Transcriptomic comparison of zebrafish JBTS mutants. (A) Schematic of a cilium, outlining ciliary subcompartments [basal body (BB), transition zone (TZ), axoneme and tip]. The localisation of the proteins encoded by the genes studied here is indicated. Specifically, Togaram1 and Talpid3 are localised to the BB, with Togaram1 additionally translocating to the ciliary tip, Inpp5e is localised along the ciliary membrane and Cc2d2a and Cep290 are found at the TZ. Bbs1, not associated with JBTS, functions as part of the octameric protein complex BBSome. Whole larval bulk RNA sequencing at 3 dpf was performed on zebrafish mutants harbouring mutations in one of the depicted genes. All were maternal zygotic mutants (mz) except for talpid3. (B) Heat map of the Pearson correlation coefficients derived from regularised log-transformed gene counts, illustrating the overall high similarity between the different samples (correlation coefficients all above 0.88). The hierarchical clustering using average Euclidean distance reveals a pronounced batch effect in both sample pairs and zebrafish lines. (C) Principal component analysis focusing on the top 25% of the most variable genes, confirming a partial clustering effect attributed to the specific zebrafish lines. (D) UpSet plot providing a visual representation of unique and shared genes that showed significant differential expression (adjusted P-value <0.05) in the paired differential expression analysis. The blue bars on the left show the total number of genes differentially expressed per mutant/control pair. The black bar plot on top indicates the number of genes that are commonly differentially expressed in the mutants indicated with the black dot below. Detailed information on individual genes and the intercept can be found in Tables S2 and S3.

Enrichment of genes involved in central nervous system development and function in all JBTS mutants. (A) Box plots showing the normalised expression levels of key actors of the Shh pathway in the different mutant lines and their respective sibling controls in the whole larval tissue analysis. Each data point represents an individual sample (of pooled larvae, see details in the Materials and Methods), where paired mutant/control sibling samples are connected by a black line. Box plots represent Q1-Q3 with median (thick line). Note the lack of a strong effect in any of the components in any of the mutant-sibling pairs. (B) Gene set enrichment analysis highlighting the commonality between the mutant lines. The blue bars on the left indicate the number of gene ontology terms enriched per mutant/control pair, while the black bar/dots indicate the number of shared terms found in all samples. The six terms shared between all samples are displayed in a heatmap showing the normalized enrichment score of these six significantly enriched ontology terms shared between all samples. (C) As an illustration, for two of the mutants (cc2d2a and talpid3), barcode plots detailing the position of individual genes associated to ‘gated channel activity’ (red) and ‘synaptic membrane’ (blue) along the fold change ranked list (shown in grey at the bottom of the graph). Note that the ontology terms ‘gated channel activity’ and ‘synaptic membrane’ for cc2d2a and talpid3 mutants are significantly enriched on the opposite ends of the ranked gene lists. Both terms are enriched in the up-regulated proportion in talpid3 mutants but in the down-regulated proportion in cc2d2a mutants (indicated by grey boxed regions).

Loss of primary cilia in cerebellar neurons of cc2d2a and talpid3 mutants. (A) Representative whole-mount maximum projection confocal immunofluorescence images showing parallel fibres of granule cells (anti-Vglut1 – cyan), Purkinje cells (anti-Pvalb7 – magenta) and olig2+ eurydendroid cells [Tg(olig2:EGFP) transgenic line – green] in the cerebellum of wild-type zebrafish larvae at 5 dpf. Images show a dorsal view with anterior to the left. (B-G) Whole-mount maximum projection confocal immunofluorescence images showing primary cilia labelled with the Tg(ubi:Arl13b-mCherry) transgenic line (Arl13b – yellow) and basal bodies (BB – green) labelled using anti-γ-Tubulin (C,E-G) or using the Tg(β-actin:GFP-centrin) transgenic line (B,D). Purkinje cells (PC – magenta) are marked using anti-Pvalb7 (B-D) and eurydendroid cells (EC – green) are marked by the Tg(olig2:EGFP) transgenic line (E-G). Note the strong decrease in cilia numbers in both mutants (cc2d2a^−/− in C and F; talpid3^−/− in D and G) compared to control (B and E). Insets show a magnified view of one representative cilium [from top to bottom: composite image (B3-G3), Arl13b (B4-G4) and BB (B5-G5)]. Images show a dorsal view of 5 dpf larvae with anterior to the left. Scale bars: 10 µm. (H-I) Violin plots showing a statistically significant reduction in Arl13b+ cilia within the mutant Purkinje cell layer (H) and olig2+ eurydendroid cell layer (I) compared to controls. Each data point represents one larva, and violin plot represents median (thick line) and quartiles (dashed line). ****P≤0.0001. Welch's ANOVA with post-hoc Dunnett's T3 multiple comparisons test. For Purkinje cell analysis, control n=25 (N=5), cc2d2a^−/−n=10 (N=3), talpid3^−/−n=15 (N=2) larvae. For eurydendroid cell analysis, control n=23 (N=4), cc2d2a^−/−n=8 (N=2), talpid3^−/−n=16 (N=2) larvae. N denotes the number of independent experiments (i.e. larvae from independent clutches) used for analysis. Analyses were carried out using mz cc2d2a^−/− and zygotic talpid3^−/− mutants.

Morphological development of the cerebellum is unaffected in cc2d2a and talpid3 mutants. (A-C) Whole-mount maximum projection confocal images showing tagRFP-T-positive Purkinje cells in control (A), cc2d2a^−/− (B) and talpid3^−/− (C) Tg(tagRFP-T:PC:GCaMP5G) larvae at 5 dpf. Note the similar morphology of the PCL in mutants and controls. (D) Violin plot showing no statistically significant difference in the number of tagRFP-T-positive Purkinje cells per cerebellar hemisphere in mutants compared to controls. Control n=24 larvae (N=4), cc2d2a^−/− n=5 larvae (N=2, P=0.99), talpid3^−/− n=16 larvae (N=2, P=0.82). (E-G) Whole-mount maximum projection confocal immunofluorescence images showing normal morphology of Vglut1-positive parallel fibres in control (E), cc2d2a^−/− (F) and talpid3^−/− (G) larvae at 5 dpf. (H) Violin plot showing no statistically significant difference in the normalised Vglut1-positive fluorescence area in mutants compared to controls. Control n=27 larvae (N=4), cc2d2a^−/− n=15 larvae (N=2, P=1.00), talpid3^−/− n=12 larvae (N=2, P=0.21). (I-K) Whole-mount maximum projection confocal images showing EGFP-positive eurydendroid cells in control (I), cc2d2a^−/− (J) and talpid3^−/− (K) Tg(olig2:EGFP) larvae at 5 dpf. The morphology of the eurydendroid cell layer is similar in mutants and controls. (L) Violin plot showing no statistically significant difference in the number of EGFP-positive eurydendroid cells per cerebellar hemisphere in mutants compared to controls. Control n=25 larvae (N=4), cc2d2a^−/− n=11 larvae (N=2, P=0.99), talpid3^−/− n=16 larvae (N=2, P=0.20). (M–O) Whole-mount maximum projection confocal images showing EGFP-positive eurydendroid cell axons in control (M), cc2d2a^−/− (N) and talpid3^−/− (O) Tg(olig2:EGFP) larvae at 5 dpf. (P) Violin plot showing quantification of the EGFP-positive axon tract thickness, where no statistically significant difference is observed in mutants compared to controls. Control n=19 larvae (N=4), cc2d2a^−/− n=10 larvae (N=2, P=0.65), talpid3^−/− n=15 larvae (N=2, P=0.50). For violin plots in D, H, L and P, each data point represents one larva. Violin plots represent median (thick line) and quartiles (dashed line). ns, not significant. Ordinary one-way ANOVA with post-hoc Dunnett's multiple comparisons test. For A-O, images show a dorsal view of 5 dpf larvae with anterior to the left. For A-K, image scale bars are 50 µm and for M-O, image scale bars are 20 µm. N denotes the number of independent experiments (i.e. larvae from independent clutches) used for analysis. Analyses were carried out using mz cc2d2a^−/− and zygotic talpid3^−/− mutants. (Q-T) Representative maximum projection mesoSPIM images showing an optical sagittal section of DAPI-stained nuclei and tagRFP-T-positive Purkinje cells in the cerebellum of Tg(tagRFP-T:PC:GCaMP5G) 11-12 wpf control (Q, n=4, N=2; S, n=10, N=4) and zygotic cc2d2a^−/− (R, n=4, N=2; T, n=6, N=3) fish. (S’,T’) Higher magnification images of boxed regions in S and T, respectively. The GCL is outlined with a cyan line, the PCL is indicated with a magenta bracket and the ML with a yellow bracket. The morphology of all layers is unchanged in mutants. (U,V) Maximum projection mesoSPIM images showing an optical sagittal section of olig2:EGFP-positive eurydendroid cells in the cerebellum of Tg(olig2:EGFP) 11-12 wpf control (U, n=3, N=2) and zygotic cc2d2a^−/− (V, n=3, N=2) fish. Eurydendroid cells and axons appear normal in mutants compared to controls. U’ and V’ show higher magnification images of boxed regions in U and V, respectively. For Q-V, images show a sagittal view with anterior to the left; scale bars: 200 µm. N denotes the number of independent experiments (i.e. larvae from independent clutches) used for analysis.

Abnormal cilia in the brain of cc2d2a and talpid3 mutants. (A) Whole-mount confocal image of a 5 dpf zebrafish larva immunostained with anti-acetylated tubulin, providing orientation for the figure. Anterior is to the left and boxes show the analysed brain regions in the subsequent panels. (B-G) Whole-mount maximum projection confocal immunofluorescence images of the forebrain (B-D) and midbrain (E-G), showing primary cilia labelled with anti-Arl13b (Arl13b – yellow) and basal bodies labelled with anti-Centrin (Centrin – magenta) in control (n=12 in B, n=13 in E, N=2), cc2d2a^−/− (n=9 in C, n=9 in F, N=1) and talpid3^−/− (n=6 in D, n=6 in G, N=2) larvae at 5 dpf. In E-G, Arl13b-positive and acetylated tubulin-positive motile cilia are visible at the midline (indicated with white arrow heads), while asterisks indicate location of Arl13b-positive primary cilia in the midbrain. In both cc2d2a^−/− and talpid3^−/− larvae, there is a marked reduction in primary cilia. (H-J) Whole-mount maximum projection confocal immunofluorescence images of the hindbrain ventricle, using anti-acetylated tubulin (AcTub – green) and anti-Arl13b (Arl13b – yellow) to label motile cilia and anti-Centrin (Centrin – magenta) to label basal bodies, in control (n=10, N=2), cc2d2a^−/− (n=8, N=1) and talpid3^−/− (n=6, N=2) larvae at 5 dpf. Motile cilia are reduced in both cc2d2a^−/− and talpid3^−/− larvae compared to controls. All images show a dorsal view of 5 dpf larvae with anterior to the left. Scale bars: 20 µm. N denotes the number of independent experiments (i.e. larvae from independent clutches) used for analysis. Analyses were carried out using mz cc2d2a^−/− and zygotic talpid3^−/− mutants.

Brain morphology and size is normal in cc2d2a and talpid3 mutants. (A-C) Whole-mount single optical slice widefield immunofluorescence images showing brain morphology of control (A), cc2d2a^−/− (B) and talpid3^−/− (C) larvae at 5 dpf, stained with anti-HuC/HuD to label neurons (HuC/HuD – magenta) and DAPI to counterstain nuclei (DAPI – cyan). Note that the morphology of the brain in cc2d2a^−/− and talpid3^−/− larvae is comparable to controls. (D) Schematic showing regions of interest used for brain area quantification. (E,F) Scatter plots showing normalised area of the forebrain, midbrain and hindbrain regions in control versus cc2d2a^−/− larvae (E) and control versus talpid3^−/− larvae (F) at 5 dpf. There is no statistically significant difference in normalised brain area in cc2d2a^−/− larvae compared to controls. In talpid3^−/− larvae, there is a minimal but statistically significant difference in normalised forebrain area compared to controls. Each data point represents one larva. Error bars are mean±s.d. ns, not significant; *P≤0.05. Unpaired t-test. For E, control n=22 (N=2), cc2d2a^−/− n=14 (N=2) larvae. For F, control n=23 (N=2), talpid3^−/− n=22 (N=2) larvae. (G-L) Whole-mount single optical slice widefield immunofluorescence images showing an overview of axonal tracts labelled with anti-acetylated tubulin (AcTub – green) and synaptic neuropil labelled with anti-synaptic vesicle 2 (SV2 – magenta) in control (G, n=35, N=4), cc2d2a^−/− (H, n=14, N=2) and talpid3^−/− (I, n=24, N=2) larvae at 5 dpf. (J-L) shows higher resolution whole-mount single optical slice confocal immunofluorescence images of boxed region in (G-I). The morphology of axonal tracts and synaptic neuropil appears unchanged in both mutants compared to controls. All images show a dorsal view of 5 dpf larvae with anterior to the left. Scale bars: 500 µm in A-C; 300 µm in G-I and 10 µm in J-L. N denotes the number of independent experiments (i.e. larvae from independent clutches) used for analysis. Analyses were carried out using mz cc2d2a^−/− and zygotic talpid3^−/− mutants.

Abnormal swimming behaviour in JBTS mutants. (A) Plot showing the mean distance travelled during three sequential 5-min periods of darkness, light, darkness in mz cc2d2a^−/− and milder zygotic cc2d2a^−/− larvae compared to controls. The distance moved by each larva is recorded during 10 s intervals, and this distance is plotted over time. Note that both mz cc2d2a^−/− and zygotic cc2d2a^−/− larvae travel less distance than controls during both dark periods. Error shown with the coloured areas represent the 95% CI. (B) Scatter plot showing the total distance travelled during each entire 5-min period of darkness, light and darkness in mz cc2d2a^−/− and zygotic cc2d2a^−/− larvae compared to controls. The total distance travelled is significantly reduced in both mz cc2d2a^−/− and zygotic cc2d2a^−/−. Error bars are mean±s.d. ns, not significant; *P≤0.05; ****P≤0.0001. Kruskal–Wallis test with post-hoc Dunn's multiple comparisons test. Control n=46 (N=3), mz cc2d2a^−/− n=16 (N=1), zygotic cc2d2a^−/− n=32 (N=2) larvae. (C) Similar plot as in (A) but for zygotic talpid3^−/− larvae compared to controls. The distance travelled throughout the experiment does not differ between mutants and controls. Error shown with the coloured areas represent the 95% CI. (D) Similar plot as in (B) but for zygotic talpid3^−/− larvae compared to controls. The total distance is not significantly reduced in talpid3^−/− larvae (trend towards decrease in the second dark phase not significant). Error bars are mean±s.d. ns, not significant. Unpaired t-test. Control n=51 (N=3), talpid3^−/− n=37 (N=3) larvae. (E) Representative 1-min traces showing locomotion in one control, one mz cc2d2a^−/−, one zygotic cc2d2a^−/− and one talpid3^−/− larva. Black traces represent inactivity, green traces represent normal activity and red traces represent fast activity. (F) Representative montages of the touch-evoked escape response (TER) in control, zygotic cc2d2a^−/− and talpid3^−/− larvae at 2 dpf. Insets show magnified view of the larvae. While the control and talpid3^−/− larvae are able to escape the stimulus, the abnormal swimming of the zygotic cc2d2a^−/− larvae (circular with partial swimming on the side or upside down) prevents it from escaping from the stimulus in the same manner. Please note the only mild curvature of the larva, which should not suffice to explain the tightly curved swimming trace observed. (G) Scatter plot showing displacement of control, zygotic cc2d2a^−/− and talpid3^−/− larvae during the TER. Zygotic cc2d2a^−/− exhibit a significantly reduced displacement compared to controls (P<0.0001). The trend towards decreased displacement in talpid3^−/− does not reach statistical significance (P=0.1146). Error bars are mean±s.d. not significant; **** P≤0.0001. Ordinary one-way ANOVA with post-hoc Dunnett's multiple comparisons test. Control n=32 (N=4), zygotic cc2d2a^−/− n=20 (N=2), talpid3^−/− n=14 (N=2) larvae. N denotes the number of independent experiments (i.e. larvae from independent clutches) used for analysis.