Author response:
The following is the authors’ response to the original reviews.
Public Reviews:
Reviewer #1 (Public review):
The central pair apparatus of motile cilia consists of two singlet microtubules, termed C1 and C2, each of which is associated with a set of projections, referred to as the C1 and C2 projections. Each projection comprises multiple distinct structural domains, designated a, b, c, and so on. Biochemical studies combined with genetic analyses in Chlamydomonas identified three proteins as the major components of the C2a projection, and subsequent cryo-EM studies confirmed these findings.
In this paper, the authors aim to study the homologues of these three proteinsCCDC108/CFAP65, CFAP70, and MYCBPAP/CFAP147-using knockout mouse models. Biochemical and cell biological analyses demonstrate that, as in Chlamydomonas, these proteins are components of the C2 projection and form a complex that depends on the presence of each other. In addition, the authors use affinity purification to identify two previously uncharacterized proteins and show that they are central pair apparatus proteins that associate with the aforementioned complex. Knockout mice lacking any of the three core proteins exhibit phenotypes consistent with primary ciliary dyskinesia (PCD).
Overall, the manuscript is clearly written, and the data are convincing and support the authors' conclusions. However, given the previous findings in Chlamydomonas, this work provides limited conceptual advances to the field. Nonetheless, it represents a useful and well-documented resource for understanding the conserved organization of the central pair apparatus in motile cilia. It will be of interest to cell and developmental biologists, biochemists, and clinicians studying and treating human ciliopathies.
We sincerely appreciate the positive feedback on our work.
Reviewer #2 (Public review):
Summary:
This manuscript investigates the protein composition and functional role of the C2a projection of the central apparatus (CA) in vertebrate motile cilia. Using three knockout mouse models (Ccdc108, Mycbpap, and Cfap70), the authors demonstrate that these genes - homologs of Chlamydomonas FAP65, FAP147, and FAP70 - are required for normal motile cilia function in ependymal and tracheal multiciliated cells. Specifically, the authors show that:
(1) Knockout mice for each gene exhibit primary ciliary dyskinesia phenotypes (hydrocephalus and sinusitis), accompanied by abnormal ciliary motion and reduced ciliary beat frequency.
(2) CCDC108, MYCBPAP, and CFAP70 physically interact and localize to the axonemal central lumen, consistent with the C2a projection.
(3) Loss of any one of these proteins destabilizes the others and disrupts CA integrity in a tissue-specific manner.
(4) ARMC3 and MYCBP are C2a-associated proteins.
Strengths:
(1) Clarity: the results are presented in a coherent sequence that facilitates understanding of both the rationale and conclusions.
(2) Genetic rigor: three independent knockout mouse lines that exhibit consistent motile cilia phenotypes provide in vivo support for the proposed role of these proteins.
(3) Integration of structural and functional analyses: combination of ultrastructural (TEM) and immunofluorescence data with CBF measurements provides convincing correlation between structural defects and impaired ciliary function.
(4) Mutual dependency model: reciprocal destabilization of CCDC108, MYCBPAP, and CFAP70 supports their interdependence in the C2a assembly.
(5) Expansion of the vertebrate C2a proteome: the identification of ARMC3 and MYCBP as C2a-associated proteins provides a foundation for future mechanistic studies.
We appreciate the valuable comments and pertinent suggestions, which provide important guidance for revising and improving this manuscript.
Weaknesses:
(1) Mechanistic depth: the data show a convincing correlation between C2a and ciliary function, but the cell type-specificity of CCDC108, MYCBPAP, and CFAP70 knockout effects is underdeveloped. This is an interesting observation that raises mechanistic/structural questions not addressed in the study, such as what is the role of C2a in CP nucleation, maintenance, or mechanical stabilization? Is C2a composition different in different cell types?
We appreciate this comment. Based on current knowledge, loss of proteins essential for CP nucleation, such as WDR47 and KIF27, typically causes severe CP loss defects [1,2]. However, only mild CP-loss defects were observed in Ccdc108, Mycbpap, or Cfap70 knockout (KO) mouse ependymal cells (mEPCs) serum-starved for 10 days (Figure 2E, F), indicating that C2a proteins are more likely to play a role in CP maintenance or mechanical stabilization. In the revision, we tested this hypothesis by examining the effects of C2a loss on CA ultrastructure in Ccdc108 KO mEPCs serum-starved for 5 days. The percentage of axonemes with defective CA decreased further (Figure 2—figure supplement 1C, D). These results further confirm that C2a proteins play a role in CP maintenance or mechanical stabilization but not in CP nucleation. We have included these results and expanded the related discussion in the revised manuscript.
To assess whether C2a composition differs across cell types, we performed co-immunoprecipitation using lysates from mouse trachea and mEPCs. We found that, in both tracheal and mEPC lysates, CFAP70, ARMC3, and MYCBP were co-immunoprecipitated with MYCBPAP (Figure 6—figure supplement 6A, B), indicating that at least the C2a core components are conserved in vertebrate motile ciliated cells. We have included these results in the revised manuscript.
(2) Cell model choice: co-immunoprecipitation was performed using mouse testis lysates. While this is a reasonable source of CA proteins from flagellated cells, the functional analyses in this study focus on ependymal and tracheal multiciliated cells. It would therefore be helpful for the authors to clarify the extent to which these interactions are expected to be conserved across ciliated cell types, and to discuss potential tissue-specific differences in CA assembly.
We thank the reviewer for the insightful suggestion. Following the reviewer’s suggestion, we performed co-immunoprecipitation using lysates from mouse trachea and mEPCs. We found that, in both tracheal and mEPC lysates, CFAP70, ARMC3, and MYCBP were coimmunoprecipitated with MYCBPAP (Figure 6—figure supplement 1A, B), indicating that at least the interactions among the C2a core components are conserved in vertebrate motile ciliated cells. We have included this result in the revised manuscript.
(3) Statistical analysis: the manuscript states "Statistical significance was defined as P < 0.5", which is likely a typo, but should be P < 0.05. In general, the statistical methods require more clarification. In several figures (e.g., 2B, 2D, 5J, 5K), multiple knockout genotypes are compared with WT, yet unpaired t-tests are reported. When more than two groups are analyzed, multiple pairwise t-tests inflate Type I error unless appropriately corrected; a oneway ANOVA with post hoc comparisons (e.g., Dunnett's test for WT-referenced comparisons) would be more appropriate. Furthermore, the analysis of ciliary movement modes (Figure 2D) involves categorical data, for which a t-test is not statistically appropriate. These comparisons could instead be evaluated using chi-square or Fisher's exact tests. Addressing these issues is important to ensure accurate statistical inference.
We thank the reviewer for identifying the error and for their suggestions on the statistical analysis. We performed a one-way ANOVA with Dunnett’s test in Prism to re-evaluate the differences between WT and each KO sample. In the revised manuscript, we have updated the statistical results and revised the Methods section.
(4) Methods section: does not sufficiently describe how image-based quantifications were performed. For example, the criteria used to define cilia number, basal body number, and rotational beating are not specified, nor is how CBF measurements were analyzed. The authors should also provide details regarding analysis software and imaging parameters used (and whether they were kept constant across genotypes).
We apologize for omitting a detailed description of image-based quantifications. For counting cilia or basal bodies, cells were immunostained with acetylated α-tubulin and CEP164 antibodies to label cilia and basal bodies, respectively, and imaged using 3D-SIM. Using these super-resolution images, cilia and basal bodies were counted in each multiciliated cell. With highspeed live-cell imaging, ciliary movements were recorded and analyzed using ImageJ. mEPCs in which the majority of motile cilia displayed rotational motility were considered ‘cells with rotational cilia’. The CBF of each cilium was calculated from the total time of 10 beating cycles. In the revised manuscript, we have included these details in the related figure legends and the methods section.
Recommendations for the authors:
Reviewer #1 (Recommendations for the authors):
(1) Line 121: "appeared decreased" should be revised to "appeared to decrease."
We thank the reviewer for the suggestion. In the revised manuscript, we have changed the text accordingly.
(2) Figure 2 legend: The statement "Arrowheads indicate the C2 projections" is misleading. The arrowheads indicate the positions of the C2 projections, as the C2 projections are absent at the locations marked by the arrowheads.
We appreciate this comment and have revised the text in accordance with the reviewer’ s suggestion.
(3) Statistical analysis: For the statistical analyses shown in Figure 2 and the other figures, a t-test was used. In general, a t-test is appropriate for comparisons between two groups. When more than two groups are compared with a single factor, a one-way ANOVA should be used, followed by appropriate post-hoc tests.
We thank the reviewer for pointing out this issue. In the revised manuscript, we have re-evaluated all statistical analyses in Figures 2 and 5 using Dunnett’s test to compare multiple treatment groups (Ccdc108 KO, Mycbpap KO, and Cfap70 KO) with a single control group (WT). We have also revised the corresponding figure legend and methods section.
(4) Docking methodology: In Figures 1A and 5L, the molecular model of the C2a projection (PDB: 7SOM) is superimposed onto the cryo-EM density map. I was unable to find a detailed description of the method used for this docking and would appreciate clarification.
We apologize for omitting a detailed description of the docking methodology. The visualizations in Figures 1A and 5L were generated using the following procedure:
(1) Generation of the Complete CA Density Map: Following the hierarchical local refinement strategy and map integration methods described in previous high-resolution studies of the Chlamydomonas central apparatus (CA) [3,4], we utilized the published density maps of the C2 microtubule and its associated projections (EMD-24191) and the C1 microtubule and its projections (EMD-24207). These maps were aligned and stitched together in UCSF ChimeraX to reconstruct a complete C1-C2 repeating unit of the central apparatus.
(2) Superimposition and Fitting (Figure 1A): To generate the molecular model shown in Figure 1A, the atomic model of the Chlamydomonas C2a projection (PDB: 7SOM) was docked into the corresponding region of the integrated C2 density map [4]. The docking was performed as a rigidbody fit using the "Fit in Map" tool in UCSF ChimeraX, which optimizes the correlation between the molecular model and the cryo-EM density.
(3) Simulation of C2a Loss (Figure 5L): For Figure 5L, we simulated the results of C2a loss observed in our mutation experiments. Using the model established for Figure 1A as a template, we selectively removed the C2a-specific density and the corresponding superimposed atomic model to schematically illustrate the structural consequences of the mutations involved in this study.
We have updated the Methods section of the revised manuscript to include these details regarding structural visualization and docking analysis.
Reviewer #2 (Recommendations for the authors):
(1) Lines 106-107: "frameshift mutation was created by introducing a 458-bp deletion of exons 6-8 in the mouse 107 Cfap70 (ENSMUST00000056073.14) (Figure 1B)". The figure indicates deletion of exons 3-8; please indicate which is correct.
We apologize for the oversight and confirm that the deletion region encompasses exons 3-8 (as shown in Figure 1B). In the revised manuscript, we have updated the text accordingly.
(2) Lines 119-121: "Genotyping at postnatal day 0 (P0) revealed that Ccdc108 KO pups, Mycbpap KO pups, and Cfap70 KO pups were all born at the expected Mendelian ratios; however, the ratio of Mycbpap KO mice at P7 appeared decreased (Figure 1E)". The authors can test whether the genotype distribution changes between P0 and P7 to directly support their claim of postnatal lethality.
We appreciate the reviewer’s comments. The P0 genotyping results were obtained from P0 neonatal mice sacrificed for mEPC culture. Therefore, the P0 and P7 genotyping distributions were from different batches of mice. Re-doing the genotyping distribution analysis would require a large number of mice and considerable time. We hope the reviewer understands the difficulty and allows us to forgo this experiment.
References
(1) Liu, H., Zheng, J., Zhu, L., Xie, L., Chen, Y., Zhang, Y., Zhang, W., Yin, Y., Peng, C., Zhou, J., et al. (2021). Wdr47, Camsaps, and Katanin cooperate to generate ciliary central microtubules. Nat Commun 12, 5796. 10.1038/s41467-021-26058-5.
(2) Park, H., Choi, M., Zhang, Y., Cheung, H.O., Makino, S., Yoshikawa, Y., Qi, H., Liu, Z., Lan, G., Fu, G., et al. (2025). The kinesin-4 protein KIF27 forms a cytoskeletal scaffold at the transi\on zone to promote mo\le cilia structural integrity. Proc Natl Acad Sci U S A 122, e2515392122. 10.1073/pnas.2515392122.
(3) Han, L., Rao, Q., Yang, R., Wang, Y., Chai, P., Xiong, Y., and Zhang, K. (2022). Cryo-EM structure of an ac\ve central apparatus. Nat Struct Mol Biol 29, 472-482. 10.1038/s41594022-00769-9.
(4) Gui, M., Wang, X., Dutcher, S.K., Brown, A., and Zhang, R. (2022). Ciliary central apparatus structure reveals mechanisms of microtubule paderning. Nat Struct Mol Biol 29, 483-492. 10.1038/s41594-022-00770-2.
