Loss of C2a proteins in mice leads to PCD-related phenotypes.

(A) Schematics of the cross-section of motile cilia and major projections associated with the C1 and C2 microtubules. The molecular model of the C2a projection (PDB: 7SOM) is superimposed onto the cryo-EM density map from Han, L. et al (Han et al., 2022). (B) Schematic strategies for the generation of Ccdc108 KO, Mycbpap KO, and Cfap70 KO mice using CRISPR/Cas9. The genomic positions of the primers used for genotyping are indicated. UTR, untranslated region; CDS, coding sequence. (C) Genotyping of WT, HET, and KO mice for each strain. (D) Immunoblotting showing the depletion efficiency in Mycbpap KO and Cfap70 KO mEPCs. GAPDH is used as a loading control. (E) Genotype distribution profiles of pups at P0 and P7 resulting from matings of HET mice with specified genotypes. Note that Mycbpap KO mice were born at Mendelian ratios but experienced early death within one week of birth. (F) Representative images of mice with specified genotypes at two weeks of age. (G) Survival curves of WT (starting number: 30), Ccdc108 KO (starting number: 18), Mycbpap KO (starting number: 10), and Cfap70 KO (starting number: 21) mice. The numbers of surviving mice for each genotype at twelve weeks of age are shown. (H) Weights of male and female WT, Ccdc108 HET, and Ccdc108 KO (survivor) mice were recorded from one to eight weeks of age. Data are presented as mean ± SD (n = 6 mice per genotype). (I) Representative images of serial vibratome sections of the brains from mice with specified genotypes at two weeks of age. (J) Periodic Acid-Schiff (PAS) staining of the nasal cavities of mice with specified genotypes at two weeks of age. Magnified images are shown on the right. The asterisks indicate mucus accumulation. (K) Incidence of hydrocephalus and sinusitis in mice with specified genotypes. Hydrocephalus and sinusitis were determined as described in (I) and (J), respectively. The cell number in each column indicates the number of mice analyzed.

C2a proteins are essential for the integrity of C2-related projections.

(A, B) Immunofluorescence (A) and quantifications (B) of the number of basal bodies and cilia per mEPC cultured from WT, Ccdc108 KO, Mycbpap KO, and Cfap70 KO mice. Cells were immunostained with acetylated α-tubulin (ace-Tub) and CEP164 antibodies, and imaged with 3D-SIM. 80 cells from 3 mice (per genotype) were scored using ImageJ. (C, D) Representative frames (C) and quantifications (D) of indicated movement modes and beat frequencies of motile cilia in WT, Ccdc108 KO, Mycbpap KO, and Cfap70 KO mEPCs. Trajectories of four cilia in each cell are shown. Diagrams illustrate the corresponding ciliary beat patterns. 60 cells from 3 mice (per genotype) were scored using ImageJ. (E–H) TEM images and quantifications of ciliary axonemes in mEPCs (E, F) and tracheal MCCs (G, H) from WT, Ccdc108 KO, Mycbpap KO, and Cfap70 KO mice. Arrowheads indicate the C2 projections. At least 50 axonemes from 3 mice (per genotype) were scored. Data in (B, D, F, H) are presented as mean ± SD. Unpaired two-tailed t-test was performed. *, P < 0.05; **, P < 0.01; ***, P < 0.001; ns, not significant.

CCDC108, MYCBPAP, and CFAP70 mutually interact with each other.

(A, B) Co-IP and immunoblotting showing the interaction between endogenous MYCBPAP and CFAP70. Co-IP was performed with a normal rat IgG control antibody and a rat polyclonal anti-CFAP70 antibody (A) or a normal guinea pig IgG and a guinea pig polyclonal anti-MYCBPAP antibody (B) in mouse testis lysates. (C, D) Co-IP and immunoblotting analyses in HEK293T cells exogenously expressing indicated proteins. HA-tagged proteins in (C) and GFP-tagged proteins in (D) were immunoprecipitated with anti-HA and anti-GFP agarose beads, respectively. Blots were probed with the indicated antibodies. Luci, luciferase. (E) Diagrams of CCDC108, MYCBPAP, and CFAP70 full-length (F) and truncated fragments showing their ability to interact. Interactions were determined through co-IP analyses. Numbers indicate amino acid positions. PPI, protein-protein interaction. (F–H) GST pull-down showing direct interactions using purified fragment proteins. Blots were probed with the indicated antibodies. CB, coomassie blue staining.

CCDC108, MYCBPAP, and CFAP70 localize to the axonemal central lumen.

(A, B) Confocal (A) and 3D-SIM (B) images of mEPCs expressing GFP-tagged proteins immunostained with the indicated antibodies. Magnified images of the cilia indicated by arrowheads are shown on the right. Line-scan graphs show the immunofluorescence intensity along the positions marked by arrows in the magnified images. The right Y-axis describes the immunofluorescence intensity of GFP-CCDC108 and GFP-CFAP70, respectively. (C, D) 3D-SIM images of mEPCs immunostained with the indicated antibodies. Magnified images of the cilia indicated by arrowheads are shown on the right. Line-scan graphs show the immunofluorescence intensity along the positions marked by the two arrows in the magnified images. The right Y-axis describes the immunofluorescence intensity of ace-Tub in (D).

CCDC108, MYCBPAP, and CFAP70 are all essential for the stable docking of each C2a protein.

(A–F) Representative confocal images of mEPC (A–C) and tracheal MCCs (D–F) from WT, Ccdc108 KO, Mycbpap KO, or Cfap70 KO mice immunostained with the indicated antibodies. Note that the ciliary staining of C2a proteins was greatly reduced in KO samples. (G) Schematic diagrams of mEPC culture and motile cilia purification. (H–J) Immunoblotting and quantification showing the C2a protein levels in motile cilia (H) or mEPCs (I) derived from Ccdc108 KO, Mycbpap KO, and Cfap70 KO mEPCs. Acetylated α-tubulin (ace-Tub) and GAPDH are used as loading controls, respectively. The MYCBPAP or CFAP70 band intensity in (I) was normalized to the corresponding GAPDH intensity. Data are presented as mean ± SD. (K) Real-time PCR analyses showing the expression levels of the indicated genes in Ccdc108 KO, Mycbpap KO, and Cfap70 KO mEPCs. The expression was normalized using the corresponding Gapdh as the reference gene and WT as the reference sample (△△CT method). Data are from three independent biological repeats and are presented as mean ± SD. (L) Proposed model showing the absence of the C2a projection in Ccdc108 KO, Mycbpap KO, and Cfap70 KO axoneme.

Identification of ARMC3 and MYCBP as new C2a components.

(A, B) Silver staining of proteins immunoprecipitated from mouse testis lysates using a normal guinea pig IgG and a guinea pig polyclonal anti-MYCBPAP antibody (A) or a normal rat IgG control antibody and a rat polyclonal anti-CFAP70 antibody (B). The bands of MYCBPAP and CFAP70 are indicated by red arrowheads. (C, D) Interactor candidates of MYCBPAP (C) and CFAP70 (D) identified by mass spectrometry analysis. (E, F) Co-IP and immunoblotting showing the interactions of endogenous MYCBPAP (E) and CFAP70 (F) with ARMC3 and MYCBP. Co-IP was performed with a normal guinea pig IgG and a guinea pig polyclonal anti-MYCBPAP antibody (E) or a normal rat IgG control antibody and a rat polyclonal anti-CFAP70 antibody (F) in mouse testis lysates. (G, H) Co-IP and immunoblotting analyses in HEK293T cells exogenously expressing indicated proteins. GFP-tagged proteins were immunoprecipitated with anti-GFP agarose beads. Blots were probed with the indicated antibodies. Luci, luciferase. (I) 3D-SIM images of mEPCs immunostained with the indicated antibodies. Magnified images of the cilia indicated by arrowheads are shown on the right. Line-scan graphs show the immunofluorescence intensity along the positions marked by two arrows in the magnified images. (J–M) Representative confocal images of mEPC from WT, Mycbpap, and Cfap70 KO mice immunostained with the indicated antibodies.

Mice lacking C2a proteins display phenotypes associated with PCD.

(A–C) Real-time PCR analyses showing the expression levels of Ccdc108, Mycbpap, and Cfap70 in various mouse tissues. The expression was normalized using the corresponding Gapdh as the reference gene and baseline 1 (heart) as the reference sample (△△CT method). Data are from three independent biological repeats and are presented as mean ± SEM. (D) Typical images of tracheal MCCs from WT and Ccdc108 KO mice immunostained with the indicated antibodies. Note that the staining of CCDC108 in Ccdc108 KO cells is invisible. (E) Representative images of WT and Ccdc108 KO mice at eight weeks of age. (F) Representative images of serial vibratome sections of the brains from WT and Ccdc108 KO mice at twelve weeks of age.

Loss of C2a proteins leads to ultrastructural defects in the cilia axoneme.

(A, B) SEM images of ependyma (A) and trachea (B) epithelia isolated from WT, Ccdc108 KO, Mycbpap KO, and Cfap70 KO mice at two weeks of age.

C2a proteins display mutual interactions.

(A–G) Co-IP and immunoblotting analyses in HEK293T cells exogenously expressing indicated proteins. GFP-tagged proteins in (A, D, and E) and HA-tagged proteins in (B, C, F, and G) were immunoprecipitated with anti-GFP and anti-HA agarose beads, respectively. Blots were probed with the indicated antibodies. Luci, luciferase.

Loss of C2a proteins has no obvious impact on non-C2a ciliary proteins.

(A–C) Representative confocal images of mEPC cultured from WT, Ccdc108 KO, Mycbpap KO, or Cfap70 KO mice immunostained with the indicated antibodies. Note that the ciliary staining of SPEF1, HYDIN, and RSPH3 was unchanged in KO samples.

Identification of ARMC3 and MYCBP as new C2a components.

Immunoblotting of mEPCs harvested at the indicated serum-starvation days for the indicated proteins. GAPDH is used as a loading control.