MUSCLE ACTIVITY AND THE LOSS OF ELECTRICAL COUPLING

0270~1X74/83/0307-1414$02.00/O

Printed in U.S.A.

The Journal of Neuroscience

Vol. 3, No. 7, pp. 1414-1421

July 1983

MUSCLE ACTIVITY AND THE LOSS OF ELECTRICAL COUPLING

BETWEEN STRIATED MUSCLE CELLS IN XENOPUS EMBRYOS’

DAVID L. ARMSTRONG,2P 3 LUCA TURIN,3

AND

ANNE E. WARNER

Department

Anatomy and Embryology, University College, London, WCIE 6BT, England

Received November 18, 1982; Revised January 11,1983; Accepted January 31,1983

Abstract

The gap junctions between embryonic striated muscle cells are lost during development. The time

course of their elimination has been examined with electrophysiological techniques in myotomes of

Xenopus laevis embryos. Gap junctions were detected by the passage of electrotonic current or the

fluorescent dye, Lucifer Yellow, from one muscle cell to another. These tracers only spread to

neighboring cells when injected intracellularly.

All the muscle cells are electrically coupled at stage 24 when neuromuscular transmission begins,

but normally many cells become uncoupled during the next 48 hr. In contrast, the muscle cells

remain electrically coupled if neuromuscular transmission is blocked during that period with tricaine

or a-bungarotoxin. When muscle activity recovers, the loss of coupling resumes. Once the coupling

has disappeared, neuromuscular blockade does not restore it. Muscle contraction is blocked during

development in a mutant of Xenopus, even though the muscle cells remain electrically excitable.

After stage 32 in these immobile embryos, the muscle cells are stimulated repeatedly by regular

bursts of neural activity. Although they never contract, the mutant muscle cells become uncoupled

at the same time as the muscle cells in normal embryos.

The results suggest that some consequence of repeated cholinergic activation, other than contrac-

tion, stimulates the loss of gap junctions between striated muscle cells during development. The

elimination of gap junctions may be required for neural control of subsequent muscle differentiation.

Developing striated muscle cells are electrically cou-

pled to one another by a direct, low-resistance pathway

(Dennis, 1975; Blackshaw and Warner, 1976; Dennis et

al., 1981). Gap junctions have been identified as the

morphological correlate of such electrical coupling, and

they connect striated muscle cells during development in

many vertebrate embryos (Kelley and Zacks, 1969;

Hayes, 1975; Keeter et al., 1975; Blackshaw and Warner,

1976; Schmalbruch, 1982). Gap junctions evidently allow

small metabolites as well as ions to pass directly from

the cytoplasm of one embryonic muscle cell to another

(Kalderon et al., 1977), but their role in myogenesis

remains to be discovered. Adult skeletal muscle fibers

are not electrically coupled (Dennis, 1975; Dennis et al.,

1981); the gap junctions disappear after neuromuscular

transmission matures (Keeter et al., 1975; Schmalbruch,

’ This work was supported by the British Medical Research Council

and the Muscular Dystrophy Association of America. We are also

grateful for the gifts of Lucifer Yellow from Dr. Walter Stewart, 01-

bungarotoxin from Dr. Stuart Bevan, and mature immobile heterozy-

gotes from Dr. Anne Droin.

’ To whom correspondence should be addressed at his new address.

’ Present address: Station Zoologique, 06230 Villefranche-sur-mer,

France.

1982). In this paper we consider the possibility that

muscle activity precipitates their elimination.

Harrison (1904) introduced the use of drugs that block

neuromuscular transmission to investigate the role of

activity in muscle development. Subsequently, several

features of nerve and muscle differentiation have been

shown to depend on functional synaptic transmission

(reviewed by Harris, 1981). For example, the extrasynap-

tic acetylcholine receptors on embryonic skeletal muscle

also disappear after muscle activity begins (Diamond and

Miledi, 1962; Dennis and Ort, 1977), but their elimination

can be delayed while neuromuscular transmission is

blocked pharmacologically (Gordon and Vrbova, 1975;

Burden, 1977). Similar observations are reported here for

the gap junctions between developing muscle cells. A

preliminary report of some of the results has been pub-

lished (Armstrong and Warner, 1980).

Materials and Methods

Xenopus laevis embryos were obtained from mature

pairs of toads by injections of chorionic gonadotropin

(Organon, Ltd.). Normal embryos were staged according

to the system of Nieuwkoop and Faber (1956) and reared

in water from the mains at room temperature. Just before

1414

The Journal

Neuroscience Gap Junctions and Muscle Development

1415

recording, embryos were anesthetized briefly in tricaine,

removed from the vitelline membrane, decerebrated, and

transferred to Ringer’s solution with the following com-

position, (in millimoles per liter): NaCl, 110; KCl, 2;

CaC12, 4; MnC12, 2; HEPES (4-(2-hydroxyethyl)-l-piper-

azine-ethanesulfonic acid), 5mM, pH 7.3. Manganese ions

were included to prevent neurally evoked muscle con-

tractions. The preparation was pinned out on Sylgard

(Dow Corning, 184) with tungsten wire pins and viewed

through a binocular dissecting microscope. The overlying

ectoderm was removed from the myotomes with glass

needles and fine forceps. The myotomes develop in a

rostrocaudal sequence, and myotome formation contin-

ues throughout the life of the tadpole. Consequently, the

timing of muscle differentiation and uncoupling may vary

among the myotomes. We have restricted our observa-

tions to the 12 most rostra1 post-otic myotomes since

these are formed before any experiments were begun.

Electrical measurements. Two muscle cells in adjacent

myotomes were each penetrated with a glass microelec-

trode. The micropipettes were filled with 0.8

potassium

citrate and had resistances between 50 and 70 megohms.

Current was injected into one cell, and the membrane

potential of the other was recorded differentially with

conventional electrophysiological techniques. The cur-

rent was measured as the voltage drop across a I-megohm

resistor between the bath and earth. Current pulses of 25

nA or less lasted about 1 set and were adjusted to

produce steady-state hyperpolarizing responses smaller

than 10 mV. No voltage responses were detected when

similar currents were passed extracellularly. In practice

the microelectrodes were several cells apart, and many

of the intervening cells must have been coupled for

current to spread between them.

One can make a rough estimate of the amount of

coupling by dividing the voltage change in one cell with

the current applied to the other cell. In circuit theory

this ratio (mV/nA) is referred to as a transfer resistance,

but the term mutual resistance has been used here to

emphasize that its value is determined by the electrical

properties of the entire cell membrane and not just the

gap junction (cf. Detwiler and Hodgkin, 1979). For ex-

ample, if a cell were damaged by the electrode, its mem-

brane resistance might be sufficiently smaller than the

gap junction’s resistance to make the cells appear uncou-

pled. This possibility was avoided by studying only mus-

cle cells with membrane potentials more negative than

-80 mV. Since extracellular voltage changes were never

detected with currents smaller than 25 nA, values of

mutual resistance greater than 0.04 mV/nA can only be

explained by a direct pathway for current flow between

the cells. Cell fusion might provide such a pathway, but

in Xenopus myotomes the muscle cells remain mononu-

cleate until stage 45 (Muntz, 1975).

Fluorescence measurements. Lucifer Yellow (Stewart,

1978) was iontophoresed into cells by repeating negative

current pulses (5 nA; 1 set) at 0.5 Hz for several minutes.

The microelectrodes were filled with a 3% solution (w/v)

of the dye in 0.1

LiCl. The spread of fluorescence was

observed through a high-pass filter (50 nm cut-off) on a

Zeiss microscope with a 50-W mercury vapor lamp and

a BG 12 excitation filter. At early stages, autofluorescence

from the numerous yolk granules precluded dye-coupling

studies. The amount of dye injected was limited by the

duration of the penetration and the modest currents

required to avoid electronic saturation with high-resist-

ance micropipettes. These factors may have been respon-

sible for some of the variability in the extent of dye

coupling. None of the muscle cells were filled with dye

when it was released extracellularly.

Pharmacological studies. Muscle activity was sus-

pended during development with ethyl m-aminobenzoate

methane sulfonate (tricaine, 1 to 2 m&f). Tricaine rapidly

crosses the embryos’ external membranes and interrupts

neuromuscular transmission. Tricaine probably acts pre-

synaptically. When tricaine was applied at these concen-

trations to adult frog sciatic nerves, the extracellularly

recorded compound action potential was abolished (n =

2). Intact embryos were exposed to tricaine for varying

periods during development. Most of the embryos re-

sumed swimming within 30 min after removing the tri-

Caine. Although tricaine noticeably slowed their devel-

opment, some tadpoles were raised up to metamorphosis

after their return to water. The experimental embryos

were allowed to recover momentarily before recording to

make sure they were still alive. Superficially normal

embryos were then re-anesthetized and prepared for

recording in the standard Ringer’s solution without tri-

Caine, exactly as described above for the controls.

Neuromuscular transmission was blocked postsynapt-

ically with a-bungarotoxin which binds irreversibly to

the cholinergic receptor on skeletal muscle (Chang and

Lee, 1963; Miledi and Potter, 1971). The muscle cells

were exposed to toxin at stage 24 by opening a flap in the

ectoderm above the myotomes while the embryos were

immersed in Ringer’s solution which contained 0.1

a-bungarotoxin. After 2 hr, the embryos were transferred

to diluted Ringer’s solution (1:9), also with 0.1

toxin,

where the ectoderm gradually healed. At the time of

transfer, intracellular recording revealed no sign of spon-

taneous neuromuscular transmission in myotomes im-

mediately underneath the flap, and embryos were dis-

carded if they showed any movement. Two hours later

still, the embryos were returned’to water. Over the next

2 days, most of the embryos recovered, presumably be-

cause toxin-bound receptors were degraded and replaced

with active ones (cf. Burden, 1977).

Results

Electrical coupling. In myotomes of early Xenopus

embryos, striated muscle cells are electrically coupled to

one another when neuromuscular transmission begins

(Blackshaw and Warner, 1976). Figure 1, left illustrates

such electrical coupling. In this case, roughly 25 nA

produced a lo-mV response in a neighboring cell. The

same current produced no response when either of the

electrodes was withdrawn to a position just outside the

cell; thus, the coupling cannot be explained by a restricted

extracellular space. This was confirmed for every record-

ing. Before stage 36, similar recordings can be obtained

from any two muscle cells in adjacent myotomes, al-

though the amount of coupling measured by the mutual

resistance declines steadily with age (Table I).

Figure 1, left was taken from an embryo at stage 35, or

1416

Armstrong et al. Vol. 3, No. 7, July 1983

about 48 hr after fertilization. By that stage the embryos

have emerged from their vitelline membrane and begun

to swim freely. The recording in Figure 1, right was made

on another embryo which was several hours older. Here

the same current produced a much smaller response in a

neighboring cell. The spontaneous quantal events which

appear can be eliminated by 10

tubocurarine. There

was an inverse correlation between the amplitudes of the

spontaneous events and the mutual resistance during

development.

Electrical coupling was observed much less frequently

in embryos that had been swimming for several hours.

Figure 2 plots the incidence of electrical coupling against

the stage of development. Although no coupling could be

detected in most embryos after stage 38, this does not

necessarily mean that all of the gap junctions have dis-

appeared. Since the electrodes were separated by several

cells during the electrical measurements, not all of the

intervening cells would have to be uncoupled for current

spread to be disrupted. Furthermore, the electrical meas-

urements of coupling require that current spread both

within the myotome and across myotomal borders.

Dye coupling. Gap junctions can also be identified by

the intercellular transfer of dyes which do not enter the

cells from the extracellular space (Payton et al., 1969).

Such dye coupling should not depend on the electrical

resistance of the nonjunctional membrane, and the dye

can be detected wherever it spreads, at least among the

surface cells. Eight individual muscle cells in six embryos

were filled with Lucifer Yellow after stage 36. In every

case but one the dye spread rapidly to at least one

neighboring cell, but never more than three cells away

from the injected cell (Fig. 3). When an electrode was

placed in adjacent cells to which the dye had diffused,

electrical coupling to the injected cell was always ob-

served (n = 5/5). The presence of dye coupling between

muscle cells makes it very unlikely that the electrical

coupling which we observe arises solely from the close

apposition of unspecialized membranes with low resist-

ance.

Lucifer Yellow seldom spread to cells within the same

myotome (n = 2/7), and neighboring cells without dye

Stage 35

25”A 1 -, r

were not usually electrically coupled to the injected cell

(n = l/5). This explains the difficulty of detecting wide-

spread electrical coupling when so many cells remain dye

coupled. We observed dye coupling between abutting

cells in adjacent myotomes as late as stage 45, when the

onset of cell fusion made further studies uninterpretable

(Muntz, 1975). Taken together, electrical coupling and

dye transfer studies suggest that many of the gap junc-

tions within each myotome disappear at the same time

t-m-l

0.1 n-l

19 26 32 35136

Nieuwkoap and Faber Stages

1 1 I 4

24 36 46

60 72

Hours after fertilization

Figure 2.

The

incidence of electrical coupling

decreases dur-

ing development. Mutual resistances greater than 0.05 mV/nA

were scored as coupled. Each

square

represents at least 10

measurements from five or more normal embryos at a particular

developmental stage (Z!I 1 stage). Each

triangle

represents at

least 5 measurements from two or more

immobile

embryos.

Hours

after fertilization are only approximate and vary with

temperature.

Stage 38

1 set

10mV ( Lr

10 mV

Figure 1.

Electrical coupling between striated muscle cells. Recordings from two

embryos at different developmental stages.

Upper traces:

current injected into one

cell; lower traces: simultaneous voltage response in another cell in an adjacent

myotome. No responses occurred when the current was released extracellularly.

Roughly the same current produces a much smaller response in the older embryo,

and this reduction in coupling is associated with the appearance of spontaneous

quantal events.

The Journal

Neuroscience

Gap Junctions and Muscle Development

Figure 3. Dye coupling between striated muscle cells in normal Xenopus myotomes at stage 40. Lucifer Yellow was injected

into the cella indicated by the arrowheads. Upper photomicrographs: phase contrast view of two preparations. The muscle cells

are roughly 100 pm in length and are organized into myotomes. Lowerphotomicrographs: diffusion of Lucifer Yellow viewed with

fluorescence in the same field. Magnification

-100.

the embryos begin using their muscle cells repeatedly to

produce sustained swimming movements. The possible

significance of this temporal relationship was revealed to

us by chance.

The role of activity. We routinely anesthetized the

embryos with tricaine before sacrificing them and pin-

ning the myotomes out in Ringer’s solution. Tricaine

rapidly and reversibly abolishes neural activity. We no-

ticed that, at about stage 36, muscle cells were substan-

tially better coupled in the embryos which had been left

in tricaine for several hours than in their counterparts

which were exposed to tricaine momentarily. Therefore,

we investigated the effects of rearing embryos in tricaine

during the period when electrical coupling normally dis-

appears.

Tricaine. Embryos were placed in tricaine between

stages 26 and 30, after the rostral myotomes had formed

and neuromuscular transmission had begun, and they

remained in tricaine past stage 36. All of the embryos

were returned to Ringer’s solution without tricaine for

measurements of electrical coupling. In contrast to nor-

mal embryos at that stage, electrical coupling was de-

tected in all of the tricaine-treated embryos. Treatment

with tricaine not only maintains electrical coupling be-

yond the stage at which it normally disappears but im-

proves the degree of coupling, too. The average mutual

resistance was 1.19 mV/nA between the tricaine-treated

myotomes (Table I). Tricaine might improve coupling by

increasing the resistance of the nonjunctional membrane,

TABLE I

Electrotonic coupling between myotomes

after neuromuscular

block

Block

Record

Coupling”

(mV/d)

Begin

End Mean* SD n

Normal embryos

Untreated - - 26-28 b 0.52 + 0.26 20

32-35 0.29 f 0.16 15

36-40 0.01 f 0.03 14

Tricaine’ 26-30 * 36-40 * 36-40 1.19 f 0.41 11

26-30 38 45 0 4

38 45 45 0 7

c+Bungarotoxin 26-28 ? 36-40 0.06 + 0.09 5

26-28 40 40

1.10 + 0.61 4

Immobile

mutants

Untreated - - 36-40 0.04 + 0.09 5

Tricaine c 26 38-41 38-41 1.34 f 0.93 7

a Average mutual resistance of adjacent myotomes in

embryos.

Values less than 0.05 mV/nA could not be detected.

* Nieuwkoop and Faber (1956) stages.

’ Measurements of electrical

coupling were made in Ringer’s solution

without tricaine.

but repeated measurements of mutual resistance during

the initial exposure to tricaine revealed no instantaneous

improvement in coupling. Inactivity itself also increases

muscle membrane resistance (Westgaard, 19751, but

widespread electrical coupling does not reappear when

the embryos are exposed to tricaine for the first time at

1418

Armstrong et al.

Vol. 3, No. 7, July 1983

stage 38. Thus tricaine only prevents the loss of coupling

if it is present when the motoneurons begin stimulating

the muscle cells to contract repeatedly.

Other properties of the muscle membrane were unaf-

fected by the prolonged exposure to tricaine. The muscle

cells retained high resting potentials (-85 to -9OmV)

throughout the exposure to tricaine. Within 30 min of

their removal from tricaine, the muscle cells were able to

spike and respond to acetylcholine. The effect of tricaine

on coupling was more slowly reversible, but the electrical

coupling eventually disappeared when the embryos were

returned to water (Table I); thus it is unlikely that

tricaine induces coupling by stimulating cell fusion.

These results raise the possibility that neuromuscular

activity is required for the elimination of gap junctions

between striated muscle cells during development. This

hypothesis is substantiated by experiments with a more

specific pharmacological probe.

a-Bungarotoxin. Embryos were exposed to a-bungar-

otoxin at stage 24 as described under “Materials and

Methods.” More than 90% of these embryos recovered

movement before stage 36, and very little electrical cou-

pling could be detected at that time (Table I); however,

in those embryos that remained paralyzed until stage 40,

the myotomes were clearly coupled electrically (Fig. 4;

Table I). Since a-bungarotoxin specifically blocks acetyl-

choline’s ability to activate the cholinergic receptor on

skeletal muscle (Chang and Lee, 1963; Miledi and Potter,

1971), this result implies that some consequence of cho-

linergic activation is required for gap junction elimina-

tion.

Immobile mutants. Droin and Beauchemin (1975) have

described a lethal recessive mutation in Xenopus which

blocks motor activity in early embryos. Apart from their

paralysis, the development of these immobile embryos is

superficially normal until stage 42, when their inability

to feed and the exhaustion of intracellular yolk granules

lead eventually to cell degeneration and death. Silver

staining reveals a normal nerve supply to the myotomes,

and ultrastructural studies confirm that the muscle cells

Stage

40.

4 toxin

0.1 set

Figure 4.

Electrical coupling remaining between striated

muscle cells in adjacent myotomes at stage 40, after exposure

to a-bungarotoxin during development (see “Materials and

Methods”). The Ringer’s solution did not contain manganese

ions during this recording.

contain regular myofibrils which, after extraction,

shorten on stimulation with adenosine 5’-triphosphate

(Droin and Beauchemin, 1975). Intracellular recordings

demonstrate that the innervation is functional: neuro-

muscular transmission is apparently normal, and the

mutant muscle cells are able to generate action potentials

(Warner, 1981).

As early as stage 32, one can record bursts of over-

shooting action potentials from the striated muscle cells

in immobile embryos. Nevertheless, the muscles do not

contract, and the embryos do not move. By stage 36 this

activity is evident in all of the mutant embryos (Fig. 5).

The bursts occur at roughly regular intervals and last for

several seconds with 20 to 50 spikes/set. The spikes are

abolished by tubocurarine (10

PM),

so we presume they

are generated by spontaneous neural activity. It is pos-

sible that similar bursts of neural activity produce the

sustained swimming which appears in normal embryos

at this stage of development. Before stage 36 one can also

record normal electrical coupling between the myotomes

in mutant embryos. This electrical coupling disappears

with the same time course as the coupling in normal

embryos, even though the mutant muscle cells never

contract (Fig. 2). Exposing the mutants to tricaine blocks

both the spikes and the loss of electrical coupling (Table

I).

Discussion

We have measured the electrical coupling between

striated muscle cells in myotomes of normal Xenopus

laevis embryos and in embryos whose activity was sus-

pended during development. Electrical coupling is wide-

spread until the embryos begin sustained and frequent

swimming at about stage 36. This increase in activity in

normal embryos is correlated with the onset of neurally

evoked bursts of overshooting spikes in the muscle cells

of immobile mutants. After stage 36 the incidence of

widespread electrical coupling declines precipitously in

both normal and mutant embryos. Electrical coupling

between cells within the same myotome evidently dis-

appears sometime before the coupling between cells in

adjacent myotomes. Some muscle cells remain coupled

as late as stage 45. Pharmacologically blocking neuro-

muscular transmission delays the loss of electrical cou-

pling within the myotomes.

Electrical coupling between muscle cells in Xenopus

myotomes is present as soon as the myotomes are formed

(Blackshaw and Warner, 1976). Gap junctions are gen-

erally recognized as the membrane structures which al-

low ions, dyes, and small metabolites to pass directly

from one cell’s cytoplasm to another (Payton et al., 1969;

Kalderon et al., 1977), and there is little doubt that they

mediate the electrical coupling which we observe. Gap

junctions have been identified in Xenopus myotomes at

these stages (Hayes, 1975; Blackshaw and Warner, 1976;

Kullberg et al., 1977), and control experiments rule out

alternative explanations for the electrical coupling. Elec-

trical coupling between embryonic striated muscle cells

seems to be a general feature of vertebrate skeletal

muscle development (Dennis, 1981). Since adult skeletal

muscle fibers are not connected by gap junctions, two

interesting questions present themselves: what role does

the coupling play in muscle development and why does

The Journal of Neuroscience

Gap Junctions and Muscle Development

Figure

5. Intracellular recordings from a striated muscle

cell

in an

immobile

mutant at stage 38. Manganese ions were only

present in the

lower right

trace. Note the spontaneous bursts of overshooting spikes in the

upper trace.

Much of the variability

in spike amplitude was an artifact of the pen recorder. Several sweeps on a faster time scale are shown at

lower left;

the spikes

decay with a different time course than do the synaptic potentials. No electrical coupling could be detected between adjacent

myotomes in this embryo.

it disappear? Our experiments are directed at the mech-

anism of gap junction elimination.

The relationship between the disappearance of electri-

cal coupling and the elimination of gap junctions is not

known. Functional uncoupling can occur without gap

junction elimination, as is found in

Aplysia

(Rayport and

Kandel, 1980), as a consequence of a reduction in the

resistance of the nonjunctional membrane. However, it

is more likely that the loss of electrical coupling which

we observe is related to an increase in gap junction

resistance. The amplitude of miniature endplate poten-

tials increases as uncoupling proceeds, and at late stages

there is a good correlation between the restricted pattern

of dye spread and the poor electrical coupling. Neverthe-

less, some muscle cells remain coupled to adjacent my-

otomes much later in development. The loss of electrical

coupling evidently proceeds faster within myotomes than

between them. There is no obvious explanation for this,

but the gap junctions also disappear in the same order,

both in

Xenopus

and axolotl myotomes (Hayes, 1975;

Keeter et al., 1975).

Muscle activity.

Xenopus

embryos, as in other ver-

tebrates (Dennis, 1975; Dennis et al., 1981), the matura-

tion of neuromuscular transmission precedes the loss of

electrical coupling during development. Since muscle

activity is required for the elimination of both the extra-

1420

Armstrong et al.

Vol.

3, No. 7,

July

1983

junctional acetylcholine receptors and the supernumer-

ary motor synapses on embryonic skeletal muscle (Den-

nis, 1981), it is easy to speculate that muscle activity also

stimulates the elimination of the gap junctions between

developing muscle cells. Our results support this hypoth-

esis; the loss of electrical coupling can be prevented while

neuromuscular transmission is blocked pharmacologi-

cally with tricaine.

At the same time, we have demonstrated that the loss

of widespread coupling within the myotomes proceeds at

the normal time in mutant embryos whose muscle cells

never contract. Other features of embryonic skeletal mus-

cle also develop normally in cultures of embryonic mouse

muscle cells which are unable to contract (Powell et al.,

1979). In both of these mutants the muscle cells remain

electrically excitable. In the immobile mutants of Xeno-

pus, the mdscle cells become electrically excitable early

in neuromuscular development (Warner, 1981); neverthe-

less, there is no decline in the incidence of electrical

coupling in the myotomes until the motoneurons begin

stimulating the muscle cells repeatedly with regular

bursts of activity. The temporal characteristics of this

neural activity are very similar to those that were most

effective for eliminating extrajunctional acetylcholine re-

ceptors by direct electrical stimulation of denervated

adult skeletal muscle (Lomo and Westgaard, 1975). Tri-

Caine blocks both the neural activity and the loss of

coupling in the immobile mutants. One concludes that

the elimination of coupling depends upon repeated mus-

cle stimulation. This may explain the apparent lack of

correlation between the development of electrical excit-

ability and the loss of electrical coupling in the embryonic

nervous system (Goodman and Spitzer, 1981; Spitzer,

1982).

The effect of activity on the muscle cells is evidently

irreversible; once it has disappeared, the widespread elec-

trical coupling cannot be restored by exposing the em-

bryos to tricaine. Morphological studies of Xenopus my-

otomes confirm that the number of gap junctions de-

creases during development (Hayes, 1975; Kullberg et al.,

1977). Since neuromuscular block with a-bungarotoxin

during that period in development also prevents the loss

of electrical coupling, this implies that gap junction elim-

ination depends on some consequence of acetylcholine

binding. Muscle excitation is the most obvious conse-

quence of cholinergic activation, but there are examples

of acetylcholine action which cannot be reproduced by

direct electrical stimulation of the postsynaptic cell (Par-

nas et al., 1974; Mathers and Thesleff, 1978; Chalazonitis

and Zigmond, 1980). In other systems the gap junctions’

conductance can be reduced by a variety of+experimental

perturbations. Acetylcholine reversibly uncouples acinar

cells in the pancreas (Iwatsuki and Petersen, 1978). In-

creases in the transjunctional voltage (Spray et al., 1981)

or the intracellular concentration of hydrogen or calcium

ions also uncouple cells in several vertebrate embryos

(Turin and Warner, 1980; Bennett et al., 1981). All of

these events may be associated with neuromuscular

transmission, but it remains to be determined whether

the repeated presentation of such perturbations could

result in the elimination of the gap junctions altogether.

In this regard it is interesting that the degree of electrical

coupling measured by the mutual resistance after pro-

longed neuromuscular blockade is substantially higher

than that observed at early developmental stages (Table

I). This implies some effect of neuromuscular activity on

gap junction permeability early on in neuromuscular

junction formation, and raises the possibility that impo-

sition of rapid stimulation early in development might

hasten the elimination of electrical coupling.

Gap junctions and muscle development. The func-

tional role of electrical coupling between developing

striated muscle cells is not known. In vitro studies of

muscle development have focused on the gap junctions

between myoblasts and their possible role in myotube

formation (Rash and Fambrough, 1973; Kalderon et al.,

1977), but recently, another cell organelle has been im-

plicated in the process of myoblast fusion (Kalderon and

Gilula, 1979). In vivo, there is less correlation between

gap junction formation and myoblast fusion. In Xenopus

myotomes, many of the gap junctions disappear before

cell fusion begins. In other species where fusion occurs

earlier in development, the myotubes remain connected

by gap junctions (Keeter et al., 1975; Schmalbruch, 1982).

Several authors have proposed that this electrical cou-

pling might allow coordinated reflex responses before

innervation is established.

It is also possible that early muscle activation is re-

quired for normal cellular differentiation. Motoneurons

control the expression of many muscle cell proteins by

regulating the pattern of muscle activity during devel-

opment (reviewed by Vrbova et al., 1978). Gap junction

elimination would be a prerequisite for such control;

otherwise all the muscle cells would experience the same

pattern of electrical activity. Now that the normal elim-

ination of gap junctions can be delayed experimentally,

these ideas can be pursued with further experiments.

References

Armstrong,

D. L., and A. E. Warner (1980) On the elimination

of gap junctions between developing muscle cells. J. Physiol.

(Lond.) 300: 64P.

Bennett, M. V. L., D. C. Spray, and A. L. Harris (1981) Electrical

coupling in development. Am. Zool. 21: 413-427.

Blackshaw, S. E., and A. E. Warner (1976) Low resistance

junctions between mesoderm cells during development of

trunk muscles, J. Physiol. (Lond.) 255: 209-230.

Burden, S. (1977) Development of the neuromuscular junction

in the chick embryo: The number, distribution and stability

of acetylcholine receptors. Dev. Biol. 55: 317-329.

Chalazonitis, A., and R. E. Zigmond (1980) Effects of synaptic

and antidromic stimulation on tyrosine hydroxylase activity

in the rat superior cervical ganglion. J. Physiol. (Lond.) 300:

525-538.

Chang, C. C., and C. Y. Lee (1963) Isolation of neurotoxins from

the venom of

Bungarus multicintus

and their modes of

neuromuscular blocking action. Arch. Int. Pharmacodyn.

Ther. 144: 241-257.

Dennis, M. J. (1975) Physiological properties of junctions be-

tween nerve and muscle developing during salamander limb

regeneration. J. Physiol. (Lond.) 244: 683-702.

Dennis, M. J. (1981) Development of the neuromuscular junc-

tion: Inductive interactions between cells. Annu. Rev. Neu-

rosci. 4: 43-68.

Dennis, M. J., and C. A. Ort (1977) The distribution of acetyl-

choline receptors on muscle fibers of regenerating salamander

limbs. J. Physiol. (Lond.) 266: 765-776.

Dennis, M. J., L. Ziskind-Conhaim, and A. J. Harris (1981)

The Journal of Neuroscience Gap Junctions and Muscle Development

1421

Development of neuromuscular junctions in rat embryos. Miledi, R., and L. T. Potter (1971) Acetylcholine receptors in

Dev. Biol. 81: 266-279. muscle fibers. Nature 233: 599-603.

Detwiler, P. B., and A. L. Hodgkin (1979) Electrical coupling Muntz, L. (1975) Myogenesis in the trunk and leg during

between cones in turtle retina. J. Physiol. (Lond.) 291: 75- development of the tadpole of Xenopus Zaeuis (Daudin 1802).

100. J. Embryol. Exp. Morphol. 33: 757-774.

Diamond, J., and R. Miledi (1962) A study of foetal and new- Nieuwkoop, P. D., and J. Faber (1956) Normal Table of Xen-

born rat muscle fibers. J. Physiol. (Lond.) 162: 393-408.

opus Zaeuis, Elsevier North-Holland, Amsterdam.

Droin, A., and M. L. Beauchemin (1975) Immobile (im), a Parnas, I., D. Armstrong, and F. Strumwasser (1974) Prolonged

recessive lethal mutation of Xenopus Zaeuis tadpoles. J. Em- excitatory and inhibitory synaptic modulation of a bursting

bryol. Exp. Morphol. 34: 435-449.

Goodman, C. S., and N. C. Spitzer (1981) The development of

pacemaker neuron. J. Neurophysiol. 37: 594-608.

Payton, B. W., M. V. L. Bennett, and G. D. Pappas (1969)

electrical properties of identified neurones in grasshopper

Permeability and structure of membranes at an electrotonic

embryos. J. Physiol. (Lond.) 313: 385-403. synapse. Science 166: 1641-1643.

Gordon, T., and G. Vrbova (1975) Changes in chemosensitivity Powell, J. A., B. A. Friedmann, and A. Cossi (1979) Tissue

of developing chick muscle fibers in relation to endplate culture study of murine muscular dysgenesis: Role of spon-

formation. Pflugers Arch. 360: 349-364. taneous action potential generation in the regulation of mus-

Harris, W. A. (1981) Neural activity and development. Annu.

Rev. Physiol. 43: 689-710.

cle maturation. Ann. N. Y. Acad. Sci. 317: 550-570.

Rash, J. E., and D. Fambrough (1973) Ultrastructural and

Harrison, R. G. (1904) An experimental study of the relation of electrophysiological correlates of cell coupling and cytoplas-

the nervous system to the developing musculature in the mic fusion during myogenesis in d-0. Dev. Biol. 30: 166-186.

embryo of the frog. Am. J. Anat. 3: 197-220.

Rayport, S. G., and E. R. Kandel (1980) Developmental mod-

Hayes, B. P. (1975) The distribution of intercellular junctions ulation of an identified electrical synapse: Functional uncou-

in the developing myotomes of the clawed toad. Anat. Em-

bryol. 147: 345-354.

pling. J. Neurophysiol. 44: 555-567.

Schmalbruch, H. (1982) Skeletal muscle fibers of newborn rats

Iwatsuki, N., and 0. H. Petersen (1978) Pancreatic acinar cells:

are coupled by gap junctions. Dev. Biol. 91: 485-490.

Acetylcholine evoked uncoupling and its ionic dependency.

Spitzer, N. C. (1982) Voltage- and stage-dependent uncoupling

J. Physiol. (Lond.) 274: 81-96. of Rohon-Beard neurones during embryonic development of

Kalderon, N., and N. B. Gilula (1979) Membrane events in- Xenopus tadpoles. J. Physiol. (Lond.) 330: 145-162.

volved in muscle fusion. J. Cell Biol. 81: 411-425. Spray, D. C., A. L. Harris, and M. V. L. Bennett (1981) Equilib-

Kalderon, N., M. L. Epstein, and N. B. Gilula (1977) Cell-to-

rium properties of a voltage-dependent junctional conduct-

cell communication and myogenesis. J. Cell Biol. 75: 788-806. ante. J. Gen. Physiol. 77: 77-93.

Keeter, J. S., G. D. Pappas, and P. G. Model (1975) Inter- and Stewart, W. W. (1978) Functional connections between cells as

intramyotomal gap junctions in the axolotl embryo. Dev.

Biol. 45: 21-33.

revealed by dye-coupling with a highly fluorescent napthal-

imide tracer. Cell 14: 741-759.

Kelley, A. M., and S. I. Zacks (1969) The histogenesis of rat Turin, L., and A. E. Warner (1980) Intracellular pH in early

intercostal muscle. J. Cell Biol. 42: 135-153.

Xenopus embryos: Its effect on current flow between blasto-

Kullberg, R. W., T. L. Lentz, and M. W. Cohen (1977) Devel- meres. J. Physiol. (Lond.) 300: 489-504.

opment of the myotomal neuromuscular junction in Xenopus Vrbova, G., T. Gordon, and R. Jones (1978) Nerue-Muscle

Zaeuis: An electrophysiological and fine structural study. Dev. Interaction, Chapman Hall, London.

Biol. 60: 101-129. Warner, A. E. (1981) Electrical properties of muscles from an

Lomo, T., and R. H. Westgaard (1975) Further studies on the

immobile mutant of Xenopus Zaevis. J. Physiol. (Lond.) 312:

control of ACh sensitivity by activity in the rat. J. Physiol.

3OP.

(Lond.) 252: 603-626.

Westgaard, R. H. (1975) Influence of activity on the passive

Mathers, D. A., and S. Thesleff (1978) Studies on neurotrophic

electrical properties of denervated soleus muscle fibres in the

regulation of murine skeletal muscle. J. Physiol. (Lond.) 282:

105-114.

rat. J. Physiol. (Lond.) 251: 683-697.