Phosphorylation, compartmentalization, and cardiac function

Summary

Protein phosphorylation is a fundamental element of cell signaling. First discovered as a biochemical switch in glycogenolysis, we now know that this posttranslational modification permeates all aspects of cellular behavior. In humans, over 540 protein kinases attach phosphate to acceptor amino acids, whereas around 160 phosphoprotein phosphatases remove phosphate to terminate signaling. Aberrant phosphorylation underlies disease, and kinase inhibitor drugs are increasingly used clinically as targeted therapies. Specificity in protein phosphorylation is achieved in part because kinases and phosphatases are spatially organized inside cells. A prototypic example is compartmentalization of the cAMP dependent protein kinase A through association with A-kinase anchoring proteins. This configuration creates autonomous signaling islands where the anchored kinase is constrained in proximity to activators, effectors, and selected substates. This article primarily focuses on AKAP signaling in the heart with an emphasis on anchoring proteins that spatiotemporally coordinate excitation-contraction coupling and hypertrophic responses.

Historical Overview

Seminal Discoveries

Phosphorylation is a posttranslational modification that is a cornerstone of many intracellular signaling cascades. This simple chemical cue has evolved into a sophisticated signaling paradigm that controls a myriad of physiological processes. This reaction occurs when the γ-phosphate group of adenosine 5’-triphosphate (ATP) is covalently linked to serine, threonine, tyrosine or histidine side chains¹. The forward reaction, carried out by protein kinases, introduces negative charge, or changes the shape of the modified sidechain to alter the biochemical properties of the substrate protein^2,3. The reverse reaction is catalyzed by phosphoprotein phosphatases that remove phosphate groups from substrates³. Although protein phosphorylation was first reported in the 1930’s, we’re only now beginning to grasp the full extent to which this covalent modification governs the biology of cells and underlies disease. Hence, pharmacological control of protein phosphorylation is a major therapeutic objective with 76 kinase inhibitor drugs currently approved for use and many more in clinical trials⁴.

Carl and Gerty Cori laid the groundwork for the study of phosphorylation through their pioneering research into the process of glycogen metabolism undertaken at the University of Washington St. Louis⁵. They discovered an intermediate compound that assisted in the breakdown of glycogen called glucose-1-phosphate, and identified glycogen phosphorylase as the rate limiting enzyme in glycogen production. They received the Nobel prize for Physiology in 1947 for this discovery.

The next breakthrough was the discovery of cyclic nucleotides as intracellular second messengers of hormone action. Sutherland and Rall, working at Vanderbilt University in 1957, showed that phosphorylase kinase was activated in a cyclic adenosine 3’,5’-monophosphate (cAMP) dependent manner^6,7. They showed that the primary messengers epinephrine and glucagon signaled the formation of cAMP by adenylyl cyclase (AC), which led to the activation of phosphorylase kinase^6,8. This work won the Nobel Prize in 1971.

Following up on earlier work into glycogen metabolism, biochemists Edmond H. Fischer and Edwin G. Krebs at the University of Washington would expand the study of glycogen phosphorylase. This enzyme exists in active and inactive forms termed phosphorylase a and b respectively. Fischer and Krebs demonstrated that the transfer of a phosphoryl group donated by Mg-ATP was necessary for the phosphorylase b to a transition^9,10. We now know that phosphorylation of Ser14 converts dormant phosphorylase b into active phosphorylase a. This phosphoryl group is removed by an enzyme which came to be known as phosphoprotein phosphatase 1¹⁰. Hence Fischer and Krebs became the first to define the reversible nature of protein phosphorylation. This was followed by the discovery of cAMP dependent protein kinase (also called protein kinase A or PKA) by Krebs, Walsh, and Perkins in 1968¹¹. Fischer and Krebs were awarded the Nobel Prize in 1992 for their discovery of protein regulation by reversible phosphorylation¹².

The utility of phosphorylation as a commonly used signaling mechanism gained further prominence in 1979 when Tony Hunter and colleagues at the Salk institute revealed phosphorylation on tyrosine while studying the Rous Sarcoma virus^13,14. Tony Pawson in Toronto and Joseph Schlessinger, now at Yale, independently defined phosphotyrosine as a recognition motif for Src homology 2 (SH2) domains in the assembly of macromolecular signaling complexes^15,16. From then onward, the study of protein phosphorylation has been inextricably linked to research advances related to cancer, diabetes, and the cardiovascular system. A timeline of key developments in our understanding of PKA phosphorylation is presented in figure 1.

Figure 1: Timeline of key discoveries pertaining to cAMP responsive phosphorylation.

(Figure created in Biorender)

cAMP Dependent Protein Kinase (PKA)

The utility of protein phosphorylation as an adaptable form of cellular regulation is underscored by evidence that 1–2% of most genomes encode protein kinases. This includes over 540 genes in humans^17–19. Protein kinase A signaling is often hailed as an archetype to understand the action of eukaryotic protein kinases. Yet, this enzyme is atypical in many respects²⁰. Of the eukaryotic protein kinases, only PKA and casein kinases exist as tetrameric holoenzymes whose catalytic activity is autoinhibited by a separate family of regulatory subunits²⁰. When steroids, neurotransmitters, or hormones engage G-protein coupled receptors (GPCRs), they mobilize adenylyl cyclases to promote a rapid rise in intracellular levels of cAMP²¹. Importantly, activation of PKA is dependent upon cAMP binding rather than autophosphorylation within the active site of the enzyme^22,23. PKA acts a master regulator of diverse cellular processes including (but not limited to) glycogenolysis, transcription, proliferation, hormone synthesis and release, and excitation-contraction coupling²⁰. Moreover, the specificity and spatiotemporal resolution of PKA action is determined by a family of A Kinase Anchoring Proteins (AKAPs) that sequester this key enzyme with preferred substrates^24–26

PKA exists as a holoenzyme composed of a dimer of regulatory subunits (R) that bind and constrain catalytic subunits (C)²⁷. In 1970, Gill and Garren were the first to resolve the regulatory and catalytic components of the holoenzyme using bovine adrenal cortical PKA²⁸. At the same time, Tao, Salas and Lipmann used sucrose density gradients to fractionate rabbit reticulocyte PKA holoenzymes into catalytic and regulatory subunits²⁹. Elegant follow-up studies by Susan Taylor, Jackie Corbin and Edwin G. Krebs showed that the R subunits existed in two forms (I and II), and that each protomer was capable of binding two molecules of cAMP^30,31–33. We now know that binding of cAMP to regulatory subunits relieves autoinhibitory control over the catalytic subunits; thereby activating the intact holoenzyme^34–36. Around the same time Walsh and Krebs uncovered a heat stable protein inhibitor of PKA catalytic activity isolated from rabbit skeletal muscle, termed PKI³⁷. Kinetic studies of PKI and peptide analogues by Scott, Fischer and Krebs defined the mechanism of kinase inhibition^38–40. Concurrent studies by Kemp, Walsh and colleagues came to the same conclusions⁴¹. PKI peptides proved to be a valuable tool utilized by Sodwaski and Taylor to lock the PKA catalytic subunit into a conformation that allowed it’s successful crystallization⁴² (Figure 1). The significance of this latter work cannot be understated, as it revealed the bilobal fold that is a recognizable hallmark of all protein kinases^43,44.

Several laboratories have amassed considerable evidence that cAMP signaling events are much more spatially controlled than originally considered. Recent work on the boundaries of cAMP nanodomains shows that discrete and autonomous pockets of cAMP are generated upon the mobilization of G-protein coupled receptors^45–47. Discerning the structure of AKAP signaling islands provides a molecular basis for how PKA is constrained within second messenger signaling compartments^34,35. These studies show that the anchored PKA holoenzyme is maintained within 200–400 angstroms of selected substrates²⁵. An implication of this revised view of cAMP signaling is that local phosphorylation events often occur within the narrow confines of AKAP signaling islands. Consequently, mutations that alter AKAP function or impact the localization of PKA underlie disease^48,49.

A Kinase Anchoring Proteins (AKAPs)

As the utility of cAMP signaling became more apparent, investigators realized that molecular mechanisms must exist to simultaneously coordinate multiple phosphorylation events inside cells. This led to the notion that PKA was compartmentalized⁵⁰. Three independent observations set the stage for the discovery of AKAPs: PKA was reported to be attached to microtubules via its regulatory subunits⁵¹, a calmodulin binding protein was identified as a contaminant with purified R subunits⁵², and RII-binding proteins were identified by far western blots using phosphorylated RII as the bait⁵³. This technique is commonly referred to as an RII overlay⁵⁴. Still, a systematic interrogation of localized PKA action was not possible until resolution of the PKA-AKAP binding interface^27,55,56 (Figure 2A). Deletion screens of RII carried out by Scott in 1990 identified the first 79 amino acids of RII as necessary and sufficient for binding to microtubule associated protein 2 (MAP2)⁵⁶. Moreover, removal of the first 14 amino acids of RII not only abolished dimerization, but also prevented interaction with MAP2⁵⁶. By the late 1990’s the structure of the AKAP binding interface with RII was solved. This is now known as the docking and dimerization domain^57,58. A hydrophobic cleft is formed by dimerization of the R subunits which fold into an antiparallel x-type helical bundle²⁷. This motif is a common structural feature of all PKA regulatory subunits, as well as a newly identified family of 16 AKAP-interacting factors called R1D2 and R2D2 proteins^59–62 (Figure 2A).

Figure 2: AKAPs localize PKA to a variety of intracellular locations.

A) Structural model of the AKAP amphipathic helix (purple) in complex with the docking and dimerization domain of RIIα. Both protomers of the RII dimer are depicted in shades of grey. Model made using PYMOL 2.5 and structures from the protein database with accession number 2izx. B) Schematic of a cell indicating the locations of compartment-specific AKAPs (schematic created in Biorender).

Reciprocal studies carried out by Scott and colleagues defined PKA binding motifs on AKAPs. Using human thyroid protein clone 31 (Ht31, now called AKAP-Lbc) as a model, mutational analysis revealed a 14–18 amino acid sequence that was both essential for RII binding and predicted to form an amphipathic α-helix⁶³. Site-directed mutagenesis of Ht31 aimed at disrupting its secondary structure confirmed that AKAPs bound RII via an amphipathic α-helix^27,54,63. Cell soluble, stapled, and chemically stabilized peptide disruptors of AKAP/RII binding have subsequently been developed based on this RII binding sequence^61,64–67. These reagents and small molecule inhibitors have become valuable tools for investigating anchored PKA function^68–72. Acquiring this fundamental understanding of the AKAP/PKA binding interface has provided a molecular toolkit to study all AKAPs and primed the field for significant advancement (Figure 2A).

One angstrom resolution crystal structures of the RIIα docking and dimerization (DD) domain in complex with AKAP peptides add molecular detail to this protein-protein interface^27,59,73. Two isoforms exist of each regulatory subunit (RI_α, RI_β, RII_α, and RII_β). The topology of the DD domain varies depending on the isoform of the R subunit present (Figure 2A). The depth of the cleft affects the number of points of contact with the AKAP α-helix. Consequently, AKAPs exhibit different binding affinities for each subtype of PKA, with the majority preferentially binding type II (so named for the R subunit)^24,27,59,73. Additionally, RI selective, and dual specificity AKAPs have been defined^74–80. Through these high affinity protein-protein interactions, PKA holoenzymes are sequestered at defined locations within cells (Figure 2B). To date, sixty human genes encoding over 150 isoforms have been classified as AKAPs²⁵. As depicted in figure 2B, multiple AKAPs can be targeted to the same subcellular compartment. This does not necessarily reflect a redundancy of function, but rather the exquisite degree of spatial organization provided by AKAPs.

AKAP Signaling Islands

Arguably the most important feature of AKAPs is their ability to organize combinations of enzymes into macromolecular signaling complexes^81,82 (Figures 3&4). In this way, these anchoring proteins spatially integrate distinct cell signaling pathways and temporally control activation and termination of cAMP signaling cascades. First evidence for the multivalence of AKAPs came from the observation that the calcium/calmodulin dependent phosphatase Calcineurin/PP2B was an interacting partner for the neuronal anchoring protein AKAP79⁸³. Functional evidence in hippocampal neurons demonstrated that localization of PKA with PP2B by AKAP79 regulated aspects of synaptic transmission^84,85. Subsequent studies on the same signaling complex identified protein kinase C (PKC) as an additional binding partner⁸⁶. Biochemical, genetic, and functional studies have gone on to show how different combinations of AKAP79-anchored enzymes modulate transmembrane receptors and ion channels in neuronal, endocrine and muscle tissues^84,87–91.

Figure 3: AKAPs coordinate excitation-contraction coupling.

A) Schematic (created in Biorender) depicting AKAP regulation of channel activity, ion flow, contraction, and relaxation of cells. B-D) Confocal images depicting the localization of AKAP18 with SERCA2 ATPase at the Z bands (see zoom panel) within normal human heart (paraffin embedded tissue sections).

Figure 4: AKAPs mediate the heart’s hypertrophic response.

A) Schematic (created in Biorender) depicting AKAP regulation of hypertrophic gene expression, reactive oxygen species generation, and mitochondrial morphology. B-D) Three-dimensional structured illumination microscopy shows the localization of dAKAP1 to the mitochondria of U2OS cells.

Additional interactions with AKAP79 preferentially place PKA in proximity to adenylyl cyclase 5, the enzyme that generates intracellular cAMP^{21 92–94}. This configuration creates an internally constrained negative feedback loop as anchored PKA phosphorylation of adenylyl cyclase 5 inhibits cAMP synthesis. Elegant work by Carmen Dessauer and colleagues has defined AKAP-AC interactions in various cellular systems^94–96. AKAPs also associate with GPCRs, as indicated by initial work focused on anchored regulation of β-adrenergic signaling^97–99. More recently it has been shown that PKA can anchor to other GPCR’s. For example, Stefan and colleagues report that PKA is directly anchored to orphan receptor Gpr161 in primary cilia¹⁰⁰.

Spatiotemporal control of cAMP signal termination is achieved by AKAP association with phosphodiesterases (PDEs). The was first reported in 2001 when it was demonstrated that muscle-selective mAKAP, maintains a cAMP signaling module, including PKA and the cAMP-specific phosphodiesterase (PDE4D3) in the heart^101–104. Around the same time, it was noted that AKAP450 constrains PKA and PDE4 at the centrosomes in Sertoli cells^105,106. Importantly the long isoforms of PDE4’s are substrates for anchored PKA, and phosphorylation enhances cAMP degradation^103,107. These reports suggest that proximity of PKA and PDE’s in AKAP signaling islands sustain a negative feedback loop to restore basal cAMP levels²⁴. Further support for tight spatial control of cAMP signaling was provided when it was shown that mAKAP anchoring of PDE4D3 not only reduces local cAMP concentrations, but also provides spatial control over the cAMP responsive guanine nucleotide exchange factor EPAC-1^102,108. Thus, AKAP signaling complexes participate in all aspects of cAMP signaling from the initiation of second messenger responses, to the termination cAMP signals, and bi-directional control of phosphorylation events.

Tissue Specific AKAP Regulation of Physiological Function

AKAPs in the Reproductive and Central Nervous Systems

Upwards of 150 tissue specific AKAP isoforms are transcribed from the human genome²⁴. AKAPs are particularly prevalent in male and female germ cells, where they contribute to sperm development and motility as well as maturation of the female genital tract¹⁰⁹. The distinctive morphology of sperm and their specialized intracellular environment favor the local activation of anchored cAMP signaling elements. AKAP110, AKAP220 and sAKAP84 are linked to sperm motility via the clustering of phosphodiesterases alongside PKA^110,111. Local cAMP production in sperm is mediated by a soluble adenylyl cyclase called sAC or AC10 to create local nanodomains of second messenger¹¹². Consideration of the aforementioned observations may help to explain the abundance of AKAPs expressed in male germ cells. In the female reproductive system, D-AKAP1 has been linked to oocyte maturation as evidenced by infertility in female knockout mice. In addition to the regulation of germ cell development, the anchoring protein Gravin is required for proper gastrulation in zebra fish¹¹³. This AKAP is thought to mediate axis elongation via inhibition of the Rho/ROCK pathway¹¹³.

One of the first neuronal AKAPs characterized, AKAP79 (or its murine ortholog AKAP150), facilitates synaptic signaling events⁷¹. AKAP79 scaffolds calcineurin (PP2B) with PKA and PKC by associating with both the plasma membrane and postsynaptic density proteins. This affects synaptic plasticity on multiple levels, principally by regulating the activity of various ion channels^86,84,114 Other AKAPs also contribute to neuronal functionality. For example, WAVE-1 anchors PKA and the tyrosine kinase Abl to the actin cytoskeleton to mediate synaptic remodeling. Analysis of WAVE-1 knockout mice revealed that loss of this signaling complex underlies defects in axonal guidance, sensory motor function and hippocampal learning and memory formation^115–117. As these topics have been extensively reviewed, we have chosen to focus on AKAP signaling in the heart.

AKAPs in the Heart

According to the American Heart Association and National Institutes of Health, cardiovascular diseases in 2020 accounted for nineteen million deaths worldwide. This represents an increase of 18.7% over the previous decade. Hence, the development of molecular therapies for the prevention and treatment of cardiomyopathies is a critical healthcare initiative. The abundance of cardiac AKAPs makes them attractive therapeutic targets. Consequently, a significant body of work has focused on characterizing AKAP signaling in the heart. Cardiac AKAPs identified to date include: AKAP79/150, AKAP18, mAKAP, AKAP-Lbc, dAKAP1, dAKAP2, p110γ, myomegalin, troponin T, AKAP95, AKAP220, gravin, ezrin, synemin, SKIP, BIG2, cypher/zasp, and yotiao. These anchoring proteins coordinate physiological processes including excitation contraction coupling, calcium homeostasis, myogenic tone, contractility, gene transcription, pathological cardiac remodeling and hypertrophy as well as oxidative stress responses^24,118

AKAP Regulation of EC Coupling

The recurrent contraction and relaxation of cardiomyocytes requires the synchronized coordination of cAMP and Ca2+ signaling events (Figure 3A). This process, known as excitation-contraction (EC) coupling, occurs when an action potential from the sinoatrial node is transduced into a chemical cue via voltage-gated ion channels localized to introversions of the sarcolemma called the transverse tubules¹¹⁹ (Figure 3B–D). Following membrane depolarization, Ca²⁺ enters the cell through L-type Ca²⁺ channels. This stimulates further Ca²⁺ release from intracellular stores via ryanodine receptors (RyRs) within the membrane of the Sarcoplasmic Reticulum (SR). This transient rise in intracellular Ca²⁺ triggers contraction of the myofibrils¹²⁰. The final phase of relaxation is largely dependent upon reuptake of calcium through the sarcoplasmic/endoplasmic reticulum Ca²⁺ ATPase (SERCA2)^119,121,122 (Figure 3A). EC coupling is a high fidelity process, with one action potential corresponding to one contraction¹²³. Consequently, disruption of signaling events at any level of this tightly controlled cascade (EC uncoupling) can result in cardiac arrhythmias and eventually heart failure.

EC coupling is regulated at many levels by PKA signaling downstream of adrenergic stimulation¹²⁴. Catecholamines engage β-adrenergic receptors coupled to the stimulatory G_αs pathway. This stimulates adenylyl cyclase to produce cAMP, followed by PKA phosphorylation of substrates¹²⁵ including L-type calcium channels (primarily Ca_v1.2)¹²⁶, phospholamban (PLN)^127,128, and ryanodine receptors¹²⁹. Anchoring of PKA is crucial for maintenance of proper calcium homeostasis and EC coupling. Anchoring proteins, including AKAP79/150, AKAP18, mAKAP, Gravin and Yotiao have been implicated in this process. Each AKAP coordinates distinct elements of this highly choreographed, vital, and repetitive physiological process.

AKAP79/150

Early biochemical and electrophysiological studies demonstrated that AKAP79 localized PKA to L-type calcium channels at the plasma membrane to mediate phosphorylation dependent modulation of ion conductance^130,131 (Figure 3A). It has also been shown that PKA inhibition in native cardiomyocytes negatively impacts calcium currents downstream of β-adrenergic stimulation¹³⁰. Yet, mouse models lacking AKAP5 (the gene encoding AKAP79/150) demonstrate that cardiomyocytes do not respond to β-adrenergic stimulation in the absence of this anchoring protein¹³². Calcium influx is mediated, in part, by the association of AKAP79 with Ca_v1.2¹³³. This heteromeric channel is comprised of a transmembrane subunit and variable accessory subunits. AKAP79/150 associates with Ca_v1.2 via the C-terminus to target kinases and phosphatases capable of modifying the channel¹³⁴. AKAP5^−/−mice display improper trafficking of β-adrenergic receptors, suggesting a role for this AKAP signaling island in receptor resensitization¹³⁵. Other regulatory events may also participate in modulating Ca_v1.2 conductance.

AKAP18

AKAP18 is concentrated at the Z bands of cardiomyocytes (Figure 3B–D). Initial studies characterized the association of AKAP18 (also called AKAP15) with cardiac L-type calcium channels^136,137. AKAP7, the gene encoding AKAP18, is subject to alternative splicing. The resultant AKAP18 isoforms vary in length, localization, tissue specific distribution, and function^138,139. Splice variation results in the formation of at least five major AKAP18 isoforms, two short and three long (AKAP18α, AKAP18β, AKAP18γ, AKAP18δ and AKAP18ε)^138,140,141. These isoforms are all expressed in the heart. AKAP18α is targeted to plasma membranes by lipid modification of its N-terminus in the form of myristoylation and palmitoylation¹³⁶. Once at the membrane, AKAP18α associates with L-type calcium channels via a conserved leucine zipper motif on the C-terminus of the channels^142,143 (Figure 3A). Numerous studies have shown that Ca_v1.2 is phosphorylated by PKA on multiple sites (Ser1700, Thr1704, Ser1928, and Ser1518) and that this modification increases calcium channel conductance and contractile force^{126,142,144–149}. Although anchoring of PKA with Cav1.2 is required for the maintenance of calcium homeostasis, it remains unclear exactly how channel activity is regulated from a molecular standpoint¹⁵⁰. For example, PKA phosphorylation of the monomeric G protein Rad has recently been shown to contribute to channel activation¹⁵¹.

As previously mentioned, the long isoforms of AKAP18 lack membrane targeting sequences. AKAP18γ and AKAP18δ contain unique localization sequences which direct them to organelles such as the sarcoplasmic reticulum (Figure 3A). It’s well established that SERCA2 activity is mediated by its association with a small transmembrane protein called phospholamban. Depending on its phosphorylation state, PLN modifies the activity of SERCA2. Unphosphorylated PLN binds SERCA2 and inhibits calcium transport activity. Phosphorylated PLN releases its inhibitory hold on SERCA2, thus allowing for calcium reuptake into the SR and cardiomyocyte relaxation. Previous studies have shown that AKAP18δ scaffolds PKA with SERCA2 and PLN¹⁵². PKA is thus positioned to phosphorylate PLN, facilitating SERCA2 function, cardiomyocyte relaxation and maintenance of excitation-contraction coupling¹⁵² (Figures 3B–D).

The localized activity of other kinases and phosphatases also participate in maintaining EC coupling. For example, calcium/calmodulin-dependent protein kinase II (CAMKII) has been linked to the regulation of calcium homeostasis. As described above, adrenergic stimulation elicits cellular calcium entry through L-type calcium channels. This influx triggers further release of calcium from intracellular stores within the sarcoplasmic reticulum into the cytosol via ryanodine receptors. Recent work has demonstrated an additional role for AKAP18δ in scaffolding CAMKII to ryanodine receptors¹⁵³. Researchers have employed the use of peptide arrays, surface plasmon resonance, structural modeling, and advanced microscopy to show that different regions of AKAP18δ differentially regulate CAMKII activity. The N-terminal region of AKAP18δ binds and inhibits activation of CAMKII. Paradoxically, binding at the C-terminal region of AKAP18δ enhances CAMKII activation and promotes PLN phosphorylation. The net effect being faster Ca²⁺ release by RyRs and reuptake into the SR^153,154. This represents a more nuanced AKAP-mediated mechanism of regulating calcium homeostasis.

Gravin and Yotiao

Gravin scaffolds protein kinase A and protein kinase C with β₂-adrenergic receptors (β₂-ARs)^155,156 (Figure 3A). This product of the AKAP12 gene positively regulates cardiac contractility. Gravin influences positive ionotropic effects by mediating β₂-AR desensitization¹⁵⁷. Interestingly, contractility is increased with and without β-adrenergic stimulation in mouse models lacking gravin^157,158. In fact, the phosphorylation state of PLN and SERCA2 remains unchanged in these animals as compared with wild type controls¹⁵⁸. While details of this molecular mechanism need to be clarified, gravin signaling islands are thought to maintain receptor sensitivity.

PKA-mediated phosphorylation of slow acting potassium channels is also important for the maintenance of EC coupling. These outward potassium currents (I_Ks) help to repolarize the membrane and contribute to myocardial relaxation or lusitropy. Elegant studies have shown that Yotiao scaffolds PKA and protein phosphatase 1 (PP1) with the KCNQ1 subunit of the channel¹³¹ (Figure 3A). Interestingly, yotiao itself is phosphorylated on Serine43 by PKA following adrenergic stimulation¹³¹. Mutational analysis revealed that substitution of Ser43 with alanine diminished channel conductance but did not alter channel phosphorylation, or the association of Yotiao with KCNQ1. Additionally, allosteric interactions that proceed through AC9 may prime KCNQ1 for phosphorylation by PKA⁹⁵. Therefore, the yotiao macromolecular complex is allosterically modulating the activity of the potassium channel¹⁵⁹.

AKAP coordination of Cardiac Hypertrophy

Cardiac hypertrophy is the heart’s primary adaptive response to pressure or volume stressors, and/or damage to the overall contractile machinery (Figure 4A). Damage to the heart can result from a variety of factors including injury and disease. Hypertrophy describes either the concentric enlargement or elongation of cardiomyocytes as a compensatory mechanism in place of mitotic proliferation. Acute hypertrophic signaling can be beneficial for the heart. For example, the muscle cells in the heart can enlarge in response to exercise in order to compensate for the increased demand for cardiac output¹⁶⁰. In contrast, chronic hypertrophic signaling underlies cardiomyopathies such as ischemia and hypertension that eventually lead to heart failure¹¹⁸. This pathological hypertrophy comprises a thickening of the myocardioum, reduction in chamber volume, interstitial fibrosis and reduced cardiac output. These systemic changes are accompanied by alterations in gene expression and cellular metabolism¹¹⁸ (Figure 4A). Adult cardiomyocytes are largely terminally differentiated and are therefore particularly susceptible to pathological hypertophy. Cardiac AKAPs can elicit pro or anti-hypertrophic effects depending upon the composition of binding partners within each signaling island (Figure 4A).

AKAP-Lbc

Multiple signaling elements converge to regulate the heart’s complex hypertrophic response. These include cytokines, MAP kinases, protein phosphatases, adenylyl cyclases, phosphodiesterases and transcription factors. The ability of AKAPs to coordinate these disparate enzymes make them key molecular platforms for the integration and propagation of intracellular signals. Activation of hypertrophic gene regulatory programs results in changes in cellular metabolism and cytoskeletal remodeling associated with disease. AKAP-Lbc signaling islands interface with the myocyte enhancer factor (MEF) pathway that is turned on in response to elevated catecholamines to promote pathological remodeling^113,161. In particular, the transcription factor MEF2 is dephosphorylated by calcineurin in complex with AKAP-Lbc, allowing for its translocation to the nucleus and upregulation of hypertrophic gene expression (Figure 4A). AKAP-Lbc is also associated with class II histone deacetylases (HDACs 4 and 5)¹⁶¹. These enzymes repress the transcription of certain genes. AKAP-Lbc positions PKC to phosphorylate PKD, allowing for the latter to be translocated into the nucleus¹⁶¹. Once there, PKD phosphorylates HDACs, eliciting their nuclear export. This anchored signaling event facilitates MEF-mediated activation of gene regulatory pathways associated with pathological hypertrophy^113,131,161.

AKAP-Lbc signaling islands mediate other aspects of hypertrophic signaling. This anchoring protein also encodes a guanine nucleotide exchange factor for Rho (Rho GEF), which then induces p38 MAPK signaling downstream of α-adrenergic receptor stimulation in cardiomyocytes to induce hypertrophy^162–164 (Figure 4A). AKAP-Lbc therefore acts as an intermediate, linking α-adrenergic signaling to the mobilization of MAP kinase activity. Oxidative stress, a scenario in which the generation of reactive oxygen species (ROS) exceeds the cell’s natural clearing capacity by antioxidant enzymes, has also been linked with the onset of hypertrophy^165,166. This imbalance can derive from physical and chemical stressors; and may lead to impaired contractility and apoptosis in addition to hypertrophy^118,167. AKAP-Lbc protects against ROS induced damage by several mechanisms. In response to α1-adrenergic stimulation, AKAP-Lbc associated protein kinase D (PKD) phosphorylates and inactivates the phosphatase slingshot 1L (SSH1L). This circumvents the activation of a cofilin2/Bax subcomplex at the mitochondria¹⁶⁸. This protects the cell from the increased ROS production that is emblematic of hypertrophy. In this way, AKAP-Lbc elicits a cardioprotective effect. AKAP-Lbc mediated activation of PKD also results in the phosphorylation of CREB which in turn upregulates expression of the anti-apoptotic gene Bcl2^118,168. Hence, the pleiotropic effects elicited by AKAP-Lbc signaling make it a complex therapeutic target for the treatment of heart disease.

dAKAP1

Mitochondrial AKAPs affect cellular respiration, ROS production, and survival^169,170. For example, the dual specificity anchoring protein dAKAP1 (a product of the AKAP1 gene, also called S-AKAP84, AKAP121 or AKAP149) has been linked to cardiac hypertrophy via its regulation of mitochondrial ROS generation^74,171–173. One isoform, AKAP121 acts in a cardioprotective manner by controlling mitochondrial morphology via the phosphorylation of Drp1 (Figure 4A–D). Further support for the cardioprotective role of AKAP121 is provided by evidence that these signaling islands repress nuclear translocation of the nuclear factor of activated T-cells NFATc3 that would otherwise enable hypertrophic gene expression¹⁷⁴. In vivo studies consolidate this notion by showing that under conditions mimicking pressure overload, AKAP121 expression is depleted. This results in mitochondrial dysfunction, oxidative stress, and cell death¹⁷². Thus, cardiac function at the level of the mitochondria may proceed through dAKAP1/AKAP121 (Figures 4B–D). However, other anchoring proteins may also participate in this process. For example, SKIP, an entirely RI-selective AKAP, is located at the inner mitochondrial membrane where it regulates metabolic processes⁸⁰. Consequently, local control of the hypertrophic response undoubtedly involves the concerted influence of multiple AKAP signaling islands (Figure 4A).

mAKAP

mAKAP is also associated with hypertrophic gene expression. This anchoring protein, originally called AKAP100, was identified in the original RII overlay screen for AKAPs conducted in the Scott laboratory¹⁷⁵. Due to splice variation, there are two main isoforms of mAKAP (mAKAPα and mAKAPβ). Full length mAKAPα is a 255 kDa protein mainly localized to neurons, with mAKAPβ (230 kDa) being the predominant isoform in the heart. mAKAPβ associates with the nuclear envelope in cardiomyocytes through direct binding to an integral membrane protein called nesprin1-α¹⁷⁶. Biochemical analyses have identified numerous mAKAP binding partners including PKA¹⁷⁷, PDE4D3¹⁰¹, adenylyl cyclases⁹⁴, PP2A and calcineurin^178,179, RyRs^{129,178,180,181}, Rap1 and EPAC1¹⁰⁸, and the MAP kinases MEK and ERK^108,182 (Figure 4A).

Extensive in vitro and in vivo studies indicate that the primary directive of the mAKAP signaling complex is to alter gene expression coincident with myocyte remodeling. The transcripton factors NFATc and MEF2 are dephosphorylated by mAKAP associated calcineurin. This triggers their translocation to the nucleus and the upregulation of gene expression associated with hypertrophic cardiac remodeling^183–185. PKD is also activated in complex with mAKAP, leading to phosphorylation of HDAC4 and the subsequent derepression of hypertophic genes^183,186. Interestingly, knockout of mAKAPβ expression has been shown to elicit a cardioprotective effect in mouse models of pressure overload and hypertrophy¹⁸³. Conversely, it has been shown that mAKAP scaffolding of E3 ubiquitin ligases with hypoxia-inducible factor 1 (HIF-1α) mediates the stability of HIF-1α. In response to hypoxia, HIF-1α is stabilized so that it can translocate to the nucleus and promote the transcription of cell survival genes¹⁸⁷. These studies demonstrate that mAKAP provides a multipurpose molecular platform to coordinate enzyme activities necessary for the processing of multiple hypertrophic signals.

Conclusions and Perspectives

From its humble beginnings as a regulatory element of glycogen metabolism, we now recognize that protein phosphorylation is a universal biochemical regulatory mechanism on par with protein ubiquitination, acetylation and proteolysis⁴. Many pivotal studies on phosphorylation were conducted in the heart, illustrating how this mode of covalent modification controls force generation and contractility. Spatiotemporal control of protein phosphorylation by AKAPs may be particularly important in this context because of the speed and repetition required for EC coupling. Likewise, modulation of adaptive responses such as physiological and pathological cardiac hypertrophy require the relay of signals from the plasma membrane to the transcriptional machinery embedded in the nucleus. Thus, mechanisms that confer local control of protein phosphorylation events have evolved to accommodate the specialized function of cardiomyocytes (Figures 3A & 4A).

Elegant fluorescent spectroscopy studies show that physiological accumulation of cAMP occurs within nanometer-sized domains^45–47. This new model is consistent with our own molecular evidence that anchored, intact, and active PKA holoenzymes operate within a 200 to 400-A° radius^34,35. This creates discrete and autonomous pockets of activity where anchored PKA phosphorylates preferred substrates; and could explain why multiple AKAPs are located at the SR, plasma membranes, mitochondria, and nuclear envelope in cardiomyocytes. For example, AKAP79, AKAP18α and Cypher/Zasp control phosphorylation of L-type Ca²⁺ channels to initiate EC coupling. Close by, mAKAP modulates ryanodine receptors to promote further calcium release, while AKAP18γ facilitates calcium reuptake by SERCA2. This highlights the remarkable degree of spatial organization provided by AKAPs that is necessary to ensure processive implementation of distinct phosphorylation events (Figure 3A).

Perhaps not surprisingly, aberrant AKAP signaling can have pathological ramifications. As previously mentioned, depletion of mitochondrial AKAP121 occurs under conditions of pressure overload. Polymorphisms in AKAP18α that abolish PKA anchoring alter regulation of cardiac L-type Ca²⁺ channels and are linked with increased susceptibility to febrile seizures¹³⁸. Likewise, polymorphisms in d-AKAP2/AKAP10 are linked to arrythmias and may be related to increased cardiac dysfunction within the aging population¹⁸⁸. In another pathological context, Cushing’s syndrome, mutations in PKAc that prevent its incorporation into AKAP signaling islands underlie hypertension, arrythmia and other cardiovascular complications⁴⁹. Taken together, these examples underscore the exquisite degree of subcellular organization provided by AKAPs. Thus, targeting drugs to AKAPs offers an appealing therapeutic strategy to exploit this sophisticated signaling terrain. Such precision pharmacological regimens would endeavor to minimize off-target effects. While local kinase inhibition is currently effective at tethering modified drug adducts to their sites of action^189,190, there is hope that this strategy can be refined to deliver small molecules that allosterically inhibit anchored kinases. In closing, Eddy Fischer and Ed Krebs would be astonished at the advances made in understanding reversible protein phosphorylation. They would be equally enthusiastic and intrigued to find out what the next generation of researchers will discover.

Abbreviations

PKA

protein kinase A

AKAPs

A-Kinase Anchoring proteins

ATP

adenosine 5’-triphosphate

cAMP

cyclic adenosine 3’,5’-monophosphate

AC

adenylyl cyclase

SH2

Src homology 2 domain

GPCR

G protein coupled receptor

R subunit

PKA regulatory subunit

C subunit

PKA catalytic subunit

Ht31

human thyroid protein clone 31

PKI

heat stable peptide inhibitor of PKA

DD

docking and dimerization domain of PKA holoenzyme

PKC

protein kinase C

PDE

phopsphodiesterase

SAC

soluble adenylyl cyclase

PP2B

protein phosphatase 2B, calcineurin

CaN

calcineurin

EC coupling

excitation-contraction coupling

Ca²⁺

calcium

RYR

ryanodine receptors

SR

sarcoplasmic reticulum

SERCA2

sarcoplasmic/endoplasmic reticulum Ca²⁺ ATPase

PLN

phospholamban

CAMKII

calcium/calmodulin-dependent protein kinase II

β₂–AR

β₂-adrenergic receptors

I _Ks

outward potassium currents

PPI

protein phosphatase 1

MEF

myocyte enhancer factor

HDACs 4 and 5

histone deacetylases 4 and 5

Rho GEF

guanine nucleotide exchange factor for Rho

ROS

reactive oxygen species

SSH1L

slingshot 1L

PKD

protein kinase D

NFAT

nuclear factor of activated T-cells

HIF-1a

hypoxia-inducible factor 1a

MAP2

microtubule associated protein 2