Structural insights into endothelin receptor signalling

Abstract

Endothelins and their receptors, type A (ET_A) and type B (ET_B), modulate vital cellular processes, including growth, survival, invasion and angiogenesis, through multiple G proteins. This review highlights the structural determinations of these receptors by X-ray crystallography and cryo-electron microscopy, and their activation mechanisms by endothelins. Explorations of the conformational changes upon receptor activation have provided insights into the unique G-protein coupling feature of the endothelin receptors. The review further delves into the binding modes of the clinical antagonist and the inverse agonists. These findings significantly contribute to understanding the mechanism of G-protein activation and have potential implications for drug development, particularly in the context of vasodilatory antagonists and agonists targeting the endothelin receptors.

Graphical Abstract

Open in new tabDownload slide

Endothelins are 21 amino acid peptides primarily derived from the endothelium (1,) and play a crucial role in maintaining vascular homeostasis. Among these peptides, the potent vasoconstrictor ET-1 was the first discovered (2,3,). The transmission of signals by ET-1 is mediated by two endothelin receptor (ETR) subtypes, ET_A and ET_B, which are class A G-protein-coupled receptors (GPCRs) (4,5,). The ET_A and ET_B receptors are promiscuous GPCRs capable of activating multiple types of G proteins (6,7,) and regulate essential cellular processes such as growth, survival, invasion and angiogenesis (8,9,). In the vascular system, both receptors are coupled to G_q and have opposing functions. ET_A primarily mediates vasoconstriction, which is prolonged upon irreversible binding to ET-1. Conversely, ET_B induces vasorelaxation through the nitric oxide-mediated pathway and functions as a clearance receptor by removing circulating ET-1 via lysosomal degradation (10,). In astrocytes, ET_B stimulates a G_i signalling pathway and thereby inhibits intercellular communication mediated by gap junctions (11,). Moreover, activation of the Rho signalling pathway in astrocytes prompts cytoskeletal reorganization and promotes the proliferation of adhesion-dependent cells (12,). This process subsequently leads to the induction of reactive astrocytes, ultimately facilitating neuroprotection (13).

Drug development for the ETRs is primarily focused on vasodilatory antagonists (8,9,). Bosentan, the first non-peptide antagonist for ET_A and ET_B, is clinically used for the treatment of pulmonary arterial hypertension. ET_A-selective antagonists are also used as therapeutic agents with fewer side effects. Endothelin-1 primarily acts through the ET_A receptor and is implicated in the neoplastic growth of multiple tumour types. Thus, ETR antagonists such as atrasentan and zibotentan have exhibited potential anticancer activities in preclinical studies (9,14,). In addition, the development of ET_B agonists is underway, as they provide therapeutic benefits such as vasodilation and neuroprotection (3,13,). IRL1620, the smallest ET_B-selective agonist and a truncated analogue of ET-1, selectively and transiently increases tumour blood flow (15,), making it a potential adjuvant cancer therapy for enhancing the delivery of anti-tumour drugs and a treatment for acute ischemic stroke (16,). However, IRL1620 is a linear peptide with exposed N- and C-termini, which pose challenges in terms of pharmacokinetics and drug delivery. Our previous crystallographic studies have elucidated the structure–activity relationships of peptide agonists and small-molecule clinical antagonists of the ET_B receptor (17–21,). Based on this structural information, researchers have developed new ETR antagonists. Moreover, cryo-electron microscopy structures of the ET_B and ET_A receptors in complex with G proteins have recently been reported (22,23). This article focuses on the mechanisms by which endothelin transmits signals via ETRs, and antagonists competitively block these signals.

Structural determination

Structural studies of the ETRs commenced with the X-ray crystallographic analyses of the ET_B receptor. By innovating methods to crystallize difficult targets (24,), a set of crystal structures of GPCRs was reported in the 2010s. Regarding those, Okuta et al. (25,26,) established a thermostabilized ET_B receptor containing five mutations (ET_B-Y5) that could be functionally expressed in a cell-free system (26,). Using the modified constructs, we determined the crystal structures of the ET_B receptor in complex with various compounds (17–21,) (Fig. 1 and Table 1). Recently, our group and others have determined the structures of wild-type ETRs in complex with G proteins (G_i or G_q) by cryo-electron microscopy (22,23,). For the structural determination of the ET-1–ET_B–G_i complex, we developed a “fusion G-system” that facilitates the efficient expression and purification of GPCR–G-protein complexes. So far, the structures of 12 different ETRs have been reported (Fig. 1 and Table 1), revealing a series of conformational changes from the inactive to fully activated state (17–23).

Endothelin binding modes

The ET_B receptor adopts the typical GPCR architecture, comprising seven transmembrane (TM) helices and intracellular helix 8 (H8). The extracellular loop (ECL)2 forms long anti-parallel β-strands with a short hairpin, which is a common feature of the peptide receptors (Fig. 2A). In addition to the TM3-ECL2 disulphide bond that is highly conserved among the class A GPCRs, ET_B has an additional disulphide bond between C90^N-ter and C358^7.25 (superscripts indicate Ballesteros–Weinstein numbering (27,)) (Fig. 2B). This disulphide bond links the receptor N-terminus and the extracellular end of TM7. The extracellular portion of the receptor is extensively involved in ET-1 binding, with the N-terminal tail, the three ECLs (ECL1–ECL3) and the six TM helices (TM2–TM7) all contributing to the formation of the orthosteric pocket, which is entirely occupied by ET-1. Furthermore, the N-terminal tail and the ECL2 β-sheet form a lid-like structure that covers the orthosteric pocket (Fig. 2B), resulting in a closed conformation that accounts for the virtually irreversible binding of ET-1 (28–30).

ET-1 adopts a bicyclic structure with two intrachain disulphide bond pairs (C1–C15 and C3–C11) and can be divided into three regions (Fig. 2C, D). In the C-terminal region, residues D18 to W21 adopt extended conformations and penetrate deeply into the receptor core, with the C-terminal W21 side chain directed towards the bottom of the pocket (Fig. 2E). Notably, the C-terminal carboxylate and the D18 side chain form an electrostatic interaction network that involves the charged residues of ET_B (K182^3.33, K273^5.38, R343^6.55 and D368^7.35). The C-terminal W21 side chain interacts not only with K182^3.33 via a π–cation interaction but also directly with W336^6.48 in the C^6.47W^6.48xP^6.50 motif, which is the essential motif for the signalling function of class A GPCRs. In the α-helical region, the central residues (D8 to L17) primarily consist of hydrophobic residues and are sandwiched between ECL2 and TM6–7 (Fig. 2F), where they form extensive van der Waals interactions with the receptor. Both ends of the α-helical region form hydrogen bonds with the bulky residues (K161^2.64, Y247^ECL2, Y350^ECL3, R357^7.24 and Y369^7.36). In contrast, the N-terminal region of ET-1, with the residues C1 to M7, is exposed to the solvent and interacts poorly with the receptor compared with the other regions (Fig. 2G).

The α-helical and C-terminal domains are remarkably conserved among endothelins and related peptides (Fig. 2C), creating the principal elements of receptor interactions. Conversely, the amino acid sequences of the N-terminal region are diverse and exhibit sparse receptor interactions. In the structures of endothelin B receptors bound to endothelin-1, endothelin-3 and sarafotoxin S6b, the agonist peptides exhibit a high degree of superimposition. Moreover, the structure of ET_A superimposes well on that of ET_B (Fig. 2H), suggesting that the binding modes of endothelins are highly conserved in both subtypes. The ET_B-selective agonist IRL1620 lacks the N-terminal region but retains the α-helical and C-terminal regions. Its negative charges, originating from Nα-succinylation and E9, are better suited to the positively charged extracellular region of ET_B compared with ET_A (Fig. 2I), resulting in preferential ET_B binding (23).

Conformational changes upon receptor activation

In the absence of ligands, the TM helices in the ET_B receptor are loosely packed against each other, resulting in a large cavity that extends to the receptor core (Fig. 3A). This broad entrance may facilitate the access of large peptide agonists to the orthosteric pocket. Upon ET-1 binding, TM2, TM6 and TM7 move inwards by 2.6, 4.1 and 4.9 Å (Fig. 3A), respectively, whereas TM1 moves outwards by about 4.4 Å, despite having no direct interaction with ET-1. These movements cause the orthosteric pocket to contract and adopt a compact conformation, enabling tight interactions with ET-1. At the intermembrane region, D147^2.50 moves downward by about 3 Å and forms a hydrogen bond with N382^7.49 (Fig. 3B), squeezing the intermembrane region. Essentially, W21 of ET-1 directly induces the downward rotation of the W336^6.48 side chain by 2.5 Å (Fig. 3B), whose nitrogen atom forms a hydrogen bond with N387^7.45. The downward motions of W336^6.48 and N387^7.45 push and induce the outward rotation of F332^6.44 in the P^5.50I/V^3.40F^6.44 motif (Fig. 3C, D), resulting in the 6.8-Å outward displacement of the intracellular portion of TM6 (Fig. 4A, B). The degree of opening is smaller than that observed in other G_i-coupled receptors (μOR: 10 Å) (31), reflecting the unique structural features of the G-protein coupling to the ET_B receptor.

In most class A GPCRs, the highly conserved D^3.49R^3.50Y^3.51 and N^7.49P^7.50xxY^7.53 motifs on the intracellular side play an essential role in G-protein coupling (32). Upon ET_B activation, the ionic lock between D198^3.49 and R199^3.50 is broken, and R199^3.50 becomes oriented towards the intracellular cavity (Fig. 4C). However, in the ET_B receptor, the N^7.49P^7.50xxY^7.53 motif is altered to N^7.49P^7.50xxL^7.53Y^7.54, where Y^7.53 is replaced by L386^7.53. The activation of class A GPCRs induces the inward movement of Y^7.53 to form a water-mediated hydrogen bond with the highly conserved Y^5.58. In contrast, L386^7.53 is a hydrophobic residue and cannot form a polar interaction, leading to the downward, instead of inward, displacement of the intracellular portion of TM7 by 2.8 Å in the ET_B–G_i complex (Fig. 4C). Y^7.54 is directed towards the membrane plane, and its rotamer does not change upon receptor activation. The substitution of Y^7.53 with L386^7.53 in ET_B affects the movement of TM7 upon receptor activation, thereby distinguishing it from other class A GPCRs.

The atypical downward motion of TM7 results in the distinctive G_i coupling mode of the ET_B receptor. L386^7.53 forms a hydrophobic contact with G352^G.H5.24 (superscript indicates the common Gα numbering system) (33,), and TM7 and H8 extensively interact with the α5-helix, which is absent in other GPCR–G_i complexes. These interactions allow the α5-helix of ET_B–G_i to assume the shallowest position relative to the receptor among the G_s-, G_i- and G_q-coupled class A GPCR structures (31,34–38) (Fig. 4D). This characteristic is commonly observed in the other G_i- or G_q-complexed structures of ETRs (Fig. 4E). Thus, the amino acid sequence of the α5-helix coupled to the ETRs would be less restricted, accounting for their G-protein promiscuity. This mechanism of G-protein promiscuity may be exclusive to ETRs, as other promiscuous GPCRs retain the N^7.49P^7.50xxY^7.53 motif. Accordingly, this information is critical for understanding the mechanism of G-protein activation by GPCRs.

Antagonist binding modes

The crystal structures of ET_B bound to the antagonist bosentan, and the inverse agonists K-8794 and IRL2500 have been determined (18,21,) (Fig. 5A–D). The antagonist-bound structures superimpose well on the apo form (17,) (Fig. 5A), and these compounds are located within the positively charged TM cleft of the receptor. Bosentan and its analogues, as well as K-8794, have many aromatic moieties consisting of a central pyrimidine template with four substituents, including 2-pyrimidyl, 4-sulfonamide, 5-phenol and 6-hydroxyl groups. The sulfonamide moiety of bosentan is specifically recognized by positively charged residues (Fig. 5B), whereas the other moieties occupy the space within the TM binding pocket surrounded by TM2–7, facilitating interactions with the receptor. Previous studies have shown that the sulfonamide substitution in bosentan analogues abolishes binding to ETRs (39,), indicating it is a pharmacophore. Bosentan superimposes well on K-8794 (Fig. 5A), and their binding modes are essentially the same (Fig. 5C). The larger substituent on the 6-position of the central pyrimidine in K-8794 leads to stronger interactions with the receptor, accounting for its higher affinity than bosentan (21,). Moreover, the K-8794-bound structure revealed the detailed water-mediated hydrogen-bonding network around D147^2.50, the putative Na⁺ binding site in class A GPCRs (40). Our structural and functional analyses suggested that, instead of the Na⁺ ion, the water molecule networks connect TMs 2, 3, 6 and 7 and stabilize the inactive conformation of the receptor.

Unexpectedly, bosentan adopts a similar binding mode to the agonists, even though it was developed without mimicking endogenous agonist peptides (21). The ET_B receptor similarly interacts with both bosentan and the C-terminal tripeptide of ET-1 (Fig. 5A), suggesting the importance of ionic and hydrophobic interactions within the TM core for antagonist and agonist binding. However, in contrast to the ET-1-induced 4-Å inward movements of both TM6 and TM7 (Fig. 5E), bosentan induces only a minor inward movement of TM6 and no movement of TM7. The 2-methoxyphenoxy group of bosentan sterically prevents the inward motion of W336^6.48, which is critical for receptor activation, and thus bosentan functions as an antagonist.

IRL2500 comprises a tryptophan, a biphenyl group and a 3,5-dimethylbenzoyl group. The biphenyl group forms a peptide bond with the tryptophan and a peptoid bond with the dimethylbenzoyl group. The carboxylate group of the tryptophan moiety in IRL2500 forms a salt bridge with R343^6.55, whereas the other aromatic moieties occupy the space within the TM binding pocket (Fig. 5D). The binding mode of IRL2500 is similar to that of bosentan. Notably, the biphenyl group penetrates deeply into the receptor core proximal to D147^2.50. Thus, the dimethylbenzoyl and biphenyl groups of IRL2500 sandwich the W336^6.48 side chain, securely preventing its inward rotation (Fig. 5E). These observations suggest that IRL2500 rigidly prevents the transition to the active state, compared with bosentan, and functions as an inverse agonist that reduces basal activity, as shown by pharmacological studies using the constitutive ET_B mutant (18).

The residues constituting the antagonist binding site, including the positively charged residues that coordinate the sulfonamide in the current structures, are highly conserved between the ET_A and ET_B receptors, suggesting that bosentan and its analogues have common binding modes for both ETRs. Furthermore, other ETR antagonists that are unlike bosentan, including the ET_A-selective antagonists ambrisentan (41,) and atrasentan (42), also possess a negatively charged group, such as a sulfonamide or a carboxylate, surrounded by two or three hydrophobic (aromatic or acyl) groups. The negatively charged group could occupy the centre of the TM pocket and form ionic interactions with the positively charged residues of the ETRs, whereas the other hydrophobic groups might fit within the pocket and facilitate these interactions, as observed in the current ET_B structures.

Future Perspectives

Over the past decade, significant advances in the structural studies of ETRs have elucidated their fundamental compositions, mechanisms of activation and diverse drug binding modes (Fig. 6). This structural information has led to the development of several ETR drugs (43,44,). Yet, two major challenges remain in the structure-based drug discovery targeting ETRs. The first is the rational design of subtype-selective ETR antagonists. The ET_A selectivity is purportedly crucial for effectiveness against pulmonary arterial hypertension and cancer (9,14,45–48,). However, the mechanism underlying ET_A selectivity remains enigmatic because of the high conservation of the residues constituting the binding pockets in both ET_B and ET_A(21,). In 2023, the first agonist-bound ET_A structure was reported (23,). Thus, the inactive ET_A structure will be determined in the future and significantly clarify the ET_A selectivities of druggable antagonists. The second challenge is the creation of small-molecule agonists for ETRs. Because ET_A activation leads to potent vasoconstriction, ET_B activation holds therapeutic potential owing to its vasodilatory effects (9,15,16,). The linear peptide IRL1620 is an ET_B-selective agonist, but small-molecule agonists have not been discovered. As evidenced by the structures, ET_B activation by endothelin-related peptides entails substantial structural changes, including the 4-Å inward displacements of TMs 6 and 7 on the extracellular side (17). It would be difficult for small-molecule drugs to induce these specific structural changes precisely. Our structural investigations propose that the binding and movement of W336^6.48 and R343^6.55 with the C-terminal region of ET-1 are of utmost importance for receptor activation. Consequently, drugs that emulate this C-terminal region may offer potential as small-molecule ETR agonists.

Funding

This work was supported by grants from the Platform for Drug Discovery, Informatics and Structural Life Science by the Ministry of Education, Culture, Sports, Science and Technology (MEXT), and JSPS KAKENHI grants 21H05037 (O.N.), 22K19371 and 22H02751 (W.S.); ONO Medical Research Foundation (W.S.); The Kao Foundation for Arts and Sciences (W.S.); The Takeda Science Foundation (W.S.); The Uehara Memorial Foundation (W.S.); and the Platform Project for Supporting Drug Discovery and Life Science Research (Basis for Supporting Innovative Drug Discovery and Life Science Research (BINDS)) from AMED, under grant number JP23ama121012.

Conflict of interest

None declared.

References

Author notes