| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
The Biotechnology Centre of Oslo, University of Oslo, Norway
ABSTRACT I. INTRODUCTION II. LOCALIZED POOLS OF cAMP A. Adenylyl Cyclases B. Phosphodiesterases C. cAMP Gradients III. cAMP EFFECTORS OTHER THAN PROTEIN KINASE A A. CNG Ion Channels B. cAMP-Regulated GEFs IV. PROTEIN KINASE A V. A KINASE ANCHORING PROTEINS VI. A MULTITUDE OF A KINASE ANCHORING PEPTIDES A. AKAPs Associated With Ion Channels 1. AKAP79 and neuronal transmission 2. AKAP15/18 3. Other ion channels regulated by PKA through AKAP interactions B. AKAPs Associated With the Cytoskeleton 1. Actin-associated AKAPs 2. Microtubule-associated AKAPs 3. Centrosome-associated AKAPs C. Mitochondria-Associated AKAPs D. AKAPs Involved in Regulation of Nuclear Dynamics and Chromatin Condensation VII. SIGNAL COMPLEXES ORGANIZED BY A KINASE ANCHORING PROTEINS VIII. cAMP SIGNALING TO THE NUCLEUS AND GENE REGULATION IX. REGULATION OF CELLULAR PROCESSES AND ORGAN FUNCTION BY cAMP AND PROTEIN KINASE A A. Regulation of Cardiovascular Function B. Regulation of Steroid Biosynthesis C. Regulation of Reproductive Function 1. Sperm/flagellar AKAPs D. Regulation of Metabolism in Adipocytes E. Regulation of Exocytotic Processes F. Regulation of Immune Function X. CONCLUDING REMARKS
 |
ABSTRACT |
|---|
 Top NextReferences |
|---|
|
I. INTRODUCTION |
|---|
|
TopPreviousNextReferences |
|---|
In this review, we discuss the different features of the cAMP-PKA pathway that provide specificity at the intracellular level and thereby convey tissue- and organ-specific effects. The question is how one single second messenger can be involved in regulation of such diverse cellular processes as regulation of the cell cycle, proliferation and differentiation and regulation of microtubule dynamics, chromatin condensation and decondensation, nuclear envelope dissambly and reassembly, as well as regulation of intracellular transport mechanisms and ion fluxes. The cAMP signaling pathway is further involved in controlling exocytotic events in polarized epithelial cells and is the primary intracellular pathway conveying 
-adrenergic signaling in the cardiovascular system and in adipose tissue. Also, cAMP pathways are involved in the regulation of steroidogenesis and reproductive function as well as in modulation of immune responses and a number of other effects elicited by hormones, neurotransmitters, and various paracrine ligands.
cAMP generation and degradation is regulated by the adenylyl cyclase and phosphodiesterase families of enzymes, respectively (305, 320). These enzymes are differentially expressed and regulated. cAMP-dependent protein kinase (PKA) is a heterotetramer composed of two regulatory and two catalytic subunits. Both the regulatory (RI
, RI, RII, RII) and the catalytic (C, C, C
) subunits possess distinct physical and biological properties, are differentially expressed, and are able to form different isoforms of PKA holoenzymes (reviewed in Refs. 300, 328). A kinase anchoring proteins (AKAPs) further contribute to the specificity as well as the versatility of the cAMP-PKA pathway by assembling multiprotein signal complexes allowing signal termination by phosphatases and cross-talk between different signaling pathways in close proximity to the substrates (Fig. 1) (84, 226). Integrating phosphodiesterases into these anchoring complexes further adds a temporal aspect to the spatial regulation of cAMP signals (303).
|
|
II. LOCALIZED POOLS OF cAMP |
|---|
|
TopPreviousNextReferences |
|---|
In mammals, nine membrane-bound isoforms of adenylyl cyclase (AC1-AC9) and one soluble sperm-specific form have been identified, all of which have distinct regulatory properties (reviewed in Ref. 134). All the membrane-bound isoforms exhibit a basal activity that is enhanced upon binding of the stimulatory G protein -subunit (Gs) and reduced upon binding of the inhibitory G protein -subunit (Gi). In addition, regulatory mechanisms including various small molecules provide means to differentially regulate the members of the family.
The membrane-bound members of the AC enzyme family comprise glycoproteins of 
120 kDa with considerable sequence homology. The suggested structure based on the amino acid sequence includes a small, cytoplasmic domain (N), two hydrophobic transmembrane domains, and two large cytoplasmic domains (C). The cytoplasmic domains are the most homologous sequences and constitute the catalytic moiety of the enzyme. C1a is the primary binding site for Gi, whereas C2a is the primary binding site for Gs and potentially the G protein -subunits. C2a also contains phosphorylation sites for protein kinase C (PKC) and calmodulin (CaM) kinase II. The various isoforms of the membrane-bound ACs can be divided into groups based on structure and regulatory properties (for a recent review, see Ref. 251).
Whereas all the membrane-bound ACs are expressed in the brain, the expression has by in situ hybridization been shown to be specific for the various structures of the central nervous system. Some of the isoforms have also been linked to specific functions. AC1 and AC2 are both highly expressed in regions associated with learning and memory as cerebral cortex, hippocampus, and cerebellum. Specifically, there is an enrichment of calcium-sensitive ACs in regions exposed to high intracellular free calcium induced by N-methyl-D-aspartate and voltage-gated Ca2+ channels, and AC1-mutated mice have affected long-term potentiation and spatial learning capabilities. Other tissues express AC isoforms at different stages of embryonic development, or in response to various stimuli such as nervous stimulation. Several tissues and cell types also display a sequential expression of AC isoforms during differentiation. Relatively little is known about the localization of various AC isoforms within subdomains of the plasma membrane. However, several AC isoforms (AC3–5) have been reported in lipid rafts and caveola and implicated in local cAMP microdomains at the membrane (289). This pertains also to G proteins and, for example, 2-adrenoceptors in the heart (314). In olfactory neurons, AC3 has been shown to be exquisitely localized to cilia, providing a clear "point source" of cAMP and presumably an associated gradient within these cells (163).
Cyclic nucleotide phosphodiesterases (PDEs) are enzymes responsible for the hydrolysis of cyclic nucleotides and play an important and highly regulated role in controlling the resting state levels of cAMP or cGMP intracellularly. Furthermore, they also contribute to establishing local gradients of cyclic nucleotides by being localized to subcellular compartments and by being recruited into multiprotein signaling complexes. This contributes to the temporal and spatial specificity of cyclic nucleotide signaling by regulating the availability of cAMP/cGMP to their effectors. The importance of the PDEs as regulators of signaling is evident from studies of PDE-deficient mice (157), and PDEs are also important drug targets in several diseases such as asthma and chronic obstructive pulmonary disease, cardiovascular diseases such as heart failure and atherosclerotic peripheral arterial disease, neurological disorders, and erectile dysfunction (69, 102, 130, 214, 310).
PDEs comprise a large superfamily of enzymes, and 11 families have been characterized on the basis of their amino acid sequences, substrate specificities, allosteric regulatory characteristics, and pharmacological properties (222, 305). In total, the superfamily of PDEs encompasses 25 genes in mammals giving rise to an estimate of more than 50 different PDE proteins (342). They share a modular architecture, with a conserved catalytic domain proximal to the COOH terminus, regulatory domains most often located at the NH2 terminus and targeting domains which we are only beginning to discover (67, 110, 147). The substrate specificites of the PDEs families include cAMP-specific, cGMP-specific, and dual-specific PDEs. We will here briefly discuss the role of PDEs in the context of generating localized pools of cAMP.
The distribution of PDEs to different subcellular localizations was proposed early on by the observation that PDE activity was found in both the soluble and particulate fractions of the cell (316). Recent evidence further supports this notion and contributes to an emerging concept of a highly organized signal pathway where specific routes of cAMP signals are formed through the localized synthesis by cell- and tissue-specific adenylyl cyclases, and where the signal is delivered to targeted effectors and terminated in a spatial and temporal manner by specific PDEs establishing local pools of cAMP close to the effector molecules.
Putative or established targeting domains have now been identified for most of the PDE families (222). PDE3s are targeted to the the endoplasmatic reticulum by a transmembrane domain consisting of six transmembrane helices (78), and PDE4D5 interacts with RACK-1, a scaffold protein which binds certain PKC isoforms after activation by diacyglycerol (363). PDE4D3 is targeted to the Golgi/centrosomal region through anchoring by myomegalin (156, 343). Some PDE4D and PDE4A variants bind Src homology 3 (SH3) domains of, e.g., Src kinases (23, 24), and via their catalytic domain PDE4 isoforms bind to and are phosphorylated by Erk (211). PDE4A1 contains a novel lipid binding domain, TAPAS, with specificity for phosphatidic acid that serves to target this PDE to specific cellular membranes (16). Most recently, the PDE4 family is reported to be recruited to activated -adrenoreceptors through interaction with -arrestin (17, 252). Furthermore, direct interaction between a PDE and two different AKAPs has recently been reported. In rat Sertoli cells, AKAP450 targets PDE4D3 to the centrosomal region together with PKA type II in a ternary complex (329). In cardiomyocytes, muscle AKAP (mAKAP) binds and targets both PDE4D3 and PKA type II to the perinuclear region (87). These are the first examples of colocalized PKA/PDE complexes providing spatial control of PKA signaling by AKAP anchoring and temporal control and termination of the cAMP signaling event by complexing PDE in the immediate vicinity. Furthermore, long PDE4 isoforms such as PDE4D3 are activated by PKA phosphorylation, effectively establishing a negative-feedback loop that terminates the cAMP signal locally (Fig. 2) (156, 212). In addition to the spatial control of PDEs by subcellular compartmentalization, PDE activity is also allosterically regulated, regulated by protein-protein interactions and by posttranslational modifications further contributing to the specificity in this signaling pathway (67, 222).
|
|
III. cAMP EFFECTORS OTHER THAN PROTEIN KINASE A |
|---|
|
TopPreviousNextReferences |
|---|
CNG ion channels have been found in a variety of cell types and tissues including kidney, testis, heart, and the central nervous system (reviewed in Refs. 42, 350, 364). These channels open in response to direct binding of intracellular cyclic nucleotides and contribute to cellular control of the membrane potential and intracellular Ca2+ levels. The first member of this family to be identified was the retinal rod photoreceptor, which is directly activated by cGMP (107, 365), and a similar channel was then subsequently identified in olfactory transduction able to bind both cAMP and cGMP (240). One of the most recently reported is the CatSper involved in cAMP-mediated sperm motility (259, 268).
The CNG ion channels are multi-subunit pore-forming channels. The different subunits are highly homologous and bear structural similarity to voltage-gated K+ channels (368). The modulation of channel activity is through allosteric binding, and maximal activation typically requires four ligands bound (207, 277). The cyclic nucleotide binding domain is connected to the last transmembrane segment of the channel by 90 amino acids called the C-linker, which also has been shown to be important for the regulation of the channel activity (108, 123, 160).
Ras-related proteins are monomeric GTPases. They cycle between an inactive GDP-bound state and an active GTP-bound state, which is achieved by the exchange of the tightly bound GDP for GTP. They then revert to the inactive state when the intrinsic GTPase activity again converts GTP to GDP (37, 38). Both these reactions are slow and are facilitated by GEFs and GTPase-activating proteins (GAPs), respectively. Ras proteins regulate downstream signaling proteins by recruitment to the plasma membrane and subsequent activation.
Rap-1, which is a small Ras-like GTPase (34, 254), was first identified as a protein that could suppress the oncogenic transformation of cells by Ras (176) and act as a suppressor of Ras (39, 68). A number of extracellular stimuli signal to Rap-1 and the more recently identified Rap-2 by induction of second messengers like cAMP, calcium, and diacylglycerol (DAG) that regulate Rap-specific GEFs. Two of these proteins called Epac (exchange protein activated by cAMP) 1 and 2 (or cAMP-GEFs) have raised considerable interest as their activities are directly regulated by cAMP, and thereby provide an additional effector system for cAMP signaling (80, 172). Epac1 has one cAMP binding site, whereas Epac2 contains two binding moities (79). The cAMP binding domains in both Epac1 and Epac2 function as inhibitors of the COOH-terminal GEF domains in the absence of cAMP, wheras cAMP binding induces a conformational change exposing and activating the GEF domain (265). The recent use of a cell-permeable cAMP agonist that is selective for Epac has provided compelling evidence for a cAMP-Epac-Rap pathway (98).
|
IV. PROTEIN KINASE A |
|---|
|
TopPreviousNextReferences |
|---|
Two classes of PKA isozymes, designated types I and II, were originally identified based on their order of elution by ion-exchange chromatography and shown to differ in the content of the R subunit, called RI or RII, respectively. Later further heterogeneity was unravelled by molecular cloning identifying RI, RI, RII, and RII as well as four C subunits C, C, C, and PRKX (reviewed in Ref. 300). PRKX (the human X chromosome-encoded protein kinase X) was recently described as a cAMP-dependent kinase that forms a catalytically inactive holoenzyme with RI, but does not bind to the RII subunit under physiological conditions (371). The R subunits exhibit different cAMP binding affinities giving rise to PKA holoenzymes with different thresholds for activation. Whereas PKA type II holoenzymes (RII2C2, RII2C2) typically activate with an activation constant (Kact) of 200–400 nM of cAMP, type I holoenzymes (RI2C2, RI2C2) have higher affinity for cAMP and activate with Kact of 50–100 nM cAMP (89). In addition, the R subunits are differentially expressed in different cells and tissues and are able to form both homo- and heterodimers generating a large number of combinations, which further contribute to diversity and presumably specificity in the cAMP signal pathway.
Subcellular localization of PKA is mainly due to anchoring of the R subunits by AKAPs, which originally were seen as contaminants of purified PKA (208, 282, 332) and later understood to enhance the efficiency and specificity of the signaling events. While PKA type I is classically known to be biochemically soluble and was thus assumed to be mainly cytoplasmic, PKA type II is typically particulate and confined to subcellular structures and compartments anchored by cell- and tissue-specific AKAPs, a field largely pioneered by the Scott laboratory (reviewed in Refs. 63, 84, 86, 226). However, a few dual-specific AKAPs (D-AKAPs) anchoring both PKA type I and type II as well as some AKAPs that selectively bind PKA type I have more recently been identified (see Table 1).
|
As evident from solution of the NMR structure, the RII subunits dimerize at the NH2 terminus in an antiparallel fashion forming an X-type, four-helix bundle that is necessary for both AKAP binding (NH2-terminal helix of both protomers) and dimerization (COOH-terminal helices of the bundle) through separate but overlapping regions involved in the two events (243–245) (Fig. 3). Dimerization is a prerequisite for AKAP binding, but deletion of residues 1–5 abolishes AKAP binding without disrupting dimer formation, and branched side chains at positions 3 and 5 are critical for the interaction with the AKAPs in a hydrophobic groove that is formed on top of the NH2-terminal helices (139, 140). The RI dimerization domain contains a similar helix-turn-helix motif recently solved by NMR which is shifted a little further from the NH2 terminus and encompasses amino acids 12 to 61 (21, 22, 195). The extreme NH2 terminus in RI is helical and believed to fold back onto the four-helix bundle and may thus contribute to differences in AKAP binding specificity between RII and RI.
|
|
V. A KINASE ANCHORING PROTEINS |
|---|
|
TopPreviousNextReferences |
|---|
|
Typically, AKAPs anchor PKA type II holoenzymes (RII2C2) with high affinity (low nanomolar range). In contrast, their interaction with PKA type I (RI2C2) appears to occur at considerable lower affinity (49, 139–141). The availability of RII versus RI may determine which subunit is tethered by AKAPs in vivo (149, 150, 266, 201). However, the AKAPCE identified in Caenorhabditis elegans is demonstrated to specifically interact with the worm's RI-like subunit and does not bind mammalian RII (11, 12). Furthermore, the mammalian AKAP82 is shown to have an RI-specific PKA binding domain (228), and both PAP7 and hAKAP220 appear to anchor PKA type I inside cells, although mapping of the RI binding domains has not been conducted (201, 266). Thus RI also appears to be capable of forming physiologically relevant anchoring interactions.
The conserved PKA tethering domain in AKAPs forms an amphipathic helix of 14–18 residues that interacts with hydrophobic determinants located in the extreme NH2 terminus of the regulatory subunit dimer (49, 50, 243, 245) (Fig. 3). The amphipathic helix of the AKAPs, with hydrophobic residues aligned along one face of the helix and charged residues along the other, binds to RII with high affinity (140, 141). Dual-specific AKAPs (149, 150) appear to bind to the RI dimerization and docking (DD) domain in a similar fashion (21, 22). Disruption of the amphipathic helix abolishes the binding to R both in vitro and in vivo, and the residues determining binding of RI and RII have been defined (8, 44, 50).
A peptide usually referred to as the Ht31 anchoring disruption peptide derived from the PKA tethering domain of the human thyroid AKAP Ht31, now called AKAP-Lbc, mimics the amphipathic helix that binds the extreme NH2 terminus of the regulatory subunit of PKA and serve as a competitive anchoring inhibitor of PKA-AKAP interactions (49, 50). The Ht31 peptide has been used extensively as a tool to analyze the effects of disrupting PKA anchoring. Interestingly, recent analysis of the RII binding domain of AKAPs by bioinformatics and peptide array approaches unravelled high-affinity peptides with specificity for binding RII (AKAP-is peptide, Ref. 8). Similarly, isoform-specific peptide disruptors of PKA type I association with AKAPs have recently been developed (44). Use of such anchoring disruptors will greatly facilitate analysis of cellular effects of anchored PKA type I and II.
|
VI. A MULTITUDE OF A KINASE ANCHORING PEPTIDES |
|---|
|
TopPreviousNextReferences |
|---|
1. AKAP79 and neuronal transmission
Protein phosphorylation and dephosphorylation by protein kinases and phosphatases play a key role in regulation of synaptic plasticity in the hippocampus (358). PKA-mediated phosphorylation potentiates the currents induced by activation of the excitatory AMPA receptor by phosphorylation of Ser-845 of the glutamate receptor 1 (GluR1) subunit (20, 128, 274, 347). The first demonstration that AKAP-mediated targeting of PKA is necessary for mediation of a biological effect of cAMP was shown by peptide-mediated disruption of a PKA-AKAP complex directing PKA toward the AMPA receptor leading to a significant reduction in the glutamate receptor activity measured by whole cell voltage clamping (275). The AKAP responsible for targeting PKA to the receptor was later identified as AKAP79 (AKAP150 and AKAP75 are murine and bovine orthologs, respectively) which is able to associate with both AMPA and NMDA receptors (41, 51, 60, 62, 122, 282). AKAP79 is targeted to the plasma membrane by three NH2-terminal basic regions that bind phosphatidylinositol 4,5-bisphosphate. Membrane-associated AKAP79 is then recruited to the NMDA and AMPA receptors by binding to the SH3 and guanylate kinase-like (GK) domains of the membrane-associated guanylate kinase (MAGUK) proteins, postsynaptic density (PSD)-95 and synapse-associated protein (SAP)-97, respectively (Fig. 5) (62). These processes are dependent on the actin cytoskeleton and recruit the AKAP79 to the NMDA and AMPA receptors localized in the postsynaptic densities of hippocampal synapses (122).
|
AKAP79 is also associated with 2-adrenergic receptors (2-AR) and recruits PKA, PKC, and protein phosphatase (PP) 2B (Fig. 5) (113). The receptor undergoes cAMP-dependent desensitization after agonist stimulation by direct PKA phosphorylation and indirectly by PKA-mediated phosphorylation and enhancement of G protein-coupled receptor kinase 2 (GRK2) (25, 66, 113). PKA phosphorylation of the 2-AR also induces a switch in the G protein coupling from Gs to Gi (74). This promotes a mitogenic signaling cascade mediated by Gi, -arrestin, and the Src-tyrosine kinase leading to mitogen-activated protein (MAP) kinase activation (210). Both receptor desensitization and MAP kinase activation can be disrupted by inhibition of PKA anchoring with Ht31 (113). Furthermore, -arrestin recruits PDE4 which control 2-AR phosphorylation by PKA and hence the Gs to Gi switch (17, 252).
In skeletal muscle transverse tubules, L-type calcium channels initiate muscle contraction by directly interacting with ryanodine receptors to cause the release of calcium from the sarcoplasmic reticulum (SR) (53). The calcium channels function both as voltage sensors to initiate excitation-contraction coupling and as a slowly activating calcium entry pathway that regulates the force of the contraction (4, 53). Repetitive high-frequency depolarizing stimuli that mimic action potentials or single long depolarizing pulses greatly enhance the activity of the L-type calcium channels (291). This enhancement is voltage dependent and requires phosphorylation by PKA (291) and can be induced by -adrenergic stimuli (Fig. 6) (14, 286).
|
The importance of PKA-mediated phosphorylation in the regulation of the calcium channel function in skeletal muscle is evident from experimental inhibition of the anchoring of PKA in the channel vicinity by Ht31 which leads to a 20-fold reduction in voltage-dependent potentiation of the calcium channel activity (126, 158, 159). The AKAP involved in this process has been identified as a 15- or 18-kDa protein that copurifies, coimmunoprecipitates, and colocalizes with the skeletal muscle calcium channel complex (114, 126, 127). AKAP15/18 (the -isoform) is an 81-residue protein containing an amphipathic helix that binds PKA and NH2-terminal myristoyl and palmitoyl lipid anchors that target the PKA-AKAP complex to the plasma membrane (114, 126). The direct interaction between AKAP15/18 and the channel involves interaction with the COOH-terminal domain of the 1-subunit of the L-type calcium channel via a leucine zipperlike mechanism providing means of localizing PKA in close proximity to a major phosphorylation site located in the 1-subunit at serine-1854 (152). This ensures specific and rapid phosphorylation of the channel.
Further studies demonstrated that the first identified AKAP15/18 is one of several splice variants, now named AKAP18, and is localized to the basolateral membrane compartment in polarized cells. Furthermore, other splice variants from this gene have apical targeting (AKAP18) and localization to cytoplasm (AKAP18) or to secretory granules (AKAP18
) (182, 338).
3. Other ion channels regulated by PKA through AKAP interactions
Several other ion channels are also regulated by PKA through AKAP interactions. The cystic fibrosis transmembrane conductance regulator (CFTR) is an epithelial Cl– channel whose activity is enhanced by PKA-dependent phosphorylation (117, 260). More than 800 mutations in CFTR have been observed in patients with cystic fibrosis, and 5% of these are in the regulatory domain of the channel containing 9 consensus sites for PKA phosphorylation. The interaction between CFTR and PKA involves targeting of PKA to CFTR by binding to the 78-kDa AKAP ezrin (92). Ezrin is the most studied member of the ezrinmoesin-radexin (ERM) family of proteins and plays structural and regulatory roles in the assembly and stabilization of specialized plasma membrane domains. Ezrin and related molecules are concentrated in surface projections such as microvilli and membrane ruffles where they link the microfilaments to the membrane. The interaction between ezrin and CFTR involves the Na+/H+ exchanger (NHE) type 3 kinase A regulatory protein (E3KARP) which binds to CFTR via a PSD-95/Disc-large/zonula occludens-1 (PDZ) binding motif (318, 319). Thus E3KARP acts as a scaffolding protein that links CFTR to ezrin. This is analogous to the targeting of ezrin to the Na+/H+ exchanger in the renal brush border by the Na+/H+ exchanger regulatory factor (NHERF or ezrin-binding phosphoprotein 50, EBP50) facilitating PKA-mediated phosphorylation and inhibition of the channel (264, 351, 352). Ezrin is also enriched in gastic parietal cells (91, 92) and plays an important role as a membrane-cytoskeletal linker in these cells (6, 135). When stimulated with gastrin, ezrin serves to recruit PKA to the secretory canaliculi (92).
PKA-mediated phosphorylation also enhances the activity of NMDA receptors (55, 262). The AKAP yotiao (splice variant from the AKAP9/AKAP450 gene) targets PKA to the receptor by binding to the NR1 subunit of the receptor (103, 204, 357). NMDA receptors are heteromultimers composed of an NR1 subunit and a variety of NR2 family members (190, 221, 232), and yotiao specifically interacts with the splice variant of NR1 that contains the C1 exon (204, 357). The functional relevance of yotiao-mediated anchoring of PKA has been demonstrated by whole cell current recording of transfected cells and by disruption of the anchoring by Ht31 (357). Yotiao also binds the PP1 which under resting conditions with low PKA activity dephosphorylates and deactivates the channel (32, 304, 346, 357). Thus yotiao coordinates the opposing kinase and phosphatase required for efficient regulation of the NMDA receptor function.
B. AKAPs Associated With the Cytoskeleton
Phosphorylation of proteins associated with the cytoskeleton plays an important role in the dynamics and functional organization of the cytoskeleton. AKAPs are emerging as facilitators of cytoskeletal events as they target PKA to sites where it can phosphorylate substrates including actin, microtubules, the centrosome, and the sperm flagella.
Actin polymerization is an essential process in all eukaryotic cells, generating the basis for establishment of cell shape, polarity of cell consitutents and membrane domains, motility, and cell division. The Rho family of small GTPases are key proteins in this process that link cell surface receptors to the organization of the actin cytoskeleton by regulating the activity of downstream effector molecules. The best studied members of the Rho family of small GTPases include Rho, Rac, and Cdc42 (31, 131). The cytoskeletal changes induced by these three molecules are associated with distinct integrin-based adhesion complexes and while Rho activation leads to assembly of stress fibers, activation of Rac and Cdc42 leads to generation of lamellipodia and filopodia, respectively (185, 246, 270, 271). The WASP family of proteins, consisting of WASP, N-WASP, and the Scar-1 orthologs WAVE1, WAVE2, and WAVE3, plays an important role in these molecular interactions by providing a molecular bridge that links Rho family members to the actin nucleation machinery, the Arp2/3 complex (143, 215, 324). Rac-1 interacts with WAVE1 to activate actin nucleation by releasing WAVE1 from a heterotetrameric complex (94, 227, 317). In addition, WAVE1 binds WRP, a Racselective GAP that specifically inhibits Rac function in vivo and functions as a signal termination factor for Rac (306). The WASP family members attach to the actin cytoskeleton through a verprolin homology (VPH) domain and a COOH-terminal acidic module that binds to the Arp2/3 complex.
WAVE1 was recently identified as an AKAP that is also able to bind the Abelson tyrosine kinase (Abl) (356). It was identified in a screen for brain AKAPs interacting with isolated SH3 domains from different signal transduction molecules. The two other WAVE isoforms, WAVE2 and WAVE3, which bear considerable sequence homology to WAVE1, lack certain key hydrophobic residues and do not bind RII. The RII-binding region of WAVE1 overlaps with a VPH domain (residue 493–510) that act as a binding site for G-actin. Although G-actin and RII recognize different determinants within the 493–510 sequence, RII and actin binding are mutually exclusive. Thus PKA anchoring by WAVE1 may be dynamically regulated by the actin concentration at sites of actin polymerization. Immunocytochemical analyses in Swiss 3T3 fibroblasts suggest that the WAVE1-kinase scaffold is assembled dynamically and translocates both PKA and Abl from focal adhesions to sites of actin reorganization such as lamellipodia and actin ring structures in response to platelet-derived growth factor treatment (356). The substrates for PKA and Abl are, however, not yet identified. Interestingly, targeted disruption of the WAVE1 gene generated mice with a complex psychomotoric and cognititive phenotype (307), and further genetic manipulation will determine the extent to which PKA-signaling events are implicated in these varied cognitive processes.
Two other actin-binding proteins have been identified as AKAPs: gravin and ezrin. Wheras ezrin is discussed above, gravin is a multivalent 250-kDa scaffold protein that interacts with PKA, PKC, and actin (124, 241). It was identified as a cytoplasmic antigen recognized by sera from patients with myasthenia gravis. Gravin localizes to filpodia in endothelial and macrophage-like cells and shares significant sequence homology with SSeCKS (Srcsuppressed C kinase substrate) which also binds PKA, PKC, and actin and mediates actin remodeling (99, 112, 119, 206, 242). In addition to playing a role in regulation of actin polymerization, gravin organizes PKA, PKC, and PP2B with the 2-AR (294) in a complex that also includes the G protein-linked receptor kinase 2 (GRK2) and transiently -arrestin and clathrin (203). Prolonged stimulation of G protein-linked receptors (GPLRs) leads to desensitization of the receptor-mediated signal and agonist-induced receptor sequestration (138). PKC and PP2B are important for the reversal of this process and thereby resensitization of the receptor, as both suppression of PKC and PP2B amplifies the agonist-induced desensitization of the receptor (294–296). Gravin is required for this event to occur (294). PKA, on the other hand, potentiates agonist-induced desensitization of the 2-AR by causing its phosphorylation and switching from Gs to Gi coupling (138, 295).
AKAP-KL is a cytoskeletal-associated AKAP expressed in lung, kidney, and cerebellum (88). There are a total of six different isoforms of AKAP-KL showing tissue-specific expression. The intracellular localization of AKAP-KL is asymmetric with an apical distribution in polarized cells such as pulmonary alveolar epithelial cells and proximal renal tubular cells. It is not yet determined whether AKAP-KL directly interacts with the actin cytoskeleton, although AKAP-KL modulates actin structure in transfected HEK293 cells (88). AKAP-KL may be involved in establishing or maintaining cellular polarity, or fascilitating transepithelial signaling processes.
2. Microtubule-associated AKAPs
The microtubule-associated protein 2 (MAP2) family of proteins stabilize microtubuli. The MAP2 proteins are predominantly expressed in neurons where they regulate microtubule nucleation, organelle transport within axons, and dendrites as well as anchoring of proteins involved in signal transduction (281). The association of MAP2 with microtubuli occurs through its tubulin binding domain, which binds to an acidic region in the COOH terminus of tubulin and is regulated by its phosphorylation status (70, 146, 292). In addition, MAP2 can also bind to and modify microfilament stability.
Numerous protein kinases and phosphatases are involved in determining the phosphorylation status of MAP2, one of which is PKA. Interestingly, MAP2 was the first AKAP to be identified and tethers one-third of the cytosolic PKA to the microtubules in neurons (332). Several PKA phosphorylation sites have been identified in MAP2 which are also conserved in the closely related MAP tau, including the KXGS motifs located in the tubulin-binding domain. Phosphorylation of these motifs leads to detachment from tubulin (154, 288). The effects of PKA phosphorylation on MAP2 proteins include decreased binding of MAP2 to tubulin and actin, reduced microtubule polymerization, and reduced proteolytic degradation of MAP2 (281). Furthermore, mice with deletion of the MAP2 NH2 terminus which includes the PKA binding site have decreased efficiency of MAP2 phosphorylation and impaired development of contextual memory (175).
3. Centrosome-associated AKAPs
The centrosome represents the major microtubule-organizing center of animal cells consisting of a pair of centrioles surrounded by the pericentriolar matrix composed of a pericentrin and -tubulin lattice (33, 370). With its crucial role in nucleation and organization of microtubules, the centrosome is important in cellular procesesses such as generating a microtubular framework for motor-protein based transport and positioning of vesicles and organelles (33, 168). In mitotic cells, centrosomes are important for the assembly and function of the mitotic spindles and thereby regulate the fidelity of chromosome segregation (65, 225). In addition, an increasing number of molecules that regulate cellular processes such as cell cycle progression and centrosome duplication are found to be localized to centrosomes.
Three AKAPs have been identified in centrosomes, AKAP450 (359), pericentrin (83), and hAKAP220, which is expressed in male germ cells (266). AKAP450 is also named AKAP350 (287) or CG-NAP (centrosome and Golgi localized PKN-associated protein) (323) and derives from the same gene as yotiao. AKAP450 is localized to the centrosome throughout the cell cycle, to the Golgi apparatus during interphase (323, 359), and in the cleavage furrow during anaphase and telophase (287). Although the roles of the pool of PKA anchored to centrosomal AKAPs are not well defined, it is possible that anchored pools of PKA participate in regulation of microtubule nucleation by targeting substrates such as stathmins (106, 125) (see Fig. 2). Moreover, the AKAP450 signal complex has a role in cell cycle progression. Displacement of endogenous AKAP450 and the molecules anchored to it by overexpression of the COOH-terminal AKAP450 targeting domain (PACT domain, pericentrin-AKAP40 centrosomal targeting domain, Ref. 120) results in cell cycle arrest and impaired cytokinesis and centriole duplication (173). In addition, the association between AKAP450 and RII appears to be under direct regulation of the mitotic kinase CDK1 (46). At the onset of mitosis, CDK1 associated with the centrosome (18, 19) phosphorylates RII on T54 leading to dissociation from its centrosomal site of anchoring (46, 174). This suggests the PKA-AKAP association in some cases may be dynamic and that CDK1 phosphorylation serves as a molecular switch that regulates RII association with AKAP450, whereas AKAP95 has an opposite role and binds the CDK1-phosphorylated PKA as described below (Fig. 7) (46, 192).
|
AKAP450 interacts with several signal transduction enzymes in addition to PKA, inluding PKN, PP1, PP2A, and the immature nonphosphorylated form of PKC
(321, 323). PKN is a serine/threonine kinase that associates with and phosphorylates intermediate filament proteins (218, 236), and AKAP450 targeting of PKN may thereby be important for cytoskeletal reorganization events. PKN is activated by Rho (9, 349) and unsaturated fatty acids such as arachidonic acid (235), or by truncation of its NH2-terminal regulatory region (322). Interaction between AKAP450 and the nonphosphorylated form of PKC is required for the phosphorylation-dependent maturation of PKC. Recently, AKAP450 was reported to anchor protein kinase CK1, which is involved in control of cell cycle progression (297). In addition, AKAP450 also anchors PDE4D (329), which allows tight control of the phosphorylation state of proteins regulated by cAMP signaling. Spatial control is achieved by targeting of PKA by AKAP450, while temporal control and inactivation of the effect of cAMP on PKA is accomplished by complexing of PDE at the same site (see Fig. 2).
Pericentrin (full-length human protein named kendrin) is an integral component of the pericentriolar matrix (90) and forms a centrosomal macromolecular complex with -tubulin (82) and dynein (258) important in the dynamic organization of centrosomes and spindles (90). Binding to -tubulin is required for microtubule nucleation during mitosis and meiosis, and the association with dynein is necessary for the transport of pericentrin--tubulin complexes along microtubules to the centrosome. Pericentrin/kendrin shares a high degree of linear homology with AKAP450. This indicates a common origin, and pericentrin/kendrin also serves as an AKAP, although the pericentrin PKA anchoring domain composed of a 100-residue, hydrophobic binding region does not exhibit the structural characteristics of the RII-binding sites found in conventional AKAPs (83). AKAP450 and pericentrin also share the common PACT domain that targets both AKAPs to the centrosome (120). However, overexpression of the AKAP450 PACT domain displaces AKAP450, but not pericentrin, and vice versa, indicating specificity in the targeting of AKAP450 and pericentrin (173). The fact that both pericentrin and AKAP450 target PKA to centrosomes may indicate an important role in microtubule trafficking and centrosome nucleation where there is a need for redundance or, as the displacement studies may indicate (173), that more than one AKAP is required to accurately position PKA versus substrates inside the centrosome. This indicates a much more sophisticated level of kinase compartmentalization than was originally conceived.
C. Mitochondria-Associated AKAPs
Several mitochondrial AKAPs have been identified. S-AKAP84 (205), AKAP121 (57), D-AKAP1 (150), and AKAP149 (337) derive from the same gene by alternative splicing and are discussed below together with Rab32 (7). PBR-associated protein 7 (PAP7) is another mitochondrial AKAP that selectively binds RI in vivo (201), discussed in section IXB.
S-AKAP84 (205), AKAP121 (57), D-AKAP1 (150), and AKAP149 (337) share a 525-amino NH2-terminal core but differ in the COOH-terminal domain as well as in their extreme NH2 termini (Fig. 4). S-AKAP84, AKAP121, and the N0 isoform of D-AKAP1 share an identical targeting motif that anchors PKA type II to the outer mitochondiral membrane, whereas alternate splicing of the NH2 terminus (N1 isoform) of D-AKAP1 directs the protein to the endoplasmic reticulum (ER) (N1) (151) and the nuclear envelope membrane network (discussed below) (313). The N1 isoform of D-AKAP1 contains an additional 30 residues responsible for anchoring this isoform to the ER (150, 151). The R-binding domains of S-AKAP84, AKAP121, and D-AKAP1 are also identical and anchor both RI and RII, although the binding affinity of RI is lower than for RII (Kd of 185 vs. 2 nM, respectively) (141).
Rab32 is a member of the Ras superfamily of small-molecular-weight G proteins that is targeted to mitochondria and involved in regulation of mitochondrial fission (7). PKA type II binds to the conserved 5-helix of Rab32 which indicates dual functions of Rab32 as an AKAP and regulator of mitochondrial dynamics (7).
D. AKAPs Involved in Regulation of Nuclear Dynamics and Chromatin Condensation
Mitotic cell division requires that the DNA is properly condensed into chromosomes. This process involves topoisomerase II (3) and a family of proteins of highly conserved ATPases called SMCs (structural maintenance of chromosomes) (144, 145, 278). SMCs participate in multiprotein chromosome condensation complexes called condensins. Purification and characterization of condensins containing XCAP-C and XCAP-E, two Xenopus members of the SMC family, revealed two major forms of condensins, 8S and 13S (144). Condensins are targeted to the chromosomes during mitosis in a topoisomerase-independent manner, and the 13S subunit is required for chromatin condensation to take place. The 13S subunit contains both XCAP-C and XCAP-E and three other subunits including pEg7/XCAPD2, which is a required component of the complex for the condensation process (71).
AKAP95 is 95-kDa protein that harbors two zinc fingers (designated ZF1 and ZF2) in its COOH-terminal half, upstream of the PKA-binding domain (59, 96). In interphase, AKAP95 is localized exclusively in the nucleus and associates with the nuclear matrix but does not anchor RII (59, 96, 61). At mitosis, AKAP95 redistributes from the nuclear matrix to chromatin and recruits the condensin complex once nuclear envelope breakdown has taken place (61, 312). Subsequently, AKAP95 anchors RII onto, or in the vicinity of, the metaphase plate. Recruitment of RII from a centrosome-Golgi localization during interphase to chromatin-bound AKAP95 at mitosis requires phosphorylation of RII on threonine-54 by CDK1 (192) (Fig. 7). Conversely, release of RII from AKAP95 upon chromosome decondenseation in vitro or mitosis exit correlates with threonine-54 dephosphorylation (192).
Distinct domains of AKAP95 are involved in binding to chromatin and in the recruitment of RII and of the condensin complex (95). Chromatin binding of AKAP95 is required for condensation to take place, and the amount of Eg7/XCAPD2 recruited correlates with the extent of chromosome condensation in vitro (312). Furthermore, disruption of the ZF1-domain abrogates chromosome condensation, but not condensin recruitment. Thus AKAP95 is essential for chromosome condensation independently of condensin recruitment (95). Interestingly, mitotic chromosome condensation does not require anchoring of PKA to AKAP95 nor PKA activity. However, both PKA activity and binding to AKAP95 are required for the maintenance of condensed chromatin during mitosis, and blocking of PKA activity or disruption of anchoring leads to premature chromatin decondensation (61).
AKAP149 is not only targeted to mitochondria and ER but also associates with the nucleus as it is an integral protein of the ER/nuclear envelope membrane network (313). In addition to anchoring PKA, AKAP149 targets a fraction of chromatin-bound PP1 to the nuclear envelope upon nuclear reformation in vitro (313). The nuclear envelope is a dynamic structure that breaks down at mitosis and reforms in an ordered manner as a result of reversible phosphorylations of membrane, lamina, and chromatin proteins. The nuclear lamina consists of intermediate filaments called A/C- and B-type lamins. Lamins mediate the interactions between the inner nuclear membrane and chromatin, participate in DNA replication, and may provide a structural role for RNA splicing (97, 155, 309). Targeting of PP1 to the nuclear envelope correlates with the nuclear assembly of B-type lamins at the end of mitosis, and disruption of AKAP149 anchoring by a peptide containing the PP1-binding domain of AKAP149 leads to failure of B-type lamin assembly, caspase-dependent proteolysis, and apoptosis (311, 313). It is not yet determined whether AKAP149 anchors PKA and PP1 in distinct complexes or in one single complex. AKAP149 may position PKA and PP1 in close proximity where they can reversibly modulate the phosphorylation status of nuclear substrates such as NPP1 (29), DNA-binding cAMP response elements (269), B-type lamins (253), and inner nuclear membrane proteins harboring PKA phosphorylation sites.
mAKAP (originally cloned and characterized as AKAP100) is a 255-kDa scaffolding protein expressed in myocytes, skeletal muscle, and brain. mAKAP assembles a signal complex consisting of PKA and PDE4D3 at the nuclear envelope, the SR of cardiomyocytes, and intercalated discs in adult rat heart tissue (87, 167, 217, 220, 361). The assembly of the mAKAP signaling complex in the perinuclear region is induced by hypertrophic stimuli in rat neonatal ventriculocytes and is thought to be associated with cellular differentiation and development of a ventricular hypertrophic phenotype (167). The induction of mAKAP expression also leads to redistribution of RII to the NE (167), which is interesting as PKA phosphorylation induces cAMP-responsive genes involved in propagation in cardiac hypertrophy (369), and the concurrent anchoring of PDE4D3 serves to establish a negative-feedback loop (87) (see discussion above and Fig. 2).
|
VII. SIGNAL COMPLEXES ORGANIZED BY A KINASE ANCHORING PROTEINS |
|---|
|
TopPreviousNextReferences |
|---|
In addition to its role in anchoring RII, studies of AKAP79 have contributed to the evolution of the model of AKAPs as scaffolding proteins able to bind and anchor multiple signal transduction proteins and also regulate their enzymatic function. From being discovered as proteins able to bind and anchor PKA, the capacity of AKAP79 to associate with other signal enzymes has led to the reevaluation of the original AKAP model. By coordinating the location of PKC and the calcium/CaM-dependent phosphatase PP2B (calcineurin) in addition to PKA, AKAP79 positions two second messenger-regulated kinases and a phosphatase near to neuronal substrates at the postsynaptic densities (see Fig. 5) (60, 177).
|
VIII. cAMP SIGNALING TO THE NUCLEUS AND GENE REGULATION |
|---|
|
TopPreviousNextReferences |
|---|
