Structural Insights into the Activation of Human Relaxin Family Peptide Receptor 1 by Small-Molecule Agonists

Abstract

The GPCR relaxin family peptide receptor 1 (RXFP1) mediates the action of relaxin peptide hormone, including its tissue remodeling and antifibrotic effects. The peptide has a short half-life in plasma, limiting its therapeutic utility. However, small-molecule agonists of human RXFP1 can overcome this limitation and may provide a useful therapeutic approach, especially for chronic diseases such as heart failure and fibrosis. The first small-molecule agonists of RXFP1 were recently identified from a high-throughput screening, using a homogeneous cell-based cAMP assay. Optimization of the hit compounds resulted in a series of highly potent and RXFP1 selective agonists with low cytotoxicity, and excellent in vitro ADME and pharmacokinetic properties. Here, we undertook extensive site-directed mutagenesis studies in combination with computational modeling analysis to probe the molecular basis of the small-molecule binding to RXFP1. The results showed that the agonists bind to an allosteric site of RXFP1 in a manner that closely interacts with the seventh transmembrane domain (TM7) and the third extracellular loop (ECL3). Several residues were determined to play an important role in the agonist binding and receptor activation, including a hydrophobic region at TM7 consisting of W664, F668, and L670. The G659/T660 motif within ECL3 is crucial to the observed species selectivity of the agonists for RXFP1. The receptor binding and activation effects by the small molecule ML290 were compared with the cognate ligand, relaxin, providing valuable insights on the structural basis and molecular mechanism of receptor activation and selectivity for RXFP1.

Figure 1

(A) Structural models of human RXFP1(hRXFP1). The active conformation is shown in green and the inactive conforamtion is shown in blue. The TM6 and ECL3 of the active and inactive form of hRXFP1 are shown in red and dark yellow. Key residues in the binding pocket of the 7TM domain are shown in a separate panel. (B) Sequence alignment of the TM6/TM7 and ECL3 of hRXFP1 with other four GPCRs: human β2 adrenergic receptor (hB2AR), turkey β1 adrenergic receptor (tB1R), human A2A adenosine receptor (hA2AR), bovine rhodpsin receptor (bRhoR).

Figure 2

MD simulations of the hRXFP structural models. (A) RMSDs of the backbone atoms of hRXFP1 in the apo and ligand-bound forms during the 50-ns simulations. (B) RMSDs of the backbone atoms of the extracellular loop ECL3 with respect to their starting structure during the 50-ns simulations.

Figure 3

Binding interaction of ML290 with hRXFP1. The agonist ML290 is shown in stick (carbon atoms in cyan, oxygen and nitrogen atoms are shown in red and blue). Key residues within hRXFP1 involved in binding interaction are shown in green/yellow/magenta (carbon atoms). The two phenyl rings of ML290 are labeled as A and B on the right panel. The β₂AR agonist BI-167107 (PDB code 3P0G) is superimposed in the binding pocket of hRXFP1 and shown in yellow on the left panel.

Figure 4

(A) Derivatives of the hRXFP1 agonist ML290 and predicted binding energies from docking study. (B) Plot of correlation between experimental EC₅₀’s and predicted binding energy (R² = 0.89).

Figure 5

Predicted binding models of derivatives of ML290 agonist within hRXFP1. Key residues of hRXFP1 are shown in stick and the two regions are circled in red and blue. The hydrophobic pocket in the binding region A is depicted in a surface representation on the left panel.

Figure 6

(A) 2D diagram of binding interaction of ML290 within hRXFP1. (B) Binding free energy decomposition of key residues of hRXFP1 involved in binding interaction with ML290.

Figure 7

Site-directed mutagenesis results. (A) The maximum response of hRXFP1 mutants stimulated by relaxin (RLN) and ML290. The activity of cAMP elevation was normalized to FSK control. (B) Surface expression of hRXFP1 mutants in transfected HEK293T cells. Data was normalized to wild-type (WT) expression. (C) mutant G659D dose-response activity by relaxin and ML290 in comparison with the wild-type hRXFP1. (D) mutant T660A dose-response activated by relaxin and ML290 in comparison with the wild-type hRXFP1.

Figure 7

Site-directed mutagenesis results. (A) The maximum response of hRXFP1 mutants stimulated by relaxin (RLN) and ML290. The activity of cAMP elevation was normalized to FSK control. (B) Surface expression of hRXFP1 mutants in transfected HEK293T cells. Data was normalized to wild-type (WT) expression. (C) mutant G659D dose-response activity by relaxin and ML290 in comparison with the wild-type hRXFP1. (D) mutant T660A dose-response activated by relaxin and ML290 in comparison with the wild-type hRXFP1.

Figure 8

(A) Binding interaction of ML290 within hRXFP2. ML290 is shown in stick. The extracellular loop ECL3 is shown in magenta color. Key residues are shown in green stick. (B) Binding interaction of ML290 within mouse RXFP1 (mRXFP1). An intramoleclular hydrogen bond is formed between residue D659 and T662, which may play a role in ligand activation by locking the receptor in an inactive state.

Figure 9

Proposed Binding Model & Activation of Mechanism of hRXFP1 by agonist ML290.