Engineered odorant receptors illuminate structural principles of odor discrimination

Abstract

A central challenge in olfaction is understanding how the olfactory system detects and distinguishes odorants with diverse physicochemical properties and molecular configurations. Vertebrate animals perceive odors via G protein-coupled odorant receptors (ORs). In humans, ~400 ORs enable the sense of smell. The OR family is composed of two major classes: Class I ORs are tuned to carboxylic acids while Class II ORs, representing the vast majority of the human repertoire, respond to a wide variety of odorants. How ORs recognize chemically diverse odorants remains poorly understood. A fundamental bottleneck is the inability to visualize odorant binding to ORs. Here, we uncover fundamental molecular properties of odorant-OR interactions by employing engineered ORs crafted using a consensus protein design strategy. Because such consensus ORs (consORs) are derived from the 17 major subfamilies of human ORs, they provide a template for modeling individual native ORs with high sequence and structural homology. The biochemical tractability of consORs enabled four cryoEM structures of distinct consORs with unique ligand recognition properties. The structure of a Class I consOR, consOR51, showed high structural similarity to the native human receptor OR51E2 and yielded a homology model of a related member of the human OR51 family with high predictive power. Structures of three Class II consORs revealed distinct modes of odorant-binding and activation mechanisms between Class I and Class II ORs. Thus, the structures of consORs lay the groundwork for understanding molecular recognition of odorants by the OR superfamily.

Figure 1.. Consensus odorant receptor strategy.

a) Consensus odorant receptor (consOR) design strategy. All 23 human OR51 subfamily sequences are aligned and the most conserved amino acid is selected at each position to create a consensus sequence. The conserved region in TM3 of the OR51 subfamily is highlighted here. b) Phylogenetic tree of the OR51 subfamily including consensus OR51 (consOR51), which occupies the root of the subfamily tree. c) Cell surface expression of HEK293 cells transiently transfected with vector control, individual OR51 family members, or consOR51. Most OR51 family members are poorly expressed at the cell surface, with the exception of OR51E2. ConsOR51 shows a dramatic increase in cell surface expression. d) Cryo-EM density map of consOR51in complex with G_s heterotrimer and stabilizing nanobody Nb35. e) Zoom in view of the putative odorant binding site in consOR51 shows a lack of identifiable density for an odorant.

Figure 2.. Structure of consOR51 provides insight into native OR51 family members.

a) Comparison of cryo-EM structure of consOR51 to cryo-EM structure of human OR51E2 indicates high degree of similarity in the 7TM domains and the extracellular loops. Close-up view of odorant binding pocket in consOR51 (b) compared to the propionate binding pocket of OR51E2 (c). Conserved side chains show similar rotamers. d) ConsOR51 is constitutively active in a Glosensor cAMP production assay. Introduction of the F110G mutation in consOR51 leads to suppression of basal activity and response to fatty acids of varying aliphatic chain length. Data points are mean ± standard deviation from n = 3 replicates. e) A homology model of human OR51E1 was constructed using consOR51. f) Docked structure of pentanoic acid in the OR51E1 homology model. g) OR51E1 recognizes long-chain fatty acids, with a preference for pentanoic acid (C5). Selectivity for fatty acid chain length is altered in OR51E1-I205A. Data points are mean ± standard deviation from n = 3 replicates.

Figure 3.. The structure of consOR1 provides insight into human OR1A1.

a) Phylogenetic tree of the human OR1 subfamily including consOR1. b) ConsOR1 is activated by diverse odorants as measured by a Glosensor cAMP production assay. Area under the dose response curve was calculated and normalized to the no odorant negative control (n = 3). c) Dose response for L-menthol activation of consOR1. d) Cryo-EM map of the consOR1-G_s complex. Insert shows cryo-EM density for L-menthol. e) View of the consOR1 odorant binding pocket within 5 Å with a single hydrogen bond shown as dashed lines. f) Mutagenesis studies of consOR1 in a cAMP accumulation assay. g) OR1A1 is activated by terpenoids L-menthol and R-carvone. h) Homology model of OR1A1 based on consOR1 i) The OR1A1-N155A mutation has a larger effect on R-carvone activity as compared to L-menthol. j) OR1A1 mutants differentially affect R-carvone or L-menthol activity. Area under the dose-response curve was calculated for each OR1A1 mutant activated by either odorant (n = 3). For each odorant, AUC values were normalized to wildtype OR1A1. Subtraction of normalized AUCs revealed a differential effect of mutations. Docking of L-menthol (k) and R-carvone (l) docked to the homology model of OR1A1. Top scoring docking results are shown for both odorants as transparent sticks. The best scoring pose is shown as solid sticks. m) Mapping the effect of mutations in (j) onto the homology modeled structure of OR1A1 shows that mutations that affect L-menthol and R-carvone cluster in distinct regions of the OR1A1 binding pocket. Every residues with a delta of more than 0.10 are reported colored by the odorant the most affected. For all cell assay, data points are mean ± standard deviation from n = 3 replicates.

Figure 4.. Structural flexibility in odorant binding.

a) Molecular dynamics simulations were performed for consOR1: with miniGα_s and L-menthol, without miniGα_s but with L-menthol, and without both miniGα_s and L-menthol. Snapshots from 100 ns intervals are shown from a representative simulation. In consOR1, TM5 and TM6 are more flexible in the absence of miniGα_s, and even more dynamic in the absence of miniGαs and L-menthol. L-menthol is dynamic in the binding pocket of consOR1 in simulations with miniGα_s (b), and shows even greater flexibility in simulations without miniGα_s (c). d) The root mean squared deviation (RMSD) of L-menthol compared to the cryo-EM pose for each simulation replicate is shown. × indicates p<0.05. e,f) In simulations of OR51E2, propionate is constrained within the ligand binding pocket and makes a persistent interaction with R262. g) The root mean squared deviation (RMSD) of propionate compared to the cryo-EM pose for each simulation replicate is shown. n.s. Indicates not significant.

Figure 5.. Structures of consOR2 and consOR4 and insights into common features of OR function.

a) CryoEM map of consOR2-G_s complex bound to activating odorant S-carvone. b) CryoEM map of consOR4-G_s complex bound to activating odorant 2-methylthiazoline (2-MT). c) Comparison of Class I and Class II OR structures in the extracellular region. Consensus OR structures of Class II ORs show variability in ECL3 conformation. d) Close-up view of ligand binding sites in Class I and Class II ORs. Class I ORs recognize carboxylic acids via the R6x59 residue in the extracellular portion of TM6. Class II ORs bind ligands via a highly conserved Y6x55 residue that further engages a conserved acidic residue in ECL2 (D/E^45x51). e) The interaction between D/E^45x51 and Y^6x55 is maintained in simulations of consOR1, consOR2 and consOR4 bound to their agonist and miniGα_s. Removal of miniGα_s and agonist leads to an increase in distance between these two positions (unpaired t-test, p < 0.0001). f) Mutation of D^45x51 and Y^6x55 in consOR1, consOR2 and consOR4 reduces OR response to odorant in a cAMP production assay. For all cell assay, data points are mean ± standard deviation from n = 3 replicates.

Figure 6.. Accessing Class I and Class II OR mechanisms and structures.

a) Model of Class I OR activation mechanism. b) Model of Class II OR activation mechanism. c) OR phylogenetic tree including the structurally elucidated consORs showing that these structures allow the homology modeling of 34% of the human native ORs. OR belonging to a consOR subfamily are highlighted by a rounded tip and ORs showing at least 60% of sequence identity with a consOR are shown in colored lines. The scale represents the amount of amino acid change for a set distance.