Distinct Effects of the mesenchymal dysplasia Gene Variant of Murine Patched-1 Protein on Canonical and Non-canonical Hedgehog Signaling Pathways

Background: Patched-1 is the primary ligand receptor in both canonical and non-canonical Hedgehog signaling.

Results: Truncation of the Patched-1 C terminus impairs regulation of Patched-1-mediated c-src activation and elevates the expression of Hedgehog target genes.

Conclusion: The Ptch1 C terminus is required for mediation of both canonical and non-canonical Hedgehog signaling.

Significance: A mutant of Ptch1 differentially regulates canonical and non-canonical Hh signaling pathways.

Abstract

Hedgehog (Hh) signaling requires regulation of the receptor Patched-1 (Ptch1), which, in turn, regulates Smoothened activity (canonical Hh signaling) as well as other non-canonical signaling pathways. The mutant Ptch1 allele mesenchymal dysplasia (mes), which truncates the Ptch1 C terminus, produces a limited spectrum of developmental defects in mice as well as deregulation of canonical Hh signaling in some, but not all, affected tissues. Paradoxically, mes suppresses canonical Hh signaling and binds to Hh ligands with an affinity similar to wild-type mouse Ptch1 (mPtch1). We characterized the distinct activities of the mes variant of mPtch1 mediating Hh signaling through both canonical and non-canonical pathways. We demonstrated that mPtch1 bound c-src in an Hh-regulated manner. Stimulation with Sonic Hedgehog (Shh) of primary mammary mesenchymal cells from wild-type and mes animals activated Erk1/2. Although Shh activated c-src in wild-type cells, c-src was constitutively activated in mes mesenchymal cells. Transient assays showed that wild-type mPtch1, mes, or mPtch1 lacking the C terminus repressed Hh signaling in Ptch1-deficient mouse embryo fibroblasts and that repression was reversed by Shh, revealing that the C terminus was dispensable for mPtch1-dependent regulation of canonical Hh signaling. In contrast to these transient assays, constitutively high levels of mGli1 but not mPtch1 were present in primary mammary mesenchymal cells from mes mice, whereas the expression of mPtch1 was similarly induced in both mes and wild-type cells. These data define a novel signal transduction pathway involving c-src that is activated by the Hh ligands and reveals the requirement for the C terminus of Ptch in regulation of canonical and non-canonical Hh signaling pathways.

Introduction

Intracellular transmission of the signals induced by the Hedgehog ligands in vertebrates is facilitated by one of two closely related, 12-pass integral membrane proteins, Ptch1² and Ptch2 (1, 2). The best studied of these, Ptch1, acts in collaboration with one of the coreceptors, BOC (Brother of CDO), CDO (CAM-related/down-regulated by oncogenes), or Gas1 (3–8), to bind one of the Hh ligands. Ligand binding causes inhibition of Ptch1, thereby activating, through an indirect mechanism, the seven-pass, small G protein-linked membrane protein Smoothened (Smo) (9–13)). In turn, Smo activity facilitates conversion of the transcriptional mediators of the canonical signaling pathway, Gli2 and Gli3, into transcriptional activators (14). In the absence of Hh ligands, Ptch1 indirectly suppresses Smo activity, leading to the proteolytic conversion of Gli2 and Gli3 into transcriptional repressors. Thus, in this basic molecular genetic pathway, Ptch1 acts as a repressor of Hh signaling, and, by virtue of the malignancies arising in specific tissues in its absence, exhibits tumor suppressor activity (15–18).

Although the mechanisms involved in the Hh ligand-dependent regulation of the transcriptional activity of the Gli proteins have been well characterized, it is apparent that a number of other signal transduction pathways are regulated by the Hh ligands. For example, as has been well described by Riobo and co-workers (19, 20) and Riobo et al. (21), a number of signaling cascades, including those involving Rac1, protein kinase C, and MAP kinase, respond to Hh signaling in a Smo-dependent manner. Some of these appear to modulate the output of the Hh pathway itself by modifying the activity of specific components of the canonical pathway. However, the overall contribution of these other pathways in the control of cell fate or, more generally, embryonic development has not been determined genetically.

In contrast to pathways stimulated by the Hh ligands in a Smo-dependent manner, several observations have also shown that Ptch1 may mediate Hh signaling independently of Smo or other components of the canonical pathway. For example, a mutation in the C terminus of Ptch1 that is found in FVB mice affects binding of the Tid1 tumor suppressor protein to this region of Ptch1 (22). This mutation segregates with susceptibility of these mice to formation of squamous cell carcinomas in response to an oncogenic H-Ras transgene. Evidence for binding of the cdk1-cyclin B1 cocomplex directly to Ptch1 in an Hh-dependent manner has also been described, implying a role for regulation of the cell division cycle through Hh ligand-dependent modulation of the localization of factors required for the transition through mitosis (23–25).

We described previously a potential signaling pathway stimulated by the Hh ligands in the absence of Smo or signaling through the canonical Hh pathway (26). Specifically, it has been demonstrated that N-Shh could activate Erk1/2 in cell lines where Smo could not be detected and where signaling through the canonical Hh pathway could not be stimulated. That Shh might signal through these pathways was initially suggested following inspection of the primary sequence of C terminus of Ptch1. This region encodes a number of highly conserved motifs that were predicted to facilitate binding of factors harboring SH2, SH3, or WW domains. Indeed, in GST pulldown and immunoprecipitation assays, we demonstrated that a number of factors encoding these domains could complex the Ptch1 C terminus (26).³

One of these was the SH3 domain of the important signal transduction factor c-src, although the interaction with native, full-length c-src was not characterized. However, the significance of this potential interaction was tested genetically using the strain of mice known as mesenchymal dysplasia (mes). These mice harbor a 32-bp deletion in one of the exons encoding the cytoplasmic C-terminal domain of Ptch1 (27, 28). This deletion alters the frame of most of the C terminus, causing a replacement of the last 220 amino acids with an unrelated, 68-amino acid nonsense peptide. Considered a hypomorphic allele of Ptch1, mes mice exhibit a limited spectrum of defects, including preaxial polydactyly, a thickened dermis because of precocious hair follicle development (29), and a bias for preadipocytes to differentiate into brown fat at the expense of white fat (30). We also demonstrated that mes exhibited a block to ductal elongation in the mammary gland at puberty (31, 32). On the basis of the potential interaction between Ptch1 and c-src, mice expressing an activated allele of c-src under the control of the MMTV promoter were bred onto the mes background (31). MMTV-c-src^Act rescued defective mammary gland development in the mes mice, albeit with altered kinetics for ductal elongation. These data suggested, therefore, that c-src played a role in the Hh signaling pathway by virtue of its ability to rescue the block of mammary gland development.

Although mes is thought to be a hypomorphic allele of mPtch1, it appears to repress Hh signaling at levels indistinguishable from wild-type Ptch1, determined in reconstitution assays in Ptch1-deficient mouse embryo fibroblasts (MEFs), and complexes N-Shh ligand with similar avidity as the wild-type protein (29). Thus, the basis for the contrasting outcomes in the dermis/hair follicle versus mammary gland development because of the mes allele is difficult to reconcile with the apparently similar activities of mes and wild-type Ptch1.

We describe here experiments that examine the regulation of both canonical and non-canonical Hh-pathways stimulated by N-Shh in mes and wild-type mPtch1-expressing cells. Our data reveal a previously uncharacterized signaling pathway that is stimulated by N-Shh, resulting in activation of c-src as well as the Hh ligand-regulated association of c-src to mPtch1. Furthermore, although both Ptch1 and mes can repress canonical Hh signaling to similar levels in transient assays, primary mammary mesenchymal cells derived from mes mice reveal the constitutive activation of c-src as well as distinct deregulation and induction of the Hh target genes Gli1 and mPtch1.

MATERIALS AND METHODS

Cell Culture

HEK 293 cells (a gift from S. Girardin) and Ptch1^−/− MEFs (a gift from S. Angers) were cultured in DMEM with 10% FBS and 1% penicillin-streptomycin. MCF10A cells were cultured in DMEM/F12 with 5% horse serum, 10 μg/ml insulin, 0.5 μg/ml hydrocortisone, 20 ng/ml EGF, 1 ng/ml cholera toxin, and 1% penicillin-streptomycin. Shh Light II fibroblasts (ATCC) were cultured in DMEM. The hybridoma cell line 5E1 (University of Iowa Hybridoma Bank) was cultured in DMEM/F12 with 10% FBS.

For serum starvation of MCF10A cells, cells were trypsinized, replated in growth medium until attached, and then changed to DMEM/F12 with 0.5% serum. After 24 h of serum starvation, the cells were stimulated with Shh- or pcDNA3-conditioned medium for 1 h and lysed in n-dodecyl-β-d-maltoside (DDM) lysis buffer.

Preparation of Shh-conditioned Media

Shh- and pcDNA-conditioned media were prepared by transfecting 40% confluent 100-mm plates of HEK 293 cells in 5% FBS with 15 μg of pcDNA3.1-N-Shh (a gift from J. Filmus, Sunnybrook Health Sciences Centre) or pcDNA3 using 2 mg/ml PEI at a 2:1 ratio. Cells were grown for 4–5 days, and the medium was harvested after centrifugation at 2500 rpm for 5 min at 4 °C. The supernatant was then sterile-filtered with a 0.22-μm syringe filter. Prior to use, conditioned medium was diluted 10× in serum-free medium to a final serum concentration of 0.5%. The activity of conditioned medium was measured by luciferase assay in Shh Light II fibroblasts as described below.

Primary Cell Culture

Wild-type and mes (The Jackson Laboratory) littermates on the C57Bl/6 (Charles River Laboratories) background were sacrificed at 3 months. Thoracic and inguinal mammary glands were then dissected and minced with surgical scissors into fragments that were as small as possible. The minced mammary glands were incubated in a solution of 3 mg/ml collagenase A in DMEM/F12 for 45 min at 37 °C. Mechanical dissociation was then performed by slowly pipetting 10–15 times with a 5-ml pipette, adding FBS to a final concentration of 2%, and then pipetting vigorously 15 times. Epithelial (pelleted) and mesenchymal (supernatant) layers were separated by centrifuging at 250 rpm for 2 min. The mesenchymal layer was then strained using a 40-μm cell strainer; washed with 5 ml of DMEM/F12; and plated in DMEM/F12 with 10% FBS, 10 μg/ml insulin, 10 μg/ml transferrin, and 20 ng/ml EGF.

For c-src and Erk1/2 activation assays, primary mouse mammary mesenchymal cells were starved for 48 h in serum-free DMEM/F12 and then stimulated for 1 h with 1 μg/ml N-Shh peptide (R&D Systems) or 200 ng/μl EGF. Cells were then lysed in 1% DDM lysis buffer.

For canonical Hh pathway activation assays, primary mammary mesenchymal cells were starved as described above and then stimulated for 24h with 1 μg/ml N-Shh peptide (R&D Systems), N-Shh peptide heat-killed at 55 °C for 30 min, or N-Shh in the presence of 20 nm SANT-1 (Toronto Research Chemicals). Loss of the activity of the heat-killed N-Shh peptide was verified using Shh Light II fibroblasts as described below.

Cloning

Wild-type, full-length mPtch1 cDNA, HA₃-tagged at its C terminus, and the full-length mes mPtch1 mutant in pcDNA3 described previously (29) were gifts from Prof. C. C. Hui (Hospital for Sick Children Research Institute). The untagged, full-length, wild-type version of mPtch1 was produced by replacing the C-terminal and HA₃ tag with untagged wild-type C terminus cDNA we isolated previously (26). The mPtch1ΔC mutant (Δ1173–1311) was produced by truncation of the full-length cDNA at the PflM1 site (nucleotide 3519). The mPtch1ΔML mutant (Δ614–709) was produced by cutting out the sequence between the StuI site (nucleotide 1845) and the XhoI site (nucleotide 2127), blunt-ending the XhoI site with Klenow, and relegating the vector. The compound mPtchΔMLΔC mutant was produced from the mPtch1ΔML mutant and cutting off the C terminus at the PflMI site as described for the mPtchΔC mutant.

Western Blotting and Immunoprecipitation

Cell lysates were prepared following addition of 0.5 ml DDM lysis buffer containing protease and phosphatase inhibitors (50 mm Tris (pH 7.6), 150 mm NaCl, 2 mm EDTA, 1% n-dodecyl-β-d-maltoside, 0.57 mm PMSF, 10 μm leupeptin, 0.3 μm aprotinin, 10 mm NaF, and 1 mm sodium orthovanadate). For straight Western blot analyses, 4× SDS loading buffer (50 mm Tris (pH 6.8), 100 mm DTT, 2% SDS, 0.1% bromphenol blue, and 10% glycerol) was added to 40 μg of lysate and incubated for 15 min at 37 °C to prevent aggregation of mPtch1 protein. Samples were resolved by 10% SDS-PAGE and blotted onto a nitrocellulose membrane. Blots were probed with primary antibody overnight at 4 °C. Antibodies and dilutions used were as follows: 1:1000 goat α-Ptch1 (Santa Cruz Biotechnology, catalog no. sc-6149), 1:1000 rabbit α-c-src (Cell Signaling Technology), 1:1000 rabbit α-p-src-416 (Cell Signaling Technology), 1:1000 mouse α-non-p-src-416 (Cell Signaling Technology), 1:5000 rabbit α-actin (Sigma), 1:1000 mouse α-p-ERK (Cell Signaling Technology), 1:1000 rabbit α-ERK (Cell Signaling Technology), 1:1000 rabbit α-phosphotyrosine (BD Biosciences), and 1:1000 mouse α-HA (Applied Biological Materials). Blots to be reprobed were stripped in stripping buffer (2% SDS, 62.5 mm Tris (pH 6.8), and 100 mm β-mercaptoethanol) for 30 min at 72 °C.

For immunoprecipitation, 150 μg (HEK 293) or 600 μg (MCF10A) of total cell lysate was incubated with primary antibody overnight at 4 °C in DDM lysis buffer. 15 μl of protein G-agarose beads (Invitrogen) was added to each sample the following day and incubated at 4 °C for 90 min. Beads were washed five times with DDM lysis buffer and resuspended in 25 μl of SDS loading buffer.

Luciferase Assays

Activity of the Shh ligand (peptide or conditioned medium) was assayed using Shh Light II fibroblasts (ATCC). Confluent cells were serum-starved in 0.5% serum for 24 h and then stimulated with conditioned medium or Shh-peptide for 24 h. Cells were then lysed in passive lysis buffer (Promega), and luciferase activity was measured by Dual-Luciferase reporter assay system (Promega). Shh Light II fibroblasts contain an 8×-Gli-Luciferase firefly reporter transgene and a constitutive Renilla luciferase transgene. Gli reporter activity was normalized to Renilla activity.

RT-PCR and Quantitative RT-PCR

For RT-PCR, total RNA was isolated from primary mouse mammary mesenchymal cells using TRIzol reagent (Invitrogen). RT-PCR was performed using a SuperScript III one-step RT-PCR kit (Invitrogen) with 260 ng of RNA per sample. Primer sequences were as follows: mPtch1, 5′ GTCTTGGGGGTTCTCAATG 3′ (forward) and 5′ ATGGCGGTGGACGTTGGGTCCC 3′ (reverse); mGli1, 5′ TGGACTCCATAGGGAGGTGAA 3′ (forward) and 5′ CTCCTCCTCGGAGTTCAGTCA 3′ (reverse); and mGAPDH, 5′ TGAGAACGGGAAGCTTGTCA 3′ (forward) and 5′ GGAAGGCCATGCCAGTGA 3′ (reverse). Reaction conditions were 1 min for each step for 30 cycles. The annealing temperature for mPtch1 was 68 °C and 56 °C for both mGli1 and GAPDH.

For quantitative RT-PCR, total RNA was isolated as described above. DNase treatment (Fermentas) was performed on 400 ng of RNA from each sample. Reverse transcription was then performed using SuperScript II reverse transcriptase (Invitrogen) and random hexamer primers (Fermentas). A 25× dilution of the resulting cDNA was subjected to quantitative PCR using iQ SYBR Green Supermix (Bio-Rad) with 10-μl reactions. Results were quantified using the ΔΔCt method corrected for primer efficiency with Arbp as a reference gene. Primers adapted from sequences published previously (33–35) were as follows: mPtch1, 5′ GGTGGTTCATCAAAGTGTCG 3′ (forward) and 5′ GGCATAGGCAAGCATCAGTA 3′ (reverse); mGli1, 5′ CCCATAGGGTCTCGGGGTCTCAAAC 3′ (forward) and 5′ GGAGGACCTGCGGCTGACTGTGTAA 3′ (reverse); and mArbp, 5′ GAAAATCTCCAGAGGCACCATTG 3′ (forward) and 5′ TCCCACCTTGTCTCCAGTCTTTAT 3′ (reverse). All reactions consisted of 40 cycles with annealing/extension at 60 °C. Primer specificity was confirmed by a combination of agarose gel electrophoresis and melt curve analysis.

RESULTS

Deletion mutant Analysis of the Ptch1-c-src Interaction

We showed previously that the C terminus of Ptch1 was capable of binding to factors harboring SH3 or WW domains (26), assessed, in part, using a GST-SH3 fusion protein derived from c-src. Furthermore, genetic work from our laboratory (31) demonstrated that mammary epithelial cell-restricted expression of an activated c-src transgene (c-src^Act, Refs. 36, 37) rescued the block in mammary gland development caused by mutation of the Ptch1 C terminus in mes mice. To characterize the potential physical interaction between full-length mPtch1 and c-src, coimmunoprecipitation assays were performed using full-length mPtch1, the mes variant of mPtch1, as well as mutants that had deleted specific cytoplasmic domains of this 12-pass integral membrane protein (Fig. 1).

FIGURE 1.

Schematic of Ptch1 structure and mutants. The top schematic shows the predicted structure of the 12-pass mPtch1 protein. Three major cytoplasmic domains, N, ML, and C, are depicted. The bottom schematics depict the wild-type and mutant mPtch1 proteins used in this work. The numbers refer to the amino acids at predicted boundaries of specific domains. ex, extracellular; cyto, cytoplasmic.

HEK293 cells were transfected with a vector expressing activated c-src (c-src^Act) as well as an amount of expression vector for the mPtch1 mutants that produced approximately equal levels of each protein, as determined by immunoblot analysis using an antibody directed to the N-terminal region of mPtch1 (Fig. 2A). Coimmunoprecipitations were then performed using an agarose-conjugated goat α-Ptch1 antibody against the Ptch1 N terminus (Fig. 2B). Immunoprecipitation of the mPtch1 mutants ΔC and ΔMLΔC revealed that loss of the entire C terminus reduced to background levels the ability of mPtch1 to coimmunoprecipitate c-src^Act. In contrast, loss of the large middle intracellular loop in the ΔML mutant did not prevent mPtch from coimmunoprecipitating c-src^Act (Fig. 2B). Wild-type c-src was also assessed for binding to full-length mPtch1. As Fig. 2D illustrates, coimmunoprecipitation of wild-type c-src to wild-type Ptch1 was evident. Interestingly, coimmunoprecipitation assays interrogating the c-src-binding activity of the mes mutant showed no apparent difference in the Ptch1-c-src interaction in wild type Ptch1 versus mes (Fig. 3, A and B). Combined with our previous results (26), these data reveal that the C-terminal cytoplasmic domain of full-length mPtch1 is sufficient and necessary for binding to c-src.

FIGURE 2.

Overexpressed Ptch1 and c-src coimmunoprecipitate via an interaction requiring the Ptch1 C terminus. Ptch1 deletion mutants were transfected into HEK 293 cells with or without activated (A and B) or wild-type (C and D) c-src. Similar expression levels of transfected plasmid were verified by Western blot analysis (A and C). B, immunoprecipitations (IP) using a goat α-Ptch1-agarose conjugate antibody against the Ptch1 N terminus were performed, and the Western blot analyses were probed for Ptch1 and c-src. D, immunoprecipitations were performed using a mouse α-HA antibody, and the resultant Western blot analysis was probed for Ptch1 and c-src. IB, immunoblot.

FIGURE 3.

Overexpressed Ptch1 is a target for tyrosine phosphorylation by overexpressed activated c-src. HEK 293 cells were transfected with the indicated wild-type mPtch1 or mes with or without activated c-src. A, immunoprecipitation (IP) of mPtch1 or mes (top panel) coimmunoprecipitated c-src (bottom panel). IB, immunoblot. B, the reciprocal immunoprecipitation of c-src coimmunoprecipitated mPtch1 and mes. C, reprobe of the immunoprecipitation in A with an anti-phospho-Tyr antibody reveals that mPtch1 and mes are tyrosine phosphorylated in the presence of activated c-src. D, the Ptch1, ΔC, and ΔML mutants were cotransfected without (left panels) or with (right panels) c-src^Act. They were immunoprecipitated and probed for phospho-Tyr (bottom panels) and then reprobed for Ptch1 (top panels).

In addition to containing potential SH3 domain-binding motifs, the C terminus and large middle intracellular loop (amino acids 586–734) of mPtch1 contain multiple tyrosine residues that are potential targets for phosphorylation by c-src. These motifs are conserved between Ptch1 proteins in multiple vertebrates (26).³ To determine whether Ptch1 is a target for tyrosine phosphorylation by c-src, immunoprecipitations were carried out in the same manner as described above, and isolated mPtch1 or mes were probed with an antibody against phospho-tyrosine (Fig. 3C). When cotransfected with c-src^Act, both wild-type mPtch1 and mes were detected using the anti-phospho-tyrosine antibody. The dependence of c-src^Act for phosphorylation is illustrated in Fig. 3D. In the absence of the cotransfected c-src^Act (Fig. 3D, left panels), Tyr phosphorylation could not be detected. The right panels in Fig. 3D show significant Tyr phosphorylation of full-length Ptch1 and a weaker signal for the ΔML deletion mutant that retains c-src^Act-binding activity. For the ΔC mutant, which does not complex c-src^Act, no signal above background for phospho-Tyr was observed. Thus, mPtch1 is a potential target for phosphorylation by c-src kinase.

Hedgehog Ligand Stimulation Inhibits c-src Binding to Ptch1

To further characterize the interaction between Ptch1 and c-src, the effect Shh on Ptch1 binding to endogenous c-src was determined. This analysis was performed using the human immortal mammary epithelial cell line MCF10A because these cells lack detectable expression of Smoothened (26, 38), signaling through the canonical Hh signaling pathway cannot be elicited with Hh ligand (31),³ and ERK1/2 is activated in response to stimulation with N-Shh (26). Thus, the effects of Hh ligand on the association of Ptch with endogenous c-src can be determined in the absence of signaling through the canonical Hh pathway.

MCF10A cells were transfected with mPtch1, grown for 48 h, serum-starved for 24 h, stimulated with Shh-conditioned or control medium for 1 h, and then mPtch1 was immunoprecipitated from 600 μg of lysate. As Fig. 4A reveals, mPtch coimmunoprecipitated endogenous c-src when MCF10A cells were treated with control-conditioned medium. This association was not detectable, however, when these cells were stimulated with conditioned medium containing N-Shh. Thus, the association of endogenous c-src with mPtch1 was disrupted upon treatment with Hh ligand in the absence of signaling through the canonical Hh pathway.

FIGURE 4.

Association of mPtch1 with endogenous c-src. MCF10A cells were transfected with Ptch1-HA, grown for 48 h, serum-starved for 24 h, and then stimulated with Shh- or pcDNA3-conditioned medium for 1 h. A, immunoprecipitation (IP) of mPtch1 (top panel). Endogenous c-Src (bottom panel) coimmunoprecipitates with mPtch1 in cells treated with control medium but not in medium containing Shh. IB, immunoblot. B, straight Western blot analyses of cells probing for mPtch1 and endogenous activated (phosphorylated) c-Src. Increased levels of activated c-Src were observed in serum-starved cells treated with medium containing Shh relative to cells treated with control medium.

The control immunoblot analyses in Fig. 4B revealed further that endogenous c-src was activated upon stimulation with Shh-conditioned medium. Specifically, Fig. 4B, center panel, shows an increase in the c-src signal when probed with an antibody directed against phospho⁴¹⁶ (activated)-c-src relative to the signal for total c-src in the bottom panel.

The ability of Hh ligand to activate c-src was examined further using primary mesenchymal cells (fibroblasts) isolated from the mammary glands of wild-type or mes mice (Fig. 5). In cells isolated from wild-type animals, activation of Erk1/2 was evident 60 min after treatment of serum-starved cells with N-Shh peptide, as we demonstrated previously (26, 31). In addition, a consistent 4-fold activation of c-src was observed for these cells within 60 min of stimulation with ligand.

FIGURE 5.

Activation of c-src in primary mesenchymal cells from wild-type but not mes mice. A, primary mammary mesenchymal cells, isolated from wild-type (left panels) and mes (right panels) littermates, were serum-starved for 48 h and then treated with Shh peptide for 1 h. Immunoblots (IB) were probed for activated (phospho)-c-Src⁴¹⁶ and reprobed for total src (top two panels) or for activated (phospho)-Erk1/2 and then reprobed for total Erk1/2 (bottom two panels). Shh activated both c-src and ERK1/2 in wild-type cells. In mes mesenchymal cells, Erk1/2 was also activated. However, constitutive activation of c-src was seen for the mes cells, this level of activation being unaffected by the addition of Shh. B, quantification of signals from the representative blot in A.

Activation of Erk1/2 and c-src by N-Shh was also examined using primary mesenchymal cells isolated from mes mammary glands (Fig. 5A, right panel). As evident in Fig. 5B, levels in mes fibroblasts of phospho-Erk1/2 were similar to those seen in wild-type cells. Furthermore, activation by N-Shh of Erk1/2 was evident in mes mesenchymal cells, typically to levels relatively higher than those seen for wild-type cells. In contrast to Erk1/2, there was a constitutively high level of activated c-src in mes mesenchymal cells. These higher levels could not be increased further upon stimulation with N-Shh. Thus, the loss of the C terminus of mPtch1 in mammary fibroblasts from mes animals resulted in constitutive activation of c-src relative to cells expressing wild-type mPtch1, whereas the resting levels of activated Erk1/2 and its activation by N-Shh was unaffected by the mes mutation in mPtch1.

Altered Canonical Hh Signaling in Mesenchymal Cells from mes Animals

To further define distinct qualities of the mes mutant, mes activity was assessed in transient assays for its ability to regulate canonical Hh signaling. First, the ability of mes and wild-type mPtch1 to repress Hh signaling was assayed (Fig. 6A). Consistent with previous reports (29), transient expression in mPtch1-deficient MEFs of both mes and wild-type mPtch1 using either 200 or 25 ng of plasmid caused potent repression of Hh signaling, as determined by the reduced expression of luciferase from the 8×-Gli1-luc reporter. Indeed, truncation of all but the first 10 amino acids of the C terminus of mPtch1 in the ΔC mutant imparted no defect on the ability of mPtch1 to repress Smo-dependent signaling. Although repression by both mPtch1 and mes has been reported and the affinity of Shh for each of these have also been shown to be similar (29), the ability of mes and mPtch1 to mediate Hh signaling in reconstituted Pcth1-deficient cells has not been determined. As Fig. 6B demonstrates, repression of Hh signaling in mPtch1-deficient mouse embryonic fibroblasts by the wild type, mes, and the ΔC variant of mPtch1 was reversed following the addition of Shh-conditioned medium. Furthermore, as Fig. 7 illustrates, transiently expressed wild-type mPtch1, mes, and ΔC all localize to cilia in serum-starved mPtch1-deficient mouse embryonic fibroblasts. Thus, the C terminus of mPtch1 appears to be dispensable in fibroblasts for both repression and activation of canonical Hh signaling as well as ciliary localization in transient assays employing overexpressed protein.

FIGURE 6.

The C terminus of mPtch1 is dispensable for the repression and activation of canonical Hh signaling in MEFs. A, Ptch1-deficient MEFs were transfected with an 8× Gli-luciferase reporter, a constitutive Renilla luciferase plasmid, and 25 or 200 ng of the indicated Ptch1 mutant. After 48 h, the cells were lysed, and relative luciferase activity was measured. At both amounts of transfected plasmid, mPtch1, mes, and ΔC all repressed canonical Hedgehog signaling. Data are displayed as mean ± S.D., n = 3. B, cells as described in A were serum-starved in control medium or Shh-conditioned medium for 24 h, and canonical Hh signaling was quantified. Wild-type mPtch1 as well as mes and the ΔC mutants all responded to Hh ligand, revealing that the C terminus is dispensable for activation of canonical Hh signaling. Data were analyzed by two-way analysis of variance, followed by pairwise comparison of means using Tukey's honest significant difference test. Data are displayed as mean ± S.D., n = 6. *, p < 0.05; ***, p < 0.001.

FIGURE 7.

Localization of mPtch1, mes, and ΔC to cilia in Ptch1-deficient MEFs. Wild-type mPtch1 (top panels), mes (center panels), or ΔC (bottom panels) were expressed at equal amounts in Ptch1-deficient MEFs. Following 48 h of serum starvation, fixed cells were probed simultaneously for mPtch1 (green channel) and acetylated tubulin (acTubulin, red channel). All three variants of mPtch1 localized to the cilia regardless of the presence of the C-terminal domain of mPtch1.

Although transient assays did not distinguish the activities of mes versus wild-type mPtch1, the different effects on activation of c-src versus Erk1/2 (Fig. 5) suggest that cells expressing endogenous wild-type mPtch1 versus mes might also differentially regulate distinct transcriptional targets of canonical Hh signaling. Previous molecular studies on the effect of the mes allele on canonical Hh pathway target activation have all involved assaying the steady-state levels of Hh target genes by end point RT-PCR (29–31). These assays have not, however, assayed the ability of the endogenous mes variant of mPtch1 to respond to stimulation by the Hh ligands. Thus, to identify whether differences exist in the activation of Shh target genes in cells from mes mice, activation assays were performed using primary mouse mammary mesenchymal cells isolated from wild-type and mes littermates. Mammary mesenchymal cells (fibroblasts) were serum-starved for 48 h and then stimulated for 24 h with N-Shh peptide, heat-killed N-Shh peptide, or Shh peptide with 20 nm of the Smo inhibitor SANT-1. Fig. 8 shows that, in a qualitative end point RT-PCR analysis, mesenchymal cells isolated from both wild-type and mes animals expressed relatively low basal levels of mPtch1, as expected for cells unstimulated by Hh ligand (39). The levels of mPtch1 were increased 24 h following addition of Shh to mesenchymal cells from both wild-type and mes animals. In contrast to wild-type cells, mesenchymal cells from the mes animals expressed a clearly detectable Gli1 message at levels that were not apparent in wild-type mammary fibroblasts. Furthermore, wild-type mesenchymal cells responded to stimulation by Shh with the expected increase in Gli1 expression, whereas Gli1 levels in the mes mesenchymal exhibited no apparent change.

FIGURE 8.

Primary mes mammary mesenchymal cells display constitutive activation of Gli1. Primary mammary mesenchymal cells isolated from wild-type and mes littermates were serum-starved for 48 h and then stimulated for 24 h with 1 μg/ml of N-Shh peptide, heat-killed N-Shh peptide, or N-Shh peptide in the presence of 20 nm SANT-1. A, semi-quantitative RT-PCR indicates that Ptch1 mRNA levels in cells from both animals respond similarly in response to Shh stimulation, whereas mes mesenchymal cells display a constitutively high level of Gli1 expression in the absence of Shh ligand that is not observed in mesenchymal cells from wild type animals. B, quantitative RT-PCR of RNA from a separate cell isolate shows increased basal levels of Ptch1 and Gli1 in mes mammary mesenchymal cells. mPtch1 is induced by Shh ∼5 to 8-fold in wild-type and mes mesenchymal cells. However, the constitutively high levels of Gli1 expression in mes mesenchymal cells are only marginally increased, whereas the expected very strong increase of Gli1 message because of stimulation by Shh was seen for wild-type cells.

Altered expression levels of these two Hh pathway target genes were then quantified using quantitative PCR (Fig. 8B). A consistent 5- to7-fold higher steady-state level of mPtch1 was seen in mes mesenchymal cells relative to wild-type cells, whereas a 5- to 8-fold-increase of expression of mPtch1 in response to N-Shh ligand was similar for cells derived from either animal. In contrast to mPtch1, the expression of mGli1 in mes appeared to be regulated distinctly. Specifically, in starved wild-type cells, the levels of mGli1 were close to background but were stimulated strongly by the addition of Shh ligand, as demonstrated previously (36, 37). However, constitutively high levels of mGli1 were present in serum-starved cells from the mes mammary gland, and less than a 2-fold increase in these levels was observed following stimulation by Shh. As also evident in the RT-PCR analysis in Fig. 8A, the expression of both mPtch1 and mGli1 were sensitive to the Smo inhibitor SANT-1, consistent with their expression in these cells to be dependent, at least in part, on the canonical Hh signaling pathway. These results indicate that the C terminus of Ptch1 plays a role in regulating canonical Hh pathway output in primary mammary mesenchymal cells.

Taken together, these data reveal the regulation of a novel non-canonical Hh signaling pathway operating through c-src as well as the Hh-regulated association of endogenous c-src with mPtch1. Furthermore, mutation of the C terminus of mPtch1, as occurs in mes mice, results in differential alterations in the activation of distinct canonical and non-canonical pathways in primary cells that are not apparent in assays using transiently expressed proteins.

DISCUSSION

The cytoplasmic C-terminal domain of the Ptch1 protein harbors a number of motifs predicted to bind to SH2-, SH3-, and WW-containing factors. Although these motifs are highly conserved among vertebrates, the sequence of the analogous region in the Ptch protein in Drosophila appears relatively unrelated, suggesting that this region of Ptch1 may have acquired novel biological activities in vertebrates. Thus, despite the well established role in the canonical Hh signaling pathway for Ptch1, sequences in the vertebrate Ptch1 C terminus predict that Ptch1 may also directly regulate a number of signal transduction cascades independently from its role in controlling Smo activity. Indeed, reports exist demonstrating this role for Ptch1 (for example, see Ref. 22), including our previous study reporting the ability of Ptch1 to complex SH3-containing factors and to activate Erk1/2 in a Smo-independent manner (26). Here, we describe another novel interaction in which c-src was activated by N-Shh and where endogenous c-src associated with mPtch1 when cells were not stimulated with N-Shh ligand. This association was observed in a cell line, MCF10A, where message for Smo cannot be detected and where signaling through the canonical Hh pathway cannot be stimulated using Hh ligand (31).³

We propose that the Ptch1 receptor represents a distinct factor that contributes to the regulation of signal transduction pathways operating through c-src. That c-src is relevant to Hh signaling was demonstrated in a genetic analysis (31). Specifically, in mes mice, a block to ductal elongation was evident in homozygous animals at puberty (31, 32). However, this block was overcome, albeit with altered kinetics relative to wild-type animals, when an activated variant of c-src was expressed in mammary luminal epithelial cells under the control of the MMTV promoter. Thus, both genetic and biochemical evidence exists for c-src activity playing a role in the Hh signaling pathway. Given the well established role for c-src in the transformation of cells in a spectrum of tissues (for reviews, see Refs. 38–40), these data further suggest that the malignant transformation associated with the loss of the Ptch1 may arise, in part, because of contributions from deregulation of signaling through c-src. Indeed, we observed that, in mammary mesenchymal cells from mes animals, constitutively high levels of activated c-src were present relative to their levels in wild-type cells. However, although c-src may contribute to transformation of Ptch1-deficient cells, it is clear that deregulation of the canonical Hh signaling pathway likely plays a more central role in transformation, as has been shown in many recent studies where growth of transformed cells was retarded using small-molecule inhibitors of Smo (41, 42). The potential contribution of c-src activity to Hh pathway-dependent cell transformation could be assessed, for example, in a manner similar to its role in ErbB2-dependent transformation (43–45).

Our data also support the significance of c-src activity in the Hh-dependent regulation of mammary gland development. Specifically, the block to pubescent mammary gland development in mes mice was associated with a strong reduction in estrogen receptor α expression. However, estrogen receptor α expression was induced in a compound mes/MMTV-C-src^Act mammary gland, suggesting that, for this tissue, pathways regulating the expression and/or activity of estrogen receptor α, a factor essential for mammary gland development, requires the wild-type activities specifically associated with the C terminus of Ptch1. Thus, we predict that components of the Hh pathway, specifically Ptch1, play an integral role in the regulation of proteins required for normal development of the mammary gland. Loss of some of the activities of Ptch1, as occurs in the mes mouse, reveals further that this potential role for Ptch1 may be mediated, at least in part, through its C terminus.

The data presented here also reveal the distinct regulation of two well characterized transcriptional targets of the Hh signaling pathway. Specifically, although both Gli1 and Ptch1 expression was induced by the Hh ligands (36, 37, 45, 46), cells expressing the mes variant of Ptch1 displayed high levels of Gli1 that showed a less than 2-fold increase upon addition of N-Shh. Increased Ptch1 levels were also observed in mes primary mesenchymal cells. However, these produced a similar relative increase to wild-type cells in response to Shh stimulation. Together, these data indicate that, in these primary mammary mesenchymal cells, the mes allele may not only exhibit weakened suppression of Smo activity but may also perturb an additional mechanism of regulating the expression of Gli1. It is interesting to note, however, that the elevated pathway activity in mes primary cells is fully sensitive to SANT-1 treatment, suggesting that Smo-independent pathways regulating Gli1 expression regulation are not involved.

In contrast to these transcriptional targets, two non-canonical signal transduction cascades were differentially regulated in mes-expressing cells. Using wild-type mammary mesenchymal cells, activation of both c-src and Erk1/2 was seen following the addition of N-Shh. Erk1/2 was also activated in mesenchymal cells from mes mice, whereas constitutive activation of c-src was evident in these same cells. Thus, these two non-canonical cascades acting through Ptch1 are regulated differentially by the activities of the C-terminal region of this receptor. We expect, therefore, that, like other classes of receptors harboring multiple docking sites for SH3 and SH2 domain proteins (for example, see Refs. 43, 44), that Ptch1 activity and transmission of signals by the Hh ligands through Ptch1 may involve sequential binding of multiple factors and modification (e.g. phosphorylation status) of regulatory sequences in Ptch1.