Multiple ligand binding sites regulate the Hedgehog signal transducer Smoothened in vertebrates

Abstract

The Hedgehog (Hh) pathway plays a central role in the development of multicellular organisms, guiding cell differentiation, proliferation and survival. While many components of the vertebrate pathway were discovered two decades ago, the mechanism by which the Hh signal is transmitted across the plasma membrane remains mysterious. This fundamental task in signalling is carried out by Smoothened (SMO), a human oncoprotein and validated cancer drug target that is a member of the G-protein coupled receptor protein family. Recent structural and functional studies have advanced our mechanistic understanding of SMO activation, revealing its unique regulation by two separable but allosterically-linked ligand-binding sites. Unexpectedly, these studies have nominated cellular cholesterol as having an instructive role in SMO signalling.

Figure 1

PTC1 regulates SMO via an unknown mechanism. In the absence of SHH, PTC1 inhibits SMO, which allows Sufu and PKA to inhibit the GLI transcription factors (left panel). In the presence of SHH, PTC1 releases its inhibition of SMO, which in turn is able to signal downstream, ultimately resulting in the transcription of target genes by GLI (right panel).

Figure 2

The overall structure of SMO. SMO consists of a large extracellular region, made up of the CRD (green) and LD (orange), and an intracellular domain (ICD, red) in addition to the seven-pass α-helical transmembrane domain (TMD, blue) (left panel). The multi-domain SMO crystal structure revealed a stacked domain arrangement with two physically separable binding sites (right panel). The approximate location of the two binding sites is marked with dashed black ovals in both left and right panels. ECL3 and TMD helix VI are also labeled. Agonist (green) and antagonist (dark grey) small molecule modulators are listed on the right and associated with a particular binding site, if known. This list is not exhaustive.

Figure 3

Multi-domain structures of SMO. (a) Close-up of the cholesterol binding site in the SMO CRD. Residues involved in binding are shown as sticks. Dotted black lines indicate potential hydrogen bonds. Two important residues also discussed in the text are labeled (Asp95 and Trp109). (b) Three multi-domain structures of human SMO, each solved with a different ligand (as indicated beneath each structure), are shown in the same orientation (PDB: 5L7D [17••], 5L7I [17••], 5V57 [18••]). The glycan occluding the CRD-site in the vismodegib complex (middle) is shown in yellow stick representation. Domains in (a) and (b) are coloured as in Figure 2. (c) Conformational changes associated with antagonist binding result in collapse of the CRD binding site, thus precluding cholesterol binding. The three multi-domain structures of SMO were aligned by their TMDs. Each structure is coloured separately with helices shown as solid cylinders and loops omitted for clarity. Red arrows indicate domain movements between structures.

Figure 4

Models for PTC1 function. (a) In the first model, the SMO TMD is constitutively associated with cholesterol but PTC1 prevents the SMO CRD from accessing cholesterol, thereby preventing activation. Upon SHH-binding, PTC1 is inactivated, allowing the CRD to acquire cholesterol and to become activated. (b) In the second model, the SMO CRD contains cholesterol as a necessary co-factor and PTC1 prevents the SMO TMD from accessing cholesterol, thereby preventing activation. Upon SHH-binding, PTC1 is inactivated, allowing the SMO TMD to acquire cholesterol and become activated. (c) In the third model, a variant of the first, PTC1 acts as a cholesterol flippase, shifting cholesterol from the outer to the inner leaflet of the lipid bilayer and thereby preventing the SMO CRD from accessing cholesterol. These models are not mutually exclusive and PTC1 could regulate cholesterol access to both sites. (d) Cholesterol acts as a co-factor for SMO activation, while PTC1 regulates a different lipidic ligand (solid square) which could either function as a SMO antagonist as shown here or as a SMO agonist.