• We are sorry, but NCBI web applications do not support your browser and may not function properly. More information
Logo of nihpaAbout Author manuscriptsSubmit a manuscriptNIH Public Access; Author Manuscript; Accepted for publication in peer reviewed journal;
Neuron. Author manuscript; available in PMC Dec 10, 2009.
Published in final edited form as:
PMCID: PMC2652502
NIHMSID: NIHMS84472

High potency olfactory receptor agonists discovered by virtual high-throughput screening: molecular probes for receptor structure and olfactory function

Abstract

The detection and discrimination of diverse chemical structures by the vertebrate olfactory system is accomplished by the recognition of odorous ligands by their cognate receptors. In the present study we used a computational high-throughput screening strategy to discover novel high affinity agonists of an olfactory G protein-coupled receptor tuned to recognize amino acid ligands. Functional testing of the top candidates validated several agonists with potencies higher than any of the receptor’s known natural ligands. Computational modeling revealed molecular interactions involved in ligand recognition by this receptor, and further highlighted interactions that have been conserved in evolutionarily divergent amino acid receptors. Significantly, the top compounds display robust activities as odorants in vivo, and include a natural product that may be used to signal the presence of bacteria in the aquatic environment. Our virtual screening approach should be applicable to the identification of new bioactive molecules for probing the structure of chemosensory receptors and the function of chemosensory systems in vivo.

Introduction

The vertebrate olfactory system receives and decodes sensory information from a myriad chemical cues. The first step in this process is the recognition of these cues by receptors expressed by the primary sensory neurons in the olfactory epithelium (Firestein, 2001). Vertebrate olfactory receptors comprise four different families of G protein-coupled receptors (Mombaerts, 2004) (GPCRs): the OR receptor gene family, the largest family with ~1,000 functional members in some mammalian species (Zhang et al., 2004); the trace amine-associated receptors (Liberles and Buck, 2006) (TAARs; <20 members); the V1R vomeronasal receptors (Zhang et al., 2004) (~150 members); and the V2R vomeronasal receptors (Yang et al., 2005) (~60 members). The V2R receptors belong to the C family of GPCRs, which includes the calcium sensing receptor (CaSR), metabotropic glutamate (mGlu) receptors, GABA-B receptors, and T1R taste receptors (Pin et al., 2003).

We previously identified a V2R-like receptor from the goldfish olfactory epithelium that is activated by all 20 natural amino acids, which are potent odorants for fish (Luu et al., 2004; Speca et al., 1999). This receptor, called receptor 5.24, responds preferentially to the long chain basic amino acids, lysine and arginine, although other amino acids can bind to this receptor with lower affinities. This broad tuning of receptor 5.24 embodies the promiscuous nature of the odorant receptors, a mechanism that allows the olfactory system to recognize a diversity of chemical structures exceeding the actual number of receptors encoded by the genome. It is therefore of great interest to elucidate the molecular determinants of ligand selectivity – using receptor 5.24 as a prototypical receptor – in order to understand how the olfactory/vomeronasal C family GPCRs have evolved to recognize their cognate ligands.

Unlike other GPCRs, members of the C family GPCRs are characterized by a large extracellular N-terminal domain (NTD), the location of the orthosteric ligand binding site. The NTD adopts a conserved clamshell-like fold – also referred to as the Venus Fly Trap Domain (VFTD) – with two lobes connected by a flexible hinge. Analysis of protein crystal structures and molecular modeling have identified ligand interactions with the inner surfaces of lobes 1 and 2 that stabilize a closed conformation of the VFTD, leading to receptor activation (reviewed by Pin et al., 2003). An inspection of the core binding residues reveals numerous potential contacts with ligand that can be sorted into two groups: the “proximal” and “distal” binding pockets. The proximal pocket residues are predicted to bind the amino acid ligand’s glycine moiety (i.e., the α-carboxyl together with the α–amino group and α–proton). Residues residing in the distal pocket interact with the amino acid ligand’s side chain and are responsible for conferring selectivity for distinct side chain structures. In the case of goldfish receptor 5.24, through homology modeling we previously identified several distal pocket residues that can account for this receptor’s preference for long chain basic amino acids (Luu et al., 2004).

In the present study, we sought to utilize novel chemical structures to probe more deeply into the structure and function of the receptor 5.24 binding pocket. We further wished to know whether high potency agonists that were selected based on their interactions with a single receptor would also be active in eliciting olfactory responses in vivo. To this end, we developed and applied a suite of computational techniques to screen for receptor 5.24 agonists. This “virtual high-throughput screening” (vHTS) approach identified numerous active compounds, with several showing significantly higher potency than any of the previously known natural ligands for this receptor. Docking of the most active compounds in three-dimensional models of the receptor confirmed the importance of several binding pocket residues in determining affinity and selectivity. Interestingly, analysis of one series of ligands reveals a conserved ligand-stabilized helix-helix interaction in lobe 1 that is associated with ligand recognition and receptor activation in evolutionarily divergent amino acid receptors. Finally, electrophysiological recordings from goldfish olfactory epithelium indicate that the computationally identified agonists can indeed elicit robust responses by olfactory sensory neurons in vivo. One novel odorant discovered, diaminopimelic acid, is a precursor in the lysine and peptidoglycan biosynthetic pathways of bacteria, suggesting that the fish olfactory system may use the presence of this metabolite to detect bacteria in their environment. Together these results demonstrate the utility of our vHTS approach in identifying novel compounds for probing C family GPCR structure and function and for interrogating olfactory function in vivo.

Results

Virtual screening for the discovery of novel receptor agonists

Elucidation of the principles underlying ligand binding and activation in the metabotropic glutamate receptors has been greatly facilitated by an array of high affinity agonists and antagonists (Pin et al., 1999). By analogy, we expect that the identification of novel high affinity ligands for receptor 5.24 will similarly provide new molecules with which to probe the structure and function of this, and potentially other chemosensory C family GPCRs. To achieve this goal, we implemented a computational screening paradigm to search for ligands based on two independent criteria: (1) chemicals that “fit” in the predicted ligand binding pocket structure (structure-based vHTS), and (2) chemicals that exhibit features identified by a pharmacophore model of activating ligands (ligand-based vHTS). Since our vHTS approach requires a detailed structural model of the receptor binding pocket as well as an extensive pharmacological profile for the receptor, we implemented this vHTS approach on goldfish receptor 5.24, for which we already have extensive data (Luu et al., 2004).

vHTS protocol design and optimization

As with any screening protocol, a balance must be found to maximize the identification of true positive “hits” while avoiding an inordinate number of false positives. We therefore designed the computational protocols according to the receiver operating characteristic (ROC) curve approach, which provides an objective measure of the ability of a given screening workflow to discriminate known active molecules from known inactive molecules – a critical test of the design of any screening protocol (Triballeau et al., 2005). Thus, the ROC curve approach can be used to facilitate decision making about the selection criteria used for computational screening.

Computational protocols were designed and optimized for both structure- and ligand-based vHTS workflows according to the ROC curve method described previously (Triballeau et al., 2005) (see Figure S1 and Supplemental Experimental Procedures for details). The structure-based vHTS protocol utilized a molecular model of the receptor 5.24 binding pocket (Luu et al., 2004). For ligand-based vHTS, we used pharmacophore modeling to identify compounds in chemical space that share characteristics similar to known receptor 5.24 ligands (Speca et al., 1999). Our resulting pharmacophore model (Figure S2) comprises features that are consistent with the conformation and interaction of L-arginine docked in a structural model of receptor 5.24 (Luu et al., 2004).

Chemical library and vHTS implementation

We used our vHTS protocols to screen a database representing a compilation of 1.6 million commercially available compounds (see Experimental Procedures). A series of computational filters (Triballeau et al., 2005) was applied to the database prior to implementation of the vHTS workflows to reduce the number of candidates to 758 unique chemical structures (see Supplement for details). Screening the 758 compounds using our vHTS protocols for receptor 5.24 resulted in 468 compounds that satisfied our selection criterion in the structure-based protocol and 50 compounds that met our criterion in the ligand-based protocol. Forty-six compounds are common to both lists (“common hits”), and include L-amino acids known to activate the receptor. The vHTS workflow and results are summarized in Table 1 and Figure S3.

Table 1
vHTS summary: number of compounds retained after each selection stage.

Functional validation of compounds identified by vHTS

We reasoned that chemical structures identified by two independent approaches should comprise the most promising candidates for subsequent validation by functional assays. Thus, we focused mainly on compounds that were retained in common by both screening protocols. Compounds were functionally validated by screening for activation of receptor 5.24 expressed in HEK 293 cells, using a calcium mobilization assay in 96 well plates. Chemicals were screened at 20 μM, a concentration ~10-times higher than the EC50 for the receptor’s most potent natural agonists, lysine and arginine (EC50 = ~2 μM). We reasoned that only those compounds showing significant activity at this concentration would be viable candidates as high affinity agonists. Thirty-six compounds were initially tested from the common hit list. After examining the structures of the most active compounds identified in this preliminary functional screen (Table 2), we selected additional compounds from the structure-based (7 compounds) or ligand-based (2 compounds) vHTS hit lists, as well as a few structural analogues of the top hits. Altogether 54 compounds were tested, with 45 from the three vHTS hit lists (Figure S4, Tables 1, ,2,2, and S1). Forty-two percent (19/45) of the compounds tested from the vHTS hit lists showed significant activity at 20 μM (operationally defined as ≥ 40% maximum activity), which is remarkable given the more typical validation rate of 10% or less in such screening protocols (Augen, 2002; Triballeau et al., 2005; Triballeau et al., 2006). In addition, of the 9 compounds selected based on their structural similarities to the active compounds, 5 exhibited significant activity as receptor 5.24 agonists. We next generated dose-response curves to calculate EC50 values for the 24 validated compounds (Tables 2 and S1).

Table 2
Activities of the top five vHTS hits and known natural agonists.

Structure-activity analysis of vHTS hits reveals a strong preference for amino acid ligands with unbranched side chains

Our previous molecular modeling study of receptor 5.24 predicted interactions between specific residues and contact sites on the bound ligand’s glycine core (the so-called proximal binding pocket) as well as on the ligand’s side chain (the “distal” binding pocket) (Luu et al., 2004). It is striking that all vHTS hits that were validated as active compounds are linear amino acids with no substitutions at the β and γ positions (Table S1). This characteristic probably reflects the restricted steric tolerance of the residues interacting with this portion of the side chain, serine-151 (S151) on lobe 1 and asparagine-310 (N310) on lobe 2. From our three-dimensional model of receptor 5.24, the bound ligand’s side chain’s β and γ methylenes are in Van der Waals contact with S151 and N310 (Luu et al., 2004) (Figure S5) These contacts would preclude substitutions on the ligand’s side chain and explain why, for example, phenylalanine, tyrosine, tryptophan and their derivatives are poor agonists of receptor 5.24.

Molecular determinants of ligand binding and receptor activation revealed by top vHTS hits

The identification of novel, high potency agonists provides an opportunity to probe the molecular interactions underlying ligand recognition and receptor activation. A close look at the side chains of the most active molecules among the hits identified by both the docking and pharmacophore models allowed us to classify them into three groups: lysine/arginine-like (hit #’s 244, 48, 677), glutamine-like (hit #’s 338, 506, 586, 137, 583) and di-amino acids (hit #’s 380, 299, 565, 611, 652, 479) (Tables 2 and S1). Most of the other validated hits (identified by the docking model alone, or selected based on structural similarity to some of the top hits) belong to one of these three categories (Table S1). To gain insight into the interactions underlying the properties of these novel receptor agonists, we generated molecular models of receptor 5.24 bound to selected compounds from the three groups of chemicals.

Lysine/arginine analogues

In the lysine- and arginine-like series, substitution of a methylene group with an oxygen atom in the alkyl chain of lysine (L-oxalysine, hit # 244) or arginine (L-canavanine, hit # 677) results in a ~2-fold increase in ligand potency (Table 2). The interactions with the ligands’ terminal amino (oxalysine) or guanidinium (canavanine) groups (Figure 1A and B) are similar to those predicted for lysine and arginine, respectively (Luu et al., 2004). In addition, each analogue’s oxygen heteroatom may provide polar interactions with a cluster of serine residues in lobe 1 (S111, S150, S151), which would explain the increased activity of oxalysine and canavanine relative to lysine and arginine, respectively (Figure 1A and B). Analogues containing sulphur heteroatoms at these positions might be expected to have a similar (but weaker) effect on ligand potency. Indeed, L-4-thialysine, which contains a sulphur heteroatom in place of the oxygen heteroatom of L-oxalysine, was retained by the structure-based vHTS protocol but with a lower score than L-oxalysine and therefore not selected for further study (data not shown). Restricting the flexibility of the alkyl chain to favor the extended bioactive conformation as in K1 or D425 does not significantly enhance the potency of the arginine- and lysine-like ligands (Table S1).

Figure 1
Molecular models of selected vHTS hits docked in the receptor 5.24 binding pocket. Lysine and arginine analogues: (A) L-oxalysine (hit # 244) and (B) L-canavanine (hit # 677); di-amino acids: (C) L-cystathionine (hit # 380) and (D) LL-α,ε-diaminopimelic ...

Di-amino acids

The strongest activation in this class was observed with hit # 380 (L-cystathionine; EC50 = 1.7 μM), and hit # 299 (LL- and DL-α,β-diaminopimelic acid; EC50 = 2.4 μM for a mixture of LL, DL, and DD enantiomers) (Table 2). Docking of L-cystathionine in a structural model of receptor 5.24 suggests ligand-receptor interactions with the distal carboxylate in addition to interactions with the distal amino group similar to those predicted for L-lysine or L-oxalysine (compare Figure 1A and 1C). In contrast, docking of LL- and DL-α,β-diaminopimelic acid suggests a reorganization of the distal binding pocket interactions in order to accommodate the distal acidic group of this di-amino acid in either diastereoisomeric configuration (the LL isomer is shown in Figure 1D). The docking model also predicts hydrogen bonds with the sulphur heteroatom of L-cystathionine (Figure 1C), as seen with the oxygen heteroatoms in the oxalysine and canavanine side chains (Figure 1A and B). However, when the adjacent δ-CH2 group of L-cystathionine is replaced by a sulphur atom (hit # 479) or selenium atom (hit # 443), the potency is markedly decreased (Table S1).

Glutamine analogues

While glutamine displays modest potency (EC50 = 9.7 μM), we discovered that its N-4-nitrophenyl derivative, L-glutamic acid-γ-p-nitroanilide (hit # 338; EC50 = 0.72 μM) is over 10 times more potent than L-glutamine and 2~3-fold more potent than lysine or arginine, the two most potent natural amino acid ligands of receptor 5.24 (Luu et al., 2004; Speca et al., 1999). This enhanced potency derives mainly from the nitro moiety, as the N-phenyl glutamine analogue (compound Q2, EC50 = 7.9 μM for the racemic mixture) is only slightly more potent than L-glutamine. None of the other substitutions of the phenyl ring or polar N-substituents such as -NH2 or –OH, induce a similar increase in potency (e.g., hit #’s 506, 583, 588, and compounds Q1 and Q3; Table S1). Several factors may be responsible for the enhanced potency of L-glutamic acid-γ-p-nitroanilide. From molecular modeling, we predict that in addition to making similar contacts as bound glutamine (Figure 1E), the distal nitro group participates in a set of novel polar interactions with Q78 and D388 in the distal pocket (Figure 1F). The distal nitro moiety may also constrain the conformation of L-glutamic acid-γ-p-nitroanilide’s phenyl ring to allow optimal interaction with M389 (and possibly other hydrophobic residues in the vicinity), which we previously demonstrated to be critical for activation by high affinity ligands (Luu et al., 2004).

Identification of a ligand-mediated helix-helix interaction conserved in divergent amino acid receptors

Among the residues predicted to contact the distal nitrophenyl moiety of L-glutamic acid-γ-p-nitroanilide (Figure 1F), K74, Q78, D388 and M389 are located within helices αI and αIX. These two helices comprise part of a conserved lobe 1 core structure in all proteins belonging to the structural periplasmic binding protein-like I superfamily, which includes the mGlu receptors and bacterial amino acid binding proteins such as LIVBP, as well as receptor 5.24 (Luu et al., 2004) (Figure 2A). From the crystal structures of mGlu1 (PDB code 1ewk) and LIVBP (PDB code 1z16), coordination of ligand by these two helices relies upon interactions with residues at positions corresponding to Q78 and M389 in receptor 5.24 (highlighted in Figure 2A). In mGlu1, bound glutamate forms a bridge between the two helices via ionic interactions with R78 on helix αI (via a bridging water molecule) and K409 on helix αIX (Figure 2B). Similarly, in LIVBP, the hydrophobic side chain of bound leucine interacts with Y18 (helix αI) and F256 (helix αIX), drawing the two helices in close apposition (Figure 2C).

Figure 2
Conservation of agonist–helix interactions in the LIVBP-like family of amino acid binding proteins. (A) Sequence alignment of helices αI and αIX (red) of the LIVBP, mGlu1 and receptor 5.24 binding pockets. Residues participating ...

Whereas bound L-arginine is not predicted to interact directly with residues on helix αI (although hydrogen bonding may occur between Q78 on helix αI and D388 on helix αIX; Figure 2D), our model predicts that bound L-glutamic acid-γ-p-nitroanilide forms a bridge between these two helices via a novel interaction with Q78 in helix αI and hydrophobic interactions with M389 (as well as polar interactions with D388) in helix αIX (Figure 2E). It is remarkable that, although the amino acid residues in mGlu1, LIVBP and receptor 5.24 are divergent at positions 78 and 389 (receptor 5.24 coordinates; Figure 2A), the function of these sites in conferring both affinity and selectivity appears to have been conserved in these amino acid binding proteins.

Functional validation of top vHTS hits as odorants in vivo

The data presented thus far demonstrate the utility of our vHTS approach in identifying novel molecules that can be used to probe the structure and function of the receptor 5.24 ligand binding pocket. We next wished to determine whether agonists identified in this way – i.e., based on their interactions with a specific receptor – could also function as odorants by eliciting responses in olfactory sensory neurons. Candidate compounds were tested for activity in vivo by measuring evoked changes in voltage across the goldfish olfactory epithelium. This measurement, known as the electro-olfactogram (EOG) (Ottoson, 1956), is an extracellular field potential that locally summates the activity of a population of cells in the sensory epithelium and is widely used as a reliable assay for olfactory neuron responses in a variety of vertebrates, including fish (Scott and Scott-Johnson, 2002).

We selected four of the top vHTS hits (# 338, L-glutamic acid γ-p-nitroanilide; # 244, oxalysine; # 299, diaminopimelic acid; and # 677, L-canavanine) for testing by EOG in the goldfish olfactory epithelium. As shown by the representative traces in Figure 3, all compounds elicited robust EOG responses at 10−5 M, showing peak EOG amplitudes comparable to the response generated by 10−5 M L-arginine. Testing of L-glutamic acid γ-p-nitroanilide, L-oxalysine, diaminopimelic acid, and L-arginine over a wide concentration range revealed detectable responses at 10−8 M (Table S2). The responses did not show saturation at 10−4 M, which is typical of amino acid-evoked EOGs in fish (Byrd and Caprio, 1982; Rolen et al., 2003) and probably reflects the recruitment of low affinity receptor interactions at higher ligand concentrations.

Figure 3
Compounds identified by vHTS function as odorants in vivo. Electro-olfactogram (EOG) recordings were performed on goldfish olfactory epithelium to measure olfactory responses to L-arginine and selected top vHTS hits. Representative traces are shown in ...

Interestingly, whereas L-glutamic acid γ-p-nitroanilide and L-oxalysine exhibit higher apparent affinities to receptor 5.24 than L-arginine (Table 2), in vivo they elicit somewhat lower peak responses across all concentrations tested (Table S2). A possible explanation for this observation is that L-glutamic acid γ-p-nitroanilide and L-oxalysine interact with a smaller number of receptors than L-arginine (i.e., they are more specific). On the other hand, diaminopimelic acid and L-canavanine are slightly more potent than L-arginine in vivo, suggesting that their interactions with the population of receptors expressed in the olfactory epithelium may be more promiscuous than those of the other compounds (including L-arginine). Whatever the case, EOG recordings clearly demonstrate that the top hits identified by our receptor 5.24-based computational screening paradigm can indeed function as odorants in vivo.

Discussion

In the present study, we applied a computational high-throughput screening approach to discover new agonists of receptor 5.24, an olfactory C family GPCR. Our strategy utilized ligand-based and structure-based models that together identified several agonists showing higher potency than any of the receptor’s previously identified natural ligands. Interestingly, compounds that bind to lobe 1 of an mGlu receptor but prevent closure of the VFTD are antagonists (Bessis et al., 2002). It therefore may be possible to identify receptor 5.24 antagonists by screening for compounds that fit a structural model of the receptor’s ligand binding domain in the open, but not closed conformation. More broadly, our vHTS approach should be applicable to the identification of ligands for other chemosensory C family GPCRs (the V2R/V2R-like vomeronasal/olfactory receptors and T1R taste receptors) (Alioto and Ngai, 2006; Shi and Zhang, 2007) as well as the OR family olfactory receptors, whose transmembrane ligand binding domains have been characterized using molecular modeling techniques (Abaffy et al., 2007; Katada et al., 2005; Schmiedeberg et al., 2007).

Conservation of molecular determinants underlying ligand recognition in the amino acid receptors: functional implications

L-glutamic acid-γ-p-nitroanilide is the most potent agonist identified in this screen. We attribute the enhanced potency of this compound to novel interactions between its distal nitrophenyl moiety and sites on helices αI and αIX of the ligand binding domain. Significantly, the positions of these determinants on helices αI and αIX – and their role in dictating affinity and selectivity –appear to be a conserved feature of highly divergent amino acid receptors, including bacterial LIVBP and a mammalian mGlu receptor (Figure 2). In this regard it is interesting that, among the top 5 vHTS hits, L-glutamic acid-γ-p-nitroanilide – a non-natural amino acid – is the only compound predicted to form direct bridging interactions with helices αI and αIX. The conserved role of these contact sites in receptor 5.24 suggests the existence of naturally occurring high potency ligands that participate in similar ligand-receptor interactions.

Receptor-based computational screening as a tool for the discovery of novel odorants

Electrophysiological recordings from goldfish olfactory epithelium confirmed that at least 4 of the compounds identified by computational screening demonstrate robust activities as odorants in vivo. These results validate the overall approach of using a receptor-specific model as a starting point in the search for novel odorants. Interestingly, one compound identified in our screen, diaminopimelic acid, is a component of gram-negative bacterial cell walls and an intermediate in the bacterial biosynthetic pathways for lysine and peptidoglycans (Born and Blanchard, 1999; Work, 1950). Moreover, the two natural stereoisomers of diaminopimelic acid (LL and DL) can be found in estuaries at concentrations ranging from ~10 to ~100 nM (Jorgensen et al., 2003), well within the sensitivity of the goldfish olfactory system for this chemical. We speculate that, as fish are thought to use metabolites such as amino acids and nucleotides as feeding cues (Hara, 1994), diaminopimelic acid may signal through the olfactory system the presence of bacteria, which presumably would be localized at or near potential food sources. Application of the high-throughput computational screening described here to other olfactory and taste receptors may reveal additional natural products that function as chemosensory cues.

Experimental Procedures

vHTS

Computational protocols were designed for both ligand-based and structure-based screening workflows using the ROC curve method (Triballeau et al., 2005). These protocols were used to screen the 1.6 million-compound CAP database (version 2004, Accelrys Inc.) following a computational pre-filtering of the database. Details of the vHTS workflows are described in the Supplement.

Molecular modeling with active compounds

Selected compounds from the vHTS hit lists, as well as structural analogues, were subjected to computational docking in a molecular model of the receptor 5.24 ligand binding domain, as described previously (Bertrand et al., 2002; Luu et al., 2004).

Receptor activation assays

For cell-based receptor activation assays, receptor 5.24 was expressed in HEK 293 cells using a CMV expression vector (Luu et al., 2004; Speca et al., 1999). Receptor activation was assayed by measuring changes in intracellular calcium using a Flexstation II fluorescence plate reader (Molecular Devices). Details of these procedures can be found in the Supplement.

Electro-olfactogram recordings

EOG recordings on goldfish olfactory epithelium were carried out essentially as described previously (Irvine and Sorensen, 1993). Odors were applied for 5 seconds with a minimum inter-stimulus interval of 3 minutes to allow the olfactory epithelium to recover full responsiveness. For each fish, a compound was tested 2–3 times and an average response was computed. The per-fish averaged responses were then normalized to the response elicited by a reference standard odorant, 10−5 M L-arginine. Responses to the reference standard were measured at least every 45 minutes and bracketed the trials with each test odorant. Normalized responses from 3–4 fish were then averaged, resulting in a mean normalized response for each compound/concentration tested. All stimuli were tested on every fish in this study.

Supplementary Material

01

Acknowledgments

This study was supported by grants from the National Institute on Deafness and Other Communication Disorders (J.N.), and the Fédération pour la Recherche sur le Cerveau, the Agence Nationale pour la Recherche, and the Fondation de France Comité Parkinson (F.A.). N.T. was the recipient of a CIFRE grant (ref. 738/2002) provided by the Association Nationale pour la Recherche Technique and Accelrys. D.C was an LSSURP participant supported by an IGERT grant from the NSF (NSF DGE-0653827). We thank Maria Waldhoer and Jennifer Whistler for their invaluable and generous assistance with the Flexstation assays, Haude Levesque for assistance with the EOG recordings, and Delphine Rigault for technical support with the commercial hits.

Footnotes

Publisher's Disclaimer: This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final citable form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.

References

  • Abaffy T, Malhotra A, Luetje CW. The molecular basis for ligand specificity in a mouse olfactory receptor: a network of functionally important residues. J Biol Chem. 2007;282:1216–1224. [PubMed]
  • Alioto TS, Ngai J. The repertoire of olfactory C family G protein-coupled receptors in zebrafish: candidate chemosensory receptors for amino acids. BMC Genomics. 2006;7:309. [PMC free article] [PubMed]
  • Augen J. The evolving role of information technology in the drug discovery process. Drug Discov Today. 2002;7:315–323. [PubMed]
  • Bertrand HO, Bessis AS, Pin JP, Acher FC. Common and selective molecular determinants involved in metabotopic glutamate receptor agonist activity. J Med Chem. 2002;45:3171–3183. [PubMed]
  • Bessis AS, Rondard P, Gaven F, Brabet I, Triballeau N, Prezeau L, Acher F, Pin JP. Closure of the Venus Flytrap module of mGlu8 receptor and the activation process: insights from mutations converting antagonists into agonists. Proc Natl Acad Sci USA. 2002;99:11,097–11,102. [PMC free article] [PubMed]
  • Born TL, Blanchard JS. Structure/function studies on enzymes in the diaminopimelate pathway of bacterial cell wall biosynthesis. Curr Opin Chem Biol. 1999;3:607–613. [PubMed]
  • Byrd RPJ, Caprio J. Comparison of olfactory receptor (EOG) and bulbar (EEG) responses to amino acids in the catfish, Ictalurus punctatus. Brain Res. 1982;249:73–80. [PubMed]
  • Firestein S. How the olfactory system makes sense of scents. Nature. 2001;413:211–218. [PubMed]
  • Hara TJ. Olfaction and gustation in fish: an overview. Acta Physiol Scand. 1994;152:207–217. [PubMed]
  • Irvine IAS, Sorensen PW. Acute olfactory sensitivity of wild common carp, Cyprinus carpio, to goldfish hormonal sex pheromones is influenced by gonadal maturity. Can J Zool. 1993;71:2199–2210.
  • Jorgensen NOG, Stepanaukas R, Pedersen AGU, Hansen M, Nybroe O. Occurrence and degradation of peptidoglycan in aquatic environments. FEMS Microbiol Ecol. 2003;46:269–280. [PubMed]
  • Katada S, Hirokawa T, Oka Y, Suwa M, Touhara K. Structural basis for a broad but selective ligand spectrum of a mouse olfactory receptor: mapping the odorant-binding site. J Neurosci. 2005;25:1806–1815. [PubMed]
  • Liberles SD, Buck LB. A second class of chemosensory receptors in the olfactory epithelium. Nature. 2006 this issue. [PubMed]
  • Luu P, Acher F, Bertrand HO, Fan J, Ngai J. Molecular determinants of ligand selectivity in a vertebrate odorant receptor. J Neurosci. 2004;24:10128–10137. [PubMed]
  • Mombaerts P. Genes and ligands for odorant, vomeronasal and taste receptors. Nat Rev Neurosci. 2004;5:263–278. [PubMed]
  • Ottoson D. Analysis of the electrical activity of the olfactory epithelium. Acta Physiol Scand. 1956;35(suppl 122):121–183. [PubMed]
  • Paulini R, Muller K, Diederich F. Orthogonal multipolar interactions in structural chemistry and biology. Angew Chem Int Ed Engl. 2005;44:1788–1805. [PubMed]
  • Pin JP, De Colle C, Bessis AS, Acher F. New perspectives for the development of selective metabotropic glutamate receptor ligands. Eur J Pharmacol. 1999;375:277–294. [PubMed]
  • Pin JP, Galvez T, Prezeau L. Evolution, structure and activation mechanism of family 3/C G-protein coupled receptors. Pharmacol Ther. 2003;98:325–354. [PubMed]
  • Rolen SH, Sorensen PW, Mattson D, Caprio J. Polyamines as olfactory stimuli in the goldfish Carassius auratus. J Exp Biol. 2003;206:1683–1696. [PubMed]
  • Schmiedeberg K, Shirokova E, Weber HP, Schilling B, Meyerhof W, Krautwurst D. Structural determinants of odorant recognition by the human olfactory receptors OR1A1 and OR1A2. J Struct Biol. 2007;159:400–412. [PubMed]
  • Scott JW, Scott-Johnson PE. The electroolfactogram: a review of its history and uses. Microsc Res Tech. 2002;58:152–160. [PubMed]
  • Shi P, Zhang J. Comparative genomic analysis identifies an evolutionary shift of vomeronasal receptor gene repertoires in the vertebrate transition from water to land. Genome Res. 2007;17:166–174. [PMC free article] [PubMed]
  • Speca DJ, Lin DM, Sorensen PW, Isacoff EY, Ngai J, Dittman AH. Functional identification of a goldfish odorant receptor. Neuron. 1999;23:487–498. [PubMed]
  • Triballeau N, Acher F, Brabet I, Pin JP, Bertrand HO. Virtual screening workflow development guided by the “receiver operating characteristic” curve approach. Application to high-throughput docking on metabotropic glutamate receptor subtype 4. J Med Chem. 2005;48:2534–2547. [PubMed]
  • Triballeau N, Bertrand H-O, Acher F. Pharmacophores and Pharmacophore Searches. Wiley-VCH; Weinheim (Germany): 2006. Are you sure you have a good model?
  • Work E. A new naturally occurring amino acid. Nature. 1950;165:74–75. [PubMed]
  • Yang H, Shi P, Zhang YP, Zhang J. Composition and evolution of the V2r vomeronasal receptor gene repertoire in mice and rats. Genomics. 2005;86:306–315. [PubMed]
  • Zhang X, Rodriguez I, Mombaerts P, Firestein S. Odorant and vomeronasal receptor genes in two mouse genome assemblies. Genomics. 2004;83:802–811. [PubMed]
PubReader format: click here to try

Formats:

Related citations in PubMed

See reviews...See all...

Cited by other articles in PMC

See all...

Links

Recent Activity

Your browsing activity is empty.

Activity recording is turned off.

Turn recording back on

See more...