Longer primary cilia in NIH3T3^Gli2−/− cells after serum starvation. (A) Immunofluorescence staining of primary cilia (red, Arl13b), the basal body [magenta, γ-tubulin (γ-tub)] and nucleus (DAPI) in NIH3T3^WT and NIH3T3^Gli2−/− cells after serum starvation for 24 h. The scatter plot shows that NIH3T3^Gli2−/− cells (n=60) possessed longer primary cilia compared to NIH3T3^WT cells (n=63), while Gli2 overexpression (OX) (green, GFP-tagged Gli2; n=68) restored the ciliary length in NIH3T3^Gli2−/− cells. The bars represent mean±s.d. ***P<0.001 (one-way ANOVA with post-hoc Bonferroni test). (B) Immunofluorescence staining of primary cilia (red, acetylated α-tubulin) and the nucleus (blue, DAPI) in NIH3T3^WT and NIH3T3^Gli2−/− cells after 24 h of serum starvation. NIH3T3^Gli2−/− cells (n=76) possessed longer primary cilia compared to NIH3T3^WT (n=110) cells. Scale bars: 2 μm. The bars represent mean±s.d. ***P<0.001 (Student's t-test). (C) Time course of ciliary length in NIH3T3 cells after serum starvation. Immunostaining the primary cilia (red, Arl13b), centriolar satellites (gray, Pcm1) and the nucleus (DAPI) in NIH3T3^WT and NIH3T3^Gli2−/− cells after serum starvation for 0, 4, 8, 12, 24 and 48 h. The ciliary lengths of NIH3T3^Gli2−/− cells were significantly longer than NIH3T3^WT after serum starvation for 4, 8, 12, 24 and 48 h. Scale bars: 2 μm. NIH3T3^WT at 0 h, n=74; 4 h, n=102; 8 h, n=105; 12 h, n=65; 24 h, n=135; 48 h, n=63; NIH3T3^Gli2−/− at 0 h, n=65; 4 h, n=156, 8 h, n=152; 12 h, n=97; 24 h, n=115; 48 h, n=65. ***P<0.001 (Student's t-test). Scatter plot displays individual cilia in NIH3T3^WT cells (white circles) and NIH3T3^Gli2−/− cells (black dots). Red lines mark the mean.

Increase in autophagic activities in NIH3T3^Gli2−/− cells. (A) Protein lysates were collected from NIH3T3^WT and NIH3T3^Gli2−/− cells after serum starvation (SS) for 0, 4, 8, 12, 24, and 48 h, and immunoblotted for LC3, p62 and Ofd1. β-actin was used as the loading control. Line charts show the relative protein levels of LC3-II (left), p62 (middle) and Ofd1 (right) in NIH3T3^Gli2−/− cells (white circles) compared to WT cells (black circles). Results are mean±s.e.m.; n=3, *P<0.05; **P<0.01 (Student's t-test). (B) Autophagosomes and autolysosomes labeled with RFP–GFP–LC3 transfected into NIH3T3^WT and NIH3T3^Gli2−/− cells. Autophagosomes (white arrows) appear as yellow whereas autolysosomes (black arrows) appear red due to GFP fluorescence being quenched in the acidic environment. Scale bar: 5 μm. (C) Bar graph showing that the relative density of both autophagosome and autolysosome puncta are higher in NIH3T3^Gli2−/− cells (n=59 cells) than those in NIH3T3^WT cells (n=21 cells). ***P<0.001 (Student's t-test). (D) Bar graph showing that there is no difference in the percentage of autolysosome versus total LC3 puncta between NIH3T3^WT and NIH3T3^Gli2−/− cells. NS, not significant (Student's t-test). Results in C and D are mean±s.e.m.

Inhibition of autophagy restores the ciliary length in NIH3T3^Gli2−/− cells. (A) Western blot of LC3 in NIH3T3 cells treated with an autophagy inhibitor, 3-methyladenine (3-MA). α-tubulin (α-tub) was immunoblotted as the loading control. Bar graph along with individual data points shows that 3-MA reduced LC3-II expression both in NIH3T3^WT and NIH3T3^Gli2−/− cells (n=3 trials). **P<0.01, ***P<0.001 (one-way ANOVA with a post-hoc Bonferroni test). (B) Immunofluorescence staining of primary cilia (red, Arl13b), basal body [magenta, γ-tubulin (γ-tub)] and the nucleus (blue, DAPI) in NIH3T3^WT and NIH3T3^Gli2−/− cells treated with DMSO or 3-MA at the first 4 h of the 24 h serum starvation period. The inset shows the boxed region in each panel. Scale bar: 2 μm. Scatter plots show that primary cilia are longer in NIH3T3^Gli2−/− cells (n=60) than NIH3T3^WT cells (n=63), while 3-MA significantly reduced the length of primary cilia in both cells (n=61 and 62 in NIH3T3^Gli2−/− cells and NIH3T3^WT cells, respectively). ***P<0.001 (one-way AVOVA with post-hoc Bonferroni test). The red line indicates the mean. (C) Immunoblots of LC3 in NIH3T3 cells transfected with sh-Ctrl-GFP or sh-Atg3-GFP and selected by puromycin treatment. α-tubulin (α-tub) serves as the loading control. Bar graph along with individual data points shows that cells transfected with sh-Atg3-GFP exhibited lower LC3-II protein compared to control group upon serum starvation. Results are mean±s.e.m.; n=3 trials. *P<0.05 (Student's t-test). (D) Immunostaining for primary cilia (Arl13b, red), centriolar satellites (Pcm1, magenta) and the nucleus (DAPI, blue) in NIH3T3^WT and NIH3T3^Gli2−/− cells transfected with sh-Ctrl-GFP or sh-Atg3-GFP. Insets show the primary cilia in the boxed regions of each panel. Scale bar: 2 μm. Scatter plots show that primary cilia were longer in NIH3T3^Gli2−/− cells (n=105) than in NIH3T3^WT cells (n=100), while sh-Atg3-GFP reduced the length of primary cilia in both cells (n=133 and 131 in NIH3T3^Gli2−/− cells and NIH3T3^WT cells, respectively). Red lines mark the mean. ***P<0.001 (one-way AVOVA with post-hoc Bonferroni test).

Delay in ciliary resorption in NIH3T3^Gli2−/− cells. (A) Time course of ciliary length in NIH3T3 cells after serum re-stimulation. The primary cilium (red, Arl13b), centrioles (gray, γ-tubulin) and nuclei (DAPI) in NIH3T3^WT and NIH3T3^Gli2−/− cells were stained after serum re-supplementation for 0, 8, 16, and 24 h following 24 h of serum starvation. The scatter plots display individual cilia in NIH3T3^WT cells (white circles) and NIH3T3^Gli2−/− cells (black dots). Red lines mark the mean. Insets represent the boxed region in each panel. Scale bars: 2 μm. NIH3T3^WT at 0 h, n=269; 8 h, n=64; 16 h, n=95; 24 h, n=64; NIH3T3^Gli2−/− at 0 h, n=196; 8 h, n=113; 16 h, n=113; 24 h, n=261. ***P<0.001 (Mann–Whitney U-test). (B) Bar graph along with individual data points representing the percentage of centrioles possessing an Arl13b-positive signal (indicating that the primary cilium is present on the centrioles) after serum starvation for 24 h (ss24), followed by serum re-stimulation for indicated conditions (+8h, +16h, +24h). n=3 trials. P-values are shown (Student's t-test).

Ablation of the primary cilia by Kif3a knockdown rescues the delay of cell cycle re-entry in NIH3T3^Gli2−/− cells. (A) Immunostaining of the primary cilia (Arl13b, red), centrioles (γ-tubulin, magenta) and nucleus (DAPI, blue) in NIH3T3^WT and NIH3T3^Gli2−/− cells transfected with sh-Ctrl-GFP or sh-Kif3a-GFP (green). The boxed regions are magnified in right of each panel, and show the primary cilia in non-transfected cells (GFP−; top) and transfected cells (GFP+; bottom). Kif3a shRNA effectively eliminates the growth of primary cilia in both NIH3T3^WT and NIH3T3^Gli2−/− cells. Scale bar: 2 μm. (B) Cell cycle analysis by flow cytometry in NIH3T3^WT and NIH3T3^Gli2−/− cells and cells transfected with sh-Ctrl-GFP or sh-Kif3a-GFP. Cells were synchronized via serum starvation for 24 h and then collected for 0 h, 8 h, 16 h, and 24 h after serum re-stimulation. (C) The line chart shows the time course of cell cycle re-entry after 0 h, 8 h, 16 h, and 24 h of serum add-back in NIH3T3^WT cells and NIH3T3^Gli2−/− cells transfected with either sh-Ctrl-GFP or sh-Kif3a-GFP. n=3 trials. While NIH3T3^Gli2−/− cells showed a delay in cell cycle re-entry, Kif3a shRNA rescued this effect. **P<0.01, ***P<0.001 (one-way ANOVA with post-hoc Bonferroni test).