Signaling Receptome: A Genomic and Evolutionary Perspective of Plasma Membrane Receptors Involved in Signal Transduction

See allHide authors and affiliations

Science's STKE  17 Jun 2003:
Vol. 2003, Issue 187, pp. re9
DOI: 10.1126/stke.2003.187.re9


Intercellular communication in multicellular organisms requires the relay of extracellular signals by cell surface proteins to the interiors of cells. The availability of genome sequences from humans and several model organisms has facilitated the identification of several human plasma membrane receptor families and allowed the analysis of their phylogeny. This review provides a global categorization of most known signal transduction-associated receptors as enzymes, recruiters, and latent transcription factors. The evolution of known families of human plasma membrane signaling receptors was traced in current literature and validated by sequence relatedness. This global analysis reveals themes that recur during receptor evolution and allows the formulation of hypotheses for the origins of receptors. The human receptor families involved in signaling (with the exception of channels) are presented in the Human Plasma Membrane Receptome database.


One of the major transitions from uni- to multicellular organisms was the evolution of unique cell-surface molecules that are essential for interactions between neighboring cells (1). Most classes of plasma membrane receptors relay extracellular signals to the interiors of the cells and allow the recognition of adjacent cells and extracellular structures. The resulting unique differentiation and growth of individual cells facilitates a "division of labor" among the cluster of cells. Most of these receptor families evolved with the advent of multicellularity and the need for coordinated cell behavior. However, a few can be traced to unicellular eukaryotes that form cell colonies. Because multicellularity has independently evolved more than 10 times (1), cell-surface receptors are unique among animals, plants, and fungi, despite the sharing of some protein domains.

Although Paul Ehrlich's original definition of receptors in 1913 deals primarily with small drugs that bind to cellular components, both nuclear and plasma membrane receptors now have been shown to interact with hormones and humoral factors essential for interorgan communication. Biochemical studies of hormones and cell-cell interactions in vertebrates yielded the terms endocrine, neuroendocrine, paracrine, juxtacrine, and autocrine, all of which involve signal transduction mediated by receptors and initiated by cognate ligands. In addition, genetic analysis using model organisms such as the fruit fly and the nematode has provided insight into multiple signaling systems that are essential for body patterning and embryonic development (2).

Recent completion of the genome sequencing of several organisms allows a genomic analysis of plasma membrane receptors and the phylogenetic tracing of receptor families. This evolutionary genomic approach provides a new understanding of extracellular ligand-mediated signaling and allows the grouping of human receptor proteins involved in signal transduction based on their phylogenetic origins. The sequence-based approach has led to the discovery of a growing number of homologous genes with sequence relatedness. Homologous genes within the same organism are termed paralogs; they are the result of gene duplication and could have similar functions. Homologous genes derived as the result of speciation are termed orthologs. Evolutionary genomic analyses allowed the identification of orthologous receptors in human and other species, as well as their related paralogs (3).

Completion of the Human Genome Project provides the opportunity to study different receptor-ligand families that have coevolved over hundreds of millions of years. The entire repertoire of genes that encode plasma membrane receptors can be called the "receptome." We have limited our analysis to the "signaling receptome," that is, those receptors implicated in cellular signal transduction. We have not included ion channel receptors in this analysis, because we are considering only those receptors that mediate a signal across the membrane without opening the membrane. Channels are also plasma membrane proteins that participate in signaling cascades. However, their mechanism of action is quite different than the receptors discussed here. Based on their evolutionary relationships and common mechanisms of action, we have outlined the major families of genes in the human signaling receptome (Table 1). It is not our intention to present a comprehensive picture of the signaling pathways for all receptor families. Instead, we emphasize the common mechanisms by which they operate. We discuss the evolutionary origin of each receptor family and their unique properties in interactions with ligands, co-receptors, and immediate downstream molecules. The classification of receptor families is based on published literature and GenBank searches. The present list is not exhaustive and omits families with few members or limited functional analysis.

Despite diverse endocrine, neuroendocrine, and paracrine modes of action, a genomic view of plasma membrane signaling molecules unifies the large number of receptors into distinct families. Numerous extracellular ligands interact with different groups of plasma membrane receptors with enzymatic, recruiting, transporter-like, or nuclear translocation properties. Because of the combinatorial use of co-receptors and downstream effectors, many receptors can activate overlapping sets of recruiters, enzymes, and transcription factors.

Analyses of the completely sequenced genomes of key model organisms allow the tracing of the phylogenetic origins of each human receptor family in the evolutionary tree (Fig. 1). We retrieved known plasma membrane receptors in the human genome by following key word searches of current literature in PubMed (National Library of Medicine) and of GenBank. We complemented our initial results by including additional genes with unique domains (for example, receptor tyrosine kinase domains and seven-transmembrane domains) that we identified by analyzing the predicted human proteome with the Simple Modular Architecture Research Tool [SMART]. After sorting known receptors into families on the basis of sequence similarity, we determined consensus sequences of major receptor families and then exhaustively searched the human genome and validated their origins in model organisms. Using multiple alignment programs [Multalin and BlockMaker], we analyzed sequence relatedness of paralogous human receptors within each subfamily; we followed this analysis with manual curation to eliminate splice variants and duplicated entries. The detailed procedure and results of our analyses can be accessed through the Human Plasma Membrane Receptome database. More than 1000 receptor gene pages in the database can be searched by their family relationships and phylogenetic origins, as well as by key words and sequences. Each gene page provides links to sequence, literature, and expression databases. Less well characterized receptors, and those belonging to families with few members, are not discussed in the present review but are listed in the database.

Fig. 1.

Evolutionary origins of the human plasma membrane signaling receptor families. Receptor families are presented in the order of their presumed appearance during evolution. Shown are the organisms in which the entire genome has been sequenced. Metazoan-specific receptors are present in nematode, fly, and chordates. Chordate-specific receptors are present in the sea squirt and vertebrates. Vertebrate-specific receptors are present in fish and human. BYA, billion years ago.

Whereas the seven-transmembrane receptor family is present in all the genomes of currently sequenced eukaryotes (4), most human receptor families evolved in ancestral metazoans, and some families are chordate- and vertebrate-specific (Fig. 1). The absence of certain genes in a present-day model organism could be the result of gene loss, and the presence of some genes could result from novel inventions. However, our evolutionary tracing of entire receptor families in model organisms enhances the confidence of the present conclusions, because it is unlikely that all members of a given receptor family became extinct in a particular evolutionary lineage. Overall, the molecular phylogeny of gene families is consistent with the innovation of superfamilies before the eumetazoan-parazoan split, and with a major intrafamily expansion after the branching of arthropods and before the emergence of jawed fish (5). Ongoing sequencing of additional genomes will allow further refinement of the present evolutionary hypotheses.

Seven-Transmembrane (7TM) Receptors

The 7TM receptor proteins are of ancient origin, first emerging in unicellular organisms. They have seven discrete and highly predictable transmembrane domains consisting of hydrophic residues (6). The central role of the 7TM heterotrimeric guanine nucleotide-binding protein (G protein)-coupled receptors (GPCRs) in multicellular organisms is reflected by their divergent structures and functions (7). Ligand occupancy of GPCRs induces a conformational change in the receptor that recruits and activates different G proteins, which stimulate the generation of adenosine 3′,5′-monophosphate (cAMP), phosphoinositides, diacylglycerol, and other second messengers. These second messengers, in turn, trigger such events as activation of kinase cascades and phosphorylation of cytosolic factors and nuclear transcription factors (8). Activated GPCRs also recruit GPCR receptor kinases (GRKs) that phosphorylate the receptors themselves to facilitate termination of signaling or receptor turnover.

The extracellular stimuli that activate 7TM receptors include light, simple ions, odorants, nucleotides, lipids, steroids, modified amino acids, peptides, and glycoprotein hormones (Fig. 2). They are the only non-channel plasma membrane receptors that are activated by inorganic chemicals and physical stimuli. Although the dimerization of most 7TM receptors is not essential for their function (7), some receptors [for example, the γ-amino butyric acid (GABA) receptors] require heterodimerization of paralogs for their proper expression and function (9).

Fig. 2.

Seven-transmembrane (7TM) receptors. Class A: (a) Rhodopsin-type receptors for the detection of photons (arrow) that interact with the 7TM core. (b) Small molecules, amines, nucleotides, and lipids also bind the 7TM core. (c) Thrombin cleaves the ectodomain and exposes a cryptic region that binds to the 7TM core. (d) Glycoprotein hormones bind to the leucine-rich repeat-containing ectodomain and interact with certain outer loops. Class B: Secretin and other peptides interact with the 7TM core and the outer loops. Class C: Receptors for certain neurotransmitters and Ca2+ have a large ectodomain that engulfs the small ligand. Frizzled class: Wnt binding to Frizzled activates the protein Dishevelled. LRP6 is a co-receptor for Frizzled but has its cognate ligand Dickkopf (Dkk). Kremen is another receptor for Dkk that, upon binding to a ternary complex with LRP6-Dkk, induces endocytosis (curved arrow).

Dictyostelium discoideum can exist as either single-celled organisms or as a colony of social amoebas. In this eukaryote, folate-sensing and cAMP-sensing are mediated by two different 7TM receptors (10). This dichotomy may represent the earliest divergence between detecting ligands of foreign origin (folate) and ligands produced by the multicellular organism itself (cAMP).

Sequence similarities of 7TM receptors that stem from phylogenetic relatedness are confined largely to the transmembrane domains (11). The current classification of human 7TM receptors includes four defined classes (A, B, C, and Frizzled) and the olfactory receptor families ( (12). Each class of GPCRs shows unique sequence features in their transmembrane regions and cannot be traced to a single evolutionary origin. The yeast Saccharomyces cerevisiae contains only two unique classes of GPCR (pheromone and glucose receptors), whereas the metazoans have developed different subtypes of 7TM receptors (Table 2). When compared with invertebrates, major expansion of class A GPCRs is evident in vertebrates. In the olfactory receptor family, many odorant and gustatory GPCRs exist in nematodes and mammals (human, 400; mouse, 1200; and worm, 800), but fewer are present in teleosts and insects, likely representing adaptation to their unique environments and the acquisition of lineage-specific functions. An evolutionary genomic reevaluation of the GPCR superfamily could more fully reveal the structure-function relationships of these proteins. This approach could also facilitate the discovery of drugs for pharmacological intervention (13) and the search for ligands for a still large group of orphan 7TM receptors (14).

The Drosophila melanogaster 7TM receptor Frizzled is a member of the 7TM receptor superfamily that is modulated by interactions with additional plasma membrane receptors. Ligand activation of the Frizzled receptor by Wingless, a member of the Wnt family (15) culminates in the accumulation of β-catenin, which, in turn, modulates gene transcription. Signal transduction by Frizzled receptors requires the participation of a low density lipoprotein (LDL) receptor-related protein, LRP5 or LRP6 (16) (Fig. 2). LRP5 and LRP6 themselves bind a ligand named Dickkopf1 (Dkk1), which inhibits Wnt signaling. Dkk1 also binds two other transmembrane receptors (Kremen1 and Kremen2), which stimulates endocytosis of the LRPs, thereby further modulating the Wnt signaling (17). Although G protein coupling is not their main signaling mechanism, Frizzled receptors may interact with G proteins in some contexts (18).

Plasma Membrane Receptors Found in Metazoans

The evolutionary emergence of multiple plasma membrane receptors coincided with the need for intercellular communication in the multicellular metazoans. In addition to the expansion of the 7TM receptor family, metazoans developed receptors containing intracellular regions with unique enzymatic, recruiting, or nuclear translocation properties (Table 1).

Families of Receptors with Intrinsic Enzymatic Activity

Four types of enzymatic domains (tyrosine kinase, serine-threonine kinase, tyrosine phosphatase, and guanylyl cyclase) are found as integral intracellular parts of plasma membrane signaling receptors, thus defining receptor families (Fig. 3). These receptors are single-transmembrane proteins and most are activated after dimerization. The subdivisions (classes) within each family rely on the sequence of their extracellular domains (19). Most receptor tyrosine kinases (RTKs), receptor-like protein tyrosine phosphatases (RPTPs), and guanylyl cyclase-natriuretic peptide receptors (GC-NPRs) form homodimers, whereas the serine-threonine kinase receptors and the unique epidermal growth factor (EGF) receptor class of RTKs form heterodimers. Ligands activate the receptor kinases to phosphorylate key intracellular residues on the receptor itself (cis), its dimer partner (trans), and downstream proteins. In contrast, RPTPs, the enzymatic activity of which can be inactivated by ligand binding, dephosphorylate cellular proteins. Furthermore, ligand-bound NPRs convert guanosine triphosphate (GTP) to guanine 3′,5′-monophosphate (cGMP).

Fig. 3.

Metazoan receptors with enzyme activity. RTKs: Ephrins are membrane-anchored ligands (Type A, GPI anchored; Type B, transmembrane) to Eph receptors that participate in bidirectional signaling. Both the receptors and ligands have fibronectin homology domains. Following ligand binding, Eph receptors phosphorylate specific sites on their own intracellular region and allow association of docking proteins. EGF receptors have a cysteine-rich domain. Their ligands are cleaved from membrane-anchored precursors by enzymes that are either membrane-anchored or soluble. EGF receptor heterodimerization is stimulated by ligand binding. Serine-threonine kinase receptors: After ligand binding, TGF-β receptor RII phosphorylates RI. RI, in turn, phosphorylates various Smads (not shown). The ligand is bound to both RI and RII, whereas RIII is a facilitator molecule. There are different types of binding proteins; chordin competes for the ligand's receptor-binding site, whereas follistatin does not. Guanylyl cyclase receptors: Natriuretic peptides (NP) activate NPR to convert GTP to cGMP. Receptor homodimerization precedes ligand binding, and each dimer binds one ligand molecule. Decoy receptors bind ligand, but lack an enzymatic domain. RPTP: These receptors are active phosphatases; ligand binding likely suppresses phosphatase activity. MAM, named for the proteins in which it was originally identified, meprin, A5, PTPμ.

RTKs may have been vital in the establishment of the first metazoans. Although no RTKs are present in yeast or plants, they do exist in sponges (20). (Tyrosine phosphorylation does occur in yeast and plants, but not through the actions of transmembrane receptors.) Several orthologs for the five major classes of human RTKs [Eph receptors, EGF receptors, insulin receptors, vascular endothelial growth factor (VEGF) and platelet-derived growth factor (PDGF) receptors, and fibroblast growth factor (FGF) receptors] are found in Caenorhabditis elegans and D. melanogaster (21-23).

The ephrin-Eph receptor complex participates in bidirectional signaling as well as cell adhesion (Fig. 3). The ligands for the Eph receptors are membrane-bound ephrins (24) capable of mediating direct cell-cell interactions. Ephrin A ligands are attached to the plasma membrane by glycosylphosphatidylinositol (GPI) linkage, whereas ephrin B ligands are transmembrane proteins (25). Ligand binding activates Eph receptor phosphorylation. Concomitantly, ephrin B ligands are also tyrosine phosphorylated themselves in the intracellular domain, thus evoking signal transduction within ligand-containing cells (26).

Most ligands for EGF receptors are produced by proteolytic cleavage of membrane-anchored precursors (Fig. 3) and the ligand-processing enzymes may themselves be membrane anchored (27). Thus, EGF signaling is generally autocrine or paracrine. EGF receptors are activated by homo- or heterodimerization, and receptor dimerization may occur without ligand binding. Although self-association of many nonliganded EGF receptors is prevented by their extracellular domain, ligand binding removes this negative constraint (28, 29). EGF receptor signaling exemplifies diversification at three levels. First, paralogous ligands exhibit varying affinities to different receptor combinations; second, high- and low-affinity receptors for the same ligand may be activated independently or in combination; and third, receptor-specific coupling activates distinct signaling cascades (30). Although all receptors in this family signal through the tyrosine-kinase-dependent pathway, nuclear translocation of one member of this class has been reported (31, 32).

Receptors for insulin and insulin-like growth factor-1 (IGF-I) consist of two identical dimers, each processed from one precursor (33, 34). Unlike the "close-encounter" model of the ephrin-Eph interaction, the insulin system represents the opposite end of the RTK receptor-ligand relationship; insulin acts on tissues remote from its origin of secretion (endocrine signaling).

In the FGF receptor class, FGFR1 through FGFR4 are active receptors, whereas FGFR5 lacks an intracellular domain and serves as a decoy receptor by heterodimerization with another member of the FGF class (35, 36). Many gain-of-function mutations in these receptors account for different skeletal abnormalities in humans, as a result of ligand-independent receptor dimerization and kinase activation, or increased ligand affinity (37).

Although phosphorylation of proteins on tyrosine residues occurs in yeast, tyrosine phosphorylation in receptor-mediated signaling is unique to metazoans. Sponges possess RTKs, suggesting their emergence before the parazoan-eumetazoan split (20). In addition to the orthologs for each RTK class in C. elegans and D. melanogaster, some intracellular signal transduction proteins are also conserved (38). However, the multiplicity of splice variants and paralogs indicates that the role and functional complexity of this receptor family have undergone profound expansion in vertebrates.

Another group in metazoans is the serine-threonine kinase receptor family for the transforming growth factor (TGF)-β family of ligands. Signaling by TGF-β uses two receptors, types I and II (RI and RII, respectively). TGF-β binds directly to RII, which induces the formation of an activated complex of the TGF-βRII and RI (Fig. 3). Transphosphorylation of RI by ligand-occupied RII, in turn, recruits and phosphorylates intracellular substrates known as receptor (R)-Smads (39). A type III TGF-β receptor (RIII), betaglycan, lacks intrinsic catalytic activity but can form a complex with RI and RII to regulate ligand binding and signaling (40, 41). The TGF-β receptor family also includes receptors for bone morphogenetic proteins (BMPs). Unlike TGF-β ligands, the BMPs bind both type I and type II receptors simultaneously (42). At key developmental stages, multiple binding proteins (follistatin, chordin, noggin, gremlin, and DAN) (43), which have minimal sequence similarity to each other, act as functional antagonists or chaperones by sequestering ligands to inhibit signaling through the TGF-β receptor family (44).

Functional conservation of TGF-β receptors is evident with orthologs found in the sponge (45). Receptor complexes from species as distant from each other as human, C. elegans, and D. melanogaster, can heterodimerize and bind orthologous ligands with high affinity (46). The downstream TGF-β-signaling mechanism may have arisen after the separation of arthropods and nematodes (1.1 billion years ago) but before the separation of arthropods and vertebrates (950 million years ago). In this window of 150 million years, a complete signaling pathway evolved (47).

In humans, the guanylyl cyclase receptor family includes the three known natriuretic peptide receptors (NPRs A, B, C) (48) and residues in their intracellular guanylyl cyclase domain show similarity to adenylyl cyclases (Fig. 3) (49). Unliganded NPRs homo-oligomerize through several extra- and intracellular domains (50), and each receptor dimer binds one ligand molecule (51). Thus, activity is not controlled by ligand-induced dimerization, but rather through conformational changes associated with ligand binding. NPR-A and NPR-B are enzymatically active. NPR-C is presumed to be a decoy receptor, because it lacks the typical cytoplasmic domain (48). Membrane guanylyl cyclases have been found in nematode, as well as sea urchin (52).

In contrast to the kinase receptors, RPTPs dephosphorylate downstream proteins. Most RPTPs contain two intracellular catalytic domains that are arranged in tandem (Fig. 3). Receptor homodimerization inhibits the constitutive activity of RPTP-α (53, 54). The majority of RPTPs are orphan receptors, and only midkine and pleiotrophin have been suggested as ligands for receptor-like protein tyrosine phosphatase-zeta (RPTP-ζ) (55). Pleiotrophin appears to suppress the receptor's constitutive activity, thus increasing the levels of tyrosine phosphorylation of various intracellular proteins (56).

RPTPs regulate neuronal migration (57). Some RPTPs have extracellular domains similar to cell adhesion molecules, suggesting potential roles in cellular adhesion. Because of their presence in sponges, RPTPs probably originated before the parazoan-eumetazoan split (58).

Families of Receptors Serving as Recruiters

In addition to receptors with intracellular enzymatic activities, several metazoan receptors lacking intrinsic enzymatic activity recruit downstream signaling molecules upon ligand binding (Fig. 4), by facilitating interactions between membrane-bound receptors and cytoplasmic proteins. Indeed, the RTKs can also be considered part of this group, because upon tyrosine phosphorylation of the receptor, various proteins are recruited to the receptor. However, this group of receptors represents those without intrinsic catalytic activity. Upon activation of these receptors, intracellular proteins are recruited to the receptor and stimulated, resulting in regulation of gene expression and initiation of other cellular processes. Many recruiter receptors require co-receptors or receptor oligomerization to provide signaling specificity. The 7TM receptors discussed previously are one type of recruiter receptor; all other recruiter-type receptors are proteins with a single transmembrane domain.

Fig. 4.

Metazoan recruiter receptors. Toll and Toll-like receptors (TLRs) form heterodimers. Each receptor has an extracellular leucine-rich domain connected to a cysteine-rich domain and the intracellular Toll-interleukin homology region (TIR). CD14 is a GPI-linked molecule and MD2 is a secreted opsonizing protein that presents exogenous bacterial LPS to the mammalian TLR. LDL receptors consist of an extracellular region with multiple EGF domains and an intracellular hinge region essential for receptor clustering. Ligand binding to monomeric LDL receptors leads to receptor clustering and internalization through coated pits. Different complex ligands interact with LRPs to recruit Dab-1 and FE65 (not shown). Integrin receptors consist of α and β subunits. Clustering of receptors leads to the recruitment of talin, thereby resulting in changes in cytoskeleton organization. The Robo receptor binds Slit and recruits the kinase Abelson, which leads to the phosphorylation of enabled. Plexin receptors have unique sema plexin/semaphorin/integrin (PSI) domains with several immunoglobulin-like folds. The intracellular region of the plexin receptors contains a GTPase-activating domain. The semaphorin ligands may be transmembrane, GPI-anchored, or secreted proteins. The plexin receptors and their ligands both have the sema domain. GTPase is recruited by the activated plexin receptor.

Toll receptors are one family of recruiter receptors with roles in development and innate immunity. In D. melanogaster, one ligand for this receptor is the cystine knot-containing protein Spätzle. Toll-like receptors (TLRs) in humans constitute a conserved family of innate immune recognition receptors activated by bacterial constituents such as the lipopolysaccharides (LPS) and peptidoglycans (59). Heterodimerization between the TLRs themselves and with cell-specific co-receptors (such as MD2 and CD14) leads to complex levels of bacterial ligand recognition (Fig. 4) (60). Receptor activation recruits adaptors such as MyD88, TIRAP, and TICAM1 that, in turn, recruit and activate downstream proteins (61-63). In C. elegans, one Toll gene (tol-1) is not required for resistance of the nematode to infection, but functions in its nutritional consumption of bacteria by enabling avoidance of pathogenic species (64).

Another family of recruiter receptors is the LDL receptors. LDL receptor and LDL receptor-related proteins (LRPs) were initially regarded as endocytosis mediators, but they also recruit downstream molecules, such as Disabled-1 (Dab1) and FE65, that are implicated in development (Fig. 4) (65, 66). Members of the LDL receptor family have been identified in D. melanogaster and C. elegans (67). The cytoplasmic parts of paralogous or orthologous receptors have little sequence similarity with the exception of a short motif that mediates clustering of the LDL receptor in coated pits before endocytosis. Signaling ligands for the LDL receptor family are usually internalized by endocytosis. Because several LDL receptors share common ligands, their function in signaling or endocytosis is determined mainly by their expression pattern (68).

LRPs work in concert with the LDL receptors to mediate the cellular uptake of chylomicron remnants. They also bind structurally diverse and unrelated ligands. Some of the LRPs are scavenger receptors or multiligand receptors (69), each interacting with a large number of ligands and participating in endocytosis and co-receptor functions. As previously mentioned, LRP5 and LRP6 are also essential for Wnt signaling through the 7TM Frizzled receptor (16, 70).

Integrins are composed of two subunits, α and β, and are another type of recruiter receptor with dual functionality. In humans, there are ~20 α and eight β chains. Each αβ combination shows unique ligand-binding and signaling properties (Fig. 4). One role of integrins is structural; they bridge the ligands of the extracellular matrix (ECM) with the cytoskeleton. Most integrins recognize several ECM proteins (71), and individual matrix proteins can bind to several integrins. By recognizing cognate ECM proteins, the integrins provide the context within which cells respond to extracellular signals. As signaling receptors, ligand-bound integrins stimulate the recruitment of a plethora of downstream molecules (for example, kinases) and promote clustering of the integrins themselves (72). Survival of a cell depends on the surrounding ECM, and cell death upon detachment from specific ECMs is termed anoikis, a mechanism that limits the spreading of cells out of their tissue boundary. Thus, cells that detach from their original location or penetrate to adjacent tissues are deterred by their contact with the nonsupportive ECM. Malignant cells, in many cases, lose their integrin specificity and thus are capable of metastasizing and spreading uncontrollably (73). The role of integrins as self-adhering molecules could have evolved during the emergence of early metazoans with two body layers, each one displaying one type of integrin (74). In C. elegans, there is one β integrin and two α integrins, and orthologs of these are found in D. melanogaster (75). In vertebrates, multiple paralogs have evolved for each of the subunits of integrins.

The Roundabout (Robo) receptors are a family of recruiter receptors important for cell migration. Robo receptors share motifs with the immunoglobulin superfamily, and orthologs exist in D. melanogaster and C. elegans. Multiple Robo genes have been described in mammals. Members of the Slit family are ligands for the Robo receptors. Robo receptors are essential for neuronal growth guidance (76, 77), and are expressed in nonneuronal tissues, where they also have roles in migration (78). The interaction between Slit-2 and Robo-1 is enhanced by cell-surface heparan sulfate (79). After ligand binding, Robo recruits such signaling components as the Abelson (Abl) tyrosine kinase and its substrate Enabled (Ena) (Fig. 4). Abl antagonizes Robo signaling, whereas Ena is required for Robo's repulsive action (80). The relative effect of each signaling component depends on concomitant input from other signaling systems. Upon Slit binding, Robo also serves as a silencer of the attractant-type netrin receptor, called deleted in colorectal carcinoma (DCC), through direct interaction of their intracellular domains (81). Thus, the Robo family of receptors shows several features of signaling receptors: modulation of activity by additional extracellular factors or membrane-bound proteins, and cross-talk with other receptors.

Another family of guidance receptors that are recruiters is the plexins, which are receptors for semaphorin ligands (82, 83). Most semaphorins bind directly to plexin receptors, but class 3 semaphorins require neuropilins as co-receptors (84). The extracellular region of the plexin receptor has a sema domain (85) and its intracellular domain recruits the monomeric small guanosine triphosphatase Rac1 (Fig. 4) and the collapsin response mediator proteins (CRMPs) (86). Semaphorins exist as both secreted and membrane-associated forms (87). They also have the sema domain, raising the possibility that these ligands and their receptors arose from a common ancestor (88). This ligand-receptor pair, much like the RTKs, spans the spectrum of interactions from direct cell-to-cell adhesion signaling by GPI-linked ligands to signaling by soluble ligands.

As seen with several other members of this recruiter group of receptors, plexins also exhibit receptor cross-talk. In epithelial cells, plexin B1 (the receptor for sema 4D) associates with Met, the RTK receptor for hepatocyte growth factor. Binding of sema 4D to plexin B1 stimulates tyrosine phosphorylation of both plexin B1 and Met; cells lacking Met do not respond to sema 4D (81). Class 4 semaphorins expressed in immune cells appear to activate another receptor, Tim-2, in addition to plexins (89).

Two types of neuronal guidance receptors activated by the netrin ligand are likely to be recruiter-type receptors. The attractant-type netrin receptor is known as DCC in humans, and heparin is a co-ligand for it. Human DCC has conserved orthologs in D. melanogaster (Frazzled), C. elegans (UNC-40), and zebrafish (zDCC) (90). The repellent-type netrin receptor is called RCM in mammals and UNC-5 in C. elegans (91); they have no sequence similarity with DCC. Although the exact mechanism of action of these receptors is unclear, intracellular levels of cAMP are critical to neuronal guidance regulated by netrin receptors (92). For the RCM receptor, ligand binding leads to the phosphorylation of the receptor by a Src family tyrosine kinase, creating potential binding sites to recruit downstream signaling proteins (93).

Receptors with Intrinsic Transcriptional Activity

Although activation of most plasma membrane receptors ultimately leads to transcriptional changes, very few plasma membrane receptors can serve as both receptors and transcription factors. Notch receptors, the best-characterized members of this family of receptors with intrinsic transcriptional activity, transduce signals primarily by nuclear translocation of a cleaved intracellular fragment. The ligands for the Notch family of receptors are Delta, Jagged, and Serrate, a group of membrane-anchored proteins. The Notch receptor is a heterodimer formed from its precursor after cleavage by secretases β and γ (94). After ligand binding, the receptor is further cleaved within its transmembrane domain by presenilin-1-associated γ-secretase (95), thereby releasing the intracellular region for nuclear translocation (Fig. 5). This region then binds to transcription factors, which together form a complex to regulate gene activity (96). In the fly, Notch controls the dorsal-ventral boundary in the developing wing; four mammalian paralogs are known (97). Fringe, a glycosyltransferase, adds O-linked fucose (98) to EGF modules of Notch receptors to regulate their ligand binding.

Fig. 5.

Notch and Patched receptors. Binding of Delta to Notch receptors leads to intramembranous proteolytic cleavage of the receptor by γ-secretase and the translocation of its intracellular region to the nucleus. The Patched receptor, a 12TM protein, and Smoothened (Smo), a 7TM protein, are important for Hedgehog (hh) signaling. Patched inhibits the action of Smo to proteolytically activate Cubitus interruptus (Ci), a transcriptional repressor. Binding of the ligand HhN (composed of hh and a cholesterol moiety) to Patched relieves this inhibition.

Transporter-Like Receptors

Patched receptors have 12 transmembrane domains. They are unique among plasma membrane receptors, because they have no identified intrinsic enzymatic activity or recruiter function. Patched receptors are activated by the Hedgehog family of ligands, which are modified by the covalent attachment of a cholesterol moiety. Ligand binding to Patched suppresses the function of Smoothened, a 7TM protein (Fig. 5). Smoothened, in turn, regulates the proteolytic cleavage of Cubitus interruptus (Ci), a transcriptional repressor (99). Ligand activation of patched suppresses proteolysis induced by Smoothened, relieving the repressor function of Ci.

In C. elegans and D. melanogaster, several genes encoding proteins related to Patched and Hedgehog exist (100, 101). Mutations in human patched and smoothened genes are associated with basal cell carcinoma (102). Patched does not dimerize with Smoothened and is structurally related to a bacterial transmembrane transporter and to human Niemann-Pick type C disease gene product, a cholesterol transporter (103). Thus, Patched may function as a transporter, and ligand binding to Patched could change the distribution or concentration of a small molecule (104). Patched may represent the fading boundary between classic receptors (which bind external ligands and activate separate intracellular components) and transporters and channels (which move components across the membrane, thus signaling by altering the concentration of signaling molecules across the membrane). Of course, many receptors, not discussed here, are ligand-gated ion channels that serve to initiate signaling cascades by altering the concentration of ions across the plasma membrane.

Plasma Membrane Receptors Found Only in Chordates

In chordates, several new recruiter-type receptors have evolved (Fig. 6). These receptor families do not have orthologs in C. elegans and D. melanogaster and are most likely chordate specific. The class 1 cytokine and tumor necrosis factor (TNF) receptors have orthologs in the sea squirt Ciona intestinalis (105). In contrast, T cell receptors and class 2 cytokine receptors, in the presently available genomes, have orthologs only in the genome of Fugu rubripes (a pufferfish), but underwent major expansion in tetrapods (106, 107).

Fig. 6.

Receptors present only in chordates. (A) Class 1 cytokine receptors consist of four types of subunits, α, β, γ, and gp130. These subunits may act as homodimers in the case of the GH receptor, or as heterodimers as in IL-6 or granulocyte macrophage-colony stimulating factor (GM-CSF) receptors. The extracellular region of these subunits contains cytokine receptor-specific domains. The intracellular region has box 1 and 2 domains. The hematopoietic class (growth hormone and prolactin receptors) consists of β-subunit homodimers; the IL-3 class (IL-3, IL-5, and GM-CSF receptors) shares an α-subunit and a β-subunit; the IL-2 class shares the α-subunit and includes some heterotrimeric members with α-, β-, and γ-subunits (IL-2 and IL-15 receptors) and some heterodimeric members with only α- and γ-subunits (IL-7, IL-9, and IL-4 receptors). The IL-6 class members all share the gp130 subunit, IL-6 and IL-11 receptors being dimeric, whereas leukemia inhibitory factor (LIF), ciliary neurotrophic factor (CNTF), oncostatin M (OSM), and cardiotrophin-1 (CT-1) receptors are trimeric (113). This class of cytokine receptors has numerous isoforms, serving as binding proteins (for example, GH binding protein) and decoy receptors (for example, prolactin decoy receptor). (B) The TNF trimeric ligand binds trimeric receptors with an intracellular DD, leading to the activation of death effector domain proteins (FADD and TRADD) and TRAF proteins. Some TNF receptors do not have a DD and recruit intracellular DD proteins and TRAFs. TRAFs then activate the caspase 8 enzyme. The p75NTR subfamily of neurotrophin receptors associates with Trks for signal transduction. The p75NTR receptor contains extracellular cysteine-rich domains, whereas Trks contain extracellular immunoglobulin-like domains and leucine-rich repeats.

The family of class 1 cytokine receptors includes receptors for many interleukins, growth hormones (GHs), and colony-stimulating factors. These receptors have a common extracellular domain but little similarity in the cytoplasmic region (108). There are four subunits--α, β, γ, and gp130--that are involved in the formation of different receptor complexes in this large family (Fig. 6) (109). Ligand binding promotes mostly hetero-oligomerization but, in some cases, can cause the homodimerization of receptor subunits (for example, the GH receptor). These receptors are recruiters that stimulate the recruitment of proteins in the Janus kinase (JAK) and signal transducer and activator of transcription (STAT) (JAK-STAT) pathway (110). This family displays many truncated receptors serving as secreted binding proteins, or functional antagonists, as a result of alternative splicing or proteolysis (111).

This ligand-receptor system shows a high degree of redundancy, and many null mutations in mice exhibit minimal phenotypes. Gene structures of the interferon (IFN)-α receptor and the type 1 cytokine receptors suggest a common origin with immunoglobulins (112). A JAK-STAT-activating receptor with limited sequence identity to gp130 (113) was found in D. melanogaster (114, 115). In addition, the JAK-STAT pathway is conserved in invertebrates (116); thus, as more signaling pathways become elucidated, the class 1 receptors may need to be reanalyzed for position in the receptome phylogeny.

The TNF receptor family is another family of chordate-specific recruiter receptors. Lymphocyte demise is initiated by activation of death-inducing molecules such as ligands for the TNF receptors (117). TNF receptors fall into two classes that are based on the presence or absence of the intracellular death domain (DD). Those with the DD directly recruit various cytoplasmic proteins with homologous domains, whereas those lacking the DD trimerize after ligand binding and recruit DD-containing TNF receptor-associated factors (TRAFs) (Fig. 6) (118, 119). The TRAFs then recruit downstream signaling components. In addition to initiating apoptosis, TNF-α may also activate a rescue pathway, and the balance between the two pathways could determine the fate of the cell (120). This family has membrane-anchored and soluble truncated proteins that act as decoys. Some TNF family ligands are also membrane-anchored, becoming soluble following proteolysis (119).

The p75 neurotrophin receptor (p75NTR) is a proapoptotic TNF receptor with a death domain that interacts with three RTKs of the neurotrophin family (Trk A, B, and C) that have antiapoptotic and differentiation roles. p75NTR binds all neurotrophins with similar affinity and may signal independently from the Trks (121). However, the Trk interaction confers ligand specificity on p75NTR (120, 122). Homodimeric Trks combine with oligomeric p75 NTR to form a multimeric complex that modifies the survival signals mediated by the Trks (123). The stoichiometry of these interactions serves to specify downstream signaling. Splicing variants of both p75 NTR and the Trks further contribute to the complexity of the system (124). A member of the TNF receptor family, Fas, exhibits a similar receptor cross-talk interaction. In hepatocytes, the association of Fas with the extracellular region of c-Met (an RTK for the hepatocyte growth factor) prevents ligand binding and the clustering of Fas, leading to cell survival (125).

Plasma Membrane Receptors Found Only in Vertebrates

As vertebrates evolved, new signaling receptors also emerged. These include receptors implicated in immune signaling such as the T cell receptors and the class 2 cytokine receptors. In addition, proteins anchored to the plasma membrane through a GPI anchor also appeared in vertebrates. These nontransmembrane receptors require a co-receptor for ligand signaling to the cell's interior.

There are four T cell receptors (α, β, γ, and δ) that form two different heterodimers (α-β and γ-δ). These receptors are similar to immunoglobulins in primary sequence, gene organization, and modes of genetic rearrangement. T cell receptors recognize antigen fragments when they are embedded in major histocompatibility complexes (MHCs) (Fig. 7) (126) and serve as recruiter receptors. Signals elicited by MHC-antigen binding of the T cell antigen receptor and the CD4 or CD8 co-receptors control the differentiation of thymic precursors into either CD4+ (helper) or CD8+ (cytotoxic) T cells. The mature T cells that emerge express only the CD4 or CD8 co-receptor and complement the MHC class-specificity of the T cell receptor. Interestingly, Notch-1 is also a co-receptor for CD8+ T-cell maturation (127). T cell receptor genes have been defined in most of the major lineages of jawed vertebrates (107).

Fig. 7.

Receptors present only in vertebrates. (left) T cell receptors. There are four different T cell receptor subunits, all with a common immune receptor-based tyrosine activation motif (ITAM). Upon ligand binding, the receptors are phosphorylated by the tyrosine kinase Lck (157). The antigen (Ag) is presented by the MHC molecules that are expressed on a different cell. The associated CD4 or CD8 molecule, together with the Notch receptor (not shown), is essential for the final differentiation of T cells. (right) Class 2 cytokine receptor family. The homodimeric IFN-γ interacts with the high-affinity receptor α chain containing two immunoglobulin-like and two fibronectin III-like domains. This, in turn, enables the recruitment of the low-affinity receptor β chain that has two fibronectin III-like domains. Recruitment of the β chain activates JAKs, culminating in the nuclear translocation of STATs.

The class 2 cytokine receptors include receptors for type II interferon (IFN-γ), type I interferons (IFN-α/β), interleukin (IL)-10, and coagulation factor VII. Four classes of type I IFNs share the same receptor complex, whereas the only type II IFN, IFN-γ, binds to a distinct type II receptor (128). IFN-γ signals through a receptor dimer of the IFN-γ receptor-α and the co-receptor IFN-γ receptor-β (129). Ligand binding to the high-affinity IFN-γ receptor-α induces conformational compatibility between the homodimeric ligand and the receptor heterodimer (Fig. 7) (130). Similarly, type I interferons also bind to a dimer composed of the receptor subunits IFNAR1 and IFNAR2. Ligand binding to these receptors recruits multiple kinases, including p135tyk2 and Jak-1 (initiators of the JAK-STAT pathway), which phosphorylate the receptor and downstream proteins (131, 132). IL-22RA2 is a naturally expressed soluble IL-22 antagonist with sequence similarity to type 2 cytokine receptors, representing a derivation of antagonists in this receptor family (133).

In contrast to membrane-traversing receptors, several vertebrate receptors are anchored on plasma membrane through a GPI linkage and require the participation of a transmembrane co-receptor for downstream signaling. The four glial cell-derived neurotrophic factors (GDNFs) bind to GPI-anchored receptors (GFR 1 through 4) that signal through a co-receptor, c-Ret (an RTK) (134). Another GPI-anchored receptor, the urokinase-type plasminogen activator receptor (uPAR), binds uPA and interacts with diverse co-receptors such as integrins, LRPs, IGF II receptor, and gp130. Because uPAR shares domains with adhesion molecules, it is situated at the boundary of cellular adhesion and the ligand-activated receptors (135). Orthologs for GFRs, but not for uPAR, are present in the Fugu genome.

Recurrent Themes in the Evolution of Plasma Membrane Receptors Involved in Signaling

Surveying the human signaling receptome reveals several fundamental principles.

Receptors are made from a limited number of structural motifs. A limited "tool kit" of similar protein motifs is found in the signaling receptome as a result of gene duplication and domain shuffling. Extracellular regions of receptors are usually composed of domains present in more than one receptor family, whereas ligand specificity is conferred by the combinatorial use of these domains. In contrast, the intracellular region of these receptors is composed of a limited number of enzymatic or recruiter domains for signal propagation. The prevalent domains in the extracellular regions of human plasma membrane receptors (deduced from the Receptome database) are fibronectin, immunoglobulin-like, and EGF domains, motifs also used for cell-cell adhesion.

Receptors may be multifunctional. Several receptor types have dual roles, serving as signaling molecules as well as co-receptors, adhesion proteins, or mediators of endocytosis. In certain families (for example, the LDL receptor family), some members are predominantly signaling molecules, whereas others have mainly nonsignaling roles. In addition, the same receptor can signal through different downstream elements in different cell types (2).

Homo- and heterodimerization of receptors is a central mechanism in signal transduction. Receptor dimerization is the rule for RTKs, but dimerization of others (such as type A GPCRs) is not obligatory for receptor function. Combinatorial interactions of receptors with co-receptors allow refinement of signaling specificity and diversity. For example, in the cytokine type 1 receptor family, several receptor subunits oligomerize in different combinations to form complexes with unique ligand specificity capable of recruiting unique downstream effectors. In addition, receptors from families that appear to have emerged later in evolution (such as p75NTR, from the TNF receptor family) interact with receptors with a more ancient origin (Trks, from the RTK family), leading to signaling diversification and robustness.

The ligand-binding domains of receptors may evolve into functional antagonists. Antagonist-binding proteins or sequestering receptors without signaling ability are most likely derived from receptor gene duplication, alternative splicing, or proteolytic cleavage. These negative regulators of signal transduction are found in both endocrine and paracrine systems, the former represented by the circulating growth hormone-binding protein and the latter by truncated RTK receptors and natriuretic peptide receptor-C. Plasma membrane-bound decoy receptors (for example, members of the TNF receptor family) may antagonize the effect of the cognate ligands by functioning in a dominant-negative manner.

Endocrine systems are evolved from ancestral paracrine systems. For almost all endocrine ligand-receptor pairs, a related autocrine or paracrine pair exists within the same receptor family. Endocrine hormones that signal over long distances likely evolved from truncated or modified local ligands. During evolution, changes in the cis-regulatory sequences of individual ligand and receptor genes confer differential spatiotemporal expression. Thus, the evolution of sophisticated endocrine systems may be the result of novel interactions that are derived from the temporal and spatial segregation of paralogous genes (136).

Evolution of Receptors and Ligands

Molecular coevolution between interacting proteins facilitates the establishment of biological novelties. Understanding the phylogeny of different ligand-receptor pairs led to the recent discovery of ligands for orphan receptors (137-139) and paralogous ligands for known receptors (140-142) based on the presumed coevolution of polypeptide ligands and receptors. Elucidation of the "subgenome" containing a limited number of coevolved genes will continue to advance ligand-receptor matching, thus "de-orphanizing" many receptors with unknown functions, including a large group of so-called viral receptors (143).

Global analysis of plasma membrane receptors suggests that selected sets of secreted or membrane-anchored molecules were likely "adopted" by different plasma membrane receptors. Prominent examples of adoptable ligands are proteins with a rigid cystine-knot structure. The cysteine arrangement in these proteins constrains their folding, exposing hydrophobic residues to facilitateinteractions with potential receptors. Remarkably, this conserved core structure is found in ligands for diverse receptor families: 7TM receptors (glycoprotein hormones), serine kinase receptors (TGF-β family), RTKs (PDGF family), and GPI-anchored receptors (GDNFs) (144) (Fig. 8).

Fig. 8.

Cystine-knot proteins are ligands for receptors from diverse families. Proteins containing the conserved cystine-knot structure interact with diverse recruiter (7TM), enzymatic (Ser-Thr kinase and RTK), and GPI-linked receptors. TGF-β, transforming growth factor β; PDGF, platelet derived growth factor; GDNF, glial cell-derived neurotrophic factor.

Further, the same ligand can activate receptors from different families. Midkine and pleiotrophin may be ligands for both the RPTP-ζ (55) and the RTK anaplastic lymphoma kinase (145), suggesting that two different enzyme-type receptors adopted the same ligands. These ligands can suppress the constitutive activity of RPTP and activate RTK, both leading to increased tyrosine phosphorylation (56). In contrast, netrins activate two recruiter-type receptors (DCC and RCM) with opposing actions (146). Cells with DCC are attracted to the netrin source, whereas cells with RCM are repelled (147). Furthermore, some peptides (FMRFamides) and small molecules (glutamate, acetylcholine, serotonin, and GABA) are not only ligands for GPCRs, but also interact with ligand-gated channels (148). Likewise, progestins, primarily known as ligands for nuclear receptors, can also activate plasma membrane GPCRs (149).

Tracing the exact evolutionary origin of plasma membrane receptors is difficult because the enormously diverse sets of ligand-receptors likely evolved independently multiple times. No simple model can address the chicken-or-egg question regarding receptor and ligand origins. Fortunately, molecular relics provide clues for the possible evolution of certain receptor groups, and two nonexclusive mechanisms can be proposed.

Common Prereceptors or Distinct Adhesion Molecules Evolved into Modern Receptors and Ligands

In the ephrin-Eph receptor and semaphorin-plexin families, ligands and receptors have sequence similarity, and these interacting pairs likely evolved from common ancestral prereceptors (Fig. 9, upper panel, A1). Ephrins are transmembrane proteins, whereas semaphorins exist in both secreted and membrane-bound forms. It is possible that bidirectional signaling by prereceptors with both ligand and receptor properties predates the modern ligand-receptor pairs involved in unidirectional signaling. After duplication of common prereceptor genes, some remained as membrane-bound signaling molecules (receptors), whereas others retained only ligand properties. The resultant directional signaling would allow for a clear division of labor between adjacent cells.

Fig. 9.

Models to explain the evolutionary origins of receptor-ligand pairs. (I) Receptors and ligands evolved from common prereceptors or interacting cell adhesion molecules. (A) The prereceptors are either (i) identical or (2) distinct transmembrane proteins capable of bidirectional signaling (stars). Unidirectional signaling emerged when one of the pairs either (B) loses activity in the intracellular domain, or (C) becomes truncated in the intracellular region but still associates with the plasma membrane through a GPI linkage. Further evolution leads to a secreted ligand following either (D) posttranslational cleavage of the extracellular region or (E) gene splicing or truncation. All these forms are present among different classes of human RTKs. (II) Constitutively active prereceptors adopt ligands. (A) The ancestral 7TM prereceptor has intrinsic (constitutive) signaling activity (star). (B) An extracellular molecule (for example, a modified amino acid) binds to the prereceptor and increases its signaling. (C) Natural selection favors a decrease in the constitutive activity of the receptor and an increase in ligand activation.

Table 1.

Classification of the human plasma membrane signaling receptome. Human receptor families are classified on the basis of their evolutionary origins and signaling properties. The listing represents the order of proposed emergence during evolution.

Table 2.

Number of 7TM receptor genes in each subgroup of 6 completely sequenced genomes.

The evolution from bidirectional to unidirectional signaling can be extended to some receptors without sequence similarity to their ligands--both receptors and ligands originating from distinct transmembrane adhesion molecules that are capable of binding to each other (Fig. 9, upper panel, A2). During evolution, the preligands developed properties to modify downstream signaling by prereceptors, and natural selection favored an increase in the signal-to-noise ratio of the signaling system. Modern soluble ligands could have derived from the alternative splicing of transcripts to remove the transmembrane region (for example, the Kit ligand for the mast cell growth factor receptor) or from the enzymatic cleavage of the extracellular domains. In the RTK superfamily, ligands range from transmembrane (ephrin B) to GPI-anchored (ephrin A), and cleaved (EGF) to secreted (insulin) forms. Thus, the primal bidirectional signaling system may have evolved into juxtacrine, paracrine, and endocrine systems (Fig. 9, upper panel, B, C, D, and E).

Intrinsically Active Receptors Adopt Ligands

In the 7TM receptor superfamily, some wild-type or mutant receptors are constitutively active (150), suggesting the existence of prereceptors that signal without ligand participation. An intrinsically active prereceptor could have adopted a binding partner through coincidental interactions with potential ligands capable of binding and signal amplification (Fig. 9, lower panel). This model may apply to GPCRs for modified amino acids (for example, glutamate and catecholamines) or metabolic byproducts (for example, prostaglandins). Subsequent evolution may have led to an increased signal-to-noise ratio of ligand signaling and the expansion of receptor-ligand sets.

In many cases, the primitive ligands were probably initially adapted for nonreceptor functions, and the thrombin receptor signaling may represent a model in which an enzyme evolved into a ligand for a GPCR. Thrombin binds to and cleaves part of the ectodomain of its receptor to reveal a cryptic region that activates the transmembrane core of the receptor (151). Some 7TM receptors (for example, the glycoprotein hormone receptors) have a large ectodomain that constrains their 7TM region (152). They represent another evolutionary mechanism through which the intrinsic activity of the prereceptor was decreased. These partly constrained GPCRs could also adopt ligands capable of interacting with the ectodomains to relieve the conformational constraint on the 7TM region, thus leading to receptor activation. If one extends this model to enzyme- and recruiter-type receptors, it is interesting to note that the extracellular region of some RTKs can suppress receptor activity by preventing receptor dimerization (153). Again, ligand binding relieves this constraint. In contrast, some RPTPs have constitutive activity that is silenced by ligand binding (56). Thus, the ancestral constitutive signaling evolved into sophisticated ligand-regulated systems.

Perspectives and Conclusions

Greater than 15% of the human genome consists of genes with predictable transmembrane domains. The present analysis represents an initial attempt to characterize the signaling receptome and does not include transmembrane proteins with mainly adhesion, transport, channel, or extracellular catalytic activity. Some receptor families with few members or nontraditional signaling receptor functions (transferrin receptors, folate receptors, selectins, and tetraspanins) are listed in the Human Plasma Membrane Receptome database and not discussed further. It is becoming clear that the existing rigid classification of transmembrane signaling molecules as receptors, adhesion molecules, channels, enzymes, and transport proteins is inadequate. Future studies on the human receptome should expand into additional cell surface signaling molecules, such as scavenger receptors, ligand-gated ion channels, and membrane-type matrix metalloproteinases, and so on. With an evolutionary genomic analysis of all transmembrane proteins involved in cell-to-cell communication, an integrated view will then emerge.

A better understanding of the evolutionary and functional relationship of different human plasma transmembrane receptors could facilitate the characterization of large numbers of membrane-spanning proteins without known function in the predicted human proteome. With multiple examples of cross-talk between receptors and adhesion molecules, as well as direct interactions between RTKs and ion channels [for example, TrkB and the sodium channel Na(V)1.9] (154, 155), the availability of the complete inventory of human plasma membrane signaling molecules will aid in the understanding of interactions among different classes of transmembrane proteins.

Paracrine regulators for any organ outnumber the endocrine regulators. Analyses based on DNA and protein microarrays allow expression profiling of all human genes. Understanding the human receptome could allow the identification of coexpressed ligand-receptor pairs to reveal paracrine networks in a given tissue. The availability of genomic and proteomic data further allows identification of receptor variants that are derived from gene splicing or proteolytic cleavage.

Loss-of-function mutations in receptors often underlie recessive diseases, whereas gain-of-function mutations are usually associated with dominant inheritance. Because many receptors function through multimeric complexes, certain loss-of-function mutations may also give rise to dominant-negative mutants (156). The chromosomal loci of human receptors and their syntenic regions in model vertebrates are being revealed. This lays the foundation to elucidate the genetic basis of idiopathic conditions involving receptor defects. A comprehensive understanding of the human receptome will allow receptor pharmacology to expand so that new agonists and antagonists can be derived. A majority of currently used drugs modulate the 7TM receptors or are recombinant agonists or antagonists for various receptor families. Further elucidation of the enzymatic and recruiting specificity of different receptors, as well as the mechanisms underlying receptor-receptor cross-talk, could facilitate new drug development. Moreover, pharmacogenomic analysis of single-nucleotide polymorphisms of receptor genes will improve the diagnosis of hormonal diseases and reveal individual variations in drug responsiveness. As additional families of hormonal receptors and ligands are discovered, new areas of research will emerge to encompass endocrinology, neurobiology, growth factor biology, extracellular matrix research, protease biology, and developmental biology.


  1. 1.
  2. 2.
  3. 3.
  4. 4.
  5. 5.
  6. 6.
  7. 7.
  8. 8.
  9. 9.
  10. 10.
  11. 11.
  12. 12.
  13. 13.
  14. 14.
  15. 15.
  16. 16.
  17. 17.
  18. 18.
  19. 19.
  20. 20.
  21. 21.
  22. 22.
  23. 23.
  24. 24.
  25. 25.
  26. 26.
  27. 27.
  28. 28.
  29. 29.
  30. 30.
  31. 31.
  32. 32.
  33. 33.
  34. 34.
  35. 35.
  36. 36.
  37. 37.
  38. 38.
  39. 39.
  40. 40.
  41. 41.
  42. 42.
  43. 43.
  44. 44.
  45. 45.
  46. 46.
  47. 47.
  48. 48.
  49. 49.
  50. 50.
  51. 51.
  52. 52.
  53. 53.
  54. 54.
  55. 55.
  56. 56.
  57. 57.
  58. 58.
  59. 59.
  60. 60.
  61. 61.
  62. 62.
  63. 63.
  64. 64.
  65. 65.
  66. 66.
  67. 67.
  68. 68.
  69. 69.
  70. 70.
  71. 71.
  72. 72.
  73. 73.
  74. 74.
  75. 75.
  76. 76.
  77. 77.
  78. 78.
  79. 79.
  80. 80.
  81. 81.
  82. 82.
  83. 83.
  84. 84.
  85. 85.
  86. 86.
  87. 87.
  88. 88.
  89. 89.
  90. 90.
  91. 91.
  92. 92.
  93. 93.
  94. 94.
  95. 95.
  96. 96.
  97. 97.
  98. 98.
  99. 99.
  100. 100.
  101. 101.
  102. 102.
  103. 103.
  104. 104.
  105. 105.
  106. 106.
  107. 107.
  108. 108.
  109. 109.
  110. 110.
  111. 111.
  112. 112.
  113. 113.
  114. 114.
  115. 115.
  116. 116.
  117. 117.
  118. 118.
  119. 119.
  120. 120.
  121. 121.
  122. 122.
  123. 123.
  124. 124.
  125. 125.
  126. 126.
  127. 127.
  128. 128.
  129. 129.
  130. 130.
  131. 131.
  132. 132.
  133. 133.
  134. 134.
  135. 135.
  136. 136.
  137. 137.
  138. 138.
  139. 139.
  140. 140.
  141. 141.
  142. 142.
  143. 143.
  144. 144.
  145. 145.
  146. 146.
  147. 147.
  148. 148.
  149. 149.
  150. 150.
  151. 151.
  152. 152.
  153. 153.
  154. 154.
  155. 155.
  156. 156.
  157. 157.
  158. 158.
View Abstract

Stay Connected to Science Signaling

Navigate This Article