Identification of specific calcitonin-like receptor residues important for calcitonin gene-related peptide high affinity binding

Abstract

Background: Calcitonin gene-related peptide (CGRP) is a vasoactive neuropeptide whose biological activity has potential therapeutic value for many vascular related diseases. CGRP is a 37 amino acid neuropeptide that signals through a G protein-coupled receptor belonging to the secretin receptor family. Previous studies on the calcitonin-like receptor (CLR), which requires co-expression of the receptor-activity-modifying protein-1 (RAMP1) to function as a CGRP receptor, have shown an 18 amino acid N-terminus sequence important for binding CGRP. Moreover, several investigations have recognized the C-terminal amidated phenylalanine (F37) of CGRP as essential for docking to the mature receptor. Therefore, we hypothesize that hydrophobic amino acids within the previously characterized 18 amino acid CLR N-terminus domain are important binding contacts for the C-terminal phenylalaninamide of CGRP.

Results: Two leucine residues within this previously characterized CLR N-terminus domain, when mutated to alanine and expressed on HEK293T cells stably transfected with RAMP1, demonstrated a significantly decreased binding affinity for CGRP compared to wild type receptor. Additional decreases in binding affinity for CGRP were not found when both leucine mutations were expressed in the same CLR construct. Decreased binding characteristic of these leucine mutant receptors was observed for all CGRP ligands tested that contained the necessary amidated phenylalanine at their C-terminus. However, there was no difference in the potency of CGRP to increase cAMP production by these leucine mutant receptors when compared to wild type CLR, consistent with the notion that the neuropeptide C-terminal F37 is important for docking but not activation of the receptor. This observation was conserved when modified CGRP ligands lacking the amidated F37 demonstrated similar potencies to generate cAMP at both wild type and mutant CLRs. Furthermore, these modified CGRP ligands displayed a significant but similar loss of binding for all leucine mutant and wild type CLR because the important receptor contact on the neuropeptide was missing in all experimental situations.

Conclusion: These results are consistent with previous structure-function investigations of the neuropeptide and are the first to propose specific CLR binding contacts for the amidated F37 of CGRP that are important for docking but not activation of the mature CGRP receptor.

Figure 1

Postulated interactions of specific CGRP domains with the CLR-RAMP1 heterodimer. Co-expression of the CLR with RAMP1 forms a mature CGRP receptor on the cell membrane surface. The characteristically long extracellular domains of the CLR-RAMP heterodimer create a high affinity binding pocket important for docking the C-terminal phenylalaninamide of CGRP. This interaction presents the CGRP N-terminus domain, essential for biological activity, to a "classical" agonist binding pocket formed in part by the transmembrane α-helices of the CLR-RAMP1 heterodimer.

Figure 2

Saturation binding characteristics of [¹²⁵I-Tyr]CGRP(8–37). Increasing amounts of [¹²⁵I-Tyr]CGRP(8–37) were used on a crude HEK293T-RAMP1 membrane preparation that had been transiently transfected with wild type CLR, in the absence and presence of 1μM CGRP to determine total and nonspecific binding, respectively. Non-linear regression analysis was used to best fit a specific binding curve to each individual binding experiment. From this best fit curve an equilibrium dissociation constant (K_d) for the peptide radioligand and total number of specific binding sites (B_max) were estimated. The K_d of [¹²⁵I-Tyr]CGRP(8–37) was calculated to be 0.9 ± 0.2 nM with a B_max estimated at 285 ± 206 fmols/mg protein. The data presented represents the mean ± S.E. for n = 3 experiments performed in duplicate.

Figure 3

Competition binding characteristics of CGRP and AcCGRP(19–37) for wild type and mutant CLRs. Increasing amounts of (A) CGRP or (B) AcCGRP(19–37) were used to compete for specific [¹²⁵I-Tyr]CGRP(8–37) binding sites on crude HEK293T-RAMP1 membrane preparations transiently transfected, individually with wild type (■), L24A (●), L34A (○) or L24A,L34A (x) CLR mutations. Non-linear regression analysis was used to best fit a sigmoidal curve from the data points of each individual competition binding experiment. From this best fit curve the concentration of competing peptide need to displace 50% of specific radioligand binding (IC₅₀) was estimated and used to calculate the equilibrium dissociation constant (K_i) of unlabled peptide for each CLR. The calculated K_i values of the competing peptides for the L24A, L34A, or L24A,L34A CLR mutants were significantly different (P < .05) from the wild type receptor and are displayed in Table 1. The data presented represents the mean ± S.E. for n = 3–5 experiments performed in duplicate.

Figure 4

cAMP production of activated wild type and CLR mutants using endogenous or truncated CGRP ligands. Increasing amounts of (A) CGRP, (B) CGRP(1–36) or (C) CGRP(1–19) were used to stimulate confluent HEK293T-RAMP1 cells that had been transiently transfected, individually with wild type (■), L24A (●), L34A (○) or L24A,L34A (x) CLR mutations. (B) Other groups of HEK-RAMP1 cells transiently transfected, individually with wild type (□), L24A (△), L34A (+) or L24A,L34A (◊) CLR mutations were treated with 1μM of the CGRP receptor antagonist, CGRP(8–37), for 30 min prior to the addition of 10μM CGRP(1–36). After 30 min all cells were lysed and the amount of cAMP generated was quantified using a radioimmunoassay according to the manufactures protocol (Amersham). Non-linear regression analysis was used to best fit a sigmoidal curve from the data points of each individual experiment. From this best fit curve a concentration of peptide agonist that produced 50% of the maximal cAMP response (EC₅₀) was estimated for each CLR. The calculated EC₅₀ values of all peptides to increase cAMP in HEK293T-RAMP1 cells transiently transfected with L24A, L34A or L24A,L34A CLR mutations were no different from values calculated for the wild type receptor and are displayed in table 2. The data presented represents the mean ± S.E. for n = 3 experiments performed in duplicate.

Figure 5

Working molecular model of CGRP phenylalaninamide interaction with CLR residues L24 and L34. Energy minimized models for CGRP amino acids 32–37 (NH₂-VGSKAF-CONH₂) and CLR amino acids 23–35 (NH₂-ELEESPEDSIQLG-COOH) were performed as described in the "Methods". Hydrophobic interactions of CLR L24 and L34 with CGRP F37 and RAMP1 (not shown) cooperatively contributes to formation of the high affinity binding pocket essential for docking the C-terminal phenylalaninamide of the neuropeptide with the mature CGRP receptor heterodimer.