Seven-Helix Receptors

Members of the largest family of plasma membrane receptors are built from a serpentine arrangement of seven transmembrane α-helices. These diverse receptors use trimeric GTP-binding proteins (see Fig. 25-9) to relay signals to effector proteins inside cells. Seven-helix receptors are found in slime molds, so the genes for these proteins originated in early eukaryotes more than 1 billion years ago. Four percent of the genes of the nematode Caenorhabditis elegans (790) encode seven-helix receptors, the largest family of proteins in the organism. In mammals, olfactory cells alone use 500 to 1000 different seven-helix receptors to discriminate odorant molecules (see Fig. 27-1). Other cells are estimated to express another 375 seven-helix receptors to respond to light, amino acids, peptide and protein hormones, catecholamines, and lipids. The chemical ligand remains to be determined for about 40% of these 375 receptors, which are termed orphan receptors. A majority of medically useful drugs bind seven-helix receptors.

Seven hydrophobic sequences traverse the plasma membrane as α-helices (Fig. 24-2). The topology is the same as that of bacteriorhodopsin (see Fig. 7-8), but this might be an example of convergent evolution. Comparative analysis of amino acid sequences suggests that all seven-helix receptors have the same arrangement of helices. For example, the minimum length of sequences connecting the helices is compatible only with the helices being arranged sequentially from I to VII in a serpentine fashion as they cross the lipid bilayer. The N-terminus is outside the cell and varies from 7 to 6000 residues. Some of the larger N-terminal domains participate in ligand binding. The C-terminal segment of the polypeptide is in the cytoplasm and varies in length from 12 to more than 350 residues. Seven-helix receptors are shown throughout this book as individual proteins, but multiple lines of evidence show that many seven-helix receptors function as dimers or larger oligomers, allowing for cross talk between the subunits.

Figure 24-2 structure of seven-helix receptors. A, Model of the generic seven-helix receptor, used throughout this text. The structure is based on the structure of bacteriorhodopsin and the amino acid sequence of a typical vertebrate seven-helix receptor. Seven hydrophobic segments cross the membrane as α-helices. Oligosaccharides are blue. The C-terminal cytoplasmic tail is anchored to the lipid bilayer by two fatty acids covalently bound to a pair of adjacent cysteines. B, Atomic structure of bovine rhodopsin, the receptor for photons in the retina. Covalently bound retinal pigment is green. Oligosaccharides on extracellular loops are blue.

(PDB file: 1F88. Reference: Palczewski K, Kumasaka T, Hori T, et al: Crystal structure of rhodopsin: A G protein–coupled receptor. Science 289:739–745, 2000.)

Most seven-helix receptors are activated by binding a soluble chemical ligand, but some interesting variations exist. Biochemical and mutagenesis experiments indicate that most small ligands bind in a central pocket among the extracellular ends of the helices. Residues lining this pocket are highly variable between receptors, providing specificity for each receptor to bind a particular ligand. The light-absorbing pigment 11- cis retinal, of the photoreceptor protein rhodopsin, is the best-characterized “ligand.” 11-cis retinal is unusual in that it is bound covalently to the receptor (Fig. 24-2B) and is activated by absorbing a photon (see Fig. 27-2). In other respects, it is a good model for other ligands (Fig. 24-3). Neurotransmitters, such as norepinephrine, and drugs also bind between the helices about one third of the way across the membrane. Peptide hormones bind deep in the helical pocket but probably also interact with residues that are more exposed on the cell surface. Receptors for some large ligands (pituitary glycoprotein hormones, such as luteinizing hormone, follicle-stimulating hormone, and thyroid-stimulating hormone) and some small ligands (glutamate, g-amino butyric acid, calcium) bind with high affinity to extracellular N-terminal domains of their seven-helix receptor. The N-terminal domain with bound ligand then stimulates the transmembrane domain of the receptor. The blood-clotting enzyme thrombin activates its receptor on platelets by proteolysis of the receptor rather than by direct binding (see Fig. 30-14). The N-terminal peptide cleaved from the receptor disso-ciates and activates other receptors; what is left of the newly truncated N-terminus folds back and activates its own receptor.

Figure 24-3 activation and adaptation of a seven-helix receptor. A, Ligand binding shifts the equilibrium from the resting conformation toward the active conformation. B, The active receptor promotes dissociation of guanosine diphosphate from the α-subunit of multiple trimeric G-proteins, allowing GTP to bind. Typically, this dissociates G_a from G_bg, each of which activates downstream effectors that produce, for example, the second messengers cAMP and diacylglycerol (DAG). cAMP and DAG activate PKA and PKC, which phosphorylate active receptors on their C-terminus. C, This attracts arrestin, putting the receptor into the inactive adapted state. PKA, protein kinase A; PKC, protein kinase C. (PDB file for arrestin: 1CF1.)

Seven-helix receptors exist in an equilibrium between two conformations: a resting state and an activated state with the ability to catalyze the exchange of nucleotide bound to trimeric G-proteins (Fig. 24-3). Without bound ligand, the resting state is strongly favored. Ligand binding to the receptor (or the isomerization of retinal after absorbing light) initiates signal transduction by shifting the equilibrium to the active state. Activation involves movement of at least two transmembrane helices, but the structural details of this conformational change are not yet well defined. In any event, activation must rearrange the cytoplasmic ends of the helices and the loops connecting them to create a binding site for a target G-protein.

Active receptors transfer the signal to the cyto-plasm by activating trimeric G-proteins. Cytoplasmic loops of active receptors catalyze the dissociation of guanosine diphosphate bound (EDP) to an inactive Ga subunit. GTP then binds and activates Ga (see Fig. 25-9). A single active seven-helix receptor can amplify the signal by activating up to 100 G-proteins. After dissociating from the receptor and each other, both Ga-GTP and Gbg stimulate downstream effector proteins, further amplifying the signal (see Fig. 27-3 for an example of amplification).

Most seven-helix receptors adapt to sustained stimulation. In the short term, phosphorylation of the C-terminal tail of the receptor provides negative feedback that inactivates receptors with ligands still bound (Fig. 24-3C). Along one pathway, second messengers—produced in response to receptor activation—stimulate protein kinases, including cyclic adenosine monophosphate (cAMP)–activated protein kinase A and protein kinase C (see Fig. 25-4). These kinases phosphorylate the C-terminal tails of active receptors, inhibiting interactions with G-proteins. This mechanism allows for cross talk between receptors, as activation of one class of receptors can inactivate other receptors. A second pathway uses Gbg subunits released in response to receptor stimulation to activate protein kinases specific for the receptors themselves, called G-protein-coupled receptor kinases. These kinases phosphorylate serines or threonines on the C-terminal tails of active (but not inactive) receptors.

Phosphorylation of receptor tails creates a binding site for arrestin, a protein with multiple functions. First, arrestin blocks interactions of the receptor with G-proteins, terminating signaling through the main pathway downstream of most seven-helix receptors. In some cases, arrestin initiates a new signal through the MAP kinase pathway (see Figs. 27-6 and 27-7 for two other pathways). Arrestin also promotes the removal of seven-helix receptors from the plasma membrane by endocytosis in clathrin-coated vesicles, a longer-term mechanism that turns down the response of a cell to continuous stimulation. Some internalized receptors recycle to the plasma membrane, but others modified by ubiquitin are directed to lysosomes for destruction. Chapter 27 covers three dramatic examples of seven-helix receptor adaptation.

Mutations of more than 30 seven-helix receptors have been linked to human diseases (Table 24-1). More than 600 mutations are known to inactivate seven-helix receptors in humans by every conceivable means from failure to synthesize the full length protein to reduced affinity for ligands to failure to activate G-proteins. These inherited loss of function mutations are recessive. For example, loss of function mutations in rhodopsin cause retinitis pigmentosa, a degeneration of photoreceptor cells. Another example is severe obesity associated with loss of function mutations of a seven-helix receptor that participates in the neural circuits controlling eating. More than a hundred different mutations produce receptors that are constitutively active without ligand. Particular mutations of rhodopsin cause night blindness, and mutations in a calcium receptor cause dysfunction of the parathyroid gland. The physiology of these activating mutations is complicated, because cells use feedback mechanisms to compensate for the continually active receptors.

Table 24-1 SEVEN-HELIX RECEPTORS AND DISEASE

Receptor Tyrosine Kinases

Many polypeptide growth factors activate cells by binding plasma membrane receptors with cytoplasmic protein tyrosine kinase activity (Fig. 24-4). Ligand binding to extracellular domains allows the cytoplasmic kinase domains of pairs receptors to activate each other and to phosphorylate each other and downstream proteins that control cellular proliferation and differentiation. Mammals have 20 families of receptor tyrosine kinases with distinct structural features.

Figure 24-4 receptor tyrosine kinases. Domain architecture of nine of the 20 families of receptor (R) tyrosine kinases, with ribbon models of several domains. The globular domain of the EphB2 receptor is a b sandwich with a ligand-binding site that includes the exposed loop on the front of this model (PDB file: 1IGY). The extracellular part of the insulin-like growth factor consists of two similar β-helical domains connected by cysteine-rich domains (PDB file: 1IGR). The cytoplasmic kinase domain from the insulin receptor is similar to most known kinases (PDB file: 1IRK). Kinase inserts and C-terminal extensions contain tyrosine phosphorylation sites. Receptor names: EphR, receptor for ephrin, membrane-bound ligands in the nervous system, the largest class of receptor tyrosine kinases; PDGFR, platelet-derived growth factor receptor; FGFR, fibroblast growth factor receptor; VEGFR, vascular endothelial growth factor; Met, receptor for hepatocyte growth factor; TrkA, receptor for nerve growth factor; RET, a cadherin adhesion receptor; Axl, receptor for the growth factor Gas6; EGFR, epidermal growth factor receptor. Domain names: Ig, immunoglobulin; F3, fibronectin-III; CAD, cadherin.

(References: Hubbard SR, Till JH: Protein tyrosine kinase structure and function. Annu Rev Biochem 69:373–398, 2000; Garrett TP, McKern NM, Lou M, et al: Crystal structure of the first three domains of the type-1 insulin-like growth factor receptor. Nature 394:395–399, 1998; and Hubbard SR, Wei L, Ellis L, Hendrickson WA: Crystal structure of the tyrosine kinase domain of the human insulin receptor. Nature 372:746–754, 1994.)

The growth factors that activate receptor tyrosine kinases regulate development and differentiation. For example, epidermal growth factor (EGF) stimulates proliferation and differentiation of epithelial cells. Platelet-derived growth factor stimulates growth of smooth muscle cells, glial cells, and fibroblasts (see Fig. 32-11). Some growth factors were discovered by biochemical purification of proteins that stimulate cellular growth or differentiation. EGF was discovered with a bioassay, as it causes the eyelids of newborn mice to open prematurely. A homolog of the EGF receptor, HER2/ErbB2, was discovered as the normal version of a cancer-causing viral oncogene. Other ligands and receptors were discovered as genes in flies or worms required for development. The Drosophila sevenless gene encodes a receptor tyrosine kinase that is related to insulin receptor. Mutations in the sevenless gene result in failure to develop photoreceptor cell number 7 in the fly’s eye.

Receptor tyrosine kinases consist of an extracellular ligand-binding domain connected to a cytoplasmic tyrosine kinase domain by a single transmembrane helix (Fig. 24-4). The ligand binding is mediated by immunoglobulin domains, fibronectin III domains (see Fig. 3-13), cadherin domains (see Fig. 30-5), and less familiar domains such as β-helical and cysteine-rich domains. This domain architecture illustrates that genes for receptor tyrosine kinases were assembled from sequences for familiar domains followed by divergence to allow for interactions with diverse ligands.

Ligand binding activates all well-characterized receptor tyrosine kinases by bringing together a pair of kinase domains on the cytoplasmic face of the membrane. Dimeric ligands such as platelet-derived growth factor recruit a pair of receptors from the pool of subunits diffusing in the plane of the membrane and connect them physically. This induced dimerization juxtaposes two kinase domains in the cytoplasm. An EGF monomer binds a receptor and induces a conformational change that favors the formation of a receptor dimer but is not a physical part of the connection between the subunits (Fig. 24-5). Insulin binding induces a conformational change in a preformed receptor dimer held together by a disulfide bond. The conformational change brings together the kinase domains (see Fig. 27-7).

Figure 24-5 subunit dimerization mechanism for activating the egf receptor tyrosine kinase. In the absence of EGF, intramolecular interactions preclude dimerization. EGF binding changes the conformation of the extracellular domains allowing dimerization of two receptors, bringing together two cytoplasmic kinase domains. Transphosphorylation activates both kinases and creates phosphotyrosine bind-ing sites for SH2 and PTB domains of downstream signal transduction proteins.

(PDB files: 2AHX and 2A91. Reference: Burgess AW, Cho HS, Eigenbrot C, et al: An open-and-shut case? Recent insights into the activation of EGF/ErbB receptors. Mol Cell 12:541–552, 2003.)

The juxtaposition of kinase domains allows the partners to activate each other by direct interaction and by phosphorylating each other on tyrosine residues. In most cases, phosphorylation of tyrosines on the activation loop of the catalytic domain refolds the loop into an active conformation (see Fig. 25-3D), although this is not required for the EGF receptor. In all cases, the paired kinases phosphorylate tyrosines on inserts and C-terminal extensions of the kinase domain, creating phosphotyrosine-binding sites for downstream effector and adapter proteins with SH2 and PTB domains (Fig. 24-5; also see Figs. 27-6 and 27-7). Each SH2 and PTB domain binds preferentially to a certain phosphotyrosine site by virtue of a pocket that recognizes both phosphotyrosine adjacent residues (see Fig. 25-11). For example, each of five phosphotyrosines of the platelet-derived growth factor receptor favors binding of a different effector or adapter protein.

Receptor tyrosine kinases activate effector proteins in two different ways. First, binding of an effector protein to a receptor phosphotyrosine favors its phosphorylation by the receptor kinase. In the case of phospholipase C g, tyrosine phosphorylation both activates its catalytic activity and dissociates the enzyme from its phosphotyrosine-binding site, allowing it to move to its site of action on the membrane. Second, binding to the receptor may promote activity by bringing an effector protein near its substrate. This applies to both phosphoinositide-3-kinase, which acts on lipid substrates in the membrane bilayer (see Fig. 27-7), and the nucleotide exchange protein that activates the Ras guanosine triphosphatase (GTPase), which is anchored to the membrane bilayer (see Fig. 27-6).

Multiple mechanisms turn off receptor tyrosine kinases. In the short term, lipid second messengers produced by phospholipase Cg activate protein kinase C, which inhibits the kinase by phosphorylation. Active kinases are also targets for the addition of a single ubiquitin to one or more lysine side chains. These single ubiquitins are the signal for the receptor to be removed from the cell surface by endocytosis in clathrin-coated vesicles (see Fig. 23-2).

Mutations in receptor tyrosine kinases cause human disease. Many cancers have activating mutations or overexpression of EGF receptors. Activating mutations in a fibroblast growth factor receptor lead to a variety of congenital abnormalities of the skeleton, including a form of dwarfism and premature fusion of the sutures between the bones of the skull. Some of these mutations activate by promoting receptor dimerization through disulfide bonds or association of transmembrane helices. Others change ligand specificity.

Cytokine Receptors

Cytokines are a diverse family of polypeptide hormones and growth factors that regulate many cellular processes. Although they differ in detail, all cytokines are four-helix bundles. Pituitary growth hormone controls body growth and development of mammals, so loss of function receptor mutations cause one type of dwarfism. Erythropoietin regulates the proliferation and differentiation of red blood cell precursors (see Fig. 28-7). Interleukins modulate cells of the immune system, so loss of function receptor mutations result in deficient immune cells.

Cytokine receptors are homodimers or heterodimers, all with two extracellular fibronectin III domains that bind the ligand (Fig. 24-6). A single polypeptide segment, likely an α-helix, crosses the membrane. Cytoplasmic domains lack enzyme activity but bind one of several protein tyrosine kinases called JAKs (“just another kinase”). It has not been settled whether inactive, ligand-free receptors diffuse separately in the plane of the membrane or form dimers with widely separated transmembrane domains, as observed in crystals of erythropoietin receptor without ligand (Fig. 24-6C). This separation of cytoplasmic domains could explain the lack of activity of ligand-free dimers.

Figure 24-6 cytokine receptors. A, Domain architecture and coupled signaling components of selected cytokine receptors. (EPO, erythropoietin; GH, growth hormone; GM-CSF, granulocyte-monocyte colony-stimulating factor; IL, interleukin.) B–C, Atomic structures of erythropoietin receptors. B, Side view of a receptor with a synthetic ligand called EMP1 (green). C, Top view of a receptor without ligand. D1 and D2 are the fibronectin III domains. The pink helices are models of the transmembrane segments. Pink bars indicate the separation of transmembrane segments. Association of erythropoietin and growth hormone with their receptors is remarkable, as a single, small, asymmetrical protein ligand binds between two identical receptor subunits using different sites on each receptor subunit.

(A, Adapted from Wells JA, de Vos AM: Hematopoietic receptor complexes. Annu Rev Biochem 65:609–634, 1996. B–C, Reference: Livnah O, Stura EA, Johnson DL, et al: Crystallographic evidence for preformed dimers of erythropoietin receptor before ligand activation. Science 283:987–990, 1999. PDB files: 1EBP and 1ERN.)

Ligand binding activates cytokine receptors by bringing together two JAK kinases bound to the cytoplasmic domains—either by changing the conformation of preformed dimers or by inducing the dimerization of separate receptors (Fig. 24-7). A conformational change of a preformed dimer would explain how erythropoietin can activate cells with low concentrations of receptors on the cell surface. Close proximity allows JAKs to activate each other by transphosphorylation. The signal is then propagated to the nucleus when JAKs phosphorylate selected members of a family of transcription factors called STATs, which migrate to the nucleus to regulate gene expression (see Fig. 27-9).

Figure 24-7 cytokine receptor activation models. A, Ligand-induced dimerization of separate subunits. B, Ligand-induced conformational change in a preformed dimer. In either case, proximity of the cytoplasmic domains allows associated JAK tyrosine kinases to activate each other by transphosphorylation. See Fig. 27-9 for details.

(Reference: Remy I, Wilson IA, Michnick SW, et al: Erythropoietin receptor activation by a ligand-induced conformational change. Science 283:990–993, 1999.)

Receptor Serine/Threonine Kinases

A third class of growth factor receptors uses cytoplasmic serine/threonine kinase domains to transduce signals (Figs. 24-8 and 24-9). Dimeric protein ligands bring together two different types of receptor subunits to turn on kinase activity. Humans have genes for seven type I receptors and five type II receptors. Active receptors phosphorylate transcription factors called Smads, stimulating their movement from cytoplasm into the nucleus, where they regulate genes controlling cellular proliferation and differentiation (see Fig. 27-10).

Figure 24-8 receptor serine/threonine kinases. A, Drawing of domain architecture of activin and transforming growth factor-b (TGF-b) receptors. (TM, transmembrane domain.) B, Mechanism of activation of receptor serine/threonine kinases. Ligand binds two type II receptors (RII) and two type I receptors (RI). Within this hexameric complex, the type II receptor phosphorylates and activates the type I receptors, which in turn phosphorylate cytoplasmic transcription factors called Smads. Phosphorylated Smads move to the nucleus to activate particular genes. See Fig. 27-10 for details. C–D, Ribbon and space-filling models of BMP7 bound to type I and type II receptors. This model is based on crystal structures of dimeric BMP7 bound to two extracellular domains of the type II activin receptor and of BMP2 bound to two extracellular domains of BMP type I receptor.

(References: Greenwald J, Groppe J, Gray P, et al: The BMP7/ActRII extracellular domain complex provides new insights into the cooperative nature of receptor assembly. Mol Cell 11:605–617, 2003; and Shi Y, Massague J: Mechanisms of TGFb signaling from cell membranes to the nucleus. Cell 113:685–700, 2003. PDB files: 1LX5 and 1S4Y.)

Figure 24-9 guanylylcyclase receptors. A, Comparison of the domain architecture of the transmembrane atrial natriuretic factor (ANF) receptor and the cytoplasmic nitric oxide receptor. B, Ribbon model of the extracellular domains of a dimer ANF receptor showing how binding of ANF brings together the cytoplasmic domains, which activates the guanylylcyclase activity by an unknown mechanism.

(PDB file: 1JDN. Reference: He X-L, Chow DC, Martick MM, Garcia KC: Allosteric activation of a spring-loaded natriuretic peptide receptor dimer by hormone. Science 293:1657–1662, 2001.)

Humans have 40 proteins that bind receptor serine/threonine kinases, including transforming growth factor- b (TGF-b), activin, inhibins, and bone morphogenetic proteins. These dimeric growth factors are particularly important during embryonic development. Activin was discovered as a releasing factor for pituitary follicle-stimulating hormone but also has a strong influence on the differentiation of early embryonic cells into primitive germ layers. Bone morphogenetic proteins influence the differentiation of osteoblasts, which lay down bone matrix, among many other cells. TGF-b has two distinct effects. First, it inhibits proliferation of most adult cells. Mice with null mutations of one of their three TGF-b genes die with inflammation in multiple organs caused by excessive proliferation of lymphocytes. Second, TGF-b stimulates production of extracellular matrix, including collagen, proteoglycans, and adhesive glycoproteins (see Chapter 29). These proteins are essential for the development of organs. Overproduction of extracellular matrix is a common feature of many chronic inflammatory diseases, so inappropriate expression of TGF-b might provide a link between inflammation and the pathological fibrosis that scars diseased organs. On the other hand, loss of TGF-b receptors during progression of some tumors makes them unresponsive to growth inhibition by TGF-b and contributes, in part, to their ability to replicate autonomously.

Serine/threonine kinase receptors are composed of two types of subunits called type I receptors and type II receptors, which are present in small numbers on the cell surface. Single transmembrane sequences link the small ligand-binding domains to cytoplasmic serine/threonine kinase domains. Signal transduction requires the kinase activities of both subunits.

Four receptor subunits bind independently to a dimeric ligand, beginning with high-affinity binding of two type II receptors followed by lower-affinity interactions with two type I receptors (Fig. 24-8B). Within this complex, the constitutively active type II receptor kinases phosphorylate the cytoplasmic domain of the type I receptors on serine and threonine residues. This activates the type I receptor kinase, which phosphorylates cytoplasmic transcription factors called Smads. Phosphorylated Smads move to the nucleus, where they cooperate with other transcription factors to regulate gene expression (see Fig. 27-10).

In addition to these transducing receptors, cells have a greater abundance of another plasma membrane TGF-β-binding protein (type III receptors) that lacks signal transduction activity. A single transmembrane sequence links the large extracellular proteoglycan domain to a small cytoplasmic domain. This receptor may concentrate TGF-b on the cell surface. Even in the absence of TGF-b, type II receptors phosphorylate type III receptors, creating a binding site for β-arrestin and promoting endocytosis of both type II and type III receptors.

Guanylyl Cyclase Receptors

Animals have a family of cell surface receptors (Fig. 24-9) with intracellular domains that catalyze the formation of 3′–5′-cGMP from GTP. Vertebrates have at least seven guanylyl cyclase receptor isoforms; nematode worms have more than 25. The gases nitric oxide and carbon monoxide activate related cytoplasmic guanylyl cyclases that participate in additional signal transduction pathways (see Fig. 26-1). Regardless of its enzymatic origin, cGMP regulates the same targets: cGMP-gated ion channels (see Fig. 10-10), cGMP-stimulated protein kinases (see Fig. 25-4), and cyclic nucleotide phosphodiesterases (see Fig. 26-1).

All known ligands for guanylyl cyclase receptors are peptides, although the ligands for some receptors are not known. Most insights regarding function have come from knowledge about the ligands, the tissue distribution of receptors, and receptor gene disruptions, as highly specific inhibitors of the cell surface guanylyl cyclases are not available.

Guanylyl cyclase receptors are homodimers with novel, ligand-binding extracellular domains and two cytoplasmic domains—an enzymatically inactive kinase domain and a guanylyl cyclase domain (Fig. 24-9). The cyclase domain is closely related to adenylyl cyclases (see Fig. 26-2). In the absence of ligand, the extracellular domains hold the cytoplasmic domains apart. Ligand binding closes the cleft between the extracellular domains, pulling together the transmembrane and cytoplasmic domains in a way that stimulates guanylyl cyclase activity.

Guanylyl cyclase receptor A (GC-A) binds atrial natriuretic factor, a polypeptide hormone that is secreted mainly by the heart to control blood pressure. It stimulates excretion of salt and water by the kidney and dilates blood vessels. Mice with null mutations for GC-A have high blood pressure and enlarged hearts and fail to respond when overloaded with fluid and salt administered intravenously. Intestinal guanylyl cyclase receptor C (GC-C) binds bacterial enterotoxin, the mediator of fluid secretion in bacterial dysentery. The bacterial toxin mimics endogenous peptides of un-known function, which are secreted by various tissues but principally by the intestine. Mice with null mutations for GC-C are completely resistant to enterotoxin but have no apparent physiological defects. Guanylyl cyclase receptors E and F (GC-E, GC-F) are restricted to the eye; a null mutation for GC-E results in loss of cone visual receptor cells. Guanylyl cyclase receptor D is restricted to olfactory neuroepithelium. Sea urchin spermatozoa use a guanylyl cyclase receptor to respond to peptides secreted by eggs.

Tumor Necrosis Factor Receptor Family

TNF and its receptor are the prototypes for a diverse group of cell-signaling partners (Fig. 24-10). Lymphocytes produce three isoforms of TNF (also called lymphotoxin and cachectin), a trimeric lymphokine with many functions, including roles in shock and inflammation, protection from bacterial infections, killing tumor cells, and wasting in chronic disease. Mice with a genetic deletion of the lymphotoxin-a gene have no lymph nodes, so TNF participates in the development of the immune system. Other ligands for the TNF class of receptors are also trimers of subunits composed of β-strands, such as nerve growth factor and Fas-ligand. These ligands and receptors regulate many processes, including cell proliferation and death (see Fig. 46-17).

Figure 24-10 tumor necrosis factor receptor family. A, Domain architecture of a sample of members from the TNF receptor family. (LT-a, lymphotoxin-a; NGF, nerve growth factor.) Fas and Fas ligand are presented in Chapter 46. B, Atomic model of TNF bound to its receptor. TNF is a trimer of three identical b sandwich subunits arranged in a pear-like structure. The four extracellular cysteine-rich domains of the receptor grasp TNF-like prongs.

(A, Adapted with permission from Beutler B, vanHuffel C: Unraveling function in the TNF ligand and receptor family. Science 264:667–668, 1994. Copyright 1994 AAAS. B, Reference: Banner DW, D’Arcy A, Janes W, et al: Crystal structure of the soluble human 55 kD TNF receptor-human TNF beta complex: Implications for TNF receptor activation. Cell 73:431–445, 1993. PDB file: 1TNR.)

Human cells express two types of TNF receptors that bind the same ligands but generate different responses. The two receptors have similar ligand-binding domains coupled by single transmembrane segments to different cytoplasmic domains. The extracellular part of these receptors consists of four similar repeats of about 40 amino acids, each with six conserved cysteines (Fig. 24-10B). The cysteines form three disulfide bridges, arranged like the rungs of a ladder, to stabilize these small domains. In the absence of ligand, individual receptor subunits are presumed to diffuse independently in the plane of the membrane.

The structure of TNF bound to the extracellular domains of its receptor showed how the receptor works. Three finger-like receptors grasp one trimeric TNF molecule by binding along the interfaces between TNF subunits. The tapered shape of TNF brings together the transmembrane segments and cytoplasmic domains of three receptors. Something about this arrangement of TNF receptors activates a plasma membrane phospholipase that hydrolyzes sphingomyelin, producing the second messenger ceramide (see Fig. 26-11). Adapter proteins associated with active TNF receptors recruit protein kinases that alter gene expression by activating the transcription factor NF-kB (see Fig. 15-22). Other receptors with linear arrays of cysteine-rich subdomains, similar to the TNF receptor, also bind to multimeric ligands, so receptor activation by clustering their cytoplasmic domains might be a general theme. For example, Fas ligand triggers cell death by clustering Fas and activating a cascade of intracellular proteolysis (see Fig. 46-17).

TNF participates in the inflammation associated with autoimmune diseases such as rheumatoid arthritis. Intercepting TNF before it reaches its receptor is remarkably successful in blunting inflammation in these diseases. This is accomplished by injections of monoclonal antibodies to TNF or with constructs containing the extracellular domains of the TNF receptor.

Toll-Like Receptors

Metazoan organisms use a small family of receptors named Toll-like receptors (TLRs) to sense and respond to infection by a wide variety of microorganisms including viruses, bacteria, fungi, and protozoa. The Toll gene was discovered in Drosophila encoding a receptor that was first linked to dorsal-ventral polarity in early development and was later found to be required for resistance to fungal infections. Mammals have about a dozen TLRs (Fig. 24-11) that bind certain macromolecules associated with microorganisms: double-stranded RNA from viruses, flagellin from bacteria, lipopolysaccharide from the outer membrane of gram-negative bacteria, and zymosan from the cell walls of fungi. Dimeric TLRs on the plasma membrane of white blood cells (including lymphocytes) and antigen-processing cells called dendritic cells sense these foreign macromolecules and stimulate the cell to respond by secreting inflammatory mediators such as TNF and interleukin-1 and -6. TLR3 is located in endosomes, where it can bind double-stranded RNA released from viruses. The signaling pathway from TLRs to TNF involves cytoplasmic adapter proteins and kinases that activate transcription factors including NF-kB (see Fig. 15-22C). TNF and interleukins then alert distant cells to respond to the infection.

Figure 24-11 toll-like receptors. Most TLRs are homodimers or heterodimers, but this figure shows a ribbon diagram of a single receptor molecule assembled from crystal structures of different receptors. Ribbon diagram of the extracellular domain of TLR3, an endosomal TLR consisting of 23 leucine-rich repeats that binds double-stranded RNAs released from viruses. A transmembrane helix connects to the cytoplasmic TIR domain from TLR2, a receptor for bacterial lipoproteins. Ligand binding to receptor dimers initiates a signal that is transmitted through adapter proteins to kinases, which activate cytoplasmic transcription factors including NF-kB. NF-kB moves to the nucleus and stimulates expression of TNF and other inflammatory mediators.

(PDB file for the extracellular domain of TLR3: 1ZIW. PDB file for the cytoplasmic domain of TLR2: 1FYW.)

Notch Receptors

Components of the Delta/Notch signaling pathway have been identified by analysis of mutations affecting early development in flies and nematodes. Ligands are transmembrane proteins called Delta in flies and vertebrates and LAG-2 in worms. These ligands and Notch receptors regulate cellular fates during early embryonic development. Typically, cells expressing Delta interact with Notch receptors on adjacent cells to force the neighboring cells to chose a different fate than their own. The actual outcome depends on the context; in each tissue, Delta/Notch signals are integrated with the actions of other signaling pathways. As a general point, Delta/Notch signals tend to reinforce differences between cells in a particular tissue. For example, Delta on the earliest neurons directs adjacent cells to other fates. Defects in Delta or Notch result in excess neurons.

Genetic analysis established that Delta/Notch signaling is vital for animal development, but less is known about the mechanisms than those of the other receptors presented in this chapter. Some Delta is active as a cell surface protein that interacts locally with adjacent cells, but a matrix metalloproteinase (see Fig. 29-20) cleaves some Delta from the membrane, allowing it to act at a distance from its cell of origin. The receptors, called Notch (flies, vertebrates) or Lin-12 (worms), consist of several extracellular EGF-like domains and leucine-rich repeats, a single transmembrane span, and an intracellular region of ankyrin repeats that lacks any known enzyme activity. Notch is synthesized as a single polypeptide chain and is cleaved once before transport to the plasma membrane. The two polypeptides remain covalently associated, presumably by a disulfide bond. Cells carrying Delta activate Notch receptors on adjacent cells. This leads to proteolytic cleavage that frees the intracellular domains from the membrane. These cytoplasmic domains move into the nucleus and join a complex of proteins, including CSL, that activate transcription of certain genes.

Hedgehog Receptors

Genetic studies of Drosophila revealed a novel class of protein ligands, called Hedgehog, and two membrane proteins, called Patched and Smoothened, that are required for signal transduction during development. The Hedgehog receptor Patched consists of 12 transmembrane segments related to proton-driven bacterial antiporters, but the transported substrate, if any, is not known. Smoothened is an unusual seven-helix receptor, because it is constitutively active. Substoichiometric quantities of Patched inhibit the activity of Smoothened, perhaps by transporting out of the cell a ligand that binds and inhibits Smoothened. In flies, the Hedgehog pathway cooperates with the Wnt system (see Fig. 30-8) to establish boundaries between segments of the embryo and maintain a pool of stem cells.

Every aspect of this novel signaling pathway established new principles. A signal sequence guides Hedgehog protein into the secretory pathway, but before it reaches the cell surface, the protein cleaves itself in two pieces. The latter half of the protein carries out the cleavage reaction. This autocatalytic reaction also adds a molecule of covalently bound cholesterol to the new C-terminus of the first half of the protein, the domain with signaling activity. This was the first example of cholesterol being used for a posttranslational modification of a protein. Cholesterol and an N-terminal palmitic acid anchor the signaling domain to membranes and lipoprotein particles, which are secreted and act on cells up to 30 cell diameters distant from the source.

The Hedgehog signal transduction pathway is complicated, in part because activation is achieved by inactivating inhibitors. In the absence of Hedgehog, active Patched inhibits Smoothened. Hedgehog binding turns off Patched and relieves the inhibition of Smoothened (the first inactivation of an inhibitor in this pathway). Active Smoothened assembles a complex of several proteins that inhibits the proteolytic inactivation of a transcription factor called Ci (the second inactivation of an inhibitor in this pathway). Active Ci controls the expression of several genes required for cell fate specification and differentiation including Patched itself.

Vertebrate orthologs of the proteins that form the insect Hedgehog pathway have similar functions, regulating cellular differentiation in many tissues, including formation of the neural tube. Mutations in one of three mammalian Hedgehog genes (sonic hedgehog) cause widespread developmental defects that range from mild to grotesque, including a single eye in the middle of the face. Mutations in the gene for Patched cause basal cell carcinoma of the skin, the most common cancer in fair-skinned people. Human Smoothened is a proto-oncogene; activating mutations prevent its inhibition by Patched and cause skin tumors. The Ci ortholog Gli1 is an oncogene that was originally discovered in brain tumors.

ACKNOWLEDGMENTS

Thanks go to Senyon Choe, Dan Leahy, and Ruslan Medzhitov for their suggestions on revisions to this chapter.

APPENDIX 24-1 Receptors and Ligands

Receptors and Ligands