L-type Ca2+ currents conducted by Cav1.2 channels initiate excitation–contraction coupling in cardiac myocytes. Intracellular Mg2+ (Mgi) inhibits the ionic current of Cav1.2 channels. Because Mgi is altered in ischemia and heart failure, its regulation of Cav1.2 channels is important in understanding cardiac pathophysiology. Here, we studied the effects of Mgi on voltage-dependent inactivation (VDI) of Cav1.2 channels using Na+ as permeant ion to eliminate the effects of permeant divalent cations that engage the Ca2+-dependent inactivation process. We confirmed that increased Mgi reduces peak ionic currents and increases VDI of Cav1.2 channels in ventricular myocytes and in transfected cells when measured with Na+ as permeant ion. The increased rate and extent of VDI caused by increased Mgi were substantially reduced by mutations of a cation-binding residue in the proximal C-terminal EF-hand, consistent with the conclusion that both reduction of peak currents and enhancement of VDI result from the binding of Mgi to the EF-hand (KD ≈ 0.9 mM) near the resting level of Mgi in ventricular myocytes. VDI was more rapid for L-type Ca2+ currents in ventricular myocytes than for Cav1.2 channels in transfected cells. Coexpression of Cavβ2b subunits and formation of an autoinhibitory complex of truncated Cav1.2 channels with noncovalently bound distal C-terminal domain (DCT) both increased VDI in transfected cells, indicating that the subunit structure of the Cav1.2 channel greatly influences its VDI. The effects of noncovalently bound DCT on peak current amplitude and VDI required Mgi binding to the proximal C-terminal EF-hand and were prevented by mutations of a key divalent cation-binding amino acid residue. Our results demonstrate cooperative regulation of peak current amplitude and VDI of Cav1.2 channels by Mgi, the proximal C-terminal EF-hand, and the DCT, and suggest that conformational changes that regulate VDI are propagated from the DCT through the proximal C-terminal EF-hand to the channel-gating mechanism.
Intracellular Mg2+ (Mgi) is not used as a signaling molecule in normal cellular function, and its concentration is thought to be nearly constant under physiological conditions. However, Mgi increases after transient ischemia in the heart (Murphy et al., 1989; Headrick and Willis, 1991) and decreases in heart failure (Haigney et al., 1998). Altered Mgi is also observed in pathophysiological conditions in the brain (Resnick et al., 2004; Mendez et al., 2005) and skeletal muscle (Resnick et al., 2004). Elucidation of the regulatory effects of Mgi under these pathophysiological conditions would be an important advance toward understanding the impairments of cell function in these disease states.
L-type Ca2+ currents initiate excitation–contraction coupling in cardiac muscle cells (Reuter, 1979; Bers, 2002). Mgi inhibits the L-type Ca2+ currents in ventricular myocytes at physiologically relevant concentrations in the range of 0.8 mM (White and Hartzell, 1988; Agus et al., 1989; Yamaoka and Seyama, 1996a; Pelzer et al., 2001; Wang et al., 2004). L-type Ca2+ currents in ventricular myocytes are conducted by CaV1.2 channels consisting of a pore-forming α11.2 subunit in association with β and α2δ subunits (Catterall, 2000). The α1 subunits are composed of four homologous domains (I–IV) with six transmembrane segments (S1–S6) and a reentrant pore loop in each. Multiple regulatory sites are located in the large C-terminal domain (De Jongh et al., 1996; Peterson et al., 1999; Zuhlke et al., 1999; Hulme et al., 2003), which is subject to in vivo proteolytic processing near its center (De Jongh et al., 1991, 1996; Hulme et al., 2005). A nearby IQ motif in the proximal C terminus is implicated in Ca2+-dependent inactivation mediated by Ca2+/calmodulin (Peterson et al., 1999; Zuhlke et al., 1999). Noncovalent interaction of the distal C terminus with the proximal C-terminal domain has an autoinhibitory effect by reducing coupling efficiency of gating charge movement to channel opening and positively shifting the voltage dependence of activation (Hulme et al., 2006). The proximal C-terminal domain also contains an EF-hand motif, a potential divalent cation-binding site and a prime candidate for mediating inhibition by Mgi.
Upon maintained depolarization, L-type calcium currents in neurons, cardiac myocytes, and other cell types inactivate by a dual mechanism dependent on both Ca2+ and voltage (Brehm and Eckert, 1978; Tillotson, 1979; Ashcroft and Stanfield, 1981; Lee et al., 1985; Nilius and Benndorf, 1986). Inactivation plays an important role in the control of the action potential duration and excitation–contraction coupling (Kleiman and Houser, 1988; Keung, 1989; Ahmmed et al., 2000). The physiological significance of voltage-dependent inactivation (VDI) is illustrated by the dramatic effects of missense mutations that impair VDI in Timothy syndrome (Splawski et al., 2004), which is characterized by prolonged QT interval, prolonged action potential duration, and severe ventricular arrhythmias in the heart, as well as by developmental abnormalities in other tissues and autism spectrum disorder in the brain (Splawski et al., 2004, 2005).
Our previous work showed that Mgi inhibits CaV1.2 channels in transfected cells in the same concentration range as in cardiac myocytes, and implicated the proximal C-terminal EF-hand motif in the reduction of peak L-type Ca2+ currents of CaV1.2 channels by Mgi (Brunet et al., 2005a,b). In the experiments described here, we have examined whether the autoinhibitory action of the distal C-terminal domain interacts functionally with the inhibition of CaV1.2 channel activity by the binding of Mgi to the EF-hand. Using Na+ as permeant ion to eliminate effects of permeant divalent cations, we confirm that Mgi reduces the amplitude and increases VDI of L-type Ca2+ currents in ventricular myocytes. With these recording methods, we find that Mgi also increases VDI of transfected CaV1.2 channels through interaction with the proximal C-terminal EF-hand. In addition, we show that an autoinhibitory complex containing the distal C-terminal domain noncovalently bound to the proximal C-terminal domain greatly enhances VDI of CaV1.2 channels in transfected tsA-201 cells. The binding of Mgi to the proximal C-terminal EF-hand is required for the distal C-terminal domain to exert its autoinhibitory effects. Our results indicate that the EF-hand with bound Mg2+ is required for inhibition of CaV1.2 channel activity by the distal C-terminal domain, and suggest cooperative regulation of channel function by the distal C-terminal domain and Mgi binding to the proximal C-terminal EF-hand.
MATERIALS AND METHODS
Ventricular myocyte isolation
Left ventricular myocytes were isolated from 8–12-wk-old female adult C57/BL6 mice as described previously (Brunet et al., 2004) and maintained at 37°C until use. All protocols were approved by the University of Washington Institutional Animal Care and Use Committee.
The C-terminal truncation of α11.2a Δ1821 and Δ1800 were generated by introducing a stop codon after amino acid residues 1,821 and 1,800, respectively, using cDNA encoding the rabbit α11.2a (Mikami et al., 1989) as template. CaV1.2 EF-hand single (D1546A/N/S/R/K) and double (E1537Q, D1546N) mutants; the mutant α11.2aΔ1800(D1546R); and the triple mutant distal1822–2171(E2103Q,E2106Q,D2110Q), abbreviated EED-QQQ, were constructed using PCR overlap extension (Brunet et al., 2005b; Hulme et al., 2006). The mutant sequence, orientation, and reading frame of all constructs were confirmed by DNA sequencing.
TsA-201 cells were grown to 80% confluence and transfected with an equimolar ratio of cDNA encoding full-length, truncated, or mutant α1.2a, CaVβ (CaVβ1b, CaVβ2a, or CaVβ2b), CaVα2δ1, and CD8 as a cell surface marker (EBO-pCD-Leu2; American Type Culture Collection) using Fugene (Roche). 15–24 h after transfection, cells were suspended, plated at low density in 35-mm dishes, and incubated at 37°C in 10% CO2 for at least 17 h before recording using the whole cell configuration of the patch clamp technique. Transiently transected cells were visualized with latex beads conjugated to an anti-CD8 antibody (Invitrogen).
Patch pipettes (2.5–3.5 MΩ) were pulled from micropipette glass (VWR Scientific) and fire-polished. Currents were recorded with an Axopatch 200B amplifier (MDS Analytical Technologies) and sampled at 5 kHz after anti-alias filtering at 2 kHz. Data acquisition and command potentials were controlled by Pulse (Pulse 8.50; HEKA), and data were stored for off-line analysis. Voltage protocols were delivered at 10-s intervals unless otherwise noted, and leak and capacitive transients were subtracted using a P/4 protocol. Approximately 80% of series resistance was compensated with the voltage clamp circuitry.
For whole cell voltage clamp recordings of CaV1.2 current in tsA-201 cells with Ba2+ as a charge carrier (ICaV1.2(Ba)), the extracellular bath solution contained (in mM): 10 BaCl2, 140 Tris, 2 MgCl2, and 10 d-glucose, titrated to pH 7.3 with MeSO4. The normal intracellular Mg (0.8 mM free Mgi) solution contained (in mM): 130 N-methyl-d-glucamine, 60 HEPES, 5 MgATP, 1 MgCl2, and 10 EGTA, titrated to pH 7.4 with MeSO4. For coupling ratio determination, the intracellular solution contained (in mM): 130 CsCl, 10 HEPES, 4 MgATP, 1 MgCl2, and 10 EGTA titrated to pH 7.3 with CsOH (Hulme et al., 2006). When Na+ was used as charge carrier (ICaV1.2(Na)), the extracellular solution contained (in mM): 150 NaCl, 10 HEPES, 0.2 MgCl2, 0.25 µM EDTA, and 10 d-glucose titrated to pH 7.3 with MeSO4. The normal intracellular Mg (0.8 mM free Mg) solution contained (in mM): 150 CsOH, 110 glutamate, 20 HCl, 10 HEPES, 5 MgATP, 1 MgCl2, and 10 EGTA titrated to pH 7.6 with CsOH (Ferreira et al., 1997).
For whole cell voltage clamp recordings of ventricular myocyte L-type Ca2+ currents with Ca2+ as charge carrier (ICa,L(Ca)), the extracellular solution contained (in mM): 1.8 CaCl2, 140 TEA-Cl, 2 MgCl2, 10 d-glucose, and 10 HEPES, pH 7.3 with CsOH. The normal Mg intracellular solution (0.8 mM Mgi) contained (in mM): 100 CsCl, 20 TEA-Cl, 10 EGTA, 10 HEPES, 5 MgATP, and 1 MgCl2 titrated to pH 7.3 with CsOH. When Na+ was the charge carrier ICa,L(Na), the extracellular recording solution contained (in mM): 100 NaCl, 40 TEA-Cl, 0.7 MgCl2, 5 CsCl, 10 HEPES, 10 d-glucose, and 0.250 µM EDTA, pH 7.3 with CsOH. The intracellular solution (0.8 mM Mgi) contained (in mM): 110 CsOH, 40 TEA-Cl, 110 glutamate, 20 HCl, 1 MgCl2, 10 HEPES, 5 MgATP, and 10 EGTA titrated to pH 7.6 with CsOH (Ferreira et al., 1997). The Mgi concentration was altered by changing the amount of MgCl2 added to the intracellular solution. Free Mgi was calculated by the MaxChelator program (Bers, 2002).
Voltage clamp data were compiled and analyzed using IGOR Pro (WaveMetrics Inc.) and Excel (Microsoft). Peak currents were measured during 300-ms (for L-type Ca2+ current) or 1,000-ms (for CaV1.2 current) depolarization to potentials between −50 and 70 mV for L-type Ca2+ currents, and −80 to 20 mV for CaV1.2 currents. To quantify inactivation, peak currents elicited by 300- or 1,000-ms depolarizations to 0 mV were normalized to 1.0, and the fraction of peak current remaining at the end of the voltage pulse (r300 or r1000) was measured. For steady-state inactivation parameters, tsA-201 cells were depolarized from a holding potential (HP) of −80 mV for 4 s to membrane potentials from −80 to 20 mV in 10-mV increments. Na+ currents were then elicited by a 30-ms depolarization to 30 mV, followed by repolarization to −40 mV to measure tail currents. Pulses were applied every 30 s. L-type Ca2+ current density (pA/pF) was defined as the peak current elicited by the voltage depolarization normalized to the whole cell membrane capacitance (within the same myocyte).
All data are presented as mean ± SEM. The statistical significance of differences between the various experimental groups was evaluated using the Student’s t test or one-way ANOVA, followed by the Newman-Keuls post-test; p-values are presented in the text.
VDI of CaV1.2 channels
In our previous studies, we found that increased Mgi reduces peak Ba2+ currents conducted by CaV1.2 channels expressed in tsA-201 cells (Brunet et al., 2005b). To determine whether Mgi modulates VDI of cloned CaV1.2 channels, we initially examined the effect of Mgi on the inactivation properties of CaV1.2 channels expressed in tsA-201 cells using Ba2+ as the charge carrier. This approach reduces Ca2+-dependent inactivation because Ba2+ does not bind with high affinity to calmodulin, which mediates Ca2+-dependent inactivation by binding to an IQ motif in the C-terminal domain (Peterson et al., 1999; Zuhlke et al., 1999). Ba2+ currents inactivated slowly as expected (Fig. 1, A and B). However, the rate of inactivation depended on the size of the Ba2+ current (Fig. 1, A and B). Using the ratio of inward current at the end of a 1,000-ms test depolarization to the peak current (r1000) as an index of inactivation, we found a significant increase in inactivation with larger peak Ba2+ current amplitude (r = −0.61; P < 0.001) (Fig. 1, A and B). Because of this Ba2+-dependent effect on inactivation, we could not use Ba2+ as a charge carrier to examine the effect(s) of Mgi on VDI of CaV1.2 channels unambiguously. To avoid such effects of permeant divalent cations, we used Na+ as charge carrier to measure VDI independently of divalent cation–dependent inactivation.
When extracellular Ca2+ is reduced below micromolar level, CaV1.1 and CaV1.2 channels become permeable to monovalent cations (Almers et al., 1984; Hess and Tsien, 1984; Hadley and Hume, 1987), allowing measurements of CaV1.2 channel activity in the absence of permeant divalent cations. As anticipated, no correlation between the peak current amplitude and r1000 was observed with Na+ as charge carrier (r = −0.11; P = NS) (Fig. 1, C and D). These results show that large Ba2+ currents induce cation-dependent inactivation, presumably caused by low affinity binding of Ba2+ to calmodulin. Similar effects of Ba2+ have been observed previously (Ferreira et al., 1997; Sun et al., 2000). Except where indicated, Na+ was used as charge carrier to examine the effects of Mgi on VDI of CaV1.2 channels in our subsequent experiments.
Reduction of peak L-type Ca2+ currents in ventricular myocytes by Mgi
Mgi inhibits L-type Ca2+ currents of cardiac myocytes (White and Hartzell, 1988; Yamaoka and Seyama, 1996a; Wang et al., 2004). As a baseline for our experiments, we examined the impact of changing Mgi on the L-type Ca2+ currents of mouse ventricular myocytes (Fig. 2). With Ca2+ as a charge carrier, increasing Mgi from the normal resting value of 0.8 to 2.4 mM, a pathophysiologically relevant concentration (Murphy et al., 1989), decreased mean ICa,L(Ca) density from −8.0 ± 0.6 pA/pf (n = 13) to −5.6 ± 0.3 pA/pf (n = 8; P < 0.05) (Fig. 2, A and B). Further reduction was observed with 7.2 mM Mgi (Fig. 2, A and B). The reduction in current density when Mgi was increased from 0.8 to 2.4 mM was greater with Na+ as charge carrier (61 ± 5%; n = 6) compared with Ca2+ as charge carrier (30 ± 3%; P < 0.001), and an additional decrease was observed at 7.2 mM Mgi (Fig. 2, C and D).
Enhancement of VDI of L-type Ca2+ currents of ventricular myocytes by Mgi
In ventricular myocytes, the rate of inactivation of ICa,L(Na) currents was reduced compared with ICa,L(Ca) currents, as observed previously (Sun et al., 2000) (Fig. 3 A). The value of r300 for ICa,L(Na) decreased from 0.44 ± 0.02 (n = 9) with 0.8 mM Mgi to 0.30 ± 0.02 (n = 13) with 2.4 mM Mgi (P < 0.01), and 7.2 mM Mgi caused a further reduction (Fig. 3, B and C). The reduction of peak CaV1.2 current (Fig. 2) plus the acceleration of VDI (Fig. 3) would work together to markedly reduce Ca2+ entry when Mgi is elevated in cardiac myocytes.
Effects of Mgi on VDI of CaV1.2 channels in tsA-201 cells
To determine whether Mgi modulates VDI of CaV1.2 channels expressed in tsA-201 cells, we measured ICa,L(Na) with a range of Mgi concentrations (0.26–7.2 mM). Increased Mgi enhanced the rate and extent of VDI of CaV1.2 current (Fig. 4 A), with an apparent Kd of 0.9 mM (Fig. 4 B). Coexpression of the calmodulin mutant CaM1234, in which mutations of all four EF-hands prevent Ca2+ binding and Ca-dependent inactivation (Peterson et al., 1999), had no effect on the measured inactivation (Fig. 4 C). These results indicate that only VDI is observed under our recording conditions.
Effects of EF-hand mutations on VDI
To determine the role of the C-terminal EF-hand in modulating VDI, we tested the EF-hand mutations D1546A/N/S/K/R, which reduce the effects of Mgi on peak Ba2+ currents compared with wild-type (WT) CaV1.2 channels (Brunet et al., 2005b). These mutations significantly reduced VDI, as assessed from the ratio of r1000 values at 0.26 and 2.4 mM Mgi (Fig. 4 D). The rank order of effects (K>R>S>N>A) was the same as previously observed for reduction of peak currents by Mgi (Brunet et al., 2005b), consistent with the conclusion that Mg2+ binds to the proximal C-terminal EF-hand and causes both effects.
To further explore the role of Mgi and the EF-hand in modulation of VDI, we studied steady-state inactivation of WT CaV1.2 or EF-hand mutant (D1546K) channels expressed in tsA-201 cells. For WT CaV1.2 channels at low Mgi concentration, steady-state inactivation was incomplete compared with higher Mgi concentrations (Fig. 5 A and Table I). Elevation of Mgi resulted in a negative shift in the voltage dependence of inactivation and an increase of maximal inactivation at positive potentials (Fig. 5 A). These effects are similar to those observed when Ba2+ was used as charge carrier for transfected CaV1.2 channels (Brunet et al., 2005b) and for myocyte L-type Ca2+ currents (Hartzell and White, 1989). Complete inactivation at depolarized potentials was observed with 2.4 and 7.2 mM Mgi in these experiments using Na+ as a charge carrier (Fig. 5 A and Table I), but not when Ba2+ was the permeant ion (Brunet et al., 2005b).
To test the role of the proximal C-terminal EF-hand in the enhancement of steady-state inactivation by Mgi, we examined the effects of Mgi on the mutant D1546K. Mgi caused a similar negative shift of the voltage dependence of inactivation for this mutant compared with WT, suggesting that the negative shift of the voltage dependence of inactivation does not require binding to the EF-hand. However, steady-state inactivation of mutant D1546K was less complete at the most positive prepulse potential (+20 mV) than WT at each concentration of Mgi (Fig. 5 B and Table I, Inon-inact). When plotted as a concentration–response curve, the dependence of the extent of inactivation at +20 mV on Mgi is shifted to higher concentrations by the mutation D1546K with a half-maximal effect at 0.78 ± 0.04 mM Mgi in WT compared with 3.6 ± 1.4 mM Mgi for the mutant, a ratio of 4.6 ± 0.39 (Fig. 5 C). This decrease in apparent affinity for Mgi in enhancing the extent of VDI caused by the mutation D1546K is similar to the decrease of 3.6-fold in the apparent affinity for Mgi in increasing the rate of VDI that is caused by the same mutation (Fig. 4 D). These results are consistent with the conclusion that both the effect of the mutation on the rate of VDI and the effect on the extent of VDI arise from the same mechanism-impaired affinity for Mgi binding to the proximal C-terminal EF-hand.
Comparison of VDI for native and transfected CaV1.2 channels
Under recording conditions with reduced Mgi (0.26 mM) and Na+ as charge carrier, the level of inactivation of ICa,L(Na) in ventricular myocytes was substantially greater than that of ICav1.2(Na) in tsA-201 cells (Fig. 6 A). At 0.26 mM Mgi, we observed a value of r300 of 0.46 ± 0.02 (n = 4) for ICa,L(Na) versus 0.87 ± 0.04 (n = 8) for CaV1.2(Na) (P < 0.0001) (Fig. 6, A and B). Several factors could contribute to the observed difference between the VDI of L-type Ca2+ current in ventricular myocytes and CaV1.2 expressed in tsA-201 cells, including differences in expressed CaVβ subunits (Colecraft et al., 2002) and/or in association with the distal C-terminal regulatory domain (DCRD) (Hulme et al., 2006; see below).
Effect of CaVβ subunits on VDI
CaVβ subunits affect the inactivation properties of CaV1.2 channels (Catterall, 2000; Colecraft et al., 2002). In agreement with previous work, we found that VDI of CaV1.2 channels is greater with the CaVβ1b subunit than with the CaVβ2a subunit in the presence of 2.4 mM Mgi (Fig. 6, C and D). It remains uncertain which CaVβ subunit is predominant in the heart (Foell et al., 2004; Pitt et al., 2006; Ter Keurs and Boyden, 2007), but increasing evidence points to the CaVβ2 family (Gao et al., 1997; Foell et al., 2004; Weissgerber et al., 2006). Previous studies of Ca2+ and Ba2+ currents suggested that CaVβ2b recapitulates the inactivation profile of L-type currents of ventricular myocytes (Colecraft et al., 2002). We tested the impact of CaVβ2b on VDI of CaV1.2 expressed in tsA-201 cells, using 0.26 mM Mgi to minimize its effect on VDI of CaV1.2 and Na+ as charge carrier to eliminate any effects of divalent ions on the inactivation process. Under these conditions, CaVβ2b significantly increased VDI compared with CaVβ1b and CaVβ2a (Fig. 6, E–H). Although the expression of CaVβ2b enhanced the inactivation of CaV1.2 (Na) currents, the level of inactivation is less than previously reported for CaVβ1b and CaVβ2b (Colecraft et al., 2002; Takahashi et al., 2003). This difference is caused by the use of 0.26 mM Mgi in our experiments because a much larger effect of the CaVβ1b subunit is observed at 2.4 mM Mgi (Fig. 6, C and D). Nevertheless, the results at 0.26 mM Mgi show that the CaVβ subunits do not have a profound effect on VDI in the presence of low Mgi when divalent cation–dependent inactivation is prevented by use of Na+ as charge carrier, and therefore indicate that CaVβ subunits cannot fully account for the difference in the rate of VDI between L-type Ca2+ currents in ventricular myocytes and CaV1.2 channels expressed in tsA-201 cells illustrated in Fig. 6 A.
Effect of the distal C terminus of CaV1.2 on VDI
The C termini of CaV1.1 and CaV1.2 channels are proteolytically processed in skeletal and cardiac muscle tissues, respectively (De Jongh et al., 1991, 1996; Gerhardstein et al., 2000; Hulme et al., 2005). The proteolytically processed distal C terminus is thought to remain tethered to the proximal C terminus via noncovalent interactions between the DCRD and the proximal C-terminal regulatory domain (PCRD) (Hulme et al., 2006). This interaction has potent autoinhibitory effects on CaV1.2 currents when the distal C terminus is expressed as a separate protein with truncated CaV1.2 channels (Hulme et al., 2006). To determine the effect of formation of this complex on VDI, we expressed the truncated CaV1.2 channel (CaV1.2Δ1821) in tsA-201 cells with and without the distal C terminus composed of amino acid residues 1,822–2,171 (distal1822–2171). There was no significant difference in inactivation between CaV1.2Δ1821/CaVβ2b (Fig. 7, A and B) and full-length CaV1.2/CaVβ2b (Fig. 6, G and H), as the r1000 was 0.54 ± 0.03 (n = 10) for CaV1.2Δ1821 and 0.54 ± 0.03 (n = 9) (P = 1.0) for full-length CaV1.2, respectively. On the other hand, the coexpression of distal1822–2171 with CaV1.2Δ1821 enhanced inactivation compared with CaV1.2Δ1821 alone (Fig. 7, A and B), with r1000 = 0.54 ± 0.03 (n = 10) for CaV1.2Δ1821 versus r1000 = 0.28 ± 0.03 (n = 21) for CaV1.2Δ1821 with distal1822–2171 (P < 0.01) (Fig. 7, B and C). VDI of CaV1.2Δ1821 with distal1822–2171 was similar to VDI of the L-type Ca2+ current in ventricular myocytes (Fig. 7 A, ICa,L). These results show that formation of the noncovalent autoinhibitory complex of the distal C-terminal domain with the truncated CaV1.2 channel greatly enhances VDI. This interaction may contribute to the more rapid rate of VDI observed in ventricular myocytes, where we estimated that at least 48% of CaV1.2 channels are in the form of proteolytically processed channels with associated C-terminal domain (Hulme et al., 2006).
Because the effect of the distal C terminus on the activity of CaV1.2 channels is mediated via an electrostatic interaction between two positively charged amino acid residues in the PCRD(RR) and three negatively charged residues (EED) in the DCRD (Hulme et al., 2006), we tested the effect of the DCRD mutant EED-QQQ(Distal1822–2171), which was shown to significantly reduce the effect of the distal C terminus on the coupling ratio of CaV1.2 channels (Hulme et al., 2006). Our results show that the EED-QQQ distal C-terminal mutant does not enhance VDI of CaV1.2Δ1821 (Fig. 7, B and C), suggesting that the electrostatic interaction between the PCRD and DCRD is important in mediating the effect of the distal C terminus on VDI of CaV1.2 channels. This effect of the C terminus was not dependent on the CaVβ subunit expressed, as similar observations were made with CaVβ1b subunit (unpublished data). These effects of distal1822–2171 on VDI of CaV1.2Δ1821 are consistent with previous results showing that distal1822–2171 reduces the current amplitude of CaV1.2Δ1821 through interaction with the EED motif in the DCRD (Hulme et al., 2006). Overall, these results suggest that the increased VDI of L-type Ca2+ current in cardiac myocytes results at least in part from the interaction with distal C terminus of CaV1.2, which enhances VDI of CaV1.2 channels through electrostatic interactions with the PCRD.
Cooperative modulation of VDI by the distal C terminus and Mgi
Because both the distal C terminus and Mgi enhance inactivation of CaV1.2 channels, we examined the impact of the distal C terminus on the Mgi modulation of VDI of truncated CaV1.2 channels. As in previous work, Mgi enhanced the inactivation of the full-length CaV1.2 channel (Fig. 8 A) (Brunet et al., 2005a), with r1000 = 0.54 ± 0.03 for 0.26 mM versus 0.15 ± 0.03 for 7.2 mM Mgi (P < 0.001). Similar to the Mgi effect on full-length channels, Mgi also enhanced VDI of CaV1.2Δ1821 (Fig. 8 B). The r1000 value was 0.54 ± 0.03 at 0.26 mM Mgi versus 0.10 ± 0.05 (P < 0.001) for 7.2 mM Mgi. In contrast, when distal1822–2171 was coexpressed with CaV1.2Δ1821 channels, 7.2 mM Mgi did not significantly enhance VDI of CaV1.2 (Fig. 8 C; 0.26 mM, r1000 = 0.28 ± 0.03; 7.2 mM, r1000 = 0.25 ± 0.03; P = 0.52). To determine whether reduced Mgi regulates this autoinhibitory complex, a lower Mgi concentration (0.1 mM) was tested. The dialysis of 0.1 mM Mgi reduced inactivation (r1000 = 0.60 ± 0.05; P < 0.01) compared with 0.26 and 7.2 mM Mgi (Fig. 8 C). This result suggests that the noncovalent interaction of the distal C terminus with the CaV1.2 channel enhances regulation of VDI by Mgi, possibly by increasing the affinity for binding of Mgi to the proximal C-terminal EF-hand.
Requirement for the EF-hand of CaV1.2 for regulation of VDI by the distal C terminus
The effect of the distal C-terminal domain on the rate of VDI can be clearly observed by comparing the rate of inactivation for CaV1.2Δ1821 without and with distal1822–2171 at 0.26 mM Mgi (Fig. 7, A and B). A similar effect on VDI is observed for truncation at position 1,800 (Fig. 9 A), the probable point of in vivo proteolytic processing determined by mass spectrometric analysis of the related CaV1.1 channel (Hulme et al., 2005), and more extensive analysis has shown that these two forms of the autoinhibitory CaV1.2 channel complex cleaved at position 1,800 or 1,821 have nearly identical functional properties when studied side-by-side (unpublished data). To determine the role of the proximal C-terminal EF-hand in the enhancement of VDI by the distal C terminus in CaV1.2Δ1800, we studied the mutation D1546R, which reduces the effects of Mgi on current amplitude and VDI of CaV1.2 (Brunet et al., 2005a,b). With EF-hand mutant CaV1.2Δ1800(D1546R) plus distal1801–2171, the distal C terminus did not enhance VDI of CaV1.2 (Fig. 9 B) in contrast to CaV1.2Δ1800 with WT distal1801–2171 (Fig. 9 A). These results indicate that interaction of Mgi with the EF-hand in the proximal C terminal is required for regulation of VDI by the distal C-terminal domain.
The distal C-terminal domain inhibits CaV1.2 channel activity by positively shifting the voltage dependence of activation and reducing the coupling ratio of gating charge movement to channel opening (Hulme et al., 2006). Using Ba2+ as a charge carrier to allow comparison of our results with the previous work, we found that the D1546R mutation in the EF-hand in the proximal C-terminal domain prevented both the reduction of the coupling ratio (Fig. 9 C) and the positive shift in the voltage dependence of activation of CaV1.2Δ1800 (Fig. 9 D). These results demonstrate that a functional EF-hand in the proximal C-terminal domain is required to mediate all of the effects of the noncovalently associated distal C terminus on CaV1.2 channels, including increased VDI, positively shifted activation, and reduced coupling ratio.
Cation-dependent inactivation and Ba2+ as charge carrier
We found that use of Ba2+ as a charge carrier resulted in a significant level of cation-dependent inactivation of CaV1.2 channels. In our transfected cell system, we observed a correlation between Ba2+ current amplitude and r1000 as an index of inactivation. This result is in agreement with previous work (Zhang et al., 1994; Ferreira et al., 1997; Sun et al., 2000), including previous studies of cation-dependent inactivation of CaV1.2 channels expressed in tsA-201 cells by Ba2+ (Ferreira et al., 1997). It was proposed that Ba2+ could bind weakly to the same sensor as Ca2+, which is now known to be calmodulin, and could activate the same inactivation mechanism when high local concentrations are achieved (Sun et al., 2000). Based on our results and this previous work, it is evident that unambiguous measurement of VDI requires use of Na+ or another monovalent cation as charge carrier, taking advantage of the ability of Ca2+ channels to conduct monovalent cations with high efficiency in the presence of low extracellular concentrations of Ca2+. Because Ca2+-dependent inactivation is much faster than VDI for CaV1.2 channels, a small contamination by this cation-dependent inactivation mechanism can have a major impact on measurements of VDI. In addition, in transfected cell systems that yield variable expression of CaV1.2 channels from different transfections and from different mutants, elimination of cation-dependent inactivation that varies with channel density is especially important. For these reasons, we used Na+ as permeant ion for most of our experiments on VDI reported here.
Mgi reduces peak amplitude and increases VDI of native cardiac L-type Ca2+ currents at physiological concentrations
Previous reports have suggested that Mgi could reduce peak L-type Ca2+ currents and enhance the VDI in ventricular myocytes, as measured with Ba2+ as the permeant ion, (Hartzell and White, 1989; Yamaoka and Seyama, 1996b; Wang et al., 2004). However, it was uncertain whether the Ba2+-dependent inactivation observed in our experiments in transfected cells might contribute substantially to those effects in ventricular myocytes. Our present results with Na+ as permeant ion clearly demonstrate that, in the absence of permeant divalent ions, increases in Mgi concentration both reduce peak Ca2+ current and enhance VDI in native ventricular myocytes.
Mgi is ∼0.8 mM in intact cardiac myocytes under physiological conditions (Murphy et al., 1989; Headrick and Willis, 1991; Haigney et al., 1998). There is a two- to threefold increase in Mgi in ventricular myocytes as ATP levels decrease during transient ischemia (Murphy et al., 1989; Headrick and Willis, 1991). Similar two- to threefold increases in Mgi are observed in cerebral ischemia (Brooks and Bachelard, 1989; Helpern et al., 1993; Williams and Smith, 1995) and in traumatic brain injury (decrease Mgi) (Vink et al., 1988; Heath and Vink, 1996). In contrast, a decrease of two- to threefold in Mgi is observed in heart failure (Haigney et al., 1998). Our results show that both the peak amplitude and VDI of L-type Ca2+ currents in ventricular myocytes would be substantially regulated by these changes in Mgi. In heart failure, reduction of Mgi would increase Ca2+ currents and slow their inactivation, which may contribute to dysregulation of Ca2+ signaling, hypertrophy, and cytotoxicity. In ischemia, the increase in Mgi would reduce L-type Ca2+ current by reducing peak current and by enhancing VDI. These effects would reduce cytotoxicity caused by Ca2+ overload under ischemic conditions. Moreover, this regulatory mechanism would be cell autonomous, reducing Ca2+ currents only in those individual cardiac myocytes that are ischemic, while leaving neighboring cells with normal ATP levels uninhibited. This unique cell-autonomous protective mechanism may contribute significantly to preventing Ca2+ overload in ischemia.
Upon maintained depolarization, the cardiac L-type calcium current inactivates by dual mechanisms dependent on Ca2+ and voltage (Lee et al., 1985; Nilius and Benndorf, 1986). Calcium-dependent inactivation accelerates the decay of the calcium current as calcium accumulates inside the cell during the action potential, and these changes in inactivation kinetics play an important role in the control of action potential duration and excitation–contraction coupling (Kleiman and Houser, 1988; Keung, 1989). However, Ca2+-dependent inactivation does not reduce the peak Ca2+ current until after Ca2+ overload has occurred and is not able by itself to reduce the Ca2+ current to zero, even at high Ca2+ levels. In contrast, VDI can reduce the peak Ca2+ current in response to sustained depolarization, even before Ca2+ overload has occurred, and it is powerful enough to reduce the Ca2+ current to zero as shown in our records. Thus, we propose that in ischemic conditions with sustained membrane depolarization, Mgi can reduce Ca2+ entry via L-type CaV1.2 channels before Ca2+ overload occurs and sustain inhibition as Ca2+ is sequestered and pumped out of the cytosol.
Effects of PKA stimulation on Mgi regulation of L-type Ca2+ channels
Previous studies have shown that Mgi modulation of the cardiac L-type Ca2+ channel depends on the state of activation of channels by cAMP-dependent phosphorylation (White and Hartzell, 1988; Wang et al., 2004; Wang and Berlin, 2006). These investigators concluded that the actions of Mgi to reduce peak currents and enhance VDI are mediated by a separate mechanism from protein phosphorylation, but the extent of modulation by Mgi is dependent on the phosphorylation state of the CaV1.2 channel. These conclusions are consistent with the work we present here, in which we find that the effects of Mgi on peak current and on VDI are caused by the binding of Mg to the C-terminal EF-hand. In other experiments (unpublished data), we found that PKA inhibitors reduced CaV1.2 current ∼50% under our basal physiological conditions; therefore, our results presented here reflect the effects of Mgi on CaV1.2 channels whose activity is partially up-regulated by basal cAMP-dependent phosphorylation.
Mgi enhances VDI of CaV1.2 channels
Our results show that altering Mgi has striking effects on VDI of CaV1.2 channels expressed in tsA-201 cells in the absence of other cardiac-specific proteins. The rate and extent of VDI increase dramatically with increased Mgi. This enhancement of inactivation occurs with an apparent Kd of 0.9 mM. CaV1.2 is the predominant Ca2+ channel in cardiac cells, and this value falls in the range of normal physiological levels of Mgi in cardiac cells (0.6–1.0 mM) (Murphy et al., 1989; Headrick and Willis, 1991; Haigney et al., 1998). Thus, the dependence of peak current amplitude and VDI of CaV1.2 channel on Mgi that we have defined in the tsA-201 cell expression system likely reflects the Mgi dependence of peak current amplitude and VDI of the CaV1.2 channel in vivo.
The proximal C-terminal EF-hand mediates the effects of Mgi on VDI
Our previous studies of the effects of Mgi on peak Ca2+ currents showed that EF-hand mutations at the −z position (D1546A/NS/K/R) reduce the inhibitory effects of Mgi because of a decrease in apparent affinity for Mgi (Brunet et al., 2005b). Our present results extend those findings to VDI and lead to the conclusion that both the reduction of peak current amplitude and the enhancement of VDI caused by Mgi result from its interaction with the proximal C-terminal EF-hand motif. This conclusion is in agreement with previous work suggesting that the structure of the EF-hand of CaV1.2 is important for VDI (Bernatchez et al., 1998). Evidently, the functional role of the EF-hand in VDI depends on its binding of Mgi as shown here.
Noncovalent association of the distal C-terminal domain increases the rate of VDI in transfected cells
Although our results describing the modulation of Mgi on L-type Ca2+ current in myocytes are qualitatively similar to our results on Mgi modulation of CaV1.2 channels in tsA-201 cells, there are some important quantitative differences. First, Mgi was more effective at enhancing the inactivation of CaV1.2 in tsA-201 cells compared with the inactivation of the L-type Ca2+ current in ventricular myocytes. Second, VDI was greater for CaV1.2 channels in cardiac myocytes than for CaV1.2 expressed in tsA-201 cells, as reported previously using Ba2+ as permeant ion (Colecraft et al., 2002). Subunit structure and composition of CaV1.2 channels in ventricular myocytes and CaV1.2 channels in tsA-201 cells may contribute to these differences.
Our results show that CaVβs have small but significant effects on the VDI of CaV1.2 measured with Na+ as the permeant ion in the presence of low Mgi. These effects of CaVβs on VDI are smaller than in previous work, and CaVβ2b did not fully restore VDI of CaV1.2 as reported previously (Colecraft et al., 2002). Our results indicate that CaVβs cannot fully account for the difference in VDI between CaV1.2 channels in ventricular myocytes and CaV1.2 channels in transfected cells in the absence of permeant divalent cations.
In myocytes, the distal C terminus of CaV1.2 is proteolytically processed in situ, whereas it is not cleaved in tsA-201 cells (De Jongh et al., 1991, 1994, 1996; Hulme et al., 2005). Our results demonstrate that expression of the distal C terminus of CaV1.2 together with the truncated CaV1.2Δ1821 channel markedly enhanced VDI under our recording conditions. This enhancement of VDI was blocked by the distal C terminus mutation EED-QQQ, supporting a role for an electrostatic interaction between the PCRD and DCDR in mediating the effect of the distal C terminus on VDI. This effect of the distal C terminus on VDI was greater than the effects of CaVβs tested, but it was comparable to the effects of Mgi. These findings support an important role of the distal C-terminal domain in regulating Ca2+ currents via control of coupling ratio, voltage-dependent activation, and VDI. These results are consistent with the autoinhibitory function demonstrated for the distal C terminus of CaV1.2 in previous studies (Hulme et al., 2006).
Mgi and the proximal C-terminal EF-hand are required for the functional effects of the noncovalently associated distal C-terminal domain
Because the effect of the distal C terminus on VDI was comparable in amplitude to that of Mgi, we investigated the possible functional interaction between Mgi and the distal C terminus. To our surprise, when the distal C terminus was expressed together with the cleaved CaV1.2 channel, increasing Mgi concentration from 0.26 to 7.2 mM had little effect. On the other hand, decreasing Mgi to lower levels reduced VDI, indicating that the apparent affinity for Mgi was increased by noncovalent association of the distal C terminus. These results demonstrate a cooperative interaction between the distal C terminus and Mgi binding to the proximal C-terminal EF-hand motif in enhancement of VDI. Consistent with this idea, the mutation D1546R reduces affinity for the binding of Mgi to the proximal C-terminal EF-hand and reduces the functional effects of the distal C terminus. Overall, these results place the EF-hand as a downstream structural element through which the distal C terminus regulates VDI of CaV1.2 channels. Evidently, Mgi enhancement of VDI is cooperatively coupled to enhancement of VDI by the distal C terminus.
A C-terminal signaling complex controlling inactivation of CaV1.2 channels
The inactivation of CaV1.2 channels is both Ca2+ and voltage dependent. Our results show that VDI of CaV1.2 channels is regulated by the distal C terminus, Mgi, and the proximal C-terminal EF-hand. At the same time, Ca2+-dependent inactivation of CaV1.2 channels is regulated by the binding of Ca2+ and calmodulin to an IQ motif just downstream from the EF-hand motif in the proximal C-terminal domain (Peterson et al., 1999; Zuhlke et al., 2000; Pitt et al., 2001; Ohrtman et al., 2008), and the effects of Ca2+/calmodulin binding to the IQ motif on CDI are blocked by mutations on the external, non-cation–binding face of the EF-hand (Peterson et al., 2000). The spatial relationship of these regulatory sites is illustrated in Fig. 10, which emphasizes that regulatory influences on both Ca2+-dependent inactivation and VDI from downstream in the C-terminal domain are mediated via the proximal C-terminal EF-hand and are potentially influenced by the binding of Mgi. We propose that conformational changes induced by noncovalent association of the distal C-terminal domain and Ca2+/calmodulin with the proximal C-terminal domain are propagated from the regulatory regions of the C terminus to the pore-lining IVS6 segment via the EF-hand motif. These results further emphasize the interactive nature of the multiple regulatory elements in the C terminus of CaV1.2 channels, and place Mgi binding to the proximal C-terminal EF-hand motif in position to serve as a key integrator of multiple regulatory signals that control Ca2+ channel function.
The authors thank Dr. Matthew Fuller (Department of Pharmacology, University of Washington) for determining the functional properties of CaV1.2aΔ1800.
The authors gratefully acknowledge the financial support provided by a National Scientist Development Grant from the American Heart Association to S. Brunet and the National Institutes of Health (P01 HL 44948 and R01 HL085372) to W.A. Catterall.
Angus C. Nairn served as editor.
Abbreviations used in this paper: DCRD, distal C-terminal regulatory domain; HP, holding potential; Mgi, intracellular Mg2+; PCRD, proximal C-terminal regulatory domain; VDI, voltage-dependent inactivation; WT, wild-type.