Ca2+/calmodulin-dependent protein kinase II regulates cardiac Na+ channels

1Department of Cardiology and Pneumology, Georg-August-University G?ttingen, G?ttingen, Germany. 2Department of Cardiology and Angiology, University Hospital Münster, Münster, Germany. 3Department of Medicine I, Division of Cardiology, University of Würzburg, Würzburg, Germany. 4Department of Pharmacology, UCSD, La Jolla, California, USA. 5Department of Physiology, Loyola University Chicago, Chicago, Illinois, USA.

Address correspondence to: Lars S. Maier, Department of Cardiology and Pneumology, Georg-August-University G?ttingen, Robert-Koch-Str. 40, 37075 G?ttingen, Germany. Phone: 49-551-399481; Fax: 49-551-398941; E-mail: lmaier@med.uni-goettingen.de.

Received for publication August 18, 2005, and accepted in revised form October 3, 2006.

    Abstract

In heart failure (HF), Ca2+/calmodulin kinase II (CaMKII) expression is increased. Altered Na+ channel gating is linked to and may promote ventricular tachyarrhythmias (VTs) in HF. Calmodulin regulates Na+ channel gating, in part perhaps via CaMKII. We investigated effects of adenovirus-mediated (acute) and Tg (chronic) overexpression of cytosolic CaMKIIC on Na+ current (INa) in rabbit and mouse ventricular myocytes, respectively (in whole-cell patch clamp). Both acute and chronic CaMKIIC overexpression shifted voltage dependence of Na+ channel availability by –6 mV (P < 0.05), and the shift was Ca2+ dependent. CaMKII also enhanced intermediate inactivation and slowed recovery from inactivation (prevented by CaMKII inhibitors autocamtide 2–related inhibitory peptide or KN93). CaMKIIC markedly increased persistent (late) inward INa and intracellular Na+ concentration (as measured by the Na+ indicator sodium-binding benzofuran isophthalate ), which was prevented by CaMKII inhibition in the case of acute CaMKIIC overexpression. CaMKII coimmunoprecipitates with and phosphorylates Na+ channels. In vivo, transgenic CaMKIIC overexpression prolonged QRS duration and repolarization (QT intervals), decreased effective refractory periods, and increased the propensity to develop VT. We conclude that CaMKII associates with and phosphorylates cardiac Na+ channels. This alters INa gating to reduce availability at high heart rate, while enhancing late INa (which could prolong action potential duration). In mice, enhanced CaMKIIC activity predisposed to VT. Thus, CaMKII-dependent regulation of Na+ channel function may contribute to arrhythmogenesis in HF.

    Introduction

Altered Na+ channel gating was shown to underlie long QT syndrome 3 (LQT3) (1), Brugada syndrome (2), and isolated cardiac conduction defects predisposing to life-threatening ventricular tachyarrhythmias (VTs). However, these mutations are relatively rare. Heart failure (HF) is associated with an increased risk of sudden death mainly caused by VT and fibrillation (3). The mechanisms are poorly understood, but altered Na+ channel gating may be involved.

Abnormal conduction is the proximate cause of sudden death in HF, and Na+ channels critically determine conduction velocity (4). A persistent (late) Na+ current (INa) was shown to cause prolongation of action potentials (APs) in HF myocytes (5). A tetrodotoxin-sensitive (TTX-sensitive) pathway was implicated in increased intracellular Na+ concentration ([Na]i) in HF (6).

It is known that calmodulin (CaM) regulates Na+ channel gating through binding to an IQ-like motif at the C terminus (7). Downstream signaling through Ca2+/CaM-dependent protein kinase II (CaMKII) may be of relevance, but little is known about CaMKII-dependent effects on INa. CaMKII is the predominant isoform in the heart (8). Upon phosphorylation, CaMKII is known to alter L-type Ca2+ channel function, providing an integrative feedback for oscillatory intracellular free Ca2+ ([Ca2+]i) (8). In human HF and in an animal HF model, expression and activity of CaMKII are enhanced 2- to 3-fold (9–11). We have shown that transgenic overexpression of cytosolic CaMKIIC induces HF (12, 13). Inhibition of CaMKII was shown to prevent remodeling after myocardial infarction and excessive ?-adrenergic stimulation (14). CaMKII has also been linked to VT in a mouse model of hypertrophy (15).

Here we explore the role of CaMKIIC on Na+ channel function using 2 models. We assessed Na+ channel function and expression in CaMKIIC-Tg mice, which develop HF. We investigated acute CaMKIIC overexpression (rabbit myocytes) to avoid unspecific adaptations occurring in HF. We show that CaMKIIC regulates Na+ channel gating and [Na]i, which may have implications for HF.

    Results

Steady-state inactivation and activation. To assess whether CaMKIIC regulates Na+ channels, we measured steady-state inactivation of INa. Figure 1 shows steady-state inactivation as a function of membrane potential (Em) in rabbit myocytes. CaMKIIC overexpression in myocytes (hereafter referred to as "CaMKIIC myocytes") but not ?-gal overexpression in myocytes (hereafter referred to as "?-gal myocytes") caused a negative voltage shift in INa steady-state inactivation (V1/2: –83.5 ± 0.8 versus –89.7 ± 0.7 mV; P < 0.05; Table 1). This reduced the fraction of available Na+ channels at a given Em. The slope factor k was unaltered. This effect was Ca2+ dependent. When [Ca2+]i was increased to 500 nM, V1/2 was further shifted toward more negative potential (P < 0.05; Table 1). All effects were reversed using KN93 or autocamtide 2–related inhibitory peptide (AIP) (Figure 1 and Table 1). Interestingly, both inhibitors increased the fraction of available Na+ channels and reversed the effects of elevated [Ca2+]i, even in ?-gal myocytes, suggesting that there may be some basal CaMKII-dependent Na+ channel regulation. Similar results were observed using physiologic extracellular Na+ concentration ([Na]o) and when investigating CaMKIIC-Tg mice (Table 2). Again, CaMKII inhibition blocked all CaMKIIC-dependent effects on the Em dependence of Na+ channel steady-state inactivation.

   Figure 1

CaMKIIc enhances steady-state inactivation of rabbit myocyte INa (10 mM [Na+]o). (A) Mean INa availability (left) and INa during conditioning pulses (right; fit parameters in Table 1). In CaMKIIc myocytes, availability was left-shifted versus ?-gal (P < 0.05), and this was reversed by CaMKII inhibitors KN93 or AIP (P < 0.05). (B and C) Original INa traces during pre-pulses (right) and test pulse (left). INa amplitudes during pre-pulses were unaltered by CaMKIIc (see Figure 2).

   Table 1

Fit parameters of the analysis of INa inactivation (rabbit, 10 mM [Na]o)

   Table 2

Fit parameters of the analysis of INa inactivation (rabbit and mouse, physiological [Na]o)

Since steady-state inactivation of Na+ channels depends on both channel activation and inactivation, the current-voltage (I-V) relation of INa in CaMKIIC-overexpressing myocytes was assessed (basic cycle length, 2 seconds; Figure 2), with no difference in maximal current density observed. There was also no alteration in the Em dependence of Na+ channel conductance (G), suggesting that Na+ channel activation is not altered by CaMKII. Therefore, the differences in steady-state inactivation are most likely ascribable to altered inactivation.

   Figure 2

Em dependence of INa activation in rabbit myocytes (10 mM [Na+]o). (A) Current-voltage relation (I-V) for CaMKIIc versus ?-gal myocytes. (B) INa activation curve, with relative conductance derived from maximal chord conductance and reversal potential (Erev) for each I-V, and peak INa/(Em – Erev). The resulting conductance was normalized to the maximal chord conductance (usually between –10 and 0 mV). CaMKIIC had no effect on voltage for half-activation V1/2 or slope factor k (mV). Maximal INa was –4.3 ± 0.2 nA and –49.8 ± 2.4 pA/pF (Cm, 88.2 ± 5.7 pF) for CaMKIIc versus –4.1 ± 0.5 nA and –42.8 ± 4.4 pA/pF (Cm, 103.8 ± 17.7 pF) for ?-gal (P = NS).

Intermediate inactivation and recovery from inactivation. We additionally investigated intermediate inactivation (IIM), a form of Na+ channel inactivation that accumulates over a few hundred milliseconds (after fast inactivation ) and recovers much more slowly at negative Em than Ifast. Enhanced IIM has been implicated in Brugada syndrome (2) and has been suggested to be a consequence of CaM-dependent Na+ channel regulation (7). IIM was measured using depolarizations of variable duration (P1) followed by a 20-ms recovery period at –140 mV, allowing for recovery from fast inactivation but not from IIM. The following test pulse (P2) to –20 mV activated all channels not in IIM (Figure 3A). Physiologically, only a small fraction of Na+ channels undergo IIM and reduce the amount of channels available for the second excitation. Figure 3 shows that CaMKIIC overexpression significantly increased the fraction of channels undergoing IIM. Curve fitting analysis revealed that the amplitude A was significantly increased, whereas the rate constant kIM was unaltered (Table 1). Thus, CaMKIIC may increase the fraction of channels that can enter IIM, but not the rate constant of development of IIM. Again, this effect was Ca2+ dependent. At high [Ca2+]i (500 nM), the amplitude of IIM was significantly increased (Table 1). All effects were reversible with KN93 or AIP, which were also effective in ?-gal myocytes.

   Figure 3

CaMKIIc increases IIM of INa in rabbit myocytes (10 mM [Na+]o). (A) Increasing conditioning pulse duration (P1) reduced peak INa assessed with a second pulse (P2). CaMKIIc increased the amount of accumulating IIM (versus ?-gal; P < 0.05), an effect reversed by KN93 or AIP (P < 0.05). Mean data were fitted to a single exponential (see Table 1). (B) Original traces of peak INa (at P2) for different P1 durations.

As for steady-state inactivation, a similarly enhanced IIM was observed using a more physiologic [Na]o and in CaMKIIC transgenics (Table 2). Again, CaMKII inhibition blocked CaMKIIC-dependent effects. In WT and ?-gal myocytes, KN93 reduced the amount of IIM even further (Table 2). Thus, CaMKII may contribute to IIM, and there may be some basal CaMKII-dependent IIM in WT mouse and control rabbit cells (with lower basal IIM in WT).

Inactivation and recovery from inactivation are closely related and critically determine channel function. Recovery from inactivation was investigated using a sustained depolarization at a time scale that initiates fast and IIM (1,000 ms), followed by recovery intervals of increasing durations and a subsequent test pulse (Figure 4A). In comparison to that of control, recovery from inactivation was slowed in CaMKIIC myocytes (Figure 4). Data were fit to a single exponential relation. The time constant for recovery from inactivation (rec) increased from 12.3 ± 1.2 ms (?-gal) to 18 ± 1.1 ms (CaMKIIC; Table 1), indicating that recovery was slower by nearly 50% in CaMKIIC myocytes. Either KN93 or AIP returned rec to control levels (Table 1). KN93 by itself slightly reduced rec in ?-gal myocytes, consistent with a basal CaMKII-dependent effect on INa recovery from inactivation. The effect of CaMKIIC on rec was Ca2+ dependent, because at high [Ca2+]i (500 nM), rec increased further, and KN93 completely reversed this (Figure 4A and Table 1).

   Figure 4

CaMKIIc slows INa recovery from inactivation in rabbit myocytes (10 mM [Na+]o). (A) Increasing durations of recovery interval between conditioning pulse (P1, causing INa and inactivation) and test pulse (P2). CaMKIIC significantly slowed recovery versus ?-gal, an effect blocked upon CaMKII inhibition (KN93; P < 0.05). Elevation of [Ca2+]i to 500 nM in myocytes overexpressing CaMKIIC further slowed recovery, and KN93 completely inhibited this effect. Mean data were fit to a single exponential (fit parameters in Table 1). (B) Original traces of INa for different recovery intervals.

These results indicate that CaMKIIC activity substantially slows INa recovery from inactivation. This may reflect, in part, slower recovery from IIM, as there was more IIM in the cases where recovery was prolonged. The slower INa recovery could also limit INa availability, especially at high heart rates.

The CaMKIIC-dependent slowing of Na+ channel recovery was also seen in physiologic [Na+]o, and the effect could be measured in CaMKII transgenics (Table 2). Consistently, the effect of CaMKIIC overexpression was completely reversible with CaMKII inhibition.

Fast, open-state inactivation and [Na]i. When Na+ channels open, they close very rapidly, within 10–20 ms, a process called fast or open-state inactivation. In contrast to IIM, for which no structural correlate has been found yet, the cytoplasmic linker between domains III and IV and the C terminus of the Na+ channel protein has been suggested to underlie Ifast. Mutations in these regions are known to disrupt this process, leading to LQT3. Since the putative target of CaMKII-dependent regulation could be located there, we investigated the effect of CaMKIIC on fast INa decay. Acute CaMKIIC overexpression significantly slowed the late component of fast INa inactivation (see 2, Figure 5A), which was sensitive to KN93.

   Figure 5

CaMKIIc slows fast decay of INa. (A) Original traces show that CaMKIIc slows fast INa inactivation in rabbit myocytes versus ?-gal. INa decay (first 50 ms) was fit with a double exponential: y (t) = A1 exp (–t/1) + A2 exp (–t/2) + y0. Fits and 1 and 2 (ms) are shown in red (?-gal) and blue (CaMKIIc). Deceleration of INa decay was significant in 2 but not in 1. KN93 could reverse the slowing of 2 by CaMKIIC. (B) In CaMKIIC-Tg mice, both 1 and 2 were slowed compared with those in WT mice. Fits to the original traces and 1 and 2 values (ms) are shown in red (WT) and blue (CaMKIIC-Tg). KN93 did not reverse these effects. Average peak INa was –10.2 ± 0.3 nA, –79.8 ± 2.7 pA/pF, Cm 132.7 ± 4 pF for rabbit and –14.2 ± 0.6 nA, –58.9 ± 2.9 pA/pF, Cm 255.3 ± 11 pF for mouse.

When myocytes from CaMKIIC-Tg mice were assessed, initial and late components (1, 2; Figure 5B) of INa decay were significantly slowed compared with those of WT myocytes. However, KN93 did not reverse the effect in mice, leaving more open the role of CaMKII in this transgenic mouse model.

Incomplete INa inactivation during the AP can influence AP duration and [Na]i and can also be arrhythmogenic. Figure 6, A and B, shows a distinct persistent INa component in both acute and chronic CaMKII overexpression that is not seen in ?-gal rabbit or WT mouse myocytes. Again, KN93 prevented this persistent INa in CaMKIIC rabbit myocytes, but not in mice.

   Figure 6

CaMKIIc enhances late INa and increases [Na]i. INa elicited at –20 mV (for 1,000 ms) was leak subtracted and normalized to peak INa. Current was integrated between 50 and 500 ms and normalized to the INa integral if no inactivation had occurred. (A) Original traces and mean data of normalized late INa. CaMKIIC overexpression significantly increased late INa (reversed with KN93). (B) CaMKIIC-Tg mice also showed late INa, but this was not reversible with KN93. Average peak INa was –10.2 ± 0.3 nA, –79.8 ± 2.7 pA/pF, Cm 132.7 ± 4 pF for rabbit and –14.3 ± 0.6 nA, –58.1 ± 2.8 pA/pF, Cm 259.7 ± 11.6 pF for mouse. (C) Mean [Na]i at different stimulation frequencies (left) and at 1 Hz (right) in rabbit myocytes. [Na]i was elevated in CaMKIIc myocytes at all frequencies (P < 0.05) and reduced by KN93 (P < 0.05). (D) Mean data for CaMKIIC-Tg mice also showing elevated [Na]i (versus WT; P < 0.05), but this was not reversed by KN93.

To assess the impact of these alterations in late INa on Na+ influx, we integrated the late INa (50–500 ms) and normalized it to the cytosolic volume. For rabbit, values were (in μM): CaMKIIC (versus ?-gal), –7.47 ± 0.77 (–3.04 ± 0.92); and CaMKIIC+KN93 (versus ?-gal plus KN93), –3.35 ± 0.36 (–2.84 ± 0.63). For mouse, integrals were Tg (versus WT), –8.23 ± 1.60 (–2.05 ± 0.60); and Tg plus KN93 (versus WT plus KN93), –7.9 ± 2.46 (–1.89 ± 0.61). If the amount of Na+ entry was extrapolated to 1 minute, the Na+ influx was 1.0 ± 0.1 versus 0.4 ± 0.1 mM/min (rabbit, CaMKIIC versus ?-gal) and –1.1 ± 0.21 versus –0.27 ± 0.08 mM/min (mouse, Tg versus WT). Interestingly, the increased amount of Na+ influx upon CaMKII overexpression strikingly resembles the TTX-sensitive Na+ entry suggested to cause elevated [Na]i in an HF model with increased CaMKII activity (6, 11).

To assess whether the CaMKII-dependent alterations in INa gating cause increased [Na]i, we measured [Na]i in field-stimulated myocytes. In CaMKIIC-overexpressing rabbit myocytes, [Na]i was significantly increased at all stimulation frequencies compared with that in control. This increase was completely reversed by KN93 (Figure 6C). In CaMKIIC-Tg mice, [Na]i was also significantly elevated (versus WT), but the increase was not reversible with KN93 (Figure 6D).

Association of CaMKII with Na+ channels and CaMKII-dependent phosphorylation. Coimmunoprecipitation studies revealed that CaMKII associates with Na+ channels in both mouse and rabbit tissue (anti-Pan Nav, an antibody recognizing all Na+ channel isoforms; Figure 7, A and B). Western blotting of the immunoprecipitates repeated with cardiac-specific anti-Nav1.5 showed similar results (data not shown), proving that the cardiac Na+ channel isoform interacts with CaMKII. Using immunocytological staining, CaMKIIC and Na+ channels were found to be colocalized in CaMKIIC-overexpressing myocytes (Figure 7C). No colocalization was seen when the anti-HA antibody was used in control-transfected myocytes (data not shown).

   Figure 7

CaMKII-dependent association with and phosphorylation of Na+ channels. (A and B) Equal amounts of CaMKII were IP from mouse hearts (A, Tg versus WT) and rabbit myocytes (B, CaMKIIc versus nontransfected cells) and subjected to immunoblotting. CaMKII IP showed CaMKII association with Na+ channels in WT and Tg hearts (n = 3 and n = 3, respectively), while IP with polyclonal rabbit anti-Cav1.2a did not. Similar results were seen in rabbit myocytes (CaMKIIC and nontransfected cells, n = 4 and n = 2, respectively). (C) Colocalization of CaMKII and Na+ channel in rabbit myocytes with immunocytological staining. (D and E) Autoradiograms measuring 32P incorporation into Na+ channels. Equal amounts of Na+ channels were immunoprecipitated from WT mouse hearts (n = 3) and directly phosphorylated with or without CaMKII, KN93 (50 μM), AIP (5 μM), or PKA/PKC inhibitor cocktail (PKA/C-I; 1 μM). (F) Endogenous CaMKII-dependent Na+ channel phosphorylation was activated in permeabilized rabbit myocytes by preincubating for 5 minutes in internal solutions of 500 nM [Ca2+] plus 2 μM CaM (versus 50 nM [Ca2+]). Sites not already phosphorylated were subsequently back-phosphorylated in Na+ channel immunoprecipitates (equal amounts) by exogenous preactivated CaMKII and ATP--32P. The intensity of 32P is inversely related to the CaMKII-dependent phosphorylation during preincubation. (G and H) Western blots (anti-Nav Pan) show upregulation of Na+ channel expression in CaMKIIC-Tg versus WT mice (relative to calsequestrin ) but unaltered expression in CaMKIIC rabbit myocytes (data pooled for MOI 10 and 100).

CaMKII-dependent phosphorylation of Na+ channels was investigated using ATP--32P. First, we exposed immunoprecipitated Na+ channels from WT mice to purified preactivated exogenous CaMKII (Figure 7D). Exogenous CaMKII clearly phosphorylated Na+ channels, and this effect was inhibited when KN93 or AIP were included, but not when PKA and PKC were blocked (Figure 7E).

Since endogenous CaMKII associates with the Na+ channel, we tested whether endogenous myocyte CaMKII can phosphorylate Na+ channels in the myocyte (Figure 7F). Saponin-permeabilized rabbit myocytes were exposed for 5 minutes to internal solution at 50 nM [Ca2+] or 500 nM [Ca2+] plus 2 μM CaM (to activate endogenous CaMKII). The extent of CaMKII-dependent Na+ channel phosphorylation was assessed afterward by Na+ channel immunoprecipitation and subsequent back-phosphorylation with exogenous preactivated CaMKII and ATP--32P. Activation of endogenous CaMKII (500 nM [Ca2+]) reduced back-phosphorylation (lane 2 versus lane 1), indicating increased Na+ channel phosphorylation upon activation of cellular CaMKII. Inclusion of AIP in the myocyte preincubation prevented the increased phosphorylation (lane 3, Figure 7F). Interestingly, overexpression of CaMKIIC increased the phosphorylation level (decreased back-phosphorylation) of Na+ channels at 50 nM [Ca2+] to a level comparable to that attained in the control at the 500 nM [Ca2+] incubation. Thus, CaMKII associates with the Na+ channel, and the endogenous CaMKII can phosphorylate the channel.

Western blotting revealed significantly increased Na+ channel expression in hearts from Tg mice (1.6-fold versus WT; Figure 7G). However, INa was not significantly different in Tg versus WT hearts (–127.1 ± 16 pA/pF versus –109 ± 22.4 pA/pF, respectively, at –30 mV; n = 5 versus 9; P = NS; average maximum INa, –15.1 ± 1 nA, –115.5 ± 15.2 pA/pF, membrane capacitance [Cm], 197.3 ± 21.7 pF). In contrast, Na+ channel expression was not altered in rabbit myocytes acutely overexpressing CaMKIIC (Figure 7H).

CaMKII and arrhythmias. To test whether CaMKIIC mice were prone to VT, we performed electrophysiological measurements in vivo (Figure 8). Application of 2 consecutive premature beats via programmed electrical stimulation induced monomorphic and polymorphic VT. In addition, 1 Tg mouse died because of spontaneous ventricular fibrillation immediately after recordings were started. In contrast, no arrhythmias were observed in WT mice (Figure 8, A and C). In separate experiments, application of isoproterenol (2 mg/kg, i.p.) increased heart rate from 460 ± 50 to 680 ± 10 bpm for WT mice, but no arrhythmias were observed (Figure 8, B and C). In contrast, in Tg mice, isoproterenol infusion induced monomorphic VT. Analysis of resting ECG parameters (Figure 8D) revealed that the corrected QT (QTc) interval and QRS duration were significantly prolonged in Tg versus WT mice. Interestingly, the PR interval was significantly shortened in Tg mice. CaMKII has been previously implicated in AV nodal conduction (16). To assess whether CaMKIIC overexpression alters AP duration depending on the heart rate, monophasic APs (MAPs) were recorded in isolated perfused hearts. After AV node ablation, Tg hearts had higher intrinsic ventricular heart rates. Spontaneous basic cycle lengths (BCLs) were 325 ± 30 ms (WT) versus 231 ± 16 ms (Tg) (n = 21 versus 22; P < 0.05). At high pacing frequencies (BCL, 100 ms), MAP duration was not significantly different for Tg versus WT hearts (Figure 8, E and F). It is possible that enhanced steady-state Na+ channel inactivation and disturbed open-state inactivation may counterbalance each other. At lower pacing frequencies physiological for mice, MAP duration was significantly prolonged in Tg hearts. This was not reversible by CaMK inhibition, suggesting that other effects, possibly related to adaptation or HF, also affect repolarization in the chronic CaMKII expression model.

   Figure 8

Arrhythmias in CaMKIIC-Tg mice. (A) Original ECG traces during programmed electrical stimulation in vivo. In WT mice, 2 consecutive premature beats (S2-S3, coupling interval 38 ms) did not induce arrhythmias (top). In contrast, in 2 different Tg mice, the same protocol induced monomorphic (middle) or polymorphic VT (bottom). (B) Representative ECG traces before and after isoproterenol (Iso) administration. P-waves (indicated by asterisks) were apparent during normal rhythms (WT with or without isoproterenol) but were absent in Tg mice after isoproterenol administration. This apparent AV dissociation indicates arrhythmia. (C) Frequency of arrhythmia induction was significant in Tg but not WT mice for both programmed electrical stimulation (left) and isoproterenol administration (right). (D) Resting ECG parameters (RR interval, corrected QT interval, QRS duration, and PR interval). (E) Right-ventricular MAPs from hearts paced at basic cycle lengths (BCLs) of 100 (top) and 150 ms (bottom) for WT mouse hearts (left), Tg mouse hearts (middle), and Tg mouse hearts during infusion with KN93 (right). (F) Mean MAP durations at 90% repolarization (APD90). At slower heart rates (BCL, 150 ms), APD90 was significantly prolonged in Tg versus WT mice (but could not be reduced by KN93). At higher heart rates, there was no significant prolongation of APD90 in Tg versus WT.

The effective refractory period (ERP) was decreased in Tg hearts (WT, 27 ± 3 ms, Tg: 20 ± 4 ms; n = 14 versus 21; P < 0.05), resulting in progressive encroachment of excitation, a scenario known to cause VT (17).

    Discussion

The present study shows for the first time to our knowledge that CaMKII regulates Na+ channel function in myocytes, most likely by association with and phosphorylation of Na+ channels. This is based on our findings that CaMKII coimmunoprecipitates with and phosphorylates Na+ channels; and that acute CaMKIIC overexpression slows fast INa inactivation, enhances late INa, increases [Na]i, but also enhances accumulation of intermediate INa inactivation, slows recovery from inactivation, and shifts steady-state inactivation of Na+ channels to more negative Em in a Ca2+-dependent manner. Moreover, all of these effects could be reversed with CaMKII inhibition. All these effects were also seen in Tg mice overexpressing CaMKIIC, although impaired inactivation and elevated [Na]i were not reversed by acute CaMKII inhibition. Overall, these effects prolong Na+ influx during depolarization (possibly explaining enhanced [Na]i) but increase steady-state inactivation of Na+ channels at shorter diastolic intervals. This could be particularly arrhythmogenic, and since CaMKII is elevated in HF (9, 10), these effects could cause an acquired form of arrhythmogenesis. Indeed, the altered Na+ channel phenotype caused by CaMKII is very similar to that caused by the 1495InsD mutation in human Nav1.5 that is linked with features common to LQT3 and Brugada syndrome (18, 19). CaMKIIC-Tg mice were also prone to VT (with reduced effective refractory period, prolonged QRS duration, and disturbed repolarization). While Na+ channel–independent effects could also contribute to this arrhythmogenesis in CaMKIIC-Tg mice, the documented Na+ channel phenotype may well be involved.

Na+ channel regulation by kinases. Na+ channels are isoform-specifically regulated by kinases including PKA and PKC. PKA-dependent phosphorylation (at Ser526 and Ser529) in the I-II cytoplasmic linker increases whole-cell conductance without altering channel gating resulting from an increased number of functional Na+ channels, possibly by altering channel trafficking, although cAMP-dependent enhancement of Na+ channel steady-state inactivation has also been reported (20). PKC-dependent phosphorylation (Ser1505) in the III-IV linker reduces maximal conductance and enhances steady-state inactivation (20).

Deschênes et al. (21), using heterologous Na+ channel expression, showed that the CaMK inhibitor KN93 slowed INa decay and shifted steady-state inactivation in the depolarizing direction, but the specific CaMKII inhibitor AIP did not alter INa gating. They concluded that a CaMK other than CaMKII (e.g., CaMKIV) might have been responsible for INa modulation. However CaMKIV expression levels in heart are low and primarily nuclear (8), so that CaMKIV is unlikely to modulate INa. Here we show that overexpression of the predominant cardiac cytosolic CaMKIIC enhanced steady-state and IIM of Na+ channels. Thus, some effects previously attributed to CaM may actually be mediated by CaMKIIC. In the present study, AIP reversed the effect of CaMKIIC overexpression on steady-state INa inactivation but also produced smaller shifts in control myocytes (Table 2). This is consistent with some baseline CaMKIIC effect on Na+ channels in ?-gal.

CaMKIIC association and CaMKII-dependent phosphorylation. CaM associates with Na+ channels (to an IQ motif in the C terminus) (7, 22). We show here that CaMKIIC also associates with Na+ channels. We also show that exogenous CaMKII can phosphorylate Na+ channels and, importantly, that endogenous CaMKII in rabbit myocytes can phosphorylate the Na+ channel at a [Ca2+]i that is physiologically relevant (500 nM). Moreover, when CaMKII was overexpressed in these myocytes, the Na+ channel was more phosphorylated even at diastolic [Ca2+]i level (50 nM). These results suggest that the observed CaMKII- and [Ca2+]i-dependent INa gating changes (Figures 1–6) may be due to Na+ channel phosphorylation by CaMKII that is associated with the Na+ channel.

It is unclear whether CaM preassociated with the Na+ channel participates in activating CaMKII. However, Na+ channels may, like many ion channels, be a multiprotein regulatory complex (23).

In the present study, acute CaMKIIC overexpression did not alter functional INa expression, according to maximal INa density and protein expression levels. This suggests that CaMKIIC-dependent Na+ channel regulation may not involve primary effects on Na+ channel expression or trafficking. However, Tg mice exhibiting HF showed significantly more Na+ channel expression but unaltered maximal INa density. This could be a consequence of the HF phenotype, but limited available data suggest that peak INa density is unaltered in HF (24). Possibly, CaMKII-dependent reduction in INa availability in rapidly beating Tg mouse hearts causes a compensatory upregulation of Na+ channel expression. However, the unaltered peak INa may also mean that some Na+ channels are less functional (e.g., due to altered channel type, internalization, gating behavior).

Enhanced steady-state and IIM. Several types of INa inactivation have been distinguished (3): (a) Ifast occurring over 2–10 ms (in our analysis with 2 time constants) and recovers rapidly at negative Em; (b) IIM, which accumulates after Ifast has occurred (over hundreds of milliseconds) and recovers more slowly; (c) slow inactivation from the open state (over hundreds of milliseconds); and (d) ultra-slow inactivation, which occurs over many seconds (not studied here). These can all influence Na+ channel steady-state inactivation, AP duration, and Na+ flux balance.

CaMKIIC overexpression shifted steady-state inactivation of INa to more negative Em and increased the fraction of channels undergoing IIM. Since these recover with slower kinetics, fewer channels would be available for the next excitation at normal heart rates. These effects are thus interrelated functionally (and possibly mechanistically). Indeed, both effects were reversed by CaMKII inhibition. A similar reduction in channel function was shown to underlie Brugada syndrome (2).

Impairment of Ifast and [Na]i. Acute CaMKIIC overexpression also slowed fast INa inactivation (Figure 5A), enhanced late INa (Figure 6A), and increased [Na]i (Figure 6C). All effects were prevented by CaMKII inhibition, implicating acute effects of CaMKII. These 3 effects are also functionally related and perhaps mechanistically associated. Slowed INa inactivation could prolong AP duration, leading to early afterdepolarizations (EADs) or LQT3-like propensity for arrhythmias. In addition, the elevated [Na]i as a consequence of slowed INa inactivation would tend to cause Ca2+ overload–induced spontaneous sarcoplasmic reticulum (SR) Ca2+ release and transient inward current (via Na+/Ca2+ exchange) that underlie delayed afterdepolarizations (DADs) (3).

In HF, these arrhythmogenic effects could be exacerbated by the already elevated [Na]i (attributed to TTX-sensitive Na+ entry) and elevated Na+ entry via slowly inactivating or noninactivating INa (5, 6, 25). The increased CaMKII expression and activation in HF (9–11) could produce the gamut of INa and [Na]i alterations described here upon CaMKIIC overexpression (enhancing arrhythmogenesis in HF). The elevated [Na]i reported here is unlikely to be explained by a Na+ "window current" (occurring at Em where steady-state activation and inactivation curves overlap) (26). This is because the activation curve is unaltered, while steady-state inactivation is shifted in the hyperpolarizing direction, a combination which would reduce window INa.

Acute and chronic CaMKIIC overexpression produced the same qualitative changes in INa gating and [Na]i, and for acute overexpression, all of the effects could be completely reversed by CaMKII inhibition. However, in CaMKIIC-Tg mice, some of these changes (slowed open state INa inactivation, persistent INa, elevated [Na]i) were not reversed by KN93 application. Thus while CaMKIIC can mediate the whole array of INa and [Na]i effects (based on the acute adenoviral studies with and without KN93 or AIP), there may be other effects in chronic HF or CaMKIIC overexpression that mimic or sustain the CaMKII effect on slowing INa inactivation. It is not clear what the cause may be, but possibilities that may merit future study include the following: (a) Conceivably another protein kinase activated during hypertrophy or HF might phosphorylate the same Na+ channel sites as CaMKII, causing slow open-state INa inactivation and enhancing late INa; (b) There could be less local phosphatase at the Na+ channel in HF (as reported for the ryanodine receptor; refs. 11, 27); (c) Normal cardiac Na+ channels can also shift into a bursting mode to produce persistent INa (28, 29), and persistent INa may have different TTX sensitivity and Em dependence from transient INa (30) and contribute to lengthening AP duration in HF (5); (d) Some of the additional Na+ channels expressed in CaMKIIC-Tg mice (Figure 7G) could be an altered form with weaker inactivation. Regardless of the underlying mechanism, an impaired fast Na+ channel inactivation mimics the functional defects of mutant Na+ channels associated with congenital LQT3 (1, 31) and therefore may not only prolong AP and increase [Na]i but also predispose to VT.

Divergent effects on gating and possible clinical implications for arrhythmias. CaMKIIC enhances steady-state and IIM, while at the same time impairing Ifast and enhancing persistent INa. Strikingly similar changes in Na+ channel gating were shown for a human mutant Na+ channel (Asp insertion at 1795 in the C terminus), which shows simultaneous LQT3-like and Brugada-like phenotypes (18). Mutant 1795insD Na+ channels expressed in mammalian cells exhibit a shift in steady-state inactivation to negative potentials (with unaltered activation), slowed inactivation, induced persistent INa, and delayed recovery from inactivation (18, 19). The net effect of both changes would be heart rate dependent. At low frequencies, impaired Ifast and persistent INa outweigh the slowed recovery from inactivation because of long-lasting diastolic intervals. This would favor AP prolongation, consistent with LQT3 syndrome (1) but also HF (4). However, at higher heart rates, there was incomplete INa recovery, reduced INa availability, and AP shortening. The consequent loss of Na+ channel function would then slow propagation and increase dispersion of repolarization. This dual scenario was also demonstrated by computational simulations that incorporate the altered INa gating properties (32). Similarly, this mixed INa phenotype was reported in a mutant Na+ channel (33) with the C terminus truncated at Ser1885 exhibiting negatively shifted steady-state channel inactivation (unchanged activation) and an increase in the fraction of channels that fail to inactivate (slowed inactivation). Therefore, it is conceivable that increased CaMKIIC activity in HF (9, 10) may alter Na+ channel gating, thereby generating the substrate for VT.

Indeed, CaMKIIC-Tg mice showed an increased propensity for VT in vivo. We show that in CaMKIIC mice, QRS duration was prolonged, indicating slowed intraventricular conduction, which is known to be proarrhythmic in Brugada syndrome. We also show that repolarization was disturbed in Tg mice, which favors VT in LQT3 patients. MAP duration was significantly prolonged, such that enhanced late INa with CaMKIIC overexpression could contribute to AP prolongation, EADs, and LQT3-like arrhythmias.

The resulting increased [Na]i could also drive up SR Ca2+ content via Na+/Ca2+ exchange, leading to spontaneous SR Ca2+ release and DADs. Interestingly, MAP prolongation was only observed at slower pacing frequencies, indicating heart rate dependence. The CaMKII-dependent INa modulation reported here may be an acquired form of combined LQT3 and Brugada syndrome, in otherwise normal Na+ channels. Such an acquired Na+ channel dysfunction may contribute to arrhythmias under conditions wherein CaMKII effects are enhanced, as in HF (9–11).

While CaMKII-dependent Na+ channel gating may contribute to the arrhythmias seen in CaMKIIC-Tg mice, changes in other ion channels may certainly also contribute and exacerbate the arrhythmogenic potential of CaMKII-modified Na+ channels. CaMKII alters L-type Ca2+ current and mediates ICa facilitation, and a reactivation of ICa during the AP can cause EADs (15, 34). CaMKII-dependent phospholamban phosphorylation also enhances SR Ca2+ ATPase activity (12), and the increase in SR Ca2+ content would tend to increase spontaneous SR Ca2+ release. Additionally, CaMKII-dependent ryanodine receptor phosphorylation increases spontaneous Ca2+ release events (12, 13, 35). Both of these SR effects could enhance DADs and trigger arrhythmias independent of Na+ channel effects. Further studies are required to isolate the relevance of these mechanisms to the arrhythmogenesis in CaMKII-Tg mice.

In summary, our study revealed that CaMKII associates with and phosphorylates cardiac Na+ channels. CaMKIIC increases steady-state Na+ channel inactivation through enhanced IIM and slows channel recovery. In contrast, CaMKII also slows open-state inactivation, thereby increasing late INa and [Na]i. This study characterizes CaMKII effects on Na+ channel gating that may have important clinical implications for a much larger poluation of patients than those documented to have genetic LQT3 or Brugada syndromes. However, further studies are required, e.g., investigating which Na+ channel sites associate with and are phosphorylated by CaMKII and precisely how important this pathway (versus others) may be arrhythmogenic in HF.

    Methods

CaMKIIC-Tg mice. Tg mice were generated as previously described (12). We used 5.5 ± 0.3–month-old CaMKIIC-Tg mice with HF (12) and age- and sex-matched WT littermates. Ventricular myocytes were isolated (13) and kept in modified Tyrode solution (in mM): 137 NaCl, 5.4 KCl, 1.2 MgSO4, 1.2 Na2HPO4, 20 HEPES, 15 glucose, 1 CaCl2 (pH 7.4). All animal studies were approved by the Ethics Committee of the Medical Faculty of the University of G?ttingen (G?ttingen, Germany).

Adenovirus-mediated CaMKIIC transfection of rabbit myocytes. Ventricular myocytes were isolated from rabbits and plated on culture dishes (36). Transfection with recombinant adenovirus encoding for HA-tagged CaMKIIC (or ?-gal as control) was performed at an MOI of 100 during plating of rabbit myocytes in supplemented M199 culture medium for 24 hours (34, 37).

Western blots. Transfected rabbit myocytes were harvested and lysed in Tris buffer (in mM: 20 Tris-HCl, 200 NaCl, 20 NaF, 1 Na3VO4, 2% 3-[(3-cholamidopropyl)dimethylammonio]-1-propanesulfonate (CHAPS), 1 DTT [pH 7.4]) and complete protease inhibitor cocktail (Roche Diagnostics) by trituration. Whole mouse hearts were homogenized in Tris buffer. Protein concentration was determined by BCA assay (Pierce Biotechnology). Denatured cell lysates and tissue homogenates (30 minutes, 37°C in 2% ?-mercaptoethanol) were subjected to Western blotting (7.5% SDS-polyacrylamide gels) using a rabbit polyclonal anti-Pan Nav or cardiac-specific anti-Nav1.5 (1:500; Alomone Labs) as primary and an HRP-conjugated donkey anti-rabbit IgG (1:10,000; Amersham Biosciences) as secondary antibody. Chemiluminescent detection was done with SuperSignal West Pico Substrate (Pierce Biotechnology).

Coimmunoprecipitation. Whole mouse hearts were homogenized in Tris buffer (in mM: 50 Tris-HCl, 200 NaCl, 20 NaF, 1 Na3VO4, 1 DTT [pH 7.4]) and protease inhibitor cocktail. Protein concentration was determined by BCA assay. Rabbit myocytes were lysed as described above. Mouse homogenates (1 mg protein) and rabbit cell lysates were suspended in dilution medium (in mM: 50 Tris-HCl, 154 NaCl, 1% CHAPS, 1 NaF, 1 Na3VO4 [pH 7.4]) and protease inhibitor cocktail. CaMKII was immunoprecipitated with rabbit polyclonal anti-CaMKII antibody (3 μg/mg protein, preincubation at 4°C overnight; M-176; Santa Cruz Biotechnology Inc.) and protein G–sepharose Fast Flow (prewashed; 2 hours, 4°C; Amersham Biosciences). As control, rabbit polyclonal anti-Cav1.2a (Alomone Labs) was used. After centrifugation, the pellets were washed with Tris buffer (in mM: 50 Tris-HCl, 154 NaCl [pH 7.4]), and immunoprecipitated proteins were eluted in 2x Laemmli sample buffer containing 4% ?-mercaptoethanol (30 minutes, 37°C) followed by centrifugation. Supernatants were subjected to Western blotting.

Immunoprecipitation and back-phosphorylation. Na+ channels were immunoprecipitated using anti-Pan Nav (5 μg/mg) and protein G–sepharose (2 hours, 4°C). Immunoprecipitated Na+ channels were resuspended in 1x CaMKII reaction buffer (in mM): 50 Tris-HCl, 10 MgCl2, 2 dithiothreitol, 0.1 Na2 EDTA (pH 7.4). CaMKII (30 U) was preactivated (New England Biolabs Inc.) in either the presence or absence of CaMKII inhibitors KN93, AIP, or PKA/PKC inhibitor cocktail (peptide inhibitors of both PKA and PKC; Upstate USA Inc.). Immunoprecipitated proteins were phosphorylated with CaMKII (30 minutes, 30°C) in the presence of ATP--32P with a specific activity of 5.7 Ci/mmol. Beads were washed with RIPA buffer, and proteins were eluted with 2x Laemmli sample buffer containing 4% ?-mercaptoethanol (30 minutes, 37°C) followed by centrifugation. Supernatants were separated on SDS–7.5% polyacrylamide gel, and phosphorylated proteins were visualized using a phosphoimager. For some experiments, rabbit myocytes were saponin permeabilized (35). After 5 minutes superfusion with a relaxation buffer containing (in mM) 150 K-aspartate, 0.1 EGTA, 5 ATP, 10 HEPES, 0.25 MgCl2, and 10 reduced glutathione (pH 7.4, 23°C), 50 μg/ml saponin was added for 30 seconds. After permeabilization, myocytes were exposed to internal solution composed of (in mM) 120 K-aspartate, 1 EGTA, 10 HEPES, 5 Na2ATP, 10 reduced glutathione, 5 U/ml creatine phosphokinase, 10 phosphocreatine, 4% dextran (Mr, 40,000), 5.7 MgCl2 (1 free [Mg]), 0.25 CaCl2 (50 nM free [Ca2+]), PKA inhibitory peptide (PKI; 15 μM), pH 7.2. After 5 minutes equilibration, okadaic acid (2 μM) was added to prevent dephosphorylation, and for 5 minutes, free [Ca2+] was either left at 50 nM or raised to 500 nM (plus 2 μM added CaM; Upstate USA Inc.) to activate endogenous CaMKII. Then myocytes were harvested and lysed, and Na+ channels were immunoprecipitated (dilution medium with 20 mM NaF) and subjected to phosphorylation by exogenous preactivated CaMKII as described above.

Coimmunocytochemical staining. Transfected myocytes were washed (PBS), fixed (4% paraformaldehyde, 30 minutes, room temperature), and blocked with 1% BSA (overnight, 4°C). Cells were incubated with primary antibodies (monoclonal mouse anti-HA, 1:100 [Roche Diagnostics]; rabbit polyclonal anti-Nav1.5, 1:60) in 0.5% BSA and 0.75% Triton X-100 (1 hour, 37°C). After washing (PBS), incubation with secondary antibodies (Texas red–conjugated goat anti-mouse IgG, fluorescein-conjugated goat anti-rabbit IgG, 1:200; Jackson ImmunoResearch Laboratories Inc.) was done in 0.5% BSA (1 hour, 37°C). Cells were covered with Vectashield HardSet Mounting Medium (Vector Laboratories) and viewed using an LSM 5 Pascal confocal microscope (Zeiss). For control, no primary antibodies were used.

Patch-clamp experiments. Ruptured-patch whole-cell voltage-clamp was used to measure INa with microelectrodes (2–3 M, room temperature). Liquid junction potentials were corrected before G seal establishment. Pipette resistance was compensated in cell-attached configuration. Cm and series resistance (Rs) were compensated automatically (using small negative pulses; where currents were capacitive, they were scaled and subtracted from currents during test pulses), largely eliminating capacitance spikes from the records. Cm and Rs were compensated after patch rupture; access resistance was typically less than 7 M. A continuous Rs compensation was done during measurements (50%). All recordings were done 5 minutes after rupture. Signals were filtered with 2.9- and 10-kHz Bessel filters and recorded with an EPC10 amplifier (HEKA). Pipettes were filled with (in mM) 40 CsCl, 80 Cs-glutamate, 10 NaCl, 0.92 MgCl2, 5 Mg-ATP, 0.3 Li-GTP, 10 HEPES, 0.03 niflumic acid, 0.02 nifedipine, 0.004 strophanthidin, 5 BAPTA (tetra-cesium salt), 1 5,5'-dibromo BAPTA (tetrapotassium salt), 1.49 CaCl2 (free [Ca2+]i, 100 nM), or 3.73 mM CaCl2 (free [Ca2+]i, 500 nM) (pH 7.2, CsOH).

The bath solution contained (in mM) 10 NaCl, 130 tetramethylammonium chloride, 4 CsCl, 1 MgCl2, 10 glucose, 10 HEPES, 0.0001 thapsigargin (pH 7.4, NaOH). Nifedipine, strophanthidin, and thapsigargin were added to the pipette and bath solution. Low [Na]o solutions were used for better voltage control. We cannot exclude voltage errors completely, but the INa amplitudes were similar among groups, so comparisons should still be valid. Some experiments were done in physiologic [Na]o with bath solution containing 140 mM NaCl without tetramethylammonium chloride. In all experiments, myocytes were mounted on the stage of a microscope (Eclipse TE2000-U; Nikon).

[Na]i. Myocytes were loaded with 10 μM sodium-binding benzofuran isophthalate–AM (SBFI-AM; Invitrogen) for 2 hours at room temperature (36, 38), mounted on the microscope stage, and superfused with Tyrode solution (in mM): 140 NaCl, 4 KCl, 1 MgCl2, 5 HEPES, 10 glucose, 1 (mouse) or 2 (rabbit) CaCl2 (pH 7.4, 37°C). Myocytes were paced at various stimulation frequencies using field stimulation (10–20 V). SBFI was alternately excited (340 and 380 nm), and emitted epifluorescence was monitored at 510 nm (F340 and F380) and recorded (IonOptix Corp.). After washout of external dye, background fluorescence was determined. The ratio F340/F380 was calculated from the background-subtracted emission intensities and converted to [Na]i. In situ calibration of SBFI was accomplished as described previously (39).

Electrophysiological studies in vivo. Electrophysiological studies in vivo were performed similarly to those in a previously described study (40). Mice were anesthetized with isoflurane and intubated endotracheally. Animals were placed on a thermoregulated table (37°C) and ventilated with isoflurane (2%) vaporized in 100% O2 using positive pressure ventilation (6 μl/g tidal volume, 140 breaths/min). Subcutaneous 27-gauge needles were placed in all limbs for ECG recordings. Resting ECGs were analyzed for RR interval, heart rate QTc (41), QRS width, and PR interval. After measurement of resting ECGs, the thorax was carefully entered by an abdominal access through the diaphragm to avoid cardiac injury, and epicardial electrodes were attached to the ventricles. Standard pacing protocols were used to determine electrophysiologic parameters (at twice the diastolic capture threshold). Induction of arrhythmias was attempted using programmed stimulation. Ten spontaneous beats were sensed, followed by 2 consecutive premature beats (S2-S3) at 38-ms intervals. Reproducibly induced VT was defined as 2 consecutive inductions of VT with at least 4 consecutive beats of ventricular origin.

For isoproterenol experiments, animals were anesthetized using 2.5% Avertin, and an equivalent of a lead II recording at 10 kHz was performed. After stabilization of the preparation, a control recording period (2 minutes) was followed by intraperitoneal injection of isoproterenol (2 mg/kg), and ECGs were recorded for an additional 5 minutes.

MAP measurements in the intact, beating heart. MAPs were recorded from right- and left-ventricular epicardium of isolated, Langendorff-perfused hearts as described previously (15, 42). Right-ventricular septal stimulation at fix pacing cycle lengths and S2 extra stimulation for determination of effective refractory periods were performed using an octapolar catheter (CIBER MOUSE; NuMED). Recordings were digitized at 2 KHz and analyzed for MAP duration.

CaMKII inhibitors. KN93 (Seikagaku Corp.) or AIP (Sigma-Aldrich) were added to the pipette solution (final concentration 1 or 0.1 μM, respectively). For back-phosphorylation concentrations in reaction buffer were 50 or 5 μM, respectively. For measurements of [Na]i and MAP recordings, KN93 (1 μM) was added to the Tyrode solution and perfusate, respectively. Wash-in–washout experiments were performed in the isolated heart.

Statistics. All data are expressed as mean ± SEM. To study Ifast, INa decay (first 50 ms) was fitted with a double exponential function. For late component of INa decay analysis, recordings at –20 mV (1,000-ms duration) were leakage subtracted using the remaining current after 1,000 ms of depolarization, followed by normalization to peak current. The current integral of the resulting current between 50 and 500 ms was calculated.

Fits were tested for significant difference using F tests. For longitudinal data, 2-way repeated-measures ANOVA was run; otherwise, Student’s unpaired t test was used. For MAP recordings only, Student’s paired t test was used. Two-sided P < 0.05 was considered significant.

    Acknowledgments

This work was funded by the Deutsche Forschungsgemeinschaft (DFG) through an Emmy-Noether grant (MA1982/1-4) and Klinische Forschergruppe grant (MA1982/2-1) to L.S. Maier; DFG SFB 656/A5 (to P. Kirchhof) and C3 (to L. Fabritz) and DFG MA2252 (to S.K.G. Maier); Interdisziplin?res Zentrum für Klinische Forschung Münster (ZPG/4a to P. Kirchhof); and NIH grants HL64724 and HL80101 (to D.M. Bers). We acknowledge the technical assistance of A. Steen, G. Müller, M. Kothe, M. Kohlhaas, and E. Hacker.

References

Bennett P., Yazawa K., Makita N., George A. 1995. Molecular mechanism for an inherited cardiac arrhythmia. Nature. 376:683–685.

Wang D.W., Makita N., Kitabatake A., Balser J.R., George A.L. 2000. Enhanced Na+ channel intermediate inactivation in Brugada syndrome. Circ. Res. 87:E37–E43.

Bers, D.M. 2001. Excitation-contraction coupling and cardiac contractile force. Kluwer Academic Publishers. Dordrecht, The Netherlands. 427 pp.

Tomaselli G.F. and Zipes D.P. 2004. What causes sudden death in heart failure? Circ. Res. 95:754–763.

Undrovinas A.I., Maltsev V.A., Sabbah H.N. 1999. Repolarization abnormalities in cardiomyocytes of dogs with chronic heart failure: role of sustained inward current. Cell. Mol. Life Sci. 55:494–505.

Despa S., Islam M.A., Weber C.R., Pogwizd S.M., Bers D.M. 2002. Intracellular Na+ concentration is elevated in heart failure but Na/K pump function is unchanged. Circulation. 105:2543–2548.

Tan H., et al. 2002. A calcium sensor in the sodium channel modulates cardiac excitability. Nature. 415:442–447.

Maier L.S. and Bers D.M. 2002. Calcium, calmodulin, and calcium-calmodulin kinase II: heartbeat to heartbeat and beyond. J. Mol. Cell. Cardiol. 34:919–939.

Kirchhefer U., Schmitz W., Scholz H., Neumann J. 1999. Activity of cAMP-dependent protein kinase and Ca2+/calmodulin-dependent protein kinase in failing and nonfailing human hearts. Cardiovasc. Res. 42:254–261.

Hoch B., Meyer R., Hetzer R., Krause E.-G., Karczewski P. 1999. Identification and expression of -isoforms of the multifunctional Ca2+/calmodulin-dependent protein kinase in failing and nonfailing human myocardium. Circ. Res. 84:713–721.

Ai X., Curran J.W., Shannon T.R., Bers D.M., Pogwizd S.M. 2005. Ca2+/calmodulin-dependent protein kinase modulates cardiac ryanodine receptor phosphorylation and sarcoplasmic reticulum Ca2+ leak in heart failure. Circ. Res. 97:1314–1322.

Zhang T., et al. 2003. The C isoform of CaMKII is activated in cardiac hypertrophy and induces dilated cardiomyopathy and heart failure. Circ. Res. 92:912–919.

Maier L.S., et al. 2003. Transgenic CaMKIIC overexpression uniquely alters cardiac myocyte Ca2+ handling: reduced SR Ca2+ load and activated SR Ca2+ release. Circ. Res. 92:904–911.

Zhang R., et al. 2005. Calmodulin kinase II inhibition protects against structural heart disease. Nat. Med. 11:409–417.

Wu Y., et al. 2002. Calmodulin kinase II and arrhythmias in a mouse model of cardiac hypertrophy. Circulation. 106:1288–1293.

Khoo M.S., et al. 2005. Calmodulin kinase II activity is required for normal atrioventricular nodal conduction. Heart Rhythm. 2:634–640.

Koller B.S., Karasik P.E., Solomon A.J., Franz M.R. 1995. Relation between repolarization and refractoriness during programmed electrical stimulation in the human right ventricle. Implications for ventricular tachycardia induction. Circulation. 91:2378–2384.

Veldkamp M.W., et al. 2000. Two distinct congenital arrhythmias evoked by a multidysfunctional Na+ channel. Circ. Res. 86:91e–97e.

Bezzina C., et al. 1999. A single Na+ channel mutation causing both long-QT and Brugada syndromes. Circ. Res. 85:1206–1213.

Wagner S. and Maier L.S. 2006. Modulation of cardiac Na+ and Ca2+ currents by CaM and CaMKII. J. Cardiovasc. Electrophysiol. 17(Suppl. 1):S26–S33.

Deschênes I., et al. 2002. Isoform-specific modulation of voltage-gated Na+ channels by calmodulin. Circ. Res. 90:49e–57e.

Kim J., et al. 2004. Calmodulin mediates Ca2+ sensitivity of sodium channels. J. Biol. Chem. 279:45004–45012.

Abriel H. and Kass R.S. 2005. Regulation of the voltage-gated cardiac sodium channel Nav1.5 by interacting proteins. Trends Cardiovasc. Med. 15:35–40.

K??b S., et al. 1996. Ionic mechanism of action potential prolongation in ventricular myocytes from dogs with pacing-induced heart failure. Circ. Res. 78:262–273.

Pieske B., et al. 2002. Rate dependence of i and contractility in nonfailing and failing human myocardium. Circulation. 106:447–453.

Attwell D., Cohen I., Eisner D., Ohba M., Ojeda C. 1979. The steady state TTX-sensitive ("window") sodium current in cardiac Purkinje fibres. Pflügers Arch. 379:137–142.

Marx S.O., et al. 2000. PKA phosphorylation dissociates FKBP12.6 from the calcium release channel (ryanodine receptor): defective regulation in failing hearts. Cell. 101:365–376.

Maltsev V.A., et al. 1998. Novel, ultraslow inactivating sodium current in human ventricular cardiomyocytes. Circulation. 98:2545–2552.

Nilius B. 1988. Modal gating behavior of cardiac sodium channels in cell-free membrane patches. Biophys. J. 53:857–862.

Saint D.A., Ju Y.K., Gage P.W. 1992. A persistent sodium current in rat ventricular myocytes. J. Physiol. 453:219–231.

Wang D.W., Yazawa K., George A.L., Bennett P.B. 1996. Characterization of human cardiac Na+ channel mutations in the congenital long QT syndrome. Proc. Natl. Acad. Sci. U. S. A. 93:13200–13205.

Clancy C.E. and Rudy Y. 2002. Na+ channel mutation that causes both Brugada and long-QT syndrome phenotypes: a simulation study of mechanism. Circulation. 105:1208–1213.

Cormier J.W., Rivolta I., Tateyama M., Yang A.-S., and Kass R.S. 2002. Secondary structure of the human cardiac Na+ channel C terminus. Evidence for a role of helical structures in modulation of channel inactivation. J. Biol. Chem. 277:9233–9241.

Kohlhaas M., et al. 2006. Increased sarcoplasmic reticulum calcium leak but unaltered contractility by acute CaMKII overexpression in isolated rabbit cardiac myocytes. Circ. Res. 98:235–244.

Guo T., Zhang T., Mestril R., Bers D.M. 2006. Ca2+/Calmodulin-dependent protein kinase II phosphorylation of ryanodine receptor does affect calcium sparks in mouse ventricular myocytes. Circ. Res. 99:398–406.

Wagner S., et al. 2003. Na+-Ca2+ exchanger overexpression predisposes to reactive oxygen species-induced injury. Cardiovasc. Res. 60:404–412.

Zhu W.Z., et al. 2003. Linkage of ?1-adrenergic stimulation to apoptotic heart cell death through protein kinase A-independent activation of Ca2+/calmodulin kinase II. J. Clin. Invest. 111:617–625 doi: 10.1172/JCI200316326.

Song Y., Shryock J.C., Wagner S., Belardinelli L., Maier L.S. 2006. Blocking late sodium current reduces hydrogen peroxide-induced arrhythmogenic activity and contractile dysfunction. J. Pharmacol. Exp. Ther. 318:214–222.

Maier L.S., Pieske B., Allen D.G. 1997. Influence of stimulation frequency on i and contractile function in Langendorff-perfused rat heart. Am. J. Physiol. 273:H1246–H1254.

Wehrens X.H., et al. 2004. Protection from cardiac arrhythmia through ryanodine receptor-stabilizing protein calstabin2. Science. 304:292–296.

Mitchell G.F., Jeron A., Koren G. 1998. Measurement of heart rate and Q-T interval in the conscious mouse. Am. J. Physiol. 274:H747–H751.

Fabritz L., et al. 2003. Effect of pacing and mexiletine on action potential duration, dispersion of repolarisation, afterdepolarisations, and ventricular arrhythmias in hearts of SCN5A delta-KPQ (LQT3) mice. Cardiovasc. Res. 57:1085–1093.