Cardiac-specific Overexpression of the α1 Subunit of the L-type Voltage-dependent Ca2+ Channel in Transgenic Mice

The l-type voltage-dependent calcium channel (l-VDCC) regulates calcium influx in cardiac myocytes. Activation of the β-adrenergic receptor (βAR) pathway causes phosphorylation of thel-VDCC and that in turn increases Ca2+ influx. Targeted expression of the l-VDCC α1 subunit in transgenic (Tg) mouse ventricles resulted in marked blunting of the βAR pathway. Inotropic and lusitropic responses to isoproterenol and forskolin in Tg hearts were significantly reduced. Likewise, Ca2+ current augmentation induced by iso- proterenol and forskolin was markedly depressed in Tg cardiomyocytes. Despite no change in βAR number, isoproterenol-stimulated adenylyl cyclase activity was absent in Tg membranes and NaF and forskolin responses were reduced. We postulate an important pathway for regulation of the βAR by Ca2+ channels.

␤-Adrenergic receptor (␤ 1 AR, ␤ 2 AR) activation regulates the L-VDCC by phosphorylating the ␣ 1 subunit, thereby causing an increase in Ca 2ϩ influx (7). This forward signal is via cAMPdependent protein kinase A phosphorylation of the channel (7). Other non-cAMP-dependent protein kinase-dependent mechanisms have also been proposed (8). However, there is little or no information regarding possible Ca 2ϩ -dependent regulation of the ␤AR signal transduction pathway via the L-VDCC.
In order to investigate possible reciprocal regulation between these pathways in vivo, we overexpressed the ␣ 1 subunit of the Ca 2ϩ channel in hearts of Tg mice and studied channel activity and the ␤AR-G protein cascade by physiological and biochemical techniques. We found a major change in cardiac function regulated by the ␤AR system and conclude that Ca 2ϩ derived from channel influx modules ␤AR activity.

EXPERIMENTAL PROCEDURES
Generation of Transgenic Mice-The overexpression construct was generated by ligating the full-length human L-VDCC ␣ 1C subunit (9) coding sequence to the ␣-myosin heavy chain (␣-MHC) promoter (clone 26) (10) and completed with a bovine growth hormone poly(A)-adenylation signal (Fig. 1A). The ␣-MHC-L-VDCC construct was cleaved with NotI and the ␣-MHC-L-VDCC fusion cDNA fragment was purified and eluted in oocyte injection buffer (5 mM Tris-HCl, 0.2 mM EDTA, pH 7.4). This construct (20 g) was then microinjected into the male pronucleus of fertilized zygotes from superovulated FVB/n mice and the surviving zygotes implanted into pseudopregnant foster mothers. Transgenic founder mice were identified with genomic DNA utilizing polymerase chain reactions and confirmed by restricted Southern blotting. Polymerase chain reaction was carried out with a sense primer (5Ј-cactcttagcaaacctcagg-3Ј) specific for the ␣-MHC gene (bp 859) and an antisense primer (5Ј-caatgcgaccatctccacagtc-3Ј) located at bp 1530 of the human ␣ 1 subunit yielding a 375-bp product from mice expressing the transgene. For Southern blots, genomic DNA was extracted from tail clips of 18-day-old pups, digested with EcoRI, and separated on a 0.7% agarose gel. The ␣-MHC-L-VDCC construct (100 pg, 500 pg, and 1 ng) was cleaved with EcoRI and loaded on the gel for quantitation. Following transfer to supported nitrocellulose (Hybond-C extra, Amersham Pharmacia Biotech), the DNA was probed with a 2295-bp EcoRI fragment from the fusion construct (Fig. 1A). The fragment contained ϳ1.5 kilobase pairs of the ␣-MHC promoter and 0.75 kilobase pair of the L-VDCC ␣ 1C cDNA. Radioactive bands were quantitated using PhosphorImager and ImageQuant software (Molecular Dynamics).
RNA Dot Blots of L-VDCC Subunits and Hypertrophic Markers-Total RNA was isolated from frozen heart samples of Tg and nontransgenic (Ntg) littermate controls using TriZol (Life Technologies, Inc.) according to the manufacturer's recommendations. Two g of RNA was loaded onto Hybond N ϩ membranes (Amersham Pharmacia Biotech) using a dot blot apparatus. Probing for the ␣ 1 subunit was completed with a 1350-bp fragment isolated by cleaving the ␣-MHC-L-VDCC vector with EcoRI (Fig. 1A). This fragment corresponds to bases 2098 -3448 of the human heart ␣ 1 cDNA. The full-length L-VDCC ␤ 2a cDNA was used to probe for gene expression levels of the ␤ subunit. The ␣ 2 /␦ subunit expression was analyzed using a 1550-bp fragment probe generated by cleaving the cDNA with HincII. Probing of the hypertrophic markers (atrial natriuretic factor, ␣-MHC, ␤-MHC, cardiac ␣-actin, skeletal ␣-actin, sarco-endoplasmic reticulum ATPase, and phospholamban) using gene-specific antisense oligonucleotides were completed as described previously (11). All probes were 32 P-labeled by the random priming technique. Radiolabeled dot blots were quantitated using a PhosphorImager and ImageQuant software (Molecular Dynamics) and normalized to the signal from glyceraldehyde-3-phosphate dehydrogenase (GAPDH), since evidence from other mouse models has demonstrated no change in gene expression of GAPDH in Tg and Ntg littermates (11,12). * This work was supported by National Institutes of Health Grants P01 HL22619 (to A. S.), RO1 R37HL 43231, and T32 HL 07382 (to J. M. and A. S.). The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked "advertisement" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
Work Performing Hearts-Eight-to ten-week-old Tg and Ntg littermates of either sex were anesthetized with sodium pentobarbital (intraperitoneal 30 mg/kg), protected with heparin, and placed in an isolated working heart mode as described (13). Following base line establishment, the heart was exposed to cumulative isoproterenol or forskolin concentrations and the inotropic, lusitropic (relaxation), and chronotropic responses measured.
Echocardiography-Analysis of left ventricular function and mass was assessed using an ATL HDI3000 broadband digital echo-Doppler machine as described previously (12). Sedated, spontaneously breathing, mice were studied using two-dimensional-guided M-mode echocardiography to estimate left ventricular dimensions and wall thickness. Fractional shortening was calculated as (EDD-ESD)/EDD, where EDD is end diastolic dimension and ESD is end systolic dimension. Echocardiographic measurements were recorded and analyzed blinded.
Isolation of Single Cardiomyocytes and Electrophysiological Recording-Single ventricular cells were isolated from the hearts of 8 -10week-old Tg and Ntg mice by enzymatic dissociation protocol as described previously (14) using Type I collagenase (Worthington). L-type Ca 2ϩ channel currents were recorded using the whole-cell mode of the patch clamp method (15).
␤-Adrenergic Receptor Density, Adenylyl Cyclase Assays, and cAMP-Measurement of adenylyl cyclase activity using a cAMP 125 I radioimmunoassay kit (NEN Life Science Products) and radioligand binding of the ␤AR were carried out as described (16). Powdered tissue for cAMP measurements was prepared by homogenizing in 20 mM Tris-HCl, pH 7.5, 250 mM sucrose, 1 mM EDTA, 10 g/ml aprotinin, 10 g/ml leupeptin, and 0.1 mM phenylmethylsulfonyl fluoride in a glass homogenizer. Membranes were pelleted by centrifugation at 100,000 ϫ g for 1 h at 4°C, resuspended in 10 mM Tris-HCl, pH 7.5, and stored at Ϫ70°C. The particulate fraction (50 g/ml of protein) was incubated for 10 min at 30°C with reagents as described (17), in a final volume of 0.5 ml of cAMP reaction buffer. cAMP measurements were performed using a competitive protein-binding cAMP 3 H assay kit (Amersham Pharmacia Biotech) according to the manufacturer's recommendations. Proteins were solubilized by adding 0.5 ml of 1 N NaOH to the trichloroacetic acid-extracted tubes and quantified using Bio-Rad assay.
Statistical Analysis-Data are reported as means Ϯ S.E. n values are equivalent to the number of mice tested except for the patch clamp experiments, which indicates the number of cardiomyocytes used. A Student's t test was used for statistical comparisons between Tg and Ntg hearts with a two-tail p value of Ͻ0.05 considered significant.

RESULTS
Using the ␣-MHC promoter, the human L-VDCC ␣ 1C subunit was overexpressed in mouse ventricles (Fig. 1A). The ␣-MHC promoter becomes active within a few days after birth and reaches steady state maximal activity around day 60 (18). We therefore used 8 -10-week-old mice for our studies. A total of five founder lines were identified, classified as M1-M5. Two of these transgenic lines died within 8 weeks of age. Pathological examination of these two lines revealed severe myocardial hypertrophy and dilatation. Other transgenic lines were maintained at heterozygosity and compared with Ntg littermates for controls. The M1 and M5 lines had similar ␣ 1 subunit protein levels. Experiments were predominantly carried out in the M1 line except as noted.
Transgenic mice had no overt phenotype by observation. Additionally, litter sizes and pup survivals were similar to Ntg littermates. Loading known amounts of plasmid construct, the M1 line was determined to carry eight copies of the transgene (Fig. 1B). A quantitative RNA dot blot showed a 2.8-fold increase in the ␣ 1 subunit and no change in the ␤ and ␣ 2 /␦ subunits (Fig. 1C, Table I). In order to determine whether the increased gene product was translated into corresponding ␣ 1 subunit protein, homogenates were probed with specific antisera that does not distinguish between endogenous and Tg ␣ 1 protein. Western analysis revealed a clear increase in ␣ 1 subunit protein (data not shown). Assays of 8-week-old ventricular RNA for potentially regulated cardiac genes showed a significant 6.6-fold increase in ANF mRNA levels, but no significant differences in ␣-MHC, ␤-MHC, cardiac ␣-actin, skeletal ␣-actin, sarco-endoplasmic reticulum ATPase, and phospholamban gene expression (Table I).
We hypothesized that the increased ␣ 1 subunit in the transgenics should result in augmented contractility. Indeed, basal contractility and relaxation (expressed as ϩdP/dt and ϪdP/dt) were significantly higher for Tg hearts (4760 Ϯ 46 mmHg/s and Ϫ3935 Ϯ 56 mmHg/s, n ϭ 5) compared with Ntg hearts (4094 Ϯ 44 and Ϫ3160 Ϯ 56, n ϭ 4, p Ͻ 0.05). Infusion of the ␤-adrenergic receptor agonist, isoproterenol, did not elicit the expected inotropic and lusitropic (relaxation, diastole) increases observed in Ntg animals (Fig. 2). Heart rate increases were, however, comparable and normal for isoproterenol effects in both sets of mice (data not shown). Additionally, increases in contractility induced by forskolin (direct AC activator) were significantly decreased in Tg hearts compared with the Ntgs, suggesting that the defect in the ␤AR signaling pathway is not limited to an "uncoupling" of the receptor (Table II).
Echocardiographic measurements indicated that left ventricular end-diastolic dimension and posterior wall thickness were minimally increased in Tg as compared with Ntg mice (Table  III). Left ventricular mass in the transgenic hearts was clearly increased (23%) compared with Ntg hearts. However, no significant differences were found in left ventricular end-systolic dimension, septal wall thickness, or fractional shortening percentage (Table III). Heart-to-body weight differences were significantly increased in Tg animals compared with Ntg littermates (Table III). These findings are consistent with a mild hypertrophy without ventricular dysfunction. On the other hand, ex vivo studies of these hearts revealed increases in basal contractility and a dramatic loss of ␤AR responsiveness.
To determine possible functional changes in the L-type Ca 2ϩ channels, whole-cell voltage clamp recordings were carried out on isolated ventricular cardiomyocytes. Average cell capacitance was significantly larger in the Tg myocytes compared with Ntg myocytes (206.6 Ϯ 16.5 pF, n ϭ 9 versus 161.1 Ϯ 7.1 pF, n ϭ 7, p Ͻ 0.05), which indicates cellular hypertrophy. L-VDCC current amplitude was larger in the Tg myocytes com- pared with Ntg myocytes (Fig. 3, A and B). Average peak current amplitude was 1.8 Ϯ 0.2 nA (n ϭ 7) for Tg myocytes and 1.3 Ϯ 0.1 nA (n ϭ 9) for Ntg myocytes (p Ͻ 0.05, Fig. 3C). When the current amplitude was normalized for cell capacitance, there was no significant difference between Tg and Ntg cardiomyocytes (8.8 Ϯ 0.6 pA/pF (n ϭ 9) and 8.3 Ϯ 0.6 pA/pF (n ϭ 7)). There were no significant differences in the voltage dependence of activation or the activation/inactivation kinetics of the channels (data not shown).
The L-VDCC has been identified as a downstream phosphorylation target of the ␤-adrenergic receptor cascade (7). To determine whether this cascade remained intact in the Tg mice, isolated cells were stimulated with isoproterenol. When cells isolated from Ntg mice were superfused with 10 Ϫ7 M isoproterenol, peak Ca 2ϩ channel currents were augmented by 72.9 Ϯ 21.6% (n ϭ 6), but only 19.4 Ϯ 11.4% (n ϭ 5) for the transgenics (p Ͻ 0.05) (Fig. 3E). Consistent with the results in the intact heart experiments, forskolin produced only a small increase in the peak Ca 2ϩ current in the Tg hearts compared with the robust increases observed in the Ntgs (Table II).     To examine the blunted isoproterenol and forskolin responses found in whole heart and single cells, the ␤AR signaling pathway was investigated using isolated membranes. No change in total ␤AR receptor density was found (Tg, 19.2 Ϯ 1.2 fmol/mg; Ntg, 19.6 Ϯ 0.6 fmol/mg; n ϭ 3). Isoproterenol-stimulated activities were 150% of basal in Ntg mice (i.e. a ϳ50% increase over basal). In marked contrast, isoproterenol failed to stimulate AC activity in membranes from the Tg mice (indeed levels were slightly below basal) (Fig. 3F). Basal AC levels were 44.8 Ϯ 17 pmol/min/mg for Ntg and 52.7 Ϯ 11 pmol/min/mg for Tg extracts. As predicted NaF and forskolin-stimulated activities were depressed compared with Ntg, but the values did not reach statistical significance (Fig. 3F). Direct cAMP measurements further supported the loss of the ␤AR responsiveness. As observed in the AC assays, Tg membranes did not respond to isoproterenol nor was forskolin able to restore cAMP levels to the Ntg levels (Table II). DISCUSSION We constructed a transgenic mouse in which the L-type voltagedependent Ca 2ϩ channel was overexpressed. The transgene is defined as an increased VDCC by four criteria, Southern analysis, RNA analysis, electrophysiological identification of an increased Ca 2ϩ current, and a higher basal contractile state in the Tg animals compared with the Ntgs. Thus, these characteristics reflect a probable small, but sustained, increase in Ca 2ϩ influx in Tg myocytes that explains the significant increase in basal contraction and relaxation we observed. Accompanying the latter, we found a surprising and striking loss of the usual effects of a well known ␤ 1,2 adrenergic agonist, isoproterenol, on myocardial contraction. In contrast, the heart rate increases secondary to isoproterenol administration were normal, which defines an interesting separation of ␤-agonist actions on contractility and heart rate. This differentiation implies that downstream effectors and/or intermediates in the G protein pathway for ␤-agonist action are different for contractility and heart rate. It seems possible that for heart rate regulation Ca 2ϩ derived from the L-VDCC may not be limiting as it is for contraction. Consistent with the loss of contractile action was a blunting of isoproterenol and forskolin stimulation of Ca 2ϩ current in isolated myocytes. Thus, the effects found on the whole heart were for the most part duplicated on single cells for both isoproterenol and forskolin.
We refer to this loss of ␤-agonist contraction effect as a defect or a "loose coupling" in the ␤AR signaling pathways. The present data provide convincing evidence for a novel "reciprocal regulation" of ␤AR signaling by an increased influx of Ca 2ϩ provided by the L-VDCC. A possible mechanism to consider is protein kinase C activation by Ca 2ϩ (19) that in turn activates ␤AR kinase leading to phosphorylation of the ␤AR resulting in a decrease in signaling activity (20). Other mechanisms possibly responsible for the loss of ␤AR signaling include an inhibition of AC activity by Ca 2ϩ (21)(22)(23) and an indirect up-regulated phosphatase activity via calcium-calmodulin-dependent protein kinase (24). The latter is supported by recent data implicating calcineurin in cardiac hypertrophy (5). Further possibilities include up-regulation of G i , changes in the ratio of ␤ 1 AR:␤ 2 AR isoforms, etc. (20). Consistent with the suggestion of a Ca 2ϩ -dependent "cascade" involving a phosphatase is the slow development of hypertrophy and subsequent cardiac failure in these animals.
Our results suggest an interesting "cross-talk" between the L-VDCC and the ␤AR in vivo. A modest increase in the ␣ 1 subunit of the L-VDCC in cardiomyocytes results in a remark-able decrease in ␤-adrenergic signaling that affects contractility but interestingly does not alter the usual responses of heart rate to ␤AR agonists. The Ca 2ϩ channel ␣ 1 overexpression also produces a setting in which late stage ventricular remodeling occurs. This is surprising, since the cardiac cell has a remarkable network of sarcoplasmic reticulum and mitochondria that one would think would be poised to sequester any increase in calcium. In all experimental models of hypertrophy in which "Ca 2ϩ overload" was produced in no case was a paradigm used in which a small but sustained Ca 2ϩ increase through the L-VDCC was provided. Clearly the present experiments show that the L-VDCC in heart is a highly sensitive conduit for Ca 2ϩ as the link between excitation and contraction. Furthermore, it is likely that the increased calcium is sequestered in a pool that is linked to a growth program. A very recent publication has revealed a pathway to gene expression and growth that requires a very low concentration of calcium (25). These Tg mice represent the first animal model with an increase in voltagedependent Ca 2ϩ channels specifically in the heart that provides a paradigm for studies of the role of Ca 2ϩ in growth adaptation (25), maladaptation, and receptor/channel regulation.