I. Rationale for a Receptor Code

A generally accepted system for defining, characterizing, and classifying transmitter/hormone receptors is desirable for several reasons. First, pharmacologists need a common language so that ambiguity is not created by varied and inconsistent terminologies. It is estimated that a significant fraction of the mammalian genome codes for receptor proteins and subunits; this leads to the prediction of several thousands of individual signal-transducing receptors, even without consideration of the very large group of olfactory and gustatory receptors that exists (see Glusman et al., 1996). Thus, an internationally acceptable classification system for receptors is both desirable and necessary. Second, the development of a systematic scheme of classification is in itself an important generic research approach in biological science. It can bring focus, highlight relationships, and stimulate the recognition and investigation of features common to various classes and groups. It also can indicate deficiencies in our information and add an evolutionary perspective, which may bring its own insights. It remains to be determined why there is, in many cases, such a large variety of receptors for a given chemical messenger transmitter. We need to determine whether all these receptors are functional. It is also interesting to consider how they evolved and how many more endogenous ligands, known and unknown, are to be anticipated. Third, drug discovery benefits greatly from the systematic analysis of the growing body of information on receptors and their subclassification, with the knowledge that each receptor subtype provides a potential new drug target. Thus, it is now well recognized that the identification of well defined receptor subtypes often leads to highly specific drugs with new clinical indications and/or less undesirable side effects. It is reasonable to assume that future computer-based analyses or predictions of drug selectivities will benefit greatly from an established universal framework of a systematic receptor classification.

The term “receptor” is in use in various broad senses, but for pharmacological purposes the term is used specifically (Barnard, 1997;Humphrey, 1997) and refers to proteins which are “signal transducing receptors,” as defined and discussed by Kenakin et al.(1992). That is, each such protein specifically recognizes and is activated by a native agonist, which is one of numerous different messenger molecules that specifically relay a signal into a cell or between compartments within a cell. Thus, receptors in the pharmacological sense may be in one of three locations: (a) in the plasma membrane (the receptors for neurotransmitters, trophins, growth factors, and cytokines, other immunomediators, morphogens, sensory stimulants, and chemoattractants, and circulating hormones); (b) in an organelle membrane (e.g., those receptors where the transduction process involves the release of Ca²⁺ from an intracellular store); (c) in the cytosol. Case c can involve migration of the ligand-bound receptor to the cell nucleus, as for the receptors that belong to the superfamily of ligand-regulated transcription factors, whose ligands are steroid hormones and certain other fat-soluble hormones (structural class 4 here; see below). For all the natural signal-transducing agonists involved, the general term “transmitter” can be used.

As a consequence of the considerations listed above, much effort has been dedicated recently to the classification of the pharmacologically better-known neurotransmitter receptors. However, molecular biology has created a dramatic increase in information on the existence and structure of receptors, often preceding any data on their function, thereby reversing the traditional order of pharmacological data acquisition. The IUPHAR Committee on Receptor Nomenclature and Drug Classification (NC-IUPHAR), with its various Subcommittees,FNb has been constituted to review current approaches to receptor characterization and to define a generic scheme of nomenclature (Vanhoutte et al., 1994, 1996; Humphrey et al., 1994). Although progress has been made in standardizing trivial names in several receptor areas (and this initiative will continue), it is suggested that an all-embracing, rational system of coding receptors should be introduced [an alphanumeric receptor code (RC)^c], analogous to the EC (Enzyme Commission) enzyme codes introduced by the International Union of Biochemistry and Molecular Biology (IUBMB). It will differ from the EC system (which was set up before the era of molecular biology) in that the proposed codes are intended to convey structural information (relevant to the mode of receptor transduction) together with additional elements relating to operational characteristics, set out in a hierarchical order. This should provide a unique, unambiguous, and authoritative descriptor for each receptor protein. Each RC, assigned by NC-IUPHAR, is intended to be used in publications for definitive identification, in conjunction with a suitable (ideally, approved) trivial name. It also will be used in The IUPHARCompendium of Receptor Characterization and Classification(July 1998), together with the planned IUPHAR Receptor Database which will link uniquely nucleotide and protein databases to comprehensive data on receptor function and drug-related operational characteristics (recognition and transduction).

The details presented here constitute a proposal, sanctioned by NC-IUPHAR (see acknowledgment), which provides a basis for a broad and full debate within the pharmacological community at large and with scientists from other disciplines and their representative bodies.

II. Criteria for Receptor Characterization

It generally is acknowledged that studies toward the pharmacological characterization and classification of neurotransmitter/hormone/autacoid receptors should involve work on both function and structure (e.g., see Humphrey et al., 1993,1994; Vanhoutte et al., 1994; Hoyer et al., 1994;Humphrey, 1997). Receptor structure, in terms of its amino acid sequence, is unambiguous and will allow the allocation of a database code to the unequivocally identified protein. Function is equally important and not necessarily sufficiently predictable from structure, although it may be. Thus, one amino acid difference may make important differences in the drug recognition characteristics of a receptor (cf., the rat and human neurokinin NK₁ receptors) or by contrast significant differences in receptor sequence homology may make little such difference (e.g., in human 5-HT_1B and 5-HT_1D receptors or somatostatin receptors sst₁ and sst₄) (see Sachaiset al., 1993; Hoyer et al., 1994, 1995). It is therefore essential to establish which drugs are specific and selective for a particular receptor, either as agonists or antagonists, and to provide appropriate quantitative measurements of key parameters (seeKenakin et al., 1992; Jenkinson et al., 1995). The affinities (as dissociation equilibrium constants) of ligands often are measured from radioligand binding studies, and such data (at least for antagonists) should be equivalent to the corresponding data from functional studies. It has been argued that parameters from both types of studies clearly cannot be considered on semantic grounds under the umbrella term “functional” and hence the term “operational” has been introduced (Humphrey et al., 1994; Humphrey, 1997). Doubts about the use of “operational” in this context have been expressed because of its prior (and currently accepted) use in reference to agonism and agonist-specific parameters, which involve an efficacy component with more than just a binding (and recognition) parameter being involved (Kenakin et al., 1992; Black and Leff, 1983). However, this strengthens the argument for using the term “operation” which, importantly, implies an added involvement of receptor “activation” and hence transduction. The alternative term “recognition” more specifically refers to the binding characteristics or binding affinities of drugs (both agonist and antagonist) for a given receptor but does not necessarily encompass all aspects of receptor function, as the term “operational” undoubtedly does.

Transduction in the context of receptor characterization is intended to refer to the steps that allow the binding of an agonist at the receptor to be linked to the transmission of a signal into a cell or between compartments of a cell. This necessarily will involve specific changes in the receptor protein (if it is in the ion-channel class) or, in other cases, relayed to associated proteins which execute the primary signaling step. Transductional data should not involve the categorization of more downstream second-message cascades themselves, although such information might be used judiciously to infer events upstream. Even if more tightly defined, the use of transductional data to characterize receptors is controversial but it has been invaluable nevertheless in the classification of 5-HT receptors (Humphrey et al., 1993; Hoyer et al., 1994). Thus 5-HT_1C (now called 5-HT_2C) and 5-HT₂ (now called 5-HT_2A) receptors were predicted to belong to the same receptor group on the basis of shared transduction mechanisms. This was confirmed later when both receptor genes were cloned and the respective proteins were shown to share a high degree of homology (Hoyer et al., 1994). In contrast, although both 5-HT₃ and 5-HT₄ receptor types can be blocked by tropisetron (albeit at different concentrations), it was obvious on the basis of transduction that the two receptors were quite distinct even before both genes had been cloned (Humphrey et al., 1993; Hoyer et al., 1994). Thus, at the very least, consideration of the transduction mechanism involved will distinguish between a ligand-gated ion channel receptor and a G protein-coupled receptor (i.e., the 5-HT₃ and 5-HT₄ receptor, respectively). It should be noted that the structural classes proposed here reflect known fundamentally different transduction mechanisms for each. It follows that a knowledge of receptor transduction mechanisms from functional studies is important but how much value should be attached to such operational data specifically for receptor characterization purposes in isolation remains to be determined. However, there is a growing view that the unique intracellular face of each membrane-bound receptor protein will dictate preferred stoichiometric interactions with adjacent proteins, which will be characteristic of the receptor type involved. On the basis of these arguments, essential operational data for receptor characterization would include all drug recognition and drug interaction data, from both functional as well as radioligand binding studies, together with data on receptor transduction mechanisms directly related to receptor activation.

In summary, we propose that the pharmacological criteria for receptor characterization will depend on the integration of data from studies on both receptor structure and receptor operation. When such an integrated pharmacological profile is detailed sufficiently it will be possible to register an IUPHAR RC with confidence.

III. Proposal for a Systematic Receptor Code

A systematic numbering system for all known pharmacological receptors allows definitive labeling of a particular receptor by an international pharmacological authority (NC-IUPHAR), while providing valuable classified information about its characteristics. Each RC would provide a unique identifier for each receptor withinThe NC-IUPHAR Receptor Database, which contains extensive pharmacological information on receptor characterization and classification. This would be analogous to the EC numbering system used for enzymes. If necessary such codes could be changed when new information and understanding dictated that change was essential. A full alphanumeric code could provide a universally accessible record of numbering and subclassification of receptors which would reflect the state of current knowledge for the structure and characteristics of each receptor. The coding system would not replace trivial names, which usually would be used, but it would circumvent some of the problems associated with attempts to standardize such trivial names.

The proposed RC would consist of a set of divisions, separated from each other by full points; each division conveys a different category of information about the receptor, as summarized in table1. Thus, according to the system proposed, the rat 5-HT_1A receptor (as a G protein-coupled recombinant receptor for 5-HT, classified pharmacologically as the 1A subtype) would have a RC of 2.1.5HT.1A.RNO.00.00 and the human 5-HT₃ (as the first ligand-gated cation channel receptor subunit for 5-HT) would have an RC of 1.1.5HT.01.HSA.00.00.S. Other examples to illustrate the coding system are the rat muscarinic acetylcholine receptor (2.1.ACH.M1.RNO.00.00) and the human P2X₁ receptor subunit (1.4.NUCT.01.HSA.00.00.S). These RCs are based on the assignment of the coded information explained in this document. (See tables2-8.)

An RC will be reserved for the polypeptide product(s) of a single orthologous gene. Species variants will share a common RC but will be differentiated by a three-letter species code (see tables9 and10). A provisional RC can be assigned without cloning of the relevant gene providing it has been characterized robustly in whole tissues (e.g., histamineH₃ receptor). This may prove difficult in the brain, or in other tissues if multiple subtypes co-occur. Where this is so, the minimum requirement for a recombinant receptor to be accepted as a functional entity will be that robust operational (which must include transductional) data are provided for the heterologously expressed receptor and that its messenger ribonucleic acid, or the protein itself, is shown to occur in vivo. Recombinant receptors which have not been shown to be functional in whole tissues (e.g., 5-ht_5b) would be assigned only a provisional RC code number designated by a terminal upper case “P” (e.g., 2.1.5HT.05B.RNO.00.00.P).

For structural classes in which hetero-oligomeric receptors occur, RCs for individual subunits often will be represented instead of the entire multimeric receptor, because the composition of the latter is usually indeterminate at present. Subunit RCs will be indicated by a terminal upper case “S” [e.g., 1.4.NUCT.01.00.00.S in the case of the P2X₁ receptor subunit for adenosine triphosphate (ATP)]. Where the subunit composition of an endogenous heteromeric receptor becomes known, that receptor is represented by a unique RC of its own (indicated by a terminal upper case “M” to represent a “multimeric” receptor); its subunit composition would be indicated in the database by listing the component RCs (S), and their stoichiometry when definitely known. For many transmitter-gated ion-channel receptors in Class 1, where a subunit combinatorial system occurs, the exact stoichiometry is not known currently. However, the situation is less complex for hetero-oligomers occurring in Class 3 where there is usually a fixed composition of subunits for each receptor type.

IV. Conclusions

We have provided justification for a systematic method of coding receptors of all structural types. The RC system proposed is designed not just to be informative but also to provide a distinct alphanumeric descriptor for each receptor protein. It is intended that each individual RC will not only define each receptor type unambiguously by way of a simple reference for publication purposes, but that it will also associate automatically with a large body of information in an authoritative pharmacological database. This database would supply an urgent need of pharmacologists and other scientists interested in the correlation and integration of data on receptor operation with that on receptor structure. The IUPHAR Receptor Databasewill link not only with existing databases already established for gene nucleotide sequences and amino acid sequences of receptor proteins, but will also uniquely provide detailed pharmacological data on the characteristics of receptor recognition and transduction. The latter data will be approved by international panels of experts in each of the many existing and future NC-IUPHAR subcommittees established for the different receptor families (see Vanhoutte et al., 1994).