Drawing of a human fetus in utero, by Leonardo da
Vinci. Human sexuality and embryonic development
represent two hormonally regulated processes of
universal interest. (Florentine Royal Collection.
Windsor, England A.K.G., Berlin/Superstock, International)

SPECIAL TOPIC

Chapter 34

The Reception and Transmission of Extracellular Information

Higher life forms must have molecular mechanisms for detecting environmental information as well as mechanisms that allow for communication at the cell and tissue levels. Sensory systems detect and integrate physical and chemical information from the environment and pass this information along by the process of neurotransmission. Control and coordination of processes at the cell and tissue levels are achieved not only by neurotransmission but also by chemical signals in the form of hormones that are secreted by one set of cells to direct the activity of other cells. In this chapter, we will address these mechanisms of information transfer, beginning with the molecular basis of hormone action and then moving to a special focus on excitable membranes, neurotransmission, and sensory systems.
            Hormones are secreted by certain cells, usually located in glands, and travel, either by simple diffusion or circulation in the bloodstream, to specific target cells. As we shall see, some hormones bind to specialized receptors on the plasma membrane and induce responses within the cell without themselves entering the target cell. Other hormones actually enter the target cell and interact with specific receptors there. By these mechanisms, hormones regulate the metabolic processes of various organs and tissues; facilitate and control growth, differentiation, reproductive activities, and learning and memory; and help the organism cope with changing conditions and stresses in its environment.

Many different chemical species act as hormones. Steroid hormones, all derived from cholesterol, regulate metabolism, salt and water balances, inflammatory processes, and sexual function. Several hormones are amino acid derivatives. Among these are epinephrine and norepinephrine, which regulate smooth muscle contraction and relaxation, blood pressure, cardiac rate, and the processes of lipolysis and glycogenolysis, and the thyroid hormones, which stimulate metabolism. Peptide hormones are a large and still expanding group of hormones that appear to regulate processes in all body tissues, including the release of yet other hormones.
            Hormones and other signal molecules in biological systems bind with very high affinities to their receptors, displaying K_D values in the range of 10^-12 to 10^-6 M. The concentrations of hormones are maintained at levels equivalent to or slightly above these K_D values. Once hormonal effects have been induced, the hormone is usually rapidly metabolized.
            Hormonal regulation depends upon the transduction of the hormonal signal across the plasma membrane to specific intracellular sites, particularly the nucleus. Many steps in these signaling pathways involve phosphorylation of serine, threonine, and tyrosine residues on target proteins. The complexity of signal transduction is thus manifested in the estimates that the human genome may contain between 1000 and 3000 protein kinases. Although only a few of these kinases—and other signaling proteins and molecules—have been identified, the principles that govern these pathways are now becoming apparent (Chapter 15).
            Signaling pathways must operate with speed and precision, facilitating the accurate relay of intracellular signals to specific targets. But how does this happen? Many protein kinases and phosphatases have relatively broad specificities, but they depend upon interactions between specialized recognition domains on the involved proteins and also upon cellular strategies for localization of the signaling molecules. As we shall see later, modular proteins having one or more protein-protein or protein-lipid recognition domains mediate these events.

34.2 · Signal-Transducing Receptors Transmit the Hormonal Message

Very often in life, the message is more important than the messenger, and this is certainly true for hormones. The structure and chemical properties of a hormone are only important for specific binding of the hormone to its appropriate receptor. Of much greater interest and importance, however, is the metabolic information carried by the hormonal signal. The information implicit in the hormonal signal is interpreted by the cell, and an intricate pattern of cellular responses ensues.
   Figure 34.1 · Nonsteroid hormones bind exclusively to plasma membrane receptors, which mediate the cellular responses to the hormone. Steroid hormones exert their effects either by binding to plasma membrane receptors or by diffusing to the nucleus, where they modulate transcriptional events.

         Steroid hormones may either bind to plasma membrane receptors or exert their effects within target cells, entering the cell and migrating to their sites of action via specific cytoplasmic receptor proteins (Figure 34.1). The nonsteroid hormones, which act by binding to outward-facing plasma membrane receptors, activate various signal transduction pathways that mobilize various second messengers—cyclic nucleotides, Ca²⁺ ions, and other substances that activate or inhibit enzymes or cascades of enzymes in very specific ways. These hormonally activated processes are the focus of this chapter.

            All receptors that mediate transmembrane signaling processes fit into one of three receptor superfamilies:

1.  The 7-transmembrane segment (7-TMS) receptors are integral membrane proteins with seven transmembrane (helical) segments, an extracellular recognition site for ligands, and an intracellular recognition site for a GTP-binding protein (see following discussion).

2.  The single-transmembrane segment (1-TMS) catalytic receptors are proteins with only a single transmembrane segment and substantial globular domains on both the extracellular and intracellular faces of the membrane. The extracellular domain in the ligand recognition site and the intracellular catalytic domain is either a tyrosine kinase or a guanylyl cyclase.

3.  Oligomeric ion channels consist of associations of protein subunits, each of which contains several transmembrane segments. These oligomeric structures are ligand-gated ion channels. Binding of the specific ligand typically opens the ion channel. The ligands for these ion channels are neurotransmitters.

34.3 · Intracellular Second Messengers

Epinephrine and glucagon activate glycogen breakdown in the liver (Chapters 15 and 23), but the mechanism of this activation was a mystery until Earl Sutherland and his colleagues showed that the glycogen phosphorylase reaction (Figure 23.16), the initial step in glycogen breakdown, was stimulated by epinephrine and glucagon. Activation of phosphorylase was eventually shown to be due to an ATP-dependent phosphorylation of the enzyme. Sutherland also showed that a phosphatase from the liver cells inactivated phosphorylase, eliminating the activation due to phosphorylation. A significant breakthrough was achieved with Sutherland’s demonstration that hormones activated phosphorylase only in the presence of plasma membrane fragments. He hypothesized that binding of epinephrine or glucagon to a receptor in the membrane activated synthesis or release of a substance that activated phosphorylation of phosphorylase. This crucial substance was eventually shown (in 1956) to be adenosine 3',5'-cyclic monophosphate, now known as cyclic AMP, denoted cAMP (Figure 34.2).

Figure 34.2 · Cyclic AMP is synthesized by membrane-bound adenylyl cyclase and degraded by soluble phosphodiesterase.

Cyclic AMP is produced by an integral membrane enzyme, adenylyl cyclase (Figure 34.2). The strain and bond distortion inherent in the bicyclic ring structure of cAMP make its formation an endergonic process, but spontaneous hydrolysis of pyrophosphate formed in this reaction drives the synthesis of cAMP forward. The phosphodiesterase reaction hydrolyzes cAMP to AMP as shown. Because this reaction relieves the strain in the cAMP structure, it is highly exergonic (DG°' = 250.4 kJ/mol; see Table 3.1). Figure 34.3 · Earl Sutherland’s simple model of hormone action, circa 1967. Phosphorylase a was eventually shown to be a phosphorylated form of glycogen phosphorylase. Although Sutherland didn’t realize it, the interaction between the hormone receptor and adenylyl cyclase is not direct, but instead involves GTP-binding proteins.

Sutherland thought of the hormone as the first messenger signaling a need for glycogen breakdown, and he called cAMP a second messenger. The basic process, as Sutherland understood it in the 1960s, is shown in Figure 34.3. In this second messenger model, the hormone causes a variety of intracellular effects without itself entering the cell. It was eventually shown that many hormonally activated processes use cAMP as a second messenger. For his remarkable achievements in the elucidation of this model, Earl Sutherland received the Nobel Prize in physiology or medicine in 1971.
            Since Sutherland’s discovery of cAMP, many other second messengers have been identified (Table 34.1). The mediators of second messenger formation for all 7-TMS receptors are GTP-binding proteins.

34.4 · GTP-Binding Proteins: The Hormonal Missing Link

Figure 34.4 · The structure of guanylylimidodiphosphate.

Two observations in the early 1970s implicated another protein in the hormonal activation of adenylyl cyclase. First, purification of adenylyl cyclase and the hormone receptor resulted in a loss of hormone stimulation of cyclase activity. Second, Martin Rodbell and his colleagues showed that GTP was necessary for hormone activation of adenylyl cyclase. Interestingly, 5'-guanylylimidodiphosphate (GMP-PNP), a nonhydrolyzable analog of GTP (Figure 34.4), was a “superactivator” of adenylyl cyclase, giving higher cyclase activities than GTP itself. This prompted Rodbell to suggest that the GTP-binding site was the active site of a GTPase. In 1977, Elliott Ross and Alfred Gilman at the University of Virginia reported the partial purification of a GTP-binding protein, which, when reconstituted with the cyclase and hormone receptor, restored hormone stimulation to the adenylyl cyclase reaction. Thus, adenylyl cyclase is not directly activated by the hormone-receptor complex. Instead, binding of hormone to the receptor stimulates a GTP-binding protein (abbreviated now to G protein), which in turn activates adenylyl cyclase.

Figure 34.5 · Activation of adenylyl cyclase by heterotrimeric G proteins. Binding of hormone to its receptor causes a conformational change that induces the receptor to catalyze a replacement of GDP by GTP on G_a. The G_a(GTP) complex dissociates from G_bg and binds to adenylyl cyclase, stimulating synthesis of cAMP. Bound GTP is slowly hydrolyzed to GDP by the intrinsic GTPase activity of G_a. G_a(GDP) dissociates from adenylyl cyclase and reassociates with G_bg. G_a and G_g are lipid-anchored proteins. Adenylyl cyclase is an integral membrane protein consisting of 12 transmembrane a-helical segments.

Typically, G proteins are heterotrimers consisting of a- (45 to 47 kD), b- (35 kD), and g- (7 to 9 kD) subunits. The a-subunit binds GDP or GTP and has an intrinsic, slow GTPase activity. The Gabg complex in the unactivated state has GDP at the nucleotide site (Figure 34.5). Binding of hormone to receptor stimulates a rapid exchange of GTP for GDP on Ga. The binding of GTP causes Ga to dissociate from Gbg and to associate with an effector protein such as adenylyl cyclase. Binding of Ga (GTP) activates adenylyl cyclase. The adenylyl cyclase actively synthesizes cAMP as long as Ga(GTP) remains bound to it. However, the intrinsic GTPase activity of Ga eventually hydrolyzes GTP to GDP, leading to dissociation of Ga(GDP) from adenylyl cyclase and reassociation with the Gbg dimer, regenerating the inactive heterotrimeric Gabg complex.
Figure 34.6 · (a) Adenylyl cyclase activity is modulated by the interplay of stimulatory (G_s) and inhibitory (G_i) G proteins. Binding of hormones to b₁- and b₂-adrenergic receptors activates adenylyl cyclase via G_s, whereas hormone binding to a₂ receptors leads to the inhibition of adenylyl cyclase. Inhibition may occur by direct inhibition of cyclase activity by G_ia or by binding of G_ibg to G_sa(GTP). (b) Two views of the complex of the VC₁-IIC₂ catalytic domain of adenylyl cyclase and G_sa (c) Details of the G_sa complex in the same orientation as the structures above. (Courtesy of Alfred Gilman, University of Texas Southwestern Medical Center.)

            Two stages of amplification occur in the G-protein-mediated hormone response. First, a single hormone-receptor complex can activate many G proteins before the hormone dissociates from the receptor. Second, and more obvious, the Ga-activated adenylyl cyclase synthesizes many cAMP molecules. Thus, binding of hormone to a very small number of membrane receptors stimulates a large increase in concentration of cAMP within the cell. The hormone receptor, G protein, and cyclase constitute a complete hormone signal transduction unit (Figure 34.6a).
            A given G protein can be activated by several different hormone-receptor complexes. For example, either glucagon or epinephrine, binding to their distinctive receptor proteins, can activate the same species of G protein in liver cells. The effects are additive, and stimulation by glucagon and epinephrine together leads to higher cytoplasmic concentrations of cAMP than activation by either hormone alone.
            G proteins are a universal means of signal transduction in higher organisms, activating many hormone-receptor-initiated cellular processes in addition to adenylyl cyclase. Such processes include, but are not limited to, activation of phospholipases C and A₂, and the opening or closing of transmembrane channels for K⁺, Na⁺, and Ca²⁺ in brain, muscle, heart, and other organs (Table 34.2). G proteins are integral components of sensory pathways such as vision and olfaction. More than 100 different G protein-coupled receptors and at least 21 distinct G proteins are known. At least a dozen different G protein effectors have been identified, including a variety of enzymes and ion channels.

Hormone-receptor-mediated processes regulated by G proteins may be stimulatory (as we have seen) or inhibitory. Each hormone receptor protein interacts specifically with either a stimulatory G protein, denoted G_s, or with an inhibitory G protein, denoted G_i. For example, the catecholamine hormones (such as epinephrine) bind to four different types of adrenergic receptor proteins, designated a₁, a₂, b₁, or b₂. Binding to a₁ receptors has no effect on adenylyl cyclase, whereas binding to b₁ or b₂ receptors is stimulatory and binding to a₂ receptors is inhibitory. The b₁ and b₂ receptors interact specifically with a G_s complex, and a₂ receptors interact only with G_i proteins. The “G protein” described in Figure 34.5 is in fact a G_s-type G protein. Two possibilities exist (Figure 34.6a). Binding of a hormone to its receptor triggers GDP^zyGTP exchange and dissociation of G_{ia(GTP) from G}ibg. Inhibition may then occur either by binding of G_ia_{(GTP) to adenylyl cyclase to directly inhibit the cyclase, or by action of G}ibg, which can compete with the cyclase for G_sa_{(GTP) complexes. The presence in liver cell membranes of much more G}i than G_s favors the competitive role for G_ibg.
            Alfred Gilman, Stephen Sprang, and coworkers have determined the structure of a complex of G_sa _{(with bound GTP) with the cytoplasmic domains (VC}1 and IIC₂) of adenylyl cyclase (Figure 34.6b). The G_sa _{complex binds to a cleft at one corner of the C}2 domain, and the surface of G_sa_{-GTP that contacts adenylyl cyclase is the same surface that binds the G}bg dimer. The catalytic site, where ATP is converted to cyclic AMP, is far removed from the bound G protein.

Figure 34.7 · ADP-ribosylation of G_sa by cholera toxin.

Many of the details of G protein mediation of hormonal effects have been elucidated through studies of the potent effects of bacterial toxins, including cholera toxin and pertussis toxin. Vibrio cholerae, the Gram-negative bacterium that causes cholera, induces severe diarrhea in its victims, leading to death if the fluids are not replenished. Cholera toxin is an 87-kD protein consisting of an A subunit and five B subunits. The B subunits act in host-cell recognition. The A subunit consists of A₁ and A₂ peptides linked by a disulfide bond. The 22-kD A₁ peptide catalyzes the ADP-ribosylation of Arg²⁰¹ in the a-subunit of G_s (Figure 34.7), using NAD⁺ as a substrate. ADP-ribosylation strongly inhibits the GTPase activity of G_sa_{, effectively trapping G}sa in the activated state and causing prolonged activation of adenylyl cyclase. Elevated levels of cyclic AMP in turn cause intestinal epithelial cells to secrete high volumes of fluid. Both the bacteria and the cholera toxin remain localized in the intestines through the course of the disease, but if fluids are actively replaced, the body’s immune system eventually gains control and destroys the bacteria.

ADP-ribosylation of a Cys residue on G_ia _{is catalyzed by pertussis toxin, a product of Bordetella pertussis, the bacterium that causes whooping cough. Pertussis toxin is a 110-kD hexameric protein (a 28-kD A subunit and five B-like subunits.) In this case, the ADP-ribosylation inhibits exchange of GDP for GTP, thereby preventing G}ia from inhibiting adenylyl cyclase. Pertussis, in contrast to cholera, is a systemic infection, and the breakdown of adenylyl cyclase regulation is felt by tissues throughout the body.

Figure 34.8 · The structure of ras p21-GTP complexed with (a) GDP and (b) GMP-PNP. The ras p21-GMP-PNP complex is the active conformation of this protein.

GTP-binding proteins are implicated in growth control mechanisms in higher organisms. Certain tumor virus genomes contain genes encoding 21-kD proteins that bind GTP and show regions of homology with other G proteins. The first of these genes to be identified was found in rat sarcoma virus, and was dubbed the ras gene. Genes implicated in tumor formation are known as oncogenes; often they are mutated versions of normal, noncancerous genes involved in growth regulation, so-called proto-oncogenes. The normal, cellular ras protein is a GTP-binding protein that functions similarly to other G proteins described above, activating metabolic processes when GTP is bound and becoming inactive when GTP is hydrolyzed to GDP. The GTPase activity of the normal ras p21 is very low, as appropriate for a G protein that regulates long-term effects like growth and differentiation. A specific GTPase-activating protein (GAP) increases the GTPase activity of the ras protein. Mutant (oncogenic) ras proteins have severely impaired GTPase activity, which apparently causes serious alterations of cellular growth and metabolism in tumor cells. Several crystallographic studies have determined the structures of ras proteins (Figure 34.8) in complexes with GDP, GMP-PNP, and GMP-PCP (nonhydrolyzable analogs of GTP in which the b-P and g-P are linked by N or C rather than by O). ras consists of a six-stranded b-sheet surrounded by a-helices. The five peptide loops (G-1 through G-5) that comprise the guanine nucleotide binding site are the most highly conserved elements in this structure. Two regions of the ras structure change conformation upon GTP hydrolysis. One of these, designated switch I, corresponds to the G-2 loop that forms part of the Mg²⁺-binding site. This segment plays a role in binding of GAP proteins (see A Deeper Look) and ras target proteins and is therefore called the effector loop. The second flexible element, switch II, includes G-3 (which comprises the GTP g-P binding site) and the a2 helix that follows.
Figure 34.9 · A complex between raps (a ras analog, green) and the ras-binding domain of c-raf-1, purple.

            A glimpse of the molecular interactions between a ras protein and an effector (target) protein is revealed in the structure (Figure 34.9) of a complex between a ras analog called raps and the ras-binding domain from a ras effector known as c-raf-1. The principal interactions in this complex involve an interprotein b-sheet formed by two antiparallel strands from the two proteins. GTP hydrolysis induces a coil-to-helix transition at the N-terminus of the switch II domain and a subsequent reorientation of the entire a2 helix, which destroys the effector-binding site.

A Deeper Look
RGSs and GAPs—Switches That Turn Off G Proteins
Nature has made ras p21 and Gsa very poor enzymes by design. For example, ras p21 hydrolyzes GTP with a rate constant of only 0.02 min21. These G proteins are active only in the GTP-bound state, and downstream targets will dissociate upon GTP hydrolysis. If ras p21 and Gsa were efficient enzymes, the GTP-bound state would be short-lived, and G protein-mediated signaling would be ineffective.             But how can G proteins be switched off if they are inherently poor GTPases? The answer is provided by GTPase-activating proteins (GAPs) and regulators of G protein signaling (RGS), which cause dramatic increases in GTPase activity when bound to G proteins.	The left figure shows ras p21 (white) with a fragment of a GAP (blue) bound to it. GAPs increase the GTPase activity of ras p21 by a factor of 105. The right figure shows RGS (blue) bound to Gai (yellow). RGS proteins accelerate Gsa-catalyzed GTP hydrolysis by nearly 100-fold. In both ras p21 and Gsa, GTPase activity and the conversion from GTP-bound to GDP-bound forms of the protein involve conformation changes in the switch regions, portions of the G protein structure that surround the GTP/GDP binding site. GAPs and RGS increase GTPase activities of their respective G proteins by binding near the active site and stabilizing the transition state of the GTP hydrolysis reaction.

34.5 · The 7-TMS Receptors

Figure 34.10 · The arrangements of the b₂- and a₂-adrenergic receptors in the membrane. Substitution of Asp¹¹³ in the third hydrophobic domain of the b-adrenergic receptor with an Asn or Gln by site-directed mutagenesis results in a dramatic decrease in affinity of the receptor for both agonists and antagonists. Significantly, this Asp residue is conserved in all other G protein-coupled receptors that bind biogenic amines, but is absent in receptors whose ligands are not amines. Asp¹¹³ appears to be the counter-ion for the amine moiety of adrenergic ligands.

The primary and secondary structures of the 7-transmembrane segment (7-TMS) receptors are similar to those of bacteriorhodopsin (see Chapter 9) and rhodopsin. The a- and b-adrenergic receptors, for which epinephrine is a ligand, are good examples (Figure 34.10). Hydropathy analysis of these receptors is consistent with seven transmembrane a-helical segments. The extracellular N-terminal segment has two glycosylation sites. The hydrophilic loops connecting the seven hydrophobic domains are not required for ligand binding. Instead, the ligand-binding site (for the cationic catecholamines) is located within the hydrophobic core of the receptor.
            The binding of epinephrine to a b-adrenergic receptor initiates the above-described G protein activation of adenylyl cyclase. a₁-Adrenergic receptors stimulate inositol phospholipid metabolism when activated (see following section). Stimulation of a₂-adrenergic receptors appears to counteract hormone-stimulated increases in [cAMP]. The b-adrenergic receptors act through G proteins. These G proteins are coupled to several pathways, including adenylyl and guanylyl cyclases, phospholipases A and C, Ca²⁺ and K⁺ channels, and phosphodiesterases.
            The b-adrenergic receptor is desensitized by phosphorylation either by a specific b-adrenergic receptor kinase (bARK) or by protein kinase A (PKA), the cAMP-dependent protein kinase. Phosphorylation sites for both kinases are located in the receptors’ C-terminal domain. The sites phosphorylated by protein kinase A are adjacent to segments of the receptor that mediate coupling to G proteins, implying that phosphorylation at these sites interrupts this coupling. The desensitization due to bARK occurs by a different mechanism; it may involve another as yet unidentified protein.

34.6 · Specific Phospholipases Release Second Messengers

Figure 34.11 · (a) The general action of phospholipase A₂ (PLA₂), phospholipase C (PLC), and phospholipase D (PLD). (b) The synthesis of second messengers from phospholipids by the action of phospholipases and sphingomyelinase.

A diverse array of second messengers are generated by breakdown of membrane phospholipids. Binding of certain hormones and growth factors to their respective receptors triggers a sequence of events that can lead to the activation of specific phospholipases. Action of these phospholipases on membrane lipids produces the second messengers shown in Figure 34.11.

Figure 34.12 · The family of second messengers produced by phosphorylation and breakdown of phosphatidylinositol. PLC action instigates a bifurcating pathway culminating in two distinct and independent second messengers: DAG and IP₃.

Breakdown of phosphatidylinositol (PI) and its derivatives by phospholipase C produces a family of second messengers. In the best-understood pathway, successive phosphorylations of PI produce phosphatidylinositol-4-P (PIP) and phosphatidylinositol-4,5-bisphosphate (PIP₂) (Figure 34.12). Four isozymes of phospholipase C (denoted a, b, g, and d) hydrolyze PI, PIP, and PIP₂. Hydrolysis of PIP₂ by phospholipase C yields the second messenger inositol-1,4,5-trisphosphate (IP₃), as well as another second messenger, diacylglycerol (DAG). IP₃ is water-soluble and diffuses to intracellular organelles where release of Ca²⁺ is activated. DAG, on the other hand, is lipophilic and remains in the plasma membrane where it activates a Ca²⁺-dependent protein kinase known as protein kinase C (see following discussion).

Activation of Phospholipase C Is Mediated by G Proteins or by Tyrosine Kinases

Figure 34.13 · Phospholipase C-b is activated specifically by G_q, a GTP-binding protein, and also by Ca²⁺.

The different phospholipase C isozymes are activated by different intracellular events (Figure 34.13). Phospholipase C-b, -g, and -d are all Ca²⁺-dependent. In addition, phospholipase C-b is stimulated by a class of G proteins known as G_q. Binding of polypeptide hormones such as vasopressin or bradykinin to 7-TMS receptors releases G_{qa(GTP) from a G}qabg trimer. In turn, G_qa_{(GTP) activates phospholipase C-b (Figure 34.13). On the other hand, phospholipase C-g is activated by receptor tyrosine kinases (Figure}

Figure 34.14 · Phospholipase C-g is activated by receptor tyrosine kinases and by Ca²⁺.

_{34.14). The primary structures of phospholipase C-b and -g are shown in Figure 34.15. The X and Y domains of phospholipase C-b and -g are highly homologous, and both of these domains are required for phospholipase C activation. The other}

Figure 34.15 · The amino acid sequences of phospholipase C isozymes b, g, and d share two homologous domains, denoted X and Y. The sequence of g isozyme contains src homology domains, denoted SH2 and SH3. SH2 domains (approximately 100 residues in length) interact with phosphotyrosine-containing proteins (such as RTKs), whereas SH3 domains mediate interactions with cytoskeletal proteins. (Adapted from Dennis, E., Rhee, S., Gillah, M., and Hannun, E., 1991. Role of phospholipases in generating lipid second messengers in signal transduction. The FASEB Journal 5:2068-2077.)

_{domains of these isozymes confer specificity for G protein activation or tyrosine kinase activation.}

The Metabolism of Inositol-Derived Second Messengers

Figure 34.16 · The pathways of phosphoinositide biosynthesis and metabolism. The Li¹-sensitive steps are shown as red lines.

A summary of PIP₂-related signaling processes is shown in Figure 34.16. IP₃ has a cellular half-life of only a few seconds. It is rapidly processed along two principal paths: (a) it can be catabolized by a series of phosphatases to yield inositol-1,4-bisphosphate, inositol-4-phosphate, and free myo-inositol, which can be subsequently reincorporated into new inositol phospholipids; or (b) it can be phosphorylated to yield inositol-1,3,4,5-tetraphosphate, which then undergoes a complex series of phosphorylations and dephosphorylations to at least six other inositol phosphate compounds. Interestingly, many of these inositol phosphates have also been shown to behave as second messengers in various cellular processes. Some of these compounds may serve as extracellular signals in the regulation of specific neural mechanisms.

Phosphatidylcholine, Sphingomyelin, and Glycosphingolipids Also Generate Second Messengers

In addition to PI, other phospholipids serve as sources of second messengers. Breakdown of phosphatidylcholine by phospholipases yields a variety of second messengers, including diacylglycerol, phosphatidic acid, and prostaglan-dins. The action of sphingomyelinase on sphingomyelin produces ceramide, which stimulates ceramide-activated protein kinase. Similarly, gangliosides (such as ganglioside G_M3 (Chapter 8)) and their breakdown products modulate the activity of protein kinases and G protein-coupled receptors.

34.7 · Calcium as a Second Messenger

Figure 34.17 · Cytoplasmic [Ca²⁺] increases occur via the opening of Ca²⁺ channels in the membranes of calciosomes, the endoplasmic reticulum, and the plasma membrane.

Calcium ion is an important intracellular signal. Binding of certain hormones and signal molecules to plasma membrane receptors can cause transient increases in cytoplasmic Ca²⁺ levels, which in turn can activate a wide variety of enzymatic processes, including smooth muscle contraction, exocytosis, and glycogen metabolism. (Most of these activation processes depend on special Ca²⁺-binding proteins discussed in the following section.) Cytoplasmic [Ca²⁺] can be increased in two ways (Figure 34.17). As mentioned briefly earlier, cAMP can activate the opening of plasma membrane Ca²⁺ channels, allowing extracellular Ca²⁺ to stream in. On the other hand, cells also contain intracellular reservoirs of Ca²⁺, within the endoplasmic reticulum and calciosomes, small membrane vesicles that are similar in some ways to muscle sarcoplasmic reticulum. These special intracellular Ca²⁺ stores are not released by cAMP. They respond to IP₃, a second messenger derived from phosphatidylinositol (PI).

Calcium-Induced Calcium Release

Figure 34.18 · IP₃ mediates Ca²⁺-induced Ca²⁺ release. Binding of IP₃ to the ER IP₃ receptor opens ER Ca²⁺ channels. Flow of Ca²⁺ through these channels induces a conformational change that opens plasma membrane Ca²⁺ channels. (Adapted from Berridge, M., 1990. Calcium oscillations. Journal of Biological Chemistry 265:9583-9586.)

Much of the Ca²⁺ entering the cytoplasm by action of IP₃ appears to come from two sources: (a) parts of the endoplasmic reticulum that are closely associated with the plasma membrane, and (b) the extracellular environment. As shown in Figure 34.18, Ca²⁺ release appears to be a two-step process. IP₃ binding to receptors on the ER membrane opens Ca²⁺ channels releasing Ca²⁺ from the ER. Released Ca²⁺ binds to the ER IP₃ receptor, causing a conformational change that induces the opening of adjacent plasma membrane Ca²⁺ channels. This conformational coupling between a receptor on one membrane and a Ca²⁺ channel on an adjacent membrane is remarkably similar to the operation of the ryanodine receptor in sarcoplasmic reticulum (Chapter 17), and the IP₃ receptor shares remarkable structural similarities with the ryanodine receptor.
Figure 34.19 · IP₃-mediated signal transduction pathways. Increased [Ca²⁺] activates protein kinases, which phosphorylate target proteins. Ca²⁺/CaM represents calci-calmodulin (Ca²⁺ complexed with the regulatory protein calmodulin).

        The actions of the two second messengers IP₃ and DAG are complementary (Figure 34.19). IP₃ elevates cytoplasmic Ca²⁺ levels, and DAG activates protein kinase C in a Ca²⁺- and phosphatidylserine-dependent manner.

Calcium Oscillations

Figure 34.20 · Induced oscillations of cytoplasmic [Ca²⁺]. (a) Carbachol-dependent oscillations in the parotid gland. (b) Response of hepatocytes to norepinephrine. (c) Endothelial cell response to histamine. (Adapted from Berridge, M., 1990. Calcium oscillations.Journal of Biological Chemistry 265:9583-9586.)

One of the most intriguing and exciting developments in the field of Ca²⁺ regulation has been the discovery by Michael Berridge and others that the increases in intracellular [Ca²⁺] induced by IP₃ are oscillatory in nature! Several examples of Ca²⁺ oscillatory behavior are shown in Figure 34.20, in-cluding patterns of near-sinusoidal oscillations and baseline spikes. Ca²⁺ oscillations are induced most often by activation of receptors that act through the phosphoinositide pathway. Examples include a₁-adrenergic receptors, vasopressin and angiotensin receptors in liver, histamine receptors in endothelial cells, cholecystokinin receptors in pancreatic cells, and B₂ bradykinin receptors in chromaffin cells.
        Several models have been proposed to account for oscillatory behavior, including models involving oscillations in IP₃ and models in which a constant IP₃ level induces fluctuations in the uptake and release of Ca²⁺. The purpose of Ca²⁺ oscillations in cells is not understood, but two possibilities include (a) the need to protect sensitive intracellular processes from prolonged high levels of Ca²⁺, and (b) the need to create spatial “waves” of Ca²⁺ in the cell. Many Ca²⁺ oscillatory systems display a spatial organization, so that Ca²⁺ transients such as those of Figure 34.20 actually spread through a cell in the form of a wave of Ca²⁺ that propagates at a rate of 10 to 100 mm/sec. There is also evidence that Ca²⁺ waves may spread from one cell to the next. For example, activated cells in the ciliated epithelium of the lung excite neighboring cells through a signal, thought to be Ca²⁺, that radiates outward in a wavelike manner at approximately 10 mm/sec.

Intracellular Calcium-Binding Proteins

Given the central importance of Ca²⁺ as an intracellular messenger, it should not be surprising that complex mechanisms exist in cells to manage and control Ca²⁺. When Ca²⁺ signals are generated by cAMP, IP₃, and other agents, these signals are translated into the desired intracellular responses by calcium-binding proteins, which in turn regulate many cellular processes. One of these, protein kinase C, is described in Section 34.8. The other important Ca²⁺-binding proteins can, for the most part, be divided into two groups on the basis of structure and function: (a) the calcium-modulated proteins, including calmodulin, parvalbumin, troponin C, and many others, all of which have in common a structural feature called the EF hand (Figure 6.26 and Figure 17.30), and (b) the annexin proteins, a family of homologous proteins that interact with membranes and phospholipids in a Ca²⁺-dependent manner.
    

Figure 34.21 · (a) Structure of uncomplexed calmodulin (gold). Calmodulin, with four Ca²⁺-binding domains, forms a dumbbell-shaped structure with two globular domains joined by an extended, central helix. Each globular domain juxtaposes two Ca²⁺-binding EF hand domains. An intriguing feature of these EF hand domains is their nearly identical three-dimensional structure, despite a relatively low degree of sequence homology (only 25% in some cases). (b, c) Complex of calmodulin (gold) with a peptide from myosin light chain kinase (white); (b) side view; (c) top view.

    More than 170 calcium-modulated proteins are known (Table 34.3). All possess a characteristic peptide domain consisting of a short a-helix, a loop of 12 amino acids, and a second a-helix. Robert Kretsinger at the University of Virginia initially discovered this pattern in parvalbumin, a protein first identified in the carp fish and later in neurons possessing a high firing rate and a high oxidative metabolism. Kretsinger lettered the six helices of parvalbumin A through F. He noticed that the E and F helices, joined by a loop, resembled the thumb and forefinger of a right hand (Figure 6.26), and named this structure the EF hand, a name in common use today to identify the helix-loop-helix motif in calcium-binding proteins. In the EF hand, Ca²⁺ is coordinated by six carboxyl oxygens contributed by a glutamate and three aspartates, by a carbonyl oxygen from a peptide bond, and by the oxygen of a coordinated water molecule. The EF hand was subsequently identified in calmodulin, troponin C, and calbindin-9K (Figure 34.21). Most of the known EF-hand proteins possess two or more (as many as eight) EF-hand domains, usually arranged so that two EF-hand domains may directly contact each other.

Calmodulin Target Proteins Possess a Basic Amphiphilic Helix

Figure 34.22 · Helical wheel representations of (a) a model peptide, Ac-WKKLLKLLKKLLKL-CONH₂, and (b) the calmodulin-binding domain of spectrin. Positively charged and polar residues are indicated in green and hydrophobic residues are orange. (Adapted from O’Neil, K., and DeGrado, W., 1990. How calmodulin binds its targets: Sequence independent recognition of amphiphilic a-helices. Trends in Biochemical Sciences 15:59-64.)

Circular dichroism measurements show that the conformations of EF-hand proteins change dramatically upon binding of Ca²⁺ ions. This change promotes binding of the EF-hand protein with its target protein(s). For example, calmodulin (CaM), a 148-residue protein found in many cell types, modulates the activities of a large number of target proteins, including Ca²⁺-ATPases, protein kinases, phosphodiesterases, and NAD⁺ kinase, as well as several proteins involved in intracellular motility. CaM binds to these and to many other proteins with extremely high affinities (K_D values typically in the high picomolar to low nanomolar range). All CaM target proteins possess a basic amphiphilic alpha helix (a Baa helix), to which CaM binds specifically and with high affinity. Viewed end-on, in the so-called helical wheel representation (Figure 34.22), a Baa helix has mostly hydrophobic residues on one face; basic residues are collected on the opposite face. However, the Baa helices of CaM target proteins, although conforming to the model, show extreme variability in sequence. How does CaM, itself a highly conserved protein, accommodate such variety of sequence and structure? Each globular domain consists of a large hydrophobic surface flanked by regions of highly negative electrostatic potential—a surface suitable for interacting with a Baa helix. The long central helix joining the two globular regions behaves as a long, flexible tether. When the target protein is bound, the two globular domains fold together, forming a single binding site for target peptides (Figure 34.21b). The flexible nature of the tethering helix allows the two globular domains to adjust their orientation synergistically for maximal binding of the target protein or peptide.

34.8 · Protein Kinase C Transduces the Signals of Two Second Messengers

Protein kinase C (PKC) elicits a variety of cellular responses by phosphorylation of various target proteins at Ser and Thr residues. PKC is specifically activated by two intracellular second messengers: diacylglycerol and Ca²⁺ (the “C” in PKC stands for Ca²⁺). Because Ca²⁺ levels increase in the cell in response to IP₃, the activation of PKC depends upon both of the second messengers released by hydrolysis of PIP₂. PKC is a cellular transducer, translating the hormonal message and the signals of second messengers into the protein phosphorylation events that control growth and development.
Figure 34.23 · The primary structure of the a-isozyme of protein kinase C. V indicates variable regions; C indicates conserved domains. The phorbol-binding domain is identical with the DAG-binding domain.

        Protein kinase C (PKC) is an 80-kD polypeptide with four conserved domains and five variable regions. The conserved regions include an ATP-binding domain, a substrate-binding domain, a calcium-binding domain, and a DAG-binding domain (Figure 34.23). The DAG-binding domain is often referred to as the “pseudosubstrate domain” because it has an amino acid sequence that closely resembles protein substrates for the enzyme. With somewhat variable sequence, there are at least eleven different members of the protein kinase C family, which probably have separate and distinguishable functions. At low levels of Ca²⁺ and in the absence of DAG, protein kinase C is inactive and is a soluble protein in the cytoplasm. In this state, the pseudosubstrate domain occupies the substrate-binding site, keeping the enzyme inactive (Figure 34.24). DAG binding causes conformation changes that dissociate the pseudosubstrate domain from the substrate-binding site and increase the affinity of the enzyme for Ca²⁺ and lipid, causing protein kinase C domain C₁ to bind to the cytoplasmic surface of the plasma membrane, whereupon the kinase becomes active. Control of enzymatic activity by the insertion of a pseudosubstrate domain into the active site has been referred to as intrasteric control, in contrast to allosteric control, in which an enzyme regulator with a structure unrelated to the substrate binds at a site separate from the active site. Many protein kinases and protein phosphatases are regulated by intrasteric control (Chapter 15).

Figure 34.24 · A model for Ca²⁺- and DAG-dependent activation of protein kinase C. C₁ through C₄ refer to domains indicated in Figure 34.23. (Adapted from Sando, J. J., Maurer, M. C., ^{Bolen, E. J., and Grisham, C. M., 1992. Role of cofactors in protein kinase C activation. Cellular Signalling 4:595-609.)}

Figure 34.25 · The structure of a phorbol ester. Long-chain fatty acids predominate at the 12-position, whereas acetate is usually found at the 13-position.

        PKC phosphorylates serine and threonine residues on a wide range of protein substrates. A role for protein kinase C in cellular growth and division is demonstrated by its strong activation by phorbol esters (Figure 34.25). These compounds, from the seeds of Croton tiglium, are tumor promoters, agents that do not themselves cause tumorigenesis but that potentiate the effects of carcinogens. The phorbol esters mimic DAG, bind to the regulatory pseudosubstrate domain of the enzyme, and activate protein kinase C.

Cellular Target Proteins Are Dephosphorylated by Phosphoprotein Phosphatases

Along with the growing appreciation of the importance of protein phosphor-ylation in the response of cells to hormones, growth factors, and other cellular control signals has come an equal appreciation for the roles of phosphoprotein phosphatases. Many phosphoprotein phosphatases, specific either for serine/threonine phosphates or for tyrosine phosphates, have now been characterized. Four different classes of serine/threonine phosphatases are known. Type 1 phosphoprotein phosphatases (PP1) dephosphorylate the b-subunit of phosphorylase kinase. Phosphoprotein phosphatase 1 is inhibited by nanomolar concentrations of two heat-stable proteins, inhibitor-1 and inhibitor-2. The other three classes of serine/threonine phosphatases are the type 2 phosphatases, and they are designated PP2A, PP2B, and PP2C. These enzymes are distinguished in part by their sensitivities to divalent cations. PP1 and PP2A do not require divalent cations, whereas PP2B is dependent on Ca²⁺ and calmodulin, and PP2C requires Mg²⁺.
        The phosphatases PP2A and PP2C are predominantly cytosolic, but PP1 binds to membranes within the cell. Evidence is accumulating that PP1 (and possibly other serine/threonine phosphatases) interacts with specialized regulatory subunits that target the enzyme to particular locations in the cell and enhance its activity toward selected substrates.

A Deeper Look
Okadaic Acid: A Marine Toxin and Tumor Promoter
A novel agent for evaluating the serine/threonine phosphatases is okadaic acid (see figure), a complex fatty acid polyketal produced by marine dinoflagellates (plankton) that accumulates in the digestive glands of shellfish and marine sponges such as Halichondria okadaii, from which it gets its name. Okadaic acid is the major cause of diarrhetic shellfish poisoning, and it is a potent tumor promoter. It is a specific inhibitor of PP1 and PP2A, but has no effect on other phosphatases or on protein kinases, including protein kinase C. Because PP1 and PP2A are the enzymes that reverse the actions of protein kinase C, it is no surprise that okadaic acid is as potent a tumor promoter as the phorbol esters. Okadaic acid is a hydrophobic molecule and easily enters cells, where it causes prolonged phosphorylation of	many cellular proteins. In so doing, it causes long-lasting contraction of vascular smooth muscle and also mimics the stimulatory behavior of insulin on glucose metabolism. The role of okadaic acid in diarrhetic shellfish poisoning appears to be related to the diarrhetic action of cholera toxin, which activates adenylyl cyclase, and the cAMP-dependent protein kinase, which phosphorylates proteins that control sodium secretion by intestinal cells. Okadaic acid probably causes diarrhea by stimulating the phosphorylation of the same sodium channel-regulating proteins. Another toxin with similar properties is calyculin A, isolated from the sponge  Discodermia calyx. It is also a strong inhibitor of PP1 and PP2 and a potent tumor promoter, although its structure differs significantly from that of okadaic acid.
Structures of okadaic acid, calyculin A, and related toxins.

34.9 · The Single-TMS Receptors

Two principal classes of hormone receptors display intrinsic enzymatic activity: receptor tyrosine kinases and receptor guanylyl cyclases. Interestingly, each of these enzyme activities is manifested in two different cellular forms. Thus, guanylyl cyclase activity is found both in membrane-bound receptors and in soluble, cytoplasmic proteins. Tyrosine kinase activity, on the other hand, is exhibited by two different types of membrane proteins. The receptor tyrosine kinases are integral transmembrane proteins, whereas the nonreceptor tyrosine kinases, which are related to a family of retroviral transforming proteins, are peripheral, lipid-anchored proteins.

Nonreceptor Tyrosine Kinases

Figure 34.26 · The soluble tyrosine kinase pp60^v-src is anchored to the plasma membrane via an N-terminal myristyl group.

The first tyrosine kinases to be discovered were associated with viral transforming proteins. These proteins, produced by oncogenic viruses, enable the virus to transform animal cells, that is, to convert them to the cancerous state. A prime example is the tyrosine kinase expressed by the src gene of Rous or avian sarcoma virus. The protein product of this gene is pp60^v-src(the abbreviation refers to phospho-protein, 60 kD, viral origin, sarcoma-causing). The v-src gene was derived from the avian proto-oncogenic gene c-src during the original formation of the virus. The cellular proto-oncogene homolog of pp60^v-src is referred to as pp60^c-src. pp60^v-src is a 526-residue peripheral membrane protein. It undergoes two post-translational modifications: (a) the amino group of the NH₂-terminal glycine is modified by the covalent attachment of a myristyl group (this modification is required for membrane association of the kinase; see Figure 34.26), and (b) Ser¹⁷ and Tyr⁴¹⁶ are phosphorylated. The phosphorylation at Tyr⁴¹⁶, which increases kinase activity two- to threefold, appears to be an autophosphorylation. The significance of nonreceptor tyrosine kinase activity to cell growth and transformation is only partially understood, but many, many cellular proteins (of mostly undetermined function) are phosphorylated by such kinases.

Receptor Tyrosine Kinases

Figure 34.27 · The three classes of receptor tyrosine kinases. Class I receptors are monomeric and contain a pair of Cys-rich repeat sequences. The insulin receptor, a typical Class II receptor, is a glycoprotein composed of two kinds of subunits in an a₂b₂ tetramer. The a- and b-subunits are synthesized as a single peptide chain, together with an N-terminal signal sequence. Subsequent proteolytic processing yields the separate a- and b-subunits. The b-subunits of 620 residues each are integral transmembrane proteins, with only a single transmembrane a-helix, and with the amino terminus outside the cell and the carboxyl terminus inside. The a-subunits of 735 residues each are extracellular proteins that are linked to the b-subunits and to each other by disulfide bonds. The insulin-binding domain is located in a cysteine-rich region on the a-subunits. Class III receptors contain multiple immunoglobulin-like domains. Shown here is fibroblast growth factor (FGF) receptor, which has three immunoglobulin-like domains. (Adapted from Ullrich A., and Schlessinger, J., 1990. Cell 61:203-212.)

The binding of polypeptide hormones and growth factors to receptor tyrosine kinases activates the tyrosine kinase activity of these proteins. These catalytic receptors are composed of three domains (Figure 34.27): a glycosylated extracellular receptor-binding domain, a transmembrane domain consisting of a single transmembrane a-helix, and an intracellular domain that includes a tyrosine kinase domain that mediates the biological response to the hormone or growth factor via its catalytic activity and a regulatory domain that contains multiple autophosphorylation sites.
        There are three classes of receptor tyrosine kinases (Figure 34.27). Class I, exemplified by the epidermal growth factor (EGF) receptor, has an extracellular domain containing two Cys-rich repeat sequences. Class II, typified by the insulin receptor, has an a₂b₂ tetrameric structure with transmembrane b-subunits and a Cys-rich domain in the extracellular a-subunit. Class III receptors, such as the platelet-derived growth factor (PDGF) receptor, have five (or sometimes three) immunoglobulin-like extracellular domains.

Receptor Tyrosine Kinases Are Membrane-Associated Allosteric Enzymes

Figure 34.28 · Ligand (hormone)-stimulated oligomeric association of receptor tyrosine kinases.

Given that the extracellular and intracellular domains of receptor tyrosine kinases are joined only by a single transmembrane helical segment, how does extracellular hormone binding activate intracellular tyrosine kinase activity? How is the signal transduced? As shown in Figure 34.28, signal tranduction occurs by hormone-induced oligomeric association of receptors. Binding of hormone triggers a conformational change in the extracellular domain, which induces oligomeric association. Oligomeric association allows adjacent cytoplasmic domains to interact, leading to phosphorylation of the cytoplasmic domains, and stimulation of cytoplasmic tyrosine kinase activity. In the case of the class II receptors (for example, the insulin receptor), hormone binding induces interactions between the two ab halves within the disulfide-linked receptor complex. By virtue of these ligand-induced conformation changes and oligomeric interactions, receptor tyrosine kinases are membrane-associated allosteric enzymes.
        Autophosphorylation at Tyr residues allows the receptor tyrosine kinase to remain active even after the activating hormone has dissociated from the receptor. On the other hand, the tyrosine kinase can be inactivated by phosphorylation of intracellular Ser/Thr residues. These inactivating phosphorylations are catalyzed by protein kinase C and by cAMP-dependent protein kinases, providing a direct linkage between receptor tyrosine kinases and several key second messengers, including cAMP, IP₃, DAG, and Ca²⁺.

Receptor Tyrosine Kinases Phosphorylate a Variety of Cellular Target Proteins

Receptor tyrosine kinases catalyze the phosphorylation of numerous cellular target proteins, producing coordinated changes in cell behavior, including alterations in membrane transport of ions and amino acids, the transcription of genes, and the synthesis of proteins. Many individual phosphorylation targets have been characterized, including the g-isozymes of phospholipase C and phosphatidylinositol-3-kinase. The latter enzyme phosphorylates PI at the 3-position, producing several phosphoinositide metabolites that are not hydrolyzed by known phospholipase C enzymes. These PI-3-P species may either act as membrane-associated second messengers or be hydrolyzed by other phospholipases.

The Polypeptide Hormones

The largest class of hormones in vertebrate organisms is that of the polypeptide hormones (Table 34.4). One of the first polypeptide hormones to be discovered, insulin, was described by Banting and Best in 1921. Insulin, a secretion of the pancreas, controls glucose utilization and promotes the synthesis of proteins, fatty acids, and glycogen. Insulin, which is typical of the secreted polypeptide hormones, is discussed in detail in Chapters 5, 15, and 23.
        Many other polypeptide hormones are produced and processed in a manner similar to that of insulin. Three unifying features of their synthesis and cellular processing should be noted. First, all secreted polypeptide hormones are originally synthesized with a signal sequence, which facilitates their eventual direction to secretory granules, and thence to the extracellular milieu. Second, peptide hormones are usually synthesized from mRNA as inactive precursors, termed preprohormones, which become activated by proteolysis. Third, a single polypeptide precursor or preprohormone may produce several different peptide hormones by suitable proteolytic processing. The following processing events are common to all preprohormones:

1.  Proteolytic cleavage of a hydrophobic N-terminal signal peptide sequence
2.  Proteolytic cleavage at a site defined by pairs of basic amino acid residues
3.  Proteolytic cleavage at the site of single Arg residues
4.  Post-translational modification of individual amino acids, including a-amidation of the C-terminal residue, phosphorylation, glycosylation, or acetylation of the N-terminal residue

Figure 34.29 · The major pathway of gastrin biosynthesis in human gastric mucosal cells. The asterisk (*) indicates a sulfate group, and P indicates a phosphorylation site. (Adapted from Dockray, G, et al., 1989. Gastrin and CCK-related peptides. In Martinez, J., ed. Peptide Hormones as Prohormones. New York, Halstead Press.)

A good example of these processing events is the biosynthesis of gastrin, a heptadecapeptide (17 residues) secreted by the antral mucosa in the stomach. Gastrin activates acid secretion in the stomach, and also stimulates growth of the acid-secreting mucosa. It is a product of preprogastrin, a peptide of 101 (human) or 104 (pig) residues (Figure 34.29). As expected for a secretory hormone, it has a signal peptide (21 residues), cleavage of which leaves progastrin. Cleavage at pairs of basic residues (noted in the figure) followed by C-terminal amidation produces gastrin (G17) and three other peptides. Several post-translational modifications occur in progastrin, including incorporation of a sulfate at Tyr¹², phosphorylation of Ser²¹, and protective modifications of both the C- and N-termini of gastrin itself. The N-terminal residue of gastrin is a pyroglutamate (a pyrrolidone carboxylic acid), formed by cyclization of the N-terminal glutamine after proteolytic cleavage of the flanking peptide. An identical N-terminal modification is found in bacteriorhodopsin (Chapter 9). At the C-terminal end, three enzymatic steps result in amidation, which protects the peptide hormone from action of carboxypeptidases. The steps involve a trypsin-like cleavage on the C-terminal side of the second arginine, a carboxypeptidase-like action to cleave the two arginines, and degradation of the glycine residue to leave a C-terminal phenylalanine amide (Figure 34.30).

Figure 34.30 · The enzyme-catalyzed proteolysis and amidation reactions at the C-terminus of progastrin that lead to gastrin.

Figure 34.31 · The conversion of prepro-opiomelanocortin to a family of peptide hormones, including corticotropin, b- and g-lipotropin, a- and b-MSH, and endorphin.

            An impressive example of the production of many hormone products from a single precursor is the case of prepro-opiomelanocortin, a 250-residue precursor peptide synthesized in the pituitary gland. A cascade of proteolytic steps produces, as the name implies, a natural opiate substance (endorphin), melanocyte-stimulating hormones (a- and b-MSH), and corticotropin (also known as adrenocorticotrophic hormone, ACTH) as shown in Figure 34.31. As noted, these proteolytic cleavages actually take place in different tissues. Cleavage of pro-opiomelanocortin in the anterior pituitary yields corticotropin and b-lipotropin, which proceed to cells of the central nervous system for the final proteolytic steps.

Figure 34.32 · (a) The primary structures of cytoplasmic and membrane-bound protein-tyrosine phosphatases (PTPases) and cell adhesion molecules (CAM). Key: S-S loops are immunoglobulin domains; blue squares are fibronectin domains; gold rectangles are PTPase catalytic domains; green and red rectangles are membrane-associated domains; blue squares on SH-PTPase are SH2 domains; olive rectangle on PTPase H1 is a cytoskeletal domain. (b) Molecular graphic of PTPase 1B. (Adapted from Streuli, M., et al., 1989. A family of receptor-linked protein tyrosine phosphatases in humans and Drosophila. Proceedings of the National Academy of Sciences, USA 80:8698-8702; and Mauro, I. J., and Dixon, J. E., 1994. Zip codes direct intracellular protein tyrosine phosphatases to the correct cellular “address.” Trends in Biochemical Sciences 19:151-155. Molecular graphic courtesy of David Barford, Oxford University.)

Phosphatases specific for phosphorylated tyrosines on proteins are different from their serine/threonine-specific cousins, and display a wide variety of structures. Some of these protein-tyrosine phosphatases (PTPases) are integral membrane proteins, whereas others are cytoplasmic in nature (Figure 34.32 and Table 34.5). The cytoplasmic PTPases consist of N-terminal catalytic domains with C-terminal regulatory regions that may guide the enzyme to specific intracellular locations and substrates and/or specifically activate the PTPase toward those substrates. The membrane-bound PTPases all consist of a cytoplasmic catalytic domain, a transmembrane domain, and various extracellular domains that are designed to convey signals across the membrane. The extracellular domains bear considerable similarity to cell adhesion molecules (CAM), with repeated immunoglobulin domains and fibronectin type III repeats. PTPases have been linked to the inhibition of cell growth by cell-cell contact and to regulation of T-cell activation and proliferation. It has been postulated that membrane-bound PTPases may act as tumor suppressors and that their loss or mutation could lead to unrestrained cell proliferation and transformation.

Figure 34.33 · The structure of membrane-bound guanylyl cyclases.

Another cellular second messenger, guanosine 3',5'-cyclic monophosphate (cGMP), is formed from GTP by guanylyl cyclase, an enzyme found in several different forms in different cellular locations. Membrane-bound guanylyl cyclases constitute a second class of single-TMS receptors with an extracellular hormone-binding domain; a single, a-helical transmembrane segment; and an intracellular catalytic domain (Figure 34.33). A variety of peptides act to stimulate the membrane-bound guanylyl cyclases, including atrial natriuretic peptide (ANP), which regulates body fluid homeostasis and cardiovascular function; the heat-stable enterotoxins from E. coli; and a series of peptides secreted by mammalian ova (eggs), which stimulate sperm motility and act as sperm chemoattractant signals.

Figure 34.34 · The primary structures of several guanylyl cyclase-activating peptides. Their receptors are glycoproteins with masses of 120 kD to 180 kD.

Speract and resact (Figure 34.34) are two such sperm chemoattractant peptides. Binding of these peptides to an extracellular site on the guanylyl cyclase in the sperm plasma membrane induces a conformational change that activates the intracellular catalytic site for cyclase activity. Activation may involve oligomerization of receptors in the membrane, as for the RTKs discussed previously. In the case of enterotoxins, the binding activity for the peptide has been separated from the guanylyl cyclase itself, suggesting that the toxin receptor is distinct from guanylyl cyclase.

Soluble Guanylyl Cyclases Are Receptors for Nitric Oxide

A soluble guanylyl cyclase in the cytoplasm is the receptor for nitric oxide, or NO ×, a reactive free radical that plays two different roles in cells. On one hand, NO × acts as a neurotransmitter and as a second messenger, inducing relaxation of vascular smooth muscle and mediating penile erection. On the other hand, NO × enables large white blood cells known as macrophages to kill tumor cells and bacteria.
     Cyclic GMP generated by the NO-stimulated, soluble guanylyl cyclase is itself a second messenger that can activate or inhibit a variety of processes. For example, cGMP can (a) regulate ion-channel gating in cerebellar glial cells, and (b) block gap junction conductivity in retinal cells.
Figure 34.35 · The synthesis of nitric oxide (NO ×) by NO synthase.

    NO × is synthesized from arginine by NO synthase in two consecutive monooxygenase reactions (Figure 34.35). As a dissoved gas, NO × is capable of rapid diffusion across membranes in the absence of any apparent carrier mechanism. This property makes NO × a particularly attracitve second messgenger because NO × generated in one cell can exert its effects quickly in many neighboring cells. NO × has a very short cellular half-life (1 to 5 seconds) and is rapidly degraded by noenzymatic pathways.
     Binding of NO × to the heme prosthetic group of soluble guanylyl cyclase causes at least a 400-fold increase in the rate of cGMP synthesis. Soluble guanylyl cyclase is a 150-kD dimer consisting of an a-subunit (82 kD), a b-subunit (70kD), and a heme prosthetic group. NO × binds to the heme group to form a nitrosoheme (Figure 34.36). Sequence analysis indicates that the carboxyl region of the 70-kD subunit is homologous with the carboxyl domain of membrane-bound forms of guanylyl cyclase.

Figure 34.36 · Binding of NO × to the heme moiety of guanylyl cyclase to form a nitrosoheme.

Critical Developments in Biochemistry
The Structure of Inducible NO Synthase
The NO synthases identified in the brain and in the linings of blood vessels produce small amounts of NO × for signaling purposes. Another form of the enzyme, found in macrophages, is termed inducible NO synthase or iNOS. This latter form of enzyme is critical for the immune response, but it is also implicated in a variety of diseases that involve overproduction of NO ×, including septic shock, Alzheimer's disease, multiple sclerosis, stroke, imflammatory bowel disease, rheumatoid arthritis, and many forms of inflammation.       Dennis Stuehr,John Trainer, and their colleagues have determined the structure of the oxygenase domain of iNOS, shown in the dimeric form in the figure. The monomer struture is elogated and curved, with an unusual single-domain a-b fold that resembles a left-handed baseball catcher's mitt. The essential heme is held in the mitt's webbed b-sheet palm with its distal face directed to a large cavity.
(Figure courtesy of John A. Tainer, Scripps Research Institute.)

34.10 · Protein Modules in Signal Transduction

Figure 34.37 · Six of the protein modules that are found in cell-signaling proteins. Shown for each are a molecular graphic image of the module, together with primary structures of several proteins in which they are found. (WW domain coordinates kindly pro- vided by Harmut Oschkinat, Forschungsinstitut für Molekulare Pharmakologie, and Marius Sudol, Mount Sinai School of Medicine.)

Signal transduction within cells occurs via protein-protein and protein-phospholipid interactions based on protein modules. Proteins with two (or more) such modules associate simultaneously with two (or more) binding partners, leading to assembly of functional complexes, either at an activated cell-surface receptor or free in the cytoplasm. The SH2 domain, which binds with high affinity to peptide motifs containing a phosphotyrosine, is a good example (Figure 34.37). The binding of a particular SH2 domain to a particular phosphotyrosine motif depends on the particular sequence of residues that are ­C-terminal to the phosphotyrosine. The SH2 domain itself is a module of about 100 residues that consists of a small b-sheet flanked by a-helices.
            SH2 domains are the prototype for a growing number of protein modules that play a role in cell signaling. PTB modules also recognize and bind to phosphotyrosine motifs, but in a manner different from that of SH2 domains. SH3 and WW modules bind proline-rich target sequences, and PDZ modules bind to the terminal four or five residues of a target protein. The PH (pleckstrin homology) module, with a fold very much like that of the PTB module, functions quite differently, associating with specific phosphoinositides and directing target proteins to the plasma membrane. Figure 34.37 illustrates structures of these modules, together with examples of proteins that contain them.

Membrane-bound receptors can amplify their signaling by means of adaptor proteins that provide docking sites for signaling modules on other proteins. Such docking proteins typically possess an N-terminal sequence that targets the docking protein to the membrane (for example, a PH domain or a myristoylation site) and a PTB domain that enables the docking protein to bind to a phosphorylated tyrosine on a receptor, as well as additional modules and phosphorylation sites that facilitate the binding of target proteins. A typical case is IRS-1 (Insulin Receptor Substrate-1), a substrate of the insulin receptor. IRS-1 has an N-terminal PH domain followed by a PTB domain and 18 potential tyrosine phosphorylation sites. The PH and PTB domains direct IRS-1 to the membrane, facilitating tyrosine phosphorylation of IRS-1 by the insulin receptor tyrosine kinase, and subsequent mediation of additional cell signaling events. Another docking protein, Shc, possesses an N-terminal PTB domain that binds both phosphotyrosine sites and phosphoinositides, several of its own tyrosine phosphorylation sites, and a C-terminal SH2 module that can bind to phosphotyrosines on other target proteins.

Figure 34.38 · A complete signal transduction pathway that connects a hormone receptor with transcription events in the nucleus. A number of similar pathways have been characterized.

Connections—Complete Pathways from Membrane to Nucleus

The complete pathway from hormone binding at the plasma membrane to modulation of transcription in the nucleus is understood for a few signaling pathways. Figure 34.38 shows a complete signal transduction pathway that connects receptor tyrosine kinases, the ras GTPase, cytoplasmic Raf, and two other protein kinases with transcription factors that alter gene expression in the nucleus. This pathway represents just one component of a complex signaling network that involves many other proteins and signaling factors. Current models of cell signaling presume that signaling pathways are redundant, because gene knockout experiments show that many cell signaling factors can be deleted (one at a time) from fertilized embryos with little or no effect on the growth and development of theorganism. Further, it is also clear that signaling pathways converge on or diverge from individual proteins and factors in a pathway. For example, ras proteins bind to, and participate in, the activation of multiple downstream signaling targets (Figure 34.39). Such multiplicities of interactions arise from the existence (and function) of multiple signaling modules in the proteins involved in these

Figure 34.39 · Cell-signaling pathways may both converge and diverge. As shown, ras proteins activate multiple downstream targets.

pathways. Figure 34.40 shows the modular interactions at work in the receptor tyrosine kinase pathway, as well as other modular interactions that operate in other presumed signaling pathways. The presumed existence of several thousand protein kinases and the growing list of signaling factors portend a complex and interwoven network of signaling interactions in nearly all cells.

Figure 34.40 · Interactions between cell-signaling modules modulate and dictate the events of cell signaling. (Adapted from Lin, D., and Pawson, T., 1997. Protein modules in signal transduction. Trends in Cell Biology 8:center.)

34.11 · Steroid Hormones

Figure 34.41 · Structures of some steroid hormones.

The steroid hormones, lipid-soluble molecules derived from cholesterol, include (Figure 34.41) the glucocorticoids (cortisol and corticosterone), the mineralocorticoids (aldosterone), vitamin D, and the sex hormones (progesterone and testosterone, for example) (see Chapter 25 for the details of their synthesis). The steroid hormones exert their effects in two ways. First, by entering cells and migrating to the nucleus, steroid hormones act as transcription regulators, modulating gene expression. These effects of the steroid hormones occur on time scales of hours and involve synthesis of new proteins. Considerable evidence has accumulated, however, that steroids can also act at the cell membrane, directly regulating ligand-gated ion channels and perhaps other processes. These latter processes take place very rapidly, on time scales of seconds and minutes.

Figure 34.42 · A model for steroid hormone action in target cells. The hormone (for example, estrogen) dissociates from plasma proteins, diffuses into the cell, and binds to receptor proteins. The active hormone-receptor complex migrates to the nucleus, where it interacts with DNA or transcription factors or both. (Adapted from Welshons, W., and Jordan, V., 1987. Heterogeneity of nuclear steroid hormone receptors with an emphasis on unfilled receptor sites. In Clark, C., ed., Steroid Hormone Receptors. Ellis Horwood, New York, VCH Publishers.)

Intracellular effects of the steroid hormones are initiated when the steroid diffuses across the plasma membrane and binds to specific receptor proteins. The binding of steroids to these receptors is typically very tight, with dissociation constants in the nanomolar range. Nearly all the receptor molecules in a steroid-sensitive cell are located in the nucleus. Nonetheless, because of the highly hydrophobic character of the steroids themselves and the low likelihood that they could migrate through the cytoplasm to the nucleus without the help of receptor proteins, it is believed that small concentrations of receptor proteins are available in the cytoplasm to ferry the steroids from the plasma membrane to the nucleus (Figure 34.42).

Figure 34.43 · The primary structures of nuclear steroid and thyroid receptor proteins. (Adapted from Gronmeyer, H., ed., 1988. Affinity labelling and cloning of steroid and thyroid hormone receptors. Ellis Horwood, VCH Publishers.)

            Steroid receptor proteins possess a strikingly similar structural organization (Figure 34.43), indicating that they are all members of a gene superfamily, and that they have evolved from a common ancestral precursor. Each of these receptor proteins contains a hydrophobic domain near the C-terminus that is presumed to bind the steroid hormone, as well as a central, hydrophilic DNA-binding domain. These DNA-binding domains are highly homologous (Figure 34.43). Perhaps the most striking feature of the DNA domains is the conserved arrangement of nine Cys residues in all steroid receptors. Three pairs of these Cys residues are in Cys-X-X-Cys sequences, which are commonly found in the DNA-binding zinc-finger proteins (Chapter 31). The steroid-binding domains of the steroid receptor proteins show less sequence homology, but certain residues are conserved at equivalent positions in all receptors. These domains are presumed to form hydrophobic pockets specific to each steroid hormone.
            The steroid-receptor complex has two functions in the nucleus. It can bind directly to DNA in order to regulate transcription, or it may combine with transcription factors such as the Jun and Fos proteins (Table A in box, page S-12). In these latter interactions, the steroid hormone receptors regulate gene expression without interacting directly with DNA.
            The receptor proteins for thyroid hormones are highly homologous with those of the steroid hormones (Figure 34.43).

Figure 34.44 · The structure of triiodothyronine and thyroxine.

These thyroid hormone receptors, which mediate the effects of thyroxine (T₄) and triiodothyronine (T₃) (Figure 34.44), possess DNA-binding domains with the same Cys-X-X-Cys motif and identically conserved Lys and Arg residues. In spite of the high degree of homology in their DNA-binding domains, these receptor proteins specifically recognize unique DNA sequences on target genes (Chapter 31).

Several effects of steroid hormones appear to involve action at the plasma membrane. For example, progesterone modulates Ca²⁺ channels in the membranes of brain stem neurons and also activates processes in Xenopus laevis oocytes by binding to the plasma membrane. Certain steroid actions occur so rapidly that activation of protein synthesis cannot be involved. The male steroid hormone testosterone quickly stimulates the transport of glucose, Ca²⁺ ions, and amino acids into rat kidney cells. Similar rapid induction of Ca²⁺ influx into heart cells by testosterone has also been demonstrated.
Figure 34.45 · A schematic model of the GABA_A receptor, indicating the interactions of steroid hormones with the receptor. (Adapted from ^{Touchette, N., 1990. Man bites dogma: A new role for steroid hormones. Journal of NIH Research 2:71-74.)}

            Much of the resistance to the idea of steroid action at plasma membrane receptors has arisen from an inability to demonstrate a steroid-receptor interaction. However, several groups have now shown conclusively that 3a-hydroxy, 5a-pregnan-20-one (3a-OH-DHP), a metabolite of progesterone, binds with high affinity to the GABA_A receptor in the brain, enhancing its inhibitory effect on neural transmission. As seen in the Special Focus that follows, the GABA_A receptor is a chloride channel (Figure 34.45) that is opened by binding g-aminobutyric acid (GABA). The demonstration that a steroid hormone can stimulate this channel is a landmark discovery and may presage a revision of the previous dogma that steroids act only at intracellular sites.

A Deeper Look
The Acrosome Reaction
Steroid hormones affect ion channels in the acrosome reaction, which must occur before human sperm can fertilize an egg. The acrosome is an organelle that surrounds the head of a sperm (see figure) and lies just inside and juxtaposed with the plasma membrane. The acrosome itself is essentially a large vesicle of hydrolytic and proteolytic enzymes. In the acrosome reaction, influx of Ca2+ ions causes the outer acrosomal membrane to fuse with the plasma membrane. These fused membrane segments separate from each other and diffuse away, freeing the acrosomal enzymes to attack the egg, and exposing binding sites on the inner	acrosomal membrane that are thought to interact with the egg in the fertilization process.         This acrosome reaction is induced by progesterone, a female hormone secreted by the cumulus oophorus, a collection of ovarian follicle cells surrounding the egg! Intracellular Ca2+ levels increase within seconds of treating human sperm with progesterone. These effects must occur via binding of the steroid to the sperm plasma membrane. A far longer time would be required for progesterone to act through an enhancement of transcription.
The acrosome reaction in human sperm.

Special Focus:

Excitable Membranes, Neurotransmission, and Sensory Systems

The survival of higher organisms is predicated on the ability to respond rapidly to sensory input such as sights, sounds, and smells. The responses to such stimuli may include muscle movements and many forms of intercellular communication. Hormones (as described earlier in this chapter) can move through an organism only at speeds determined by the circulatory system. In most higher organisms, a faster means of communication is crucial. Nerve impulses, which can be propagated at speeds up to 100 m/sec, provide a means of intercellular signaling that is fast enough to encompass sensory recognition, movement, and other physiological functions and behaviors in higher animals. The generation and transmission of nerve impulses in vertebrates is mediated by an incredibly complicated neural network that connects every part of the organism with the brain—itself an interconnected array of as many as 10¹² cells.
            Despite their complexity and diversity, the nervous systems of higher organisms all possess common features and common mechanisms. Physical or chemical stimuli are recognized by specialized receptor proteins in the membranes of excitable cells. Conformational changes in the receptor protein result in a change in enzyme activity or a change in the permeability of the membrane. These changes are then propagated throughout the cell or from cell to cell in specific and reversible ways to carry information through the organism. This section describes the characteristics of excitable cells and the mechanisms by which these cells carry information at high speeds through an organism.

Figure 34.46 · The structure of a mammalian motor neuron. The nucleus and most other organelles are contained in the cell body. One long axon and many shorter dendrites project from the body. The dendrites receive signals from other neurons and conduct them to the cell body. The axon transmits signals from this cell to other cells via the synaptic knobs. Glial cells called Schwann cells envelop the axon in layers of an insulating myelin membrane. Although glial cells lie in proximity to neurons in most cases, no specific connections (such as gap junctions, for example) connect glial cells and neurons. However, gap junctions can exist between adjacent glial cells.

Neurons and neuroglia (or glial cells) are cell types unique to nervous systems. The reception and transmission of nerve impulses are carried out by neurons (Figure 34.46), whereas glial cells serve protective and supportive functions. (“Neuroglia” could be translated as “nerve glue.”) Glial cells differ from neurons in several ways. Glial cells do not possess axons or synapses and they retain the ability to divide throughout their life spans. Glial cells, which outnumber neurons by at least 10 to 1 in most animals, may be of several types. Certain glial cells, such as Schwann cells, envelop and surround neurons, forming a protective sheath. Other glial cells are phagocytic in nature and remove cellular debris from nervous tissue. Glial cells also form linings in the cavities of the brain and around the spinal cord.
            Neurons are distinguished from other cell types by their long cytoplasmic extensions or projections, called processes (Figure 34.46). Most neurons consist of three distinct regions: the cell body (called the soma), the axon, and the dendrites. The axon ends in small structures called synaptic terminals, synaptic knobs, or synaptic bulbs. Dendrites are short, highly branched structures emanating from the cell body that receive neural impulses and transmit them to the cell body. The space between a synaptic knob on one neuron and a dendrite ending of an adjacent neuron is the synapse or synaptic cleft.
            Three kinds of neurons are found in higher organisms: sensory neurons, interneurons, and motor neurons. Sensory neurons acquire sensory signals, either directly or from specific receptor cells, and pass this information along to either interneurons or motor neurons. Interneurons simply pass signals from one neuron to another, whereas motor neurons pass signals from other neurons to muscle cells, thereby inducing muscle movement (motor activity).

Ion Gradients: Source of Electrical Potentials in Neurons

Figure 34.47 · The concentrations of Na⁺, K⁺, and Cl^- ions inside and outside of a typical resting mammalian axon are shown. Assuming relative permeabilities for K⁺, Na⁺, and Cl^- are 1, 0.04, and 0.45, respectively, the Goldman equation yields a membrane potential of -60 mV. (See A Deeper Look, this page.)

The impulses that are carried along axons, as signals pass from neuron to neuron, are electrical in nature. These electrical signals occur as transient changes in the electrical potential differences (voltages) across the membranes of neurons (and other cells). Such potentials are generated by ion gradients. The cytoplasm of a neuron at rest is low in Na⁺ and Cl^- and high in K⁺, relative to the extracellular fluid (Figure 34.47). These gradients are generated by the Na⁺, K⁺-ATPase (see Chapter 10). A resting neuron exhibits a potential difference of approximately 260 mV (that is, negative inside). Consider the potassium gradient in isolation (Figure

Figure 34.48 · The K⁺ concentration inside a resting axon is approximately 400 mM, whereas the concentration outside is 20 mM. In the absence of an electrical potential difference across the membrane, there would be a strong thermodynamic tendency for K⁺ to flow out of the axon. According to the Nernst equation, the concentration gradient of K⁺ would be exactly balanced (at 25°C) by a potential difference of -77 mV (negative inside). If DY = -77 mV, the net flow of K⁺ would be zero.

34.48). The electrochemical potential for an ion distributed across a membrane was given in Equation 10.2 as

For the particular potassium gradient in Figure 34.48, what is the equilibrium potential at which no net ion flow would occur? At equilibrium, DG=0, and Equation 10.2 becomes

                                                                                                                                              (34.1)

This is a form of the Nernst equation, which relates concentration differences and potential differences. For the case in Figure 34.48, with Z=1,                                                                                                                                               (34.2)

and, at 25°C, with  kJ/V-mol,

At a potential difference of -77 mV, the net flow of K⁺ across this membrane is zero. What is the equilibrium potential for Na⁺ in Figure 34.47? Applying the Nernst equation again, an equilibrium potential of +53.4 mV for Na⁺ in Figure 34.47 can be calculated.

The Action Potential

Nerve impulses, also called action potentials, are transient changes in the membrane potential that move rapidly along nerve cells. Action potentials are created when the membrane is locally depolarized by approximately 20 mV—from the resting value of about -60 mV to a new value of approximately -40 mV. This small change is enough to have a dramatic effect on specific proteins in the axon membane called voltage-gated ion channels. These proteins are ion channels that are specific either for Na⁺ or K⁺. These ion channels are normally closed at the resting potential of -60 mV. When the potential difference rises to -40 mV, the “gates” of the Na⁺ channels are opened and Na⁺ ions begin to flow into the cell, as expected from the electrochemical potential exerted on Na⁺.

Figure 34.49 · The time dependence of an action potential, compared with the ionic permeabilities of Na⁺ and K⁺. (a) The rapid rise in membrane potential from -60 mV to slightly more than +30 mV (a) is referred to as a “depolarization.” This depolarization is caused (b) by a sudden increase in the permeability of Na⁺. As the Na⁺ permeability decreases, K⁺ permeability is increased and the membrane potential drops, eventually falling below the resting potential—a state of “hyperpolarization”—followed by a slow return to the resting potential. (Adapted from Hodgkin, A., and Huxley, A., 1952. A quantitative description of membrane current and its application to conduction and excitation in nerve. Journal of Physiology 117:500-544.)

As Na⁺ enters the cell, the membrane potential continues to increase, and additional Na⁺ channels are opened (Figure 34.49). The potential rises to more than +30 mV. At this point, Na⁺ influx slows and stops because the actual potential is approaching the Na⁺ equilibrium potential. As the Na⁺ channels close, K⁺ channels begin to open and K⁺ ions stream out of the cell, returning the membrane potential to negative values. The potential eventually overshoots its resting value a bit. At this point, K⁺ channels close and the resting potential is eventually restored by action of the Na⁺,K⁺-ATPase and the other channels. These transient increases and decreases, first in Na⁺ permeability and then in K⁺ permeability, were first observed by Alan Hodgkin and Andrew Huxley. For this and related work, Hodgkin and Huxley, along with J. C. Eccles, won the Nobel prize in physiology or medicine in 1963.

Figure 34.50 · The propagation of action potentials along an axon. Figure 34.49 shows the time dependence of an action potential at a discrete point on the axon. This figure shows how the membrane potential varies along the axon as an action potential is propagated. (For this reason, the shape of the action potential is the apparent reverse of that shown in Figure 34.49.) At the leading edge of the action potential, membrane depolarization causes Na⁺ channels to open briefly. As the potential moves along the axon, the Na⁺ channels close and K⁺ channels open, leading to a drop in potential and the onset of hyperpolarization. When the resting potential is restored, another action potential can be initiated.

These changes in potential in one part of the axon are rapidly passed along the axonal membrane (Figure 34.50). The sodium ions that rush into the cell in one localized region actually diffuse farther along the axon, raising the Na⁺ concentration and depolarizing the membrane, causing Na⁺ gates to open in that adjacent region of the axon. In this way, the action potential moves down the axon in wavelike fashion. This simple process has several very dramatic properties:

1.  Action potentials propagate very rapidly—up to and sometimes exceeding 100 meters per second.

2.  The action potential is not attenuated (diminished in intensity) as a function of distance transmitted.

The input of energy all the way along an axon—in the form of ion gradients maintained by Na⁺,K⁺-ATPase—ensures that the shape and intensity of the action potential is maintained over long distances. The action potential has an all-or-none character. There are no gradations of amplitude; a given neuron is either at rest (with a polarized membrane), or is conducting a nerve impulse (with a reversed polarization). Because nerve impulses display no variation in amplitude, the size of the action potential is not important in processing signals in the nervous system. Instead, it is the number of action potential firings and the frequency of firing that carry specific information.

The Voltage-Gated Sodium and Potassium Channels

Figure 34.51 · Na⁺ channels are infrequently and randomly distributed in unmyelinated nerve. In myelinated axons, Na⁺ channels are clustered in large numbers in the nodes of Ranvier, between the regions surrounded by myelin sheath structures.

The action potential is a delicately orchestrated interplay between the Na⁺,K⁺-ATPase and the voltage-gated Na⁺ and K⁺ channels, initiated by a stimulus at the postsynaptic membrane. The density and distribution of Na⁺ channels along the axon are different for myelinated and unmyelinated axons (Figure 34.51). In unmyelinated axons, Na⁺ channels are uniformly distributed, although they are few in number—approximately 20 channels per mm². On the other hand, in myelinated axons, Na⁺ channels are clustered at the nodes of Ranvier. In these latter regions, they occur with a density of approximately 10,000 per mm². Elucidation of these distributions of Na⁺ channels was made possible by the use of several Na⁺-channel toxins (Figure 34.52).

Figure 34.52 · Effectors of Na⁺ channels include (a) tetrodotoxin and saxitoxin, which block the Na⁺ channel in a closed state, and (b) veratridine and batrachotoxin, which block the Na⁺ channel in an open state. K⁺ channel blockers include (c) 4-aminopyridine, tetraethylammonium ion, mast cell degranulating peptide (MCDP), dendrotoxin (DTX), and charybdotoxin (CTX). (See A Deeper Look boxes, “Tetradoxin and Other Na⁺ Channel Toxins” and “Potassium Channel Toxins.”)

Figure 34.53 · The Na⁺ channel comprises three subunits, denoted a, b₁, and b₂. A disulfide bridge links a and b₂ as shown. All three subunits are glycosylated, and the a subunit can be phosphorylated on the cytoplasmic surface.

        The purified Na⁺ channel from mammalian brain is a heterotrimer consisting of a- (260 kD), b₁- (36 kD), and b₂- (33 kD) subunits (Figure 34.53). All three subunits are exposed to the extracellular surface and are heavily glycosylated. The b₂-subunit is attached to the a-subunit by disulfide bonds. The 260-kD a-subunit contains the binding site for toxins. The Na⁺-channel a-subunit contains four domains of 300 to 400 amino acids each (Figure 34.54), with

approximately 50% identity or conservation in their amino acid sequences. Each domain contains six regions (denoted
Figure 34.54 · Model for the arrangement of the Na⁺ channel a-subunit in the plasma membrane. The a-subunit consists of four domains (I through IV), each of which contains six transmembrane a-helices, designated S1 through S6. Phosphorylation sites (P) and location of positive charges on helix S4 are indicated.

S1 to S6) of probable a-helical structure, which are long enough to be membrane-spanning segments. Segments S5 and S6 are uniformly hydrophobic, S1 and S2 are hydrophobic with an occasional hydrophilic residue, and S3 has several charged residues. S4 has both hydrophobic and positively charged residues; its sequence is highly conserved among voltage-gated Na⁺, K⁺, and Ca²⁺ channels. Nearly every third residue in S4 segments is a lysine or an arginine. Studies with mutant forms of the Na⁺ channel have provided evidence that this segment may be part of the voltage-sensing mechanism in the Na⁺ channel. Shosaku Numa and his collaborators have produced Na⁺ channels with one or more of the positively charged lysines and arginines of S4 missing and found that a reduction in the net positive charge of S4 results in a
Figure 34.55 · A model for the formation of an ion channel from the four membrane-spanning domains of the Na⁺ channel. The ion channel is located in the center of the structure. Four corresponding helical segments (purple), one from each domain, form the wall of the channel. Voltage-dependent gating is thought to be mediated by the S4 segments (green).

reduction in the steepness of the voltage dependence of sodium channel activation. S4 senses the transmembrane electric field, thereby controlling the gating of the sodium channel. Numa has suggested that the four S2 segments of an a-subunit form the walls of the Na⁺ channel (Figure 34.55).

A Deeper Look
Tetrodotoxin and Other Na+ Channel Toxins
Tetrodotoxin and saxitoxin are highly specific blockers of Na+ channels and bind with very high affinity (KD>1 nM). This unique specificity and affinity have made it possible to use radioactive forms of these toxins to purify Na+ channels and map their distribution on axons. Tetrodotoxin is found in the skin and several internal organs of puffer fish, also known as blowfish or swellfish, members of the family Tetraodontidae, which react to danger by inflating themselves with water or air to nearly spherical (and often comical) shapes (see figure). Although tetrodotoxin poisoning can easily be fatal, puffer fish are delicacies in Japan, where they are served in a dish called fugu. For this purpose the puffer fish must be cleaned and prepared by specially trained chefs. Saxitoxin is made by Gonyaulax catenella and G. tamarensis, two species of marine dinoflagellates or plankton that	are responsible for “red tides” that cause massive fish kills. Saxitoxin is concentrated by certain species of mussels, scallops, and other shellfish that are exposed to red tides. Consumption of these shellfish by animals, including humans, can be fatal. In addition to these toxins, which prevent the Na+ channel from opening, there are equally poisonous agents that block the Na+ channel in an open state, permitting Na+ to stream into the cell without control, destroying the Na+ gradients. Included in this group of toxins (Figure 34.52) are veratridine from Schoenocaulon officinalis, a member of the lily family, and batrachotoxin, a compound found  in skin secretions of a Colombian frog, Phyllobates aurotaenia. These skin secretions have traditionally been used as arrow poisons.
	Tetrodotoxin is found in puffer fish, which are prepared and served in Japan as fugu. The puffer fish on the left is unexpanded; the one on the right is inflated. (left, Zig Leszczynski/Animals, Animals; right, Tim Rock/Animals, Animals.)

 Figure 34.56 · (a) A model for the insertion of the potassium channel a subunit in the plasma membrane. The transmembrane domain consists of six putative membrane-spanning helical segments and is highly homologous with the corresponding domains of the Na⁺ channel. (b) Structure of the K⁺ channel from Streptomyces lividans, top view and side view. (Images courtesy of Roderick MacKinnon, Rockefeller University.)

       Voltage-gated potassium channels have now been identified in many tissues and species. Although the primary structures of these proteins are highly homologous, they display a variety of functional properties. Just as for the voltage-gated sodium channels, the high-affinity binding of several specific K⁺-channel blockers has aided in the identification and characterization of K⁺ channels. The K⁺ channel from rat synaptosomal membranes consists of an a-subunit of 76 to 80 kD and a b-subunit of 38 kD. Phosphorylation of the a-subunit, either by a cAMP-dependent protein kinase or by an endogenous protein kinase, leads to activation of the K⁺ channel. The a-subunit possesses binding sites for dendrotoxin I and for MCDP (Figure 34.52). TheK⁺-channel a-subunit from Drosophila consists of 616 amino acids. Hydropathy analysis reveals six or seven sequences with hydrophobic character and helix-forming potential. The fourth of the six putative transmembrane segments has striking homology with S4 of the sodium channel, and the proposed arrangement of the K⁺-channel protein in Figure 34.56a is similar to that for each of the four homologous domains of the sodium channel. The role of S4 of the sodium channel as a voltage sensor and its similarity to the fourth segment of the K⁺ channel raise the question of whether this fourth segment might play a similar voltage-sensing role in the K⁺ channel. Roderick MacKinnon and coworkers have determined the structure of a simpler, two-helix K⁺ channel from Streptomyces lividans (Figure 34.56b).

Cell-Cell Communication at the Synapse

How are neuronal signals passed from one neuron to the next? Neurons are juxtaposed at the synapse. The space between the two neurons is called the synaptic cleft. The number of synapses in which any given neuron is involved varies greatly. There may be as few as one synapse per postsynaptic cell (in the midbrain) to many thousands per cell. Typically, 10,000 synaptic knobs may impinge on a single spinal motor neuron, with 8000 on the dendrites and 2000 on the soma or cell body. The ratio of synapses to neurons in the human forebrain is approximately 40,000 to 1!
        Synapses are actually quite specialized structures and there are several different types. A minority of synapses in mammals, termed electrical synapses, are characterized by a very small gap—approximately 2 nm—between the presynaptic cell (which delivers the signal) and the postsynaptic cell (which receives the signal). At electrical synapses, the arrival of an action potential on the presynaptic membrane leads directly to depolarization of the postsynaptic membrane, initiating a new action potential in the postsynaptic cell. However, most synaptic clefts are much wider—on the order of 20 to 50 nm. In these, an action potential in the presynaptic membrane causes secretion of a chemical substance—called a neurotransmitter—by the presynaptic cell. This substance binds to receptors on the postsynaptic cell, initiating a new action potential. Synapses of this type are thus chemical synapses.
        Different synapses utilize specific neurotransmitters. The cholinergic synapse, a paradigm for chemical transmission mechanisms at synapses, employs acetylcholine as a neurotransmitter. Other important neurotransmitters and receptors have been discovered and characterized. These all fall into one of several major classes, including amino acids (and their derivatives), catechol-amines, peptides, and gaseous neurotransmitters. Table 34.6 lists some, but not all, of the known neurotransmitters.

The Cholinergic Synapses

Figure 34.57 · Cell-cell communication at the synapse (a) is mediated by neurotransmitters such as acetylcholine, produced from choline by cholineacetyltransferase. The arrival of an action potential at the synaptic knob (b) opens Ca²⁺ channels in the presynaptic membrane. Influx of Ca²⁺ induces the fusion of acetylcholine-containing vesicles with the plasma membrane and release of acetylcholine into the synaptic cleft (c). Binding of acetylcholine to receptors in the postsynaptic membrane opens Na⁺ channels (d). The influx of Na⁺ depolarizes the postsynaptic membrane, generating a new action potential.

In cholinergic synapses, small synaptic vesicles inside the synaptic knobs contain large amounts of acetylcholine (approximately 10,000 molecules per vesicle; Figure 34.57). When the membrane of the synaptic knob is stimulated by an arriving action potential, special voltage-gated Ca²⁺ channels open and Ca²⁺ ions stream into the synaptic knob, causing the acetylcholine-containing vesicles to attach to and fuse with the knob membrane. The vesicles open, spilling acetylcholine into the synaptic cleft. Binding of acetylcholine to specific acetylcholine receptors in the postsynaptic membrane causes opening of ion channels and the creation of a new action potential in the postsynaptic neuron.

Acetylcholine Release Is Quantized

The action of Ca²⁺ ions on this exocytotic release of acetylcholine is not fully understood. The membranes of synaptic vesicles contain a 75-kD protein called synapsin-I, which binds the Ca²⁺-regulatory protein calmodulin. Synapsin-I, which can be phosphorylated by a cAMP-dependent protein kinase, appears to facilitate fusion of synaptic vesicles with the knob membrane. The release of acetylcholine in discrete quanta (10,000 molecules per vesicle) was a novel idea when it was proposed by Bernard Katz in the early 1950s. Katz based his hypothesis on the occurrence of small electrical pulses in resting neuromuscular junctions. The size of these pulses—3 mV and less—was about what would be expected from the release of a single vesicle’s contents of acetylcholine. The size of the pulses observed in response to an action potential corresponds to the release of up to 400 acetylcholine-containing vesicles.
        A variety of toxins can alter or affect this process. The anaerobic bacterium Clostridium botulinum, which causes botulism poisoning, produces several toxic proteins that strongly inhibit acetylcholine release. The black widow spider, Lactrodectus mactans, produces a venom protein, a-latrotoxin, that stimulates abnormal release of acetylcholine at the neuromuscular junction. The bite of the black widow causes pain, nausea, and mild paralysis of the diaphragm but is rarely fatal.

Figure 34.58 · Two types of acetylcholine receptors are known. Nicotinic acetylcholine receptors are locked in their open conformation by nicotine. Obtained from tobacco plants, nicotine is named for Jean Nicot, French ambassador to Portugal, who sent tobacco seeds to France in 1550 for cultivation. Muscarinic acetylcholine receptors are stimulated by muscarine, obtained from the intensely poisonous mushroom, Amanita muscaria.

Two different acetylcholine receptors are found in postsynaptic membranes. They were originally distinguished by their responses to muscarine, a toxic alkaloid in toadstools, and nicotine (Figure 34.58). The nicotinic receptors are cation channels in postsynaptic membranes and the muscarinic receptors are transmembrane proteins that interact with G proteins. The receptors in sympathetic ganglia and those in motor endplates of skeletal muscle are nicotinic receptors. Nicotine locks the ion channels of these receptors in their open conformation. Muscarinic receptors are found in smooth muscle and in glands. Muscarine mimics the effect of acetylcholine on these latter receptors.
        The nicotinic acetylcholine receptor is a transmembrane glycoprotein with an approximate molecular mass of 270 kD, consisting of four different subunits, a (54 kD), b (56 kD), g (58 kD),

Figure 34.59 · A model for the arrangement of the acetylcholine receptor a subunit in the postsynaptic membrane.

and D (60 kD), with a quaternary structure of a₂bgd. Each a-subunit possesses a binding site for acetylcholine. The four different subunits have homologous sequences and may have evolved via gene duplication. Each subunit includes five hydrophobic regions that are postulated to be helical, membrane-spanning segments (Figure 34.59). Three-dimensional reconstructions indicate that the receptor has a cylindrical shape, with approximate fivefold symmetry and a central pore (Figure 34.60). Models indicate that several charged residues are clustered on one face of the fourth helix on each subunit. These charged helical faces may line the wall of the transmembrane channel.

Figure 34.60 · (a) Side view and (b) top view of the structure of the nicotinic acetylcholine receptor Na⁺ channel, as determined by image reconstruction from electron micrographs. The diameter of the pore at the mouth of the channel (facing the synapse) is 2.2 nm. (Adapted from Brisson, A., and Unwin, P. N. T., 1985. Quaternary structure of the acetylcholine receptor. Nature 315:474-477.)

The Nicotinic Acetylcholine Receptor Is a Ligand-Gated Ion Channel

The nicotinic acetylcholine receptor functions as a ligand-gated ion channel, and, on the basis of its structure, it is also an oligomeric ion channel. When acetylcholine (the ligand) binds to this receptor, a conformational change opens the channel, which is equally permeable to Na⁺ and K⁺. Na⁺ rushes in while K⁺ streams out, but, because the Na⁺ gradient across this membrane is steeper than that of K⁺, the Na⁺ influx greatly exceeds the K⁺ efflux. The influx of Na⁺ depolarizes the postsynaptic membrane, initiating an action potential in the adjacent membrane. After a few milliseconds, the channel closes, even though acetylcholine remains bound to the receptor. At this point the channel will remain closed until the concentration of acetylcholine in the synaptic cleft drops to about 10 nM.

Acetylcholinesterase Degrades Acetylcholine in the Synaptic Cleft

Figure 34.61 · Acetylcholine is degraded to acetate and choline by acetylcholinesterase, a serine protease.

Following every synaptic signal transmission, the synapse must be readied for the arrival of another action potential. Several things must happen very quickly. First, the acetylcholine left in the synaptic cleft must be rapidly degraded to resensitize the acetylcholine receptor and to restore the excitability of the postsynaptic membrane. This reaction is catalyzed by acetylcholinesterase (Figure 34.61).

Figure 34.62 · Following a synaptic transmission event, acetylcholine is repackaged in vesicles in a multistep process. Synaptic vesicles are formed by endocytosis, and acetylcholine is synthesized by cholineacetyltransferase. A proton gradient is established across the vesicle membrane by an H¹-transport ATPase, and a proton-acetylcholine transport protein transports acetylcholine into the synaptic vesicles, exchanging acetylcholine for protons in an electrically neutral antiport process.

        When [acetylcholine] has decreased to low levels, acetylcholine dissoci-ates from the receptor, which thereby regains its ability to open in a ligand-dependent manner. Second, the synaptic vesicles must be reformed from the presynaptic membrane by endocytosis (Figure 34.62) and then must be restocked with acetylcholine. This occurs through the action of an ATP-driven H¹ pump and an acetylcholine transport protein. The H¹ pump in this case is a member of the family of V-type ATPases. It uses the free energy of ATP hydrolysis to create an H¹ gradient across the vesicle membrane. This gradient is used by the acetylcholine transport protein to drive acetylcholine into the vesicle, as shown in Figure 34.62.

     Figure 34.63 · Tubocurarine, obtained from the plant Chondrodendron tomentosum, is the active agent in “tube curare,” named for the bamboo tubes in which it is kept by South American tribal hunters. Atropine is produced by Atropa belladonna, the poisonous deadly nightshade. The species name, which means “beautiful woman,” derives from the use of atropine in years past by Italian women to dilate their pupils. Atropine is still used for pupil dilation in eye exams by ophthalmologists. Cobratoxin and a-bungarotoxin are produced by the cobra (Naja naja) and the banded krait snake (Bungarus multicinctus), respectively.

   Antagonists of the nicotinic acetylcholine receptor are particularly potent neurotoxins. These agents, which bind to the receptor and prevent opening of the ion channel, include d-tubocurarine (Figure 34.63), the active agent in the South American arrow poison curare, and several small proteins from poisonous snakes. These latter agents include cobratoxin from cobra venom, and a-bungarotoxin, from Bungarus multicinctus, a snake common in Taiwan.

Muscarinic Receptor Function Is Mediated by G Proteins

There are several different types of muscarinic acetylcholine receptors, with different structures and different apparent functions in synaptic transmission. However, certain structural and functional features are shared by this class of receptors. Muscarinic receptors that have been isolated or cloned and sequenced are 70-kD glycoproteins (of which about 25 to 27% are carbohydrates) and are members of the 7-transmembrane segment (7-TMS) family of receptors.
Figure 34.64 · Muscarinic acetylcholine receptors are typical 7-transmembrane segment receptor proteins. Binding of acetylcholine to these receptors activates G proteins, which inactivate adenylyl cyclase, activate phospholipase C, and open K⁺ channels.

            Activation of muscarinic receptors (by binding of acetylcholine) results in several effects, including the inhibition of adenylyl cyclase, the stimulation of phospholipase C, and the opening of K⁺ channels. As shown in Figure 34.64, all of these effects of muscarinic receptors are mediated by G proteins, and it appears that muscarinic receptor activation involves several different G proteins. Many antagonists for muscarinic acetylcholine receptors are known, including atropine from . the deadly nightshade plant, whose berries are sweet and tasty but highly poisonous (Figure 34.63).

Figure 34.65 · The structures of decamethonium ion and succinylcholine.

            Both the nicotinic and muscarinic acetylcholine receptors are sensitive to certain agents that, in effect, overstimulate the receptor. This effect can happen in two ways. Certain substances, such as the cations decamethonium and succinylcholine (Figure 34.65) bind to (and activate) the receptors but are not rapidly degraded by acetylcholinesterase. Thus, they remain bound to the receptors for long periods, keeping the ion channel in the open conformation and preventing the reestablishment of receptor sensitivity. Similar effects can be produced by agents that inactivate acetylcholinesterase itself. Acetylcho­linesterase is a serine esterase similar to trypsin and chymotrypsin (Chapter 16). The reactive serine at the active site of such enzymes is a vulnerable target for organophosphorus inhibitors (Figure 34.66). DIPF and related agents form stable covalent complexes with the active-site serine, irreversibly blocking the enzyme. Malathion and parathion are commonly used insecticides, and sarin and tabun are nerve gases used in chemical warfare. All these agents effectively block nerve impulses, stop breathing, and cause death by suffocation.

Figure 34.66 · Covalent inhibitors (blue) of acetylcholinesterase include DIFP, the nerve gases tabun and sarin, and the insecticides parathion and malathion. Milder, noncovalent (pink) inhibitors of acetylcholinesterase include physostigmine and neostigmine.

            Milder inhibitors of the acetylcholinesterase reaction are useful therapeutic agents. Physostigmine, an alkaloid found in calabar beans, and neostigmine, a synthetic analog, contain carbamoyl ester groups (Figure 34.66). Reaction with the active-site serine of acetylcholinesterase leaves an intermediate that is hydrolyzed only very slowly, effectively inhibiting the enzyme. Physostigmine and neostigmine have been used to treat myasthenia gravis, a chronic disorder that causes muscle weakness after the muscle has been exercised, and eventual paralysis. Myasthenia gravis is an autoimmune disease, in which individuals produce antibodies that bind to their own acetylcholine receptors, blocking the response to acetylcholine. By blocking acetylcholinesterase (thus allowing acetylcholine levels in the synaptic cleft to remain high), physostigmine and neostigmine can suppress the symptoms of myasthenia gravis.

Other Neurotransmitters and Synaptic Junctions

Synaptic junctions that use amino acids, catecholamines, and peptides (see Table 34.6) appear to operate much the way the cholinergic synapses do. Presynaptic vesicles release their contents into the synaptic cleft, where the neurotransmitter substance can bind to specific receptors on the postsynaptic membrane to induce a conformation change and elicit a particular response. This common mechanism notwithstanding, the synapses using these various neurotransmitters display markedly different properties. Some of these respond quickly to neuronal signals, whereas others respond slowly. Some of these neurotransmitters are excitatory in nature and stimulate postsynaptic neurons to transmit impulses, whereas others are inhibitory and prevent the postsynaptic neuron from carrying other signals. Moreover, it is also becoming clear that each known neurotransmitter acts on a family of postsynaptic receptors. Just as acetylcholine acts both on nicotinic and muscarinic receptors, so most of the known neurotransmitters act on several (and in some cases, many) different kinds of receptors. Biochemists are just beginning to understand the sophistication and complexity of neuronal signal transmission.

Glutamate and Aspartate: Excitatory Amino Acid Neurotransmitters

Figure 34.67 · The excitatory neurotransmitter glutamate is cleared from the synaptic cleft by either of two pathways. K⁺-dependent reuptake transporters may transport it back into the presynaptic neuron, or it may accumulate in nearby glial cells via similar transport proteins. Glutamate is repackaged in synaptic vesicles by means of a transport protein that exchanges glutamate for protons. This latter uptake into vesicles depends upon the establishment of a proton gradient across the vesicular membrane. In the glial cells, glutamate is converted to glutamine and then transported back to the presynaptic neuron, where it is converted back to glutamate in the mitochondria.

The common amino acids glutamate and aspartate act as neurotransmitters. Like acetylcholine, glutamate and aspartate are excitatory and stimulate receptors on the postsynaptic membrane to transmit a nerve impulse. The details of glutamate processing at an excitatory synaptic junction are shown in Figure 34.67. In the resting state, the glutamate concentration in the extracellular space is approximately 1 mM, whereas in the presynaptic cytoplasm and the lumen of the synaptic vesicles, the glutamate concentrations are about 10 mM and 100 mM, respectively. These concentrations are maintained by specific carrier proteins. The carriers in the presynaptic and glial plasma membranes are Na⁺-dependent and have K_m values for glutamate as low as 2 mM. ATPases in the presynaptic vesicle membranes establish a proton gradient, which provides the driving force for accumulation of glutamate anions by a vesicle membrane transporter that has a K_m for glutamate in the millimolar range. Arrival of a nerve impulse triggers Ca²⁺-dependent exocytosis, in a manner similar to acetylcholine release, followed by binding to glutamate receptors on the postsynaptic membrane. However, because no enzymes that degrade glutamate (in analogy to acetylcholinesterase) exist in the extracellular space, glutamate must be cleared by the high-affinity presynaptic and glial transporters—a process called reuptake. Glutamate taken up by glial cells is converted to glutamine by glial glutamine synthetase. Glutamine, which does not display neurotransmitter activity, is present in the extracellular space at about 0.5 mM and is taken into the presynaptic neuron by an Na⁺-independent carrier. It can be reconverted to glutamate by mitochondrial glutaminase and then accumulated in synaptic vesicles.

Figure 34.68 · Four classes of glutamate receptors. (a) The NMDA receptor is an Na⁺ and Ca²⁺ channel, which is regulated by Zn²⁺ and glycine, stimulated by N-methyl-D-aspartate, and inhibited by phencyclidine (PCP) and the anticonvulsant drug MK-801. (b) The kainate receptor is an Na⁺ and K⁺ channel. (c) The AMPA receptor is an Na⁺ channel activated by a-amino-3-hydroxy-5-methyl-4-isoxazolepropionic acid (AMPA), and (d) the metabotropic receptor is a G protein-dependent receptor that is stimulated by ibotenic acid. (Adapted from Young, A., and Fagg, G., 1990. Excitatory amino acid receptors in the brain. Membrane binding and receptor autoradiographic approaches. Trends in Pharmacological Sciences 11:126-133.)

            There are at least five subclasses of glutamate receptors known. Four of these are identified by specific antagonist effects on glutamate receptors (Figure 34.68). The N-methyl-D-aspartate (NMDA), kainate, and AMPA receptors are ligand-gated ion channels. The metabotropic receptors are G protein-mediated receptors that are coupled to phosphatidylinositol metabolism, much in the manner of the muscarinic acetylcholine receptors. The best understood of these excitatory receptors is the NMDA receptor. The NMDA receptor is a ligand-gated channel that, when open, allows Ca²⁺ and Na⁺ to flow into the cell and K⁺ to flow out of the cell. A unique property of these ion channels is that they are closed by Mg²⁺ ions in a voltage-dependent fashion. Phencyclidine (PCP) is a specific antagonist of the NMDA receptor (Figure 34.68), and, unlike Mg²⁺, its inhibition is not voltage-dependent. Phencyclidine was once used as an anesthetic agent, but legitimate human use was quickly discontinued when it was found to be responsible for bizarre psychotic reactions and behavior in its users. Since this time, PCP has been used illegally as a hallucinogenic drug, under the street name of angel dust. Sadly, it has caused many serious, long-term psychological problems in its users.

g-Aminobutyric Acid and Glycine: Inhibitory Neurotransmitters

Certain neurotransmitters, acting through their conjugate postsynaptic receptors, inhibit the postsynaptic neuron from propagating nerve impulses from other neurons. Two such inhibitory neurotransmitters are g-aminobutyric acid (GABA) and glycine. These agents make postsynaptic membranes permeable to chloride ions and cause a net influx of Cl^-, which in turn causes hyperpolarization of the postsynaptic membrane (making the membrane potential more negative). Hyperpolarization of a

Figure 34.69 · GABA (g-aminobutyric acid) and glycine are inhibitory neurotrans­mitters that activate chloride channels. Influx of Cl^- causes a hyperpolarization of the post­synaptic membrane.

neuron effectively raises the threshold for the onset of action potentials in that neuron, making the neuron resistant to stimulation by excitatory neurotransmitters. These effects are mediated by the GABA and glycine receptors, which are ligand-gated chloride channels (Figure 34.69). GABA appears to operate mainly in the brain, whereas glycine acts primarily in the spinal cord. The effects of ethanol on the brain arise in part from the opening of GABA receptor Cl^- channels. GABA is derived by a decarboxylation of glutamate and is broken down by successive aminotransferase and dehydrogenase reactions (Figure 34.70). The GABA receptor is a hetero-oligomer of 220 to 400 kD.

Figure 34.70 · Glutamate is converted to GABA by glutamate decarboxylase. GABA is degraded by the action of GABA-glutamate transaminase and succinate semialdehyde dehydrogenase to produce succinate.

Figure 34.71 · Glycine receptors are distinguished by their unique affinity for strychnine. A model for the arrangement of the 48-kD subunit of the glycine receptor in the postsynaptic membrane is shown.

           The glycine receptor can be distinguished on the basis of its specific affinity for the convulsive alkaloid strychnine (Figure 34.71). Purified glycine receptor from spinal cord tissue is a glycoprotein containing polypeptides of 48 kD, 58 kD, and 93 kD. The 48-kD subunit contains the strychnine-binding site. The 48-kD and 58-kD subunits are homologous integral membrane proteins that are thought to form the Cl^- channel core of the receptor, whereas the 93-kD subunit is a peripheral membrane protein located on the cytoplasmic face of the postsynaptic glycine receptor complex. The sequence of the 48-kD subunit is homologous with that of the GABA receptor and the nicotinic acetylcholine receptor.

The Catecholamine Neurotransmitters

Figure 34.72 · The pathway for the synthesis of catecholamine neurotransmitters. Dopa, dopamine, noradrenaline, and adrenaline are synthesized sequentially from tyrosine.

Epinephrine, norepinephrine, dopamine, and L-dopa are collectively known as the catecholamine neurotransmitters. These compounds are synthesized from tyrosine (Figure 34.72), both in sympathetic neurons and in the adrenal glands. They function as neurotransmitters in the brain and as hormones in the circulatory system. However, these two pools operate independently, thanks to the blood-brain barrier, which permits only very hydrophobic species in the circulatory system to cross over into the brain. (In spite of this, the brain receptors are referred to as adrenergic receptors.) Hydroxylation of tyrosine (by tyrosine hydroxylase) to form 3,4-dihydroxyphenylalanine (L-dopa) is the rate-limiting step in this pathway. Dopamine, a crucial catecholamine involved in several neurological diseases, is synthesized from L-dopa by a pyridoxal phosphate-dependent enzyme, dopa decarboxylase. Subsequent hydroxylation and methylation produce norepinephrine and epinephrine (Figure 34.72). The methyl group in the final reaction is supplied by S-adenosylmethionine.
            Each of these catecholamine neurotransmitters is known to play unique roles in synaptic transmission. The neurotransmitter in junctions between sympathetic nerves and smooth muscle is norepinephrine. On the other hand, dopamine is involved in other processes. Either excessive brain production of dopamine or hypersensitivity of dopamine receptors is responsible for psychotic symptoms and schizophrenia, whereas lowered production of dopamine and the loss of dopamine neurons are important factors in Parkinson’s disease.
            At least three different dopamine receptors (denoted D₁, D₂, and D₃) have been characterized. These dopamine receptor subtypes are homologous with one another, as well as with the b-adrenergic receptors, and are putative 7-TMS proteins of 446 residues. All possess the conserved Asp residue in the third transmembrane segment, which is found in the b-adrenergic receptor sequences. In addition, they contain two Ser residues in the fifth transmembrane segment that are conserved among catecholamine receptors and are critical for the recognition of agonist ligands possessing a catechol group. D₁ receptors stimulate adenylyl cyclase. D₂ receptors have been linked to inhibition of adenylyl cyclase, inhibition of phosphatidylinositol turnover, activation of K⁺ channels, and inhibition of Ca²⁺-channel activity. D₃ receptors are similar to D₂ receptors.

Many relatively small peptides have been shown to possess neurotransmitter activity (see Table 34.6). One of the challenges of this field is that the known neuropeptides may represent a very small subset of the neuropeptides that exist. Another challenge arises from the small in vivo concentrations of these agents and the small number of receptors that are present in neural tissue. Physiological roles for most of these peptides are complex. For example, the endorphins and enkephalins are natural opioid substances and potent pain relievers. The endothelins are a family of homologous regulatory peptides, synthesized by certain endothelial and epithelial cells, that act on nearby smooth muscle and connective tissue cells. They induce or affect smooth muscle contraction; vasoconstriction; heart, lung, and kidney function; as well as mitogenesis and tissue remodeling. Vasoactive intestinal peptide (VIP) produces a G protein-adenylyl cyclase-mediated increase in cAMP, which in turn triggers a variety of protein phosphorylation cascades, one of which leads to conversion of phosphorylase b to phosphorylase a, stimulating glycogenolysis. Moreover, VIP has synergistic effects with other neurotransmitters, such as norepinephrine. In addition to increasing cAMP levels through b-adrenergic receptors, norepinephrine acting at a₁-adrenergic receptors markedly stimulates the increases in cAMP elicited by VIP. Many other effects have also been observed. For example, injection of VIP increases rapid eye movement (REM) sleep and decreases waking time in rats. How these complex effects are mediated is not understood, but it has been shown that VIP receptors exist in regions of the central nervous system involved in sleep modulation.

Human Biochemistry
The Biochemistry of Neurological Disorders
Defects in catecholamine processing are responsible for the symptoms of many neurological disorders, including clinical depression (which involves norepinephrine) and Parkinsonism (involving dopamine). Norepinephrine (NE) and dopamine (DA) (see Figure 34.72) are processed much the way glutamate is (Figure 34.67). Once these neurotransmitters have bound to and elicited responses from postsynaptic membranes, they must be efficiently cleared from the synaptic cleft (see figure below, part a). Clearing can occur by several mechanisms. NE and DA transport or reuptake proteins exist both in the presynaptic membrane and in nearby glial-cell membranes. On the other hand, catecholamine neurotransmitters can be metabolized and inactivated by two enzymes: catechol-O-methyl-transferase in the synaptic cleft and monoamine oxidase in the mitochondria (see figure part b). Catecholamines transported back into the presynaptic neuron are accumulated in synaptic vesicles by the same H1-ATPase/H1-ligand exchange mechanism described for glutamate. Clini­cal depression has been treated by two different strategies. Monoamine oxidase inhibitors act as antidepressants by in­creasing levels of catecholamines in the brain. Another class of antidepressants, the tricyclics, such as desipramine (see figure part c) act on several classes of	neurotransmitter reuptake transporters and facilitate more prolonged stimulation of postsynaptic receptors. Prozac is a more specific reuptake inhibitor and acts only on serotonin reuptake transporters.         Parkinsonism is characterized by degeneration of dopaminergic neurons, as well as consequent overproduction of postsynaptic dopamine receptors. In recent years, Parkinson’s patients have been treated with dopamine agonists such as bromocriptine (see figure part d) to counter the degeneration of dopamine neurons.         Catecholaminergic neurons are involved in many other interesting pharmacological phenomena. For example, reserpine (see figure part e), an alkaloid from a climbing shrub of India, is a powerful sedative that depletes the level of brain monoamines by inhibiting the H1-monoamine exchange protein in the membranes of synaptic vesicles.         Cocaine (see figure part f), a highly addictive drug, binds with high affinity and specificity to reuptake transporters for the monoamine neurotransmitters in presynaptic membranes. Thus, at least one of the pharmacological effects of cocaine is to prolong the synaptic effects of these neurotransmitters.
	(a) The pathway for reuptake and vesicular repackaging of the catecholamine neurotransmitters. The sites of action of desipramine, tranylcypromine, and reserpine are indicated. (b) Norepinephrine can be degraded in the synaptic cleft by catechol-O-methyltransferase or in the mitochondria of presynaptic neurons by monoamine oxidase. (c) The structures of tranylcypromine, desipramine, and Prozac. (d) The structure of bromocriptine. (e) The structure of reserpine. (f) The structure of cocaine.