Fig 2: HH-GLI2 signaling maintains GCP in an undifferentiated state and promotes their survival. a Schematic representation of the MASTR approach. The R26MASTR allele expresses a GFPcre fusion upon Flp induced deletion of a neo (STOP) cassette. When the R26MASTR allele and the Atoh1-FlpoER transgene are combined with a floxed gene such as Gli2, recombination of the loxP sites occurs in > 98% of GFP+ cells within 3 days of administrating tamoxifen (Tm) at P2. The mutant cells and their progeny can subsequently be identified by the continuous expression of eGFP from the R26 allele. b-c Fluorescent Immuno-Histo-Chemical (FIHC) detection of the indicated proteins and dapi on mid-sagittal sections (lobule VII and VIII) of P8 control R26MASTR/+; Atoh1-FlpoER/+; Gli2lox/+ (Atoh1-M-Gli2 het, b) and R26MASTR/+; Atoh1-FlpoER/+; Gli2lox/lox (Atoh1-M-Gli2 CKO, c) mice treated Tm at P2. d-f Graphs of the proportion of GFP+ cells in the outer (o) EGL at P8 (n = 3) (d), the proliferation index at P8 (% [GFP+ EdU+] cells of all [GFP+] cells in the oEGL) (n = 3) (e) and the number TUNEL+ particles per section at P4 (n = 3) (f) in Atoh1-M-Gli2 het (control, black) and R26MASTR/+; Atoh1-FlpoER/+; Gli2lox/lox (Atoh1-M-Gli2 CKO, red) mice treated with Tm at P2. All of the analyses were performed on 3 midline sections per brain. All graphical data are presented as means ± SEM and significance determined using two-tailed T-test. g-h In situ hybridization of Gli1 mRNA on P0 mid-sagittal cerebellar sections of Gli2lox/lox (control, g) and Atoh1-Cre/+; Gli2lox/lox (Atoh1-Gli2 CKO, h) mice. Black arrowhead indicates the loss of Gli1 expression in the anterior mutant EGL, black arrow Gli1 remaining in the posterior EGL and yellow arrowheads indicate Gli1 expression in Bergmann glia in the Purkinje Cell Layer (PCL). i-j FIHC detection of the indicated proteins and dapi on P0 mid-sagittal cerebellar sections of Gli2lox/lox (control, i) and Atoh1-Cre/+; Gli2lox/lox (Atoh1-Gli2 CKO, j) mice. High power images are shown of the areas indicated by white rectangles and the thickness of the EGL is indicated by yellow bracket. Scale bars represent 100 µm (b-c) and 500 µm (g-j)

Fig 3: ULK3 participates in regulation of GLI1/2 transcriptional activators in ASCs. (A) Expression of genes encoding mediators of the Shh signaling pathway was analyzed using RT-PCR in ASCs isolated from four donors. (B) Primary cilia (red) and nuclei (blue) were stained using antibody against acetylated tubulin and DAPI, respectively. (C) Levels of GLI1 (a), PTCH1 (b), GLI2 (c) and ULK3 (d) mRNAs were measured by qPCR in ASCs stimulated with Shh or TGF-ß1. (D) ASCs were transfected with Neg. or ULK3-specific siRNAs. The levels of ULK3 (a), GLI2 (b) and GLI1 (c) transcripts were measured by qPCR; (d) GLI1 and GLI2 proteins were analyzed by immuno-blotting. (E) ASCs were transfected with Neg. or ULK3-specific siRNAs and 24 h later treated with TGF-ß1 for 24 h. The levels of ULK3 (a), GLI2 (b) and GLI1 (c) transcripts were measured by qPCR. (F) (a) GLI2 and GLI1 proteins were analyzed by WB in cells transfected with Neg. or ULK3-specific siRNAs and induced with TGF-ß1 for 30 h. GLI2, GLI1 and GAPDH images were quantified. Normalized GLI2 (b) and GLI1 (c) protein levels were set as 1 in cells transfected with Neg. siRNA. Data of qPCR and quantified WBs are presented as an average mean ± S.D.; *—p < 0.05, **—p < 0.01, ***—p < 0.001, n = 3.

Fig 4: The accumulation of Gli2 at the ciliary tip is not just a consequence of increased proteins levels, but requires activation of the Hh pathway.(A) WB showing Gli2 levels in control cells, cells stimulated with Shh-N for 90 min or 18 h and cells treated with the proteasome inhibitor MG132 for 90 min. Gli2 levels were normalized against a-tubulin and the graph shows mean ± s.d of biological duplicates. (B) Analysis of Gli2 ciliary localization in transduced NIH/3T3 expressing GFP-Gli2 treated as described in (A). Cells were stained for cilium (anti-Ac.Tub, red), GFP-Gli2 (anti-GFP, green) and nucleus (TOPRO, blue). Small pictures show amplification of the selected region. Yellow and white arrows indicate cilia with or without Gli2 at the ciliary tip respectively. Scale bar: 10 µm. (C) Quantification of Gli2 ciliary localization. Results are expressed as the fraction of cilia that showed Gli2 staining (Gli2+ cilia) with 95% confidence interval (CI). At least 85 cilia were analysed for each sample. ** p<0.001 (hypothesis test for proportions).

Fig 5: Imp-a/ß1 transports Gli2 into the nucleus.(A) Two segments of a T-Coffee [69] alignment of four Gli-like sequences: GLI2_Mm (Mus musculus), GLI2_Hs, GLI1_Hs, GLI3_Hs (Homo sapiens) and Ci_Dm (Drosophila melanogaster). Residue similarity is colour coded according to the Risler substitution matrix, using ESPript [70]. Dots mark 10-residue intervals of the top sequence. Black bars indicate the two bipartite cNLS (cNLS-1 and cNLS-2) predicted by cNLS mapper [40] for GLI2_Mm. (B) GFP-Gli2-NIH/3T3 cells were treated with Importazol (IPZ) in order to inhibit Imp-ß1-mediated nuclear transport. DMSO was used as control. Hh signalling was activated for 90 min using SAG and non-stimulated cells were treated with DMSO. Cells were stained for cilium (anti-Ac.Tub, red), GFP-Gli2 (anti-GFP, green) and nucleus (TOPRO, blue). The same pictures are also shown with the channel corresponding to GFP-Gli2 in grey scale to help the visualization. Scale bar: 10 µm. (C) Quantification of nuclear Gli2 was performed as explained in Materials and Methods. The mean nuclear fluorescence was normalised against the mean total GFP fluorescence of the cell, so as to correct for variations in Gli2 expression among different cells. Results are representative of three experiments and at least 60 cells were analysed for each condition. ** p<0.001, *** p<0.0001 (Kruskal-Wallis test). (D) Activation of a luciferase-based Hh reporter gene in NIH/3T3 cells stimulated with SAG (or DMSO as control) in the presence of IPZ or DMSO. As explained in Materials and Methods, RLU values from SAG treated cells are normalised against RLU of DMSO treated ones and expressed as mean ± s.d. from triplicates from two independent experiments. * p< 0.01 (Mann-Whitney test). (E) Western blot (WB) showing GFP-Gli2 levels in control and IPZ-treated cells. GFP-Gli2 was detected using an anti-GFP antibody. (F-G) WB detecting Imp-ß1 (F) and Imp-a1 (G) after precipitating GFP-Gli2 with GFP-Trap (left) from HEK293FT transfected with pEGFP-Gli2 (left panels). Membranes were cut at different levels so as to detect in the same samples the precipitated GFP-Gli2 and GFP (left panels). The levels of GFP-Gli2, Imp-ß1, Imp-a and GFP in the lysates used for immunoprecipitation were assessed by WB and are shown in the right panels. The band corresponding to GFP looks distorted because the protein migrates with the dye front.