Biochemical Society Transactions

Biochemical Society Focused Meetings

The role of the extracellular loops of the CGRP receptor, a family B GPCR

James Barwell , Michael J. Woolley , Mark Wheatley , Alex C. Conner , David R. Poyner

Abstract

The CGRP (calcitonin gene-related peptide) receptor is a family B GPCR (G-protein-coupled receptor). It consists of a GPCR, CLR (calcitonin receptor-like receptor) and an accessory protein, RAMP1 (receptor activity-modifying protein 1). RAMP1 is needed for CGRP binding and also cell-surface expression of CLR. There have been few systematic studies of the ECLs (extracellular loops) of family B GPCRs. However, they are likely to be especially important for the interaction of the N-termini of the peptide agonists that are the natural agonists for these receptors. We have carried out alanine scans on all three ECLs of CLR, as well as their associated juxtamembrane regions. Residues within all three loops influence CGRP binding and receptor activation. Mutation of Ala203 and Ala206 on ECL1 to leucine increased the affinity of CGRP. Residues at the top of TM (transmembrane) helices 2 and 3 influenced CGRP binding and receptor activation. L351A and E357A in TM6/ECL3 reduced receptor expression and may be needed for CLR association with RAMP1. ECL2 seems especially important for CLR function; of the 16 residues so far examined in this loop, eight residues reduce the potency of CGRP at stimulating cAMP production when mutated to alanine.

  • alanine scan
  • calcitonin
  • mutagenesis
  • receptor activity-modifying protein
  • secretin-like G-protein-coupled receptor
  • structure–activity relationship

Family B GPCRs (G-protein-coupled receptors)

GPCRs form a number of families. The best characterized of these are the rhodopsin-like or family A GPCRs. The secretin-like or family B GPCRs form receptors for several 30–50 amino acid peptides such as secretin, glucagon and corticotropin-releasing factor. They are characterized by large N-termini of approximately 100–150 residues. They have virtually none of the conserved motifs found in family A receptors, instead having their own set of distinctive residues [1]. Peptide agonist binding to these receptors follows a two-step model. The C-terminus of the ligand interacts with the N-terminus of the receptor. However, by itself, this is generally not enough to cause receptor activation. This is produced by the interaction of the N-terminus of the ligand, interacting with the seven-helical bundle of the receptor, with contacts to either the TM (transmembrane) helices or the ECLs (extracellular loops) [2,3]. There are no crystal structures for the TM domains of family B GPCRs, although a number of structures are available for the isolated N-termini [4]. Consequently, the details of binding of the N-termini of peptide agonists to their cognate receptors for family B GPCRs remains obscure, although the problem has been addressed by photoaffinity cross-linking, disulfide trapping and mutagenesis [58]. It has been suggested that when the endogenous peptide agonists bind to their receptors, they do so as a helix and this is important for receptor activation [9]; it is noteworthy that most of the peptide agonists are able to form an N-cap structure, which would stabilize such a helix [10].

CGRP (calcitonin gene-related peptide) and its receptor

CGRP is a widespread 37-amino-acid peptide that is an extremely potent vasodilator. It is part of the calcitonin family of peptides that also includes adrenomedullin, adrenomedullin2/intermedin, amylin and calcitonin. The peptide is the subject of considerable attention at the moment as a CGRP antagonist is undergoing clinical trials for the treatment of migraine [11]. CGRP acts via a family B GPCR, the CLR (calcitonin receptor-like receptor). This receptor has a number of unusual features. Although it has the structure of a typical family B GPCR, by itself it will not bind any known endogenous ligand. In order for this to happen it must interact with a second protein from the RAMP (receptor activity-modifying protein) family. These have a single TM helix and an extracellular domain of approximately 100 amino acids; RAMP1 associates with CLR to give the CGRP receptor, whereas RAMP2 and RAMP3 give adrenomedullin receptors [12,13]. Crystal structures are available showing how the extracellular domains of CLR and RAMP1 interact [14]. Two of the structures include small molecule antagonists but there is no structure to show how CGRP or any of its peptide derivatives bind, although by analogy with other family B GPCRs, it is possible to speculate how this might happen [4,13]. It is clear that the interaction of the extracellular domains is with the C-terminal portion of CGRP, meaning that the N-terminus of CGRP must interact with the TM helices and ECLs.

There is no structural information on how the N-terminus of CGRP might interact with its receptor. The calcitonin family of peptides are distinct from other ligands to family B GPCRs in that they have a disulfide-bonded ring at their N-termini; however, it has been suggested that this performs the same function as the N-cap in other family B receptors [10]. Removal of the first seven amino acids to delete the ring creates an antagonist, at least on cAMP production, confirming the importance of the extreme N-terminus for agonist activity [15]. An alignment of the N-terminal portion of the calcitonin family of peptides shows that in addition to the two cysteine residues that take part in the disulfide bond, the threonine at position 6 of CGRP is also absolutely conserved in all members of the family (Figure 1). There is no published information on the role of Thr6 in CGRP, but when its equivalent in adrenomedullin, Thr20, was substituted by alanine, there was a significant decrease in the potency of the analogue at reducing blood pressure [16]. This is consistent with Thr6 playing an important role in receptor activation. Given its location, it is highly likely that it interacts with a residue in the ECLs of CLR or the adjacent TM regions.

Figure 1 N-terminal sequences of CGRP and members of the calcitonin family of peptides

Sequences are 1–16 for CGRP, CRSP (calcitonin receptor-stimulating peptide) and amylin, 1–15 for calcitonin, 14–29 for adrenomedullin and 16–31 for adrenomedullin2/intermedin. Shading shows identity.

ECL1 and ECL3 in CLR

We have carried out alanine scans of ECL1 and ECL3 of CLR [17] (Figure 2). Owing to difficulties in ascertaining the precise start and finish of the ECLs in family B GPCRs, the alanine scan was extended into what is likely to be the adjacent TM helices. The most significant results are shown in Figure 3. In ECL1, Leu195, Val198 and Ala199 all reduced CGRP potency at cAMP production upon mutation; in addition, L195A and A199L also reduced the potency of CGRP. The hydrophobicity of these residues suggest that they are most likely to be at the top of TM helix 2 rather than in ECL1 itself. They cluster together to form a patch that is clearly important for CGRP binding and also subsequent signal transduction. It is difficult to tell if they form an actual contact with CGRP or whether they play a secondary role in maintaining the architecture of this part of the receptor in such a way that it can bind the peptide. Within what is more likely to be ECL1, A203L and A206L both increased CGRP potency although the precise mechanism of these effects is unclear. However, it is obvious that the upper portions of ECL1 are in a position to influence the binding of CGRP. This loop is also important in the calcitonin receptor, where an insertion of 37 amino acids into ECL1 of the rat calcitonin receptor reduces agonist binding [18]. A third set of residues that influenced CGRP potency were H219A, L220A and L222A. These are likely to be in the middle third of TM helix 3. H219A reduced cell-surface expression of the receptor as well as CGRP potency, possibly suggesting increased internalization and perhaps a better coupling to β-arrestin. The two leucine residues both increased CGRP potency on cAMP production and L222A increased the affinity of the receptor for CGRP. Given the role of TM3 in family A GPCRs, it is tempting to suggest that these residues play a role in an agonist-mediated conformational switch that can be activated by CGRP, leading to enhanced receptor activation. As with the hydrophobic cluster in TM helix 2, it remains difficult to pinpoint the exact mechanistic role of these important residues.

Figure 2 Schematic snake plot of CLR showing the location of mutated residues

ECL1, green; yellow, ECL2; grey, ECL3. The exact boundaries of the loops are difficult to identify and this illustration is diagrammatic only.

Figure 3 Model of CLR showing residues in ECL1 and ECL3 that alter function

The model was produced as described previously [17]. Left-hand panel: side view; right-hand panel: view of the extracellular surface of the receptor

Within ECL3, fewer residues had effects on CGRP potency at cAMP production. The most notable effect on this parameter was seen with the mutant I360A, which caused approximately 10-fold change in EC50. This seems to be the main residue in ECL3 that influences CGRP potency; our modelling suggested that it could be in close vicinity to ECL2 and may stabilize the structure of that loop (see below). In the calcitonin receptor, Leu368, the equivalent of Val364 in CLR, is a photoaffinity contact for residue 8 of salmon calcitonin [19]. This suggests that the top of ECL3 may be in proximity to CGRP, perhaps just C-terminal to the start of the disulfide-bonded ring [although it should be pointed out that photoaffinity contacts can operate over a distance of 10 Å (1 nm), making it very difficult to generate high-resolution data from such experiments]. A number of mutants reduced receptor expression, especially L351A and E357A. Modelling suggested that these lay on the outer surface of ECL3 and TM6. As TM6 has been implicated in RAMP binding [20], they may form a contact surface for RAMP1.

ECL2

We have carried out a preliminary characterization of ECL2 by means of an alanine scan (Figure 2). ECL2 is likely to be the longest loop in CLR, extending to approximately 20 amino acids. ECL2 has one of the very few conserved structural motifs found in both family A and family B GPCRs; a disulfide bond between the top of TM3 to the middle of the loop. Our data indicate that eight residues in the loop (Arg274, Tyr277, Tyr278, Asp280, Cys282, Trp283, Ser285 and Thr288) reduce the potency of CGRP at stimulating cAMP production when mutated to alanine. This is a much higher percentage than found in ECL1 or ECL3, suggesting that ECL2 may be of particular importance for the binding of CGRP. Of the residues, Arg274, Tyr277, Cys282 and Trp283 are conserved in all family B GPCRs. The CW (Cys282Trp283) motif is of particular interest. An alanine scan of the CRF1R (corticotropin-releasing factor-1 receptor) has shown that they, with an adjacent phenylalanine, are essential for the binding of the agonist sauvagine [8]; furthermore Lys262 of the CRF1R (the equivalent of Ser286 in CLR) makes a photoaffinity contact with a derivative of sauvagine [21]. In the glucagon-like peptide 1 receptor, the equivalent of Trp283 is also a photoaffinity contact for a glucagon-like peptide 1 derivative [5]. Trp283 and its neighbours may produce a hydrophobic patch that is utilized by the peptide ligands for a number of family B GPCRs.

An overview of CGRP binding?

A detailed model of how CGRP interacts with the ECLs of CLR cannot yet be produced. It is not possible to produce a reliable model of the ECL domain for any family B GPCR and there is no information on any specific CLR–CGRP contact. It is, however, possible to produce some general observations. The C-terminus of CGRP interacts with the extracellular domains of CLR and RAMP1, almost certainly fitting into the interface between the two molecules [4,13,14]. It is not known for any family B GPCR how the extracellular domain is orientated with respect to the ECLs and TM bundle, although models incorporating photoaffinity distance constraints based on the glucagon-like peptide 1 receptor and secretin receptor are consistent with the extracellular domain sitting immediately above the TM bundle with its long-axis parallel to that of the TM domain [5,22]. By analogy with other family B GPCRs, this will present the N-terminus of CGRP as a rod, also roughly parallel to the main long-axis of the receptor [9,10]. In this case, the ECLs may be imagined as forming a funnel, to hold the N-terminal activator region of CGRP. It is likely that they will make contacts well beyond the disulfide-bonded N-terminal ring. Residues 8–18 of CGRP show a strong tendency to form an α-helix; within this region there are a number of residues that are important for binding [23,24] and they may interact with the upper reaches of the ECLs, as seen in photoaffinity studies of the glucagon-like peptide 1, secretin and parathyroid hormone receptors [5,22,25]. It is not clear how exactly the extreme N-terminus of any family B peptide agonist is orientated with respect to the TM bundle or whether there is even a common mode of binding [5,7,22]. The mutagenesis data implicate residues in the upper third of the TM bundle in CGRP binding and these are at a similar depth to those identified from cross-linking and mutagenesis experiments as important for the binding of glucagon, glucagon-like peptide 1, secretin and parathyroid hormone [5,7,22,26,27]. Thus, regardless of its orientation, the extreme N-terminus of CGRP may be quite deeply buried within the TM bundle, allowing for a direct effect on TM helix orientation and consequent coupling to G-proteins,

Although CGRP can activate multiple second messengers, [28], virtually all work on CGRP antagonists has concentrated on measurements of cAMP production. It is not known if they antagonize other pathways; interestingly, an N-terminally truncated PTH (parathyroid hormone) derivative that would be expected to be an antagonist promotes β-arrestin coupling and subsequent signalling [29]. In family A GPCRs, a distinction has been made between ligands that occupy the ‘major pocket’ on the surface of the TM bundle (formed by helices 4, 5, 6 and 7) and the minor pocket (helices 1, 2, 3 and 7); it has been suggested that occupation of the minor pocket may bias the response to activation of β-arrestin [30]. It is not known if these pockets exist on family B GPCRs, but it is interesting to speculate that interactions of ligands such as CGRP with the upper reaches of the ECLs may trigger conformational changes beyond those normally associated with activation of Gs [3]. If this is correct, it could open the way to the development of biased agonists.

Conclusions

Although molecular details remain elusive, there is no doubt that the ECLs of CLR play a pivotal role in binding CGRP and subsequent receptor activation. It will be of particular interest to compare the binding of CGRP with other family B GPCRs, to see how far receptor activation mechanisms are conserved in family B GPCRs and also to see whether differential interaction of the ECLs with CGRP allows the design of biased agonists.

Footnotes

  • Dynamics Within and Between Proteins: A Biochemical Society held at the University of Essex, 31 August–2 September 2011. Organized and Edited by Christopher Cooper, Neil Kad, Jody Mason, Phil Reeves and Jon Worrall (University of Essex, U.K.).

Abbreviations: CGRP, calcitonin gene-related peptide; CLR, calcitonin receptor-like receptor; CRF1R, corticotropin-releasing factor-1 receptor; ECL, extracellular loop; GPCR, G-protein-coupled receptor; RAMP, receptor activity-modifying protein; TM, transmembrane

References

View Abstract