Research Article

Two Patched molecules engage distinct sites on Hedgehog yielding a signaling-competent complex

See allHide authors and affiliations

Science  05 Oct 2018:
Vol. 362, Issue 6410, eaas8843
DOI: 10.1126/science.aas8843

It takes two to signal

The Hedgehog (HH) signaling pathway is important in development, and excessive HH signaling is associated with cancer. Signaling occurs through the G protein–coupled receptor Smoothened. The pathway is repressed by the membrane receptor Patched-1 (PTCH1), and this inhibition is relieved when PTCH1 binds the secreted protein HH. Two recent papers have described structures of HH bound to PTCH1, but surprisingly, each described a different binding epitope on HH. Qi et al. present a cryo–electron microscopy structure that explains this apparent contradiction by showing that a single HH protein uses both of these interfaces to engage two PTCH1 receptors (see the Perspective by Sommer and Lemmon). Functional assays suggest that both interfaces must be bound for efficient signaling.

Science, this issue p. eaas8843; see also p. 26

Structured Abstract

INTRODUCTION

Appropriate Hedgehog (HH) signaling is important for human health; excessive HH signaling leads to various cancers, whereas insufficient HH signaling results in birth defects. The mature N-terminal domain of the HH protein with N-terminal palmitate and C-terminal cholesterol modifications (referred to as “native HH-N” here) binds to its receptor, Patched-1 (PTCH1), to release its repression of the HH pathway, presumably by regulating the transport of a ligand that regulates the downstream oncoprotein Smoothened. The palmitate moiety of HH-N is essential for HH signaling; however, the mechanism of how the palmitate moiety of HH induces the signaling is not clear. The means by which HH protein recognizes PTCH1 is also controversial: A recent 1:1 PTCH1–HH-N complex structure visualized a palmitate-mediated binding site on HH-N, which was inconsistent with previous studies that implied a distinct, Ca2+-mediated interface for PTCH1 and HH co-receptors.

RATIONALE

Cell biological studies indicated that mutations on either the palmitate-dominated interface or the Ca2+-mediated interface of HH-N eliminated signaling, indicating that both HH-N interfaces are necessary for HH to inhibit PTCH1 activity. We managed to assemble a 2:1 PTCH1–HH-N complex using a biologically functional PTCH1 variant (PTCH1*) and native Sonic HH-N (SHH-N) in the presence of 1 mM CaCl2 (a physiological extracellular calcium concentration) for structural investigation.

RESULTS

The cryo–electron microscopy (cryo-EM) structure of native SHH-N in complex with PTCH1* at 3.5-Å resolution demonstrates that one SHH-N molecule concomitantly engages both epitopes and binds two PTCH1 receptors (PTCH1-A and PTCH1-B) in an asymmetric manner. PTCH1-A and PTCH1-B exhibit several conformational differences and bind distinct portions of SHH-N. Structural comparisons suggest that the Ca2+-mediated interface of SHH-N physiologically binds different HH binders, including PTCH1, HH co-receptors, and the SHH-N antagonist 5E1 antibody, to regulate HH signaling. Functional assays using PTCH1 or SHH-N mutants that disrupt the individual interfaces illustrate that simultaneous engagement of both interfaces between SHH-N and PTCH1 is essential for maximal signaling in cells.

Calculations using MOLE reveal a tunnel, ~150 Å in length with a minimum radius of 4 Å, that stretches through PTCH1-B but not PTCH1-A, which interacts with the palmitate of SHH-N. The path of this PTCH1-B tunnel includes the extracellular domain I (ECD-I), the cavity of the ECDs that accommodates the palmitate, and the transmembrane region. We introduced two point mutations (L427R and F1017E) on full-length human PTCH1 to block this tunnel by a ~3-Å salt bridge instead of by palmitate insertion. Cell biological assays show that unlike wild-type PTCH1, this variant was not able to repress the HH pathway; this finding supports our hypothesis that the palmitate can abolish the activity of PTCH1 by blocking the cavity of the ECDs.

CONCLUSION

Our structure of a signaling-competent complex consisting of SHH-N and an asymmetric arrangement of two PTCH1* molecules reconciles previous disparate findings of how HH-N interacts with PTCH1. The structural analysis reveals a tunnel in the PTCH1 molecule that can be blocked using mutations in the ECDs that mimic the palmitate insertion, thus abolishing HH pathway repression by PTCH1. Thus, our work illuminates how two sites on SHH-N serve to unite two PTCH1 receptors, thereby initiating a full HH signal.

Cryo-EM structure of 2:1 human PTCH1*–SHH-N signaling-competent complex.

PTCH1-A, PTCH1-B, and SHH-N are colored yellow, light blue, and cyan, respectively, and are viewed from the side of the membrane. Palmitate (PLM) is shown as magenta sticks, calcium as green spheres. Cholesterol in the cell membrane is shown as ball-and-stick models colored yellow and red.

Abstract

Aberrant Hedgehog (HH) signaling leads to various types of cancer and birth defects. N-terminally palmitoylated HH initiates signaling by binding its receptor Patched-1 (PTCH1). A recent 1:1 PTCH1-HH complex structure visualized a palmitate-mediated binding site on HH, which was inconsistent with previous studies that implied a distinct, calcium-mediated binding site for PTCH1 and HH co-receptors. Our 3.5-angstrom resolution cryo–electron microscopy structure of native Sonic Hedgehog (SHH-N) in complex with PTCH1 at a physiological calcium concentration reconciles these disparate findings and demonstrates that one SHH-N molecule engages both epitopes to bind two PTCH1 receptors in an asymmetric manner. Functional assays using PTCH1 or SHH-N mutants that disrupt the individual interfaces illustrate that simultaneous engagement of both interfaces is required for efficient signaling in cells.

Appropriate Hedgehog (HH) signaling is important for human health: Excessive HH signaling leads to various cancers, whereas insufficient HH signaling results in birth defects (16). Nascent HH proteins enter the secretory pathway. After signal peptide cleavage, HH undergoes autocatalytic processing between its N- and C-terminal domains (HH-N and HH-C) in the endoplasmic reticulum (7). During this reaction, cholesterol is covalently linked to the C terminus of HH-N, which is necessary for its trafficking and distribution (8). Hedgehog acyltransferase subsequently adds palmitate to the α-amino group of the N-terminal cysteine of HH-N (9, 10). This modification is indispensable in activating the HH pathway (914). Activation occurs when the lipid-modified HH-N binds to its receptor, Patched-1 (PTCH1) protein (15). Notably, a palmitoylated SHH-N-terminal peptide bound to PTCH1 was found to partially stimulate HH signaling while the globular portion of HH could increase the binding to PTCH1 and facilitated PTCH1 export from cilia (16). In the absence of HH-N binding, PTCH1 represses the HH pathway presumably by regulating the transport of a ligand of its downstream protein, Smoothened (SMO) (2, 17, 18). Patched is a tumor suppressor involved in basal cell carcinoma and medulloblastoma (19), and SMO is an oncoprotein and a target of antitumor agents (20, 21).

We recently determined the cryo–electron microscopy (cryo-EM) structure of a functional human PTCH1 protein (PTCH1*) alone and in complex with Sonic Hedgehog (SHH-N) (22). Although PTCH1* lacks the internal loop between transmembrane helices 6 and 7 (TM6 and TM7) and the C-terminal cytoplasmic domain, it can still repress the HH pathway and be regulated by SHH-N, comparable to wild-type full-length PTCH1 in Ptch1–/– mouse embryo fibroblasts (MEFs) (22, 23). The structure shows that PTCH1* comprises two structurally homologous extracellular domains (ECDs) and 12 TMs. Notably, an endogenous ligand observed in the sterol-sensing domain (SSD), consisting of TMs 2 to 5, suggests that the SSD may play a role in the transport and/or regulation of the SMO ligand. The structure of the complex shows a palmitate-dominated interface between SHH-N and PTCH1 (22); this interface is opposite the interface of SHH-N with its co-receptors (24), which had been implicated in PTCH1 binding in previous studies (2528) and can be regulated by calcium (24). Consistent with this, we reported biochemical assays showing that SHH-N lacking palmitate interacts with PTCH1 through the Ca2+-mediated interface. However, unpalmitoylated SHH-N fails to stimulate HH signaling in the cell (16, 22).

Assembly of 2:1 PTCH1*–SHH-N complex

The extracellular Ca2+ concentration is typically in the range of 1 to 2 mM (29), and mutations on either the palmitate-dominated interface (22) or the Ca2+-mediated interface (30) attenuated HH signaling, indicating that both SHH-N interfaces are necessary for maximal HH signaling in the cell. Previously, Ca2+ was not included in our purification buffer. To enable the formation of either or both interfaces, we mixed PTCH1* with native SHH-N containing its N-terminal palmitate and C-terminal cholesterol modifications in the presence of 1 mM CaCl2. A 2:1 PTCH1*–SHH-N complex was eluted ahead of the 1:1 complex by gel filtration and subjected to cryo-EM studies (Fig. 1A and fig. S1A).

Fig. 1 Overall structure of 2:1 PTCH1*–SHH-N complex.

(A) Cryo-EM 2D classification from RELION. (B) Cryo-EM map after Frealign refinement sharpened using BFACTOR.EXE with a resolution limit of 3.5 Å and a B-factor value of –100 Å2. (C) Ribbon representation of the structure viewed from the side of the membrane. In this and succeeding figures, calcium and zinc are shown as spheres in green and gray, respectively; palmitate (PLM) is shown in magenta. (D) The ECDs (top view). (E) The transmembrane domains (top view).

We determined this structure at 3.5-Å resolution (Fig. 1B, fig. S1, B to D, and table S1). One SHH-N forms a complex with two PTCH1* molecules (PTCH1-A and PTCH1-B) using two distinct interfaces (Fig. 1C). The SHH-N palmitate inserts into the ECD cavity of PTCH1-A, and its Ca2+-mediated interface directly binds the ECDs of PTCH1-B (Fig. 1C). PTCH1-A uses its ECD-I tobind the globular portion of SHH-N, while PTCH1-B engages both of its ECDs to bind the Ca2+-mediated epitope (Fig. 1D). The interface between PTCH1-A and SHH-N is ~2300 Å2 (including 350 Å2 to accommodate the palmitate), whereas the interface between PTCH1-B and SHH-N is ~1100 Å2. The transmembrane domains of the two PTCH1* molecules are arranged asymmetrically, with the TMs of PTCH1-B rotated about 150° relative to the TMs of PTCH1-A (Fig. 1E).

Interfaces between SHH-N and two PTCH1* molecules

The current map allows us to visualize the interactions between SHH-N and the two PTCH1* molecules (figs. S2 and S3). There are several hydrogen bonds and salt bridges between the N-terminal residues of SHH-N and the PTCH1-A ECDs, which form the palmitate-dominated interface. Only five residues of ECD-II (Asn802, Tyr804, Asn940, Gln973, and Tyr1013) interact with the N-terminal peptide (Np) and the palmitate of SHH-N; however, more than 10 residues from ECD-I contribute to this interaction (Fig. 2, A and B). A previous study showed that the mutation Pro26 → Ala on SHH-N impairs signaling (16). Our structure reveals the molecular basis of this effect: The proline facilitates palmitate insertion into the hydrophobic cavity by keeping the N terminus of SHH-N in a bent position (Fig. 2C). In addition to the palmitate insertion, the palmitate-dominated interface includes α5 of PTCH1-A recognizing α1 and the N terminus of α2 of SHH-N (Fig. 2A). We previously showed that mutations on α5 of PTCH1 abolished PTCH1*–SHH-N binding and HH signaling (22).

Fig. 2 The two interfaces of SHH-N bound to PTCH1*.

(A) Details of α1 and α2 of SHH-N binding to PTCH1-A. (B) Details of the interaction between the N terminus of SHH-N and PTCH1-A. (C) The bend in the N terminus of SHH-N. (D to F) Interaction detail of the Ca2+-mediated interface of SHH-N to PTCH1-B residues. (G) Summary of the residues of PTCH1* bound to the distinct interfaces of SHH-N. M, main chain–involved interaction. (H) Structural comparison of SHH-N in 2:1 and 1:1 PTCH1*–SHH-N complexes. Amino acid residues: A, Ala; C, Cys; D, Asp; E, Glu; F, Phe; G, Gly; H, His; I, Ile; K, Lys; L, Leu; M, Met; N, Asn; P, Pro; Q, Gln; R, Arg; S, Ser; T, Thr; V, Val; W, Trp; Y, Tyr. Residues of SHH-N are underlined.

The PTCH1-B molecule also uses its α5 helix and several loops of ECD-I to form the Ca2+-mediated interface of SHH-N (Fig. 2, D and E). In addition, three residues (Glu947, Asp951, and Asp954) in an ECD-II loop (D loop) form salt bridges with Arg153 and Arg155 of SHH-N (Fig. 2F). The D-loop is not used in the PTCH1-A palmitate-mediated interaction. In the context of SHH-N without palmitate, mutation of Arg153 causes loss of binding to PTCH1 (23). In the context of SHH-N with palmitate, the mutant also fails to stimulate HH signaling (30), consistent with the requirement for both interfaces. A structure of PTCH1 with unpalmitoylated SHH-N was recently published (31), which is consistent with our observation of the interface between PTCH1-B and native SHH-N.

PISA (proteins, interfaces, structures, and assemblies) server analysis (32) calculates that the ΔiG between the palmitate and its dominated interface to PTCH1-A is –12.1 kcal/mol (P = 0.232), more stable than the calcium-mediated interface to PTCH1-B (ΔiG = –1.1 kcal/mol, P = 0.717). Overall, across both interfaces, residues in ECD-I contribute to more than 80% of the interactions with SHH-N. Residues Asp217 and Glu221 in helix α5 and Gly378 in the loop connecting α3 and α3′ function in both interfaces; however, no residues in ECD-II are used in both of the two interfaces at the same time (Fig. 2G). Residues in ECD-II bind to the globular portion of SHH-N on the Ca2+-mediated interface, and different residues in ECD-II are involved in the palmitate insertion. This distinction introduces an asymmetric arrangement of two PTCH1* molecules in the complex.

Superimposing PTCH1-A and SHH-N from the 2:1 complex with the PTCH1*–SHH-N 1:1 complex (PDB code 6D4J) reveals a slight change in the palmitate-dominated interface. SHH-N in the 2:1 complex has moved away from α5 of PTCH1-A as a result of the interaction of PTCH1-B with the Ca2+-mediated interface (Fig. 2H). Intriguingly, there is a rod-like density of unknown identity among helices α5 to α7 of ECD-I in both PTCH1* molecules (fig. S4). This ligand may stabilize α5 of ECD-I in the conformation that binds SHH-N or could be a ligand that is regulated by PTCH1 in the HH pathway; however, this molecule and its precise function require identification and further analysis.

Structural comparison of PTCH1-A and PTCH1-B after SHH-N binding

PTCH1-A uses its ECD-I for binding the globular domain of SHH-N, and PTCH1-B uses a similar area of its ECD-I for SHH-N recruitment (Fig. 1C). To compare the conformations of the two PTCH1* molecules in the complex, we superimposed PTCH1-A and PTCH1-B: The α3 and α3′ helices of ECD-I in PTCH1-A are moved outward relative to PTCH1-B to accommodate the N terminus of SHH-N that inserts into the cavity of PTCH1-A (Fig. 3A). The Np and the palmitate of SHH-N trigger a loop in ECD-I to move, allowing insertion (Fig. 3A, red frame). The α5 in ECD-I is in distinct conformations in PTCH1-A and PTCH1-B because it is involved in different interfaces with SHH-N (Fig. 2, A and E, and Fig. 3A). The loop between α6 and α7 in ECD-I of PTCH1-B is shifted as a result of its interaction with Glu89 in SHH-N (Figs. 2D and 3B). In the transmembrane domain, TM4 and TM12 have moved slightly in PTCH1-A relative to PTCH1-B (Fig. 3, C and D). There are some conformational changes between the two small α helices following TM1 and TM7 (pre-α1) and α4 of ECD-I, which form the cavity that accommodates the palmitate (Fig. 3C and fig. S3).

Fig. 3 Structural comparison of the two PTCH1* molecules.

(A) Movement of structural elements of ECD-I in PTCH1-A after SHH-N insertion. (B) Superimposition of the ECDs of PTCH1* molecules (top view). (C) Superimposition of transmembrane domains of PTCH1* molecules (top view). (D) Superimposition of transmembrane domains of PTCH1* molecules from the side of the membrane.

As seen in the 1:1 complex (22), two endogenous sterol-like ligand densities can be detected in the transmembrane domain of each PTCH1* molecule, one in the SSD and the other close to TM12 (figs. S2 and S5). The palmitate moiety in the cavity between the ECDs, together with TM4 and TM6, provides a hydrophobic environment for the endogenous ligand (Fig. 3D). The palmitate may affect the dynamics of the SSD by inserting into the center of the protein; a similar mechanism was postulated to regulate cholesterol egress from the lysosome in the homologous NPC1 protein (3335). The cavity between the ECDs might also serve as part of a tunnel for the transport of endogenous ligands from ECD-I or the TM domain (figs. S4 and S5). Blocking the tunnel by palmitate could then inactivate PTCH1.

Physiological importance of the Ca2+-mediated interface of SHH-N

The palmitate of SHH-N facilitates an interaction with PTCH1 that is distinct from interactions with other HH-N partners (Fig. 4A). In contrast, the Ca2+-mediated interface of SHH-N, which is required for PTCH1-B binding, has been reported to recruit other HH-N binders such as the co-receptors CDO and BOC (Fig. 4, B to F) (24). Consistent with this, previous experiments showed that the fibronectin type III (FnIII) domain of CDO competes with PTCH1 for HH-N binding in vitro (27). HH co-receptors (e.g., CDO and BOC) are essential for HH-mediated cell proliferation (30, 36): By binding the Ca2+-mediated interface of HH-N, they may increase the local concentration of HH-N, which could promote the PTCH1–HH-N interaction (Fig. 4, C and D).

Fig. 4 Physiological importance of the Ca2+-mediated interface of SHH-N.

(A) Interaction details of PTCH1-A and the globular portion of SHH-N. (B) Interaction details of PTCH1-B and SHH-N. (C) Structure of IHH-N with FnIII domain of CDO (PDB code 3D1M). (D) Structure of SHH-N with FnIII domain of BOC (PDB code 3N1M). (E) Structure of SHH-N with 5E1 Fab (PDB code 3MXW). (F) Structure of SHH-N with Hedgehog-interacting protein (HHIP) (PDB code 3HO5).

The commonly used monoclonal antibody 5E1 binds to the Ca2+-mediated interface of SHH-N, blocking HH signaling (37, 38). Structural analysis indicates that the interface between 5E1 and SHH-N overlaps with that of PTCH1-B and SHH-N (Fig. 4, B and E). However, 5E1 does not compete with the palmitate-dominated interface; we previously detected a PTCH1*–SHH-N–5E1 complex in pull-down assays (22). Hedgehog-interacting protein (HHIP) serves as a negative regulator of the HH pathway and blocks HH signaling (39) by directly binding the Ca2+-mediated interface of SHH-N (28, 40) (Fig. 4F). Therefore, we speculate that the Ca2+-mediated interface, rather than the palmitate-dominated interface, may be primarily regulated by the different HH binders in cells to prevent or promote the formation of the 2:1 complex, affecting signaling.

2:1 PTCH1–SHH-N complex in HH signaling

The structural analysis shows that most of the SHH-N surface is occupied by two PTCH1 molecules and that the N terminus of SHH-N is buried in the cavity formed by the ECDs of PTCH1-A (Fig. 1C). Interestingly, PTCH1-A and PTCH1-B use a similar area to recognize the globular portion of SHH-N, making it impossible for SHH-N to mediate the formation of a PTCH1 oligomer of higher order than a dimer. Our finding suggests that 2:1 PTCH1-HH complex could be the functional organization in HH signaling.

To investigate this further, we performed cell-based assays. SHH-light II cells are transfected with GLI-responsive firefly luciferase reporter and a constitutive renilla-luciferase expression vector for the quantitative measurement of biologically active HH proteins. The palmitoylated SHH-N and its variants were collected from the medium and detected by an antibody to SHH. Using the SHH-N with either I111E/N115K or R153E mutations—located in the two interfaces in our 2:1 complex—substantially reduced signaling in cells (Fig. 5, A and B). The C24S palmitoylation site SHH-N mutant completely lost its activity for HH signaling (Fig. 5B). Together, these studies show that both interfaces of SHH-N and its palmitate moiety are necessary for maximal HH signaling in cells and that the palmitate moiety is absolutely required.

Fig. 5 Functional characterization of HH signal biogenesis.

(A) Location of the key residues for HH signal biogenesis. The residues of SHH-N are underlined. (B) Mutation on either interface or palmitoylation site of SHH-N abolishes HH signaling. HH activity was measured by luciferase assay. Data are means ± SD (n = 3). SHH-N in conditioned medium was detected by Western blotting (inset). (C) HH signaling in Ptch1–/– MEFs transfected with PTCH1 and treated with SHH-N ligand. Luciferase activity was measured as described in the methods section. Data are means ± SD (n = 3). The expression of PTCH1 was detected by Western blotting (inset).

Using Ptch1–/– MEFs with the same luciferase reporting system, we expressed full-length human PTCH1 with either N940R or E947R/D951R mutations on ECD-II to specifically block the interaction on either side of SHH-N (Figs. 2G and 5A). Expression of either of the mutants alone repressed the HH pathway, and in neither case could the repression be released by introducing SHH-N (Fig. 5C). Additionally, our previous work showed that PTCH1 with a mutation of Glu221, a residue that is necessary for the PTCH1–SHH-N interaction in both interfaces (Fig. 5A), can repress the HH pathway but does not respond to HH-N (22), supporting our current observations. Therefore, these assays validated our structure-based proposal that the 2:1 PTCH1–SHH-N complex is essential for maximal HH signaling.

The putative mechanism of palmitoylated HH–mediated signaling

The palmitate of SHH-N inserts into the cavity created by several hydrophobic residues of the ECDs, preventing any access of a putative ligand through the protein. MOLE calculations (41) show that there is a tunnel, ~150 Å in length with a minimum radius of 4 Å, that stretches throughout the PTCH1-B protein but not PTCH1-A, which interacts with the palmitate of SHH-N (Fig. 6A). The path of this tunnel includes the area where the unidentified ligand is located in ECD-I (fig. S4), plus the cavity of the ECDs that accommodates the palmitate and the transmembrane region. There are two gates of this tunnel in the transmembrane domain; one is near TM3 of the SSD and the other is near TM12. Notably, two unidentified sterol-like densities were detected in these two areas in the 2:1 PTCH1SHH-N complex (fig. S5) as well as in the 1:1 PTCH1SHH-N complex (22), implying that these gates are involved in the transport and/or regulation of the sterol-like molecules.

Fig. 6 Putative mechanism of palmitoylated HH–mediated signaling.

(A) Structure of the 2:1 PTCH1*–SHH-N complex with the putative tunnel in blue; the two gates in the transmembrane domain are indicated. (B) Location of the residues for mimicking the palmitate insertion to block the ECDs’ cavity. (C) HH signaling in Ptch1–/– MEFs transfected with PTCH1 and treated with SHH-N ligand. Luciferase activity was measured as described in the methods section. Data are means ± SD (n = 3). The expression of PTCH1 was detected by Western blotting (inset).

To test the function of the palmitate insertion, we introduced two point mutations (L427R and F1017E) on full-length human PTCH1 to block this tunnel by a ~3-Å salt bridge instead of by a palmitate insertion (Fig. 6B). This variant was transfected to Ptch1–/– MEFs, and unlike wild-type PTCH1, it was not able to repress the HH pathway (Fig. 6C). This finding supports our hypothesis that the palmitate could abolish PTCH1 repression of HH signaling by blocking the cavity of the ECDs. It also could explain how abolishing the palmitoylation of HH-N leads to embryonic defects in both Drosophila and mice (10, 12, 13, 4244), which is a direct biological consequence of insufficient HH signal.

Before the HH-N recognizes PTCH1, co-receptors might capture the Ca2+-mediated interface of HH-N by their flexible extracellular domains (24). This would increase the likelihood of forming a PTCH1–HH-N complex, a function reminiscent of the interleukin-2 receptor α-subunit (45). It is also possible for a PTCH1 molecule to engage the Ca2+-mediated interface if the palmitoylated N terminus of HH-N is bound to other binders, such as various heparan sulfate proteoglycans (46, 47). This would allow PTCH1 to serve as a possible co-receptor to promote extracellular transport of HH-N. Because the globular portion of SHH-N triggers the internalization of PTCH1 (16), once HH-N recognizes PTCH1, the co-receptor–PTCH1–HH-N or 2:1 PTCH1–HH-N trimetric complex might be internalized, therefore creating a PTCH1-free environment that further promotes HH signaling (22).

Conclusions

Our structure of a signaling-competent complex consisting of SHH-N and an asymmetric arrangement of two PTCH1* molecules reconciles the disparate findings of how HH-N interacts with Patched. Our cell biological assays revealed that both PTCH1–SHH-N interactions are required for efficient signaling. The CTD and cytosolic loop between TM6 and TM7 have been removed from PTCH1*; therefore, further investigation is required to determine whether the complex results in any coupling between the cytosolic regions of full-length PTCH1.

The structural analysis reveals a tunnel in the PTCH1 molecule. Interestingly, the palmitate of SHH-N directly disrupts this tunnel when bound to PTCH1 (Fig. 6A). Mutations in the ECDs that mimic the palmitate insertion by blocking the ECD cavity abolish the repression of the HH pathway by PTCH1 (Fig. 6, B and C). Knowledge of the specific identity of the endogenous sterol-like molecules in ECD-I and the transmembrane region of PTCH1 would substantially facilitate our understanding of PTCH1 function and SMO activation.

Previous studies also show that HH-N co-receptors are necessary for HH signaling by binding to HH-N and forming a complex with PTCH1 (30, 36, 4850). These HH-N co-receptors might play a role in increasing the local concentration of HH-N, thereby promoting PTCH1–HH-N interaction and HH signaling. The structure of PTCH1–HH-N in complex with its co-receptors will be informative in furthering our understanding of their function in the HH pathway.

Materials and methods

Protein expression and purification

Human PTCH1* and SHH-N proteins were purified according to our previously published protocol (22). To assemble the PTCH1* and SHH-N complex, native SHH-N (purchased from R&D Systems, Cat # 8908-SH/CF) was mixed with purified PTCH1* at a 1:2 molar ratio and purified by size-exclusion chromatography using a Superdex-200 column (GE Healthcare) in buffer A (20 mM Hepes pH 7.5, 150 mM NaCl) with 0.06% Digitonin and 1 mM CaCl2. The peak fractions were collected and concentrated to 5 mg/ml for grid preparation. The mutated and truncated DNA constructs were generated using Gibson Assembly (NEB) or QuikChange Mutagenesis kit (Agilent).

EM sample preparation and imaging

A freshly purified protein sample was added to Quantifoil R1.2/1.3 400 mesh Au holey carbon grids (Quantifoil), blotted using a Vitrobot Mark IV (FEI), and frozen in liquid ethane. The grids were imaged in a 300 keV Titan Krios (FEI) with a Gatan K2 Summit direct electron detector (Gatan). Data were collected at 1 Å per pixel with a dose rate of 8 electrons per physical pixel per second. Images were recorded for 10-s exposures in 50 subframes to give a total dose of 80 electrons per Å2.

Image processing and 3D reconstruction

Dark-subtracted images were first normalized by gain reference that resulted in a pixel size of 1 Å per pixel. Drift correction was performed using the program MotionCorr (51). The contrast transfer function (CTF) was estimated using CTFFIND4 (52). To generate 2:1 PTCH1*–SHH-N templates for automatic picking, around 3000 particles were manually picked and classified by 2D classification in RELION (53). After auto-picking in RELION, the low-quality images and false-positive particles were removed manually. About 275k particles were extracted for subsequent 2D and 3D classification. We used the cryo-EM structure of human PTCH1* (EMD-7795) low-pass filtered to 60 Å as the initial model for 3D classification in RELION. The PTCH1* model of best class after 3D classification was used as the initial model for the final 3D classification in RELION. The best class, containing ~77k particles, provided a 5-Å map after 3D auto-refinement in RELION. Refinement was performed in FREALIGN (54) using this best class as the initial model. The global search was performed once without mask followed by another global search using the mask, which was generated using “relion_mask_create” with 6-Å extensions excluding the micelle. The final map is estimated to be 3.5 Å using the 0.143 cutoff criteria.

Model construction

To obtain better side-chain densities for model building, we sharpened the map using BFACTOR.EXE (author: N. Grigorieff) with a resolution limit of 3.5 Å and a B-factor value of –100 Å2. The structure of human PTCH1*–SHH-N complex was docked to the map. The density of residues 1 to 75 (N-terminal domain), 608 to 618 and 721 to 730 (TM6-TM7 linker), and 1177 to 1188 (C terminus) of PTCH1* is not resolved nor built. Residues 890 to 915 of PTCH1* were built with polyalanine due to limited local resolution.

Model refinement and validation

The model was refined in real space using PHENIX (55) and also in reciprocal space using Refmac with secondary-structure restraints and stereochemical restraints (56, 57). Structure factors were calculated from a half-map (working) using the program SFall (58). Fourier shell correlations (FSCs) were calculated between the two half maps, the model against the working map, the other (free) half map, and full (sum) map (59). Local resolutions were estimated using Blocres (60). MolProbity (61) was used to validate the geometries of the model. Structure figures were generated using PyMOL (www.pymol.org) and Chimera (62).

HH reporter assays

Human SHH-N (24-197) was constructed into pcDNA3.1 vector with the signal sequence of human calreticulin at the N terminus as described (16). Secreted SHH-N was produced in HEK293 cells by transient transfection for 72 hours and collected in DMEM, with 0.5% fetal bovine serum (FBS). SHH Light II cells, a stable cell line expressing firefly luciferase with an 8X-Gli promoter and Renilla luciferase with a constitutive promoter (from B. Chen and J. Kim), were used to measure HH pathway activity. SHH Light II cells were treated with the conditioned medium diluted in fresh DMEM with 0.5% newborn calf serum for 30 hours. To measure the activity of PTCH1 variants in HH signaling, the 8X-Gli-Firefly luciferase reporter transgene, a constitutive Renilla luciferase transgene, and a pcDNA3.1 vector encoding PTCH1 variants were transfected to Ptch1–/– MEFs (from B. Chen and J. Kim) using TransIT reagent (Mirus Bio LLC). After 24 hours, cells were serum-starved in DMEM with 0.5% FBS; 24 hours later, cells were treated with SHH-N conditioned medium for another 24 hours. Firefly and Renilla luciferase activity were measured using the Dual-Luciferase Reporter Assay System (Promega). All reporter assays were performed at least three times. The conditioned medium added was normalized based on Western blotting with anti-SHH antibody. The expression of PTCH1 variants and internal calnexin in MEF cells were detected by Western blotting with anti-PTCH1 antibody (GeneTex, 83771) and anti-calnexin antibody (Novus, NB100-1965), respectively. Each assay was reproduced at least three times.

Supplementary Materials

References and Notes

Acknowledgments: We dedicate this work to Dr. Günter Blobel. We thank M. Ebrahim and J. Sotiris at The Rockefeller University Evelyn Gruss Lipper Cryo-Electron Microscopy Resource Center for data collection; D. Stoddard at the UT Southwestern Cryo-EM Facility (funded in part by CPRIT Core Facility Support Award RP170644) for grid screening; L. Beatty and A. Hassan for technical help; B. Chen and J. Kim for luciferase plasmids, SHH Light II cells, and Ptch1–/– MEFs; and M. Brown, E. Debler, J. Goldstein, and J. Jiang for discussion during manuscript preparation. Funding: Supported by the Endowed Scholars Program in Medical Science of UT Southwestern Medical Center, O’Donnell Junior Faculty Funds, Welch Foundation (I-1957) (to X.L.) and NIH grant P01 HL020948 (to Tissue Culture Core) by The Rockefeller University (to E.C.). X.L. is the Rita C. and William P. Clements Jr. Scholar in Biomedical Research at UT Southwestern Medical Center. Author contributions: X.L. conceived the project and designed the research with X.Q.; X.Q. and P.S. purified the protein; X.Q., E.C., and X.L. performed the functional characterization; all authors carried out cryo-EM work; X.Q. built the initial model and refined the structures; all authors analyzed the data and contributed to manuscript preparation; and X.L. wrote the manuscript. Competing interests: The authors declare no competing financial interests. Data and materials availability: The 3D cryo-EM density maps have been deposited in the Electron Microscopy Data Bank under accession number EMD-8955. Atomic coordinates for the atomic model have been deposited in the Protein Data Bank under accession number 6E1H.
View Abstract

Stay Connected to Science

Navigate This Article