G protein-coupled receptors (GPCRs) are the largest superfamily of membrane proteins, which transduce signals from a wide range of stimuli across the plasma membrane to activate intracellular effector proteins, most importantly G proteins, GPCR kinases, and the β-arrestin family of proteins. Ligand binding to the extracellular, so-called orthosteric, ligand-binding pocket of GPCRs induces global conformational alterations in the seven transmembrane helix domain (TMD), including the prominent outward movement of helix 6 (TM6) and inward movement of TM7 in the cytosolic region. These conformational changes open the intracellular crevice for the binding and subsequent activation of G proteins. Activation of G protein is followed by desensitization of the receptor largely via GPCR receptor kinase (GRK) phosphorylation-induced β-arrestin-mediated receptor internalization.
GPCRs usually exhibit a basal level of activity in the apo-state that is modulated in an efficacy-dependent manner upon ligand binding to the extracellular orthosteric binding site, where agonists fully activate the receptor and inverse agonists suppress the basal activity. The functional output of the receptor can be further modulated by binding of allosteric ligands to secondary binding pockets or by biased ligands that selectively activate one of the signalling pathways over the others. Central concepts for understanding such pharmacological behaviour in GPCR signalling are (1) a pre-existing equilibrium of fully inactive to fully active conformations in the conformational landscape of apo-state receptor and (2) a population shift towards the active conformation upon agonist binding that has been supported by NMR, EPR, single molecule studies as well as MD simulations. For example, recent NMR studies on Adenosine 2 A (A 2A) receptor indicate sampling of active state conformations of the receptor that are further populated in the presence of ligand and heterotrimeric G proteins. The population-shift model addresses the equilibrium aspects of GPCR signalling and allostery, irrespective of the kinetic pathways underlying the coupling of population shift and ligand binding. In the case of two pre-dominant conformations, such as an inactive receptor conformation R 1 and an active conformation R 2, there are two pathways along which a population shift from R 1 to R 2 during ligand binding can occur (Supplementary Fig.1): an induced-fit pathway, along which ligand binding to R 1 and formation of encounter complexes precede the conformational change to R 2 and induce the population shift; and a conformation-selection pathway in which the conformational change from R 1 to R 2 precedes ligand binding. Despite the pharmaceutical importance of GPCRs, the kinetic mechanism underlying population shifts in GPCRs is poorly understood. Insight into this mechanism may aid tailoring of selective designer molecules with desired pharmacological output against GPCRs.
Neurotensin receptor 1 (NTS 1) is a class A GPCR that is primarily expressed in the central nervous system and gastrointestinal tract and activated by the endogenous 13-residue linear peptide neurotensin, pELYENKPRRPYIL. The last six residues of NT (NT8-13) have been demonstrated to be the primary epitope of the peptide for high affinity receptor binding and activation and have been used as a scaffold for development of NTS 1 targeting drug candidates. NTS 1 regulates neurological processes including dopamine transmission and GABAergic system modulation and is considered as a promising target for treatment of addiction and schizophrenia. Crystal structures of NTS 1 in complex with different ligands showed efficacy-dependent modification of the volume of the ligand-binding pocket, where agonists contract the binding pocket and inverse agonists expand its volume. The conformational changes of the binding pocket of NTS 1 have been further investigated by NMR. However, how receptor conformational changes and dynamics are kinetically linked to ligand recognition remains unclear.
In this study, we combined 19 F NMR experiments, hydrogen–deuterium exchange mass spectrometry (HDX-MS) and stopped-flow fluorescence kinetics to address the mechanism underlying NT recognition and activation of NTS 1. Our ligand-observed 19 F-NMR experiments, on fluorinated full agonist analogues of NT, and receptor-observed experiments, on NTS 1 labelled with fluorinated unnatural amino acids, revealed formation of NTS 1 conformers upon ligand binding that are in slow conformational exchange. HDX-MS demonstrated that this conformational heterogeneity arises from the interaction between the N-terminal region of the receptor and extracellular loop 2 (ECL2). Further kinetic analysis of binding of NT to NTS 1 using stopped-flow fluorescence proposes an induced-fit mechanism of binding underlying NT recognition by NTS 1.
NMR is well-suited to obtain atomic resolution insight into the dynamics of biomolecular systems. However, common isotope labelling schemes, including 13 C and 15 N, are cumbersome to study the dynamics of large systems such as GPCRs. Recently, 19 F-labelled aromatics have re-gained popularity as NMR probes to study conformational dynamics of GPCRs due to the sensitivity of 19 F-aromatics as chemical microenvironmental sensors, resulting in high resolution NMR spectra. For example, fluorinated ligands have been used to unravel the conformational heterogeneity in the orthosteric pocket of the neurokinin receptor. Moreover, the large gyromagnetic ratio of 19 F enables working with low concentrations of the receptor. In this context we were inspired to develop 19 F-NT analogues, by substituting Tyr11 in NT, to investigate the mechanisms of ligand recognition by the receptor. We substituted Tyr11 with para-trifluoromethyl-phenylalanine (tfmF) to produce a sensitive fluorinated analogue of NT (Y11tfmF-NT) (Supplementary Fig.2) with minimal modification. This substitution reduced the affinity of the peptide for wt-rNTS 1 by 100-fold (Supplementary Fig.3a), which may be due to the steric bulk of the CF 3 group and its low tendency to form hydrogen bonds. Perturbing hydrogen bonding of Tyr11 by phenylalanine substitution shows similar effects, but less pronounced than tfmF probably due to the smaller size of phenylalanine compared to tfmF. However, the presence of the CF 3 moiety leaves the affinity of Y11tfmF-NT for the engineered NTS 1 variant, enNTS 1, unaffected (Supplementary Fig.3b), suggesting that the introduced mutations during thermostabilization of the receptor (Supplementary Fig.4) may account for this gain of affinity. Nonetheless, these probes are equally efficacious in BRET assays measuring G protein (Supplementary Fig.3c) and β-arrestin (Supplementary Fig.3d) recruitment and
