12) The MT nodes - can you segment the tyrosinated MTs and define nodes and then quantify theDEUP1 presence on them?
Please see answer to reviewer 1 regarding this point.
(13) Figure 8 supp 1 (E): Representative XY distribution of CEN2-GFP+ centrioles at the end of migration (Sas6 negative) in brain MCCs treated with DMSO, Nocodazole 1µM and 5µM (48h). Scale bar, 5µm Bit more detail on how you define fully migrated vs still migrating centrioles in z. You say you are using Sas-6 negativity to define fully migrated cells in the legend, yet you say noco treatment leads to premature sas-6 negativity, and yet the apical migration takes longer upon noco treatment?
Nocodazole does not lead to premature SAS6 negativity but to a partial disengagement which lead to SAS6 negative “mature” centrioles being still connected to deuterosomes. We define complete migration when all the centrioles are on the apical side of the nucleus. We now clearly define what “apical” migration stands for in the main text and changed the pictograms in Fig. 8G to clarify this.
(14) Figure 8H and video 18 - it isn't obviously clear to me that the noco-treated cells are "more erratic" or how you decide what counts as apically migrated successfully. How do you control for drift in z? Can you track individual centrioles as you did in untreated and define what is "erratic about their movement?
Erratic means that the centrioles are moving away from each others, and back, in a non-predictable way, instead of migrating up and gathering. The drift in z of the whole cell is visible because there is always some centrioles, that are apically located at the beginning, that remains on the apical membrane, probably because they are already docked.
We have indeed followed the centrioles individually in the nocodazole condition. However, in the control, the XYZ coordinates of one of the centrioles of the centrosome, which normally don’t move, are substracted to the coordinates of all the other centrioles as explained in the method section. This allows to have a subcellular reference, and to circumvent the movements of the cell, which are non-negligible at all at this timescale. In the nocodazole treated cells, the centrosomal centrioles share the erratic movements of the other centrioles and can migrate up and down, which exclude them as a reference. Since the nucleus is also moving a lot, we were left with no reference point.
(15) Figure 8 supplement 1E can you quantify the final area of centriole patch in XY upon noco treatment?
It was in main Fig. 8J and is now in Fig. 8 Supplementary 1F.
(16) Figure 8J legend- MBB is never defined as an acronym.
Thank you for pointing this.
(17) Define what is the frequency and how is it calculated - Figure 8J.
This is the MBB patch area in µm<sup>2</sup>
Text edits:
(1) "Altogether, these results suggest that, in this non-tissue-specific proxy of MCC progenitors, microtubules organize the onset of centriole amplification in the pericentrosomal region."
Sentences have changed.
(2) "Increasing the temporal resolution to 5-15s reveals that DEUP1+ foci observe an exhibit oscillatory dynamics to at the centrosome (Figure 2E, colored arrows, Video 3, 5/10 cells observed for 1-4min)."
Sentences have changed.
(3) "stage procentrioles were involved in this perinuclear migration and distribution. In fact, this dynamic is reminiscent of the centrosome migration that occurs during the G2-to-M progression in cycling cells in preparation for mitotic spindle organization. In cycling cells, this" Grammar - maybe change to "stage procentrioles were involved in this perinuclear migration and distribution. This is reminiscent of the centrosome migration that occurs during the G2-to-M".
Sentences have changed.
(4) "We then wondered whether these microtubule-dependent dynamics was were required for an efficient subsequent centriole disengagement during the following D-stage."
Sentences have changed.
(5) "Then, monitoring tens of disengagement movies, we identified a transient stage during which disengaging procentrioles redistribute isotropically in the 3 dimensions, along the nuclear membrane (Figure 6A, 4:30, Video 7) before losing its contact to migrate to the apical surface (Figure 6A, 6:30 to 14:00)."
Sentences have changed.
(6) Discussion: "Since pioneer electron microscopy studies on basal body production in quail oviduct MCC 35 years ago (Boisvieux-Ulrich et al., 1987, 1990; Boisvieux-Ulrich et al., 1989), this work is the first to assess the role of microtubules in the now finely described centriole amplification process. This"
Sentences have changed.
(7) "Using live imaging on brain MCC, we highlight the existence of a nest composed of DEUP1, PCNT and Centrin2, pre-assembled before the onset of centriole amplification onset."
Sentences have changed.
(8) "Recently, formation of DEUP1 pure condensates in solution as well as FRAP experiments after overexpression of DEUP1 in MCC progenitors suggested that deuterosomes where are not liquidlike structures (Yamamoto & Kitagawa, 2019). Consistently, we never observed fusion events of DEUP1."
Sentences have changed.
(9) "This reminds is reminiscent of the centriole-to-centrosome conversion occurring at the G2-M transition followed by the associated microtubule dependent nuclear migration of new centrosomes at mitosis onset (Agircan et al., 2014)."
Sentences have changed.
(10) "Following individual trajectories requires high resolutive resolution spatio-temporal live imaging while avoiding excessive light exposure which disturbs centriole migration (Boudjema et al., 2024)."
Sentences have changed.
(11) "Using high temporal resolution microscopy, we further identify that individual dynamics is are complex and can be splitted between divided into the baso-apical migration, where centrioles move in a processive and more..."
Sentences have changed.
Reviewer #3 (Recommendations For The Authors):
(1) Growing MEF-MCCs on micropatterns has successfully mimicked the dynamics of centriole amplification in brain MCCs, allowing the authors to study the spatial origin of procentrioles. Since this is a powerful system, a more quantitative description of the system will be informative and beneficial for future studies. For example: What is the efficiency of this system? Do the cilia that form in MEF-MCCs motile?
The system of MEF-MCCs has been described in a previous paper from the Kintner lab. It seems that growing the MEF-MCCs on micropatterns did not ameliorate the ciliation which is partial, probably due to the absence of an apico-basal polarity.
(2) Figure 2: The analogy drawn by the authors between DEUP1 oscillatory dynamics and centriolar satellites is intriguing. In early amplifying cells within the cloud, do these DEUP1 structures co-localize with the satellite marker PCM1?
We have added immuno stainings of PCM1 in mRuby-DEUP1 / CEN2-GFP cells in Fig. Supplementary 2E. Within the centrosomal cloud, DEUP1 colocalizes with PCM1. Interestingly, this PCM1 concentration at the centrosome is dependent, at least in part, on dyneins. Then, PCM1 can localize around the deuterosomes, but it is never colocalized with deuterosomes (not shown). This is also showed by immuno-EM in Zhao et al., 2019. Although it was shown that PCM1 is a proximity interactor of DEUP1 (called ccdc67 at that time) by Firat-Karalar et al., 2014., absence of PCM1 staining on deuterosomes does not favor the hypothesis of PCM1 and DEUP1 being part of the same entities. One could hypothesizes that DEUP1 is transcribed locally within the satellites, explaining the colocalization of the 2 proteins and the + BioID results, and then form PCM1negative deuterosomes.
(3) The authors propose a physical link between deuterosomes and centrosomes based on their oscillatory behavior. How are the oscillatory dynamics of DEUP1 affected by nocodazole treatment or inhibition of microtubule motors (i.e ciliobrevin treatment)?
These oscillations are inhibited by nocodazole (Fig. 4D). They are also inhibited by dynapyrazole (Fig. 4D). We never succeeded in having a nice disruption of the Golgi apparatus with ciliobrevin and therefore we did not used it.
(4) In addition to nocodazole treatment, it would be important to determine the consequences of microtubule stabilization by taxol and inhibition of microtubule motors during critical stages of centriole amplification where microtubules are reported to play a role for the first time in this manuscript. Another interesting area of investigation will be to study the extent to which microtubule PTMs contribute to these processes.
We now blocks dyneins during the different stages of amplification. The results are in main and associated Fig. 4, 5, 7, 8. The role of microtubule PTM, is not in the scope of this manuscript.
(5) Describing microtubule dynamics along with Centrin/DEUP1 dynamics will be informative in assessing whether these structures associate and/or move along microtubules? Have the authors performed their imaging experiments with SIR tubulin?
Yes, we have tried hard! But we have encountered different obstacles:
3-color video microscopy is phototoxic,
siRTubulin is bleaching very rapidly
The density of microtubules in MCC makes the observation hardly informative
(6) Figure 5: The role of PLK1 in centriole-centrosome conversion and generation of multiple MTOCs can be tested with a PLK1 inhibitor for further confirmation.
We have also tried but inhibiting Plk1 blocks the A-to-G and G-to-D transitions so it was not possible to uncouple the role of Plk1 in stage transitions versus centriole maturation.
(7) Figure 6: The tight co-localization of nuclear pore proteins with centrioles poses questions about the role of nuclear pore proteins or other nuclear proteins that are associated with centrioles during centriole disengagement and migration. Considering the existing literature on centrosome-nucleus attachments, can there be a way to test this question within the scope of this manuscript?
We have tried to deplete Nup133 but it’s killing the cells. Our additional experiments now show that the nuclear migration of centrioles during G-stage is dynein dependent, reinforcing the parallel with centrosome migration in prophase. We also added results from our scRNA sequencing (Fig. 5 Supplementary 1) showing that some key players of centriole migration to the nuclear membrane are conserved in the MCC cell cycle variant, and expressed with a comparable dynamics as to the canonical cell cycle.
(8) Figure 8: Manually tracking a subset of migrating centrioles to define their dynamics during centriole migration and docking provides valuable analysis for determining the molecular mechanism of these processes. In addition to microtubules, does actin contribute to this process? Since centrioles eventually migrate to the apical side in nocodazole-treated cells, there should be other molecular players involved in this process.
We did block actin polymerization but we found that the different stages were affected and that it would be better to dedicate a whole manuscript on the role of actin during each stage of amplification. We discuss the migration mechanism, and the putative role of actin, in the discussion.
(9) The legends for Supplementary Figures 1 and 2 in Figure 3 are mixed and need correction.
Figures have been remodelled.
(10) In Figure 3P, the term "PLK4+" is labeled in bright green, which is not clearly visible. It maybe beneficial to change the color of this label for better visibility.
We have tried to correct this.
(11) Figure 6F quantifies "% tethered flowers" on the nuclear membrane. When quantifying, is the3D localization of DEUP1 flowers in both DMSO- and Noc-treated cells considered? A flower may appear to be on the nucleus in 2D, but it could be detached from the membrane in a 3D view.
The quantifications are done in 3D. However, flowers that are below or above the nucleus are not quantified since the space is confined and the resolution in z to small to see whether they are connected or not. This is now precised in the legend.
Before the editors proceed with an updated assessment, they've requested that we pass on some of the comments that have arisen as part of the evaluation of your revised manuscript. They feel that these concerns should be addressed before we proceed with issuing a formal assessment and publishing the revised Reviewed Preprint:
We thank the reviewers and the editors for the corrections and insighfull comments. We apologize for our delayed answer and hope our corrections in the main text and some of the figures will give them satisfaction.
The revised manuscript is greatly improved with nice new data regarding the role of microtubules. It also has changed quite a bit including the title. The new focus is on the cell and centriole cycle variants in MCC. While this helped to focus the study, there remains an important issue related to the interpretation of the data and the proposed 2-in-1 cycle model. Before providing the final updated assessment, we ask you to address the following points (which were raised already in the first round of review): The manuscript still contains statements that are not aligned with published work and the current view in the field regarding the timing of events during canonical centriole biogenesis. These timings are in conflict with your model that 2 centriole cycles are "superposed" in the MCC cell cycle variant, as currently presented. An alternative straightforward interpretation would be that multiciliogenesis uses an accelerated centriole duplication cycle where key steps occur concomitantly or in short succession instead of being separated by mitotic divisions as in the canonical cycle.
We do agree with the acceleration of all steps into only one cycle, this is actually what we think we have proposed. When correcting our confusions as regard to centriole-to-centrosome conversion (as explained below) and putting the events in a scheme, this reveals that the events of the two canonical cycles nicely superpose, both in term of molecular composition and dynamics (corrected Fig. 9). We therefore maintain that the null hypothesis is that the acceleration is done through a superposition of events that; although driven by the same molecular machinery, are normally occuring in two consecutive cell cycle. We explain ourself briefly in two paragraphs, before answering point by point to the questions of the reviewers.
As regard to centriole-to-centrosome conversion:
We thank the reviewer for pointing out that we used “MTOC conversion” for what is normally called “centrosome maturation”. We have removed the term “centriole-to-centrosome conversion” during the first round of revision but we now realize that “MTOC conversion” leads to the same misinterpretation as regard to the literature on centriole duplication.
The reviewer asks us to refer to the work of the Tsou lab (Wang 2011, reference now added in the manuscript) showing that daughter centrioles are “modified” (e.g. recruit PCM, become competent for MT nucleation and duplication) during late M/early G1. This “centriole-to-centrosome conversion” can’t occur for our procentrioles at this stage since they are not even born during the mitosis that precedes MCC differentiation. Also, in our cells, such modification does not include the capacity to become competent for duplication since we know that procentrioles become basal bodies without making any round of duplication (Al Jord et al., 2014).
Also, we have not done the experiments to tackle the question on when our centriole become “modified-like”. What we can say is that during A-stage, they become progressively positive for PCM (Fig. 5 Supplementary 2) and a weak signal shows that some MT are seen emerging from them (Fig. 5 and Fig. 5 Supplementary 2, and see point by point answer).
What we do see is that, at the A-to-G transition, they increase their PCM recruitment, show clear and strong MTOC ability (sometimes as strong as the centrosomal centrioles), and that this is associated with migration and separation of centrosome/deuterosomes around the nuclear membrane (Fig. 5). We therefore connect this to what occurs at the G2/M transition which is an increased recruitment of PCM protein, an increased ability to nucleate MT, associated with centrosome migration and separation at the nuclear membrane. Since this process in the canonical cell cycle is called “centrosome maturation”, we therefore should refer to this term in our study. However, centrioles in the MCC variants are not organized in centrosomes, so we now compare what we see to the “centrosome maturation” of the canonical cell cycle with an associated reference (Joukov et al., 2018), but name it “centriole maturation”.
We have modified the text (track changes visibles) and the schemes (Fig. 5, Fig. 5 Supplementary 1 and 2, Fig. 9, Fig. 9 Supplementary S1; new versions uploaded) accordingly.
As regard to 1.5 or 2 cell cycles
Except for the “MTOC conversion” that we have now changed, as explained above, we think our work does suggest (depicted on Fig. 9) what the reviewer states for centriole duplication: “In the current view, centriole biogenesis starts in early S, elongation proceeds through G2/M and by early G1 it is complete. During M/early G1 centrioles disengage and newly formed daughters recruit PCM (centrosome conversion). Then these centrioles go through another complete cell cycle and when they reach early G1 again they have acquired DAs and SDAs. Key here is that biogenesis and disengagement/centrosome conversion are separated by the first mitosis (ensuring duplication occurs only once), and acquisition of DAs and SDAs is separated by another mitosis (ensuring that cells only form a single cilium)”.
We feel that going from early S to a G1 phase, after 2 mitosis, is what one can call “2 cell cycles”. One of the paper that inspired us a lot when studying how the cell cycle machinery can drive centriole amplification in MCC is a paper from Jadranka Loncarek team (Kong et al., 2014) where they also state that “nascent centrioles gradually mature through 2 cell cycles”. Very interestingly, in this study they show that when they enhance Plk1 activation, they could erase centriole age and new procentrioles are able to recruit PCM and appendages within only 1 cell cycle, without mitotic progression, like what we see in MCC. We have added the reference in our discussion.
Point by point answer
(1) Original work on canonical centriole disengagement and centriole-to-centrosome conversion should be cited (e.g. PMID: 16862117, PMID: 21576395)
As explained earlier, we used the wrong term since the begining. We do not speak about the centriole-to-centrosome (nor MTOC) conversion since we do not test when centriole modification (Wang et al., 2011) occurs in the MCC cell cycle variant. We know that PCNT is present on the procentrioles during A-stage (as shown in Fig. 5 Supplementary 2B), but we do not know when it is recruited (UExM did not work properly with this antibody). We quantify a weak MT staining in regrowth experiment during A-stage and see that procentrioles can be connected to MT in both brain MCC and MEFs (as shown in Fig. 5D, E for brain MCC and Fig. 5 Supplementary 2F for MEFs) , but we do not know when during A-stage they become competent for nucleation. We therefore did not speak about this process that we do not document. What we clearly document/quantify is the enhanced MT nucleation capacities at the A-to-G transition, concomitent with the nuclear migration (easily defined with Cen2-GFP or GT335 stainings) and that we compare to centrosome maturation occuring at the canonical G2/M transition.
(2) The authors state in several places that canonical centriole formation and maturation takes two iterations of the canonical cell cycle. This is imprecise. Based on the above work and work by others, the broadly accepted view is that it takes 1.5 cell cycles. This difference matters for the final proposed model (see below). Reviewed e.g. here: PMID: 20869612; PMID: 30601682
Our answer is in the preamble.
(3) "Centriole maturation cycle superposes with centriole elongation cycle in the MCC cell cycle variant": Your description of the canonical cycle differs from the current view in the field. In the current view, centriole biogenesis starts in early S, elongation proceeds through G2/M and by early G1 it is complete. During M/early G1 centrioles disengage and newly formed daughters recruit PCM (centrosome conversion). All this occurs in 0.5 cycles. Then these centrioles go through another complete cell cycle and when they reach early G1 again they have acquired DAs and SDAs (total of 1.5 cell cycles). Key here is that biogenesis and disengagement/centrosome conversion are separated by the first mitosis (ensuring duplication occurs only once), and acquisition of DAs and SDAs is separated by another mitosis (ensuring that cells only form a single cilium).
(4) Fig 5A, B and Fig. 9
(a) Are 2 separate figures needed for the model? They seem redundant.
We find it easier not to wait Fig. 9 to have the first part depicted.
(b) The model shows loss of SAS6 throughout G1, but this already occurs during M/early G1
Thanks. It was already ok in Fig. 9, we have modified for Fig. 5.
The model shows "MTOC capacity/conversion" during S phase, but this occurs during early G1
Thanks a lot, as explained earlier, we used the term MTOC conversion occurring in G1 for what is normally called centrosome maturation occurring in G2/M, as explained earlier. We do not speak anymore of MTOC conversion since we have not tackled this question (explained above). We have therefore removed MTOC conversion in the texts and the schemes and replaced it by “centrosome maturation” for the duplication cycle, and by “enhanced MT nucleation capacity” for the MCC cycle. To be clearer and schematize that procentrioles are competent for MT nucleation before G2/M or A/G transitions, we have added some MT nucleated from G1 procentrioles during the canonical cycle, and from late A-stage procentrioles during the MCC cycle.
The model shows disengagement only in the second M phase, but this occurs already at the first M phase, directly following centriole biogenesis, right before centosome conversion.
This is a big edition error in both Fig. 5 and 9. Of course the daughter centriole disengage during the first M-phase. This has been changed. Thanks a lot for spotting it. This, however does not contradict the hypothesis of superposition.
We also added the acquisition of distal appendage which was written in Fig. 5 but not in Fig.
9 for duplication during the second M-phase.
When the correct timings are incorporated in the figure, the proposed superposition of two cycles is not an accurate description of the events. Instead, your data seem consistent with a model where MCC incorporates all steps in one cell cycle variant that lacks mitoses, so that disengagement and MTOC conversion occur together with centriole elongation, followed immediately by acquisition of DAs and SDAs.
We do agree with the acceleration of all steps into only one cycle, this is actually what we tried to propose. When putting the events in a scheme, this reveals that the events of the two canonical cycles nicely superpose, both in term of molecular composition and dynamics (Fig. 9). We therefore maintain that the null hypothesis is that the acceleration is done through a super opposition of events that; although driven by the same molecular machinery, are normally occurring in two consecutive cell cycle. This is notably consistent with the findings of Kong et al., 2014 cited previously.
(5) While all reviewers felt that there was no need to introduce the new term "nest", they leave it to the authors to keep it. However, the authors may want to consider that the term is still not introduced and explained properly, which may confuse readers. For example, while this section reads like an introduction to the term: "Correlative DEUP1 live-imaging and EM highlights the existence of a pericentrosomal "nest" in brain MCC", the term is already used two times before without explanation. The first mentioning is at the beginning of the results section and is followed by citations, which gives the impression that these studies describe the nest, which is not the case.
The first mention of “nest” is in the end of introduction resuming the findings of the paper where the term is in the following context: “we found that centriole amplification emerges in a pericentrosomal “nest” concentrating core centriole/deuterosome elements”. We looked at nest definition in the Collins Dictionnary : “a structure or other place where creatures, esp. birds, give birth or leave their eggs to develop”, we felt this was clear. We added quotation marks around the term nest.
Then, the result section opens with this sentence: “The origin of amplified centrioles in MCC remains controversial. Some live imaging experiments and electron microscopy suggest that the centrosome could constitute a nest for centriole and deuterosome biogenesis (Al Jord et al., 2014; Kalnins et al., 1972; Mori et al., 2017), but others have proposed that procentriole-loaded deuterosomes emerge independently from the centrosome location, all over the cytoplasm (Nanjundappa et al., 2019; Sorokin, 1968; Zhao et al., 2013, 2019).”. Here, the term nest is again used as a place of birth for centrioles and deuterosomes which is what is actually proposed in these papers. First, Kalnins el al., in 1969 (we made an error on the reference date, this has been changed), resume in their abstract “This observation suggests that all of the clusters may form initially in close association with the diplosomal centrioles”. Then, not to mention Al Jord 2014 which comes from our lab, the title of Mori et al. is “Cytoplasmic E2f4 forms organizing centres for initiation of centriole amplification during multiciliogenesis”, and in the paper, they show that E2F4 accumulates at the centrosome. This is now also proposed by collaborators for MCIDAS (Lu et al., 2025). We feel that these references, which are often omitted, are appropriated at this location.
Then we continue with: “To test whether microtubules drive the organization of a centrosomal nest from which procentrioles emerge”, which keeps the notion of the place of birth.
Then the title "Correlative DEUP1 live-imaging and EM highlights the existence of a pericentrosomal "nest" in brain MCC" arrives. In this section we first speak about a pericentriosomal cloud on which we zoom in using CLEM, to then conclude at the end of the section “Altogether live imaging mRuby-DEUP1/CEN2-GFP during early A-stage suggests that core deuterosome and centriole components are concentrated in a primordial cloud around the centrosome, which constitutes a nest where centrioles and deuterosomes concomitantly form before they move away from the centrosomal region (Fig. 2F)”.
Finally, we begin the discussion section regarding the nest by: “We named this transitory compartment a “nest” since deuterosomes and procentrioles emerge specifically in this region and grow while moving away from it.”
During the first revision, we tried to make it clearer. If this is still not the case after and the reviewer has another proposition of definitions/phrasing, we will be glad to consider it.
As replied to the other reviewer, the term “nest” does not need to be retained as a new terminology. It is just a way for us to identify the transitory region and to best define one of its function/characteristic which is to host the birth of new deuterosomes and centrioles.
The following comments from Reviewer #3 may also provide further context regarding the editors' remaining concerns:
The authors have done an excellent job addressing the points I raised overall, and the revision is substantially improved in focus and clarity. That said, some concerns raised by other reviewers, particularly regarding terminology and statistical analysis, could have been addressed more fully. One issue remains insufficiently resolved. Several quantitative analyses (for example Fig. 5C and 5E) still appear to rely on pooled single-event measurements collected across three independent experiments. This approach can overstate statistical significance. The authors indicate in their rebuttal that they use chi-square tests to compare proportions and to justify pooling across replicates. However, I am not convinced this addresses the issue for the intensity-based and single event distributions shown in the panels specified above. I recommend that these key analyses be represented with biological replicates shown explicitly (superplot-style, with replicates distinguished).
Our reply was for the comparison of proportions and not the intensity-based and single event distributions shown in the panels Fig. 5C and Fig. 5E. We have now changed our plots to represent biological replicates explicitly (superplot-style, with replicates distinguished). As for the statistical analysis: we evaluated differences in marker intensity between A-stage and G-stage samples using a linear regression model, with stages as the main effect and replicate as a fixed covariate, to account for batch variation. Statistical significance was assessed using Type II ANOVA.
Separately, I continue to feel that some newly introduced terminology (for example, the "nest") may not be necessary at this stage. It may be sufficient to describe these structures and focus on their spatiotemporal behavior, composition, and measurable features, rather than assigning new names. Having read the authors' response, I understand that they would like to retain this terminology, which is acceptable; however, it may not be readily adopted by the field.
The term “nest” does not need to be retained as a new terminology. It is just a way for us to identify the region and to best define one of its function/characteristic which is to host the birth of new deuterosomes and centrioles.
Minor correction (remove "in MCCs" part from the following sentence):
In MCC, PCM1 depletion alters deuterosome formation and centriole production in brain and airway MCC (Hall et al., 2023; Zhao et al., 2021).
Done
