Cardiovascular Research

Cardiovascular Research 1998 38(3):559-588; doi:10.1016/S0008-6363(98)00063-7
© 1998 by European Society of Cardiology

This Article Abstract FREE Full Text (PDF) Alert me when this article is cited Alert me if a correction is posted Services Email this article to a friend Similar articles in this journal Similar articles in PubMed Alert me to new issues of the journal Add to My Personal Archive Download to citation manager Request Permissions Google Scholar Articles by Rapundalo, S. T Search for Related Content PubMed PubMed Citation Articles by Rapundalo, S. T Social Bookmarking          What's this?

Copyright © 1998, European Society of Cardiology

Cardiac protein phosphorylation: functional and pathophysiological correlates

Stephen T Rapundalo^*

Department of Biochemistry, Parke-Davis Pharmaceutical Research, Division of Warner-Lambert, 2800 Plymouth Road, Ann Arbor, MI 48105, USA

^* Tel.: +1 (734) 622 5170; Fax: +1 (734) 622 5987; E-mail: stephen.rapundalo@aa.wl.com

Received 19 June 1997; accepted 4 February 1998

	Abstract

Top Abstract 1 Introduction 2 Cardiac phosphoproteins 3 Coordinated regulation of... 4 Pathophysiological alterations... 5 Summary and perspectives References

 
Protein phosphorylation acts a pivotal mechanism in regulating the contractile state of the heart by modulating particular levels of autonomic control on cardiac force/length relationships. Early studies of changes in cardiac protein phosphorylation focused on key components of the excitation-coupling process, namely phospholamban of the sarcoplasmic reticulum and myofibrillar troponin I. In more recent years the emphasis has shifted towards the identification of other phosphoproteins, and more importantly, the delineation of the mechanistic and signaling pathways regulating the various known phosphoproteins. In addition to cAMP- and Ca²⁺-calmodulin-dependent kinase processes, these have included regulation by protein kinase C and the ever-emerging family of growth factor-related kinases such as the tyrosine-, mitogen- and stress-activated protein kinases. Similarly, the role of protein dephosphorylation by protein phosphatases has been recognized as integral in modulating normal cardiac cellular function. Recent studies involving a variety of cardiovascular pathologies have demonstrated that changes in the phosphorylation states of key cardiac regulatory proteins may underlie cardiac dysfunction in disease states. The emphasis of this comprehensive review will be on discussing the role of cardiac phosphoproteins in regulating myocardial function and pathophysiology based not only on in vitro data, but more importantly, from ex vivo experiments with corroborative physiological and biochemical evidence.

KEYWORDS Phosphoproteins; Regulatory proteins; Protein kinases; Protein phosphatases; Dephosphorylation; Excitation-contraction coupling

	1 Introduction

Top Abstract 1 Introduction 2 Cardiac phosphoproteins 3 Coordinated regulation of... 4 Pathophysiological alterations... 5 Summary and perspectives References

 
Phosphorylation and dephosphorylation of proteins is widely recognized as an important mechanism for regulating cellular function by a variety of physiological stimuli. It is clear that many hormones, neurotransmitter substances, and other extracellular stimuli mediate their physiological actions by altering either directly, or indirectly through regulatory proteins, the phosphorylation and dephosphorylation of many intracellular proteins. In general, phosphorylation of specific residues on target substrates triggers small conformational changes in protein structure which alter biological properties. Processes as diverse as membrane transport and permeability, metabolism, ionic fluxes, contractility, and the transcription and translation of genes, are all regulated by this versatile post-translational mechanism [1].

Protein phosphorylation is intimately involved in the regulation of myocardial contraction and metabolism [2, 3]. A number of the proteins phosphorylated in the heart have now been identified, but in many cases the precise mechanisms by which phosphorylation modulates their behavior is only beginning to be elucidated. Historically, the cAMP-dependent processes in heart have received the most attention as modulators of cardiac protein phosphorylation [4–7]. More recently however, investigation has expanded to include other protein kinases such as Ca²⁺-calmodulin protein kinase (Ca²⁺/CAM-PK) [8, 9], protein kinase C (PKC) [10–12], cGMP-dependent protein kinase (PKG) [13, 14], tyrosine protein kinases (PTKs) [15], extracellularly regulated kinases (ERKs) [16], as well as mitogen-activated protein kinase (MAPKs) [15, 17], and the related stress-activated or c-jun N-terminal protein kinases (SAPK/JNKs) [18]. Similarly, the process of dephosphorylation by protein phosphatases has taken on new emphasis as investigators have recognized the integral part these enzymes may play in modulating normal cardiac cellular function [19–21], given that the steady-state level of phosphorylation of any protein is therefore a reflection of the relative activities of protein kinases and phosphatases that mediate the interconversion process.

The objectives of this review are to present a comprehensive survey of the literature on cardiac phosphoproteins, their role in regulating cardiac function, and their potential roles in myocardial pathophysiology. In so doing, some general perspectives on mechanisms of cardiac protein phosphorylation will be presented.

The phosphorylation of cardiac phosphoproteins will be discussed in this review from an experimental viewpoint. There are two general experimental approaches used to study protein phosphorylation of candidate substrates. The first is in vitro phosphorylation of purified proteins or mixtures of proteins (e.g. homogenates) by the exogenous addition of specific kinases, phosphatases, modulating agents and/or cofactors, plus []-ATP. This type of reaction defines proteins that are substrates, provides clues on the specificity of activation, delineates the identity of phosphorylation sites, and may indicate effects of phosphorylation on the functional properties of the protein. The second approach is protein phosphorylation of in situ organs or intact cells preincubated with []-ATP to label cellular ATP pools. The tissue/cells are then exposed to various physiological stimuli or pharmacological to elicit phosphorylation and, it is hoped, clearly measurable biological effects. Information regarding the conditions required for phosphorylation, the extent to which proteins are phosphorylated in response to stimuli and, sometimes more clearly than in in vitro studies, the consequences of phosphorylation, can be gleaned from these ‘ex vivo’ experiments. In vitro or ex vivo, phosphorylated proteins are often analyzed using one- or two-dimensional (2-D) gel electrophoresis [22], the increased labeling of proteins being used to indicate stimuli-mediated phosphorylations.

The investigation of protein phosphorylation is not straightforward. Problems with the in vitro approach include the use of nonphysiological doses of modulating agents and cofactors, as well as an altered availability of substrates due to cell disruption or dilution. Each may lead to protein phosphorylations not seen ex vivo, and several examples of this will be described in the context of cardiac phosphoproteins. Yet another problem may be that the previous phosphorylation state of a putative protein substrate, as it is extracted from the cell, may influence the ability of the enzyme itself to utilize the protein as a substrate. The ex vivo assay suffers from the possible lack of specificity when using certain modulating agents, and this may require separate corroborative evidence. Lastly, it should be noted that phosphorylated proteins cannot be positively identified as direct substrates, as there may be a ‘cross-talk’ or ‘cascade effect' between cellular signal transduction pathways. Technically, the use of one-dimensional gels to analyze protein phosphorylation have been characterized by poor resolution, particularly when working with protein mixtures, or when trying to identify small, novel proteins. On the other hand, 2-D gels have been somewhat problematic in the past because of replication errors and difficulty in analysis, although recent technological advances have largely ameliorated these issues. It is important then that the limitations of each type of experimental approach be recognized when interpreting demonstrated results.

Nonetheless, the valuable information that in vivo or ex vivo experiments can provide is critical, and in conjunction with corroborative in vitro evidence, a very powerful means for determining the properties of putative phosphoproteins. For instance, preparations of isolated cardiac myocytes offer the advantage of performing multiple and temporal measurements on a uniform cell population. With perfused hearts sufficient tissue is available to study various proteins and associated biochemical activities. In both cases useful measurements of mechanical activity are possible, particularly with the advent of sophisticated devices to quantify contractile parameters in isolated cells. Thus, the emphasis in this review will be on discussing cardiac phosphoproteins that have been confirmed through not only in vitro data, but more importantly, from ex vivo experiments with corroborative physiological and biochemical evidence.

	2 Cardiac phosphoproteins

Top Abstract 1 Introduction 2 Cardiac phosphoproteins 3 Coordinated regulation of... 4 Pathophysiological alterations... 5 Summary and perspectives References

 
2.1 Sarcolemma
The plasma membrane or sarcolemma is a semipermeable membrane critically involved in the regulation of the myocardial contraction/relaxation process and general cell homeostasis. Most important are the mechanisms by which intracellular Ca²⁺ concentrations i.e. [Ca²⁺]_i in the cardiac myocyte are controlled. As a result the sarcolemma contains several key activities for translocating Ca²⁺ into and out of the cell. These include the voltage-dependent, slow Ca²⁺ channel for inward Ca²⁺ conductance, Na⁺/Ca²⁺-exchanger and Ca²⁺-ATPase pump for extruding Ca²⁺. In addition, there are a number of other proteins crucial to myocyte function including membrane-bound protein receptors for neurohormonal factors, enzymes such as adenyl cyclase and the Na⁺,K⁺-ATPase pump, numerous ion channel types like the Na⁺ and K⁺ channels, and a variety of other proteins which may have regulatory roles associated with the aforementioned major components. A few of the sarcolemmal proteins listed have now been identified as phosphoproteins and their properties as such are described below.

2.1.1 Phospholemman
Phospholemman (PL) is a 72-amino-acid plasma membrane protein of apparent 15 kDa molecular mass as analyzed by SDS-PAGE [23, 24]. It is a highly basic protein and possesses a single membrane-spanning domain consisting entirely of uncharged residues [24]. The amino-terminus end is oriented extracellularly, whereas the positively charged C-terminus end, which contains protein kinase phosphorylation sites, projects into the cytoplasm. PL shows many similarities to phospholamban, the main regulatory protein of the sarcoplasmic reticulum (see section below), including sharing seven out of nine sequence residues in the critical phosphorylation sites area. In the transmembrane region, PL exhibits 52% amino acid homology to the -subunit of Na⁺,K⁺-ATPase [24].

Previous in vitro studies have shown that PL can be phosphorylated in cardiac sarcolemma at multiple sites by the catalytic subunit of the cAMP-dependent protein kinase (PKA) [10, 25], as well as being the major plasma membrane substrate for PKC [10, 23, 25]. In beating hearts this protein has been demonstrated to be phosphorylated upon perfusion with various inotropic agents. Stimulation of β₁-adrenergic receptors in intact guinea pig myocardium with either isoproterenol [26]or denopamine [27]resulted in a rapid onset and a 2–3 fold increase in []P_i incorporation that was correlated to an increase in the maximal rate of developed tension following drug treatment. Dephosphorylation of PL has been demonstrated in intact guinea pig hearts, but it occurs very slowly in contrast to other cardiac phosphoproteins suggesting differential regulation of its activity [28]. Adenosine agonist treatment of intact guinea pig hearts was also found to reduce the isoproterenol-stimulated phosphorylation state of PL [29].

Activation of -adrenergic receptors has been associated with an increased phosphorylation state of PL, although some disparities in the literature do exist. Several groups have found that -adrenergic stimulation in isolated rat [25, 30]and rabbit hearts [11]results in an increased phosphorylation of PL, and in at least the latter study, was shown to be mediated by PKC activation. The role of PL as a key plasma membrane substrate for PKC was also supported by Hartmann and Schrader who showed that direct treatment with the PKC activator TPA could stimulate phosphorylation of PL in intact rat cardiac myocytes [30]. Based on this premise, Kranias and colleagues [10, 31]further examined the role of -adrenergic and direct PKC activation in effecting PL phosphorylation of isolated guinea pig hearts, but were unable to observe any detectable changes in []P_i incorporation. The discrepancy between this study and others noted above is unresolved, though it may simply be due to species differences in signal transduction processes, or perhaps to the complexities of protein kinase/phosphatase interaction(s) in the intact cell.

The functional role of PL remains unknown, however it is possible that alteration of membrane surface charge secondary to phosphorylation may play a role in its function and may result in effects on activities of various channels, pumps and/or antiporters [24]. Indeed, expression of PL in Xenopus oocytes leads to the occurrence of a unique chloride current [32], suggesting that the protein may be itself be an ion channel. Other investigators have instead proposed that PL may be a prototypic member of a new family of membrane proteins capable of regulating ion channel activity [33]. In this regard, it has also been suggested that this 15 kDa sarcolemmal protein may modulate increases in the cardiac slow inward Ca²⁺ current [33].

As a major phosphoprotein localized in the sarcolemma and by virtue of its activation by - and β-adrenergic processes, PL is most probably involved in positive inotropic effects by as yet undetermined molecular mechanisms. The discrepant findings that this protein may also mediate negative inotropic actions via PKC-stimulated increases in phosphorylation are puzzling, but may suggest differential site phosphorylation resulting in opposing physiological responses.

2.1.2 Na⁺/Ca²⁺ exchanger
Very few reports have addressed the role of protein phosphorylation in the regulation of the cardiac Na⁺/Ca²⁺ exchanger (NCX1) thereby effecting extrusion of Ca²⁺ from myocytes [34, 35]. When bovine NCX1 was stably overexpressed in COS or CHO cells no phosphorylation was detected [34]. In contrast, more recent data has clearly demonstrated significant basal phosphorylation of canine NCX1 transfected into both CCL39 cells and rat cardiomyocytes, that was further enhanced by treatment with endothelin-1 (ET-1), acidic fibroblast growth factor (aFGF), the phorbol ester, PMA, or the phosphatase inhibitor, okadaic acid [35]. Additionally it was observed that the PKC inhibitors, calphostin C and K252a, or EGTA, inhibited phosphorylation. All treatments that increased NCX1 phosphorylation also significantly increased both forward and reverse modes of Na⁺/Ca²⁺ exchange. It appears then that the cardiac NCX1 could play an integral role in some of the reported negative inotropic actions of PKC-activating agents.

2.1.3 Ion channels
Considerable attention has been given to phosphorylation of cardiac ion channels as a means whereby the activity of ion channels can be regulated, and in most cases result in the alteration of the myocardial contractile state. Much of this evidence is based on electrophysiological studies involving whole-cell patch clamp techniques under phosphorylatable conditions. It has been assumed in many cases that the regulation is mediated by phosphorylation(s) of the channel protein directly or of an associated regulatory protein. Several excellent reviews have summarized in detail the current level of understanding of ion channel function, particularly that of slow calcium channels, and the role of phosphorylation in modulating ion fluxes [36, 37]. Only a brief mention of the pertinent information is made here.

2.1.3.1 Calcium channels
Cardiac Ca²⁺ channels have received most of the scrutiny relating to effects of phosphorylation on channel function, and thereby the contractile state of the heart [38]. This is because the Ca²⁺ channels are the key determinant of intracellular [Ca²⁺]_i levels, which in turn directly effect force of contraction, as well as acting as a cellular second messenger for other processes, including Ca²⁺-dependent protein kinases. A model by Sperelakis et al. [37]postulated that cAMP elevation either by receptor-mediated stimulation or indirectly through phosphodiesterase inhibition, activates PKA which phosphorylates either the slow Ca²⁺ channel or a contiguous regulatory protein, such that a greater number of channels are available for voltage activation. Two mechanisms have been proposed to account for the enhanced channel activation following phosphorylation, the first being a conformational change in the channel protein allowing the activation gate to be opened, or secondly, by directly increasing the effective channel pore diameter. In this model, phosphorylation markedly increases the probability of channel opening during depolarization. A number of studies utilizing direct intracellular injections of various agonist agents, including the catalytic subunit of PKA [39–41], and protein inhibitor of PKA [42], are consistent with the phosphorylation model of Ca²⁺ channel activation and stimulated I_Ca. Furthermore, channel activity can be reversed upon withdrawal of PKA, suggesting that regulatory components of the slow Ca²⁺ channels are either washed away or lose their affinities over time, most likely through phosphatase action [43]. Thus, any agent that increases the cellular cAMP level of the myocardial cell will tend to potentiate I_Ca, [Ca²⁺]_i, and contraction. Studies in intact canine myocardium have demonstrated that isoproterenol and norepinephrine treatment induced substantial []P_i incorporation into the β-subunit of cardiac L-type calcium channels using back-phosphorylation techniques, and that this correlated well with positive inotropic and chronotropic responses and tissue levels of cAMP [44, 45]. Phosphorylation of the Ca²⁺ channel ₁ subunit via PKA-mediated mechanisms has also been demonstrated, both in liposomes reconstituted with the Ca²⁺ channel [46]and in CHO cells overexpressing the ₁ subunit protein [47], which resulted in enhanced Ca²⁺ efflux.

Other protein kinase systems in addition to PKA also appear to be involved in the regulation of cardiac slow Ca²⁺ channels. Direct PKC activation with phorbol esters and indirect PKC activation by angiotensin II (AII) treatment have been shown to stimulate I_Ca in chick and rat hearts [48, 49], but not in guinea pig hearts [50]. Inhibitors of calmodulin (CAM) can inhibit slow Ca²⁺ channel activity, and this effect can be reversed by subsequent microinjection of CAM [51]. Thus, it seems that slow Ca²⁺ channel activation can be achieved by at least three apparent phosphorylation mechanisms namely, PKA, Ca²⁺/CAM-PK, as well as PKC. It remains unclear if all these phosphorylations occur on the same protein or on separate proteins, or even if similar phosphorylation sites are involved.

The myocardial slow Ca²⁺ channels are also regulated by cGMP, by use of the non-hydrolyzable cGMP analog and potent PKG stimulator, 8-Br-cGMP, in a manner that is antagonistic to that of cAMP. This has been demonstrated at both the whole-cell voltage clamp and single channel level [52, 53]. Direct introduction of PKG into neonatal myocytes is apparently associated with a rapid inhibition of I_Ca [37]. A single protein with approximately 47 kDa mass has been demonstrated to be specifically phosphorylated by PKG in guinea pig sarcolemmal membranes [54], suggesting the existence of a putative protein mediator involved in Ca²⁺ channel regulation by a cGMP pathway. There has however, been no confirmation of a cGMP-dependent sarcolemmal phosphoprotein at an in vivo level.

A brief mention should also be made that protein phosphatases may also have a direct role in the signal transduction cascade of Ca²⁺ channel activation. Treatment with the catalytic subunit of protein phosphatases inhibited PKA-mediated activation of Ca²⁺ channels [19, 55], whereas exposure to phosphatase inhibitors such as okadaic acid and microcystin the phosphatase-dependent dephosphorylation [56, 57]. Furthermore, in a recent study Herzig et al. demonstrated that stimulation of protein phosphatase type 2A activity abolished muscarinic-receptor-mediated inhibition of the β-adrenergic stimulating effects on Ca²⁺ channel activity [58]. This evidence supports the concept that protein phosphatase stimulation can take part in the functional antagonism between adrenergic and cholinergic stimuli in the intact myocardium [20].

2.1.3.2 Potassium (K⁺) channels
A second important class of ion channels involved in the heart is the K⁺ channels. While less is generally known about the specific mechanism(s) by which these channels are regulated, sufficient electrophysiological evidence now exists to at least partially implicate phosphorylation as a key modulatory step. Injection of Xenopus oocytes, co-expressing a cloned cardiac delayed rectifier K⁺ channel (RAK) and the human β-adrenergic receptor (βAR), with isoproterenol caused a significant increase in the I_RAK current [59], comparable to that previously seen by adrenergic stimulation of K⁺ currents in isolated frog myocytes [60]. In separate experiments, similar findings were obtained in oocytes expressing only the cloned K⁺ channel and treated with PKA catalytic subunit [59]. Koumi et al. demonstrated that the inwardly-rectifying K⁺ channel (I_Kl) in human ventricular myocytes could be inhibited by PKA-mediated phosphorylation, and that this response was antagonized by either a PKA inhibitor or phosphatase treatment [61]. Since I_Kl inhibition led to observed increases in action potential duration and depolarization of the resting membrane potential, it is presumed that these were a result of phosphorylation.

A third type of K⁺ channel, the ATP-sensitive K⁺ (K_ATP) channel, has recently received increasing attention because these channels may regulate cardiac function during cellular injury [62]. Little direct evidence exists demonstrating the role of protein phosphorylation in cardiac K_ATP channel activity [63]. Phosphorylation of K_ATP channels is presumed to occur based on the strict requirement for a MgATP-dependent process to maintain these channels operative [62]. It has been proposed that the K_ATP channel possesses two phosphorylation sites which are differentially regulated by MgATP, and that their different phosphorylation states may describe the various activities displayed by K_ATP channels, including resting state (closed channel), spontaneous activity, rundown (dephosphorylation of the channel), and reactivation (rephosphorylation of the channel) [64]. All this may have further implications on K_ATP channel activity with regards to how cellular phosphorylation-dephosphorylation ratios may be effected by various endogenous neurohormonal and metabolic influences that would be expected in altered inotropic states.

It is unclear presently which specific phosphorylation sites are involved in the activation or inhibition of K_ATP channel activity. Recent evidence by Kwak et al. showed that K_ATP channel activity in rat cardiac myocytes is reciprocally modulated by phosphorylation of both Tyr and Ser/Thr residues [65].

The most direct electrophysiological evidence for K_ATP channel phosphorylation to date are the observations that treatment of ventricular myocytes with various phosphorylating and dephosphorylating agents resulted in modulation of K_ATP channel activity [65–67]. Kwak et al. have shown that the K_ATP channel run-down process was suppressed by the protein phosphatase inhibitor, okadaic acid, and accelerated by the protein tyrosine phosphatase inhibitor, sodium orthovanadate [65]. Following run-down, the ATP-induced reactivation was enhanced by genistein, a tyrosine kinase inhibitor. Furthermore, this study provided the first direct evidence for a regulatory link between K_ATP channel activity and specific phosphorylation sites as activity was reduced by protein phosphatase 2A (PP2A) and increased by tyrosine phosphatase 1B. In a separate study, PKC phosphorylation resulted in inhibition of ventricular K_ATP channel activity at low cellular ATP levels and alteration in the stoichiometry of ATP binding to the channel [66]. However, at physiological ATP levels, PKC upregulated K_ATP channel activity and the reversal of this effect was dependent on the activity of a membrane-associated PP2A [67]. Thus, it appears that the extent of channel phosphorylation and therefore cardiac K_ATP channel activity may be dependent on reciprocal modulation of specific sites via different signaling mechanisms.

2.1.3.3 Other channels
A brief mention should be made regarding other cardiac ion channel activities that have been reported to be regulated by phosphorylation-dephosphorylation mechanisms, including chloride (I_Cl) currents [57, 68–72], which are critical for action potential repolarization, as well as connexin43 and connexin45 gap junction channel (I_J) conductances [73–75], that are essential for cell-to-cell communication. For both channel types, conductances were enhanced following activation of PKA- [57, 68–71, 74, 76]and PKC-dependent processes [70, 71, 74]. Complete deactivation of channel conductances resulted from phosphatase action [57, 74], and treatment with phosphatase inhibitors enhanced currents and reversed their deactivation [57, 70, 74]. Confirmation that connex43 is a phosphoprotein was obtained by Laird and coworkers who showed that the protein is typically present in neonatal rat cardiac myocytes as a 42 kDa band and constantly incorporates []P_i [77]. A second phosphorylated form, connexin45, was also observed by []P_i labeling at 44 kDa. Phosphatase treatment of cell lysates eliminated the 42 kDa phosphoprotein band, revealing a non-phosphorylated 40 kDa form. It remains unclear as to whether in intact myocytes connexin43 is phosphorylated by different kinases or by a single kinase at several sites, although the primary sequence data indicates consensus sites for PKC and possibly Ca²⁺/CAM-PK [78]. Laird et al. postulated that multiple phosphorylation sites and forms of connexin43 may control various aspects of gap junction metabolism/function, thereby facilitating cell-to-cell communications [77].

2.1.4 Membrane receptors
Past studies have proposed protein phosphorylation as a general mechanism in the regulation of receptor function. The mechanisms involved in receptor phosphorylation are most likely diverse and still poorly understood [79–81]. In some cases receptors that signal the formation of second messengers are themselves regulated by protein kinases activated by the second messenger [79]. In other instances, receptors are regulated by soluble receptor-‘specific’ protein kinases or so-called G-protein-coupled receptor kinases (GRKs) [82], as for example, the β-adrenergic receptor kinase or βARK [81, 83, 84]. The consequences of receptor phosphorylation include desensitization, i.e. the diminished responsiveness of receptors to agonists, as well as receptor internalization and activation [85, 86]. The sites involved in GRK phosphorylation have not been unambiguously identified, though it appears that differential activation can occur subsequent to adrenergic- or PKC-induced phosphorylation by βARK1 on specific serines of both the β- and -adrenergic receptors [87–90]. The apparent role of GRKs in myocardial function based on indirect studies has recently been reviewed [91].

The muscarinic-cholinergic receptors (mAChR) have been the only receptor family directly studied at the intact cardiomyocyte or tissue level in terms of their possible functional regulation by phosphorylation. Hosey and colleagues have demonstrated that the 79 kDa mAChR protein is phosphorylated in an agonist-dependent manner in intact chick and porcine tissue [92–94]. In both cases, stimulation of []P_i-labeled tissue with muscarinic agonists led to increased phosphorylation of the mAChRs, and specifically occurred on Ser and Thr residues [93, 94]. Activation of either PKC or PKA had no effect on receptor phosphorylation or agonist affinity, nor did the treatment of CAM antagonists. Furthermore, agonist-dependent phosphorylation of cardiac mAChRs appeared to correlate with a decreased agonist affinity and ability to produce a negative inotropic response. Taken together, the studies by Hosey and coworkers support the idea that cardiac mAChRs require agonist occupancy of the receptor and may involve the participation of a receptor-specific protein kinase. In addition it appears that phosphorylation of mAChRs in the heart may be a critical step leading to their desensitization.

In this regard, the mAChRs possess features that are strikingly similar to that observed for several other members of the G-protein coupled superfamily of receptors, most notably the β-adrenergic receptors [85]. Studies using purified mAChRs from chick [95]and porcine heart [96]have been found to be excellent substrates in vitro for βARK. Further work is required however, to confirm that βARK or a related receptor kinase directly phosphorylates mAChRs in vivo.

2.2 Sarcoplasmic reticulum
Cardiac sarcoplasmic reticulum (CSR) consists of a complex network of anastemosing membrane-limited intracellular tubules which surround the myofilaments as a network. The CSR, whose main function is the regulation of cytosolic Ca²⁺ or [Ca²⁺]_i, is the most important system in the cardiac cell that delivers activator Ca²⁺ needed during contraction for binding to the myofilaments. From both a structural and a functional standpoint this membrane is divided into two general regions. These are (a) the subsarcolemmal cisternae or junctional SR, which refers to that portion of the CSR that comes into close apposition to the sarcolemma and transverse tubules, and that contains the Ca²⁺ release channels through which Ca²⁺ flows to initiate contraction, and (b) the much more extensive sarcotubular network or longitudinal (free) SR that contains the Ca²⁺-ATPase pump/phospholamban protein complexes which regulate active Ca²⁺ transport into the CSR lumen, and that forms the tubular network around the A and I bands of myofilaments.

The CSR membrane contains a number of intrinsic proteins that are key regulators in cardiac excitation-contraction coupling, and specifically modulate [Ca²⁺]_i in determining rates of myofilament contraction and relaxation. These proteins are the Ca²⁺-ATPase pump (SERCA, predominantly SERCA2 in the myocyte) [97], phospholamban (PLB) [98], Ca²⁺ release channel (CCRC) or ryanodine receptor [99, 100], and several Ca-binding proteins that include calsequestrin (CSQ) [101], calreticulin, and the 26 and 170 kDa Ca-binding proteins. Of this group, the only phosphoproteins identified to date are PLB, the CCRC and CSQ.

2.2.1 Phospholamban
PLB is the key CSR phosphoprotein involved in the regulation of the SERCA2 pump, and hence Ca²⁺ transport [102]. PLB is currently viewed as a functional inhibitor of the SERCA2 when it is in an unphosphorylated state. Once phosphorylated, the inhibition of pump activity is removed, and the process of active Ca²⁺ transport into the CSR lumen is allowed to occur. Indeed, recently it has been demonstrated by Kranias and colleagues that after ablation of the PLB gene, the Ca²⁺-uptake rate into CSR was enhanced in the PLB-deficient hearts compared with the wild-type mice hearts, with an ensuing elevation in basal contraction [103]. Despite a considerable amount of investigation on this protein in recent years, its basic functional unit and structure in the CSR has not yet been clearly defined, nor have the molecular mechanisms fully defined its interactions with the SERCA2 pump [104]. Nonetheless, recent advances in our understanding of the SERCA2 protein structure have allowed for the development of a fairly precise mechanistic model of Ca²⁺ transport with phosphorylation playing a key role [105].

The primary structure of PLB has been deduced from sequence data obtained from cDNA clones [106]. It is generally assumed that the quaternary structure of PLB corresponds to a pentameric structure with an apparent M_r of 25 000 based on SDS-PAGE [107]. The oligomeric structure has been confirmed by the observation that dissociation of PLB into five lower M_r species can occur following boiling in SDS [108]. Each subunit of PLB contains 52 amino acids, with a calculated M_r of 6080 [106], and is made up of two major structural domains [106, 109]. The highly hydrophobic C-terminal domain most likely inserts itself into the CSR membrane. The N-terminal sequence of PLB is comprised of an amphiphilic -helical cytoplasmic domain which contains the target sequences for kinase phosphorylation sites. The cytoplasmic domain also possesses the domain(s) that interact(s) with the SERCA2 pump protein. Direct interaction between these two molecules has been demonstrated by cross-linking studies [110]. Recent data by Cornea et al. have revealed that the oligomeric states of PLB in an artificial lipid bilayer environment are typically found in a dynamic equilibrium which is perturbed following in vitro phosphorylation by PKA catalytic subunit [111].

A number of studies involving a variety of techniques have led to a proposed structural model of PLB whereby the protein exists in its pentameric structure in such a manner as to form a hydrophobic Ca²⁺-selective ion channel pore [112–116]. It is unclear at this time how the putative PLB channel pore is integrated functionally with the interaction between PLB and the SERCA2 pump.

The concept that phosphorylation of PLB allows for the removal of the molecule's inhibition upon SERCA2 activity is supported by several lines of evidence. Proteolysis studies by Kirchberger et al., demonstrated a correlation between tryptic digestion of unphosphorylated PLB and an activation of Ca²⁺ transport by the SERCA2 pump [117]. Upon phosphorylation, digestion of PLB by proteolytic enzymes is greatly reduced. Huggins and England suggested that the hydrophobic domain of PLB may in fact undergo a conformational change [118, 119]. This was further supported by Simmerman et al. when discussing their two-domain model of PLB, where each contains a stable -helix, and upon phosphorylation, these helices are believed to rotate relative to each other so as to remove its inhibition on the SERCA2 [120]. Studies by Wang and coworkers have implicated residues 7–16 of the N-terminal region of PLB as those essential in the direct regulation of the SERCA2 [121]. While the exact molecular mechanisms defining the interactions of PLB and SERCA2 are not currently well understood, it appears certain that the interaction between them is completely dependent on the phosphorylation state of PLB [122]. Recent data would suggest that PLB phosphorylation is associated with enhanced interactions between individual SERCA2 polypeptide chains due to spatial rearrangement and protein-protein interactions [123].

Phosphorylation of PLB in vitro by PKA [124, 125], Ca²⁺/CAM-PK [126, 127], PKC [128, 129], or PKG [13, 130], has been shown to markedly increase CSR Ca²⁺ uptake. The substrate site for phosphorylation of PLB by PKA has been demonstrated to be Ser¹⁶, whereas Thr¹⁷ is phosphorylated by the Ca²⁺/CAM-PK [114]. No definitive phosphorylation sites on PLB have been identified for PKC, although it did not appear to occur at either Ser¹⁶ or Thr¹⁷ [128]. In the case of PKG, the Ser¹⁶ residue was suggested as the target substrate site, since in the presence of PKA stimulatory effects of both kinases on the SERCA2 activity and Ca²⁺ transport were not additive [130]. The characteristics of PLB phosphorylation appear to be based on changes in structural conformation, most likely at the Ser¹⁶ site [131, 132]. Detailed studies have revealed that PLB phosphorylation of Ser¹⁶ by PKA proceeds via a random mechanism, while that of Thr¹⁷ by Ca²⁺/CAM-PK proceeds via a cooperative mechanism [133]. The latter process is apparently unaffected by the phosphorylation status of Ser¹⁶, and Ca²⁺ accumulation was stimulated in proportion with the stoichiometry of PLB phosphorylation, regardless of the site of phosphorylation.

Phosphorylation of CSR by endogenous or exogenous PKA was first shown to be associated with PLB by Kirchberger et al. [134]. Subsequent studies have confirmed that PKA-dependent phosphorylation of PLB results in the stimulation of the initial rate of Ca²⁺ transport by lowering the half-maximal concentration of Ca²⁺ needed for stimulation of SERCA2 activity [125, 135, 136]. These findings strongly suggested that the apparent affinity of the SERCA2 pump for Ca²⁺ was increased during PLB phosphorylation.

Two possible mechanisms have been proposed to explain the observed stimulation of Ca²⁺ uptake by PKA dependent phosphorylation, namely an enhanced SERCA2 turnover rate, or an increased efficiency of the SERCA2 for Ca²⁺ (i.e. an increased coupling ratio) [137, 138]. Studies of SERCA2 activity have demonstrated that PKA-dependent phosphorylation of PLB correlated well with stimulation of the SERCA2 while maintaining a stoichiometric ratio of 2:1 for mole of Ca²⁺ uptake/mole ATP hydrolyzed [139]. The enhanced turnover rate of SERCA2 activity by PKA stimulation can be ascribed to changes at two major reaction steps of the ATP hydrolysis-Ca²⁺ transport model. Rapid kinetic experiments demonstrated that PKA phosphorylation of PLB produced a marked increase in the reaction associated with Ca²⁺ binding to the ATPase enzyme, and in the rate at which the phosphorylated intermediate (EP) was subsequently formed [137, 138]. Stimulation of EP formation was found to be associated with a decrease in the dissociated constant for Ca²⁺ binding (i.e. an increase in the affinity of the SERCA2 pump for Ca²⁺). The observed alterations in Ca²⁺ affinity are probably due to an increase in the rate of the slow conformational change of the SERCA2 enzyme upon Ca²⁺ binding [140]. Whether the conformational change of the SERCA2 and its subsequent alteration of intrinsic rate constants are due to a direct steric interaction of PLB with the enzyme has not be completely elucidated. Recent kinetic studies have revealed that PKA-mediated phosphorylation of PLB are correlated to an accelerated rate of decomposition of the phosphorylated SERCA2 intermediate (E₂P) which contributes to the increase in V_max(Ca) [141]. These observations may have some implications for the possible stimulatory role of PKA-dependent phosphorylation on CSR function in vivo, as will be discussed later, since it is a slow and rate-limiting reaction step in the SERCA2 reaction scheme.

PLB phosphorylation by PKA in vitro has also been shown to stimulate Ca²⁺ efflux from the CSR [142, 143]. A reduction in the amount of Ca²⁺ needed to attain half-maximal activation of Ca²⁺ efflux was also observed. These findings suggest that a component of the SERCA2 pump participates in the release of Ca²⁺ from CSR, but no definitive evidence is available to describe the possible molecular mechanisms for this process.

Phosphorylation of PLB in situ, either in isolated perfused hearts or in isolated cell culture systems, has now been established in different species using the paradigm of β-adrenergic stimulation, in conjunction with measurement of associated functional and biochemical parameters. In this regard, several groups have made major contributions over the last two decades to our understanding of PKA-mediated stimulation of cardiac regulatory phosphoprotein activity, including Kranias/Solaro et al. [5, 7, 28, 144–148], Watanabe and coworkers [6, 8, 149, 150], England and colleagues [151–153], and more recently, Neumann et al. [29, 154]. In most studies, β-adrenergic stimulation was found to enhance Ca²⁺ transport activities, and observed PLB phosphorylation appeared to parallel the temporal change in myocardial relaxation [6, 147, 155]. Other PKA-stimulating agents such as phosphodiesterase inhibitors [5, 156], adenyl cyclase activators [151, 156], β₂-adrenergic receptor ligands [157, 158], and Ca²⁺-sensitizing agents [159], have also been studied for their effects on enhancing PLB phosphorylation. However, there appears to occur a compartmentalization of cAMP or PKA, or both, in heart that is differentially coupled (in some cases not at all) to PLB phosphorylation and/or contraction, based on the use of PKA-dependent agents [5, 348].

In PLB-ablated transgenic mouse hearts, baseline or isoproterenol stimulation resulted in similar levels (albeit attenuated in the case of isoproterenol treated animals) in tissue cAMP levels and the degree of phosphorylation of other cardiac phosphoproteins when compared to wild-type hearts [148]. These data that has emerged confirms the concept of PLB acting as a key regulator of myocardial relaxation during sympathetic or catecholamine stimulation.

Parasympathetic control through muscarinic/cholinergic-mediated processes have been studied as a corollary to β-adrenergic stimulatory effects on PLB. Several groups have provided evidence suggesting that muscarinic/cholinergic action can reverse β-adrenergic effects of PLB phosphorylation and function, without much if any alteration to elevated cAMP levels [21, 150, 156, 160–163]. The effect of adenosine agonists has also been examined on β-adrenergic-stimulated PLB phosphorylation in intact cardiac myocytes [21, 29, 162, 164–166]. The resulting data showed that elevated PLB phosphorylation levels were attenuated following adenosine receptor stimulation, which supports the idea that adenosine receptor-mediated events may share similar signal transduction processes to that of muscarinic/cholinergic pathways in reversing β-adrenergic responses. Similarly, treatment of cardiac myocytes or isolated hearts with ₁-adrenoceptor stimulation was found to inhibit β-adrenergic agonist-induced increases in protein phosphorylation and myocardial contractility [167]. It was speculated that this modulation may occur via the PKC-mediated pathway, but no data were provided to substantiate this.

As in the case with PKA, Ca²⁺/CAM-PK-dependent PLB phosphorylation in vitro has been shown to enhance the initial rates of CSR Ca²⁺ uptake [126, 127, 137, 168]. This stimulatory effect was most pronounced at low (<1 µM) Ca²⁺ levels and appears to be due to an increase in the apparent affinity of the SERCA2 for Ca²⁺ [127, 168]Ca²⁺/CAM-PK-dependent phosphorylation of PLB requires the absolute presence of free Ca²⁺ over a concentration range of 10^–7 to 10^–5 M [137], in addition to exogenous CAM (EC₅₀=50 nM) [127]. Kranias and coworkers have demonstrated that calmodulin can activate an endogenously bound Ca²⁺/CAM-PK which phosphorylates PLB, and in turn stimulates the initial rates and maximal levels of the phosphorylated intermediate, EP, of the SERCA2 [168]. In this regard, the Ca²⁺/CAM-PK dependent system is similar to the PKA dependent process in its kinetic effects on the SERCA2 reaction sequence. This was confirmed by several groups but it is not clear whether the cytoplasmic or CSR luminal Ca²⁺ pools are involved in the Ca²⁺/CAM-dependent stimulation [169–171]. Studies by Karczewski et al. have shown recently that Ca²⁺/CAM-PK phosphorylated PLB on Thr¹⁷ exclusively, though this was predicated on the use of a synthetic PLB peptide substrate [172]. The Ser³⁸ residue on SERCA2 has been identified as the specific site for phosphorylation by Ca²⁺/CAM-PK [169].

Studies in intact hearts indicate that direct alterations in [Ca²⁺]_i levels do not by themselves stimulate PLB phosphorylation, even though Ca²⁺/CAM levels are raised [8]. As a result neither CSR Ca²⁺ transport or SERCA2 activities were altered, nor was the observed rate of myocardial relaxation. However, PLB phosphorylation was attenuated under conditions where activation of Ca²⁺/CAM-dependent processes was inhibited [8, 9, 173, 174]. Thus, it is possible that Ca²⁺/CAM-dependent mechanisms may be partially mediating the β-adrenergic cardiac relaxant effect. More recent data has given new insights into the mechanisms underlying Ca²⁺/CAM-dependent PLB phosphorylation in the intact heart [175]. Under maximal β-adrenergic stimulation activation of Ca²⁺/CAM-PK accounted for approximately 50% PLB phosphorylation exclusively at the Thr¹⁷ site and was closely associated with increased myocardial relaxation. A recent study by Baltas et al. demonstrated that CSR Ca²⁺/CAM-PK is activated in response to β-adrenergic stimulation, prompting autophosphorylation of its regulatory domain and conversion to an active Ca²⁺-independent species [349]. It was suggested that this could form the basis for potentiation of Ca²⁺ transients in the heart.

Since CSR function is regulated in vitro by both PKA and Ca²⁺/CAM-PK pathways, a clear understanding of the possible contribution each assumes becomes critical in determining the interrelationship of these two regulatory mechanisms. It has been demonstrated that PKA stimulation of Ca²⁺ uptake can occur independently of Ca²⁺/CAM-PK dependent phosphorylation [125, 176], with the converse being true as well [127]. When both protein kinase systems are operating, they appear to have an additive effect [125]. These findings suggested that at least in vitro, the presence of dual control systems operating in an intricate manner to regulate CSR function.

Evidence from intact myocardium studies have suggested that a coordinated dual control system involving PKA and Ca²⁺/CAM-PK may not necessarily be the case. This conclusion is supported by Wegener et al. who demonstrated sequential PLB phosphorylation, meaning that Ca²⁺/CAM-PK phosphorylation occurred only after PLB had been phosphorylated by PKA-dependent mechanisms. [177]Thus, multi-site and sequential phosphorylation of PLB by two different protein kinases in response to β-adrenergic stimulation in beating hearts illustrates the complex and integrated regulation that occurs in the modulation of CSR Ca²⁺ fluxes, and of the myocardial relaxation process.

Mechanistic information on the in vitro stimulatory effects of PKG on PLB phosphorylation are limited. Based on kinetic characterization, PLB seems to be an excellent substrate for PKG with V_max values approximately four times than that seen for PKA [13]. Furthermore, the kinetics of the phosphorylation appear to be cooperative. At the intact heart or myocyte level the biochemical mechanisms of how cGMP modulates cardiac contraction/relaxation is not completely understood. This is due to disparity in reported data on effects of PKG on PLB phosphorylation. In one study, Huggins et al. used cGMP analogs but were unable to observe enhanced PLB phosphorylation [13], whereas Sabine et al. have recently demonstrated that PKG-dependent agents increased PLB phosphorylation with an accompanying rise in cellular cGMP levels [14]. Currently, there is no explanation for the disparity in results from the two studies.

Studies in vitro have demonstrated that PKC can also phosphorylate PLB [128, 178, 179], in both junctional and free CSR.[179]This phosphorylation increases the CSR SERCA2 activity and thus Ca²⁺ uptake [128]. However, PKC-mediated phosphorylation of PLB appears not to occur in vivo. Indeed, both Talosi and Kranias [11], and Hartmann and Schrader [30]did not observe any PLB phosphorylation following perfusion of guinea pig hearts or incubation of isolated rat myocytes with the phorbol ester, PMA.

In order for the phosphorylation of PLB to play a physiological role in the regulation of CSR function, and thereby myocardial contraction/relaxation, some mechanism(s) must exist to dephosphorylate the protein and return it to its role as a functional inhibitor or the SERCA2. Such a mechanism is fulfilled by protein phosphatases which hydrolyze the phosphoester bonds formed by protein kinases. As a result, protein phosphatases are generally viewed as being intricately involved in regulating a variety of signal transduction pathways and cellular proteins [180].

Studies in the cardiac system have shown that endogenous protein phosphatase activity in CSR could dephosphorylate the PKA sites on PLB, and thus cause a decrease in the degree of stimulation of Ca²⁺ transport activity [135, 181]. Subsequent investigations have confirmed these initial observations using CSR membrane-bound protein phosphatase, and were also able to demonstrate that dephosphorylated PLB could be rephosphorylated to full recovery of Ca²⁺ transport activity [182]. Three distinct types of protein phosphatases have now been demonstrated to dephosphorylate PLB to some degree in vitro [183–186]. Kranias and coworkers have been able to partially purify a ‘PLB-specific’ phosphatase that was capable of dephosphorylating both PKA and Ca²⁺/CAM-PK activated phosphorylation sites [183, 184, 186]. A recent in vitro investigation using rat cardiac microsomal preparations has shown that PP1 is capable of dephosphorylating PLB when the latter protein is phosphorylated by PKA, but not by Ca²⁺/CAM-PK, and that under certain conditions PP2B is also able to dephosphorylate PKA-activated PLB [187].

The findings from intact heart studies are in general agreement with in vitro results on the effects of various protein kinases, and in particular PKA-dependent phosphorylation, on CSR function. They support the hypothesis that PLB phosphorylation plays a pivotal role in mediating sympathetic and other neurohormonal inotropic effects on the heart. Furthermore, PLB phosphorylation appears to be only one aspect of a more complex and coordinated regulatory system whereby PKA-mediated effects modulate cardiac function. The full breadth of mechanisms and effects of this regulatory system are only now beginning to be understood. As for the roles of Ca²⁺/CAM-, PKC, or PKG-dependent protein phosphorylation, their contributions to the overall regulation of myocardial contractile and relaxation process remain to be fully elucidated, though some emerging mechanisms are discussed in a later section.

2.2.2 Ca²⁺ release channel/ryanodine receptor
In cardiac muscle, Ca²⁺ release from the CSR is mediated by a Ca²⁺-activated channel called the cardiac Ca²⁺-release channel (CCRC) or ryanodine receptor [99, 100, 188]. The CCRC is regulated by Ca²⁺ influx through voltage-gated Ca²⁺ channels in the sarcolemma. This process, termed Ca²⁺-induced Ca²⁺ release [189], is fundamental to cardiac excitation-contraction coupling, the mechanism that links surface membrane depolarization to Ca²⁺ activation of the contractile apparatus [190].

Cloning and sequence analysis of cDNA have suggested that the CCRC protein is a large polypeptide of approximate 565 kDa mass. The cytoplasmic region of the CCRC appears to correspond to the ‘foot’ structure (the part of the junctional CSR that bridges the gaps between the CSR and surface membrane transverse T-tubules) [191]. The C-terminus region probably forms the CCRC [192].

The cardiac CCRC appears in vitro to be regulated by Ca²⁺ [193, 194], Mg²⁺ [194], adenine nucleotides [195], and several protein kinases, including Ca²⁺/CAM-PK [196, 197]and PKA [197, 198]. It has been proposed that protein kinase-dependent phosphorylation of the CCRC may be physiologically important for the regulation of the protein's activity, and thereby cardiac muscle excitation-contraction coupling. Direct support of in vivo phosphorylation of CCRCs was reported by Yoshida et al., who observed that []P_i incorporation into CCRCs was enhanced upon isoproterenol treatment of neonatal rat cardiac myocytes [199]. Thus, the PKA-dependent phosphorylation of the CCRC may at least partially account for the acceleration of the rising phase of transient [Ca²⁺]_i in β-agonist-treated cardiac myocytes.

2.2.3 Calsequestrin
Calsequestrin (CSQ) is the major Ca²⁺-binding protein of CSR with an apparent mass of 55 kDa based on SDS-PAGE analysis [101, 200]. The protein is mostly localized in the lumen of the junctional SR [201], where it has been shown to bind to some protein constituents, for example the 'foot' protein and the CCRC [202, 203]. Ikemoto et al. have suggested that Ca²⁺-dependent conformational changes in CSQ affect the junctional SR proteins and in turn regulate CCRC function [204]. In general, the primary physiological function of CSQ is thought to be sequestration of large amounts of Ca²⁺ in the lumen of the CSR, reducing luminal levels of free Ca²⁺ and facilitating further Ca²⁺ uptake by the CSR SERCA2 pump, though details of this process have not been fully elucidated [205].

Cardiac CSQ has been proposed to be a phosphoprotein based on the identification of several consensus phosphorylation sites in its primary structure. A unique feature of the protein is a highly acidic, 31 amino acid C-terminal tail (residues 361–391) which contains three closely spaced Ser residues that were proposed to act as excellent substrates for casein kinase II [206]. Moreover, cardiac CSQ was shown to contain endogenous P_i localized to the same cluster of Ser residues identified in the primary sequence [207]. In this latter study, Cala and Jones demonstrated that the same Ser sites of CSQ were rapidly phosphorylated in vitro by casein kinase II. Similar observations were made in cultured rat myotubes, validating that CSQ can be phosphorylated in vivo [208]. There did not appear to be any marked effect of phosphorylation on CSQ Ca²⁺-binding capacity or affinity either in vitro or in intact cells. Thus the functional significance of the phosphorylation event by casein protein kinase II or similar enzyme upon cardiac CSQ is presently unknown.

2.2.4 CSR membrane phospholipids
In addition to the phosphorylation of membrane proteins, there have been several in vitro studies suggesting that phospholipids may also be phosphorylated by both PKA- and Ca²⁺/CAM-dependent kinases in cardiac muscle [209, 210]. This type of phosphorylation has been implicated in the modulation of [Ca²⁺]_i by CSR Ca²⁺ transport activity, probably via diacylglycerol stimulation of PKC which is dependent on phospholipid moieties for its activation [128]. There is in vivo evidence that demonstrated an association between β-adrenergic stimulation of intact guinea pig hearts with phosphorylation of whole cardiac and CSR membrane polyphosphoinositides [146]. Specifically, isoproterenol stimulation increased phosphorylation of phosphatidyl mono- and bi-phosphate, as well as phosphatidic acid. In an associated study from the same laboratory, Edes et al. quantified changes in phosphoinositide turnover under similar inotropic conditions [144]. Their findings revealed increases in phosphoinositide cycle intermediates that were not correlated with either increases in regulatory protein phosphorylation or cAMP levels. A concomitant decrease in inositol triphosphates was observed which was apparently related to lowered in phosphoinositol-phospholipase C enzymatic activity. The underlying mechanisms of phospholipid phosphorylation in situ are not presently known, but it suggests that there may be a complex interrelationship between membrane protein phosphorylation (e.g. PLB) and phosphoinositide phosphorylation, and respective functional correlates.

2.3 Myofibrillar proteins
The contractile properties of the heart are determined by the interaction of three major classes of proteins, namely, contractile proteins (myosin and actin), regulatory proteins (tropomyosin and troponin complex) and structural proteins (C-protein, -actinin, etc.) [211]. A necessary requirement for cardiac contraction is the ability of myosin to hydrolyze ATP, liberating the terminal phosphate as a source of energy for cardiac contraction. The contractile proteins convert the chemical energy of ATP hydrolysis into mechanical work through physicochemical changes. Regulatory proteins bound to actin serve to regulate the actomyosin cycle such that when [Ca²⁺]_i levels are high (100 µM), the regulatory protein complex is inhibited, and ATP hydrolysis is increased resulting in ‘shortening’ or contraction. Conversely, when [Ca²⁺]_i levels are low (0.1 µM), the regulatory protein complex blocks the actinmyosin crossbridge attachment, ATP hydrolysis is reduced, and relaxation ensues. Structural proteins generally do not participate in the active contractile process, but are believed to provide some mechanical linkage and stability properties to the contractile and regulatory proteins.

The three myofibrillar proteins described below represent each of the major classes of proteins mentioned above. More importantly, they all are proteins in which the phosphate groups are in rapid equilibrium with [ATP]_i, and whose phosphorylation states can change quickly following neurohormonal stimulation.

2.3.1 Troponin I
Troponin I (TnI) is the inhibitory subunit of the regulatory troponin complex, along of troponin C (TnC) and troponin T (TnT). Tissue-specific isoforms derived from different gene products have been identified, including the cardiac muscle type [212]. TnI has a molecular weight of about 27 000 daltons and serves as the specific inhibitor of the actomyosin Mg²⁺-ATPase.

There are now several lines of evidence indicating that TnI phosphorylation is a physiologically significant event in the regulation of myocardial contraction. Early in vitro studies by several groups demonstrated that TnI was an excellent substrate for PKA [212–215], with the N-terminal Ser²⁰ as a specific phosphorylation site [216]. Several studies indicated that phophorylation of TnI by PKA decreased the Ca²⁺ sensitivity or activity of actomyosin Mg²⁺-ATPase [145, 217, 218], and rate or affinity of Ca²⁺ binding to TnC [219]. Recent in vitro data from Solaro and colleagues suggest that phosphorylation of TnI at Ser²³ and Ser²⁴ residues in the unique NH₂-terminus domain is both necessary and solely sufficient for the decrease in myofilament Ca²⁺-sensitivity associated with PKA-dependent phosphorylation [220]. The latter observations have been extended by Keane et al. who noted that phosphorylation at these two sites is sequential with Ser²⁴ being rapidly phosphorylated followed by a slower phosphorylation of Ser²³ that occurs only after Ser²⁴ phosphorylation is almost complete [221]. Furthermore, the Arg²² residue appeared to be critical in determining the reaction kinetics of phosphorylation by PKA resulting in conformational changes around the paired Ser region. Evidence for conformational changes in N-terminal extension of TnI has recently been demonstrated whereby phosphorylation caused reductions in the distance between sites located at the N- and C-terminal portion of TnI [222]. In addition, there appears to be a direct transduction of a PKA-induced phosphorylation signal from TnI to the regulatory site of TnC involving a global change in TnI structure [223]. Together these studies provide a molecular basis for the change in Ca²⁺ sensitivity of the troponin complex, most likely through an enhanced off rate for Ca²⁺ exchange with TnC, following its activation by phosphorylation.

Pioneering work on protein phosphorylation in the intact heart actually began with the demonstration that β-adrenergic (catecholamine) stimulation resulted in TnI phosphorylation [4, 224–226], specifically at the Ser²⁰ site [4, 226]. Over the years other groups [7, 145, 156, 227]have confirmed these early observations and have extended them to include other PKA-dependent agonists including forskolin [151, 156], cAMP phosphodiesterase inhibitors [5, 152, 154, 156], and Ca²⁺-sensitizers [159]. An enhanced sensitivity to β-adrenergic stimulation of TnI as compared to PLB was observed in isolated hearts, suggesting possible compartmentation of cAMP [5, 172]. In all cases where it was measured, an increase in TnI phosphorylation was associated with a rise in force of contraction.

The relationship between TnI protein phosphorylation and the temporal features of associated functional activity has been well documented. Increases in TnI phosphorylation parallel the rise in force observed with β-stimulation, but levels of phosphorylation remain high in spite of a return of force to pre-stimulation levels [28, 145]. It may be then, that TnI phosphorylation is not an absolute requirement for the increase in myofibrillar force, but may serve to act in concert with other possible mechanisms. Demonstration that the two Ser residues in TnI capable of being phosphorylated by PKA have differential effects on the decrease in myofilament Ca²⁺-sensitivity, may provide an explanation for the lack of correlation between TnI phosphorylation levels and function during wash-out of β-adrenergic effects [220].

Dephosphorylation of TnI is promoted by the addition of acetylcholine [156, 165, 225]or adenosine agonists [29, 162, 165, 228]to hearts or cells stimulated with isoproterenol, in which case force and TnI phosphorylation fall more or less in parallel. However, in one study no reduction in TnI phosphorylation was noted following adenosine agonist treatment [164]. The reason for these disparate results is unknown. The demonstrated effects of adenosine and cholinergic agents on β-adrenergic stimulation of TnI phosphorylation have been found to be pertussis toxin sensitive and cAMP-independent, since activated cellular cAMP levels were unaltered [162].

Recent experiments by Neumann and coworkers have focused on the role of phosphatase inhibition on stimulating TnI phosphorylation. Several phosphatase inhibitors have been characterized, including okadaic acid, calyculin A, and cantharidin, and all were observed to enhance TnI phosphorylation and the force of contraction [229–231].

For the most part it seems that TnI can also act as superb substrate for PKC, especially in vitro [232–234]. The major site of phosphorylation for PKC appears to be Thr¹⁴⁴, with secondary phosphorylation sites identified at Ser⁴³, Ser⁴⁵ and Thr⁷⁸ [233]. These latter three residues are located in the N-terminal region where most of the binding to TnC occurs [211]. Recent studies have shown that isozymes of PKC exhibited discrete specificities in phosphorylating distinct sites in TnI [235]. For instance, PKC- was uniquely able to phosphorylate Ser²³ and Ser²⁴, previously recognized as being PKA phosphorylation sites only, and thereby reduced myofibrillar Ca²⁺-sensitivity. In addition, PKC-, like PKC- and PKC-, but not PKC-, phosphorylated Ser⁴³ and Ser⁴⁵ and reduced maximal Mg²⁺-ATPase activity.

In intact cardiac myocytes, direct activation of PKC with phorbol esters, or indirectly with ET-1 and arachidonic acid, thought to be mediated by a PKC-dependent pathway, served to increase TnI phosphorylation in a Ca²⁺-independent manner [12, 236, 237]. Venema and Kuo were able to show that the PKC-mediated TnI phosphorylation was associated with the inhibition of myofibrillar actomyosin Mg²⁺-ATPase, similar to that seen with PKA [12]. However, in beating heart studies (rat or rabbit) indirect PKC activation using either ET-1 treatment [238]or -adrenergic stimulation [11], was insufficient to stimulate TnI phosphorylation, despite being an excellent substrate for PKC in vitro in the latter study. One explanation for this may be that activation of specific PKC isozymes are needed to demonstrate in vivo phosphorylation of TnI, and hence distinct functional effects. Recent studies by Jideama seem to support this hypothesis since sites on TnI exclusively phosphorylated by PKC- were only minimally phosphorylated in a myocyte model [235].

The phosphorylation of TnI in situ appears to be specific for PKA-dependent and PKC-dependent processes, and is unaffected by inotropic interventions thought to act by variations in [Ca²⁺]_i. Both Frearson et al., using oubain and variable amounts of Ca²⁺ [239], and Ezrailson et al., using Ca²⁺ ionophores and paired pulse stimulation [240], have showed that TnI is not phosphorylated in beating heart preparations under these conditions.

2.3.2 Troponin T
Troponin T (TnT) is the largest component of the Tn complex and is named for its ability to bind to tropomyosin [211]. Early studies by Perry and colleagues found that TnT existed as a phosphoprotein in cardiac muscle [213, 241]. Like TnI, TnT has also been shown to be phosphorylated stoichiometrically and at multiple sites by PKC in vitro [232–234], leading to reduced Ca²⁺-stimulated actinomyosin MgATPase activity [234, 242]. These sites are all located at the C-terminus of TnT where binding to tropomyosin and TnC occurs. More recent studies have revealed that PKC- selectively phosphorylated two previously unknown sites in TnT, leading to a slight increase in Ca²⁺-sensitivity without affecting the Mg²⁺-ATPase activity [235]. It now appears that PKC- is the major isozyme responsible for phosphorylation of TnT in cardiac myocytes in situ and perhaps in vivo, though further work is needed utilizing pharmacological and molecular approaches to decipher the role of specific PKC isozymes in myofibrillar and cardiac function.

2.3.3 Myosin light chains
The myosin molecule is one of the largest proteins known, having a mass of 500 000 daltons, and is composed of two classes of subunits held together by non-covalent forces, namely two heavy chains (each 200 kDa) and four light chains associated with the head region of the heavy chain [211]. In cardiac muscle myosin, two types of light chains are found, one with a M_r of about 27 000 daltons and termed LC-I (essential light chains), and a second with a M_r of 19 000 daltons and referred to as LC2, regulatory light chain, or phosphorylated-light chain (MLC-2). Functionally MLC-2 is not essential for enzymatic hydrolysis of ATP by myosin, rather it may act to modulate the ATPase activity. The actin-activated ATPase of cardiac myosin changes several-fold with no change in MLC-2 composition, although removal of MLC-2 caused an enhanced ATPase, which was restored with the readdition of MLC-2 [243]. Questions remain as to whether or not phosphorylation of MLC-2 impacts its interaction with myosin, but readdition of phosphorylated MLC-2 to a MLC-2 free myosin preparation does not inhibit actin-activated ATPase to the same extent as nonphosphorylated MLC-2 [243]. It should also be noted that MLC-2 possess Ca²⁺-binding regions in its structure homologous to that seen with other Ca²⁺-binding proteins like TnC and calmodulin. Not surprisingly, the kinase that phosphorylates MLC-2 is a specific, Ca²⁺/CAM-dependent enzyme termed myosin light chain kinase (MLCK), and one that phosphorylates MLC-2 at a single specific Ser residue.

Much of the understanding for a role of phosphorylation on MLC-2 function has evolved due to studies on its phosphorylation properties in intact hearts following inotropic interventions [147, 244, 245]. Various studies have found that MLC-2 phosphorylation levels are unaltered subsequent to treatment with adrenaline [246, 247], or with increased [Ca²⁺] or [K⁺] levels [246, 248]. However, in several instances changes in MLC-2 phosphorylation have been observed under β-adrenergic stimulation [152, 248–250], but in some of these studies the changes in MLC-2 phosphorylation may be due to experimental artifact (i.e. uncontrolled heart rate effects and lower than expected baseline MLC-2 phosphorylation levels) [248–250]. Despite a general lack of effect on MLC-2 phosphorylation with short and long term changes in contractility, perfusion studies do show that the P_i group in MLC-2 is rapidly being turned over under these conditions [246, 251]. This indicates that the MLCK and putative phosphatases are both active at the level of MLC-2 in the heart. In fact, Neumann et al. demonstrated in a recent study that MLC-2 phosphorylation levels were increased following treatment with okadaic acid, along with other cardiac regulatory phosphoproteins [229]. Lastly, a mention should be made of an interesting study whereby arachidonic acid and ET-1 treatment of cardiac myocytes resulted in enhanced PL-C phosphorylation, via a presumed PKC-dependent process [237]. Since there have been no other reports of PKC effects of PL-C phosphorylation or function it is difficult to interpret these observations.

2.3.4 C-protein
C-protein is one a group of structural proteins found in striated muscle, including heart, and has a 150 000 dalton molecular weight by SDS-PAGE [2]. The function of C-protein is unclear, although both structural and regulatory roles have been suggested. For instance, it has been reported that C-protein serves to stabilize the thick filament via interactions with the myosin tail regions [252]. Since C-protein at low ionic strength can inhibit actomyosin Mg²⁺-ATPase activity, and activate the Mg²⁺-ATPase at high ionic strength, it has been suggested that perhaps it may be involved in some sort of regulatory capacity [253].

Characterization of cardiac C-protein phosphorylation in vitro showed that it is an excellent substrate for PKA [254]. Early evidence suggested that C-protein phosphorylation was important in the physiological regulation of myofilament activity. Jeacocke and England were the first to provide in vivo evidence for C-protein phosphorylation mediated by PKA [255]. Since then a number of other groups have confirmed these results in a variety of paradigms, including β-adrenergic receptor stimulation [145, 227, 256], and indirect activation of PKA-dependent processes [151]. Properties of the phosphorylation of C-protein appear to be similar to that seen with TnI. Recent reconstitution studies though demonstrated that the decrease in Ca²⁺-sensitivity associated with PKA phosphorylation is not altered by myofibrils lacking C-protein [220], thereby obscuring the role C-protein may play in regulating myofibrillar function and cardiac contractility.

Dephosphorylation of C-protein tracks slowly with TnI after withdrawal of β-agonist treatment [145], and could be stimulated by ACh or adenosine treatment [165, 225]. Phosphatase inhibition with calyculin A resulted in a time- and concentration-dependent increase in C-protein phosphorylation, suggesting that both it and TnI are dephosphorylated by the same protein phosphatase [257]. No evidence currently exists demonstrating PKC-dependent activation of the C-protein phosphorylation state. The only study to examine this found no change in the phosphorylation state of any cardiac regulatory protein except for PL [11]. Thus, while C-protein shares many similarities to TnI, little is known about its function, thereby making it difficult for any speculation.

2.4 Other cardiac phosphoproteins
2.4.1 Protein phosphatase inhibitor-1
Phosphorylation states of proteins depend on the relative rates of phosphorylation and dephosphorylation [180]. It is now accepted that protein phosphatases can counteract the phosphorylation of proteins, and more importantly, are themselves under strict control by endogenous phosphatase inhibitor proteins. One such regulatory protein is termed protein phosphatase inhibitor-1 (PPI-1), and is active only when phosphorylated by PKA-dependent processes. Thus, these types of phosphatase regulatory proteins are envisioned as participating in a positive feedback system wherein PKA-induced protein phosphorylation is enhanced when the PPI-1, itself activated by PKA, can inhibit the activity of protein phosphatases [180].

Several reports describe the putative role of PPI-1 in mediating PKA-dependent phosphorylation processes in the heart [20, 149, 258]. Iyer et al. reported that in rat heart membrane preparations PPI-1 phosphorylation was increased by isoproterenol treatment [258]. Furthermore, PLB dephosphorylation by exogenous protein phosphatase type 1 was reduced by the enhanced phosphorylation state of PPI-1. In separate studies, Watanabe and coworkers demonstrated that isoproterenol and forskolin increased PPI-1 activity two- and three-fold respectively, effects that could be antagonized by ACh co-administration [20]. Both PKA-dependent agents also reduced protein phosphatase type 1 activity intrinsic to the CSR using either []-ATP-labeled membrane vesicles or phosphorylase a as substrates. Co-administration of ACh antagonized these effects as well, and were reflected as an increase in protein phosphatase type 1 activity. The mechanisms for the inhibition of PPI-1 activity by muscarinic agents are presently unclear, but appear to involve a PKA-independent process since cAMP levels were not effected when hearts were treated with ACh in the absence or presence of isoproterenol [21]. This same study demonstrated that similar PPI-1 inhibitory properties were shared by adenosine agonists. In a subsequent study, direct in vivo evidence was provided for the phosphorylation of the 26 kDa PPI-1 in response to isoproterenol [149]. Thus, PPI-1 can be modulated in intact heart by autonomic regulatory mechanisms.

2.4.2 Nuclear components
A growing body of evidence now exists that implicates the phosphorylation of nuclear proteins in the modulation of gene expression [259]. Many of these nuclear proteins are known to be transcriptional factors that when subjected to a variety of stimuli results in post-translational modifications and changes in gene expression. Identification of these nuclear proteins and characterization of their phosphorylation mechanisms in vivo may lend great significance to their role in normal and diseased cell function.

In cardiac muscle, attention has recently been given to the characterization of phosphorylation properties of the cAMP response element binding protein (CREB-P) [260, 261]. CREB-P is a key mediator of gene transcription activation in response to a stimulated cAMP-signaling pathway [262]. It is a dimeric protein of about 43 000 daltons and contains phosphorylation sites for PKA, PKC, casein kinase II, and other kinases. These phosphorylation sites lie within a specific transactivation domain region which is critical for induction of gene expression. CREB-P is phosphorylated in vitro by PKA, PKC and Ca²⁺/CAM-PK at the same Se¹³³ site. It is thought that activation of CREB-P, at least by PKA, causes a conformational change in the protein which allows the transactivation domain to interact with its target protein, a TATA-binding factor. In turn, the TATA-binding factor can effect downstream changes in gene expression.

CREB-P has been shown to be expressed in the nuclei of primary embryonic chick heart cell cultures [260]and in human heart tissue [261]. Following treatment with either isoprotere