Rem-GTPase regulates cardiac myocyte L-type calcium current

Abstract

Rationale: The L-type calcium channels (LTCC) are critical for maintaining Ca²⁺-homeostasis. In heterologous expression studies, the RGK-class of Ras-related G-proteins regulates LTCC function; however, the physiological relevance of RGK–LTCC interactions is untested.

Objective: In this report we test the hypothesis that the RGK protein, Rem, modulates native Ca²⁺ current (I_Ca,L) via LTCC in murine cardiomyocytes.

Methods and Results: Rem knockout mice (Rem^−/−) were engineered, and I_Ca,L and Ca²⁺-handling properties were assessed. Rem^−/− ventricular cardiomyocytes displayed increased I_Ca,L density. I_Ca,L activation was shifted positive on the voltage axis, and β-adrenergic stimulation normalized this shift compared with wild-type I_Ca,L. Current kinetics, steady-state inactivation, and facilitation was unaffected by Rem^−/−. Cell shortening was not significantly different. Increased I_Ca,L density in the absence of frank phenotypic differences motivated us to explore putative compensatory mechanisms. Despite the larger I_Ca,L density, Rem^−/− cardiomyocyte Ca²⁺ twitch transient amplitude was significantly less than that compared with wild type. Computer simulations and immunoblot analysis suggests that relative dephosphorylation of Rem^−/− LTCC can account for the paradoxical decrease of Ca²⁺ transients.

Conclusions: This is the first demonstration that loss of an RGK protein influences I_Ca,L in vivo in cardiac myocytes.

Introduction

L-type Ca channels (LTCC) are critical for maintaining Ca-homeostasis, providing trigger Ca for Ca-induced Ca-release, and regulating electrical function in cardiac myocytes. Therefore, precise control of LTCC function has profound implications for cardiac myocyte physiology. LTCC exist as a multi-protein complex in cardiac myocytes.¹ Thus, understanding of LTCC function requires evaluation of LTCC Ca current (I_Ca,L) in native complexes. Cardiac myocyte LTCC consists of the main pore-forming subunit in complex with accessory proteins including Ca_Vβ2 family members, α2δ, and calmodulin (CaM; reviewed by 2). In addition, we, and others have recently extensively characterized the ability of the RGK-family of monomeric G-proteins to modulate I_Ca,L.^3-6

The RGK family of Ras-related monomeric G-proteins includes Rem, Rem2, Rad and Gem/Kir.⁷ Rem and Rad are expressed in myocardium.^8,9 Rad co-expressed with recombinant LTCC show block of current,¹⁰ and therefore, reduction of Rad in vivo is expected to increase I_Ca,L. Rad is downregulated in experimental induced heart failure (thoracic aorta constriction, TAC) in mice.¹¹ Rad knockout mice show no functional phenotype at baseline, but following TAC CaMKII increases.¹¹ Rad knock-down via RNAi results in an increase of I_Ca,L, Ca transients, and contractility.¹²

Heterologous expression studies document Rem blockade of I_Ca,L^5,10,13-15; however, there are no reports that assess in vivo Rem function. Overexpression of Rem, and other RGK proteins, results in prominent blockade of cardiac myocyte I_Ca,L.^4,16 Therefore, we tested the hypothesis that depletion of Rem results in an increase of I_Ca,L. In this report, we examined cardiac myocyte I_Ca,L in Rem-null mice (Rem^−/−). Cardiac myocyte I_Ca,L density increases in Rem^−/− compared with that from wild-type mice. Nevertheless, Rem-deletion results in no frank cardiac phenotype at baseline. The increased I_Ca,L density may be offset by a positive shift of the steady-state activation curve. Thus, our data support the notion that compensatory changes in Ca homeostasis preserve cardiac function.

Results

I_Ca,L in Rem^−/− cardiac myocytes

Rem overexpression blocks I_Ca,L in heterologous expression systems,¹⁴ and in cardiac myocytes.¹⁶ These results suggested that Rem governs I_Ca,L in vivo. To test this idea, we generated a Rem1 knockout mouse (Fig. 1). Rem null mutant mice (Rem^−/−) were born at the expected Mendelian ratio, and grew to adulthood without showing any discernible abnormalities.

Figure 1. Rem knockout mouse. (A) Schematic representation of the strategy used to generate the Rem knockout mouse. The coding region of Exon 2 was targeted and replaced by a LacZ expressing cassette, followed by the neomycin gene flanked by FRT sites. Arrows indicate the location of genotyping primers (a, b, c, and d). (B) Genomic DNA was isolated from mouse tail biopsies, and subjected to PCR analysis using the oligonucleotide primer pairs indicated in panel A. A 348 base pair band represents the mutant allele (lower panel; amplification using primer pair c/d), while a 500 base pair band represents the wild-type allele (upper panel; amplification using primer pair a/b). (C) Total protein was isolated from hearts obtained from Rem mutant or WT-littermates, and subjected to anti-Rem immunoblot analysis. Wild-type Rem expressed in HEK293 cells was included as a positive control. Note that Rem protein levels are completely eliminated in Rem^−/− cardiac muscle, while reduced in Rem^+/− animals.

To examine the contribution of Rem to LTCC function, we measured I_Ca,L in Rem^−/− cardiac myocytes. The expectation is that I_Ca,L should increase in the absence of Rem blockade of current. As the analog traces of Figure 2A shows the I_Ca,L density was significantly greater in Rem^−/− compared with age-matched wild-type controls (peak I_Ca,L density amplitudes at + 10 mV: -19.0 ± 0.5 pA/pF (n = 55) for Rem^−/−, and -16.0 ± 1.0 pA/pF (n = 40) for the age-matched wild-type controls; p < 0.05). The current-voltage relationships (I(V)) show that the differences in calcium current densities were statistically significant in the membrane potential range between 0 mV and +45 mV (Fig. 2B). These data are consistent with Rem governing I_Ca,L amplitude in cardiomyocytes.

Figure 2. I_Ca,L density for wild-type and Rem^−/− cardiac ventricular myocytes. (A) Representative original traces from wild-type and Rem^−/− mice. (B) Average I_Ca,L-V relationships for wild-type and Rem^−/− cardiac myocytes. (C) The voltage dependence of steady-state activation and steady-state inactivation of I_Ca,L for wild-type and Rem^−/− cardiac myocytes. V_½ of steady-state activation is significantly shifted +4mV for Rem^−/− compared with wild type. V_½ was obtained from fitting the current-voltage data as in panel B to a modified Boltzmann distribution of the form: I(V) = G_max*(V-E_rev)/(1+exp(V_½ - V)/k), where G_max is maximal conductance, E_rev is reversal potential, V_½ is activation midpoint potential, and k is the slope factor. The voltage dependence of steady-state inactivation of I_Ca,L was measured by a double-pulse protocol. The overlap of steady-state activation and steady-state inactivation curves (cross-hatched) illustrates decrease of window I_Ca,L in Rem^−/− compared with wild-type cardiac myocytes.

The maximal I_Ca,L was greater in Rem^−/− compared with wild-type control, and examination of the I(V) curves reveals a shift in the voltage dependence of I_Ca,L (Fig. 2B). Figure 2C shows that the midpoint of activation of Rem^−/− I_Ca,L is significantly shifted relative to control myocytes (-7.2 ± 0.3 mV and -3.3 ± 0.2 mV at wild-type and Rem^−/− myocytes respectively, p < 0.05). The voltage dependence of steady-state inactivation was measured by a double pulse protocol. As Figure 2C demonstrates Rem deletion does not affect the voltage dependence of steady-state inactivation. Thus the theoretical steady-state window Ca-current was reduced in Rem^−/− cardiac myocytes (Fig. 2C, shaded area). Taken together, there is an increase of maximal-activated I_Ca,L that is potentially compensated by a shift of steady-state activation.

I_Ca,L kinetics are key determinants of total Ca²⁺-entry into the cytosol, and in heterologous expression systems Rem alteration of kinetics depends on exogenous CaM expression.⁵ To evaluate whether Rem alters native I_Ca,L we measured I_Ca,L kinetics in cardiomyocytes. I_Ca,L kinetics were unaffected by Rem^−/−. The effect of Rem on the time dependent inactivation of I_Ca,L was determined by bi-exponential fitting to the decay of current. The tau₁ values were 24.7 ± 2.2 ms, and 25.5 ± 1.8 ms for wild type, and Rem^−/−, respectively. The tau₂ values were 5.0 ± 0.8 ms, and 5.3 ± 0.6 ms (Fig. S1). As expected, the fast and slow component amplitudes of Rem^−/− were significantly larger than that of the age-matched wild-type controls (p < 0.05) but the relative amplitudes of the fast/slow components were not different between Rem^−/− and control. The effect of Rem was also tested on the facilitation of I_Ca,L . During these measurements I_Ca,L was elicited by 16 consecutive pulses at 0.5Hz to 0 mV from -80 mV holding potential. The current amplitudes were normalized to the I_Ca,L amplitude elicited by the first pulse. Facilitation was not significantly different for Rem^−/− (8+4%) compared with that in wild-type cardiac myocytes (6+/−3%; Fig. S2).

β-adrenergic stimulation increases Rem blockade of I_Ca,L in heterologous expression systems.¹⁶ Also, it is well-established that β-adrenergic stimulation increases the I_Ca,L amplitude, and causes a negative shift of I_Ca,L activation. Thus, in the absence of Rem, we predict there is a greater dynamic range of I_Ca,L responsiveness to β-adrenergic stimulation. To test this hypothesis, we superfused isoproterenol onto cardiomyocytes during the whole-cell recording. The I_Ca,L response to isoproterenol was greater for Rem^−/− compared with wild-type cardiac myocytes. 10⁻⁸ M isoproterenol caused a 28 ± 4% and 22 ± 3% increase in I_Ca,L amplitude of Rem^−/− and wild-type myocytes, respectively. Isoproterenol treatment caused no significant shift in the midpoint (V_1/2) of steady-state inactivation curve. In the presence of 10⁻⁸ M isoproterenol inactivation V_1/2 was -29.8 ± 1.0 mV for Rem^−/−, and -28.9 ± 0.5 mV for wild-type cells. Isoproterenol shifted the voltage dependence of steady-state activation by -6.2 ± 1.0 mV for Rem^−/−, but only -3.9 ± 1.3mV for wild type (Fig. 3). As a consequence, isoproterenol-modulated channel voltage-dependence was no longer significantly different between Rem^−/− and wild type. These data are consistent with the notion that Rem serves as a governor of β-adrenergic stimulated I_Ca,L.

Figure 3. Isoproterenol normalizes the shift of activation V_½ for Rem^−/− compared with wild type. (A) Activation curves for wild-type I_Ca,L in control bath and 10nM isoproterenol (V_½ for control and isoproterenol were -7.2 ± 0.3mV and -11.7 ± 0.3mV, respectively; p < 10⁻³; n = 12). (B) Activation curves for Rem^−/− I_Ca,L in control bath, and following 10 nM isoproterenol (V_½ for control and isoproterenol were -3.3 ± 0.2mV and -9.4 ± 0.3mV, respectively; p < 10⁻⁴; n = 24).

The Na⁺-Ca²⁺ exchange current (I_NCX) is the major efflux pathway that homeostatically balances cytosolic Ca-entry via I_Ca,L.^17,18 The increased peak I_Ca,L can be offset by a positive shift of LTCC activation. Therefore, as a surrogate for assessing net trans-sarcolemmal I_Ca,L entry we measured I_NCX. I_NCX was recorded as a Ni²⁺-sensitive current using the descending limb of a ramp pulse changing from +60 to −100 mV. Outward and inward I_NCX were determined at +40 and −80 mV, respectively (Fig. 4). Rem^−/− have a significantly smaller inward I_NCX (0.90 ± 0.16 pA/pF) than for the wild type (1.92 ± 0.4 pA/pF; p < 0.05, n = 7; Figure 4B). The mean outward current was larger in wild-type cells compared with that of Rem^−/− cells, but the difference was not significant statistically (1.19 ± 0.24 pA/pF vs. 0.73 ± 0.11 pA/pF for wild-type and for Rem^−/− myocytes, respectively, N.S., n = 7; Figure 4A). These results are consistent with the notion that Rem^−/− net LTCC trans-sarcolemmal Ca flux may be reduced relative to wild-type cardiac myocytes.

Figure 4. I_NCX in wild-type and Rem^−/− ventricular myocytes. Upper panel shows voltage protocol. Outward I_NCX measured at +40 mV, and inward I_NCX measured at -80mV as indicated on hyperpolarizing ramp. Mean I_NCX is greater in wild type compared with control for (A) outward and (B) inward current. *p < 0.05

Ca²⁺-transients and contractility in Rem^−/− cardiac myocytes

Despite the increase of maximal peak I_Ca,L in Rem^−/− cardiomyocytes, the shift of activation, and the decrease of I_NCX suggest that in intact cells Rem deletion leads to reduced LTCC function. To determine the impact of Rem deletion on Ca²⁺-handling, we simultaneously measured whole-cell Ca²⁺ transients and cell shortening (Fig. 5). Isolated ventricular myocytes were paced with 1 Hz field stimulation. The amplitude of twitch Ca²⁺ was significantly smaller for Rem^−/− compared with wild type (0.35 ± 0.03 vs. 0.27 ± 0.02 for wild-type and for Rem^−/− myocytes, respectively, p < 0.01, n = 25 and n = 19 for Rem^−/− and wt, respectively; Figure 5A, B). Ten nM isoproterenol significantly increased twitch Ca amplitude in both Rem^−/− and wild type, but in the presence of isoproterenol the twitch Ca amplitude remained significantly smaller for Rem^−/− relative to wild type.

Figure 5. Ca-transients in intact ventricular myocytes. One Hz pacing reveals decreased twitch amplitude, without an effect on SR load in Rem^−/− cardiomyocytes. (A) Representative single twitch Ca²⁺-transient from myocytes 1 paced at 1Hz. (B) Rem^−/− twitch Ca-amplitude is significantly smaller than that in wild type. SR Ca-load was assessed by rapid application of 10mM caffeine. SR Ca load is not significantly different between Rem^−/− and wild type. n = 25 and 19, Rem^−/−, and wild type, respectively; *p < 0.01 control; *p < 0.05 in isoproterenol. (C) Decay kinetics of twitch Ca transients are not significantly different between Rem^−/− and wild type. (D) Simultaneous measure of fractional cell shortening revealed no significant different between Rem^−/− and wild type in either control or isoproterenol bath solution. (E and F) Computer simulations capture the decreased Ca transient amplitude in response to a +4mV shift of I_Ca,L steady-state activation. (E) Superimposed steady-state inactivation curve normalized to maximal conductance, and steady-state activation curve. For the activation curve, Gmax = 0.45 pS/pF for Rem^−/− and 0.35pS/pF for wt. Inset shows activation curves expanded for voltage ranging from -40 to -10mV. (F) Simulated Ca²⁺-transients shows smaller peak Ca²⁺-transient amplitude in Rem^−/− compared with wt.

The decaying phase of Ca transients was fitted to an exponential function to summarize the kinetics of Ca extrusion. The absence of Rem had no effect on the decay time constant (0.14 ± 0.01 sec for Rem^−/− and 0.15 ± 0.01 sec, for wild-type cells, N.S., n = 25 and n = 19 for Rem^−/− and wt, respectively, Figure 5C). Ten nM isoproterenol did not significantly alter the rate of Ca extrusion. The calcium load of SR was evaluated by a pulse of 5 mM caffeine, which was applied at the end of each experiment. SR Ca²⁺ load was unchanged by Rem knockout (Fig. 5B).

In spite of the smaller twitch Ca amplitude fractional cell shortening was not significantly different between Rem^−/− and wild type (Fig. 5D). Time to peak and relaxation kinetics of cell shortening was faster in Rem^−/− compared with wild-type myocytes (Fig. S3). The time to peak values of Rem^−/− and wild type were 110 ± 4 ms and 132 ± 7ms (n = 21; p < 0.05), while the half relaxation time was 77 ± 5 ms and 124 ± 16 ms (n = 21; p < 0.05), respectively. 10nM isoproterenol treatment decreased the time-to-peak values compared with normal pre-drug treatment values in both animal groups (Fig. S3B). Isoproterenol treatment reduced the half relaxation time by 20 ± 8 ms in wild-type animals (n = 21; p < 0.05), but we did not observe significant change of this parameter in of Rem^−/− (Fig. S3B). Cell shortening kinetics is a measure of a complex process. Shortening time reflects Ca-induced Ca release, Ca diffusion, and contractile apparatus response. Relaxation kinetics is dominated by cytosolic Ca clearance. Taken together these results suggest that the effect of the chronic absence of Rem extends beyond modulation of I_Ca,L.

Computer simulations

To explain our results we employed a newly developed computational model of mouse ventricular myocytes.¹⁹ Our approach was to re-adjust LTCC-related parameters based on data from I_Ca,L recordings in Rem^−/− vs. wild-type myocytes, and then have the model generate Ca²⁺ transients. Figure 5E shows the activation and steady-state inactivation curves for I_Ca,L superimposed with the model output. The computer simulations quantitatively capture the resulting decrease of 1-Hz stimulated Ca transients (Fig. 5F). In addition, the tendency for the mean sarcoplasmic reticulum Ca reduction was also present in the model output. Cell-size adjustments had only modest effects on Ca transients (data not shown). These results are consistent with the contention that a +4mV shift of LTCC activation results in a decreased Ca-transient, despite a larger maximal conductance.

Compensatory responses

The positive shift of basal I_Ca,L activation suggests that Ca_V1.2 protein might be less phosphorylated in Rem^−/− compared with wild type. The Ca_V1.2 distal carboxyl terminus (DCT) Ser1928 is a known substrate for the β-adrenergic-PKA signaling axis in cardiomyocytes.^20,21 Cardiomyocyte lysates show that DCT migrates at about 37kDa.²² The specificity for the antibody used to probe for Ser1928^P is demonstrated in Figure S4. Fractional DCT-S1928^P/total-DCT was significantly less in Rem^−/− (51+/−3%; n = 4) compared with that in wild-type ventricular lysates (78+/−4%; n = 4; p < 0.002; Figure 6A). We also tested phospholamban (PLB) phosphorylation to assess whether this was a general elevation of PKA signaling. PLB-Ser16 is a PKA substrate site,²³ and there was no difference in PLB-Ser16^P between wt and Rem^−/− (Fig. 6B and C). PLB-Thr17 is an established CaMKII substrate. Although the mean PLB-Thr17^P was greater in wt than Rem^−/−, the differences were not statistically significant (Fig. 6B and C).

Figure 6. Analysis of Ca_V1.2 and phospholamban phosphorylation in wt and Rem^−/−. (A) Protein extracts from wild-type and Rem^−/− hearts were subjected to western blotting and probed with antibodies recognizing Ca_V1.2 distal carboxyl terminus (DCT) and DCT-Ser1928^P. Size marker (horizontal line) is 37kDa. n = 4 Rem^−/− and n = 4 wild type. (B) Protein extract was not boiled prior to gel loading to preserve multimeric PLB structure. Six bands for PLN are resolvable; Ser16^P, and Thr17^P are confined to upper two bands. (C) Bar graph summarizes fractional phosphorylation. *p < 0.002; n = 4 Rem^−/− and n = 4 wild type. PLB phosphorylation was not significantly different. n = 6. (D) Ca_V1.2 expression is downregulated in Rem^−/− vs. wt heart; 74+/− 4% decrease in Rem^−/− vs. wild type; n = 3 Rem^−/− and n = 3 wild type; p < 0.01.

We postulated that an increase of I_Ca,L density in the absence of Rem expression, could in theory, be offset by a compensatory decrease of Ca_V1.2 protein expression. To test this straightforward notion, we evaluated immunoblots from isolated ventricular membrane fractions. We observed a significant decrease of full-length Ca_V1.2 protein for Rem^−/− compared with wild type (74+4%; p < 0.01; Figure 6D). The other major RGK protein in the myocardium, Rad, was not different in Rem^−/− hearts (Fig. S5).

Discussion

Overexpressed Rem blocks I_Ca,L in myocytes suggesting that endogenous Rem might provide tonic blockade of I_Ca,L. Thus, eliminating Rem expression was expected to increase I_Ca,L. The major finding of this study is that Rem knockout results in an increased I_Ca,L density. This is the first time that RGK GTPase modification of I_Ca,L has been shown following in vivo knockout. Rem deletion also resulted in a depolarizing shift of channel activation, and this shift was normalized to control by β-adrenergic stimulation.

Rem is one of three RGK proteins (Rem, Rad, and Gem) expressed by cardiac myocytes. To date, all RGK proteins share the functional property of blockade of I_Ca,L when overexpressed in either cardiac myocytes,^4,16 or other cells,^6,14 including those expressing recombinant proteins.³ In contrast, loss-of-function studies reveal that Rad knockdown via shRNA,¹² or Rad-dominant negative mutant overexpression²⁴ leads to increased I_Ca,L density. A Rad-null mouse has been described,¹¹ but unfortunately no information regarding I_Ca,L is available in these studies. At baseline, Rad-null mice have no phenotypic differences from control cardiac myocytes and hearts¹¹; however, in view of our present findings it would be reasonable to re-evaluate Rad-null mice. Such a study might shed light on possible non-redundant functions of Rad vs. Rem in cardiac myocytes.

A partial decrease of Rad via shRNA increases I_Ca,L and leads to a similar increase of SR Ca-release.¹² The parallels between depressing Rad function and the present report of Rem knockout are partial. Elimination of Rem also increases I_Ca,L density, but does not alter SR Ca-load. However, twitch Ca release was diminished. There are at least two non-competing mechanistic interpretations of how Ca-transients can be diminished despite elevated I_Ca,L density in Rem^−/− cardiac myocytes. Rem may interfere with excitation-contraction coupling directly; alternatively, I_Ca,L driven by the action potential may be reduced in Rem^−/− cardiac myocytes. Computer simulations support this notion. The +4mV activation shift is sufficient to decrease I_Ca,L, despite an elevated maximally-activated conductance, and this activation shift may responsible for the reduced twitch Ca amplitude. Interestingly the reduced twitch Ca amplitude did not cause a change of shortening amplitude. The activation shift in Rem-null cardiac myocytes is consistent with relatively fewer de-phosphorylated LTCC. Elevated I_Ca,L leads to cardiac hypertrophy.²⁵ Thus, it is plausible to view the shift of I_Ca,L activation as a compensatory protective effect against increased LTCC G_max in the absence of Rem expression.

Other compensatory responses that might occur in Rem^−/− mice are more complex. For example, there is an elevated in vivo heart rate (data not shown) consistent with elevated basal sympathetic tone. In this light, the relative de-phosphorylation of Ca_V1.2 could protect against Ca-overload via hyper-active I_Ca,L. However, we could not detect a change in PLB phosphorylation status, despite assessing conditions that preserved pentameric PLB. Note that such conditions show more pronounced changes in the phosphorylation status of Ser16 and Thr17.²⁶ The difference in relative phosphorylation status of Ca_V1.2 and PLB can reside in a number of factors, including but not limited to differential accessibilities to PKA and differential modulation by a variety of phosphatases.

We used Ca_V1.2-Ser1928 phosphorylation as a surrogate for the capacity of LTCC to serve as a substrate for PKA activity. The function of Ca_V1.2-Ser1928^P is controversial, but the conservation of Ser1928 across species indirectly argues for physiological relevance. Two recent studies argue convincingly that Ser1700^P and Ser1704^P confer β-adrenergic receptor initiated modulation,²¹ whereas Ser1928^P is not necessary for I_Ca,L modulation.^27,28 Nonetheless, data are also convincing that Ser1928 is a substrate for PKA in cardiac myocytes.²⁰

The mechanism for Rem alteration of I_Ca,L has been the focus of several recent studies. Early suggestions were that Rem blocks current via ‘sponging’ Ca_Vβ; however, several lines of evidence argue in favor of a model whereby a Rem-LTCC interaction occurs at the cell surface. Despite RGK overexpression, Ca_V1.2 can be detected at the cell surface in heterologous expression systems,¹³ neurons,²⁹ and other excitable cells.¹³ In addition, Ca²⁺-channel agonists causes rescue of I_Ca,L despite RGK overexpression in cardiac myocytes.⁴ Thus, it follows that in the absence of Rem, Ca_V1.2 channels at the surface are unmasked. This does not preclude alternative mechanisms of RGK modulation of LTCC function, particularly changes in stability of LTCC complexes on the plasma membrane whereby Rem promotes Ca_V1.2 internalization.¹⁵ We measured less total Ca_V1.2 protein concomitantly with more maximally activated I_Ca,L. This would be consistent with less Ca_V1.2 internalization in the absence of Rem expression. Unfortunately we were not able to directly assess endogenous Ca_V1.2 – this will be an important prediction to evaluate in future studies. To counterbalance higher maximal I_Ca,L the activation curve shift prevents excessive Ca²⁺ load as discussed above. It will be interesting to explore additional compensatory mechanisms possibly operating to maintain Ca-homeostasis. In summary our data are consistent with Rem null resulting in increased I_Ca,L; this is the opposite result to that in response to Rem overexpression in cardiac myocytes.¹⁶

In conclusion, the major finding of this study is that Rem-null mice display increased I_Ca,L. The increased I_Ca,L density is accompanied by a positive shift of the activation curve resulting in a predicted, yet paradoxical decrease of I_Ca,L entry at voltages corresponding to the plateau of the action potential. Thus, Rem modulates I_Ca,L in vivo, and cardiac myocytes compensate for modulated I_Ca,L to maintain Ca-homeostasis.

Methods

Generation of Rem^−/− mice

Mice were housed in a pathogen-free facility and handled in accordance with standard use protocols, animal welfare regulations, and the NIH Guide for the Care and Use of Laboratory Animals. All protocols were approved by the University of Kentucky Institutional Animal Care and Use Committee. A targeting vector (Fig. 1) containing approximately 11.4 kb of the murine Rem1 gene was constructed from 129/Sv strain mouse genomic DNA. Additional details are provided in the Supp. Methods.

Adult ventricular myocyte cell isolation, electrophysiology, and calcium imaging

Single ventricular myocytes from female mice were enzymatically isolated following a modified AfCS protocol PP00000125 as previously described.³⁰ External Ca was added incrementally back to the solution to 2.0 mM. Only rod-shaped, quiescent myocytes with clear edges were selected for current recording.

Ionic currents were recorded from Ca-tolerant female mouse ventricular cells at 37°C, 1 to 6 h post-isolation. Currents were recorded with the whole-cell configuration of the patch-clamp technique.

I_Ca,L was elicited by 500 ms long depolarizations to various test potentials arising from the holding potential of −40 mV. The stimulation frequency was 0.2 Hz. Peak current amplitude was defined as a difference between the peak value of I_Ca,L and its pedestal measured at the end of the pulse. I_Ca,L was normalized to cell capacitance.

Na⁺/Ca²⁺ exchanger (NCX) current was recorded in voltage-clamped ventricular myocytes as described previously.³¹ Briefly, NCX current was recorded using ramp pulses (shown in inset to Figure 3). Outward and inward NCX currents were determined during the descending limb of the ramp at +40 and –80 mV, respectively. After taking the control record in K⁺-free solution, the cell was superfused with 10 mM NiCl₂ in order to fully block the current. Thus, total NCX current was determined at both membrane potentials as a Ni²⁺-sensitive current.

The measurements of intracellular calcium concentration were done on isolated cardiac myocytes. The cells were loaded with 2 μM fura-2-a.m. for 45 min. All recordings were performed at 20–22°C. Cells were paced for a minimum of 2 min at 1Hz before recordings to allow steady-state Ca loading. At the end of each measurement, the intracellular calcium store was emptied by a pulse of 5 mM caffeine. During the caffeine pulse the cells were not stimulated. Cell shortening was recorded simultaneously using IonOptix myocam and software.

Additional recording conditions are presented in the Supp. Methods. In all recordings the mouse of origin genotype was encoded to blind the investigator.

Computer simulations

Computer modeling was performed using the framework of the previously developed electrophysiology model for murine ventricular myocytes at 35°C. The difference in cell volume between the wild type and Rem^−/− was calculated based on average measurements of the cell surface area and cell length and under the assumption that the myocytes are cylindrical in shape. The volumes of the junctional SR, network SR and dyadic space were adjusted according to their percentages (0.35, 1.05 and 0.1%, respectively) of the myoplasmic volume. The differences in the cell capacitance, myoplasmic volume and the volumes of the intracellular compartments between the wild type and Rem^−/− were incorporated into the model during parameterization of the I_Ca,L.

Parameters for I_Ca,L were fitted to experimental measurements obtained from the wild type and Rem^−/−, including the voltage-dependence of activation and inactivation kinetics, the current-voltage relationship, and the decay kinetics of the current, as described previously.³¹ All other parameters for the wild-type and the Rem^−/− models were unchanged from the existing model.¹⁹

Western blot analysis

Hearts were excised from wild-type and Rem^−/− mice and the left ventricles were isolated. Additional details are in Supp. Methods. Anti-pan phospholamban (PLB) was obtained from Thermo Scientific (USA) and S16^P and T17^P-phospholamban antibodies were purchased from Badrilla Ltd (UK). The Ca_V1.2 distal carboxyl terminus (DCT) antibodies (pan-DCT and S1928^P-DCT) custom designed (ECM Biosciences, Versailles, KY, USA). Figure S4 shows a representative blot demonstrating phospho-selectivity of the S1928P-antibody; for these experiments a group of treated cells was exposed to 10μM forskolin and 30 μM 3-isobutyl-1-methylxanthine (IBMX) for 10 min immediately prior to cell harvest. The Rad antibody was the kind gift of Dr. CR Kahn (Joslin Diabetes Center, Harvard Medical School, Boston). The anti-Rem antibody was monoclonal anti-Rem (24E4) from gamma-one laboratories (Lexington, KY). All antibodies were diluted in 5% milk blocking buffer and incubated either over night at 4°C or for at least 1 h at room temperature with rocking. The membranes were washed and incubated with secondary antibody for use on the Li-Cor Odyssey Imaging System (Lincoln, NB, USA) and analyzed using their software.

Glossary

Abbreviations:

LTCC

L-type calcium channel

Ca

Ca²⁺ or calcium ion

RGK

Rem, Rad, Rem2, Gem/Kir subfamily of Ras-related small GTPase

I_Ca,L

L-type calcium channel current

TAC

thoracic aorta constriction

CaM

calmodulin

CaMKII

calmodulin kinase II

Rem^−/−

Rem-null mouse

ECG

electrocardiogram

NCX

sodium-calcium exchange protein

PLB

phospholamban

DCT

distal carboxyl terminus of Ca_V1.2

Disclosure of Potential Conflicts of Interest

No potential conflicts of interest were disclosed.

Sources of Funding

Supported by NIH HL072936 (DAA, JS), HL07491 (JS), 2P20 RR020171 (DAA), OTKA 73160 (JM), MRC G0800980 (NS), and Challenge Grant ES018636.

Supplemental Material

Supplemental material may be found here: http://www.landesbioscience.com/journals/channels/article/20192/