Rational development of a high-affinity secretin receptor antagonist

Abstract

The secretin receptor is a prototypic class B GPCR with substantial and broad pharmacologic importance. The aim of this project was to develop a high affinity selective antagonist as a new and important pharmacologic tool and to aid stabilization of this receptor in an inactive conformation for ultimate structural characterization. Amino-terminal truncation of the natural 27-residue ligand reduced biological activity, but also markedly reduced binding affinity. This was rationally and experimentally overcome with lactam stabilization of helical structure and with replacement of residues with natural and unnatural amino acids. A key new step in this effort was the replacement of peptide residue Leu²² with L-cyclohexylalanine (Cha) to enhance potential hydrophobic interactions with receptor residues Leu³¹, Val³⁴, and Phe⁹² that were predicted from molecular modeling. Alanine-replacement mutagenesis of these residues markedly affected ligand binding and biological activity. The optimal antagonist ligand, (Y¹⁰,c[E¹⁶,K²⁰],I¹⁷,Cha²²,R²⁵)sec(6-27), exhibited high binding affinity (4 nM), similar to natural secretin, and exhibited no demonstrable biological activity to stimulate cAMP accumulation, intracellular calcium mobilization, or β-arrestin-2 translocation. It acts as an orthosteric competitive antagonist, predicted to bind within the peptide-binding groove in the receptor extracellular domain. The analogous peptide that was one residue longer, retaining Thr⁵, exhibited partial agonist activity, while further truncation of even a single residue (Phe⁶) reduced binding affinity. This sec(6-27)-based peptide will be an important new tool for pharmacological and structural studies.

Keywords: Antagonist; G protein-coupled receptor; Secretin; Secretin receptor.

Fig. 1.. Primary sequences of human secretin analogues used in the current study.

Residues present in natural secretin are illustrated in grey, while substituted non-natural residues are in black. Residue numbers correspond to numbering of natural sec(1-27), and all peptides were carboxyl-terminally amidated. X represents L-cyclohexylalanine (Cha). Brackets indicate the location of a lactam bridge linking the side chains of glutamic acid and lysine. Shown in black with white lettering is a pseudopeptide bond (CH₂-NH) between Gly⁴ and Thr⁵ in peptide 2. Peptides 4 to 12 all incorporated a Tyr (Y) in position 10 to replace Leu (L) as a site for radioiodination that has previously been demonstrated to be well tolerated [19,20].

Fig. 2.. Enhancing binding affinity of sec(5-27) analogues as potential SecR antagonists.

A. Concentration-dependent inhibition of secretin-like radioligand (¹²⁵I-(Y¹⁰)sec(1-27)) binding to CHO cells expressing 32,400 human secretin receptors per cell (CHO-SecR-low) by natural sec(1-27) (1) and analogues that were previously reported to be receptor antagonists; [ψ^4,5]sec(1-27) (2) and amino-terminally truncated sec(5-27) (3). B. cAMP accumulation following stimulation with sec(1-27) (1), [ψ^4,5] sec(1-27) (2), sec(5-27) (3), or the best sec(5-27) analogue developed in the current structure-activity series, (Y¹⁰,c[E¹⁶,K²⁰],I¹⁷,Cha²²,R²⁵)sec(5-27) (9). C. Effects of Cha substitution of Leu in the lactam-containing secretin analogue, (Y¹⁰,c[E¹⁶,K²⁰])sec(5-27) (4) on SecR binding affinity. Incorporation of Cha into position 22 (6) increased affinity, while substitution of positions 19 (5), 23 (7), or 26 (8) had no significant effect (see Table 1 for quantification). D. Incorporation of Ile and Arg into positions 17 and 25, respectively, of (Y¹⁰,c[E¹⁶,K²⁰],Cha²²)sec(5-27) (6) produced (Y¹⁰,c[E¹⁶,K²⁰],I¹⁷,Cha²²,R²⁵)sec(5-27) (9) that had a binding affinity equivalent to that of natural sec(1-27) (1). All data are mean ± SEM of at least three independent experiments performed in duplicate. * indicates data points for the putative existing antagonists that are significantly different from the basal levels.

Fig. 3.. Molecular model of secretin peptide occupation of SecR.

SecR transmembrane domain is shown in light blue, while the extracellular domain (ECD) is shown in deep blue. A. Shown is the proposed docking of natural sec(1-27) (1), with amino acids 1 to 4 shown in light green and 5 to 27 shown in magenta. Asp³* of sec(1-27) is shown in sticks, while interacting Arg¹⁸⁸ of SecR (numbering of receptor residues based on full sequence, including signal peptide) is shown in spheres. B. Shown is the proposed docking of the optimized sec(5-27) analogue, (Y¹⁰,c[E¹⁶,K²⁰],I¹⁷,Cha²²,R²⁵)sec(5-27) (9). This peptide is illustrated in magenta, with mutated Tyr, Ile, and Cha residues are indicated with spheres in grey, orange and yellow, respectively. c[E¹⁶,K²⁰] modifications are represented with sticks in lime green for Cα atoms. C. Close up of this peptide (9) bound to the secretin holoreceptor. The hydrophobic receptor residues surrounding the modification Cha²²* (yellow spheres) are Leu³¹, Val³⁴ and Phe⁹², indicated with green sticks, while residue Arg²⁵ is shown in blue.

Fig. 4.. Pharmacological characterization of (Y¹⁰,c[E¹⁶,K²⁰],I¹⁷,Cha²²,R²⁵)sec(5-27).

A. Concentration-dependent inhibition of photoaffinity labeling of the secretin receptor. Illustrated is a typical autoradiograph (of 3 independent experiments) of a 10% SDS-PAGE gel used to separate the products of secretin receptor-bearing CHO cell membranes photoaffinity labelled with the Bpa²⁶-secretin probe in the absence and presence of increasing concentrations of (Y¹⁰,c[E¹⁶,K²⁰],I¹⁷,Cha²²,R²⁵)sec(5-27) (9). B. Densitometric quantification of the inhibition of receptor photolabeling performed in A. C. Concentration-dependent inhibition by (Y¹⁰,c[E¹⁶,K²⁰],I¹⁷,Cha²²,R²⁵)sec(5-27) (9) of cAMP responses in CHO-SecR-low cells stimulated by 0.01 or 0.1 nM sec(1-27) (1). These concentrations of sec(1-27) elicited 21 ± 2% and 68 ± 2% of the maximal response to secretin, respectively. D. Secretin concentration-response curves for stimulation of cAMP accumulation in CHO-SecR-low cells in the presence of 0.1 or 1 μM (Y¹⁰,c[E¹⁶,K²⁰],I¹⁷,Cha²²,R²⁵)sec(5-27) (9) are shifted to the right. Peptide (9) alone does not alter cAMP levels in these cells. E and F. Analogous experiments to those shown in panel D illustrating that (Y¹⁰,c[E¹⁶,K²⁰],I¹⁷,Cha²²,R²⁵)sec(5-27) (9) does not inhibit cAMP accumulation stimulated by either human calcitonin in CHO-CTR expressing cells (E), or human GLP-1(7-36)NH₂ in CHO-GLP-1R expressing cells (F). Data points are mean ± SEM of data from three independent experiments performed in duplicate, normalized relative to the controls. CT, calcitonin. GLP-1, glucagon-like peptide-1.

Fig. 5.. Impact of SecR density on the pharmacology of (Y¹⁰,c[E¹⁶,K²⁰],I¹⁷,Cha²²,R²⁵)sec(5-27) (9).

Left, competition for [¹²⁵I]-Y¹⁰-sec(1-27) binding to CHO cell lines expressing 83,500 (high) (A), 54,000 (medium) (C) and 32,400 (low) (E) sites per cell, or neuroblastoma/glioma cell line, NG108-15, expressing 9,000 sites per cell (G) revealed equivalent affinity for natural secretin and peptide 9 for all cell lines. Values are percentages of saturable binding, expressed as mean ± SEM from three independent experiments performed in duplicate. Right, Shown are peptide-stimulated cAMP responses to sec(1-27) (1) or (Y¹⁰,c[E¹⁶,K²⁰],I¹⁷,Cha²²,R²⁵)sec(5-27) in each of the cell lines (B, D, F, and H, respectively). The (Y¹⁰,c[E¹⁶,K²⁰],I¹⁷,Cha²²,R²⁵)sec(5-27) analogue exhibited no significant intrinsic agonist activity in NG108-15 cells (H) or in the medium (D) or low (F) expressing CHO cell lines, but exhibited significant stimulation of cAMP accumulation in the high SecR-expressing cell line (B) (p = 0.04 at 0.1 μM and 0.03 at 1 μM), revealing partial agonist activity. Also illustrated is the ability of 1 μM (Y¹⁰,c[E¹⁶,K²⁰],I¹⁷,Cha²²,R²⁵)sec(5-27) to right shift the secretin concentration-response curve in NG108-15 cells (H). Data points are mean ± SEM of data from three independent experiments performed in duplicate, normalized relative to the maximal responses stimulated by secretin. * indicates data points for the putative antagonist that are significantly different from the basal levels in t-test, p < 0.05.

Fig. 6.. Further amino-terminal truncation of peptide (9) generates antagonist peptides without discernable intrinsic efficacy.

A. cAMP accumulation in CHO-SecR-high cells stimulated with sec(1-27) (1), (Y¹⁰,c[E¹⁶,K²⁰],I¹⁷,Cha²²,R²⁵)sec(6-27) (10) or (Y¹⁰,c[E¹⁶,K²⁰],I¹⁷,Cha²²,R²⁵)sec(7-27) (11) (n = 3). B. Inhibition of [¹²⁵I]-Y¹⁰-sec(1-27) binding to CHO-SecR-high cells by each of the peptides. Values are percentages of saturable binding, expressed as the mean ± SEM from three independent experiments performed in duplicate. C. Peptide-stimulated mobilization of intracellular calcium (n = 4). D. Peptide-stimulated cAMP accumulation in COS cells transiently transfected with both SecR and Gαs constructs for either (Y¹⁰,c[E¹⁶,K²⁰],I¹⁷,Cha²²,R²⁵)sec(5-27) (9) (n = 5) or (Y¹⁰,c[E¹⁶,K²⁰],I¹⁷,Cha²²,R²⁵)sec(6-27) (10) (n = 3). Peptide 9 was a partial agonist under these conditions, with p = 0.03 at 10 nM, 0.006 at 100 nM, and 0.006 at 1 μM. Transfection with Gαs resulted in higher basal levels of cAMP in these cells than in control cells. E. Effects of sec(1-27) (1), (Y¹⁰,c[E¹⁶,K²⁰],I¹⁷,Cha²²,R²⁵)sec(5-27) (9) or (Y¹⁰,c[E¹⁶,K²⁰],I¹⁷,Cha²²,R²⁵)sec(6-27) (10) on β-arrestin-2 translocation to the plasma membrane in CHO-SecR-high cells. Shown are representative fluorescence images of translocation of β-arrestin-2-eGFP after 5 min at 37 °C, stimulated by buffer (left panels), 10 nM and 100 nM sec(1-27) (1) (next column of panels), 10 and 100 nM (Y¹⁰,c[E¹⁶,K²⁰],I¹⁷,Cha²²,R²⁵)sec(5-27) (9) (next column of panels), or 10 and 100 nM (Y¹⁰,c[E¹⁶,K²⁰],I¹⁷,Cha²²,R²⁵)sec(6-27) (10) (right panels). Data are representative of three independent experiments. Arrows indicate fluorescent β-arrestin-2 translocated to the plasma membrane in response to stimulation with sec(1-27). F. Concentration-dependent inhibition of 0.01 nM or 0.1 nM secretin-stimulated cAMP responses in CHO-SecR-high cells by (Y¹⁰,c[E¹⁶,K²⁰],I¹⁷,Cha²²,R²⁵)sec(6-27) (10). These concentrations of sec(1-27) elicited 21 ± 2% and 68 ± 2% of the maximal response to secretin, respectively. G. Inhibition of secretin-stimulated cAMP accumulation in CHO-SecR-high cells by 0.1 and 1 μM (Y¹⁰,c[E¹⁶,K²⁰],I¹⁷,Cha²²,R²⁵)sec(6-27) (10). Data points are mean ± SEM of data from three independent experiments performed in duplicate, normalized relative to controls. * indicates data points significantly different from the basal levels in t-test, p < 0.05.

Fig. 7.. Ability to transform the SecR antagonist to a full agonist.

Concentration-dependent stimulation of cAMP accumulation (A and B) and mobilization of intracellular calcium (C and D)) in CHO-SecR-high (A and C) CHO-SecR-low (B and D) cells by sec(1-27) (1), (Y¹⁰,c[E¹⁶,K²⁰],I¹⁷,Cha²²,R²⁵)sec(6-27) (10) and sec(1-5)-(Y¹⁰,c[E¹⁶,K²⁰],I¹⁷,Cha²²,R²⁵)sec(6-27) (12). Values represent means ± SEM of data from at least three independent experiments performed in duplicate, normalized relative to controls. Shown also are curves illustrating concentration-dependent inhibition of secretin-like radioligand binding to CHO-SecR-high (E) and CHO-SecR-low (F) cells by these peptides. Values represent percentages of saturable binding, expressed as the means ± SEM of data from at least three independent experiments performed in duplicate.