Facilitation and Ca2+-dependent Inactivation Are Modified by Mutation of the Cav1.2 Channel IQ Motif*

The heart muscle responds to physiological needs with a short-term modulation of cardiac contractility. This process is determined mainly by properties of the cardiac L-type Ca2+ channel (Cav1.2), including facilitation and Ca2+-dependent inactivation (CDI). Both facilitation and CDI involve the interaction of calmodulin with the IQ motif of the Cav1.2 channel, especially with Ile-1624. To verify this hypothesis, we created a mouse line in which Ile-1624 was mutated to Glu (Cav1.2I1624E mice). Homozygous Cav1.2I1624E mice were not viable. Therefore, we inactivated the floxed Cav1.2 gene of heterozygous Cav1.2I1624E mice by the α-myosin heavy chain-MerCreMer system. The resulting I/E mice were studied at day 10 after treatment with tamoxifen. Electrophysiological recordings in ventricular cardiomyocytes revealed a reduced Cav1.2 current (ICa) density in I/E mice. Steady-state inactivation and recovery from inactivation were modified in I/E versus control mice. In addition, voltage-dependent facilitation was almost abolished in I/E mice. The time course of ICa inactivation in I/E mice was not influenced by the use of Ba2+ as a charge carrier. Using 1,2-bis(o-aminophenoxy)ethane-N,N,N′,N′-tetraacetic acid as a chelating agent for intracellular Ca2+, inactivation of ICa was slowed down in control but not I/E mice. The results show that the I/E mutation abolishes Ca2+/calmodulin-dependent regulation of Cav1.2. The Cav1.2I1624E mutation transforms the channel to a phenotype mimicking CDI.

The heart muscle adapts to physiological needs with an autonomic modulation of cardiac contractility. This process is determined mainly by the intrinsic properties of the cardiac L-type Ca 2ϩ channel (Ca v 1.2). These properties include Ca 2ϩdependent facilitation (CDF) 3 and Ca 2ϩ -dependent inactivation (CDI). CDF serves to potentiate Ca 2ϩ influx through Ca v 1.2 channels during repeated activity, which contributes to increased force-frequency relationship of the heart muscle during exercise (1,2). CDI terminates Ca 2ϩ entry after opening of the channel to avoid Ca 2ϩ overload and arrhythmias (3,4). Both types of regulation involve, as a key step, direct binding of the Ca 2ϩ sensor protein calmodulin (CaM) to the Ca v 1.2 channel (4 -7), alongside with the action of CaM on Ca 2ϩ /CaM-dependent protein kinase II (CaMKII) (2, 6, 8 -11).
Strong evidence exists that CaM binds to the IQ motif of the Ca v 1.2 channel that is located at amino acids 1624 -1635 of the Ca v 1.2 C terminus (5,6,12,13). Especially Ile-1624 determines the amount of CaM binding. For example, the mutation I1624E decreases the affinity of the IQ sequence for CaM by ϳ100-fold (6,11). In addition, the I/E mutation abrogates CDF and CDI of L-type Ca 2ϩ currents expressed in Xenopus oocytes (11).
It is unclear if the binding of CaM to the Ca v 1.2 channel is relevant for the function of the cardiac calcium channel in the behaving mouse. Therefore, we mutated Ile-1624 to Glu in the IQ motif of the murine Ca v 1.2 channel gene. The results show that this mutation is lethal. To overcome this problem, we generated mice with a conditional heart-specific I/E mutation in the Ca v 1.2 channel gene. Electrophysiological analysis of Ca v 1.2 channel currents in cardiomyocytes (CMs) from I/E mice revealed that the I/E mutation blocks CaM/CaMKII-mediated regulation of the Ca v 1.2 channel in the heart and induces a channel phenotype with permanent CDI characteristics.

EXPERIMENTAL PROCEDURES
All substances used were of the highest purity available. The Ca v 1.2-specific antibody used in this study has been described previously (14). Amino acid numbering is according to the Oryctolagus cuniculus Ca v 1.2 sequence (GenBank TM accession number X60782.1).
Generation of Mice with the I1624E Mutation-To construct the targeting vector, a 8.0-kb fragment containing exons 39 -44 of CACNA1C was isolated from 129/Svj mouse genomic DNA. The targeting vector was composed of a 1.1-kb short 5Ј-arm, a 4.5-kb fragment containing the neo-tk and loxP sequences, a 743-base fragment including the I/E mutation at position 1624, and the long a 5.5-kb long 3'-arm. All mutation procedures were carried out by site-directed PCR mutagenesis (Stratagene). The targeting construct was electroporated into R1 embryonic stem cells (129/Sv ϫ 129/Sv-CP F1) (15). Positive clones were identified by PCR and confirmed by Southern blotting. The neo-tk cassette was removed from the germ line through cre/loxP-mediated excision in a second targeting step. A positive embryonic stem cell clone was injected into C57BL/6 blastocysts, and chimeras were crossed with C57BL/6 mice. After confirmation of successful targeting by PCR and Southern blot analysis, heterozygous mice were bred with mice expressing Cre under the control of the ␣-myosin heavy chain * This work was supported by grants from the Deutsche Forschungsgemeinschaft and the Fond der Chemischen Industrie. 1 Both authors contributed equally to this work. 2  Electrophysiological Recordings-The whole cell L-type Ca 2ϩ channel current (I Ca ) was measured at 35°C. Stimulation and data acquisition were performed as described (17). Time constants of I Ca inactivation were obtained by a fit from the peak current to the current value at the end of the voltage pulse by a two-exponential function using pCLAMP 9 (Molecular Devices). All fits showed a correlation coefficient Ͼ0.98.
Statistics-Data plotting and statistical analysis were carried out using Prism 5 (GraphPad Software). The null hypothesis was rejected if p was Ͻ0.05. Data are presented as means Ϯ S.E.

RESULTS
Ile-1624 of the CACNA1C gene was mutated to glutamate using transgenic gene knock-in techniques (Fig. 1A). The resulting homozygous mice (genotype Ca v 1.2 I1624E on both alleles) were not viable. Therefore, we crossbred heterozygous Ca v 1.2 I1624E mice with mice expressing the floxed Ca v 1.2 gene and ␣-myosin heavy chain-MerCreMer (16,18), allowing tissue-and time-dependent inactivation of the Cav1.2 gene by the tamoxifen-controlled ␣-myosin heavy chain-MerCreMer recombinase. The mutation in the resulting I/E mice (genotype Ca v 1.2 Ϫ/I1624E ) was confirmed by genomic sequencing (Fig.  1B). I/E mice had a reduced life span and died within 3 weeks after treatment with tamoxifen ( Fig. 1C). Western blot analysis of cardiac muscle using anti-Ca v 1.2 antibody detected reduced protein levels in the ventricles of I/E mice compared with littermatched Ctr mice (genotype Ca v 1.2 Ϫ/ϩ ) at day 10 ( Fig. 1D). Reduced expression of the Ca v 1.2 I1624E cDNA was confirmed in To test the physiological significance of the mutation of Ile to Glu in the cardiac Ca v 1.2 channel, I Ca was recorded in isolated ventricular CMs from I/E and Ctr mice using the patch-clamp technique ( Fig. 2A). Cell capacitance was similar in CMs from I/E and Ctr mice (221 Ϯ 14 (n ϭ 52) and 213 Ϯ 10 (n ϭ 60) picofarads, respectively). The current-voltage relation of I Ca was significantly reduced in CMs from I/E mice versus CMs from Ctr mice (Fig. 2B). These findings indicate expression of Ca v 1.2 I1624E channels in the membrane of CMs from I/E mice and an even functional availability of these channels.
CaM regulates CDF and CDI of Ca v 1.2 (4 -7). Facilitation is mainly due to activation of multifunctional CaMKII (8,10,17,19). Inhibition of CaMKII has been shown to influence recovery from inactivation and steady-state inactivation of I Ca (10). In accordance with these studies, recovery from inactivation of I Ca was slowed down in CMs from I/E mice compared with CMs from Ctr mice (Fig. 2C). Fits of the recovery from inactivation by one-exponential functions revealed time constants of 54 ms (Ctr) and 72 ms (I/E). Likewise, the steady-state inactivation curve of I Ca was shifted to the left in CMs from I/E mice compared with CMs from Ctr mice (Fig. 2D). Using the Boltzmann equation, the voltage for half-maximal inactivation was calculated to be Ϫ16 mV (Ctr) and Ϫ27 mV (I/E). KN-93 (1 M) did not influence recovery from inactivation and steady-state inactivation in CMs from I/E mice. These results indicate that the I/E muta-FIGURE 2. Analysis of L-type I Ca in ventricular CMs. A, original recordings of I Ca in a CM from a Ctr mouse and from an I/E mouse. CMs were stimulated by the voltage protocol depicted at 0.2 Hz. Current traces and voltage protocols are superimposed. A prepulse from Ϫ80 to Ϫ40 mV was used to inactivate fast Na ϩ currents. Traces are corrected for cell capacitance. pF, picofarad. B, current-voltage relation of I Ca . Peak current density is plotted against the voltage pulse. Data points represent means Ϯ S.E. with n ϭ 60 for Ctr and n ϭ 52 for I/E mice. Data sets from Ctr and I/E mice were statistically different as revealed by two-way analysis of variance (repeated measurement design, p Ͻ 0.05). C, recovery from inactivation of I Ca . CMs were stimulated by the twin-pulse protocol depicted at 0.03 Hz. Fractions of current (I 2 /I 1 ) are plotted against the interval duration. Data points represent means Ϯ S.E. with n ϭ 13 for Ctr, n ϭ 14 for I/E mice, and n ϭ 3 for I/E mice with KN-93 (1 M). Data sets from Ctr and I/E mice were statistically different as revealed by two-way analysis of variance (repeated measurement design, p Ͻ 0.01). D, steady-state inactivation of I Ca . CMs were stimulated by the twin-pulse protocol depicted at 0.03 Hz. Fractions of current (I 2 /I 1 ) are plotted against the prepulse voltage. Data points represent means Ϯ S.E. E 0.5 was calculated to be Ϫ16 Ϯ 1 mV (n ϭ 11) for Ctr mice and Ϫ27 Ϯ 4 mV (n ϭ 17) for I/E mice after fitting the data sets with a Boltzmann equation. Some data points from I/E mice (n ϭ 3) were obtained in the presence of KN-93 (1 M). Data sets from Ctr and I/E mice were statistically different as revealed by two-way analysis of variance (repeated measurement design, p Ͻ 0.01). tion in the Ca v 1.2 channel mimics the effects of CaMKII inhibitors on Ca v 1.2 channel properties (10).
Several studies have shown that facilitation of Ca v 1.2 currents depends on CaM/CaMKII (6, 9, 17, 19 -21). Facilitation of I Ca was almost abolished in CMs from I/E mice (Fig. 3, A-C). As suggested previously (9,17,20,21), facilitation depended on CaMKII. Inhibition of CaMKII activity by KN-93 (1 M) abolished facilitation in CMs from Ctr mice but had no effect in CMs from I/E mice, supporting a specific effect of KN-93 in Ctr CMs (Fig. 3D). Taken together, these results support the notion that the I/E mutation of the Ca v 1.2 channel abolishes CaM/ CaMKII-mediated effects on facilitation.
In general, CDI of I Ca is attenuated by the use of Ba 2ϩ as a charge carrier or by high concentrations of intracellular Ca 2ϩ buffers (22). Consequently, we recorded current through L-type Ca 2ϩ channels in the same CMs using Ca 2ϩ and Ba 2ϩ as the charge carrier. As expected, the fast component of inactivation observed with Ca 2ϩ was slowed down with Ba 2ϩ as the charge carrier, resulting in poorly inactivating currents in CMs from Ctr mice (Fig. 4, A and C). In contrast, a fast component of inactivation was still present in CMs from I/E mice with both Ca 2ϩ and Ba 2ϩ as the charge carrier (Fig. 4, B and C). Next, we compared the effects of buffering intracellular Ca 2ϩ by the Ca 2ϩ chelators EGTA and BAPTA. BAPTA has been shown to bind Ca 2ϩ more efficiently than EGTA, thus attenuating CDI of I Ca (23). Indeed, inactivation of I Ca was slowed down in BAPTAversus EGTA-dialyzed CMs from Ctr mice (Fig. 4, D  and F). However, inactivation of I Ca was not slowed down in BAPTAversus EGTA-dialyzed CMs from I/E mice (Fig. 4, E  and F), in which the fast component of inactivation was even faster in BAPTAversus EGTA-dialyzed CMs. Slow compo-nents of inactivation were not different in CMs from Ctr and I/E mice. These results suggest that the mutation of Ile to Glu at position 1624 of the Ca v 1.2 channel abolishes the effects of Ca 2ϩ on inactivation of I Ca , most likely because the channel has already been transformed to a phenotype mimicking CDI.

DISCUSSION
In this study, we have shown that exchange of Ile with Glu in the CaM-binding motif (IQ) of the Ca v 1.2 channel gene is lethal  to mice. Electrophysiological analysis of CMs from mice with a conditional heart-specific I/E mutation in the Ca v 1.2 gene revealed that the mutation abolishes CaM/CaMKII-mediated regulation of the Ca v 1.2 channel. In addition, the mutation transforms the Ca v 1.2 channel to a phenotype that recapitulates the properties of a Ca 2ϩ -inactivated channel.
In heart muscle, Ca v 1.2 is strongly associated with a number of regulator proteins, building up a macromolecular signaling complex (24,25). Among the association partners, CaM is permanently bound to the channel and acts as a resident Ca 2ϩ sensor (26,27). Ca 2ϩ -bound CaM regulates both CDI and CDF of Ca v 1.2 (5), the latter by regulating the activity of CaMKII that is tethered to the Ca v 1.2 channel (21,28). The major binding site for CaM is the IQ motif (amino acids 1624 -1635) located in the C-terminal tail of the Ca v 1.2 channel (5,7). Mutations in the IQ motif have been shown to inhibit CaM binding to the Ca v 1.2 channel, thus reducing facilitation and CDI (11). Especially exchange of Ile-1624 with Glu in the IQ motif of the Ca v 1.2 channel reduces CaM affinity by ϳ100-fold and prohibits effectively facilitation and CDI of I Ca in the Xenopus oocyte expression system (11). This work clearly demonstrated that exchange of Ile-1624 with Glu in the cardiac murine Ca v 1.2 channel gene likewise altered the electrophysiological properties of I Ca in CMs and reduced the life span of the mutant mice.
Experiments using peptides containing the entire IQ motif of Ca v 1.2 or the I/E mutation showed an ϳ100-fold decreased affinity of the I/E mutation for CaM (11). However, no in vivo quantitative measurements of affinity changes are available for the full-length channel with this mutation. Therefore, this number must be viewed with caution. The I/A mutant, which has as strong an effect on CDI as the I/E mutant but leaves CDF intact in heterologous expression studies, showed no measurable changes in its association with CaM, as shown by both biochemical studies (11) and crystal structure (29). Therefore, one cannot rule out a possibility that the effects of the I/E mutation also result from some distortion in the structure and correspondingly in the function of the IQ domain and not only from a reduction in CaM binding.
CaMKII is a major modulator of I Ca activity (2). Inhibition of CaMKII by inhibitory peptides or blockers such as KN-93 prolongs recovery from inactivation (10,30), shifts the steady-state inactivation curve to more negative voltages (10,31), and reduces facilitation of I Ca (17,28,30). In addition, knock-out of CaMKII␦ slows down recovery from inactivation and reduces facilitation of I Ca (9). The I/E mutation of the Ca v 1.2 channel likewise prolonged recovery from inactivation, shifted the steady-state inactivation curve to more negative voltages, and reduced facilitation of I Ca . Thus, we conclude that the I/E mutation abolishes the effects of CaMKII on the Ca v 1.2 channel because the I/E mutation shows a reduced affinity for CaM (11), preventing activation of CaMKII. At present, we cannot rule out the possibility that the I/E mutation distorts the C terminus of Ca v 1.2 in vivo and thereby reduces the affinity for CaMKII.
The fundamental role of CaM in mediating CDI has been discussed in several excellent reviews (4,5,20,32). Unfortunately, pharmacological inhibitors of CaM are not useful to characterize the role of CaM in regulating cardiac I Ca (33). Instead, the role of CaM in cardiac I Ca is assessed mainly by reducing intracellular [Ca 2ϩ ], namely by the use of Ba 2ϩ as the charge carrier for currents through Ca v 1.2 channels, by the use of high concentrations of intracellular Ca 2ϩ buffers, or the replacement of intracellular CaM with a CaM that does not bind Ca 2ϩ (7). Each experimental condition attenuates CDI, as has been observed in part in this study using CMs from Ctr mice. In contrast, CDI was no longer observed in CMs from I/E mice. Instead, inactivation of I Ca in I/E CMs was not different from that in Ctr CMs under all conditions tested. Thus, although the I/E mutation decreases significantly the affinity of the IQ motif for CaM, the mutant channel inactivates in a way that recapitulates the binding of a fully activated CaM.
In this study, we have shown that adult mice carrying the I/E mutation in the cardiac Ca v 1.2 channel gene are not viable. Preliminary experiments suggest that these mice develop dilated cardiomyopathy, in concert with a reduced contractility at an unchanged heart rate. 4 At present, we can only speculate about the reasons for this phenotype. One reason may be that the I/E channel inactivates fast and thereby decreases the amount of Ca 2ϩ entry. This lack of Ca 2ϩ is not compensated and decreases cardiac contraction, which increases sympathetic tone and initiates cardiac dilation.
Another reason may be that a hindered association of CaM with the Ca v 1.2 channel reduces trafficking of the channel to the membrane during biosynthesis, as shown for cultured neurons (34), which could account for the observed reduction in channel expression and in contractility. A further reason may be the missing facilitation due to the absence of CaMKII-mediated regulation of I Ca , which may reduce the ability of heart muscle to adapt to exercise. Indeed, mice deficient in CaMKII␦ show a reduced heart rate in response to work load or ␤-adrenergic stimulation (9). This phenotype, together with a fully inactivated I Ca , may limit Ca 2ϩ entry and thus Ca 2ϩ -induced Ca 2ϩ release, leading to an insufficient contraction and finally to death of the mice. In conclusion, the mutation of Ile to Glu at position 1624 of the Ca v 1.2 channel abolishes CaM/CaMKIIdependent regulation of I Ca but simultaneously transforms the channel to a phenotype mimicking CDI.