An ether -à-go-go K⁺ current, I_h-eag, contributes to the hyperpolarization of human fusion-competent myoblasts

Abstract

1. Two early signs of human myoblast commitment to fusion are membrane potential hyperpolarization and concomitant expression of a non-inactivating delayed rectifier K⁺ current, I_K(NI). This current closely resembles the outward K⁺ current elicited by rat ether-à-go-go (r-eag) channels in its range of potential for activation and unitary conductance.

2. It is shown that activation kinetics of I_K(NI), like those of r-eag, depend on holding potential and on [Mg²⁺]_o, and that I_K(NI), like r-eag, is reversibly inhibited by a rise in [Ca²⁺].

3. Forced expression of an isolated human ether-à-go-go K⁺ channel (h-eag) cDNA in undifferentiated myoblasts generates single-channel and whole-cell currents with remarkable similarity to I_K(NI).

4. h-eag current (I_h-eag) is reversibly inhibited by a rise in [Ca²⁺]_i, and the activation kinetics depend on holding potential and [Mg²⁺]_o.

5. Forced expression of h-eag hyperpolarizes undifferentiated myoblasts from −9 to −50 mV, the threshold for the activation of both I_h-eag and I_K(NI). Similarly, the higher the density of I_K(NI), the more hyperpolarized the resting potential of fusion-competent myoblasts.

6. It is concluded that h-eag constitutes the channel underlying I_K(NI) and that it contributes to the hyperpolarization of fusion-competent myoblasts. To our knowledge, this is the first demonstration of a physiological role for a mammalian eag K⁺ channel.

The skeletal muscle fibre is a postmitotic multinucleated cell formed by the fusion of mononucleated myoblasts. Myoblast fusion is essential to skeletal muscle development and repair, and understanding this process is important for myoblast-based grafts or gene therapies. Our recent work on human fusion-competent myoblasts indicates that an early step in the fusion process is a hyperpolarization of the resting potential. This hyperpolarization is a two-step mechanism involving the sequential expression of two voltage-gated potassium currents. The first is a non-inactivating delayed rectifier current (I_K(NI)) that hyperpolarizes the cell to an intermediate resting potential of approximately −32 mV (Bernheim et al. 1996). Then, slightly before fusion, an inward rectifier current is expressed (I_K(IR)), which drives the potential further down to approximately −65 mV, i.e. similar to that measured for multinucleated myotubes (Liu et al. 1998).

We recently isolated the full-length coding region of a human ether-à-go-go K⁺ channel (h-eag) from myoblasts undergoing differentiation (Occhiodoro et al. 1998). We describe here the properties of the current elicited by h-eag and present data indicating that it constitutes the K_(NI) channel responsible for the first hyperpolarization step linked to myoblast differentiation.

METHODS

Dissociation and cell cultures

Samples of human skeletal muscle were obtained during corrective orthopaedic surgery of young patients (9 months to 17 years old) without any known neuromuscular disease, in accordance with the guidelines of the ethical committee of the University Hospital of Geneva (written informed consent was obtained from patients or their legal guardians). Myoblasts (Baroffio et al. 1993) and fusion-competent myoblasts (Krause et al. 1995) were prepared as described. The differentiation-promoting medium consists of Dulbecco's modified Eagle's medium (DMEM, Gibco) supplemented with 0.5 mg ml⁻¹ bovine serum albumin, 10 ng ml⁻¹ epidermal growth factor, 10 μg ml⁻¹ insulin and 1 μg ml⁻¹ gentamicin.

h-eag expression in human myoblasts and COS-7 cells

Transfections of human proliferating myoblasts with a bicistronic h-eag-Green Fluorescent Protein (GFP) vector are described elsewhere (Occhiodoro et al. 1998). Electrophysiological recordings were made in cells labelled with GFP 36–48 h after transfection. Controls were mock-transfected using the same vector but devoid of the h-eag first cistron.

Electrophysiology

Whole-cell and single-channel (outside-out) configurations of the patch-clamp technique (Hamill et al. 1981) were used as described (Bernheim et al. 1996). Leak current subtraction procedures are discussed in the figure legends.

Whole-cell recording

Extracellular solution contained (mm): 100 N-methyl-d-glucamine chloride (NMG-Cl), 5 KCl, 3 MgCl₂, 5 Hepes, 50 NaOH, 50 acetic acid, and 8 glucose. The pH was adjusted to 7.4 with NMG. Intracellular (pipette) solution contained (mm): 110 KCl, 5 NaCl, 1 MgCl₂, 5 Hepes, 20 BAPTA and 5 glucose. The pH was adjusted to 7.4 with KOH.

Single-channel recording

Extracellular solution contained (mm): 50 NaOH, 50 acetic acid, 100 NMG-Cl, 5 KCl, 3 MgCl₂, 5 Hepes, 0.5 CaCl₂ and 0.03 nifedipine. The pH was adjusted to 7.4 with NMG. Intracellular (pipette) solution contained (mm): 70 KCl, 80 KF, 4 NaCl, 1 MgCl₂, 5 Hepes and 10 BAPTA. The pH was adjusted to 7.4 with KOH.

Chemicals

Hepes was obtained from Sigma, and BAPTA and NMG were from Fluka.

Intracellular free Ca²⁺ measurements

Intracellular free Ca²⁺ was measured using fura-2 (Molecular Probes, Inc., Eugene, OR, USA) and an Attofluor RatioVision recording system (Atto Instruments Inc., Rockville, MD, USA). Myoblasts were pre-incubated for 30 min (light protected) at room temperature (20–22°C) in a solution (mm: 150 NaCl, 5 KCl, 5 CaCl₂, 2 MgCl₂, 5 Hepes and 8 glucose) containing 1 μm fura-2 AM, then rinsed three times, and incubated for 20 min (light protected) at 37°C to allow hydrolysis of fura-2 AM. Intracellular free Ca²⁺ (and simultaneous electrophysiological) measurements were performed at 20–22°C with a pipette solution composed of (mm): 140 KCl, 5 NaCl, 1 MgCl₂, 10 Hepes, 5 glucose, 0.1 BAPTA and 50 μm fura-2. The pH was adjusted to 7.3 with KOH. Intracellular free Ca²⁺ was calculated using the equation:

where R is the ratio of the two fluorescence intensities (340 nm/380 nm excitation light), and K* is a constant that defines the calibration curve near resting [Ca²⁺]_i. R_min was measured after dialysis of myoblasts with a solution containing (mm): 110 KCl, 5 NaCl, 2 MgCl₂, 10 Hepes, 5 glucose, 30 BAPTA and 50 μm fura-2; pH 7.3. From three recordings, R_min was 0.06 ± 0.02. R_max was estimated by dialysing the myoblasts with the standard BAPTA (0.1 mm), fura-2 (50 μm) pipette solution and then lysing the myoblasts with a −150 mV potential step. R_max was 1.5, the highest of three measurements. K* was measured by dialysing the myoblasts with a pipette solution containing 105 nm free [Ca²⁺]_i (mm: 105 KCl, 7 NaCl, 2 MgCl₂, 20 Hepes, 0.5 glucose, 20 BAPTA, 6.5 CaCl₂ and 50 μm fura-2; pH 7.3). From five measurements, K* was 1270 ± 258 nm.

Statistics

All data are expressed as means ±s.e.m.

RESULTS

I_K(NI) recorded in human fusion-competent myoblasts closely resembles r-eag current

When human myoblasts are transferred to differentiation-promoting medium, they rapidly express a voltage-gated non-inactivating outward potassium current (I_K(NI); Bernheim et al. 1996). This current shares many characteristics with the rat ether-à-go-go (r-eag) potassium current (Ludwig et al. 1994; Stansfeld et al. 1996; Terlau et al. 1996), and with currents elicited by the b-eag channels recently isolated from bovine retina (Frings et al. 1998). The kinetics of activation of the r-eag current are influenced by the holding potential: hyperpolarization slows down the kinetics of activation whereas depolarization accelerates them (Ludwig et al. 1994). Another property is that [Mg²⁺]_o slows the activation of this current (Terlau et al. 1996).

Figure 1A illustrates the effect on I_K(NI) of holding potential and [Mg²⁺]_o. A fusion-competent myoblast was held at various potentials for 200 ms and then stepped to +40 mV. In nine cells, the half-maximum amplitude of I_K(NI) (I_1/2, see inset) was reached in 76 ± 5 ms and 6 ± 1 ms when the myoblast was stepped from −140 and −60 mV, respectively. Removal of extracellular Mg²⁺ accelerated I_K(NI) activation about 10-fold (Fig. 1B). I_1/2 was reached in 8 ± 1 and 1 ± 0.3 ms when the myoblast was stepped from −140 and −60 mV, respectively (n = 7).

Figure 1. Modulation of I_K(NI) by voltage, [Mg²⁺]_o and [Ca²⁺]_i.

A, whole-cell current was recorded from a fusion-competent myoblast after 2 days in differentiation medium. Cell was held at −80 mV (steady-state potential), then stepped to potentials ranging between −60 and −140 mV for 200 ms before the final step to the test potential (+40 mV). Recordings were made in a bath medium containing 3 mm Mg²⁺ and in the presence of 30 μm nifedipine and 2 μm 4-aminopyridine to block other K⁺ currents. Leak current was estimated in presence of 90 mm TEA and 5 mm Ba²⁺, and subtracted. Cell capacitance was 12 pF. The relationship between the half-activation time course (I_1/2) and the prepulses are illustrated in the inset, in which symbols indicate the means from nine cells and were connected by straight lines. B, recordings from the same cell as in A. Conditions were the same as in A except that Mg²⁺ was removed from the bath medium; inset, symbols represent the means from seven cells and were connected by straight lines. C, voltage protocol used in A and B. D, simultaneous recordings of whole-cell current during steady state holding at +40 mV (top panel) and [Ca²⁺]_i (lower panel). The fusion-competent myoblast (2 days in differentiation medium) was held at +40 mV for 3 min prior to data acquisition and perfused with extracellular medium containing 3 mm TEA to block large conductance calcium-activated K⁺ channels (Hamann et al. 1994). Ionomycin (0.5 μm) was added to the extracellular medium and perfused as indicated (horizontal bar). Leak current was estimated by linear extrapolation from the current-voltage relationship from −90 to −60 mV, and subtracted. [Ca²⁺]_i was determined using fura-2. Inset, outward current at +40 mV plotted against [Ca²⁺]_i, during the recovery phase. The continuous line is a Hill equation with IC₅₀ of 144 nm and a Hill coefficient (n_H) of 4.5. In five cells, IC₅₀ and n_H were 135 ± 28 nm and 3.6 ± 0.3, respectively.

Another characteristic of r-eag current is its blockade by an elevation of [Ca²⁺]_i (Stansfeld et al. 1996). Using the calcium ionophore ionomycin (0.5 μm) to increase [Ca²⁺]_i, a drastic reduction of I_K(NI) was demonstrated (Fig. 1D) by monitoring [Ca²⁺]_i with fura-2 and I_K(NI) amplitude at +40 mV (the inset in Fig. 1D shows I_K(NI) as a function of [Ca²⁺]_i).

These electrophysiological resemblances of I_K(NI) to r-eag current led us to search for a human eag homologue in a cDNA library made from mRNA of myoblasts undergoing differentiation (Occhiodoro et al. 1998). The isolated clone encodes an eag protein of 962 amino acids showing 97 % overall amino acid identity with the rat eag protein, confirming its identity as human eag.

Properties of h-eag current recorded in h-eag-transfected undifferentiated myoblasts

A bicistronic h-eag-GFP vector was constructed by inserting the h-eag sequence upstream of an internal ribosomal entry site (Jang et al. 1989) followed by the sequence encoding GFP (Chalfie et al. 1994). h-eag was overexpressed in undifferentiated myoblasts, which do not express active K_(NI) channels, or only at a very low density (Bernheim et al. 1996). As a control, COS-7 cells, which do not express endogenous K_(NI) channels, were transfected in parallel and results on currents elicited by h-eag in COS-7 cells were identical to those obtained with transfected myoblasts (data not shown).

Figure 2 illustrates whole-cell and single-channel characteristics of the outward current (I_h-eag) recorded in transfected undifferentiated myoblasts. This current resembles I_K(NI) observed in fusion-competent myoblasts (Bernheim et al. 1996), except for its larger amplitude. A whole-cell outward current of the amplitude shown in Fig. 2A (inset) was never observed in untransfected myoblasts, nor in mock-transfected cells or myoblasts transfected with an identical vector expressing an inward rectifier potassium channel in place of h-eag (data not shown). The mean conductance at +40 mV in h-eag-transfected myoblasts is approximately 100 times larger than currents occasionally observed in some undifferentiated myoblasts expressing a very low density of I_K(NI).

Figure 2. Properties of h-eag current in transfected human undifferentiated myoblasts.

A, recordings were performed in cells labelled with GFP 48 h after transfection of undifferentiated myoblasts with the bicistronic h-eag-GFP vector. Whole-cell outward currents were measured at the end of 600 ms steps and leak current subtracted. The leak current was determined in presence of 90 mm TEA and 5 mm Ba²⁺. Steady-state conductances were plotted against step potentials and were well described by a Boltzmann equation. Cell capacitance was 24 pF. Inset, voltage-dependent activation of h-eag channels at different voltage steps. B, membrane patches (outside-out) were excised from undifferentiated myoblasts transfected with the h-eag-GFP vector. Symbols (▪) represent the mean single-channel amplitudes obtained from all-points amplitude histograms computed from current traces at potentials between −40 and +40 mV (n = 4). Examples of single-channel recordings are shown for each potential. The continuous line is fitted using the Goldman-Hodgkin-Katz equation.

Like I_K(NI), I_h-eag activates above −50 mV (Fig. 2A) and does not inactivate during sustained depolarization (data not shown). The steady-state conductances were calculated by dividing I_h-eag measured during various voltage steps (see Fig. 2A, inset) by the driving force on potassium ions (E_K=−78 mV). The steady-state conductance can be described by a Boltzmann equation (Fig. 2A):

where V_m is the membrane potential, G_∞,max is the maximum conductance, Q is the gating charge, V_o is the voltage of half-activation, k is the Boltzmann constant and T is the temperature. Results in ten h-eag-transfected myoblasts were: Q = 1.55± 0.03 elementary charge, V_o=−4.1 ± 2.1 mV, and G_∞,max= 1401 ± 365 pS pF⁻¹. Except for G_∞,max, which is about ten times larger than for I_K(NI) in fusion-competent cells, reflecting a high level of h-eag expression in transfected cells, these results are similar to those obtained for I_K(NI) recorded in fifteen fusion-competent myoblasts (Bernheim et al. 1996).

Outside-out patch recordings were made in h-eag-transfected myoblasts in order to compare unitary I_h-eag with unitary I_K(NI) (Fig. 2B). Non-inactivating K⁺ channels were looked for at potentials between −40 and +40 mV. Ca²⁺-dependent potassium channels (Hamann et al. 1994) were blocked by 10 mm BAPTA in the pipette, and nifedipine was added to block delayed rectifier K⁺ channels (30 μm; Bernheim et al. 1996). Figure 2B represents the mean amplitudes of unitary I_h-eag obtained from four patches and plotted against voltage, with sample recordings at each voltage. A Goldman-Hodgkin-Katz (GHK) equation (Hille, 1992) was used to fit the data points. The GHK equation was adjusted with the single-channel permeability coefficient as the only free parameter ([K⁺]_i and [K⁺]_o were 150 mm and 5 mm, respectively). The quality of the fit indicates that unitary I_h-eag is carried by K⁺ ions (single-channel permeability coefficient is 3.7 × 10⁻¹⁴ cm³ s⁻¹). The elementary conductance of the channel at +40 mV is 8.4 ± 0.7 pS (E_K=−86 mV). The unitary conductance and the single-channel permeability coefficient of I_h-eag are identical to those reported previously for I_K(NI) (Bernheim et al. 1996).

h-eag current is modulated by voltage, [Mg²⁺]_o and [Ca²⁺]_i

Figure 3A illustrates that hyperpolarizing a h-eag-transfected myoblast for 200 ms before stepping to +40 mV slows I_h-eag activation kinetics, as seen with I_K(NI). In the presence of 3 mm[Mg²⁺]_o, the half-maximum amplitude of I_h-eag (I_1/2, Fig. 3A, inset) was reached in 58.2 ± 5.2 ms and 3.5 ± 0.5 ms when the myoblast was stepped from −140 and −60 mV, respectively (n = 4). As for I_K(NI), removing extracellular Mg²⁺ accelerated I_h-eag activation (Fig. 3B), with I_1/2 reached in 4.2 ± 1.8 ms and 1.5 ± 0.2 ms, when the myoblast was stepped from −140 and −60 mV, respectively (n = 4; Fig. 3B, inset).

Figure 3. Modulation of h-eagpotassium channels by voltage, [Mg²⁺]_o and [Ca²⁺]_i.

A and B, whole-cell outward currents were recorded from an undifferentiated myoblast 48 h after transfection with h-eag-GFP. Data were acquired either in the presence (A) or absence (B) of 3 mm Mg²⁺. All recordings were performed in the same conditions as for I_K(NI), with 30 μm nifedipine and 2 μm 4-aminopyridine added to the bath medium. Leak current was estimated in presence of 90 mm TEA and 5 mm Ba²⁺, and subtracted. Cell capacitance was 17 pF. Insets, half-activation time course (I_1/2) of I_h-eag were plotted against the prepulses. Symbols are the means from four cells and were connected by straight lines. C, voltage protocol used in A and B. D, effect of a rise of [Ca²⁺]_i on I_h-eag recorded from a h-eag-transfected undifferentiated myoblast. The cell was steadily held at −80 mV and then stepped to +40 mV for 400 ms. Ionomycin (0.5 μm) was added to the bath medium and perfused as indicated (horizontal bar). Current amplitudes were measured at the end of 400 ms steps and not corrected for leak current. Cell capacitance was 50 pF. Inset, current traces recorded before (▪), during (♦), and after (*) ionomycin application.

Like I_K(NI) (Fig. 1D), I_h-eag was inhibited by an increase in [Ca²⁺]_i induced by application of 0.5 μm ionomycin (Fig. 3D). Short duration (400 ms) steps to +40 mV every 10 s were preferred to the sustained holding at +40 mV used for I_K(NI) (Fig. 1D) due to the magnitude of the current that mobilized a substantial fraction of the intracellular potassium ions present. Figure 3D (inset) shows I_h-eag current traces before, during and after the ionomycin application.

Therefore, with respect to current activation, kinetics of activation-deactivation, unitary conductance and modulation by Mg²⁺ and Ca²⁺, the properties of the h-eag current elicited in transfected cells are strikingly similar to those of I_K(NI) recorded in fusion-competent myoblasts.

I_K(NI), I_h-eag and myoblast membrane resting potential

We have previously suggested that I_K(NI) plays a role in setting the resting potential of fusion-competent myoblasts to approximately −32 mV (Bernheim et al. 1996). However, no specific blocker of this current is yet available that can firmly establish this relationship. A I_K(NI)-linked hyperpolarization of fusion-competent myoblasts is supported by the analysis presented in Fig. 4A, which shows the relationship between I_K(NI) density and resting potential, determined in 134 fusion-competent myoblasts. Myoblast resting potentials ranging between −4 and −45 mV were distributed in five bins of 10 mV, and the corresponding I_K(NI) density was plotted for each bin. It can be seen that a larger hyperpolarization corresponds to a larger I_K(NI) density, as is expected for a resting potential linked to I_K(NI) expression.

Figure 4. I_K(NI), I_h-eag and membrane resting potential of myoblasts.

A, whole-cell I_K(NI) and resting membrane potentials were measured in fusion-competent myoblasts after 2–3 days in differentiation medium. I_K(NI) was measured at +40 mV. Resting membrane potentials were determined in current-clamp mode. Symbols represent the means of current density and resting potential of the indicated number of cells. The dotted line was drawn by eye. B, ▵ represents the means of I_h-eag density and resting potential measured from h-eag-transfected undifferentiated myoblasts, 48 h after transfection. I_h-eag was measured at +40 mV. Parallel measurements were performed on undifferentiated myoblasts that were mock-transfected and ▴ (error bars are not apparent because they are smaller than symbol) represents the means of the resting membrane potential and I_K(NI) density.

According to the results presented in Fig. 4A, forcing undifferentiated myoblasts to overexpress h-eag should result in a resting potential even more hyperpolarized than that of fusion-competent myoblasts, as the activation zone of the current begins near −50 mV (Fig. 2). Figure 4B shows that, due to the high density of I_h-eag (142 ± 29 pA pF⁻¹, n = 10), the resting potential of h-eag-transfected undifferentiated myoblasts is, indeed, more hyperpolarized (−51 ± 5 mV, n = 10) than fusion-competent myoblasts, and is close to the potential at which I_h-eag activates. As a control, we tested undifferentiated myoblasts that were mock-transfected. These cells had resting membrane potentials (−9 ± 1 mV, n = 9) identical to non-transfected undifferentiated myoblasts (−8 ± 1 mV, n = 218; Hamann et al. 1994).

DISCUSSION

Human myoblast fusion is preceded by major changes in membrane electrical properties. We have recently shown that an initial hyperpolarization of the resting potential is associated with the expression of a non-inactivating delayed rectifier outward potassium current, I_K(NI) (Bernheim et al. 1996). The I_K(NI) expressed by fusion-competent myoblasts closely resembles the non-inactivating outward K⁺ current elicited by r-eag channels in its range of potentials for activation, kinetics of activation-deactivation, and unitary conductance (Ludwig et al. 1994; Stansfeld et al. 1996; Terlau et al. 1996). Furthermore, as was demonstrated here, the activation kinetics of I_K(NI), like those of r-eag, depend on the holding potential and on [Mg²⁺]_o, and I_K(NI) is reversibly inhibited by a rise in [Ca²⁺]_i. These properties of I_K(NI), all shared with r-eag, prompted us to search for an eag homologue in human myoblasts undergoing differentiation and led to the isolation of the complete coding sequence of h-eag (Occhiodoro et al. 1998).

h-eag constitutes the K_(NI) channel expressed in fusion-competent myoblasts

Expression of the human eag sequence produces a functional non-inactivating delayed rectifier potassium channel. The electrophysiological properties of I_h-eag were analysed in undifferentiated myoblasts transfected with a h-eag expression vector, and were found to be indistinguishable from those of the endogenous I_K(NI) expressed in fusion-competent myoblasts, except for G_∞,max, which simply reflects overexpression of the channel.

Single-channel recordings of I_h-eag demonstrate that K⁺ is the carrier and that the elementary conductance of I_h-eag is identical to that of I_K(NI) (Bernheim et al. 1996). Moreover, the holding potential and [Mg²⁺]_o both influence the activation kinetics of I_h-eag, and a provoked increase in [Ca²⁺]_i blocks I_h-eag. Thus electrophysiological analyses of I_h-eag suggest that h-eag is the molecular entity constituting the endogenous K_(NI) channel expressed at the onset of human myoblast fusion.

I_K(NI), I_h-eag and myoblast membrane resting potential

When human myoblasts are induced to differentiate, their resting potential rapidly drops from about −8 mV to approximately −32 mV. Data presented here confirm that this hyperpolarization is due to the sudden increase in I_K(NI). Analysis of current density versus resting potential of fusion-competent myoblasts demonstrates that the higher the current density, the more hyperpolarized the cell. As current density increases, the resting potential clearly moves towards the range of potentials at which K_(NI) channels activate, indicating that the balance between I_K(NI) and leak current (estimated reversal potential near 0 mV; Bernheim et al. 1996) is in favour of I_K(NI). The fact that the resting potential of undifferentiating myoblasts forced to overexpress I_h-eag reaches a value (−51 ± 5 mV; n = 10) near the threshold for I_h-eag activation further strengthens our assumption that the first hyperpolarization step observed at the onset of myoblast commitment to fusion is due to the expression of h-eag channels.

The resting potential of proliferating cells, in general, is depolarized, while more negative potentials have been observed to be associated with differentiation and cell-cycle arrest (Binggeli & Weinstein, 1986). This has raised the hypothesis that the G1-S transition may be inhibited by highly negative values of membrane potential (Binggeli & Weinstein, 1986; Day et al. 1993; Arcangeli et al. 1995). Observations that r-eag currents heterologously expressed in Xenopus oocytes are suppressed by mitosis-promoting factor (Brüggemann et al. 1997) and that an eag current expressed in the human neuroblastoma SH-SY5Y cell line is down-regulated when the cells are blocked in the G0/1 phase by exposure to retinoic acid (Meyer & Heinemann, 1998) suggest that eag channel activity may be linked to the cell-cycle. In human myoblasts, evolution of the resting potential towards hyperpolarized values following h-eag/I_K(NI) expression, coincides with cell-cycle arrest, and appears to be an important event to favour their commitment to the differentiation pathway.