Analytical and Bioanalytical Chemistry

, Volume 377, Issue 3, pp 427–433

Olfactory receptors: molecular basis for recognition and discrimination of odors


    • Institute of PhysiologyUniversity of Hohenheim

DOI: 10.1007/s00216-003-2113-9

Cite this article as:
Breer, H. Anal Bioanal Chem (2003) 377: 427. doi:10.1007/s00216-003-2113-9


The daunting task of our nose to detect and discriminate among thousands of low-molecular-weight organic compounds with diverse chemical structures and properties requires an enormous molecular recognition capacity. This is based on distinct proteins, capable of recognizing and binding odorous compounds, including odorant-binding proteins, which are supposed to shuttle odorous compounds through the nasal mucus, and most notably the odorant receptors, which are heptahelical membrane proteins coupling via G-proteins onto intracellular transduction cascades. From more than a thousand genes each olfactory neuron is supposed to express only one receptor subtype. Receptors appear to be selective but rather non-specific—i.e. a distinct odorant activates multiple receptors and individual receptors respond to multiple odorants. It is the molecular receptive range of its receptor type which determines the reaction spectrum of a sensory neuron. Populations of cells equipped with the same receptor type project their axons to common glomeruli, thereby transmitting the molecular receptive range of a receptor type into the receptive field of glomerulus. Recent insight into the molecular basis of odor recognition and the combinatorial coding principles of the olfactory system may provide some clues for the design and development of technical sensors, electronic noses. In this review more emphasis has been placed on physiological rather than analytical aspects.


Odorant receptorsBinding-proteinsGene familySensory neuronsProjectionPlfactory bulb


The chemical senses are responsible for the perception of signaling molecules, such as odorants, tastants and pheromones; they create a representation of the chemical world of our ambience. The sense of smell is specialized to detect small airborne molecules; volatile, largely hydrophobic compounds can be perceived at low concentrations, some at concentrations as low as a few parts per trillion: moreover, myriads of distinct compounds can be discriminated. Thus, the nose can be considered as an efficient chemodetector exposed to the outside world which is capable of permanently monitoring the ever changing chemical composition of our adjacency, which is perceived as distinct smells. The processes of odor perception begin when volatile chemicals, typically small organic molecules carried by the respiratory air stream, reach the nasal neuroepithelium where millions of distinct olfactory sensory cells reside. Each of these sensory cells is a bipolar neuron which projects an axon to the olfactory bulb and extends a dendritic process to the nasal lumen. The tip of the apical dendrite carries 5–10 immotile cilia, the actual chemosensory structures; they are embedded in the protecting nasal mucus. Olfactory cilia are considered as a large expansion of the surface area that can interact with odorous molecules and are the sites of primary olfactory processes. The interaction of odorous molecules with the chemosensory membrane elicits a cascade of transduction events that ultimately lead to an increase in membrane conductance; the resulting generator potential is converted to a distinct frequency of action potentials which are conveyed to the olfactory bulb. Thus, the strength, duration, and quality of odorant stimuli are encoded into patterns of neuronal signals. Since olfactory neurons differ in their responsiveness to odorants and because different odorants stimulate only a certain population of olfactory cells, a given odorant elicits a distinct pattern of neuronal activity. From the olfactory bulb odor signals are conveyed to the primary olfactory cortex; from there olfactory information is relayed both to higher cortical areas, which enable the conscious perception of odors, and to the limbic system, which generates their emotional and motivational effects. Thus, odor perception is the brain's interpretation of neuronal activity patterns, which are elicited due to a specific interplay of odorous molecules with distinct sensory cells, i.e. olfaction is primarily based on molecular recognition processes [1, 2, 3].

Perireceptor events

Odorant-binding proteins

Terrestrial animals smell volatile, hydrophobic molecules; these airborne compounds must traverse the aqueous milieu of the mucus layer covering the nasal epithelium before contacting the recognition sites on the olfactory cilia. The entry, exit and residence time of lipophilic odorants in the receptor environment are considered an important part of the chemical sensing process, although the mechanisms for these "perireceptor events" are still poorly understood. The discovery of small, water-soluble proteins in the mucus fluid surrounding the sensory dendrite and cilia, which are produced by glands of the nasal cavity, has led to the concept that these odorant-binding proteins (OBP) may accommodate hydrophobic molecules in an aqueous environment and enhance their access to the receptor sites. Molecular cloning approaches have revealed that OBP belong to a family of proteins, called lipocalins, which serve as carriers for small lipohilic molecules in other body fluids. Structure analysis revealed that the polypeptide chain is folded into eight antiparallel β-sheets with a α-helical domain located near the carboxy terminal. The β-sheets form a continuously hydrogen-bonded β-barrel, with the internal cavity accommodating the hydrophobic ligands (Fig. 1). Several distinct OBP-subtypes have been identified and each OBP-subtype appears to have a unique ligand-binding profile [4] suggesting a more specific role of these proteins than acting as a general solubilizer or unspecific carrier for all hydrophobic compounds. Based on the diversity of OBP, it seems conceivable that they are involved in preselecting those volatile compounds that finally interact with the olfactory sensory cells, i.e. OBP may play a role as a specific filter rather than a passive shuttle protein for odorants in the mucus layer.
Fig. 1.

Schematic representation of an odorant-binding protein. The globular protein is constituted of eight β-sheets and a helical domain at the C-terminus. The polypeptide chain forms a β-barrel structure; the resulting cavity comprises the binding site for small hydrophobic compounds

The occurrence of binding proteins in chemosensory systems of terrestrial vertebrates and insects implicates that OBP may be a molecular adaptation to terrestrial life. The observation that OBP from vertebrates and insects show neither sequence homology nor structural similarity favors the view that both animal groups evolved binding proteins for odorants independently, i.e. that OBP in vertebrates and insects represent an evolutionary convergence.

Biotransformation enzymes

Much olfactory-related behavior is based on monitoring the chemical environment through consecutive sniffing. It is therefore imperative that the odorants received by a given sniff will rapidly extinguished, thus allowing to perceive a "fresh" wave of odorants. So, the rapid inactivation and clearance of odorous molecules is a prerequisite for the capability of the olfactory system to receive iterative incoming signals. Odorous compounds appear to be inactivated by biotransformation or detoxification enzymes. The reaction of phase I enzymes, e.g. cytochrome P-450 mono-oxygenases, which introduce chemical changes such as hydroxylation, is followed by phase II enzymes, such as UDP-glucuronosyl transferase or glutathione-S-transferase catalyzing the conjugation of glucuronic acid or of glutathione to phase I-modified odorants. Odorous molecules modified by this sequential biotransformation are no longer lipid soluble and are incapable of receptor stimulation; finally mucus flow and blood flow are supposed to clear the modified odorants effectively (Fig. 2).
Fig. 2.

Perireceptor events in the mucus layer of the olfactory epithelium influence the entry, exit, or residence time of odorous molecules in the receptor environment. These ancillary processes are integral components of the chemical sensing systems and include the interaction of odorants with specific soluble binding proteins which may act as shuttles for the volatile, lipophilic odorous molecules through the aqueous mucus layer and lead to inactivation of odorants by degrading and/or biotransformation enzymes, thus clearing the system between consecutive sniffs

Olfactory receptors

Receptor-encoding multigene family

The mechanisms by which thousands of different odorants that vary widely in structure are readily detected and discriminated has been a longstanding puzzle. The accuracy of odor discrimination depends on the specificity with which odorants interact with appropriate olfactory neurons via specific receptors in the plasma membrane. Thus, olfactory receptor proteins are considered as molecular entities at the interface between the environment and the nervous system. Understanding the nature, diversity and specificity of odorant receptors is therefore considered as a key for understanding the molecular basis of olfaction.

The discovery of a large family of genes which encode heptahelical transmembrane proteins and are expressed in the olfactory epithelium [5] was the ground-breaking work which opened new avenues of research for better understanding of odorant recognition. The candidate odorant receptors (OR) contain all the characteristic hallmarks of G-protein-coupled receptors (GPCR), including seven hydrophobic transmembrane domains; they are classical G-protein-coupled receptors belonging to class A GPCR, which also includes opsins and catecholamine receptors [6]. The genes encoding OR are devoid of introns within their coding region and are typically organized in clusters of ten or more members, located on all but a few chromosomes. Sequence database mining determined in mice a repertoire of about 1300 putative OR genes, including a few hundred pseudogenes [7]. This enormous number indicates that apparently more genes are devoted to perceiving odors than to any other single purpose, thus underscoring the importance of smell to most mammals. In the human genome about 900 OR genes were identified, but two-thirds of these turned out to be non-functional or "pseudogenes", which have lost their function during evolution; a total of 347 putative functional OR genes in man was determined [8]. In spite of the reduced number of OR-subtypes the human olfactory system retained the ability to recognize a broad spectrum of chemicals, however, it is probably less discriminating than in mice. The high proportion of pseudogenes [9, 10] and the unusually high rate of single nucleotide polymorphisms (SNP) in human receptor genes [11] indicate a variable repertoire of functional OR genes in the human population. Many specific anosmia, i.e. the inability to smell particular odors, could be due to hereditary defects of OR genes.

Structural features of olfactory receptors

Certain sequence features appear to be unique to odorant receptors. Conserved sequence motifs are apparent in intracellular loops, most obvious at the junction of transmembrane domain TM3 and intracellular loop IC2; this region includes the DRY-motif which is conserved in all GPCR. Other conserved regions include TM1, TM2 and TM7. Structural diversity is most notably in the central transmembrane domains and it has been proposed that about 20 variable amino acid residues in TM3, TM4, and TM5 may constitute the binding pocket, "the ligand complementary-determining region" [12]. However, in view of the structural feature of the EC-loops and binding sites in other GPCR, the TM-regions may not be the only determinants for the binding specificity of olfactory receptors. Amino termini as well as carboxy termini are short (about 20 amino acids). The N-termini comprise consensus sequences for N-glycosylation sites as in most other GPCR. The C-termini of most OR contain putative phosphorylation sites, which are also found in the third IC-loop, they seem to be phosphorylated by second messenger-activated kinases and receptor kinases, thus uncoupling the transduction cascade. A putative palmitylation site, which allows the formation of a "fourth cytoplasmic loop" and appeared to be typical for class A GPCR is lacking in most of the OR (Fig. 3). As other GPCR, the OR comprise highly conserved cysteine residues in EC1 and EC2, which may contribute to intra- or inter-molecular disulfide bonds; however, as a unique feature shared only by few other GPCR, OR sequences contain three cysteines in the EC2-loop. A very recent study suggests a particular function for one of the cysteines in EC2. It was proposed that the sequence motif HXXC[DE] which is present in the EC2-loop of about three quarters of all OR, constitutes a metal binding site and that a complex with Zn(II) or Cu(II) will turn the EC2-loop into an α-helical structure. In this view chemicals with high affinity to the metal ions could replace one of the metal-ligated amino acid residues, resulting in a structural rearrangement of the receptor protein, a prerequisite for an activation of the cytoplasmic G-proteins [13]. Interestingly, certain chemicals which are very good ligands for metal ion coordination complexes, such as amines or thiols, are particularly strong smelling odorants.
Fig. 3.

Proposed membrane topology of an odorant receptor. The seven transmembrane domains (TMD) are supposed to form a "funnel" which accommodates the odorous ligands. The specificity of the binding site is determined by amino acid side-chains of the central TMD

Evolution of OR gene families

Since the genome projects has provided the sequences for numerous OR genes from a variety of vertebrate and invertebrate species, including mammals, birds, amphibia, teleosts, cyclostomes, nematodes, and insects, detailed comparisons across species borders are now possible. The most obvious feature that emerged upon sequence comparison is the enormous diversity of OR genes both within and across species. The small genome of C. elegans appears to contain several hundred OR genes, none of them share significant sequence identity with vertebrate OR genes [14]. This seems also true for insects; OR genes from fly, mosquito, and moth have been studied [15, 16, 17] indicating that the repertoire appears to be below 100 and the sequences are diverse, sharing no identity with vertebrate or nematode OR genes. An extreme sequence diversity within and across the species seems to be a characteristic feature of insect OR genes. Unlike mammalian OR genes most of the insect OR cannot be grouped into families and subfamilies based on sequence similarities. However, there is an exception; one receptor type has been identified that is conserved across a variety of insect species [18].

Interestingly, OR genes of the invertebrate species all comprise introns; in contrast to the intronless vertebrate OR genes.

In a search for the phylogenetic origin of the vertebrate OR genes it was found that lampreys, one of the most ancient surviving vertebrate species, already contains a family of OR genes with all of the sequence motifs typical of higher vertebrate OR; however, phylogenetic analysis suggest that the lamprey receptors may not represent the ancestral form of modern, vertebrate OR or any intermediate form between invertebrate OR and fish OR [19, 20].

In lower vertebrates, such as fish, the repertoire of receptor genes ranges between 50 and 100; based on sequence divergence fish receptors form a separate group, non-overlapping with typical rodent receptor families [21]. It seems plausible that the structural and numerical differences of the receptor gene repertoire between fish and mammals reflects the phylogenetic distance; however, it is also conceivable that it may be the result of adaptive processes, allowing the fish to deal with a limited number of water-soluble molecules and mammals to detect a large variety of hydrophobic, volatile compounds. The identification of both fish-like (class I) and mammalian-like (class II) receptor genes in the amphibian Xenopus laevis [22], expressed in two different compartments of the frog's nose, which are employed in the selective recognition of either water-soluble or airborne odorants supports the notion that class I receptors may be specialized for detecting water-soluble odorants and class II receptors for recognizing volatile odorants. This view is supported by functional analysis of class I and class II receptors heterologously expressed in Xenopus oocytes [23].

Overall, the available data suggest that throughout evolution, genes encoding olfactory receptors, i.e. receptors of chemosensory neurons, emerged several times in an independent manner, at least until the appearance of fish OR. Moreover, it becomes evident that OR gene-families in ancient vertebrates and invertebrates appear to be relatively small in size, however, they exhibit an enormous diversity of sequences; in contrast, mammals have a largely expanded number of genes encoding receptors and based on sequence similarities, certain receptor genes can be categorized in families and subfamilies.

Control of OR genes expression

The large number of OR genes and their organization in multiple gene clusters poses a great challenge to the mechanisms controlling gene activity. Expression of OR is restricted to mature olfactory neurons and each cell seems to express only a single OR gene. Moreover, expressing of a given receptor derives exclusively from either the maternal or the paternal allele involving mechanisms of allelic inactivation [24, 25]. It has been proposed that the choice of a given OR gene from an array of genes in a genomic cluster could either be accomplished by DNA-rearrangements or by an enhancer which can only activate one gene in its cluster. Sequence inspection of OR gene loci have identified a number of more or less conserved motifs located upstream of the putative transcription start sites [26, 27, 28]. The functional implications of these motifs in gene control mechanisms are still elusive, however a recent study has demonstrated that the 395 bp sequence region upstream of the predicted transcriptional start site is essential for gene expression [29]; interestingly, several of the conserved motifs are located in this 5′-region.

Ligand specificity of olfactory receptors

Physiological single and multi-unit recordings have demonstrated that individual olfactory sensory neurons typically respond to a variety of different odorants and that each cell shows a unique order of agonist potency, indicating that olfactory neurons are highly diverse and broadly tuned. Based on the notion that each olfactory sensory cell expresses only one OR subtype, it seems likely that a relatively non-specific ligand spectrum is a characteristic feature of olfactory receptors. As with other orphan receptors, elucidation of the ligand/receptor specificity requires assessment of the responsiveness of a distinct receptor type; this is usually performed by expression in heterologous cell types and high-throughput screening assays. OR genes have been proven to be exceptionally difficult to express in heterologous systems, presumably due to the failure of proper folding or membrane targeting of the receptor proteins. However, some of these problems were circumvented by approaches using either a homologous in vivo expression system transfected by means of recombinant adenovirus and assessed by electrophysiological recordings [30] or using engineered OR chimeric receptors with the N-terminal "membrane-import-sequence" of either rhodopsin or serotonin receptors in heterologous cells monitored by imaging approaches [31, 32, 33]. The general consensus from these studies is that a distinct OR is activated by multiple odorants and that the range of agonist potency of dissimilar ligands for a distinct OR subtype, the molecular receptive range [34], resembles that of individual olfactory sensory neurons. This view was confirmed and extended in studies monitoring the responsiveness of isolated individual olfactory neurons and determining their receptor type by single cell RT-PCR [35, 36]; the results led to the conclusion that a single receptor can recognize multiple odorants and a single odorant is typically recognized by multiple receptors. Thus, all data indicate that the nose uses a combinatorial coding scheme to discriminate the vast number of different smells. Analogous to the visual system which uses three receptor types (three opsin-subtypes of the three cone populations) to make sense of all perceivable colors, the olfactory system computes information from combinations involving any of about a thousand receptor types. The numerous possible combinations explain the capacity of the system to encode an unlimited number of odors. Instead of dedicating an individual odor receptor to a specific odor, the olfactory system uses an "alphabet" of receptors to create a specific odor response; in this view, a distinct receptor type participates in encoding very different odors much the same way as a distinct letter participates in forming very different words.

The principle of combinatorial coding implies that odorants of nearly identical structure are recognized by different but overlapping sets of receptors, thus explaining why even a slight change in the structure of an odorant can cause a dramatic shift in its perceived odor. For example, when the hydroxyl group of octanol is replaced by a carboxy group to make octanoic acid, its odor changes from orange to rancid. It may also explain the phenomenon that the perceived quality of an odorant can differ with a change in its concentration; for example, indole has a putrid smell when concentrated but is perceived as floral when diluted.

Spatial patterns of OR-expression

The notion that a given sensory neuron expresses only one type of receptor implies that the supposedly about 20 million olfactory neurons of a mouse can be subdivided into about a thousand subpopulations. Within the olfactory epithelium, subpopulations of cells, expressing the same receptor type, i.e. groups of 10,000–20,000 olfactory sensory neurons, are confined to one of several broad expression zones. Within a zone, neurons of a subpopulation are broadly distributed and surrounded by cells expressing different receptor types [37, 38]. The functional implications of this zonal segregation are still elusive, but it is maintained in the olfactory bulb with each zone of the epithelium projecting to a distinct region of the olfactory bulb, suggesting that the subpopulation of olfactory neurons which expresses the same receptor and the subpopulations which express related receptors project to the same region of the bulb. Besides the principle of zonal distribution, realized for most subpopulations, a relatively small number of neuron populations which express receptor subtypes of the OR37 subfamily are segregated in a small area on the tip of central turbinates; these cell populations are all spatially organized in clusters [39], moreover, OR37 subpopulations project to distinct glomeruli which are grouped together in a small circumscribed area of the bulb [40] (Fig. 4).
Fig. 4.

Distribution and projection patterns of olfactory neurons. Olfactory sensory cells expressing a distinct receptor type are spatially segregated in an anterior-posterior zone of the olfactory epithelium (OE) and project their axons to common glomeruli (glom) in the olfactory bulb (OB)

Projection patterns of olfactory neurons

To inform the brain, each olfactory neuron projects a single unbranched axon to the olfactory bulb where the axon synapses onto the dendrites of mitral and tufted cells. Several thousand olfactory neurons synapse in each of the approx. 2000 glomeruli on the surface of the bulb. Studies using gene-targeted mouse lines which coexpress marker proteins, such as lacZ or GFP, together with a distinct receptor type, allowed the visualization of individual axons. By these approaches it was demonstrated that all neurons expressing the same receptor type converge their axons onto the same glomerulus; usually two glomeruli which are located on the lateral and medial hemisphere of the bulb, respectively [41]. Thus, olfactory neurons that express the same receptor type are widely scattered in a distinct zone of the nasal neuroepithelium but their axons converge at two specific sites in the olfactory bulb. Exploring the projection of two neuron populations which express highly related receptor genes, transgenic approaches with different markers for each populations were employed. It was found that neuron populations with very similar receptor types nevertheless project to distinct glomeruli; however, the glomeruli are located in immediate vicinity. The relative position of each glomerulus appears to be not stereotypic but rather displays some inter-individual variability [40].

The wiring processes and the interplay between axon populations in finding a common target area in the bulb but projecting onto distinct glomeruli is largely unknown. Examining the trajectory of axons during target finding as well as the timing of glomerulus formation for neuron populations which express highly related receptor subtypes and innervate neighboring glomeruli in the olfactory bulb revealed that axons project to largely overlapping areas within the developing olfactory bulb during the prenatal phases. Only during an early postnatal phase, the overlapping axonal-dendritic networks of several populations segregate into distinct glomeruli [42]. This is partly reminiscent of the situation in the visual system, where axons of retinal ganglion cells from both eyes initially invade overlapping areas in the lateral geniculate nucleus and are gradually retracted from inappropriate regions to form eye-specific domains [43].

High-tech noses; electronic sensing

Although electronic sensors to detect toxic gases have been used for years, the recent progress in understanding the mechanisms and principles underlying our sense of smell have intensified efforts to design devices capable of monitoring the ambient odorous world with areas of application ranging from registration of environmental hazards and explosives to assessment of food quality or even of exhaling. Electronic odor devices or "artificial noses" are based on sensors that detect volatile chemicals and respond by producing electrical signals that are passed on to an artificial computing system, which is programmed to interpret them. With a large number of different sensors arranged in an array—like the multiple detectors in the nose—a system can combine the varying signals producing a pattern that can be matched to a database of known scents—much like the brain does. Polymers, metal oxides, or fluorescent compounds which change their properties in response to specific volatile compounds have been used as sensors [44]. Attempts are currently made not only to mimic the principles of the biological nose but also to employ some of its molecular elements as functional entities in biosensors; e.g. isolated olfactory receptor proteins coated onto the surface of a piezoelectric crystal, which serves as a signal transducer, have been tried [45]. Whether an "e-nose" will ever match the capacity of its biological counterpart is unclear but electronic odor sensing has already become a valuable tool in quality control, classifying a collection of odors according to their similarity to a standard.


The enormous sensory capacity of the nose as a highly sensitive detector system recognizing and discriminating myriads of extraneous volatile compounds has long been considered as a scientific mystery; the application of modern biological techniques has, however, opened new avenues of research in olfaction. This has already led to the deciphering of some of the principles and mechanisms underlying the sense of smell. In particular, characterization of the detector elements, the receptors, has greatly contributed to unveiling some of the secrets of olfaction. Moreover, due to the availability of distinct receptors, the possibilities emerge for designing novel compounds which will activate defined sets of receptors and thus elicit desirable smells; conversely, it is also conceivable that compounds can be generated which will act as anti-scents, by blocking a defined set of receptors and thus prevent the perception of certain smells. In addition, molecular elements of the sensory cells may be produced by means of biotechnology, integrated into biosensors and thus operating as functional modules in artificial noses.

Copyright information

© Springer-Verlag 2003