J. Physiol. (1975), 244, pp. 129-143 With 5 text-figures Printed in Great Britain
129
DEVELOPMENTAL CHANGES OF MEMBRANE ELECTRICAL PROPERTIES IN A RAT SKELETAL MUSCLE CELL LINE
BY YOSHIAKI KIDOKORO From the Salk Institute, P.O. Box 1809, San Diego, California 92112, U.S.A. (Received 19 April 1974) SUMMARY
1. The developmental changes of the cell membrane electrical properties were studied with micro-electrodes in a rat skeletal muscle cell line. 2. The resting potentials in myoblasts were -71 + 3 mV (mean + S.D.) and those in myotubes which were formed by fusion of myoblasts were -69 + 3 mV. There was no developmental change in the resting potential during the period examined. 3. The ionic mechanism for the resting potential was the same in myoblasts and myotubes. The resting membrane was almost exclusively permeable to K ions, while permeability to Na ions was not detectable. There was a small permeability to Cl ions. 4. The specific membrane resistance and capacitance were 8 kQ. cm2 and 1 #sF/cm2 for myoblasts and 12 kQ. cm2 and 5 ,uF/cm2 for myotubes,
respectively. 5. Action potentials in myoblasts were evoked by anode break stimulation. They were small and did not overshoot zero membrane potential. The action potentials in myotubes were larger, and had an average overshoot of 32 + 7 mV and a maximum rate of rise of 93 + 28 V/sec. 6. The current-voltage relation was examined. Delayed rectification was found in myotubes but not in myoblasts. INTRODUCTION
Recent advances in tissue culture techniques have made it possible to grow and differentiate muscle cells and neurones in vitro. Clonal cell lines from various origins are now available for study. A clonal muscle cell line is a suitable system for studying the developing excitable membrane. Cells can be identified at early stages in development, which is difficult for cells in vivo. Cells grown in a monolayer are clearly seen with a phase contrast microscope, which makes it easy to penetrate the cells with 5-2
YOSHIAKI KIDOKORO micro-electrodes. The cellular homogeneity of clonal lines is obviously advantageous for biochemical work such as changes in enzyme and synthesis of specific proteins during development (Schubert, Tarikas, Humphreys, Heinemann & Patrick, 1973), and this system should be useful for a future analysis of the molecular structure responsible for the action potential. A skeletal muscle cell line was used in this study of action potential development because it is easy to define the developmental stage of individual cells on the basis of morphology. Myogenic precursor cells (myoblasts) divide and increase their number exponentially (Yaffe, 1968). When they become confluent in the tissue culture dish, they fuse and form multinucleate myotubes. The production of myosin heavy chain was found to increase five- to sevenfold after myoblasts fused forming multinucleate myotubes. Cross-striation has been observed in some of the old myotubes (Schubert et al. 1973). Multinucleate myotubes are found to be electrically excitable (Harris, Heinemann, Schubert & Tarikas, 1971). The morphological, biochemical and electrophysiological differentiation of this clonal cell line shows that it has retained the essential developmental properties of normal myogenic cells. A preliminary report of the present work has been published (Kidokoro, 1973). 130
METHODS Cell culture. A muscle cell line derived from rat thigh muscle and adapted to culture conditions by Yaffe (1968) was used in these experiments (line L6). Myoblasts were grown in plastic tissue culture dishes in modified Eagle's medium containing 10% foetal calf serum at 370 C in an atmosphere of 12% C02 and 88 % air (Vogt & Dulbecco, 1963). The development of the cultured muscle cell was morphologically classified into two stages: the stage of mononucleate myoblasts in which the cells increase their number exponentially, and the stage of multinucleate myotubes which are formed by fusion of myoblasts. Some of the myotubes become partially striated and some of them contract spontaneously. Electrophysiology. Cells for electrophysiology experiments were grown attached to a round cover glass (diameter 25 mm) at the bottom of a tissue culture dish (diameter 60 mm). After the appropriate time of incubation, the cover glass was picked up and placed in a glass-bottomed chamber containing about 3 ml. of a saline. In some experiments cells grown in tissue culture dish were used in 8itu. The composition of the salines used are listed in Table 1. Myoblasts were examined at least 2-3 days after they attached to the bottom of the culture dish. The chamber was placed on the stage of a compound microscope (Reichart, Austria) which was modified to move the objective lens instead of the specimen stage for focusing (McBain Instruments, Inc., Calif., U.S.A.). The cells were observed with a 40 x water immersion phase contrast objective and 10 x ocular lens (Carl Zeiss, West Germany). Glass micro-electrodes were filled with 3 M-KC1 and had resistances of 80-150 MCI. Tippotentials of these electrodes were between -2 and - 20mV. In different solutions the tip potentials changed slightly (+ 7 to -4 mV). For penetration the microelectrode tip was pressed on the cell making a small dimple. A gentle tap on the base
DEVELOPMENTAL CHANGES OF MUSCLE MEMBRANE 131 of the manipulator (Leitz, West Germany) or on the table was enough to push the electrode tip into the cell. Usually two electrodes were placed inside a cell, one for passing current and the other for recording the resulting potential. Each of them was connected to the input stage of a preamplifier (M-4A, W. P. Instruments, U.S.A.) which included a bridge circuit for passing current and recording voltage simultaneously. Sometimes the bridge circuit to pass current and record resulting potentials was used with a single electrode to reduce cell damage. The bridge circuit could be balanced when a small amount of current (less than 0 5 nA) was used and small cells which had a large input impedance were chosen. The current passed through the electrode was measured by a conventional current-voltage converter circuit having a feed-back resistance of 1 or 10 MCI, inserted between an Ag-AgCl pellet in the bath and earth. When the external Cl concentration was changed, an agar-3 m-KC1 bridge was used between the bath and the Ag-AgCl pellet. The value of 19 V/sec for the maximum rate of rise for myotube action potentials given in a preliminary report (Kidokoro, 1973) was found to be an underestimate for several reasons: (1) cells which had been penetrated with two electrodes had a smaller maximum rate of rise than those which had been penetrated with a single electrode: this can be explained by damage to the membrane; (2) the stray capacity of high resistance micro-electrodes (100 MQ or more) had not been appropriately compensated. The rise time of the potential response due to a square pulse of current was 200-300 lusee after proper adjustment; (3) the resistance and capacitance which limit the frequency characteristics of the differentiator were not appropriate to record high frequency components. The differentiator in the preliminary report had highest gain at 50 Hz. For the present report the peak gain was at 50 kHz. Myoblasts and myotubes had variable shapes. If it is assumed that the membrane potential is equal everywhere inside the cell in the steady-state condition, the specific membrane resistance can be calculated from the effective input resistance and the cell surface area. Therefore, cells were chosen which were less than 300 #m in length and were relatively thick. The thickness of these cells was 8-15 jam in the region where the nuclei were. This was measured by the travel of the microscope fine adjustment between the position at which the upper surface of the cell was focused and that at which the glass surface was focused. A cylinder 300 ,gm in length, 8 #am in diameter, with an internal resistance of 180 L. cm (Albuquerque & Thesleff, 1968) and an input impedance of 50 MO, was chosen as a model to calculate the steady-state potential distribution inside a cell, using a short cable model (Weidmann, 1952). It was found that the steady-state potential at the end of the cell would be 98 % of the potential recorded at the middle of the cell where current was injected. The experimental cells selected for calculations were smaller in length and larger in width than the model, which favours an even better isopotential condition. However, in the case of myoblasts, the peripheral part of the cell was usually thin, which might cause an error in calculation. Since myoblasts (Schubert, Harris, Heinemann, Kidokoro, Patrick & Steinbach, 1974) and possibly myotubes are electrically coupled with each other, it was necessary to choose a cell well isolated from others. A single electrode was used both for passing current and recording membrane voltage. If the electrodes were carefully chosen and only a small amount of inward current (less than 0-5 nA) was used, the resistance of the electrodes behaved in a linear fashion. After a proper compensation of electrode capacity, the voltage transient had a time constant of less than 1 msec when an electrode was in the saline, whereas the cell membrane time constant was more than 6 msec in myoblasts and more than 20 msec in myotubes. Therefore it was not difficult to obtain a well-balanced bridge circuit. The small size of the cells grown in tissue culture is advantageous since they have high input impedances. The input
132
YOSHIAKI KIDOKORO
resistance increases after penetration for 2-5 min with a slight increase in the resting potential (5-10 mY). Therefore, at least a few minutes were allowed for the input resistance to reach the highest level. To calculate specific membrane constants, the cell was measured from an enlarged micrograph and the whole surface area was taken as twice the value measured. Since cultured cells are flat, this simplification should not produce a large error. All experiments were done at room temperature (20-24O C).
RESULTS
Resting membrane potentials. The tissue culture cells were fragile. In most cases, especially in myoblasts, the resting potential decreased quickly after penetration and cells appeared damaged under the microscope. There were occasions in which the resting potential stayed at its largest value for more than one minute and the tip potential of the recording electrode did not change more than 10 mV before and after the penetration. On such occasions, the average resting potential for myoblasts was -71 mV in normal saline, which was similar to that found for myotubes, -69 mV (mean and S.D. of the resting potential in various solutions are listed in Table 2). When the Ca concentration in the external saline was increased to 10 mm, the average resting potential was -74 mV in myoblasts and -72 mV in myotubes. It appeared to be easier to obtain a successful penetration in 10 mM-Ca saline. The ionic mechanism for the resting potential was studied by changing the external saline with solutions having various ionic compositions. Na ions. In order to determine the contribution of Na ions to the resting potential, the membrane potential was measured in Na-free saline where all NaCl was replaced by equiosmolar Tris-HCl or choline chloride. The average resting potential was -72 mV in myoblasts and -73 mV in myotubes. There was no significant difference in resting potentials between normal and Na-free saline. This was true in either 1-8 or 10 mm external Ca concentration (Table 2). K ions. The dependence of the resting potential on K ion concentration was examined in the absence of Na ions in order to avoid any complication due to a permeability to Na ions. Na-free saline and K-saline (Table 1) were mixed to obtain the desired concentration of K ions. The resting membrane potential was dependent upon the K ion concentration. When the resting potential was plotted against the logarithm of K concentration, it was linear both in myoblasts and myotubes (Fig. 1). The change was 50 mV/decade of K concentration in myoblasts and 52 mV/decade in myotubes. When the line was extrapolated to zero membrane potential, the internal K concentration could be estimated as 151 mm in myoblasts and as 150 mM in myotubules. The K equlibrium potential in normal
DEVELOPMENTAL CHANGES OF MUSCLE MEMBRANE 133 medium was calculated as -83 mV which is more negative than the measured resting potential. The discrepancy may be explained by a leak of the membrane due to micro-electrode penetration. 0
10
Myoblast
g
E~~~~~
a.
.0 -50
K concentration (mM) 0 100
10
100
Myotube
/
C~~~~~
-50
/
*/ II/ Fig. 1. Relation between external K concentration and resting membrane potentials in myoblasts and myotubes. K concentration is plotted on abscissa in logarithmic scale. Straight lines were drawn by eye. TABLE 1. Composition of salines used (mm) Ca Na K Tris 1.8 157-0 5-6 2-0 Normal saline 0 Na-free saline 5-6 198-0 1P8 144-0 100 2-0 5-6 High Ca saline 0 K saline 162-6 2-0 1P8 10 0 144-0 5-6 2-0 High Ca, Cl-free saline * MeSO3; methane sulphonate
Cl, 167-9 173-5 171-3 167-9 0
MeSO3* 0 0 0 0 171-3
C1 ions. The cells were immersed in Cl-free saline for more than 10 min to allow equilibration to occur. The resting potential in Cl-free media with 10 mM-Ca was -75 mM in myoblasts and -72 mV in myotubes. These values were not different from those in control solution either in myoblasts or in myotubes (Table 2). It is known for frog skeletal muscle that the Cl permeability is greater than that of K and when the external Cl concentration is decreased at constant K concentration, internal KCl (with water) leaks out to attain a new equilibrium [K]0 [Cl]o = [K]i[Cl]t. The resting potential, therefore, depolarizes suddenly at the time of solution change, and gradually returns to the resting potential (Hodgkin & Horowicz, 1960). It is possible to perfuse the external solution while the recording and current electrodes are maintained in a myotube. When the solution was
134 YOSHIAKI KIDOKORO changed from normal to Cl-free saline the resting potential did not change significantly but a slight increase in effective input resistance (10-20%) was noticed. Thus, in the myotube membrane Cl ions are less permeable than K ions. TABLE 2. Resting potentials of myoblasts and myotubes in various salines
Myoblasts
Myotubes
mean + S.D.
mean + S.D.
-69 ± 3 (18) mV -71 ± 3 (20)* mV Normal saline -72+2 (11) -74+ 2 (9) High Ca saline -73 + 3 (12) -72 + 4 (8) Na-free saline High Ca -75±4 (10) -72±2 (10) Cl-free saline High Ca -72±3 (12) -69±3 (8) Na-free saline * Figures in parentheses are the number of cells examined.
Specific membrane resistance and capacitance. The input resistance ranged from 53 to 140 MU in myoblasts and 80 to 120 MCI in myotubes (RI in Table 3). The specific membrane resistance (Rm) was calculated by multiplying the surface area with input resistance. The calculated (Rm) was 8'1 + 2*1 kQ. cm2 for myoblasts and 12-3 + 4-2 kf2. cm2 for myotubes. The potential transient caused by a square pulse of current fit a single exponential curve, showing that the isopotential condition is maintained within the limit of this detection procedure. The time constant (r) was 5-7-10 0 msec in myoblasts and 26-5-123-0 msec in myotubes (Table 3). The membrane capacitance was calculated by dividing the time constant (Xr) by the specific membrane resistance obtained above. The specific membrane capacitance for myoblasts was 1.0 + 0-2 ,uF/cm2 and 4*7 + 2-1 ,uF/cm2 for myotubes. Action potentials. In the following experiment, two intracellular electrodes were used to avoid artifacts due to improper bridge-balance which tended to occur when a large amount of current was passed (more than 1 nA). In myoblasts, the resting potential was small (-30 to -50 mV) when two electrodes were placed inside the cell. When the membrane was hyperpolarized to about - 100 mV, a small spike appeared after the end of the current pulse (Fig. 2A1). There was a clear inflexion at the foot of the spike (indicated by an arrow), consistent with the notion that this was a regenerative potential. The inflexion was at -30 to -40 mV. The average spike height measured from this inflexion to the top was 23 mV, so the spike did not overshoot beyond zero membrane potential. The maximum rate of rise was 3-13 V/sec. These small action potentials could be the result of damaging the cell membrane, but even when the resting potential
DEVELOPMENTAL CHANGES OF MUSCLE MEMBRANE 135 was about -50 mV, the action potential in myoblasts was much smaller than that in myotubes under comparable conditions. The action potential was larger and overshot zero membrane potential in multinucleate myotubes (Fig. 2B1). The resting potential was -50 to -65 mV with two electrodes in a cell. Again an action potential could be evoked only by anode-break excitation. The inflexion at the foot of the action potential was at -65 + 4 mV (indicated by an arrow). The amount of TABLE 3. Electrical membrane properties in myoblasts and myotubes
Myoblasts 1 2 3 4 5 Mean ± S.D.
Rest (mV) -57 - 46 - 39 -62 -37
RI (MQ) 104 53 84 140 110
r (msec)
Rm (ku)
6.5 7.8 10.0 5.7 7-1
11.4 5.8 8.4 7-0 7.7 8.1+2-1
Cm
(#ttF/cm2) 06
1*3 1.2
0*8 0*9
10±+0.2
Myotubes 1 2 3 4 5* Mean ± S.D.
-50 -58 -58 -67 -55
87 93 80 110 120
26-5 29-8 56 0 55.0 123.0
5.9 14.0 15*5 10.1 15*8 12.3±4.2
4-5
2*1 3.6 5-4 7.8
4.7+2±1
* Saline buffered with 5 mm HEPES-NaOH.
overshoot and maximum rate of rise of the action potential were dependent upon the amount of the preceding hyperpolarization (Fig. 2B1, 2). When the membrane potential at the steady state of hyperpolarization was plotted on the X axis and percent of the maximum rate of rise on the Y axis, a curve analogous to the inactivation curve in voltage clamp experiments in squid axon and other excitable membranes was obtained (Fig. 3). The potential at which half of the maximum rate of rise could be obtained was -69 to -86 mV. This shows that inability to evoke an action potential with depolarizing current was due to an inactivation process going on at the resting potential. The average overshoot of the maximal action potential was 32+7 mV (n = 15) and the average maximum rate of rise was 93 + 28 V/sec (n = 15) (these values were obtained with a single electrode for recording and injecting current to reduce damage). As shown in Fig. 2B1, the action potential had a long duration of 62+ 30 msec (n = 10) measured at half height. The action potential was composed of a fast initial spike, followed by a relatively slower hump (marked with a dot). The spike was followed by a hyperpolarizing potential
YOSHIAKI KIDOKORO 136 (positive after-potential) of 1030 mV in amplitude and about 5 sec in duration, which indicated that delayed rectification had developed at this stage. The duration was shorter (about 2 sec) when the myotube was old (about 1 month after fusion). Still it is considerably longer than that found in adult rat muscles (Albuquerque & Thesleff, 1968). A.
B.
Myoblast
Myotube
2
3
3 ~~~~t
I
_nA
-
I
_A
4~~~~~ 2OmVse
100 nisec
2m
Fig. 2. Responses to various applied currents in a myoblast (A) and a myotube (B). Upper trace in each pair of records shows current (downward deflexion indicates inward current), and zero membrane potential level. Lower trace is membrane potential record. Calibration shown in A4 applies for all A records. Calibration shown in B4 applies for all B records.
There was no obvious effect of increasing the Ca concentration (10 mm) myoblast action potential. Myotube action potentials in high Ca solution often have two peaks. The action potential was analysed and it was found that the initial peak is due to the inward movement of Na ions and the later peak to that of Ca ions (Kidokoro, 1975). The threshold for spike initiation in 10 mm Ca (measured by the level of inflexion at the foot on the
DEVELOPMENTAL CHANGES OF MUSCLE MEMBRANE 137 of an action potential indicated by an arrow in Fig. 2B1) was less negative than that in normal saline by 14 mV. The 'inactivation curve', obtained as described above, was shifted to the left along the voltage axes by about 10 mV (Fig. 3). The potential at which half of the maximum rate of rise could be obtained was -67 to -72 mV. The effect of high Ca on the K system (delayed rectification) was not studied. 10mm-Ca
18mM-Ca
100
0'1~
%-.
0
._
50 E E
%& o
0 0
-50 -100 Membrane potential (mV)
Fig. 3. Relation between maximum rate of rise of myotube action potentials and the level of preceding hyperpolarization in normal (open circles) and 10 mm-Ca saline (filled circles). Maximum rate of rise is standardized to the largest value and plotted on the Y axis. The membrane potential level is plotted on the X axis. Open arrow at the X axis indicates the resting potential in normal saline (-60 mV) and filled arrow that (-52 mV) in 10 mM-Ca saline.
Current-voltage relation. The steady-state current-voltage relation was examined in myoblasts and myotubes. Myoblasts were penetrated with two electrodes which usually lowered the resting potential (-30 to -56 mV). However, the input impedance of some of the cells was 10-49 MU which was not very low compared with the input impedance measured with single electrodes, considering some of the cells chosen for this experiment were not completely isolated and could have been electrically coupled to nearby cells. Therefore, the I- V curve obtained is probably valid and is not distorted too much by the existence of a leak due to
138 YOSHIAKI KIDOKORO micro-electrode penetration. Between - 100 and 0 mV, the I-V curve was straight (Fig. 4A). However, at a membrane potential above 0 mV, the resistance increased 2-7 times over that in the resting state. The myoblasts did not exhibit delayed rectification. This is not due to a small resting membrane potential (-30 to -56 mV) which can inactivate the delayed rectification (Nakajima, Iwasaki & Obata, 1962). Even in cells which have resting potentials between -60 and -70 mV (single electrode penetration), a large depolarizing current did not cause any hyperpolarization after the pulse as in myotubes. On the other hand, A
mV 50
,,
O
*
8
mV 50 0~~~~~~~~~~
,
5
-s
; -50 *
B
Depolarization
..
0°
.0
-
nA
00
so-50
-100
5
*
nA
*
-100
Hyperpolarization
Fig. 4. Current-voltage relation in a myoblast (A) and a myotube (B). Current is plotted on the X axis (outward current is plotted as plus), membrane potential on the Y axis. The potential change was measured 250 msec after the onset of current pulse in the myoblast and 250 msec (open circles), and 500 msec (filled circles) in the myotube. The intersection of the I-V curve with the Y axis shows the resting potentials (-48 mV in the myoblast and -50 mV in the myotube).
delayed rectification was clearly demonstrated in multinucleate myotubes by the I-V curve. When the depolarization exceeded -40 mV (-34 to -46 mV), an undershoot after the pulse occurred. When small pulses of inward current were applied during the after-spike hyperpolarization (positive after-potential), the potential change decreased to 7-22 % of that in the resting state. This is probably an underestimate of the conductance increase during the after-spike hyperpolarization (positive after-potential), since the isopotential condition may not be satisfied during the conductance increase. The potential level at the peak of the hyperpolarization was -70 to -77 mV. After an action potential or depolarization by a brief current pulse, the hyperpolarization lasted 4-8 sec. The rectification started
DEVELOPMENTAL CHANGES OF MUSCLE MEMBRANE 139 with a considerable delay which resulted in a large initial hump in response to a depolarizing current pulse (Fig. 2B4). Anomalous rectification was observed in the hyperpolarizing region of the steady-state I-V curve (Fig. 4B, filled circles). The slope conductance was higher in this region than in the resting potential region. Anomalous rectification was clear in the I-V curve in which membrane potentials were measured at 250 msec after the onset of current pulse, at which time delayed rectification had not yet started (Fig. 4B, open circles). The slope conductance for the inward current was larger than that for the outward current. A
B Depolarization
rV
mV 50 -
50-
/
1
nA
S -5
o
5
nA
-so
-S.0 Hyperpolarization
Fig. 5. Current-voltage relation in high K saline ([K]o = 1108 mM) inamyoblast (A) and a myotube (B). The voltage change was measured 250 msec after the onset of the current pulse in A and after 300 msec in B.
Currentvoltage relaim in high K saline. When myoblasts were soaked in high K saline (K, 110-8 mM) the resting potential was -2 to -8 mV and the steady-state I-V curve was straight between -40 and + 40 mV (Fig. 5A). On the other hand, in similarly treated myotubes the steadystate I-V curve showed a slope resistance to outward current 6-8 times larger than that to inward current (Fig. 5B). In all three cases examined, the slope resistance in the region above + 50 mV became smaller, as has been observed in frog skeletal muscle (Nakajima et al. 1962). DISCUSSION
It was found that in this muscle cell line, myoblasts and myotubes had similar resting potentials. In a previous study the resting potentials of myotubes and striated fibres were found to be the same in this tissue culture system (Kidokoro, 1973). There was no developmental change of resting potentials during the period examined. Boethius & Knutsson (1970) reported a gradual increase of resting potentials during differentiation of chick skeletal muscle from the 3rd day in ovo to the adult. The furthest
140 YOSHIAKI KIDOKORO developed stage examined here probably corresponds to their myocytes at 19th day in ovo where nuclei have migrated to the cell periphery. In their report, resting potentials earlier than the 15th day in ovo were distributed in a wide range and were small in average. Still they recorded around -60 mV resting potentials in several cells at these stages. The resting potentials of cells in which the micro-electrode penetration does not cause large damage do not vary greatly (for example - 82-4 + 2-4 mV in frog twitch muscle according to Stefani & Steinbach, 1969). The real resting potential in the early developmental stages might be more negative than -60 mV and the potentials recorded by Boethius & Knutsson (1970) might be small because of mechanical damage due to the relatively large electrodes (10-20 MU) they used. Thus, there may be no significant increase of the resting potential of chick muscle cells during development in ovo. The early developmental change of the resting potential in chick skeletal muscle has also been studied in culture (Fischbach, Nameroff & Nelson, 1971). They reported a small resting potential (-10 to -15 mV) in mononucleate muscle precursor cells and a relatively large resting potential (-60 to -80 mV) in old multinucleate myotubes. Similar results have been reported in rat skeletal muscle in vitro (Fambrough & Rash, 1971). Fischbach et al. claim that there is an increase in resting potential during development. The discrepancy in the resting potentials in their report and in the present one might be due to the differences in cell cultures. They used primary cultures whereas here a clonal cell line was used. There is a possibility that cells in a clonal cell line have different properties than those dissected from a foetus. The resting membrane of both myoblasts and myotubes is predominantly permeable to K ions. The permeability to Na ions seems to be small. At this stage in development the muscle membrane was relatively impermeable to Cl ions. There are no data concerning Cl ion permeability in adult rat muscle, but it is known that in goat muscle the Cl conductance is 85 % of the total. A similar proportion of Cl conductance has been found in human muscle (Bryant & Morales-Aguilera, 1971). It is well known that Cl ions are more than twice as permeable as K ions in frog sartorius muscle (Hodgkin & Horowicz, 1960; Hutter & Warner, 1967). Therefore it is reasonable to assume that Cl ions are more permeable than K ions in adult rat muscle. The resting membrane of a barnacle muscle fibre is mostly permeable to K ions in a solution of pH 7-7 (Hagiwara, Toyama & Hayashi, 1971). On the other hand, low pH is known to suppress the Cl permeability in frog sartorius muscle (Hutter & Warner, 1967). The normal saline used in this experiment has only 2 mm Tris-HCl which might be too weak to maintain the pH especially in the vicinity of the cells. An increase
DE VELOPMENTAL CHANGES OF MUSCLE MEMBRANE 141 in the buffer concentration to 5 mm (HEPES-NaOH) did not cause any increase in membrane conductance (myotube no. 5 in Table 3). The specific membrane resistance and capacitance of myoblasts were found to be 8 kQ. cm2 and 1 /tF/cm2. The capacitance increased to 5 ,cuF/cm2 in myotubes whereas the membrane resistance (12 kQ. cm2) was only slightly higher than that of myoblasts. The large capacitance is probably due to the development of T-system in myotubes as seen by electron microscopy (Ezerman & Ishikawa, 1967). The membrane resistance and capacitance for chick skeletal muscle in primary cultures are reported to be 2 6 kQ . cm2 and 3 9 /XF/cm2 respectively (Fischbach et al. 1971). Those for tissue cultured mouse muscles are 0-7 kQ.cm2 and 8-4,uF/Cm2 respectively (Powell & Fambrough, 1973). The membrane resistance given here for myotubes is large compared with these values, while the capacitances are more or less similar. The large membrane resistance can be explained by the low Cl conductance. The membrane resistance of goat external intercostal muscle in low Cl medium is reported to be 1O kQ. cm2 (Bryant & Morales-Aguilera, 1971), which is reasonably close to the value obtained here. Therefore, the low values of membrane resistance obtained in tissue cultured chick and mouse muscle cells may result from either a high Cl permeability or membrane damage by the micro-electrode penetration. A small regenerative potential was seen in myoblasts which are dividing exponentially. The height of the action potential increased abruptly after the time of fusion, since an overshooting action potential was observed even in binucleate cells. The largest overshoot (+ 43 mV) found in myotubes is not different from that in vivo (Albuquerque & Thesleff, 1968), suggesting that the Na concentration inside the cell is similar in both cases. The maximum rate of rise of action potential in myotubes (93 V/sec at 20-24o C) may be comparable to that of rat skeletal muscle after denervation, around 300 V/sec at 290 C (Redfern & Thesleff, 1971; Harris & Thesleff, 1971), when the difference in temperature is taken into account. No delayed rectification was found in myoblasts, although a mechanism for voltage sensitive inward current has been demonstrated. The later development of delayed rectification compared with voltage sensitive inward currents is also found in a tunicate muscle (Miyazaki, Takahashi & Tsuda, 1972; Takahashi, Miyazaki & Kidokoro, 1971). The delayed rectification found in myotubes is qualitatively similar to that of muscles in vivo, but is much slower (Nakajima et al. 1962). Inward-going rectification in myoblasts is different from that found in frog skeletal muscle (Katz, 1949) in that the I-V curve in normal saline was straight between -80 and 0 mV and in high K saline the curves were also straight between -40 and + 40 mV. On the other hand, in myotubes
142 YOSHIAKI KIDOKORO I-V curves showed inward-going rectification similar to 'anomalous rectification' found in frog muscles (Katz, 1949). The author thanks Drs S. Hagiwara, A. D. Grinnell, S. Heinemann, K. Takahashi and B. L. Brandt for their comments in preparing this manuscript. This work was supported by a Muscular Dystrophy Associations of America grant and a National Science Foundation grant to Dr Stephen Heinemann and a Sloan Foundation Grant to the Neurobiology Department, Salk Institute. REFERENCES
ALBUQUERQUE, E. X. & THESLEFF, S. (1968). A comparative study of membrane properties of innervated and chronically denervated fast and slow skeletal muscles of the rat. Acta phy8iol. 8cand. 73, 471-480. BOETHUS, J. & KNUTSSON, E. (1970). Resting membrane potential in chick muscle cells during ontogeny. J. exp. Zool. 174, 281-286. BRYANT, S. H. & MORALES-AGUILERA, A. (1971). Chloride conductance in normal and myotonic muscle fibres and the action of monocarboxylic aromatic acids. J. Physiol. 219, 367-383. EzEmRM., E. G. & Ism AwA, H. (1967). Differentiation of the sarcoplasmic reticulum and T system in developing chicks skeletal muscle in vitro. J. cell Biol. 35, 405-420. FAMBROUGH, D. & RASH, J. E. (1971). Development of acetylcholine sensitivity during myogenesis. Devl Biol. 26, 55-8. FISCHBACH, G. D., NAMEROFF, M. & NELSON, P. G. (1971). Electrical properties of chick skeletal muscle fibres developing in cell culture. J. cell Phy8iol. 78, 289-300. HAGIWARA, S., ToYAmA, K. & HAYASHI, H. (1971). Mechanism of anion and cation permeations in the resting membrane of a barnacle muscle fiber. J. gen. Phy8iol. 57, 408-434. HARRIS, A. J., HEiNEm.&NN- S., SCHUBERT, D. & TAREis, H. (1971). Trophic interaction between cloned tissue culture lines of nerve and muscle. Nature, Lond. 231, 296-301. HARRis, J. B. & THESLEFF, S. (1971). Studies on tetrodotoxin resistant action potentials in denervated skeletal muscle. Acta physiol. 8cand. 83, 382-388. HODGKIN, A. L. & HoRpowcz, P. (1960). The effect of sudden changes in ionic concentrations on the membrane potential of single muscle fibres. J. Phyaiol. 153, 370-385. HUTGER, 0. F. & WARNER, A. E. (1967). The pH sensitivity of the chloride conductance of frog skeletal muscle. J. Physiol. 189, 403-425. KATz, B. (1949). Les constantes 6lectriques de la membrane du muscle. Arch8 Sci. physiol. 3, 285-300. KIDOKORO, Y. (1973). Development of action potentials in a clonal rat skeletal muscle cell line. Nature, Lond. 241, 158-159. KIDOKORO, Y. (1975). Sodium and calcium components of the action potential in a developing skeletal muscle cell line. J. Physiol. 244, 173-187. MIYAZAKI, S. I., TAKAHASHI, K. & TSUDA, K. (1972). Calcium and sodium contributions to regenerative responses in the embryonic excitable cell membrane. Science, N.Y. 176, 1441-1443. NAKA.IMA, S., IWASAKI, S. & OBATA, K. (1962). Delayed rectification and anomalous rectification in frog's skeletal muscle membrane. J. gen. Phy&iol. 46, 97-115. PowELL, J. A. & FAMBROUGH, D. M. (1973). Electrical properties of normal and dysgenic mouse skeletal muscle in culture. J. cell Phy8iol. 82, 21-38.
DEVELOPMENTAL CHANGES OF MUSCLE MEMBRANE 143 REDFERN, P. & THESLEFF, S. (1971). Action potential generation in denervated rat skeletal muscle. I. Quantitative aspects. Acta phy8iol. 8cand. 81, 557-564. SCHUBERT, D., HARRIS, A. J., HEn EmANN, S., KIDOKORO, Y., PATRICK, J. & STEIINBACH, J. H. (1974). Differentiation and interaction of clonal cell lines of nerve and muscle. In Tisue Culture of the Nervou8 Sy8tem, ed. SATO, G. London: Plenum Press. SCHUBERT, D., TARTIAs, H., HUMPHREYS, S., HEINEMANN, S. & PATRICK, J. (1973). Protein synthesis and secretion in a myogenic cell line. Devl Biol. 33, 18-37. STEFANI, E. & STEINBACH, A. B. (1969). Resting potential and electrical properties of frog muscle fibres; effect of different external solutions. J. Physiol. 203, 383-401. TAKAHASHI, K., MIYAzAKI, S. & KIDOKORO, Y. (1971). Development of excitability in embryonic muscle cell membrane in certain tunicate. Science, N. Y. 171, 415-418. VOGT, M. & DuiLBECCO, R. (1963). Steps in the neoplastic transformation of hamster embryo cells by polyoma virus. Proc. natn. Acad. Sci. U.S. 49, 171-179. WEIDMANN, S. (1952). The electrical constraints of Purkinje fibres. J. Phy8iol. 118, 348-360. YAFFE, D. (1968). Retention of differentiation potentialities during prolonged cultivation of myogenic cells. Proc. natn. Acad. Sci. U.S.A. 61, 477-483.
129
DEVELOPMENTAL CHANGES OF MEMBRANE ELECTRICAL PROPERTIES IN A RAT SKELETAL MUSCLE CELL LINE
BY YOSHIAKI KIDOKORO From the Salk Institute, P.O. Box 1809, San Diego, California 92112, U.S.A. (Received 19 April 1974) SUMMARY
1. The developmental changes of the cell membrane electrical properties were studied with micro-electrodes in a rat skeletal muscle cell line. 2. The resting potentials in myoblasts were -71 + 3 mV (mean + S.D.) and those in myotubes which were formed by fusion of myoblasts were -69 + 3 mV. There was no developmental change in the resting potential during the period examined. 3. The ionic mechanism for the resting potential was the same in myoblasts and myotubes. The resting membrane was almost exclusively permeable to K ions, while permeability to Na ions was not detectable. There was a small permeability to Cl ions. 4. The specific membrane resistance and capacitance were 8 kQ. cm2 and 1 #sF/cm2 for myoblasts and 12 kQ. cm2 and 5 ,uF/cm2 for myotubes,
respectively. 5. Action potentials in myoblasts were evoked by anode break stimulation. They were small and did not overshoot zero membrane potential. The action potentials in myotubes were larger, and had an average overshoot of 32 + 7 mV and a maximum rate of rise of 93 + 28 V/sec. 6. The current-voltage relation was examined. Delayed rectification was found in myotubes but not in myoblasts. INTRODUCTION
Recent advances in tissue culture techniques have made it possible to grow and differentiate muscle cells and neurones in vitro. Clonal cell lines from various origins are now available for study. A clonal muscle cell line is a suitable system for studying the developing excitable membrane. Cells can be identified at early stages in development, which is difficult for cells in vivo. Cells grown in a monolayer are clearly seen with a phase contrast microscope, which makes it easy to penetrate the cells with 5-2
YOSHIAKI KIDOKORO micro-electrodes. The cellular homogeneity of clonal lines is obviously advantageous for biochemical work such as changes in enzyme and synthesis of specific proteins during development (Schubert, Tarikas, Humphreys, Heinemann & Patrick, 1973), and this system should be useful for a future analysis of the molecular structure responsible for the action potential. A skeletal muscle cell line was used in this study of action potential development because it is easy to define the developmental stage of individual cells on the basis of morphology. Myogenic precursor cells (myoblasts) divide and increase their number exponentially (Yaffe, 1968). When they become confluent in the tissue culture dish, they fuse and form multinucleate myotubes. The production of myosin heavy chain was found to increase five- to sevenfold after myoblasts fused forming multinucleate myotubes. Cross-striation has been observed in some of the old myotubes (Schubert et al. 1973). Multinucleate myotubes are found to be electrically excitable (Harris, Heinemann, Schubert & Tarikas, 1971). The morphological, biochemical and electrophysiological differentiation of this clonal cell line shows that it has retained the essential developmental properties of normal myogenic cells. A preliminary report of the present work has been published (Kidokoro, 1973). 130
METHODS Cell culture. A muscle cell line derived from rat thigh muscle and adapted to culture conditions by Yaffe (1968) was used in these experiments (line L6). Myoblasts were grown in plastic tissue culture dishes in modified Eagle's medium containing 10% foetal calf serum at 370 C in an atmosphere of 12% C02 and 88 % air (Vogt & Dulbecco, 1963). The development of the cultured muscle cell was morphologically classified into two stages: the stage of mononucleate myoblasts in which the cells increase their number exponentially, and the stage of multinucleate myotubes which are formed by fusion of myoblasts. Some of the myotubes become partially striated and some of them contract spontaneously. Electrophysiology. Cells for electrophysiology experiments were grown attached to a round cover glass (diameter 25 mm) at the bottom of a tissue culture dish (diameter 60 mm). After the appropriate time of incubation, the cover glass was picked up and placed in a glass-bottomed chamber containing about 3 ml. of a saline. In some experiments cells grown in tissue culture dish were used in 8itu. The composition of the salines used are listed in Table 1. Myoblasts were examined at least 2-3 days after they attached to the bottom of the culture dish. The chamber was placed on the stage of a compound microscope (Reichart, Austria) which was modified to move the objective lens instead of the specimen stage for focusing (McBain Instruments, Inc., Calif., U.S.A.). The cells were observed with a 40 x water immersion phase contrast objective and 10 x ocular lens (Carl Zeiss, West Germany). Glass micro-electrodes were filled with 3 M-KC1 and had resistances of 80-150 MCI. Tippotentials of these electrodes were between -2 and - 20mV. In different solutions the tip potentials changed slightly (+ 7 to -4 mV). For penetration the microelectrode tip was pressed on the cell making a small dimple. A gentle tap on the base
DEVELOPMENTAL CHANGES OF MUSCLE MEMBRANE 131 of the manipulator (Leitz, West Germany) or on the table was enough to push the electrode tip into the cell. Usually two electrodes were placed inside a cell, one for passing current and the other for recording the resulting potential. Each of them was connected to the input stage of a preamplifier (M-4A, W. P. Instruments, U.S.A.) which included a bridge circuit for passing current and recording voltage simultaneously. Sometimes the bridge circuit to pass current and record resulting potentials was used with a single electrode to reduce cell damage. The bridge circuit could be balanced when a small amount of current (less than 0 5 nA) was used and small cells which had a large input impedance were chosen. The current passed through the electrode was measured by a conventional current-voltage converter circuit having a feed-back resistance of 1 or 10 MCI, inserted between an Ag-AgCl pellet in the bath and earth. When the external Cl concentration was changed, an agar-3 m-KC1 bridge was used between the bath and the Ag-AgCl pellet. The value of 19 V/sec for the maximum rate of rise for myotube action potentials given in a preliminary report (Kidokoro, 1973) was found to be an underestimate for several reasons: (1) cells which had been penetrated with two electrodes had a smaller maximum rate of rise than those which had been penetrated with a single electrode: this can be explained by damage to the membrane; (2) the stray capacity of high resistance micro-electrodes (100 MQ or more) had not been appropriately compensated. The rise time of the potential response due to a square pulse of current was 200-300 lusee after proper adjustment; (3) the resistance and capacitance which limit the frequency characteristics of the differentiator were not appropriate to record high frequency components. The differentiator in the preliminary report had highest gain at 50 Hz. For the present report the peak gain was at 50 kHz. Myoblasts and myotubes had variable shapes. If it is assumed that the membrane potential is equal everywhere inside the cell in the steady-state condition, the specific membrane resistance can be calculated from the effective input resistance and the cell surface area. Therefore, cells were chosen which were less than 300 #m in length and were relatively thick. The thickness of these cells was 8-15 jam in the region where the nuclei were. This was measured by the travel of the microscope fine adjustment between the position at which the upper surface of the cell was focused and that at which the glass surface was focused. A cylinder 300 ,gm in length, 8 #am in diameter, with an internal resistance of 180 L. cm (Albuquerque & Thesleff, 1968) and an input impedance of 50 MO, was chosen as a model to calculate the steady-state potential distribution inside a cell, using a short cable model (Weidmann, 1952). It was found that the steady-state potential at the end of the cell would be 98 % of the potential recorded at the middle of the cell where current was injected. The experimental cells selected for calculations were smaller in length and larger in width than the model, which favours an even better isopotential condition. However, in the case of myoblasts, the peripheral part of the cell was usually thin, which might cause an error in calculation. Since myoblasts (Schubert, Harris, Heinemann, Kidokoro, Patrick & Steinbach, 1974) and possibly myotubes are electrically coupled with each other, it was necessary to choose a cell well isolated from others. A single electrode was used both for passing current and recording membrane voltage. If the electrodes were carefully chosen and only a small amount of inward current (less than 0-5 nA) was used, the resistance of the electrodes behaved in a linear fashion. After a proper compensation of electrode capacity, the voltage transient had a time constant of less than 1 msec when an electrode was in the saline, whereas the cell membrane time constant was more than 6 msec in myoblasts and more than 20 msec in myotubes. Therefore it was not difficult to obtain a well-balanced bridge circuit. The small size of the cells grown in tissue culture is advantageous since they have high input impedances. The input
132
YOSHIAKI KIDOKORO
resistance increases after penetration for 2-5 min with a slight increase in the resting potential (5-10 mY). Therefore, at least a few minutes were allowed for the input resistance to reach the highest level. To calculate specific membrane constants, the cell was measured from an enlarged micrograph and the whole surface area was taken as twice the value measured. Since cultured cells are flat, this simplification should not produce a large error. All experiments were done at room temperature (20-24O C).
RESULTS
Resting membrane potentials. The tissue culture cells were fragile. In most cases, especially in myoblasts, the resting potential decreased quickly after penetration and cells appeared damaged under the microscope. There were occasions in which the resting potential stayed at its largest value for more than one minute and the tip potential of the recording electrode did not change more than 10 mV before and after the penetration. On such occasions, the average resting potential for myoblasts was -71 mV in normal saline, which was similar to that found for myotubes, -69 mV (mean and S.D. of the resting potential in various solutions are listed in Table 2). When the Ca concentration in the external saline was increased to 10 mm, the average resting potential was -74 mV in myoblasts and -72 mV in myotubes. It appeared to be easier to obtain a successful penetration in 10 mM-Ca saline. The ionic mechanism for the resting potential was studied by changing the external saline with solutions having various ionic compositions. Na ions. In order to determine the contribution of Na ions to the resting potential, the membrane potential was measured in Na-free saline where all NaCl was replaced by equiosmolar Tris-HCl or choline chloride. The average resting potential was -72 mV in myoblasts and -73 mV in myotubes. There was no significant difference in resting potentials between normal and Na-free saline. This was true in either 1-8 or 10 mm external Ca concentration (Table 2). K ions. The dependence of the resting potential on K ion concentration was examined in the absence of Na ions in order to avoid any complication due to a permeability to Na ions. Na-free saline and K-saline (Table 1) were mixed to obtain the desired concentration of K ions. The resting membrane potential was dependent upon the K ion concentration. When the resting potential was plotted against the logarithm of K concentration, it was linear both in myoblasts and myotubes (Fig. 1). The change was 50 mV/decade of K concentration in myoblasts and 52 mV/decade in myotubes. When the line was extrapolated to zero membrane potential, the internal K concentration could be estimated as 151 mm in myoblasts and as 150 mM in myotubules. The K equlibrium potential in normal
DEVELOPMENTAL CHANGES OF MUSCLE MEMBRANE 133 medium was calculated as -83 mV which is more negative than the measured resting potential. The discrepancy may be explained by a leak of the membrane due to micro-electrode penetration. 0
10
Myoblast
g
E~~~~~
a.
.0 -50
K concentration (mM) 0 100
10
100
Myotube
/
C~~~~~
-50
/
*/ II/ Fig. 1. Relation between external K concentration and resting membrane potentials in myoblasts and myotubes. K concentration is plotted on abscissa in logarithmic scale. Straight lines were drawn by eye. TABLE 1. Composition of salines used (mm) Ca Na K Tris 1.8 157-0 5-6 2-0 Normal saline 0 Na-free saline 5-6 198-0 1P8 144-0 100 2-0 5-6 High Ca saline 0 K saline 162-6 2-0 1P8 10 0 144-0 5-6 2-0 High Ca, Cl-free saline * MeSO3; methane sulphonate
Cl, 167-9 173-5 171-3 167-9 0
MeSO3* 0 0 0 0 171-3
C1 ions. The cells were immersed in Cl-free saline for more than 10 min to allow equilibration to occur. The resting potential in Cl-free media with 10 mM-Ca was -75 mM in myoblasts and -72 mV in myotubes. These values were not different from those in control solution either in myoblasts or in myotubes (Table 2). It is known for frog skeletal muscle that the Cl permeability is greater than that of K and when the external Cl concentration is decreased at constant K concentration, internal KCl (with water) leaks out to attain a new equilibrium [K]0 [Cl]o = [K]i[Cl]t. The resting potential, therefore, depolarizes suddenly at the time of solution change, and gradually returns to the resting potential (Hodgkin & Horowicz, 1960). It is possible to perfuse the external solution while the recording and current electrodes are maintained in a myotube. When the solution was
134 YOSHIAKI KIDOKORO changed from normal to Cl-free saline the resting potential did not change significantly but a slight increase in effective input resistance (10-20%) was noticed. Thus, in the myotube membrane Cl ions are less permeable than K ions. TABLE 2. Resting potentials of myoblasts and myotubes in various salines
Myoblasts
Myotubes
mean + S.D.
mean + S.D.
-69 ± 3 (18) mV -71 ± 3 (20)* mV Normal saline -72+2 (11) -74+ 2 (9) High Ca saline -73 + 3 (12) -72 + 4 (8) Na-free saline High Ca -75±4 (10) -72±2 (10) Cl-free saline High Ca -72±3 (12) -69±3 (8) Na-free saline * Figures in parentheses are the number of cells examined.
Specific membrane resistance and capacitance. The input resistance ranged from 53 to 140 MU in myoblasts and 80 to 120 MCI in myotubes (RI in Table 3). The specific membrane resistance (Rm) was calculated by multiplying the surface area with input resistance. The calculated (Rm) was 8'1 + 2*1 kQ. cm2 for myoblasts and 12-3 + 4-2 kf2. cm2 for myotubes. The potential transient caused by a square pulse of current fit a single exponential curve, showing that the isopotential condition is maintained within the limit of this detection procedure. The time constant (r) was 5-7-10 0 msec in myoblasts and 26-5-123-0 msec in myotubes (Table 3). The membrane capacitance was calculated by dividing the time constant (Xr) by the specific membrane resistance obtained above. The specific membrane capacitance for myoblasts was 1.0 + 0-2 ,uF/cm2 and 4*7 + 2-1 ,uF/cm2 for myotubes. Action potentials. In the following experiment, two intracellular electrodes were used to avoid artifacts due to improper bridge-balance which tended to occur when a large amount of current was passed (more than 1 nA). In myoblasts, the resting potential was small (-30 to -50 mV) when two electrodes were placed inside the cell. When the membrane was hyperpolarized to about - 100 mV, a small spike appeared after the end of the current pulse (Fig. 2A1). There was a clear inflexion at the foot of the spike (indicated by an arrow), consistent with the notion that this was a regenerative potential. The inflexion was at -30 to -40 mV. The average spike height measured from this inflexion to the top was 23 mV, so the spike did not overshoot beyond zero membrane potential. The maximum rate of rise was 3-13 V/sec. These small action potentials could be the result of damaging the cell membrane, but even when the resting potential
DEVELOPMENTAL CHANGES OF MUSCLE MEMBRANE 135 was about -50 mV, the action potential in myoblasts was much smaller than that in myotubes under comparable conditions. The action potential was larger and overshot zero membrane potential in multinucleate myotubes (Fig. 2B1). The resting potential was -50 to -65 mV with two electrodes in a cell. Again an action potential could be evoked only by anode-break excitation. The inflexion at the foot of the action potential was at -65 + 4 mV (indicated by an arrow). The amount of TABLE 3. Electrical membrane properties in myoblasts and myotubes
Myoblasts 1 2 3 4 5 Mean ± S.D.
Rest (mV) -57 - 46 - 39 -62 -37
RI (MQ) 104 53 84 140 110
r (msec)
Rm (ku)
6.5 7.8 10.0 5.7 7-1
11.4 5.8 8.4 7-0 7.7 8.1+2-1
Cm
(#ttF/cm2) 06
1*3 1.2
0*8 0*9
10±+0.2
Myotubes 1 2 3 4 5* Mean ± S.D.
-50 -58 -58 -67 -55
87 93 80 110 120
26-5 29-8 56 0 55.0 123.0
5.9 14.0 15*5 10.1 15*8 12.3±4.2
4-5
2*1 3.6 5-4 7.8
4.7+2±1
* Saline buffered with 5 mm HEPES-NaOH.
overshoot and maximum rate of rise of the action potential were dependent upon the amount of the preceding hyperpolarization (Fig. 2B1, 2). When the membrane potential at the steady state of hyperpolarization was plotted on the X axis and percent of the maximum rate of rise on the Y axis, a curve analogous to the inactivation curve in voltage clamp experiments in squid axon and other excitable membranes was obtained (Fig. 3). The potential at which half of the maximum rate of rise could be obtained was -69 to -86 mV. This shows that inability to evoke an action potential with depolarizing current was due to an inactivation process going on at the resting potential. The average overshoot of the maximal action potential was 32+7 mV (n = 15) and the average maximum rate of rise was 93 + 28 V/sec (n = 15) (these values were obtained with a single electrode for recording and injecting current to reduce damage). As shown in Fig. 2B1, the action potential had a long duration of 62+ 30 msec (n = 10) measured at half height. The action potential was composed of a fast initial spike, followed by a relatively slower hump (marked with a dot). The spike was followed by a hyperpolarizing potential
YOSHIAKI KIDOKORO 136 (positive after-potential) of 1030 mV in amplitude and about 5 sec in duration, which indicated that delayed rectification had developed at this stage. The duration was shorter (about 2 sec) when the myotube was old (about 1 month after fusion). Still it is considerably longer than that found in adult rat muscles (Albuquerque & Thesleff, 1968). A.
B.
Myoblast
Myotube
2
3
3 ~~~~t
I
_nA
-
I
_A
4~~~~~ 2OmVse
100 nisec
2m
Fig. 2. Responses to various applied currents in a myoblast (A) and a myotube (B). Upper trace in each pair of records shows current (downward deflexion indicates inward current), and zero membrane potential level. Lower trace is membrane potential record. Calibration shown in A4 applies for all A records. Calibration shown in B4 applies for all B records.
There was no obvious effect of increasing the Ca concentration (10 mm) myoblast action potential. Myotube action potentials in high Ca solution often have two peaks. The action potential was analysed and it was found that the initial peak is due to the inward movement of Na ions and the later peak to that of Ca ions (Kidokoro, 1975). The threshold for spike initiation in 10 mm Ca (measured by the level of inflexion at the foot on the
DEVELOPMENTAL CHANGES OF MUSCLE MEMBRANE 137 of an action potential indicated by an arrow in Fig. 2B1) was less negative than that in normal saline by 14 mV. The 'inactivation curve', obtained as described above, was shifted to the left along the voltage axes by about 10 mV (Fig. 3). The potential at which half of the maximum rate of rise could be obtained was -67 to -72 mV. The effect of high Ca on the K system (delayed rectification) was not studied. 10mm-Ca
18mM-Ca
100
0'1~
%-.
0
._
50 E E
%& o
0 0
-50 -100 Membrane potential (mV)
Fig. 3. Relation between maximum rate of rise of myotube action potentials and the level of preceding hyperpolarization in normal (open circles) and 10 mm-Ca saline (filled circles). Maximum rate of rise is standardized to the largest value and plotted on the Y axis. The membrane potential level is plotted on the X axis. Open arrow at the X axis indicates the resting potential in normal saline (-60 mV) and filled arrow that (-52 mV) in 10 mM-Ca saline.
Current-voltage relation. The steady-state current-voltage relation was examined in myoblasts and myotubes. Myoblasts were penetrated with two electrodes which usually lowered the resting potential (-30 to -56 mV). However, the input impedance of some of the cells was 10-49 MU which was not very low compared with the input impedance measured with single electrodes, considering some of the cells chosen for this experiment were not completely isolated and could have been electrically coupled to nearby cells. Therefore, the I- V curve obtained is probably valid and is not distorted too much by the existence of a leak due to
138 YOSHIAKI KIDOKORO micro-electrode penetration. Between - 100 and 0 mV, the I-V curve was straight (Fig. 4A). However, at a membrane potential above 0 mV, the resistance increased 2-7 times over that in the resting state. The myoblasts did not exhibit delayed rectification. This is not due to a small resting membrane potential (-30 to -56 mV) which can inactivate the delayed rectification (Nakajima, Iwasaki & Obata, 1962). Even in cells which have resting potentials between -60 and -70 mV (single electrode penetration), a large depolarizing current did not cause any hyperpolarization after the pulse as in myotubes. On the other hand, A
mV 50
,,
O
*
8
mV 50 0~~~~~~~~~~
,
5
-s
; -50 *
B
Depolarization
..
0°
.0
-
nA
00
so-50
-100
5
*
nA
*
-100
Hyperpolarization
Fig. 4. Current-voltage relation in a myoblast (A) and a myotube (B). Current is plotted on the X axis (outward current is plotted as plus), membrane potential on the Y axis. The potential change was measured 250 msec after the onset of current pulse in the myoblast and 250 msec (open circles), and 500 msec (filled circles) in the myotube. The intersection of the I-V curve with the Y axis shows the resting potentials (-48 mV in the myoblast and -50 mV in the myotube).
delayed rectification was clearly demonstrated in multinucleate myotubes by the I-V curve. When the depolarization exceeded -40 mV (-34 to -46 mV), an undershoot after the pulse occurred. When small pulses of inward current were applied during the after-spike hyperpolarization (positive after-potential), the potential change decreased to 7-22 % of that in the resting state. This is probably an underestimate of the conductance increase during the after-spike hyperpolarization (positive after-potential), since the isopotential condition may not be satisfied during the conductance increase. The potential level at the peak of the hyperpolarization was -70 to -77 mV. After an action potential or depolarization by a brief current pulse, the hyperpolarization lasted 4-8 sec. The rectification started
DEVELOPMENTAL CHANGES OF MUSCLE MEMBRANE 139 with a considerable delay which resulted in a large initial hump in response to a depolarizing current pulse (Fig. 2B4). Anomalous rectification was observed in the hyperpolarizing region of the steady-state I-V curve (Fig. 4B, filled circles). The slope conductance was higher in this region than in the resting potential region. Anomalous rectification was clear in the I-V curve in which membrane potentials were measured at 250 msec after the onset of current pulse, at which time delayed rectification had not yet started (Fig. 4B, open circles). The slope conductance for the inward current was larger than that for the outward current. A
B Depolarization
rV
mV 50 -
50-
/
1
nA
S -5
o
5
nA
-so
-S.0 Hyperpolarization
Fig. 5. Current-voltage relation in high K saline ([K]o = 1108 mM) inamyoblast (A) and a myotube (B). The voltage change was measured 250 msec after the onset of the current pulse in A and after 300 msec in B.
Currentvoltage relaim in high K saline. When myoblasts were soaked in high K saline (K, 110-8 mM) the resting potential was -2 to -8 mV and the steady-state I-V curve was straight between -40 and + 40 mV (Fig. 5A). On the other hand, in similarly treated myotubes the steadystate I-V curve showed a slope resistance to outward current 6-8 times larger than that to inward current (Fig. 5B). In all three cases examined, the slope resistance in the region above + 50 mV became smaller, as has been observed in frog skeletal muscle (Nakajima et al. 1962). DISCUSSION
It was found that in this muscle cell line, myoblasts and myotubes had similar resting potentials. In a previous study the resting potentials of myotubes and striated fibres were found to be the same in this tissue culture system (Kidokoro, 1973). There was no developmental change of resting potentials during the period examined. Boethius & Knutsson (1970) reported a gradual increase of resting potentials during differentiation of chick skeletal muscle from the 3rd day in ovo to the adult. The furthest
140 YOSHIAKI KIDOKORO developed stage examined here probably corresponds to their myocytes at 19th day in ovo where nuclei have migrated to the cell periphery. In their report, resting potentials earlier than the 15th day in ovo were distributed in a wide range and were small in average. Still they recorded around -60 mV resting potentials in several cells at these stages. The resting potentials of cells in which the micro-electrode penetration does not cause large damage do not vary greatly (for example - 82-4 + 2-4 mV in frog twitch muscle according to Stefani & Steinbach, 1969). The real resting potential in the early developmental stages might be more negative than -60 mV and the potentials recorded by Boethius & Knutsson (1970) might be small because of mechanical damage due to the relatively large electrodes (10-20 MU) they used. Thus, there may be no significant increase of the resting potential of chick muscle cells during development in ovo. The early developmental change of the resting potential in chick skeletal muscle has also been studied in culture (Fischbach, Nameroff & Nelson, 1971). They reported a small resting potential (-10 to -15 mV) in mononucleate muscle precursor cells and a relatively large resting potential (-60 to -80 mV) in old multinucleate myotubes. Similar results have been reported in rat skeletal muscle in vitro (Fambrough & Rash, 1971). Fischbach et al. claim that there is an increase in resting potential during development. The discrepancy in the resting potentials in their report and in the present one might be due to the differences in cell cultures. They used primary cultures whereas here a clonal cell line was used. There is a possibility that cells in a clonal cell line have different properties than those dissected from a foetus. The resting membrane of both myoblasts and myotubes is predominantly permeable to K ions. The permeability to Na ions seems to be small. At this stage in development the muscle membrane was relatively impermeable to Cl ions. There are no data concerning Cl ion permeability in adult rat muscle, but it is known that in goat muscle the Cl conductance is 85 % of the total. A similar proportion of Cl conductance has been found in human muscle (Bryant & Morales-Aguilera, 1971). It is well known that Cl ions are more than twice as permeable as K ions in frog sartorius muscle (Hodgkin & Horowicz, 1960; Hutter & Warner, 1967). Therefore it is reasonable to assume that Cl ions are more permeable than K ions in adult rat muscle. The resting membrane of a barnacle muscle fibre is mostly permeable to K ions in a solution of pH 7-7 (Hagiwara, Toyama & Hayashi, 1971). On the other hand, low pH is known to suppress the Cl permeability in frog sartorius muscle (Hutter & Warner, 1967). The normal saline used in this experiment has only 2 mm Tris-HCl which might be too weak to maintain the pH especially in the vicinity of the cells. An increase
DE VELOPMENTAL CHANGES OF MUSCLE MEMBRANE 141 in the buffer concentration to 5 mm (HEPES-NaOH) did not cause any increase in membrane conductance (myotube no. 5 in Table 3). The specific membrane resistance and capacitance of myoblasts were found to be 8 kQ. cm2 and 1 /tF/cm2. The capacitance increased to 5 ,cuF/cm2 in myotubes whereas the membrane resistance (12 kQ. cm2) was only slightly higher than that of myoblasts. The large capacitance is probably due to the development of T-system in myotubes as seen by electron microscopy (Ezerman & Ishikawa, 1967). The membrane resistance and capacitance for chick skeletal muscle in primary cultures are reported to be 2 6 kQ . cm2 and 3 9 /XF/cm2 respectively (Fischbach et al. 1971). Those for tissue cultured mouse muscles are 0-7 kQ.cm2 and 8-4,uF/Cm2 respectively (Powell & Fambrough, 1973). The membrane resistance given here for myotubes is large compared with these values, while the capacitances are more or less similar. The large membrane resistance can be explained by the low Cl conductance. The membrane resistance of goat external intercostal muscle in low Cl medium is reported to be 1O kQ. cm2 (Bryant & Morales-Aguilera, 1971), which is reasonably close to the value obtained here. Therefore, the low values of membrane resistance obtained in tissue cultured chick and mouse muscle cells may result from either a high Cl permeability or membrane damage by the micro-electrode penetration. A small regenerative potential was seen in myoblasts which are dividing exponentially. The height of the action potential increased abruptly after the time of fusion, since an overshooting action potential was observed even in binucleate cells. The largest overshoot (+ 43 mV) found in myotubes is not different from that in vivo (Albuquerque & Thesleff, 1968), suggesting that the Na concentration inside the cell is similar in both cases. The maximum rate of rise of action potential in myotubes (93 V/sec at 20-24o C) may be comparable to that of rat skeletal muscle after denervation, around 300 V/sec at 290 C (Redfern & Thesleff, 1971; Harris & Thesleff, 1971), when the difference in temperature is taken into account. No delayed rectification was found in myoblasts, although a mechanism for voltage sensitive inward current has been demonstrated. The later development of delayed rectification compared with voltage sensitive inward currents is also found in a tunicate muscle (Miyazaki, Takahashi & Tsuda, 1972; Takahashi, Miyazaki & Kidokoro, 1971). The delayed rectification found in myotubes is qualitatively similar to that of muscles in vivo, but is much slower (Nakajima et al. 1962). Inward-going rectification in myoblasts is different from that found in frog skeletal muscle (Katz, 1949) in that the I-V curve in normal saline was straight between -80 and 0 mV and in high K saline the curves were also straight between -40 and + 40 mV. On the other hand, in myotubes
142 YOSHIAKI KIDOKORO I-V curves showed inward-going rectification similar to 'anomalous rectification' found in frog muscles (Katz, 1949). The author thanks Drs S. Hagiwara, A. D. Grinnell, S. Heinemann, K. Takahashi and B. L. Brandt for their comments in preparing this manuscript. This work was supported by a Muscular Dystrophy Associations of America grant and a National Science Foundation grant to Dr Stephen Heinemann and a Sloan Foundation Grant to the Neurobiology Department, Salk Institute. REFERENCES
ALBUQUERQUE, E. X. & THESLEFF, S. (1968). A comparative study of membrane properties of innervated and chronically denervated fast and slow skeletal muscles of the rat. Acta phy8iol. 8cand. 73, 471-480. BOETHUS, J. & KNUTSSON, E. (1970). Resting membrane potential in chick muscle cells during ontogeny. J. exp. Zool. 174, 281-286. BRYANT, S. H. & MORALES-AGUILERA, A. (1971). Chloride conductance in normal and myotonic muscle fibres and the action of monocarboxylic aromatic acids. J. Physiol. 219, 367-383. EzEmRM., E. G. & Ism AwA, H. (1967). Differentiation of the sarcoplasmic reticulum and T system in developing chicks skeletal muscle in vitro. J. cell Biol. 35, 405-420. FAMBROUGH, D. & RASH, J. E. (1971). Development of acetylcholine sensitivity during myogenesis. Devl Biol. 26, 55-8. FISCHBACH, G. D., NAMEROFF, M. & NELSON, P. G. (1971). Electrical properties of chick skeletal muscle fibres developing in cell culture. J. cell Phy8iol. 78, 289-300. HAGIWARA, S., ToYAmA, K. & HAYASHI, H. (1971). Mechanism of anion and cation permeations in the resting membrane of a barnacle muscle fiber. J. gen. Phy8iol. 57, 408-434. HARRIS, A. J., HEiNEm.&NN- S., SCHUBERT, D. & TAREis, H. (1971). Trophic interaction between cloned tissue culture lines of nerve and muscle. Nature, Lond. 231, 296-301. HARRis, J. B. & THESLEFF, S. (1971). Studies on tetrodotoxin resistant action potentials in denervated skeletal muscle. Acta physiol. 8cand. 83, 382-388. HODGKIN, A. L. & HoRpowcz, P. (1960). The effect of sudden changes in ionic concentrations on the membrane potential of single muscle fibres. J. Phyaiol. 153, 370-385. HUTGER, 0. F. & WARNER, A. E. (1967). The pH sensitivity of the chloride conductance of frog skeletal muscle. J. Physiol. 189, 403-425. KATz, B. (1949). Les constantes 6lectriques de la membrane du muscle. Arch8 Sci. physiol. 3, 285-300. KIDOKORO, Y. (1973). Development of action potentials in a clonal rat skeletal muscle cell line. Nature, Lond. 241, 158-159. KIDOKORO, Y. (1975). Sodium and calcium components of the action potential in a developing skeletal muscle cell line. J. Physiol. 244, 173-187. MIYAZAKI, S. I., TAKAHASHI, K. & TSUDA, K. (1972). Calcium and sodium contributions to regenerative responses in the embryonic excitable cell membrane. Science, N.Y. 176, 1441-1443. NAKA.IMA, S., IWASAKI, S. & OBATA, K. (1962). Delayed rectification and anomalous rectification in frog's skeletal muscle membrane. J. gen. Phy&iol. 46, 97-115. PowELL, J. A. & FAMBROUGH, D. M. (1973). Electrical properties of normal and dysgenic mouse skeletal muscle in culture. J. cell Phy8iol. 82, 21-38.
DEVELOPMENTAL CHANGES OF MUSCLE MEMBRANE 143 REDFERN, P. & THESLEFF, S. (1971). Action potential generation in denervated rat skeletal muscle. I. Quantitative aspects. Acta phy8iol. 8cand. 81, 557-564. SCHUBERT, D., HARRIS, A. J., HEn EmANN, S., KIDOKORO, Y., PATRICK, J. & STEIINBACH, J. H. (1974). Differentiation and interaction of clonal cell lines of nerve and muscle. In Tisue Culture of the Nervou8 Sy8tem, ed. SATO, G. London: Plenum Press. SCHUBERT, D., TARTIAs, H., HUMPHREYS, S., HEINEMANN, S. & PATRICK, J. (1973). Protein synthesis and secretion in a myogenic cell line. Devl Biol. 33, 18-37. STEFANI, E. & STEINBACH, A. B. (1969). Resting potential and electrical properties of frog muscle fibres; effect of different external solutions. J. Physiol. 203, 383-401. TAKAHASHI, K., MIYAzAKI, S. & KIDOKORO, Y. (1971). Development of excitability in embryonic muscle cell membrane in certain tunicate. Science, N. Y. 171, 415-418. VOGT, M. & DuiLBECCO, R. (1963). Steps in the neoplastic transformation of hamster embryo cells by polyoma virus. Proc. natn. Acad. Sci. U.S. 49, 171-179. WEIDMANN, S. (1952). The electrical constraints of Purkinje fibres. J. Phy8iol. 118, 348-360. YAFFE, D. (1968). Retention of differentiation potentialities during prolonged cultivation of myogenic cells. Proc. natn. Acad. Sci. U.S.A. 61, 477-483.
