Interaction of smoothened with integrin-linked kinase in primary cilia mediates Hedgehog signalling

Abstract

Here we report that ILK localizes in the mouse primary cilium, a sensory organelle required for signalling by the Hedgehog (Hh) pathway. Genetic or pharmacological inhibition of ILK blocks ciliary accumulation of the Hh pathway effector smoothened (Smo) and suppresses the induction of Gli transcription factor mRNAs by SHh. Conditional deletion of ILK or Smo also inhibits SHh-driven activation of Gli2 in the embryonic mouse cerebellum. ILK regulation of Hh signalling probably requires the physical interaction of ILK and Smo in the cilium, and we also show selective cilia-associated interaction of ILK with β-arrestin, a known mediator of Smo-dependent signalling.

Figure 1

Integrin-linked kinase (ILK) is expressed in granule cell precursors (GCPs) and mediates Hedgehog (Hh) signalling in vivo and in vitro. (A) GCPs isolated from post-partum day (dpp) 7 mice were positive for HNK1, Math1 and ILK and have primary cilia (Arl13b, red). Scale bar, 5 μM. (B, upper) Representative sagittal sections of H&E-stained E18.5 cerebellae from E18.5 ilk^loxP/+ (control), Math1-Cre;ilk^loxP/loxP and Math1-Cre;Smo^loxP/loxP mice. Asterisks indicate shallow invaginations corresponding to fissure locations in control cerebellum. (B, lower) Sagittal sections as indicated, stained with antibody specifically recognizing activated Gli2^A. Arrows indicate GCPs in the EGL, bounded by a dashed line. Scale bar, 50 μm. (C) Induction of alkaline phosphatase activity in ILK-depleted 10T1/2 cells by SHh-CM, recombinant Sonic Hedgehog (rhSHh) and SAG (smoothened agonist). Controls included conditioned medium (CM) from HEK293 cell cultures (293-CM) and SHh-CM pre-incubated with 5E1 SHh-neutralizing monoclonal antibody. IB: western blotting for ILK protein levels in control and ILK siRNA-treated cells. Data are presented as mean±s.e.m., *P<0.05, **P<0.01 (Student’s t-test), n=4 independent experiments with triplicate determinations in each. (D) Primary cultures of ilk^loxP/loxP MEFs were infected with Ad-GFP control or Ad-Cre^GFP for 24 h, serum starved for an additional 24 h before stimulation with rhSHh for 0, 12 or 24 h. Cells were harvested and RNA subjected to qRT–PCR to quantify ILK, Gli1 and Gli2 expression as described in Methods. (E) Primary MEF cultures were serum starved for 24 h, then pretreated with the indicated concentrations of QLT-0267 for 1 h before stimulation for 24 h with rhSHh. Cells were harvested and RNA subjected to quantitative RT–PCR (qRT–PCR) to quantify Gli1 and Gli2 induction. Results are inclusive of two independent experiments. Columns indicate the mean of two independent experiments with duplicate determinations in each experiment.

Figure 2

Integrin-linked kinase (ILK) localizes in primary cilia of epithelial and fibroblast cells. Stacked confocal images of IMCD3 (A) and C3H10T1/2 (B) cells transfected with mCherry-tagged ILK. Primary cilia were labelled with anti-αtub^acet antibody (green). Channels are shown as a single overlay (merge), with signal in the primary cilium shown as shifted-overlay (inset). (C) Confocal images of C3H10T1/2 cells labelled for ILK with an anti-ILK mouse monoclonal antibody (mILK, green) and centrosomes using γ-tubulin antibody (γ-tub, red). (D) Confocal images of IMCD3 cells labelled with an anti-ILK rabbit monoclonal antibody (rILK, red) and α-tub^acet antibody (green). Lower panel represents the corresponding confocal image in the x–z plane. Arrows indicate primary cilia. Scale bar, 10 μm.

Figure 3

Integrin-linked kinase (ILK) and smoothened (Smo) associate in the primary cilium. Protein–protein interactions were determined using a Venus-based BiFC assay. (A,B) Z-stack confocal images of (A) β-arrestin 2 (arrB2) and Smo and (B) ILK and Smo protein–protein interactions (green). Primary cilia and basal bodies were identified by immunofluorescence using anti-α-tub^acet (magenta) and anti-γ-tub (blue) antibodies, respectively. Transfected cells were identified by immunofluorescence with anti-GFP antibody (red) specifically detecting the V2 tag. (C) Representative images of Venus signal in cilia and its quantification using 3D isosurface rendering and interaction density maps as detailed in Methods. (D) Quantification of the mean Venus interaction intensity relative to the amount of split Venus components present. The signal ratio of reconstituted Venus (green) to total GFP (red) in cilia of transfected cells is indicated below graph. Results from one representative experiment are graphed. Scale bar, 5 μM for panels A and B.

Figure 4

β-arrestin 2 interacts with integrin-linked kinase (ILK) and Kif3a at base of the cilium. Z-stack confocal images of V2-β-arrestin 2 (arrB2) co-transfected with (A) V1-ILK or (B) V1-Kif3a. Interaction of V1- and V2-tagged proteins (green) is localized relative to the primary cilium (magenta, α-tub^acet) and basal bodies (blue, γ-tub), as in Fig 3. 3D isosurface renderings for A and B were generated using IMARIS software. Scale bar, 5 μM (A), with the same magnification shown in B.

Figure 5

Integrin-linked kinase (ILK) mediates ciliary translocation of smoothened (Smo). (A) Single-plane confocal images of IMCD3 cells untreated or transfected with either scrambled siRNA (siScr) or ILK siRNA (siILK) and treated with either SHh-CM, 1 μg/ml recombinant Sonic Hedgehog (rhSHh) or 200 nM smoothened agonist (SAG) for 6 h. Smo protein (red) is detected by immunofluorescence. Primary cilia are labelled with anti-α-tub^acet antibody (green), the centrosome with γ-tubulin (magenta) and nuclei with DAPI (blue). Scale bar, 10 μM. (B,C) Number of cilia positive for Smo as detected by immunofluorescence was quantitated in IMCD3 (B) and C3H10T1/2 (C) cells and expressed as a percentage of cilia count. Data are shown as mean±s.e.m., asterisk indicates significance, P<0.01 (Student’s t-test), n=3 independent experiments.