HOME HELP FEEDBACK SUBSCRIPTIONS ARCHIVE SEARCH TABLE OF CONTENTS		QUICK SEARCH:   [advanced] Author: Keyword(s): Year:  Vol:  Page:

This Article Abstract Full Text (PDF) Alert me when this article is cited Alert me if a correction is posted Citation Map Services Email this article to a friend Similar articles in this journal Similar articles in ISI Web of Science Similar articles in PubMed Alert me to new issues of the journal Download to citation manager Citing Articles Citing Articles via HighWire Citing Articles via ISI Web of Science (71) Citing Articles via Google Scholar Google Scholar Articles by Seino, S. Articles by Shibasaki, T. Search for Related Content PubMed PubMed Citation Articles by Seino, S. Articles by Shibasaki, T.

PKA-Dependent and PKA-Independent Pathways for cAMP-Regulated Exocytosis

Division of Cellular and Molecular Medicine, Kobe University Graduate School of Medicine, Kobe, Japan

ABSTRACT

I. INTRODUCTION

II. FUNDAMENTAL FEATURES OF REGULATED EXOCYTOSIS

A. Neurons

B. Endocrine and Neuroendocrine Cells

C. Exocrine Cells

III. INTRACELLULAR cAMP METABOLISM

A. Adenylyl Cyclases

B. Phosphodiesterases

IV. EFFECTORS OF cAMP AND THEIR PROPERTIES

A. Structure of cAMP-Binding Proteins

B. cAMP-Dependent Protein Kinase

C. Cyclic Nucleotide-Gated Channels

D. cAMP-GEF/Epac

E. CRP/CAP

V. PHARMACOLOGICAL AGENTS TARGETING THE cAMP PATHWAY

A. cAMP Analogs

B. Inhibitors Targeting Catalytic Subunits of PKA

VI. PKA-DEPENDENT EFFECTS OF cAMP ON REGULATED EXOCYTOSIS

A. Neurons

B. Pituitary Cells

C. Adrenal Chromaffin Cells

D. Pancreatic {beta}-Cells

E. Exocrine Acinar Cells

VII. PKA-INDEPENDENT EFFECTS OF cAMP ON REGULATED EXOCYTOSIS

A. Pancreatic {beta}-Cells

B. Neurons

VIII. A MODEL FOR cAMP-REGULATED EXOCYTOSIS

IX. CONCLUDING REMARKS

GRANTS

ACKNOWLEDGMENTS

REFERENCES

	ABSTRACT

Top Next References

	I. INTRODUCTION

Top Previous Next References

 
The fusion of secretory vesicles to the plasma membrane in multicellular organisms is a crucial event in regulated exocytosis and is tightly controlled to release vesicle contents in response to specific signals, often in a specialized region of the plasma membrane (78). There are two major types of regulated exocytosis: secretory granule (large dense-core vesicle, LDCV) exocytosis, as in neuroendocrine, endocrine, and exocrine cells, and synaptic vesicle (SV) exocytosis, which occurs in neurons. In some neurons and endocrine cells, both LDCV and SV exocytosis are found (32, 205, 358, 362, 380, 445, 476, 512, 562, 582, 620). They can be distinguished by morphological appearance of secretory vesicles and by kinetics of release (79, 522). Stimulus-secretion coupling is the major biological process in the many secretory cells in which regulated exocytosis occurs. While a rise in intracellular Ca²⁺ concentration ([Ca²⁺]_i) is generally the trigger for exocytosis, other intracellular signals including cAMP, diacylglycerol (DAG), phospholipids, and ATP also regulate or modulate Ca²⁺-triggered exocytosis (85, 306, 433). Among these, cAMP is well known to regulate exocytosis in a variety of secretory cells (79, 85, 112, 177, 306, 322, 334, 353, 365, 393, 433, 469, 477, 553). In neurons, cAMP has been shown to induce long-term potentiation (LTP) (44, 329, 400) by increasing neurotransmitter release at mossy fiber synapses in the hippocampus of cerebrum (552, 581) and parallel fiber-Purkinje neuron synapses in the cerebellum (104, 472). cAMP increases transmitter release at many synapses in vertebrate peripheral ganglia and invertebrate nervous system, including sympathetic (64, 317) and parasympathetic ganglion neurons (55), neuromuscular junctions of crayfish (618), central synapses of Aplysia (73, 90, 279, 308, 505), and neuromuscular junctions of Drosophila melanogaster (Drosophila) (322, 604). cAMP also regulates release of various hormones in endocrine cells, including pancreatic hormones such as insulin from pancreatic -cells (60, 226, 230, 365, 433, 496, 524), glucagon from pancreatic -cells (140, 208), pituitary hormones such as adrenocorticotropin (ACTH) from pituitary corticotrophs (10, 357, 588), and catecholamines from adrenal chromaffin cells (107, 413, 434, 586). In exocrine parotid acinar cells, cAMP rather than Ca²⁺ is the primary signal in amylase release (177, 441).

PKA has been thought to be the major target of cAMP in cAMP-regulated exocytosis in multicellular organisms. However, cAMP is now known to have other targets as well, including cyclic nucleotide-gated (CNG) channels (294), hyperpolarization-activated cyclic nucleotide-gated (HCN) channels (43), and cAMP-specific guanine nucleotide exchange factors (cAMP-GEF)/exchange proteins directly activated by cAMP (Epac) (hereafter cAMP-GEF/Epac) (53, 514). Although the PKA-dependent mechanisms of regulation and modulation of exocytosis by cAMP have been studied extensively (79, 85, 112, 177, 306, 322, 334, 353, 365, 393, 477, 553), the PKA-independent mechanisms are currently being unveiled. In this review, we discuss both the PKA-dependent and the PKA-independent pathways of cAMP-regulated exocytosis and suggest a mechanism of differential implementation of these two pathways within the cell.

	II. FUNDAMENTAL FEATURES OF REGULATED EXOCYTOSIS

Top Previous Next References

 
Regulated exocytosis has been well studied in both neurons and nonneuronal cells. In neurons, neurotransmitters are released by fusion of SVs with the presynaptic plasma membrane. Nonneuronal secretory cells possess LDCVs (dense-core granules or secretory granules) containing various bioactive substances such as peptide hormones, amines, and enzymes, the contents of which exert diverse biological effects on cellular functions. Despite differences in the time course, Ca²⁺ dependency, and signal input between SV and LDCV exocytosis, both involve common processes: vesicle recruitment to the plasma membrane, docking of vesicles at the plasma membrane, priming of fusion machinery, and fusion of vesicles with the plasma membrane. Although these processes are central to understanding exocytosis, they are difficult to distinguish by currently available methods. In neurons, release of neurotransmitters is triggered primarily by Ca²⁺ influx through voltage-dependent Ca²⁺ channels (VDCCs) in response to action potentials, while in secretory cells such as neuroendocrine, endocrine, and exocrine cells, various intracellular signals in addition to the Ca²⁺ signal also participate in stimulus-secretion coupling. Exocytosis has been extensively investigated in neurons (18, 86, 522), but the detailed molecular mechanism is still largely unknown. Although basic components of the regulated exocytotic apparatus are highly conserved among different secretory cell types, the secretory processes are specialized and distinct (79, 235). Since the molecular mechanisms of regulated exocytosis have been extensively reviewed recently (79, 343, 346, 392, 452, 522), we describe only the principal components here (Fig. 1). We discuss characteristic features of stimulus-secretion coupling in various cell types.

View larger version (23K): [in this window] [in a new window]   FIG. 1. Model of regulated exocytosis. Key components involved in docking, priming, and fusion steps are indicated. Proteins are color coded: VAMP (light blue), synaptotagmin (red), syntaxin (green), SNAP-25 (orange), Munc18 (yellow), unknown tethering protein (light green), Rab3 (purple), Rim (gray), Munc13 (dark blue), and voltage-dependent Ca2+ channel (pink).

Action potential-induced Ca²⁺ influx is the principal signal that triggers synaptic vesicle exocytosis. In some neurons, SVs and LDCVs, which store low-molecular-weight transmitters and neuropeptides, respectively, are present together (32, 205, 358, 562, 620). For example, in certain neurons in which ACh and vasoactive intestinal polypeptide (VIP) coexist, stimulation of muscarinic cholinergic autoreceptors inhibits ACh and VIP release, while VIP enhances ACh release, suggesting that release of one of the coexisting transmitters modulates release of the other (32). Exocytosis of SVs and LDCVs occurs in response to distinct and specific signals, and the Ca²⁺ threshold for initiation, kinetic properties, and requirement for release sites differ in the two types of secretion, suggesting distinct regulatory mechanisms (562).

The conserved protein components of the exocytotic machinery in neurons include SNAREs [soluble N-ethylmaleimide-sensitive factor attachment protein (SNAP) receptors], ATPase N-ethylmaleimide-sensitive factor (NSF), and its cofactor -SNAP, Munc18/Sec1, Munc13, synaptotagmins, and Rab3 and its effectors. SNAREs are membrane proteins characterized by an -helical coiled-coil domain consisting of 60 amino acids that is called the SNARE motif (269, 459, 465, 580). The human genome contains 36 SNAREs (45, 269). SNAREs were initially classified as v-SNAREs and t-SNAREs, based on their localization on the vesicular and target plasma membrane, respectively (511). More recently, they were reclassified as arginine (R)-SNAREs and glutamine (Q)-SNAREs, depending on whether conserved arginine or glutamine is present in the center of the SNARE motif (161). Formation of the SNARE complex is proposed to mediate membrane fusion (38, 465). The SNARE complex comprises three proteins: vesicle-associated membrane protein (VAMP-2) (synaptobrevin), syntaxin 1, and the 25-kDa synaptosomal-associated protein (SNAP-25) (Fig. 1). VAMP-2 is an R-SNARE originally identified as an integral membrane protein of synaptic vesicles (520, 551). Syntaxin 1 is a Q-SNARE that was found as a presynaptic plasma membrane protein (30, 262). SNAP-25 is a Q-SNARE originally identified as a brain-specific neuronal membrane protein (414). VAMP-2 and syntaxin 1 each contains a single SNARE motif, while SNAP-25 contains two SNARE motifs. The four motifs from these proteins form an extremely stable ternary complex (SNARE complex). Neuronal SNAREs have been shown to interact with many other vesicle-associated proteins including Munc18/Sec1 (224), Munc13 (41), synaptotagmins (200), and complexin/synaphin (263, 377). Genetic disruption of neuronal SNAREs in Drosophila, Caenorhabditis elegans, and mice shows that SNARE is essential for evoked synaptic transmission but is not always involved in spontaneous synaptic events (67, 133, 403, 484, 487, 572), indicating that the exocytotic processes of evoked and spontaneous release differ in their requirement for the SNARE complex.

To recycle the individual proteins of the SNARE complexes, they must be disassembled after exocytosis. NSF and -SNAP act to disassemble the complexes (29). NSF is a hexameric protein that belongs to the AAA+ ATPase superfamily of chaperone-like ATPases (221, 372, 558, 622). -SNAP, an adaptor protein, interacts with the assembled SNARE motifs, allowing subsequent binding of NSF (221, 239, 583). Hydrolysis of ATP by NSF then dissociates the complex into its individual components. -SNAP stimulates the ATPase activity of NSF as well as recruits NSF to the SNARE complex (29, 225, 371, 388, 515).

Munc18/Sec1 is a family of hydrophilic proteins with no recognizable motifs (217, 268). The first member of the family was discovered in C. elegans in the UNC-18 mutant (Munc18) (63, 247) and was later found in yeast in the first secretory mutant (Sec1) (405). The mammalian homologs, Sec1 and Munc18, were also found (217). The human genome contains seven UNC-18/Sec1 homologs, three of which are Munc18–1, 18–2, and 18–3. Munc18–1 is predominantly expressed in neuronal and endocrine tissues, while Munc18–2 and Munc18–3 are ubiquitously expressed (194, 196, 224, 427, 542). Disruption of Munc18–1 in mice completely abolishes neurotransmission without affecting SV docking, suggesting a role of Munc18 in a postdocking step (561). In contrast, in adrenal chromaffin cells of Munc18–1 knockout mice, LDCV exocytosis is reduced, with fewer docked vesicles, suggesting that Munc18 functions at the docking step in LDCV exocytosis (564). Munc18 specifically binds to the H_abc domain of syntaxin 1 in the closed conformation, which prevents formation of the SNARE complex (147, 384, 599). Although Munc18 was initially proposed to function as a negative regulator of membrane fusion by inhibiting SNARE complex assembly (426, 488), recent studies favor the possibility that Munc 18 regulates transition of syntaxin 1 from closed to open conformation, thereby facilitating SNARE complex assembly.

UNC-13 was originally identified in C. elegans (475). Munc13 (four isoforms, Munc13–1, 13–2, 13–3, and 13–4) is a mammalian homolog of C. elegans UNC-13 (70, 117, 311). Munc13 interacts with the H_abc domain of syntaxin 1 (41). Munc13 is thought to prime synaptic vesicles by hindering the interaction between syntaxin 1 and Munc18, thereby promoting activation of syntaxin 1 for the formation of the SNARE complex in the active zone at the presynaptic terminal membrane (16, 41, 203, 475). Synaptic transmission of Drosophila UNC-13 has also been studied, and found to be essential for synaptic transmission (13). Munc13 is also involved in DAG-mediated, PKC-independent neurotransmitter release through its C₁ domain (325, 383, 406, 454). In addition to interacting with syntaxin 1, Munc13 interacts with several other proteins such as Rim1 (42) and Doc2 (412, 471).

Rab proteins are members of the Ras superfamily of small G proteins that function in vesicular transport (428, 490, 614). Rab proteins cycle between an inactive state (the GDP-bound form localized in the cytosol) and an active state (the GTP-bound form localized in the membrane). In the human, there are at least 60 Rab isoforms (45, 422, 614). Of these, Rab3 has been implicated both in synaptic and secretory granule exocytosis (124, 197, 522). In mammals, there are four structurally related Rab3 isoforms, Rab3A, B, C, and D, which are differentially expressed. Rab3A is expressed at high levels in the brain (166, 269) and is involved in the regulation of various steps in synaptic vesicle trafficking including targeting, docking, and postdocking processes (197, 335, 404). The amount of evoked transmitter release per stimulus is enhanced in Rab3A knockout mice (198), and LTP in mossy fiber synapses of the CA3 region in the hippocampus is abolished (91). Quadruple knockout mice (lacking all Rab3 isoforms) are born alive but exhibit respiratory failures (480). These knockout mice exhibit no apparent changes in synapse structure or expression levels of proteins associated with SV exocytosis, except for the loss of Rabphilin3, a Rab3-binding protein. Analysis of cultured hippocampal neurons from these knockout mice revealed that Rab3 is not essential for synaptic membrane trafficking, but modulates basic release machinery.

The diverse effects of Rab3 are mediated by interaction with its multiple effectors (81, 119, 124, 199, 503, 522). Rab3 has several potential effectors, including Rabphilin3 (502), Rims (Rim1 and Rim2) (415, 570, 571), Noc2 (369), and Granuphilin (118). Among these, Rabphilin3 and Rim1 are expressed predominantly in the brain. Rabphilin3, a SV-associated protein, is reversibly recruited to synaptic vesicles by Rab3 (504). Rabphilin3 knockout mice do not exhibit any of the obvious phenotypes of regulated exocytosis. (481). Rim (now called Rim1) was discovered originally as a Rab3-interacting molecule by yeast hybrid screen and is an active-zone scaffolding protein that binds to Rab3 when synaptic vesicles dock with GTP-bound Rab3 (570). Rim2, which was found later as an isoform of Rim (415, 571), is expressed in endocrine cells as well as in the brain (see sect. IIB for details). Rim contains a Zinc finger domain and two C₂ domains. Rim1 knockout mice exhibit an altered short-term synaptic plasticity and an absence of LTP in mossy fiber synapses in the hippocampus (92). The Rim null mutation in C. elegans decreases the number of fusion-competent vesicles, suggesting a role in the postdocking process (315). Rim1 interacts with Munc13–1 (42). Disruption of this interaction causes a drastic decrease in the size of the readily releasable pool, suggesting that the interaction of Rim1 and Munc13–1 is involved in regulating synaptic vesicle priming (42). Rim1 and Rim2 interact with cAMP-GEFII/Epac2, a cAMP-binding protein exhibiting GEF activity toward Rap (415). In pancreatic -cells, the interaction of Rim2 and cAMP-GEFII/Epac2 mediates cAMP-dependent, PKA-independent insulin granule exocytosis (see sect. VII). This mechanism involving cAMP-GEFII/Epac2 may also be present in certain neurons in which cAMP modulates transmitter release, since cAMP-GEFII/Epac2 binds to Rim1, which is expressed predominantly in the brain (see sect. VII).

An increase in [Ca²⁺]_i almost universally triggers the fusion of secretory vesicles to the plasma membrane. Among many Ca²⁺-binding proteins that might function as Ca²⁺ sensors for vesicle fusion, synaptotagmin is the best characterized candidate (17, 98, 521). Synaptotagmin was originally reported as p65, a 65-kDa synaptic vesicle protein, and renamed after its cloning (370, 424). Synaptotagmins include 13 isoforms in humans (521), and the presence of an additional 6 potential isoforms has been suggested by database search (120). While most synaptotagmins are localized on transport vesicles, some (synaptotagmin III, VI, and VII) are present at the plasma membrane (23). Synaptotagmin has an NH₂-terminal transmembrane region followed by two C₂ domains (C₂A and C₂B) and binds phospholipids with considerable variations in Ca²⁺ dependence. In addition, synaptotagmin I also binds to syntaxin 1, SNAP-25, Ca²⁺ channels, and Rim1 (119). Studies of mutations in synaptotagmin I in mice, Drosophila, and C. elegans have shown that synaptotagmin I is essential only for fast, Ca²⁺-triggered release, but not for other modes of the exocytotic process such as stimulus-induced asynchronous release or spontaneous release, which suggests that synaptotagmin I functions at the Ca²⁺-sensing step for fast vesicle fusion (23, 125, 350). Synaptotagmin IX, which has close homology to synaptotagmin I, also binds to Ca²⁺, but does not associate with the SNARE complex, suggesting a role distinct from that of synaptotagmin I (500). The presence of different synaptotagmin isoforms with distinct Ca²⁺-binding affinities might account in part for variations in the effective concentration of Ca²⁺ in exocytosis seen in various types of secretory vesicles (288, 305). Vesicular synaptotagmins with low Ca²⁺ affinities may be more important for fast synaptic vesicle exocytosis, while plasma membrane synaptotagmins with a higher Ca²⁺ affinity may be more important in slower, secretory granule exocytosis (521).

B. Endocrine and Neuroendocrine Cells

In endocrine cells there is a large, reserved pool of secretory granules and a readily releasable pool constituting only a small fraction of the secretory granules. Stimulus-secretion coupling has been characterized in many endocrine and neuroendocrine cells, particularly insulin-secreting pancreatic -cells and catecholamine-secreting adrenal chromaffin cells. The pancreatic -cell plays a central role in glucose homeostasis by regulating insulin secretion. The details of the mechanism of insulin secretion have been reviewed recently (3, 202, 273, 463, 587). Ca²⁺, ATP, cAMP, phospholipids, and DAG are major intracellular signals in stimulus-secretion coupling in insulin granule exocytosis. The generally accepted model of glucose-induced insulin secretion is depicted in Figure 2. An increase in the ATP concentration (or ATP-to-ADP ratio) due to elevated glucose metabolism closes ATP-sensitive K⁺ (K_ATP) channels and depolarizes the -cell membrane, leading to opening of the VDCCs and resultant Ca²⁺ influx. The rise in [Ca²⁺]_i triggers exocytosis of the insulin granules. Sulfonylureas, widely used for treatment of diabetes mellitus, stimulate insulin secretion by closing the K_ATP channels directly. Thus the K_ATP channels are critical in glucose- and sulfonylurea-induced insulin release by coupling metabolic changes to electrical activities (15, 116, 381, 382, 491). Opening of the VDCCs represents a common step in insulin secretion induced by glucose, sulfonylureas, and amino acids (14). Modulation of VDCC activities affects insulin secretion (35, 252, 336). Although Ca²⁺ influx through VDCCs is indispensable in glucose-induced insulin granule exocytosis, it has been suggested that mobilization of intracellular Ca²⁺ from ryanodine-sensitive Ca²⁺ stores by cyclic ADP-ribose generated by glucose stimulation also contributes (292, 531). On the other hand, incretins such as gastric inhibitory polypeptide/glucose-dependent insulinotropic peptide (GIP) and glucagon-like peptide-1 (GLP-1) potentiate insulin secretion through cAMP-PKA signaling (see sect. VID for details) (72, 243, 302). ACh, a major parasympathetic neurotransmitter, generates phospholipid-derived messengers. One of these, DAG, activates PKC (202). Activation of DAG-sensitive PKC by a DAG analog, phorbol ester, induces a prolonged insulin secretory response in pancreatic islets and insulin-secreting cell lines (344, 518, 555). Inositol 1,4,5-trisphosphate (IP₃), which is also produced by ACh stimulation, mobilizes Ca²⁺ from intracellular stores (202, 344). Free fatty acids act as signaling molecules as well as an energy source in insulin secretion (216). Among free fatty acids, long-chain free fatty acids in insulin-secreting MIN6 cells potentiate glucose-induced insulin secretion through direct activation of GPR40 (264), an orphan G protein-coupled receptor (GPCR) (65, 264). Peptide agonists including somatostatin, norepinephrine, and galanin inhibit insulin secretion through the activation of pertussis toxin (PTX)-sensitive G protein G_i/o (3, 497, 595, 596).

View larger version (29K): [in this window] [in a new window]   FIG. 2. Model of insulin secretion. Glucose-induced insulin secretion and its potentiation constitute the principal mechanism of insulin release. Glucose is transported by the glucose transporter (GLUT) into the pancreatic -cell. Metabolism of glucose increases ATP production (and the ATP-to-ADP ratio), closing the KATP channels, resulting in membrane depolarization (), opening of the voltage-dependent Ca2+ channels (VDCCs), and Ca2+ influx, which triggers insulin granule exocytosis. Insulin granule exocytosis is also regulated by hormones and neurotransmitters, which generate intracellular signals such as cAMP, diacylglycerol (DAG), and inositol trisphosphate (IP3).

Because the GTPase-deficient mutation of each of the four Rab3 isoforms inhibited nutrient-induced secretion in hamster -cell line HIT-T15 cells (255), Rab3 was suggested to be involved in insulin granule exocytosis (255, 342, 446). This is confirmed by a finding in mice that knockout of Rab3A impairs glucose-induced insulin secretion (594). Noc2, a target of Rab3, suppresses the inhibitory effect of G_i protein signaling on insulin secretion, thereby playing an important role in the maintenance of normal insulin secretion (369). Disruption of the interaction between Rab3 and Noc2 in mice unmasks G_i protein signaling, resulting in reduced insulin secretion under conditions such as stress. Rab27, another member of the Rab family, also has been implicated in regulated exocytosis of LDCVs (602) and melanosomes (181). Overexpression of Rab27 increases insulin secretion in the insulin-secreting cell line MIN6 (602). Granuphilin, a recently identified protein associated with insulin granules (569), has been suggested to be a target of Rab27 (602). In addition to Rab27, Granuphilin also binds to syntaxin 1A (549) and Munc18 (118).

In pancreatic -cells, the secretory granules are mainly LDCVs, but synaptic-like microvesicles (SLMVs), which contain GABA, are also present (445). There are two major components, fast and slow, in Ca²⁺-triggered exocytosis in pancreatic -cells (526). It is proposed that fast exocytosis is associated with release of GABA, while slow exocytosis is associated with secretion of insulin, as assessed by capacitance measurement and amperometric detection of vesicular contents (526). GABA released from SLMVs may act as a paracrine inhibitor on adjacent glucagon-secreting -cells and somatostatin-secreting -cells, and serve as an autocrine regulator on pancreatic -cells (380, 476, 512, 582). The inhibitory effect of GABA on insulin secretion in pancreatic -cells occurs through G_i/o protein signaling (62). GABA release is regulated by glucose (362), suggesting that GABA-containing SLMVs also undergo regulated exocytosis. Overexpression of -SNAP in rat pancreatic islets enhances glucose-induced insulin secretion (395). Overexpression of mutant -SNAP lacking the binding region to syntaxin 1A in insulin-secreting MIN6 cells inhibits glucose-induced insulin release, but does not alter GABA release, indicating that the regulatory mechanism of exocytosis differs in insulin-containing LDCVs and GABA-containing SLMVs (395).

The adrenal gland plays an essential role under various stresses (19, 306). The chromaffin cells in the adrenal medulla secrete catecholamines (epinephrine and norepinephrine) and a number of neuropeptides, all of which are stored in dense-core vesicles called chromaffin granules. Chromaffin cells are excitable, generating action potentials in response to ACh (59) or electrical stimulation of the splanchnic nerve (26, 278). Innervation of adrenal chromaffin cells is principally cholinergic, by preganglionic sympathetic fibers in the splanchnic nerve, which originate in the intermediolateral cell column of the thoracic spinal cord (11, 241). Peptidergic nerves containing enkephalin, substance P, or VIP are also present (349, 486). Cholinergic fibers all seem to innervate both epinephrine- and norepinephrine-secreting cells, while enkephalin-containing fibers surround only epinephrine-secreting cells (242), thus regulating epinephrine secretion. In addition, pituitary adenylyl cyclase activating polypeptide (PACAP)-containing nerve terminals, the cell bodies of which are located in the sensory neurons of dorsal root ganglia and nodose ganglia, are present in the adrenal medulla. Components of SNARE proteins are also present in chromaffin cells. VAMP is found on the membrane of chromaffin granules (238, 240). VAMP-2 mediates MgATP-dependent catecholamine exocytosis from permeabilized chromaffin cells (331). In addition, syntaxin 1A and 1B are expressed in these cells (238, 240). In permeabilized chromaffin cells, an anti-syntaxin antibody inhibits the Ca²⁺-triggered catecholamine release, indicating that syntaxin also functions in exocytosis of chromaffin granules (215). SNAP-25 is expressed at high and low levels in noradrenergic and adrenergic chromaffin cells, respectively (283). The following mechanism of stimulus-secretion coupling in chromaffin cells is generally accepted. Upon activation of nicotinic ACh receptors, ACh opens the ionophore of the receptor, allowing influx of Na⁺ and, to a lesser extent, Ca²⁺, which results in a depolarization of the membrane. Depolarization opens the voltage-dependent, tetrodotoxin-sensitive Na⁺ channels (93), inducing further depolarization of the membrane and opening of the VDCCs (193).

Opening of both Na⁺ and Ca²⁺ channels enhances Ca²⁺ influx. The resultant rise in [Ca²⁺]_i triggers exocytosis of catecholamine secretory granules (59, 300, 301). The time constant for exocytosis in chromaffin granules has been reported to be 150–1,000 ms, much longer than that for exocytosis of synaptic vesicles (286). cAMP increases norepinephrine secretion in response to depolarization by high K⁺ or stimulation by ACh in a pathway distinct from that controlling internal Ca²⁺ levels in PC12 cells (368).

C. Exocrine Cells

Although many exocrine glands including salivary, gastric and intestinal glands, and exocrine pancreas are known to possess systems of regulated exocytosis, the molecular basis for stimulus-secretion coupling in exocrine cells has been characterized in only a few exocrine glands.

Pancreatic acinar cells synthesize, package, and release a variety of digestive enzymes and are the system in which the secretory pathway was first examined (416). The cells are highly polarized, with two distinct plasma membrane domains, an apical domain and a basolateral domain. Digestive enzymes are stored in secretory vesicles known as zymogen granules, the majority of which are present at the apical pole of the cell. Various secretagogues including a gastrointestinal hormone, cholecystokinin (CCK), neurotransmitters, ACh and VIP, and a mammalian bombesin homolog, neuromedin C, stimulate enzyme secretion by acting through the intracellular signals Ca²⁺, cAMP, and DAG (573, 585). A rise in [Ca²⁺]_i is the primary signal triggering fusion of the zymogen granule membrane to the apical membrane. The mode of Ca²⁺ signaling seems to depend on the concentration of the secretagogue (573). Low concentrations of secretagogues evoke an oscillatory pattern of [Ca²⁺]_i depending largely on Ca²⁺ release from intracellular stores (611). In contrast, high concentrations induce a completely different pattern of [Ca²⁺]_i change, consisting of a rapid initial rise followed by a decline to a sustained plateau (87). This pattern requires initial Ca²⁺release from intracellular stores followed by both Ca²⁺ extrusion from the cell and Ca²⁺ influx into the cell. Ca²⁺ extrusion is mediated by Ca²⁺-ATPase localized to the apical membrane (37), while Ca²⁺ influx is thought to be through transient receptor potential (TRP) channels, most likely residing in the basolateral membrane (437), by a mechanism of capacitative or store-mediated Ca²⁺ entry. Cyclic ADP ribose (545), arachidonic acid (327), and NAADP also modulate Ca²⁺ signals (88). cAMP-increasing ligands such as VIP and PACAP have been shown to potentiate CCK-induced zymogen granule exocytosis in isolated pancreatic acinar cells, as assessed by enzyme secretion (80, 227) and membrane capacitance measurements (477).

SNARE proteins are also present in pancreatic acinar cells. VAMP-2 is present on the membrane of zymogen granules (186). Three syntaxin isoforms are localized in acinar cells: syntaxin 2 on the apical plasma membrane, syntaxin 4 on the basolateral membrane, and syntaxin 3 on the zymogen granule membrane (184). The finding that botulinum toxin C (BoNT/C) treatment of plasma membrane completely cleaves syntaxin 2 and blocks granule-plasma membrane fusion indicates that syntaxin 2 mediates granule-plasma membrane fusion in pancreatic acinar cells (220). On the other hand, syntaxin 3 is involved in both granule-granule and granule-plasma membrane fusion, as BoNT/C treatment of granules completely cleaves syntaxin 3 and blocks both granule-granule and granule-plasma membrane fusion (184). SNAP-23, a widely expressed homolog of SNAP-25, is found on the basolateral plasma membrane domain, where it has been proposed to play a role in granule-membrane fusion in pancreatitis (187). On the other hand, SNAP-25 is not found in pancreatic acinar cells. Two isoforms of Munc18 (b and c) are present in pancreatic acinar cells. Both Munc18-b and Munc18-c appear to be localized on the basolateral plasma membrane, while Munc18-b is also present on the granule membrane (185). Among Rab3 members, Rab3D is present on the zymogen granule membrane (408). It has been shown in transgenic mice overexpressing Rab3D in pancreatic acinar cells that Rab3D is important in the initial phase of amylase secretion induced by CCK (409). On the other hand, an experiment overexpressing a dominant-negative form of Rab3D in pancreatic acinar cells indicates that Rab3D regulates terminal steps in exocytosis of zymogen granules (105, 409). However, a study of Rab3D knockout mice has suggested that Rab3D is not required for exocytosis of zymogen granules, but rather plays a role in the maintenance of granule maturation (408). Genetic disruption in mice of Noc2, a target of Rab3, Rab27, and Rab8 (180), causes no amylase response to stimuli and a marked accumulation of zymogen granules, suggesting that Noc2 is essential in stimulus-secretion coupling in pancreatic acinar cells (369).

In parotid acinar cells, stimulation of -adrenergic receptors results in intracellular cAMP accumulation, inducing exocytosis of amylase-containing granules without a rise in [Ca²⁺]_i, whereas a rise in [Ca²⁺]_i by activation of -adrenergic, cholinergic, or substance P receptors has only a mild effect on amylase release (25, 436, 532, 606). In perifusion experiments using isolated parotid acinar cells, amylase release induced by stimulation with isoproterenol reaches maximum at 6 min, followed by a gradual decrease in release (606). On the other hand, both carbachol and substance P cause a transient increase in amylase release (30–60 s), after which release is maintained at a steady-state level (606). Both cAMP signaling and Ca²⁺ signaling are proposed to act at different steps in amylase release (605). The combination of isoproterenol and substance P evokes biphasic amylase release (a first, large peak followed by a sustained plateau), and the amount of released amylase is greater than that induced by each agonist alone (605).

The basic components involved in exocytosis (including SNARE proteins, Rabs, and their related proteins) also are expressed in parotid acinar cells. Syntaxins, except for syntaxin 1, VAMPs (except for VAMP-7), SNAP-23, -SNAP, and NSF are all expressed (259). The major t-SNARE proteins, syntaxin 1 and SNAP-25, both of which are involved in regulated exocytosis in neuronal, neuroendocrine, and endocrine cells, are not expressed (179, 259). The SNARE-related proteins, synaptotagmins (III, IV, and XI) and Munc18s (1, 2, and 3), are also expressed (259). As cleavage of VAMP-2 by botulinum toxin B in parotid acinar cells inhibits isoproterenol-induced amylase release, SNARE proteins are involved in fusion of granules to plasma membrane in cAMP- as well as Ca²⁺-triggered exocytosis (179).

Rab3D and Rab27B are present in parotid acinar cells (260, 443). Introduction of wild type or the dominant active form of Rab3D into permeabilized parotid acinar cells inhibits cAMP and Ca²⁺-triggered amylase release (399). However, Rab3D knockout mice show that Rab3D is not essential for regulation of exocytosis in parotid acinar cells, but exerts regulatory effects on granule maturation. Rab27B regulates amylase release in parotid acinar cells through formation of a complex with its effector protein MyRIP/Slac-2, which was identified as a myosin Va/VIIa and actin-binding protein (151, 260). In Noc2 knockout mice, there is also a marked accumulation of secretory granules in salivary glands (571). Thus Rab3, Rab27, and their target proteins participate in exocytosis in parotid acinar cells.

	III. INTRACELLULAR cAMP METABOLISM

Top Previous Next References

 
Ligands such as hormones and neurotransmitters require at least four components to regulate the intracellular cAMP concentration: GPCR, heterotrimeric G protein, adenylyl cyclase, and cyclic nucleotide specific phosphodiesterase (PDE). Intracellular cAMP metabolism is determined primarily by activities of adenylyl cyclase and PDE. However, the intracellular cAMP concentration ([cAMP]_i) varies in amplitude, duration, and gradient within the cell in response to the different ligands. In addition, the presence of multiple isoforms of adenylyl cyclase and PDE, distinct kinetic and regulatory properties of each isoform, and unique distribution of the various isoforms in the cells all contribute to cAMP compartmentation (see sect. VIII) as well as to great diversity in the cAMP synthesis and degradation processes (see Refs. 114, 117, 128, 157, 250, 536, for recent reviews).

A. Adenylyl Cyclases

Adenylyl cyclase, an enzyme that synthesizes cAMP from ATP (431, 432, 461, 462), comprises a large superfamily (117, 128, 219, 523). Activation of adenylyl cyclase is induced mainly by G_s, an -subunit of heterotrimeric G proteins. In mammals, nine membrane-bound forms (type I-IX) and one soluble form expressed specifically in sperm have been identified (316, reviewed in Refs. 117, 128, 523). The membrane-bound forms of 120 kDa share a common structure composed of two cytosolic domains (C1 and C2) and two transmembrane domains (TM1 and TM2). Both C1 and C2 domains have 230 amino acid regions (designated C1a and C2a, respectively) that share more than 50% similarity and contribute to ATP binding and formation of the catalytic core (598). Types II, IV, VI, VII, and IX are widely expressed in tissues, while type V is expressed predominantly in heart. Ca²⁺/calmodulin (CaM)-regulated isoforms (type I, III, and VIII) are expressed in neurons and nonneuronal secretory cells, including hippocampus, cerebellum, pancreatic acini, and pancreatic islets (546, 568, 574, 591). G_s protein stimulates adenylyl cyclase activities of all isoforms, while G_i, G, Ca²⁺, Ca²⁺/CaM, PKA, PKC, and Ca²⁺/CaM-dependent kinase (CaMK) regulate the activities in an isoform-specific manner. G_i protein inhibits activities of types I, V, VI, and VIII (539). G subunits negatively regulate activity of type I and positively regulate activities of types IV and VII (188, 534, 610). Ca²⁺/CaM stimulates activities of types I and VIII by binding to the CaM-binding site in the cytosolic domain (212, 339, 565). Mice lacking type I or VIII exhibit defects in synaptic plasticity, including LTP, and in long-term memory formation (1, 478, 574, 589, 590). Types I and VIII are activated by Ca²⁺/CaM at the EC₅₀ values of 150–200 and 800 nM, respectively (402). Ca²⁺/CaM also inhibits activities of types I, III, and IX via CaMK IV (575), CaMK II (578), and calcineurin (421), respectively. At physiological submicromolar concentrations, Ca²⁺ directly inhibits activities of types V and VI (485, 608). Ca²⁺also stimulates activities of type V via its phosphorylation by PKC (295). Other intracellular messengers involved in regulation of adenylyl cyclase activity include DAG and cAMP. A DAG analog, phorbol ester, stimulates activities of types I, II, III, and VII by phosphorylation via PKC (110, 149, 266, 359, 507, 609). cAMP inhibits activities of types V and VI by phosphorylation via PKA (106, 265). Thus feedback inhibition of adenylyl cyclase by PKA phosphorylation may be involved in ligand-induced desensitization.

B. Phosphodiesterases

Cyclic nucleotide-specific PDEs are enzymes that hydrolyze the phosphodiester bond of cyclic nucleotides, thus playing an important role in the regulation of intracellular cyclic nucleotide levels (114, 157, 510). There are 25 PDE genes in mammals that encode a large number of isoenzymes, resulting in more than 50 different PDE proteins (114, 157, 439). This large superfamily of PDEs is subdivided into 11 distinct families based on structure, regulation, and kinetic properties. PDE families also are distinguished functionally by their unique pharmacological inhibitors (157, 298, 373, 556). They share common structural features, having targeting domains and regulatory domains in the NH₂-terminal region and a conserved catalytic domain consisting of 270–300 amino acids, usually located toward the COOH-terminal half of the protein. Among PDE isoforms, PDE1C, PDE3A and B, PDE4 (A, B, C, and D), PDE7A and B, PDE8A and B, PDE10A, and PDE11A are specific for cAMP (167, 223, 249, 378, 379, 438, 439, 509, 550, 597). The Michaelis constant (K_m) of these PDEs for cAMP is 0.06–4.00 µM. PDE4s, PDE7s, PDE8s, and PDE10A are expressed in the brain (167, 249, 379, 509, 550, 597). PDE1C, PDE3B, PDE4s, and PDE8A are expressed in a pancreatic -cell line and islets (2, 439, 494, 597).

PDE1s contain a Ca²⁺/CaM-binding site. Activities of PDE1s are strongly enhanced by Ca²⁺/CaM (99, 597). Activities of PDE4 are enhanced by its PKA-mediated phosphorylation in TSH-responsive thyroid, vascular smooth muscle, and lymphocytic cell lines (5, 150, 352, 493). Phosphorylation of PDE4D3 and PDE4D5 is known to contribute to cAMP homeostasis and cAMP signaling at a specialized region in the cell through their interaction with the complex of the A-kinase anchoring protein (AKAP) and PKA (115, 250, 411). cAMP-specific PDEs are involved in regulation of central nervous system activities, including hormone secretion, inflammatory response, and fertility and growth of mice (270, 407, 439, 550).

	IV. EFFECTORS OF cAMP AND THEIR PROPERTIES

Top Previous Next References

 
cAMP has been shown to be a universal intracellular messenger that mediates a wide variety of biological responses in almost all tissues in mammals (34). An elevation of [cAMP]_i activates several effectors of cAMP, leading to various cellular responses. PKA is the first effector that was characterized and has been studied extensively. cAMP is now known to have several other effectors, including cAMP receptor protein (CRP)/catabolite gene activator protein (CAP) (present in E. coli), CNG channels, HCN channels, and cAMP-GEF/Epac (135, 144, 155, 195, 294, 297, 356, 473, 538, 541, 624).

A. Structure of cAMP-Binding Proteins

cAMP-binding proteins share a common cyclic nucleotide monophosphate (cNMP)-binding domain (Fig. 3) consisting of a stretch of 120 amino acids. The three-dimensional structure of CRP/CAP has been determined by X-ray crystallography (375) and may be a model of other cAMP-binding proteins. The cAMP-binding domain of CRP/CAP comprises three -helices (A, B, and C) and an eight -stranded anti-parallel -barrel structure (1 to 8). cNMP-binding occurs in a phosphate-binding cassette formed by the COOH-terminal -helices and the -barrel (375) (Fig. 4). Cyclic nucleotide binds to this domain through a network of polar and nonpolar interactions. In addition to the cAMP-binding domain, these cAMP-binding proteins have other functional domains (Fig. 4). Binding of cAMP induces activation of these functional domains in CRP/CAP and cAMP-GEF/Epac (420, 448), and also causes dissociation and activation of the catalytic subunit from the regulatory subunit and the catalytic subunit complex of PKA (271, 538, 541).

View larger version (20K): [in this window] [in a new window]   FIG. 3. Functional domains of cNMP-binding proteins. cAMP, cAMP-binding domain; DNA, DNA-binding domain; cNMP, cNMP-binding domain; CaM, CaM-binding domain; Dimerization, dimerization domain; AKAP, AKAP-binding domain; C, region of interaction with the catalytic subunit; cAMP-A, cAMP-binding domain A; cAMP-B, cAMP-binding domain B; DEP, domain found in Dishevelled, Egl-10, and Plekstrin; REM, Ras exchange motif; GEF, guanine nucleotide exchange factor.

View larger version (47K): [in this window] [in a new window]   FIG. 4. Alignment of cNMP-binding domains. A: secondary structure of PKA RI and CRP/CAP are indicated according to their crystal structure: , -helix; , -barrel. Phosphate-binding cassette is indicated by red letter. [Modified from Su et al. (519).] B: amino acid sequence of phosphate-binding cassette and hinge region from various cNMP-binding domains. Conserved amino acid residues are shown with black background. Leucine (L) and phenylalanine (F) (indicated by dot) in the phosphate-binding cassette and hinge region, respectively, are proposed to play an important role in conformational change upon cNMP binding (448). cAMP-binding domain A of rat PKA RI (accession no. M17086), cAMP-binding domain A of rat PKA RII (J02934), cAMP-binding domain B of rat PKA RI (M17086), cAMP-binding domain B of rat PKA RII (J02934), cAMP-binding domain of rat cAMP-GEFI (AAD12739), cAMP-binding domain A of mouse cAMP-GEFII (AB021132), cAMP-binding domain B of mouse cAMP-GEFII (AB021132), cAMP-binding domain of E. coli CAP (J01598), cNMP-binding domain of rat CNGA1 (Q62927), cNMP-binding domain of rat CNGB1a (CAA04133), cNMP-binding domain of rat HCN1 (AF028737), and cGMP-binding domain A of mouse PKG IB (AAD16044).

In eukaryotic cells, PKA is the best-characterized protein kinase and has served as a model of the structure and regulation of cAMP-binding proteins as well as of protein kinases (541). Numerous endogenous and exogenous ligands activate PKA by binding to GPCR. PKA is a heterotetramer composed of two regulatory and two catalytic subunits. PKA isozymes were originally identified as type I and type II, which differ in the content of the regulatory subunits RI and RII. Molecular cloning revealed the presence of heterogeneity in four different types of regulatory subunit (RI, RI, RII, and RII) and four different types of C subunit (C, C, C, and PrKX) (541, 621). In addition, the regulatory subunits can form both homodimers and heterodimers, further contributing to the molecular diversity of PKA. The inactive holoenzyme is an R₂C₂ tetramer. The catalytic subunits inhibit binding of cAMP to the phosphate-binding cassette in the regulatory subunits (62). cAMP binds cooperatively to two sites, A and B, which are present on the regulatory subunit. In the inactive holoenzyme, only the B site is exposed for cAMP binding. Upon stimulation, occupation of the B site by cAMP induces binding of cAMP to the A site (172). Binding of four cAMP molecules to the inactive R₂C₂ tetramer (two cAMP molecules to each regulatory subunit) leads to a conformational change and dissociation into a regulatory subunit dimer (with four bound cAMP molecules) and two catalytic subunit monomers (314), which in turn become catalytically active (271, 508). PKA type I holoenzymes (RI₂C₂, RI₂C₂) have relatively higher affinities for cAMP, while PKA type II holoenzymes (RII₂C₂, RII₂C₂) have lower affinities. The presence of PKA isozymes having distinct biochemical properties and tissue distributions is the basis for the functional diversity and specificity of the effects of PKA. PKA activities are modulated by the expression level and localization of the regulatory or catalytic subunit (94, 248). Intracellular targeting and compartmentation of PKA is determined mainly by association with AKAPs, a family of structurally related proteins consisting of more than 50 members (537). AKAPs target PKA to specific substrates and specialized subcellular compartments (31, 113, 537). AKAPs also serve as scaffolding proteins that assemble PKA with other PKA signaling regulatory proteins such as phosphatases and cAMP-specific PDE. Thus spatial and temporal integration of PKA signaling components as a complex in a particular compartment of a cell is required for the precise regulation of PKA signaling (31, 113, 537).

C. Cyclic Nucleotide-Gated Channels

CNG channels were first discovered in the plasma membrane of the outer segment of rod photoreceptors in vertebrates, in which they are essential for generation of the primary electrical signal in photoreceptor response to light (163). CNG channels were found later in various tissues including kidney, testis, heart, and brain (68, 579, 601). CNG channels are nonselective cation channels that mediate Ca²⁺ and Na⁺ influx in response to direct binding of intracellular cyclic nucleotides (601). In vertebrates, six members of the CNG channel gene family have been identified (163). Based on sequence similarity, they are classified into two groups: CNGA (A1, A2, A3, and A4) and CNGB (B1 and B3) (56, 294). The pore-forming CNGA subunits form functional channels by themselves, while the CNGB subunits do not (351, 560). However, CNGA4 does not form a functional channel by itself, and functions as a modulator of olfactory CNG channels (50, 57, 138, 430). CNGA1 and CNGB1 are subunits of rod channels (616, 617), while CNGA3 and CNGB3 are subunits of cone channels (49, 201). The cNMP-binding domain located in the COOH-terminal region regulates activities of these channels (169, 272). CNG channels exhibit a high degree of cyclic nucleotide specificity. CNGA1 channels have the highest, intermediate, and lowest affinity for cGMP, cIMP, and cAMP, respectively, while CNGA2 channels have higher and lower affinity for cGMP and cAMP, respectively. Native olfactory CNG channels and heteromeric channels composed of CNGA2 and CNGA4 have similar affinities for both cAMP and cGMP (9, 57, 397, 495). Olfactory sensory neurons express a variety of GPCRs that activate the G protein G_olf, which is similar to the stimulatory G_s protein. Activated G_olf then increases the activity of adenylyl cyclase type III, leading to cAMP production. The resultant increase in cAMP concentration opens the CNG channels (174, 320, 397), allowing influx of Ca²⁺ and Na⁺, which depolarizes the neuron and causes outward Cl^– flux (307, 320, 355). In addition to the electrical excitation of neurons, influx of Ca²⁺ through CNG channels induces hormone release (563) and protein phosphorylation in presynaptic terminals (389).

Recently, other members of the CNG channel superfamily, called HCN channels, have been identified. HCN channels (I_h) play a role in initiation and modulation of cardiac and neuronal pacemaker depolarization. In vertebrates, the HCN channel family comprises four members (HCN1–4) (195, 293, 356, 473, 474), all of which are expressed in the brain and three of which (HCN1, HCN2, and HCN4) are also expressed in the heart. HCN channels are similar to CNG channels in many respects. HCN channels have six transmembrane regions and a cyclic nucleotide-binding domain in the intracellular COOH-terminal region. HCN channels function as homotetramer or heterotetramer and are permeable to both Na⁺ and K⁺. Opening of HCN channels occurs in response to membrane hyperpolarization rather than depolarization. HCN channels are regulated by various neurotransmitters including norepinephrine and ACh (417). cAMP enhances the intrinsic rate of firing by a positive shift in the voltage dependence of I_h activation through direct binding (48). cGMP, which has 10–100 times lower affinity than cAMP for the cNMP-binding domain of HCN channels, modulates HCN channel activity only at very high concentrations (356). cAMP also modulates HCN channel activity via a PKA-dependent mechanism (97).

D. cAMP-GEF/Epac

A family of novel cAMP-binding proteins, cAMP-GEF/Epac, has been discovered recently (134, 135, 297). cAMP-GEF/Epac consists of four functional domains: cAMP-binding domains, a DEP (Dishevelled, Egl-10, and Pleckstrin) domain responsible for its localization to the plasma membrane, a REM (Ras exchange motif) domain required for stabilizing GEF (guanine nucleotide exchange factor) activity, and the GEF domain, which exerts GEF activity toward the Ras-like small GTP-binding proteins Rap1 and Rap2 (134, 135, 297). The GTP-bound forms of Rap interact specifically with their effector proteins and activate downstream targets in various cells (54, 89, 517, 529, 566, 603). There are two isoforms of cAMP-GEF/Epac, cAMP-GEFI/Epac1 and cAMP-GEFII/Epac2, which are coded by different genes (557; http://www.ncbi.nlm.nih.gov/LocusLink/LocRpt.cgi?l=10411, http://www.ncbi.nlm.nih.gov/LocusLink/LocRpt.cgi?l=11069). The molecular mass values of cAMP-GEFI/Epac1 and cAMP-GEFII/Epac2 are 96.9 and 111.0 kDa, respectively. The dissociation constants of cAMP-GEFI/Epac1 and domain B of cAMP-GEFII/Epac2 for cAMP are 4.0 and 1.2 µM, respectively (134), while PKA binds cAMP in the range of 5.0–24.6 nM (75, 319, 457). Accordingly, cAMP-GEF/Epac may be a cAMP sensor in a range at which PKA activity is fully saturated. Structurally related proteins, PDZ-GEF1/nRapGEF/RA-GEF1/CNrasGEF and PDZ-GEF2/RA-GEF2, also have been identified that exhibit GEF activity toward Rap1 and Rap2 (133, 191, 345, 410, 429). Although these proteins contain a PDZ domain, a GEF domain, and a domain that resembles a cAMP-binding domain, they do not bind cAMP, and the GEF activity toward Rap is not regulated by cAMP (133, 318, 410).

Studies using X-ray crystallography and biochemical analyses of cAMP-GEF/Epac have suggested the molecular mechanism by which cAMP controls cAMP-GEF/Epac function (448–450). In the absence of cAMP, a regulatory region containing the cAMP-binding domain interacts directly with a catalytic region containing REM and GEF domains. The interaction of cAMP and cAMP-GEF/Epac is proposed to induce a conformational change of cAMP-GEF/Epac through the VLVLE sequence, thereby inducing GEF activity (53). Thus the cAMP-unbound form of cAMP-GEF/Epac inhibits GEF activity.

cAMP-GEFI/Epac1 mRNA is ubiquitously expressed, at high levels in adult tissues containing thyroid, kidney, ovary, skeletal muscle, and heart, and at low levels in the brain (135, 297, 415). cAMP-GEFII/Epac2 mRNA is predominant in the brain and neuroendocrine and endocrine tissues. A short form of cAMP-GEFII/Epac2 protein lacking the first cAMP-binding domain and DEP domain is expressed specifically in the liver (626). cAMP-GEFI/Epac1 mRNA is expressed widely but at low levels in the adult brain, while it is expressed at high levels in septum and thalamus of the neonatal brain (297). On the other hand, cAMP-GEFII/Epac2 mRNA is expressed at high levels in the cerebral cortex, hippocampus, cerebellum, olfactory bulb, thalamus, habenula, and pituitary (297, 415). The significance of signal transduction mediated by cAMP-GEF/Epac has just begun to emerge. cAMP-GEF has been suggested to play roles in many PKA-independent processes in eukaryotes, including cell proliferation, cell adhesion, apoptosis, and secretion (111, 121, 160, 182, 324, 330, 340, 363, 444, 482, 547). The role of cAMP-GEF/Epac in exocytosis is discussed in section VII.

E. CRP/CAP

CRP/CAP has been found in a variety of prokaryotes including both eubacteria and archaebacteria and regulates transcription of >150 genes (82, 127, 453). CRP/CAP is a 45-kDa dimer composed of two identical subunits (313). The large NH₂-terminal domain (amino acid residues 1–139) is responsible for both dimerization of CRP/CAP and binding to cAMP. cAMP binding induces an allosteric conformational change of CRP/CAP, resulting in binding of the small COOH-terminal DNA-binding domain to a specific DNA sequence and transcriptional activation of RNA polymerase. Both the cAMP- and DNA-binding domains show similarities to the respective domains found in a variety of other proteins from prokaryotes to eukaryotes. The cAMP-binding domain of CRP/CAP has sequence and structural similarities to the regulatory subunit of PKA (519, 576, 577), the cNMP-binding domain of CNG (396), and the cAMP-binding domain of cAMP-GEF/Epac (134, 135, 297, 448). Thus the structure of CRP/CAP has been a useful model for studies of both cyclic nucleotide sensitivity and the mechanism of activation by cyclic nucleotides (206, 613).

	V. PHARMACOLOGICAL AGENTS TARGETING THE cAMP PATHWAY

Top Previous Next References

 
To clarify the physiological roles of the cAMP signal, cAMP-binding protein-specific pharmacological agents are useful. Based on analysis of the cAMP-binding properties of PKA regulatory subunits and CRP/CAP, various cAMP analogs have been developed (489). These cAMP analogs are generally cAMP-specific, PDE-resistant, and membrane-permeable compounds and bind to PKA and cAMP-GEF/Epac selectively or nonselectively and modulate their activities positively or negatively. Modulators of the catalytic subunits of PKA have also been developed (109, 132, 289, 540). Agents that modulate cAMP production and degradation are also useful for analysis of cAMP signaling. Readers are referred to recent reviews of these agents (136, 148, 157, 407, 535).

A. cAMP Analogs

On the basis of site-directed mutagenesis and structural analysis, cAMP has been shown to interact with amino acid residues in a highly conserved phosphate-binding cassette motif [GELAL(X)_3–5PR(A/T)A(T/S)] in the loop linking 6 and 7 of the PKA regulatory subunit and cAMP-GEFII/Epac2 (448, 519) (Fig. 4). The phosphate and ribose rings of cAMP interact with the binding cassette through a network of hydrogen bonds and electrostatic interactions. The interaction of the adenine ring of cAMP with the binding cassette occurs in and near -helix C through hydrophobic and stacking binding. Glu²⁰² in domain A of the PKA regulatory subunit and Glu³²⁶ in domain B interact with 2'-OH of ribose. Glu in the phosphate-binding cassette is highly conserved among PKA, CNG, HCN, and CAP (294, 356, 519). Lys⁴²³ in domain B of cAMP-GEFII/Epac2 is a cAMP-GEFII/Epac2-specific residue and is thought not to participate in the interaction with 2'-OH of ribose (448). Arg and the first Ala in the PRAA(S/T) motif of the PKA regulatory subunit and the cAMP-GEFII/Epac2 domain B interact with cAMP (Fig. 5) (448, 519). Most cAMP analogs are modified at 6- and 8-positions in adenosine, 2'-position in the ribose ring, and the cyclophosphate ring (Fig. 5).

View larger version (14K): [in this window] [in a new window]   FIG. 5. Chemical structure of cAMP and its analogs. A: structure of cAMP. Positions of adenine ring and ribose ring are numbered. Solid line indicates covalent bond. Dotted line indicates hydrogen bond. B: various cAMP analogs. Modifications at side chains are indicated (see text for details).