Heart and Circulatory Physiology

Properties and expression of Ca2+-activated K+ channels in H9c2 cells derived from rat ventricle

Wei Wang, Makino Watanabe, Takeshi Nakamura, Yoshihisa Kudo, Rikuo Ochi


H9c2 is a clonal myogenic cell line derived from embryonic rat ventricle that can serve as a surrogate for cardiac or skeletal muscle in vitro. Using whole cell clamp with H9c2 myotubes, we observed that depolarizing pulses activated slow outward K+ currents and then slow tail currents. The K+ currents were abolished in a Ca2+-free external solution, indicating that they were Ca2+-activated K+ currents. They were blocked by apamin, a small-conductance Ca2+-activated K+ (SK) channel antagonist (IC50 = 6.2 nM), and byd-tubocurarine (IC50 = 49.4 μM). Activation of SK channels exhibited a bell-shaped voltage dependence that paralleled the current-voltage relation for L-type Ca2+ currents (I Ca,L).I Ca,L exhibited a slow time course similar to skeletalI Ca,L, were unaffected by apamin, and were only slightly depressed byd-tubocurarine. RT-PCR analysis of the mRNAs revealed that rSK3, but not rSK1 or rSK2, was expressed in H9c2 myotubes but not in myoblasts. These results suggest that rSK3 channels are expressed in H9c2 myotubes and are primarily activated byI Ca,L directly or indirectly via Ca2+-induced Ca2+ release from sarcoplasmic reticulum.

  • SK channels
  • L-type calcium channels
  • apamin
  • d-tubocurarine
  • reverse transcriptase-polymerase chain reaction

the myogenic cell line H9c2 was established by Kimes and Brandt in 1976 from embryonic rat ventricle (14); since that time it has been used as an in vitro model of cardiac muscle in a variety of biochemical and pathophysiological studies (13, 18, 29). Although H9c2 cells express SmN protein, a cardiac muscle-specific splicing factor (6), evidence suggests that H9c2 cells possess properties of skeletal and cardiac muscle: they depolarize in response to ACh, yet they can exhibit rapidly activating cardiac L-type Ca2+ currents (I Ca,L) (9). In fact, skeletal and cardiacI Ca,L have been recorded from single H9c2 myotubes, and the corresponding α1-subunit mRNAs for the respective L-type Ca2+ channels were detected by RT-PCR (17). Sipido and Marban (26) demonstrated a delayed rectifying outward current in H9c2 cells. Beyond that, however, the properties of K+ currents in H9c2 cells remain largely unknown.

Ca2+-activated K+ channels are widely distributed among tissues and exhibit a wide range of conductances (10). Small-conductance Ca2+-activated K+ (SK) channels are characterized by their small unitary conductance (∼10 pS), high Ca2+ sensitivity, very weak voltage sensitivity, and susceptibility to blockade by apamin, a polypeptide toxin isolated from bee venom, andd-tubocurarine (d-TC) (2, 22). Because SK channels are highly Ca2+ sensitive, even at resting potentials, they generate afterhyperpolarizations that are important in controlling spike frequency in excitable cells (15) and Ca2+ influx in nonexcitable cells (19). Recently, the molecular structures of several SK channels, including hSK1, rSK1, rSK2, rSK3, and hSK4 (12, 16), were described. Heterologously expressed SK2 channels are highly sensitive to apamin and d-TC, whereas SK1 channels are much less sensitive (16). Heterologously expressed SK3 channels, which have only one of the two key amino acid residues necessary for high apamin sensitivity, are less susceptible than SK2 channels to apamin blockade (11).

In the present study we characterized the Ca2+-dependent activation, ion selectivity, and susceptibility to blockade by apamin andd-TC of SK channel currents (I SK) in H9c2 cells. We then used RT-PCR to determine that rSK3 mRNA is expressed in myotubular H9c2 cells.


Cell culture.

H9c2 cells derived from embryonic rat ventricle (passage 14; American Type Culture Collection, Rockville, MD) were cultured in DMEM (GIBCO BRL, Gaithersburg, MD) supplemented with 10% fetal bovine serum (Irvine Scientific, Santa Ana, CA) under an atmosphere of 95% air-5% CO2 at 37°C. Stocks of myoblasts were propagated in culture flasks for successive passage. For electrophysiological recordings, cells were plated to a density of 104cells/cm2 on small pieces of coverslip. The cells initially grew as flat, spindle-shaped mononucleated myoblasts that were several tens of micrometers long and 10–20 μm wide. Within 1–2 wk, however, they began to form multinucleated myotubes that were often several hundred micrometers long with ≥10 nuclei. Mononucleated myoblasts cultured for 4–9 days and myotubes (<400 μm long with 3–8 nuclei) cultured for 2–6 wk were used in these experiments. The myotubes did not possess sarcomere-like structures and did not contract spontaneously.

Electrophysiological recordings.

Patch pipettes were constructed from soft glass capillaries (Propper) that were double pulled, coated with Sylgard (Dow Corning, Midland, MI), and fire polished; resistances were 2–3 MΩ when pipettes were filled with solution. Membrane currents were recorded using an EPC-7 (List Electronic, Darmstadt, Germany) or EPC-8 (HEKA Elektronic, Lambrecht, Germany) patch-clamp amplifier and standard whole cell patch-clamp techniques (7). The recorded data were stored for later analysis on DAT tape after processing by a PCM recorder (model 110, TEAC, Tokyo, Japan) or on a hard disk utilizing the MacLab Chart system (version 3.5s, AD Instruments, Castle Hill, NSW, Australia). Command pulses were applied at 0.2 Hz with use of a step-pulse generator (model SET-1201, Nihon Kohden, Tokyo, Japan) or pClamp software (version 6, Axon Instruments, Foster City, CA). Data analysis was performed using pClamp or Patch Analyst Pro (version 1.21, MT, Tokyo, Japan) running on a Macintosh computer. Whole cell membrane capacitance was calculated from the peak amplitudes and the time constants of decay of capacitative transients elicited by 10-mV, hyperpolarizing voltage pulses from the holding potential (HP, −50 mV). In myotubes the membrane capacitance was 276.2 ± 22.4 (SE) pF (n = 59).

Solutions and chemicals.

Under control conditions, cells were superfused with Tyrode solution containing (in mM) 135 NaCl, 5.4 KCl, 1.8 CaCl2, 1.0 MgCl2, 5 HEPES, and 10 glucose; pH was adjusted to 7.4 with Tris. Ca2+-free Tyrode solution was made by adding 0.5 mM EGTA to nominally Ca2+-free Tyrode solution. Equimolar KCl or tetraethylammonium chloride (TEA; Sigma Chemical) replaced NaCl in some experiments. The pipette solutions contained (in mM) 140 KCl, 1 MgCl2, 10 HEPES, 5 Mg-ATP, and 0.1 or 10 EGTA; pH was adjusted to 7.3 with KOH. Ca2+ currents were recorded using TEA bath solution containing (in mM) 135 TEA, 5.4 CaCl2, 1 MgCl2, 5 Tris, and 10 glucose; pH was adjusted to 7.4 with HEPES. Pipette solution used for recording Ca2+ currents contained (in mM) 140 CsCl, 1 MgCl2, 1 EGTA, 5 HEPES, and 5 Mg-ATP; pH was adjusted to 7.3 with Tris. Apamin (Peptides Institute, Osaka, Japan) and d-TC (Sigma Chemical) were stored as aqueous stock solutions. All experiments were carried out at room temperature (23–25°C).


Total RNA was prepared from rat brain, undifferentiated H9c2 myoblasts, and differentiated myotubes by use of the guanidine thiocyanate method. Strand cDNA was then synthesized from 1 μg each of the respective total RNA samples by oligo(dT)-primed reverse transcription in a 20-μl reaction volume. Aliquots (0.5 μl) of the RT reaction mixtures were then subjected to PCR amplification with use ofTaq polymerase. PCR was carried out in a thermal cycler at 94°C for 5 min, then 21–34 cycles at 94°C for 1 min, 55–60°C for 30 s, and 72°C for 1 min. This protocol was followed by a final 10-min extension step at 72°C. The following specific primers for the SK channel family were designed on the basis of previously reported sequences (16): rSK1, CAGGC CCAGCA GGAGG AGTT (forward) and GGCGG CTGTG GTCAG GTG (reverse); rSK2, TCCGA CTTAA ATGAA AGGAG (forward) and GCTCA GCATT GTAGG TGAC (reverse); rSK3, GTGCA CAACT TCATG ATGGA (forward) and TTGACA CCCCT CAGTT GG (reverse). The primers used for myogenin were CTGGG GACCC CTGAG CATTG (forward) and ATCGC GCTCC TCCTG GTTGA (reverse). For β-actin the primers were CATGC CATCC TGCGT CTGGA (forward) and CCACA TCTGC TGGAA GGTGG (reverse). After PCR amplification, equivalent volumes of each PCR reaction mixture were subjected to electrophoresis on 5% polyacrylamide gels, stained with ethidium bromide, and visualized by ultraviolet fluorescence. The nucleotide sequence of each PCR product was analyzed by direct DNA sequencing.

Data analyses.

K+ current amplitudes were measured as the difference between the maximal current amplitude at each test potential and the current at the HP (−50 mV). Averaged and normalized data are presented as means ± SE. Dose-response curves were fit by one-site competition with use of Prism 2.0 (GraphPad Software, San Diego, CA). The statistical significance of differences between calculated means was evaluated using Student’st-test for unpaired samples;P < 0.05 was considered to be significant.


Slow outward currents are activated by depolarization and blocked by SK channel antagonists.

In H9c2 myotubes a series of 500-ms depolarizing steps from the HP (−50 mV) elicited a family of outward currents with a biphasic time course: an early rapid phase followed by a late slow phase (Fig.1 A). On repolarization, slow outward tail currents were observed. Switching from normal Tyrode to Ca2+-free solution abolished the slow phase of the outward currents as well as the slow tail currents (Fig. 1 B;n = 23). In addition, the relation between current density (pA/pF) and membrane voltage, calculated from current amplitudes measured at the end of the 500-ms pulses, showed that membrane conductance was almost halved in the absence of extracellular Ca2+, and the shape of the curve was linearized at large depolarized potentials (Fig.1 C).

Fig. 1.

Effects of Ca2+-free solution on outward K+ currents in H9c2 myotubes. Depolarizing voltage steps to −40 mV and up to 40 mV in 10-mV increments were applied from holding potential (HP, −50 mV). A: normal Tyrode (NT) solution;B: Ca2+-free Tyrode solution. Currents are calibrated differently inA andB. C: current density expressed as a function of membrane voltage; currents were measured on completion of test pulses. Values are means ± SE from 13 cells.

The remaining outward currents in Ca2+-free solution were activated rapidly, and then they slowly decayed (Fig.1 B). Similar fast outward currents with no superimposed slow current components were also recorded in myotubes when the pipette solution contained 10 mM EGTA (not shown,n = 22) and in myoblasts in the presence of 0.1 mM EGTA solution (not shown,n = 13). Because TEA (135 mM), a blocker of voltage-dependent K+channels, markedly suppressed the fast outward currents in myotubes dialyzed with 10 mM EGTA (n = 4) and in the myoblasts (n = 5), they were considered to be voltage-gated K+currents (I Kv). The slower currents, which were seen only in myotubes, were dependent on an increase in the intracellular Ca2+ concentration ([Ca2+]i) and were thus considered to beI K(Ca).I K(Ca) andI Kv accounted for 61.1 ± 14.2% (n = 13) and 38.9 ± 14.2% of total delayed outward current, respectively, at 20 mV (Fig. 1 C).

Figure 2 illustrates the effects of the SK channel blocker apamin on a family of depolarization-activated outward currents in a small myotube with a membrane capacitance of 60 pF, which enabled homogenous space clamping of the myotube. Under control conditions the slow currents continued to increase throughout the 500-ms depolarization steps, and again slow tail currents were elicited by repolarization (Fig. 2 A). Apamin (30 nM) suppressed the slow phase of the depolarization-induced currents and completely blocked the tail currents (Fig.2 B). Apamin at 10–100 nM blocked slow outward currents (n = 16), and the remaining current exhibited the rapid activation and slow inactivation characteristic ofI Kv. The apamin-sensitive currents (I SK) were isolated by subtracting theI Kv from the total currents. The tail currents, which on this time scale appeared to result entirely from theI SK, were first detected at depolarization steps more positive than −20 mV and were maximal at 10 mV (Fig. 2 C). The isochronal current-voltage (I-V) relationships of the total currents indicate thatI SK predominated after 500 ms in this cell and that peak amplitudes of the bell-shapedI-V curve forI SK occurred at 10–20 mV (Fig. 2 D). The voltage dependence of I SKactivation was assessed from the I-Vcurve obtained by plotting tail current amplitudes as a function of the preceding depolarization step (Fig.2 E). The resultant bell-shaped curve paralleled the I-V curve forI Ca,L (see Fig.5), and identical bell-shaped I-Vcurves for I SKtail currents were observed in four other cells. These results indicate that activation ofI SK was dependent on a rise in [Ca2+]i, elicited directly via Ca2+ entry through membrane Ca2+ channels or indirectly via Ca2+-induced Ca2+ release from internal stores.

Fig. 2.

Voltage dependence and apamin sensitivity of Ca2+-activated K+ current [I K(Ca)] in H9c2 myotubes. A: control;B: in presence of 30 nM apamin;C: apamin-sensitive currents obtained by subtracting apamin-resistant currents (b) from control currents (a) at corresponding voltage steps. Depolarizing steps of 500 ms to −40 mV and up to 50 mV in 10-mV increments were applied at 0.2 Hz from HP (−50 mV).D: current-voltage (I-V) curves constructed from traces in A (●),B (•), andC (○). Current amplitudes were measured at 500 ms from 0 current. E:I-V curves constructed fromI K(Ca) tail current traces depicted in C. Tail currents were measured as maximal positive deflection from 0 current. Membrane capacitance was 60 pF.

Dependence of ISK reversal potential on extracellular K+concentration.

To verify that K+ served as the charge carrier forI SK, the reversal potential of the tail currents was estimated in the presence of 5.4 and 30 mM extracellular K+ by use of a double-pulse protocol (Fig.3 Aa). In the presence of 5.4 mM extracellular K+, the polarity of the outward tail currents was eventually reversed by successively increasing the amplitudes of the hyperpolarizing second steps (Fig.3 Ab). The resultant inwardI SK tail currents decayed at a rate similar to that of the outward currents, indicating that the gating of the channel was voltage independent.

Fig. 3.

Effect of extracellular K+concentration on reversal potentials of evoked outward currents in H9c2 myotubes. A andB: in presence of 5.4 and 30 mM extracellular K+, respectively.a, Pulse protocol: 500-ms hyperpolarizing (A) and depolarizing (B) pulses applied in 5-mV increments after 500-ms conditioning pulses to 20 mV from HP (−50 mV); b, control currents;c, currents evoked in presence of 300 μM d-tubocurarine (d-TC);d, subtracted currents.C:I-V curves depicting tail current amplitudes as a function of membrane potential.

When extracellular K+concentration was increased to 30 mM, repolarization to the HP elicited large inward currents, the polarity of which was reversed by successively increasing the depolarizing second steps (Fig.3 Bb). In addition, the rate of decay of the tail currents became slower with depolarization, and it seemed likely to us thatI Kv were superimposed on the SK tails during the depolarizing second steps. Therefore, d-TC (300 μM), an SK channel antagonist, was utilized to separate SK tail currents fromI Kv.d-TC blocked the slow currents within several minutes (Fig. 3 Ac), and considerable I Kvwere seen during the depolarizing second steps of the double-pulse protocol (Fig. 3 Bc). When currents elicited in the presence of d-TC were subtracted from control currents, the slow activation and large slow tails of the remaining currents were characteristic ofI SK (Fig.3 Bd). TheI-V relationships of the tail currents were approximately linear in the presence of 5.4 or 30 mM extracellular K+, and the reversal potentials (E RV) were −72 and −28 mV, respectively. In seven experiments,E RV were calculated to be −66.6 ± 1.7 and −23.5 ± 1.5 mV in 5.4 and 30 mM extracellular K+, respectively. The 43.1 ± 0.9 mV positive shift inE RV approximated the Nernst shift ofE RV for a K+ electrode (45 mV), indicating that I SK was carried primarily by K+.

Dose dependence of apamin- and d-TC-induced inhibition of ISK.

Dose-response curves for apamin andd-TC were plotted on the basis of the fractional decreases in the amplitudes of theI SK tail currents elicited after 500-ms depolarization steps to 10 or 20 mV from the HP (−50 mV) in the presence of selected concentrations of the inhibitors (Fig. 4). Assessing the inhibitory potency of apamin was complicated by the fact that apamin was not easily washed out. In contrast, blockade byd-TC developed rapidly and was readily reversed by washing it from the chamber. The dose-dependent decreases in I SK were fit by a dissociation curve, with the assumption of a single binding site for apamin or d-TC; IC50 was 6.2 nM for apamin (Fig.4 A) and 49.4 μM ford-TC (Fig.4 B).

Fig. 4.

Pharmacological blockade of slow tail currents by apamin (A) andd-TC (B). SK channel tail currents (I SK) were obtained after 500-ms depolarizations to 10 or 20 mV from HP (−50 mV). Insets: cumulative depression ofI SK in presence of 30 mM K+ elicited by increasing concentrations of apamin (0, 1, 3, 10, 30, 100, and 300 nM) ord-TC (0, 1, 3, 10, 30, 100, 300, and 1,000 μM). Dose-inhibition curves were obtained by plotting fractional decrease of controlI SK against log concentrations of apamin (A;n = 6–11) andd-TC (B; n= 4–6); values are means ± SE. Data were fitted by the following equation: y =y 1 + (y 2y 1)/[1 + 1 Embedded Image ], where y 1 andy 2 are minimal and maximal values, respectively. InA,y 1 = 0.0657,y 2 = 0.763, IC50 = 6.23 × 10−9 M; inB,y 1 = 0.0464,y 2 = 0.888, IC50 = 4.94 × 10−5 M.

Effect of apamin and d-TC on Ca2+ currents.

Apamin is a highly potent fetal cardiac L-type Ca2+ channel antagonist (3). To determine the extent to which inhibition ofI Ca,L by apamin or d-TC contributed to their suppression ofI SK, we examined their effects onI Ca,L in H9c2 myotubes. I Ca,Lwere elicited by 500- or 1,000-ms depolarizing steps from the HP (−50 mV) in the presence of TEA solution containing 5.4 mM Ca2+ and recorded using Cs+ pipettes. As shown by the representative traces, the time courses of theI Ca,L were slow (Fig.5 A): in 11 cells the time required for currents elicited by depolarization to 10 mV to reach one-half peak amplitude was 27.9 ± 3.5 ms, and the time to peak was 94.7 ± 9.5 ms. Current amplitudes then decayed slowly to 59.4 ± 3.8% of peak in 500 ms. TheI-V relations forI Ca,L were obtained by plotting the amplitudes of evoked currents, normalized to the maximal I Ca,Lelicited by depolarization to 10 mV under control conditions, against membrane voltage (Fig. 5 B).I Ca,L were detected at potentials more positive than −10 mV and were maximal at 10 mV; the average maximum amplitude ofI Ca,L was 833.3 ± 162.7 pA (n = 11), andE RV was more positive than 50 mV.I Ca,L tail currents deactivated much more rapidly thanI SK tail currents.

Fig. 5.

Effect of apamin on L-type Ca2+currents (I Ca,L) in H9c2 myotubes. A: maximalI Ca,L evoked in tetraethylammonium chloride solution containing 5.4 mM Ca2+ in presence and absence (control) of apamin (300 nM). Note slow time course of inactivation and rapid decay of tail current. B:I-V curve depictingI Ca,L amplitudes in presence and absence of apamin.I Ca,L were measured as peak amplitudes of inward currents from 0 current during 1,000-ms depolarization steps from HP (−50 mV). Currents were normalized to peak currents evoked by depolarization to 10 mV under control conditions (I/I peak). Data from 7 cells are expressed as means ± SE.

Apamin (300 nM) had no effect on the Ca2+ I-V curves or the time courses of evoked currents: the ratio of the maximalI Ca,L amplitude in the presence to that in the absence of apamin was 1.06 ± 0.06 (n = 7). Moreover, time to one-half peak amplitude, time to peak, and percent decay at 500 ms were unaffected by apamin. d-TC (1 mM) also did not shift the I-V curve (not shown) and did not significantly affect the time courses of activation or deactivation, but it decreased maximal current amplitudes at 10 mV to 75 ± 9% of control (n = 4).

RT-PCR analysis of SK channels.

To confirm the expression of the SK channels that were functionally demonstrated by the patch-clamp study, we used RT-PCR analysis to determine whether SK channel mRNA was expressed in H9c2 myotubes. When total RNA, separately isolated from undifferentiated and differentiated H9c2 cells (Fig. 6 D), was subjected to RT-PCR with rSK3-specific primers, a PCR product with the predicted, 182-bp length of rSK3 mRNA was detected in differentiated myotubes but not in undifferentiated cells (Fig.6 A). Expression of rSK3 mRNA was also detected in adult rat brain, which is consistent with earlier in situ hybridization experiments (16). In H9c2 cells the expression of rSK3 was well correlated with that of myogenin, a myogenic regulatory factor, the expression of which is increased when myoblasts terminally differentiate and fuse into multinucleated myotubes (28).

Fig. 6.

RT-PCR analysis of SK channel mRNA in H9c2 cells.A: RT-PCR detection of rSK3, myogenin, and β-actin expression in brain (B), and undifferentiated (U) and differentiated (D) H9c2 cells. For rSK3 and myogenin, PCR protocol called for 30 amplification cycles; for β-actin, indicated numbers of cycles were selected to avoid saturating amplification.B: PCR amplification of templates with (+) and without (−) RT reaction.C: RT-PCR detection of rSK1 and rSK2 channels in brain and H9c2 cells. PCR protocol entailed 34 cycles.D: ribosomal 28S and 18S RNA in RNA samples prepared from H9c2 cells. Total RNA (5 μg) was subjected to electrophoresis on 2% agarose gels and stained with ethidium bromide.

As shown in Fig. 6 B, total RNA fractions isolated from cultured H9c2 myotubes were intact: the PCR products for rSK3 and myogenin were not detected unless they were reverse transcribed, which verified that the DNA fragments were not derived from contaminating genomic DNA sequences. In addition, the nonsaturating amplification of β-actin showed that there were no significant differences among samples with respect to the amount of first-strand cDNA within each sample (Fig.6 A). Using a competitive PCR method for β-actin, we confirmed that the quantities of PCR template in each sample were equivalent (data not shown). Apamin-insensitive rSK1 (159-bp PCR product) was not detected in H9c2 myotubes, and apamin-sensitive rSK2 (190-bp PCR product) was expressed to a minimal degree in differentiated myotubes (Fig.6 C), whereas rSK1 and rSK2 mRNA were detected in adult rat brain. Nucleotide sequences of the PCR products were identical to those of the cloned sequences (16). Thus apamin-sensitiveI SK recorded in H9c2 myotubes resulted virtually entirely from activation of rSK3 channels.


SK channels are voltage independent but highly Ca2+ sensitive (2, 16, 20, 22). Apamin-sensitiveI SK were identified in H9c2 cells by their Ca2+ dependence and their sensitivity to apamin and d-TC. The activation of theI SK was dependent on I Ca,L, and their activation and deactivation time courses were slow, as has been reported for I SKin rat chromaffin cells (22). Analysis of mRNA expression using RT-PCR revealed that I SKflowed through SK3 channels.

In a previous study of 10- to 20-day-old H9c2 cell cultures, Sipido and Marban (26) showed that depolarization-induced, whole cell outward currents exhibited variable time courses. In some cases, currents rose rapidly and then slowly decayed in a manner analogous to theI Kv described here. In other cases, currents rose rapidly and then continued to gradually increase throughout the depolarization steps, corresponding to the present myotubular outward currents composed ofI Kv andI SK. The slow tail currents characteristic ofI SK were not observed by Sipido and Marban, although the absence of tail currents on repolarization to −80 mV is to be expected for currents carried by K+. In the present study we characterized the slowly increasing outward current component and identified it asI SK. The voltage-dependent nonspecific cation currents observed at the early developmental stages (26) were absent in the myotubes.

Activation of SK channels in H9c2 myotubes exhibited a bell-shaped voltage dependence that paralleled theI-V relation forI Ca,L, strongly suggesting that Ca2+ influxes during depolarization steps were involved in the activation of SK channels. Although apamin blocksI Ca,L in fetal cardiac muscle (3),I Ca,L in H9c2 cells were unaffected by apamin and only slightly depressed byd-TC, indicating that these antagonists inhibitedI SK not by depressing I Ca,L, but by directly affecting SK channels. Consistent with the slow time course of skeletal muscleI Ca,L (4),I Ca,L recorded in H9c2 cells rose slowly (time to peak = 95 ms), continued to flow during the entire 500-ms depolarization, and decayed to only 59% of peak by the end of the pulse.

It may be that the slow activation of SK channels results from the slow rise in [Ca2+]iproduced by Ca2+ entry via L-type Ca2+ channels during depolarization. For instance, if an H9c2 myotube is assumed to be a 300-μm-long, 20-μm-diameter cylinder with Ca2+ flowing into it at a constant current of 0.4 nA Ca2+ during a 500-s depolarization pulse and distributing homogeneously, [Ca2+]iat the end of a depolarization pulse would be 11 μM in the presence of 5.4 mM. Even in the presence of 1.8 mM Ca2+, [Ca2+]iwould exceed the IC50 for SK channels (0.5–0.8 μM) (16, 22). Furthermore, increases in [Ca2+]imay be augmented if Ca2+ influxes induce Ca2+ release via Ca2+-induced Ca2+ release (5) from sarcoplasmic reticulum. On the other hand, because [Ca2+]iis the product of the equilibrium between Ca2+ influx, Ca2+ release from and uptake into internal stores, Ca2+ extrusion by the plasma membrane Ca2+ pump and Na+/Ca2+exchange, and cytosolic Ca2+buffering, influx-induced changes in [Ca2+]iare certainly attenuated. Rapid buffering by abundant (several mM), endogenous Ca2+ buffers with dissociation constants ∼100 μM has been reported in bovine chromaffin cells (30). Such buffering could account for the observed slow activation ofI SK, despite the large Ca2+ influx. Similarly, the slow decay of I SKtail currents and their characteristic plateaulike initial phase (20,22) may reflect [Ca2+]ibuffering and the time-dependent decline in [Ca2+]imediated by sequestration by internal stores and extrusion by Na+/Ca2+exchange. The contribution of Na+/Ca2+exchange to the extrusion of Ca2+was apparent from the reduction of the rate of decline inI SK tail currents seen with repolarization to more positive voltages (Fig.3 B).

The structures of several SK channels, including hSK1, hSK4, rSK1, rSK2, and rSK3, have recently been identified (12, 16). SK channels form a separate branch of the K+channel superfamily and are homologous with other K+ channels only in areas of the pore region. In situ hybridization has shown that rSK1, rSK2, and rSK3 mRNAs are broadly distributed in overlapping patterns (12, 16): rSK1 mRNA is present in brain and heart, rSK2 mRNA is present in brain and adrenal gland, and hSK4 mRNA is present in placenta and lung. We observed that mRNA encoding rSK3, but not rSK1 or rSK2, was expressed in H9c2 myotubes. In agreement with an earlier in situ hybridization study (16), we also found that rSK1, rSK2, and rSK3 mRNAs were expressed in rat brain.

The pharmacology of SK channels is distinct, in that some channels are blocked by apamin, and the apamin-sensitive channels are blocked also by d-TC. Heterologously expressed SK2 channels are blocked by apamin with an IC50 of 60 pM, whereas SK1 channels are unaffected by as much as 100 nM apamin (16). SK2 channels are also blocked by d-TC with an IC50 of 2.4 or 5.4 μM in heterologous expression systems, whereas SK1 channels are blocked byd-TC with an IC50 of 76.2 or 354 μM (11, 16). By introducing point mutations at specific sites on the cloned channels, it was shown that two amino acid residues on either side of the channel pore are the primary determinants of the sensitivity to apamin and d-TC (11): SK2 channels contain both residues, whereas SK1 channels lack the residues; SK3 channels, which exhibit intermediate sensitivity to apamin (IC50 = 2 nM), contain one of the residues.

SK channels in H9c2 cells were blocked by apamin with an IC50 of 6.2 nM and byd-TC with an IC50 of 49.4 μM. The sensitivity of H9c2 I SK to apamin was considerably lower than that of rSK2 channels but was on the same order as that of heterologously expressed rSK3 channels (11). Moreover, d-TC sensitivity of H9c2I SK is comparable to that of E330D (IC50 = 62.6 μM), an rSK1 mutant channel with one apamin-sensitive residue, similar to the rSK3 channel (11). We, therefore, conclude that myotubular H9c2I SK primarily passed through SK3 channels. Similarly, SK channels in rat chromaffin cells are blocked by apamin and d-TC with IC50 of 4.4 nM and 20 μM, respectively (22), which suggests that chromaffin cells also express SK3 channels, although only rSK2 mRNA is currently known to be expressed in rat adrenal gland (16).

Apamin-sensitive SK channels and radiolabeled apamin binding sites are expressed in denervated skeletal muscle and in skeletal muscle from patients with myotonic muscular dystrophy, but not in normal adult skeletal muscles (24, 25). In cultured rat skeletal muscle, expression of SK channels is seen in myoblasts, and its levels increase with the fusion of myoblasts into myotubes (27). This scenario is somewhat different from that in H9c2 cells, where rSK3 mRNA was detected only in multinuclear myotubes. Apamin binding can be enhanced during fusion of cultured rat skeletal muscles by suppressing spontaneous excitation through blockade of voltage-gated Na+ channels (27). The absence of innervation and spontaneous excitation may also be prerequisites for expression of SK channels in H9c2 cells.

Myogenesis in skeletal muscle is known to be regulated by skeletal muscle-specific transcription factors, such as the MyoD family of muscle-specific basic-helix-loop-helix proteins (21, 28). The expression of rSK3 mRNA in H9c2 cells was well correlated with expression of myogenin, a myogenic basic-helix-loop-helix transcription factor, the expression of which is increased when myoblasts terminally differentiate and fuse into multinucleated myotubes (1, 8, 28). Thus the expression of rSK3 in H9c2 is regulated by a differentiation program similar to that for other skeletal muscle-specific proteins known to be expressed in H9c2 myotubes (23).

The H9c2 cell line has served as a useful surrogate of cardiac and skeletal muscles. Although the cardiac phenotype is expressed in H9c2 cells [e.g., cardiac L-type Ca2+ channels are present (9, 17,26)], skeletal muscle-specific transcription factors and proteins are also expressed (Fig. 6) (17, 26). The H9c2 cell line, therefore, appears to be an interesting model for studying the regulation of gene expression of cardiac and skeletal muscle-specific proteins, although it is still not entirely clear how expression of either phenotype is regulated.


This work was supported by a Grant-in-Aid for Scientific Research on Priority Areas from the Ministry of Education, Science, Sports, and Culture.


  • Address for reprint requests and other correspondence: R. Ochi, Dept. of Physiology, Juntendo University School of Medicine, Tokyo 113-8421, Japan (E-mail: ochir{at}med.juntendo.ac.jp).

  • The costs of publication of this article were defrayed in part by the payment of page charges. The article must therefore be hereby marked “advertisement” in accordance with 18 U.S.C. §1734 solely to indicate this fact.


View Abstract