CaV1.3 L-type Ca2+ channel contributes to the heartbeat by generating a dihydropyridine-sensitive persistent Na+ current

The spontaneous activity of sinoatrial node (SAN) pacemaker cells is generated by a functional interplay between the activity of ionic currents of the plasma membrane and intracellular Ca2+ dynamics. The molecular correlate of a dihydropyridine (DHP)-sensitive sustained inward Na+ current (I st), a key player in SAN automaticity, is still unknown. Here we show that I st and the L-type Ca2+ current (I Ca,L) share CaV1.3 as a common molecular determinant. Patch-clamp recordings of mouse SAN cells showed that I st is activated in the diastolic depolarization range, and displays Na+ permeability and minimal inactivation and sensitivity to I Ca,L activators and blockers. Both CaV1.3-mediated I Ca,L and I st were abolished in CaV1.3-deficient (CaV1.3−/−) SAN cells but the CaV1.2-mediated I Ca,L current component was preserved. In SAN cells isolated from mice expressing DHP-insensitive CaV1.2 channels (CaV1.2DHP−/−), I st and CaV1.3-mediated I Ca,L displayed overlapping sensitivity and concentration–response relationships to the DHP blocker nifedipine. Consistent with the hypothesis that CaV1.3 rather than CaV1.2 underlies I st, a considerable fraction of I Ca,L was resistant to nifedipine inhibition in CaV1.2DHP−/− SAN cells. These findings identify CaV1.3 channels as essential molecular components of the voltage-dependent, DHP-sensitive I st Na+ current in the SAN.


Results
Identification of I st in mouse SAN cells. The magnitude of I st varies depending on SAN cell types with distinct morphologies 5 . Mouse SAN cells used for I st recordings were typically spindle-or spider-shaped with no obvious striations. These cells were small (C m , 34.8 ± 1.2 pF, n = 42) compared to rod-shaped atrial-like cells and were spontaneously beating when superfused with normal Tyrode solution. To confirm the presence of I st in these cells, the late currents elicited by 1-s depolarizing voltage-clamp steps to various test potentials from a holding potential of −90 mV were examined for the characteristics of I st (Fig. 1). In order to avoid contamination of recordings by K + currents, we employed a Cs + -rich internal solution. I f was removed by substituting K + with Cs + in the external Tyrode solution, which contained 1.8 mM Ca 2+ . To confirm the sensitivity of the sustained current to DHPs, the typical hallmark of I st 19 , we tested the sensitivity of the current to the potent DHP L-type Ca 2+ -channel blocker isradipine, which has not previously been tested. In Fig. 1Aa, membrane currents recorded under control conditions (black trace), after lowering [Ca 2+ ] o from 1.8 to 0.1 mM (blue trace) and during subsequent application of 1 µM isradipine (red trace) are superimposed at individual test potentials. In the control bathing solution, membrane depolarization positive to −60 mV evoked a large transient inward current attributable to the activation of I Na and I Ca,T (note that peaks are not to scale in the figure), followed by a late inward current sustained during the entire period of 1-s depolarizing pulses. An inward current with a slow current decay was observed at test potentials of >−40 mV, as expected for I Ca,L activation. The current-to-voltage (I-V) relationship obtained by plotting the current amplitude measured near the end of test pulses indicated that the late current level becomes more inward with increasing depolarization between −70 and −50 mV (black circles, Fig. 1Ab), generating a negative slope conductance in the range of the diastolic depolarization. Lowering external Ca 2+ reduced a considerable fraction of I Ca,L at membrane voltages positive to −30 mV (inset, Fig. 1Aa), whereas the sustained inward current was not reduced. It is thus unlikely that the sustained inward current was generated by a window component of I Ca,L . However, bath application of isradipine readily inhibited the sustained inward current and unmasked an almost linear background conductance (Fig. 1Aa,b). Under conditions of low [Ca 2+ ] o , the DHP-sensitive sustained inward current peaked at −50 mV and the current direction was reversed at ~+ 26 mV (Fig. 1Ac).
Since I st has been shown to be carried by Na +5 , we tested the permeability of the sustained inward current component for Na + . The external Na + was replaced with an equimolar amount of N-methyl-D-glucamine (NMDG) in the presence of 1.8 mM Ca 2+ (Fig. 1B). Perfusion of SAN cells with Na + -free NMDG solution readily suppressed the sustained inward current as well as Na + -dependent background conductance ( Fig. 1Ba and b). Subsequent application of 1 µM nifedipine did not affect the late inward current component, indicating that Na + was the predominant ion carrying the DHP-sensitive sustained inward current. At voltages positive to −20 mV the outward current component was partially reduced by nifedipine (average current density of the DHP-sensitive outward current, 0.34 ± 0.08 pA/pF at + 20 mV; n = 4, two independent experiments: N = 2), suggesting that Cs + was carrying the DHP-sensitive current component.
We next evaluated the kinetics of inactivation of the sustained inward current in mouse SAN cells in further detail ( Fig. 1C and D). In the experiment shown in Fig. 1C, the inactivation time course was determined by measuring the fractional change of the sustained current elicited by a depolarizing step to −50 mV from a holding potential of −80 mV in 0.1 mM [Ca 2+ ] o , immediately (0.05 s) before (I ref ) and after (I test ) conditioning pulses to 0 mV of variable duration (0.5-4.5 s). Nifedipine was then applied to acquire the background current (red trace) at −50 mV, which was used to evaluate the net amplitude of the DHP-sensitive inward current. In Fig. 1C the ratio of I test /I ref is plotted as a function of the conditioning pulse duration, indicating that while I st displayed slow inactivation (τ = 1.94 ± 0.57s, n = 3, N = 1), a considerable current fraction remained available even after a 4.5-s conditioning pulse (0.48 ± 0.04, n = 3, N = 1). In addition, recovery from inactivation was assessed by applying a 5-s conditioning prepulse followed by test pulses to −50 mV after varying intervals of recovery (0.05-8.05 s) at −80 mV (Fig. 1D). Recovery of the sustained current proceeded exponentially with a time constant of 2.66 ± 0.63s (n = 3, N = 1).
These properties (low voltage for activation, DHP sensitivity, Na + permeability and slow inactivation), clearly identified the sustained inward current in our mouse SAN cell preparations as I st 5, [15][16][17][18]26 . We only failed to record the sustained current in five of 24 experiments (~20%), which is likely to indicate inhomogeneous expression of I st in SAN cells 16 . Four of five I st -deficient cells were nearly indistinguishable from clear striated atrial-like myocytes.
I Na and I NCX do not contribute to I st in mouse SAN. To assess whether voltage-gated Na + currents could interfere with I st recordings in mouse SAN cells, we investigated the effect of the I Na blocker TTX on the membrane current ( Fig. 2A). Since I st exhibited little inactivation during the 1-s square pulse, the I-V relationship was measured using a slow (150 mV/s) voltage-ramp protocol in 0.1 mM [Ca 2+ ] o . Under these conditions, the contributions of I Ca,L and I Ca,T to the total membrane current were minimized 5,15,18 . Figure 2Aa shows a superimposition of the original current traces in response to the voltage ramp in the control (black trace), during 10 µM TTX application (blue trace) and after nifedipine application (red trace). Figure 2Ab displays the corresponding I-V relationships obtained from the descending limb of the voltage ramp. As evidenced in the current recordings and in the corresponding I-V curve, bath application of TTX readily inhibited the transient inward I Na at the beginning of the pulse (see expanded traces in the inset), but did not affect the subsequent current. Application of nifedipine (1 µM) then revealed that the TTX-insensitive and DHP-sensitive current component could be attributed to I st (Fig. 2Ac). It should be noted that partial inhibition of the late inward current by TTX was observed in six of 23 cells (26%) (average current density of TTX-sensitive current, 0.82 ± 0.17 pA/pF at −50 mV; n = 6, N = 4), suggesting that some mouse SAN cells also express a TTX-sensitive persistent Na + current 27 .
The involvement of I NCX was also investigated (Fig. 2B). I st was hardly affected by total replacement of external Na + with an equimolar concentration of Li + to abolish I NCX . This result is consistent with the previously characterized selectivity of I st to monovalent cations 26 . In conclusion, I NCX did not contaminate our recordings of I st .  (Table 1) provided a strong rationale for testing the hypothesis that these currents share common molecular determinants. It is now generally accepted that I Ca,L in SAN cells is composed of two separate current components mediated by distinct pore-forming alpha subunits, Ca V 1.2 and Ca V 1.3 11,12 . To directly examine the possibility of a functional link between I st and Ca V 1.3, we recorded I Ca,L and I st in SAN cells from mice lacking Ca V 1.3 channels (Ca V 1.3 −/− mice, Fig. 3). Since most SAN cells obtained from Ca V 1.3 −/− mice were quiescent, we selected single cells for recordings based on morphological criteria rather than spontaneous activity. After the control recording in Cs + -substituted, K + -free Tyrode solution with 1.8 mM [Ca 2+ ] o (black traces), I st was separated from I Ca,L by switching the bath solution to 0.1 mM [Ca 2+ ] o containing 10 µM TTX (blue traces). I st was identified as a current component inhibited by subsequent application of 1 µM nifedipine (red traces). Consistent with previous studies, genetic ablation of Ca V 1.3 channels resulted in considerable reduction of I Ca,L 12 as well as a shift in the current half-activation voltage 11,12 (Fig. 3A-C). Indeed, the peak density of I Ca,L was significantly reduced from −6.97 ± 0.85 pA/pF in wild-type SAN cells (n = 19, N = 6) to −4.81 ± 0.45 pA/pF in Ca V 1.3 −/− cells (n = 18, N = 6, p = 0.0336), and was accompanied by a positive shift in the peak of the I-V relationship by ~20 mV (Fig. 3C). The calculated half-maximal activation voltage (V 0.5act ) was shifted from −29.3 mV in wild-type cells to −12.8 mV in Ca V 1.3 −/− SAN cells.

Sensitivity of I st to I
I st was evident in wild-type SAN cells after I Ca,L removal by lowering [Ca 2+ ] o , as manifested by the increase in the sustained inward current with depolarization between −70 and −50 mV that was finally blocked by nifedipine ( Fig. 3A and B). By contrast, the late current obtained from Ca V 1.  cells (Fig. 4). A slow ascending ramp (−65 to −35 mV, 100 mV/s) voltage command was employed to mimic the diastolic depolarization. We first recorded I Ca,L using the 0 Na + , 1.8 mM Ca 2+ external solution (Fig. 4A). Under these recording conditions, the voltage ramp gradually activated an inward current yielding negative slope conductance in wild-type SAN cells. This current was strongly augmented by Iso and inhibited by subsequent application of nifedipine. The nifedipine-sensitive difference current showed that the net I Ca,L started to activate clearly within the diastolic depolarization range, as expected for the low-voltage activation of Ca V 1.  (Fig. 4B). Similar to wild-type cells, Ca V 1.2-mediated I Ca,L could be elicited by subsequent depolarization at +10 mV. We did not find a statistically significant difference in the response of I Ca,L at +10 mV to Iso between wild-type (112.0 ± 8.7%, n = 7, N = 3) and Ca V 1.3 −/− SAN cells (94.6 ± 6.1%, n = 6, N = 3, p = 0.1319). Taken together, these observations indicated that Ca V 1.3 channels alone fully accounted for I Ca,L in the pacemaker potential range. We then switched to an external recording solution containing 140 mM Na + , 0.1 mM Ca 2+ and 10 µM TTX to record I st in the same cells. As illustrated in Fig. 4E, slow replacement of the bathing solution enabled the monitoring of gradual changes in membrane currents. These changes included a marked inward shift in the holding current and a reduction in I Ca,L at +10 mV (indicated by arrows). In contrast to I Ca,L , we observed an increase in the inward current accompanied by a negative shift in the peak potential, which indicated that increased I st offsets the loss of I Ca,L along the voltage ramp. Similar to I Ca,L , I st was enhanced by Iso and blocked by nifedipine (Fig. 4C). I st was detected in all wild-type SAN cells (0.228 ± 0.039 and 0.453 ± 0.076 pQ/pF in the absence and presence of Iso, respectively, n = 7, N = 3). However, we failed to record I st in Ca V 1.3 −/− SAN cells (0.012 ± 0.002 and 0.014 ± 0.004 pQ/pF in the absence and presence of Iso, respectively, n = 6, N = 3, Fig. 4D).  (Fig. 5A). I Ca,L was recorded after elimination of I st and I Na by Na + removal from the external recording solution. Bath application of nifedipine (0.03-1 µM) reduced the peak amplitude of I Ca,L in a concentration-dependent manner to a maximum of ~64% even at a saturating concentration of DHP (1 µM). This residual DHP-resistant I Ca,L was completely blocked by application of verapamil (3 µM), in line with previous data showing that the T1066Y mutation preserves the high sensitivity of Ca V 1.2 to phenylalkylamines 29 . Peak inward current of the nifedipine-resistant component activated more slowly (Fig. 5A), as expected for Ca V 1.2-mediated . The presence of this DHP-insensitive component is consistent with our earlier finding 11, 12 that I Ca,L in SAN cells is mediated by both Ca V 1.2 and Ca V 1.3.
We then examined whether altered sensitivity of Ca V 1.2 to DHP also affected the response of I st to nifedipine. The concentration-dependent inhibition of I st by nifedipine was investigated in SAN cells isolated from wild-type and Ca V 1.2 DHP−/− mice (Fig. 5B)

Discussion
Here we have demonstrated, for the first time, that voltage-gated L-type Ca V 1.3 Ca 2+ channels are essential for the expression of a DHP-sensitive, voltage-dependent Na + conductance, previously described as I st . Our finding is based on the observations that (1) I st is consistently identified in wild-type SAN cells but not in Ca V 1.3-deficient cells; (2) block of I st by nifedipine was unaffected upon ablation of Ca V 1.2 DHP sensitivity in Ca V 1.2 DHP−/− SAN cells; (3) DHP sensitivity of I st overlapped that of Ca v 1.3-mediated I Ca,L in Ca V 1.2 DHP−/− SAN cells; and (4) I st could not be attributed to late I Na or I NCX . Sensitivity to Ca 2+ -channel blockers such as DHPs, verapamil and diltiazem, as well as activators such as Bay-K8644 and FPL-64176, is based on highly specialized structural motifs conserved in Ca V 1 α 1 -subunits of L-type Ca 2+ channels 21,31 . Thus, our genetic and pharmacological evidence showing overlapping properties between I st and Ca V 1.3-mediated I Ca,L indicates a close functional relationship between these currents in SAN cells ( Table 1). The demonstration of Ca V 1.3 α 1 -subunits as essential molecular determinants of a voltage-dependent DHP-sensitive Na + conductance is a novel and unexpected finding and constitutes a fundamental step in elucidating the molecular nature of I st .
Our patch-clamp recordings clearly show that under physiological conditions I st is predominantly carried by Na + rather than Ca 2+ ions. Indeed, while lowering external Ca 2+ did not affect I st , removal of extracellular Na + abolished the current even in the presence of a physiological concentration of Ca 2+ (Fig. 1). L-type Ca 2+ channels are permeable to Na + in the absence of extracellular divalent cations 32,33 . However, L-type Ca 2+ channels are highly selective for Ca 2+ over Na + with a permeation ratio (P Ca /P Na ) of ~1000 under physiological conditions 32 . Therefore, Na + influx through L-type Ca 2+ channels is blocked by extracellular Ca 2+ in the submicromolar range 32,33 . It is thus unlikely that the "classical" permeation pathway of L-type Ca 2+ channels mediates I st . Indeed, currently available recombinant Ca V 1.3 channels with canonical channel pore sequence are Ca 2+ -selective I Ca,L , with poor permeability to Na + at least in the experimental solutions used for the I st recording in the present study (Toyoda et al., unpublished observation). Our results therefore suggest that Ca V 1.3 α 1 -subunits in the SAN cell not only form Ca V 1.3 L-type channels but also contribute to the formation of voltage-gated Na + conductance through an unknown mechanism.
However, revealing the molecular mechanism allowing Ca V 1.3 α 1 -subunits to form I st is challenging. We favour the hypothesis that Ca V 1.3 α 1 -subunits themselves form the I st pore due to the observation that I st and Ca V 1.3 possess an essentially indistinguishable pharmacological profile and Ca 2+ -channel blockers have been shown to exert their pharmacological modulation exclusively by binding to α 1 -subunits 34 . In this case the different ion selectivity of I st in SAN cells would require a modification of the ion permeation pathway. Substitution of negatively charged residues forming the ion selectivity filter of voltage-gated Ca 2+ channels by lysine can indeed induce persistent Na + currents similar to I st 35,36 , suggesting that increased Na + permeability per se could reproduce I st properties. To date, analysis of Ca V 1.3 transcripts has not identified alternatively spliced Ca V 1.3 variants with a modified selectivity filter [37][38][39] . Since Na + conductance through such modified channels may be larger than for Ca 2+ 36, 40 , I st transcripts may be present at low levels. This would make their detection particularly difficult in tissues with low cell numbers such as the SAN. On the other hand, the possibility that I st could be generated by alternative splicing of Ca 2+ channels is also suggested by a recent report that T-type (Ca V 3) Ca 2+ channels of the snail heart have high permeability to Na + due to unique splicing in the outer pore region 41 . Although this possibility cannot be excluded for mammalian SAN, splicing of T-type α 1 -subunits appears an unlikely explanation for I st because of the L-type channel-specific pharmacology. Another possible explanation for Na + selectivity of Ca V 1.3 α 1 -subunits could be structural modifications of the ion conducting pathway through RNA editing, which so far has only been detected in the brain and in the cytoplasmic C-terminal tail of the channel 39,42 .
Alternatively, a cationic channel functionally coupled to Ca V 1.3 activity could mediate I st . In mouse SAN cells Ca V 1.3 is co-localized with sarcoplasmic reticulum ryanodine receptors (RyRs) and controls diastolic RyR-dependent Ca 2+ release 25,43 . RyR-dependent Ca 2+ release could then activate an inward Na + current. However, the possibility that I st is mediated by Ca 2+ -dependent opening of a cationic channel appears unlikely, because I st density did not decrease upon lowering extracellular Ca 2+ (Fig. 1), as one would expect for Ca 2+ -dependent opening of a Na + -selective channel associated with Ca V 1.3. Finally, the possibility that I st could be generated by direct opening of a Na + channel physically coupled to Ca V 1.3 channel gating is also unlikely, because I st activates negative to Ca V 1.3-mediated I Ca,L (Fig. 1).
Ca V 1.3 loss-of-function in mice or humans results in SAN dysfunction, which indicates that Ca V 1.3 channels play a major role in pacemaker activity 11,12,[44][45][46] . Consequently, the present findings also suggest that the loss of I st could contribute to the SAN dysfunction induced by Ca V 1.3 gene inactivation. In addition, our results indicate that the heart rate reducing effect of Ca 2+ channel antagonists can be explained by drug binding to Ca V 1.3 channels and reduction of Ca V 1.3-mediated I Ca,L and I st . Consistent with previous observations 11,12 , our recordings in Ca V 1.3 −/− SAN cells show that Ca V 1.3 underlies a low-threshold I Ca,L activated at voltages spanning the diastolic depolarization range. I st differs from Ca V 1.3-mediated I Ca,L in the charge carrier and shows a more negative voltage for half-activation. I st and Ca V 1.3-mediated I Ca,L could thus differentially contribute to the generation of the diastolic depolarization. For example, I st could generate a persistent Na + influx in the diastolic depolarization, while Ca V 1.3-mediated I Ca,L could generate inward Ca 2+ current 12 and control RyR-dependent Ca 2+ release 43 . Notably, both I st and Ca V 1.3-mediated I Ca,L are strongly potentiated by β-adrenergic activation, which suggests a dual role of Ca V 1.3 in the sympathetic control of heart rate via I Ca,L and I st .
In conclusion, we provide novel evidence supporting the involvement of Ca V 1.3 in the generation of I st in SAN cells. Our work provides valuable new insights into the molecular basis of I st as well as the diverse functional significance of Ca V 1.3 in cardiac pacemaker activity.

Methods
Ethics. The investigation conforms to the Guide for the Care and Use of Laboratory Animals (8 th edition,