JOURNAL OF NEUROBIOLOGY VOL. 6, NO. 6, PP. 531-548
Neuromuscular Junctions in the Buccal Mass of Aplysia: Fine Structure and Electrophysiology of Excitatory Transmission P. M. DRKAND* and R. K. ORKANDt Marine Biology Laboratory, Kotor, Yugoslavia, and Departments of Anatomy and Biology, Uniuersity of California, Los Angeles, California
SUMMARY
The lower extrinsic protractor muscle in the buccal mass of Aplysia consists of bundles of muscle fibers 4-12 in diameter, containing thick and thin filaments that are not arranged in a transversely striated pattern. Individual fibers come close to one another and form specialized junctional regions. Electrophysiological evidence indicates that the muscle fibers form an electrical syncytium. Muscle bundles are innervated by more than one excitatory axon a t a number of points along their length. The presynaptic terminals contain spherical electron-lucent vesicles and a few larger electron-dense vesicles. There are no obvious structural postsynaptic specializations. Graded contraction can result from summation of excitatory junctional potentials in separate axons or from summation and facilitation of junctional potentials from a single axon. The buildup of facilitation during a train of stimuli results from the linear summation of facilitation remaining from preceding impulses. INTRODUCTION
The central nervous systems of gastropod molluscs, especially Aplysia, have been the subject of intensive anatomical and physiological study. By contrast, little is known about the neuromuscular systems in these animals (for reviews see Gerschenfeld, 1973; Heyer, Kater, and Karlsson, 1973). Despite the small size of the muscle fibers and tough connective tissue (Hoyle, 1964), Kater, Heyer, and Hegmann (1971) and Carew, Pinsker, Rubinson, and Kandel (1974) were able to obtain intracellular records of electrical responses of gastropod muscle fibers to the stimulation of motor axons. Kater et al. (1971) studied three muscles in the snail (posterior jugalis, buccal retractor, and columellar), and reported that they are comparable to the vertebrate “twitch” system; i.e., they are innervated by a single excitatory motor-
* Present address: Department of Anatomy, University of Pennsylvania, School of Medicine, Philadelphia, Pennsylvania 19174. + Present address: Department of Physiology and Pharmacology, University of Pennsylvania, School of Dental Medicine,. Philadelphia, Pennsylvania 19104. 531 01975 by John Wiley & Sons, Inc.
532
ORKAND AND ORKAND
neuron and produce overshooting action potentials. Carew et al. (1974) found summation and potentiation of excitatory junctional potentials (ejp’s) with repetitive stimulation of an appropriate motorneuron innervating the gill muscles in Aplysia. The present paper reports the results of our study of some of the main anatomical features and of excitatory neuromuscular transmission in the buccal mass of Aplysia. In a separate communication, preliminary evidence is presented for peripheral inhibitory transmission in the same muscle. Banks (1975) has found neurons in the buccal ganglion which give rise to hyperpolarizing junctional potentials in the muscle fibers, and decrease the amplitude and hasten the decay of the ejp’s. This neuromuscular system is of particular interest because much is already known about the neuronal organization and electrophysiological properties of neurons in the buccal ganglion in Aplysia (Gardner, 1971) and the closely related Nuuanax (Levitan, Tauc, and Segundo, 1970; Spira and Bennett, 1972) and Pleurobranchea (Davis and Mpitsos, 1971). METHODS Aplysia depilans 8 to 15 cm long were collected from the Adriatic Sea near Budva (Yugoslavia) and maintained in tanks of running seawater obtained from the Bay of Kotor. The animal was pinned out while immersed in seawater and opened by a ventral incision. The buccal mass and associated buccal ganglia were removed and placed in a petri dish half-filled with paraffin wax. Two longitudinal cuts in the midsagittal plane of the buccal mass permitted the removal of the radula and separated the two symmetrical preparations of buccal musculature each with its associated buccal ganglion. We used the muscle described by Huxley (1853, p. 58) as the lower extrinsic protractor. The spontaneously contracting esophogeal muscle was cut away and the
I P
Fig. 1. Diagram of the muscle preparation from one-half of the buccal mass of Aplysia. Muscle studied is the lower extrinsic protractor muscle. I t is usually about 1 cm2. Two nerve roots leave the buccal ganglion to innervate the muscle. b.g., Buccal ganglion; c.e.r., circumesophageal ring; e, esophagus; l.e.p.m., lower extrinsic protractor muscle; m, mouth; n.r., nerve roots.
APLYSIA NEUROMUSCULAR JUNCTIONS
533
preparation pinned out flat in the dish as shown in Figure 1. The second half of the buccal mass was stored in a beaker of seawater in the refrigerator to be used a few hours later or even the next morning without obvious deterioration. Some of the observations were repeated in Los Angeles with the buccal mass from Aplysia californica. We did not find any difference in the preparations from the two species. Hist o 1ogy After pinning out the buccal mass, and immobilizing the muscle and the two nerve roots entering it, the seawater was replaced by a fixative containing 1%glutaraldehyde, 4% paraformaldehyde in 0.1 M phosphate buffer and 3% NaCI, 0.2% CaC12. This fluid was replaced several times during the first hour of fixation and then allowed t o remain overnight. After a rinse in buffer, thin slices of the surface fan-shaped protractor musculature and pieces of the nerve root were postosmicated, dehydrated, and embedded in EPON 812. In several cases, the tissue was stained en bloc in aqueous 4% uranyl acetate before dehydration. Thick (about 1 p ) sections of the plastic embedded blocks were made with a Porter-Blum MT-2 ultramicrotome and stained with toluidine blue for light microscopic observation. Thin sections, stained with lead citrate and uranyl acetate, were examined in the JEOL 7T electron microscope. Electrophysiology Intracellular recording was accomplished with conventional 3 M KCl microelectrodes (resistance 10-30 MQ) connected to a high impedance solid state preamplifier with negative capacitance. The output of the amplifier was displayed on a Tektronix 565 oscilloscope and photographed with a Grass camera. A second, independently manipulated 3 M KCl electrode was occasionally inserted into a cell to pass current. The Ag/AgCl indifferent electrode in the bath was connected to virtual ground a t the summing junction of an operational amplifier used t o measure the amount of current passed through the second microelectrode. The buccal ganglion was cut off, and each of the two nerves leading to the muscle was mounted on a separate pair of platinum stimulating electrodes placed just below the surface of the natural seawater bathing solution. A frequent problem was the spontaneous movement of the preparation. One Ource of this mechanical activity was the small pieces of esophogus attached to the buccal mass; a second source was the muscle itself. Intracellular recording revealed two types of spontaneous electrical activity. The first was small spontaneous ejp’s (0-2 mV in amplitude), which on gross inspection appeared similar to spontaneous epp’s in frog slow muscle fibers (Burke, 1957). Th at is, they appeared to occur a t random intervals and to vary greatly in size and timecourse; the smallest ones disappearing into the noise level. The second kind of potential was larger, up to the size of neurally evoked ejp’s and occurred in rhythmic bursts. Since the buccal ganglion was removed, and we found no evidence for more peripheral nerve cells, we suspect they arose from spontaneous action potentials in motor nerve axons. The possibility that peripheral cell bodies exist has not been ruled out. Often these latter potentials disappeared within a couple of hours after dissection and it was possible to work with mechanically quiescent preparations. RESULTS
His to logy Light microscopy. Upon gross examination or under low magnification, the lower extrinsic protractor musculature of the buccal mass appears to consist of rather broad bands (50-600 p in width) which seem to be large muscle fibers. However, transverse sections have revealed that these bands are actually bundles of much smaller muscle fibers surrounded by perimysial connective tissue. Those in Figure 2a are around 8 p in diameter, but in larger animals the fiber diameters are greater. The muscle cells have centrally placed nuclei (arrow in Fig. 2a), usually accompanied by a central core of clear cyto-
534
ORKAND AND ORKAND
Fig. 2. (a) A light micrograph of transversely sectioned Aplysia muscle from the buccal mass. Small muscle fibers (here only 5-10 p in diameter) are grouped into fascicles about 50-500 thick. In some fibers, the section passes through a centrally placed nucleus (filled arrow). A nerve bundle containing 2 axons is also in this field (open arrow). Bar = 10 p. (b) A light micrograph of a transverse section through one of the nerve roots which originate in a buccal ganglion and end in the buccal mass. It has several large axons (up to about 20 p ) (e.g., a t arrow) and many hundreds of tiny axons not resolvable by light microscopy. Bar = 10 p.
APLYSIA NEUROMUSCULAR JUNCTIONS
535
plasm. Other fibers have a dense core. Large axons can often be identified in the perimysium (Fig. 2a). The two nerve roots which enter the muscle on each side arise from one of the buccal ganglia. Light microscopy shows that they contain several large
Fig. 3. An electron micrograph of transversely sectioned Aplysia muscle. The three fibers shown here have central cores containing mitochondria surrounded by contractile filaments. Notice the array of dense bodies within the fibers and applied to their cell membranes. The cells are surrounded by a rather dense concentration of collagen fibrils. Bar = 1k .
536
ORKAND AND ORKAND
Fig. 4. An electron micrograph of transversely sectioned Aplysia muscle. In several areas (e.g., circled) an ordered array of thin filaments surround a thick one. Tubules of sarcoplasinic reticulum can also be seen. Bar = 1fi.
axons (arrows in Fig. 2b). The rest of the root is occupied by a myriad of tiny axons (about 0.1 p ) not resolvable by light microscopy. Electron microscopy: Morphology of the muscle cells. Electron micrographs of transverse sections show that the muscle fibers are bound within their bundles by a rather dense collagenous connective tissue (Fig. 3). Thley are irregularly polygonal in shape, and invariably display a thick outer rind of contractile material surrounding a central core containing mitochondria, as well as nuclei. The contractile material consists of thick (-300 A) and thin (-70 A) filaments in an interdigitating array. In muscle fibers that are relatively stretched, thin filaments can often be seen in rather orderly orbits around thick ones (Fig. 4). The thick (Fig. 5) filaments display an axial periodicity of 150-200 A. They are not grouped into distinct myofibrils; rather there are irregular, branching spaces between clusters of filaments. Electron-dense, granular bodies, usually circular in cross section and about 0.1-0.2 p in diameter, are scattered among the clusters of filaments, and are also found heneath the cell membrane at irregular intervals (Fig. 3). In longitudinal sections one can demonstrate that they are attachment sites for thin myofilaments (Fig. 5). Longitudinal sections also reveal that the thick and thin filaments and the dense bodies are not arranged in any repeating units; i.e., the muscle is not transversely striated (Fig. 5). The sarcoplasmic reticulum of these muscle fibers consists of caveolae and flattened cisternae lying immediately beneath the plasma membrane, and of tubules which run, more or less, longitudinally throughout the rest of the fiber (Figs. 3, 5, and 6). We have found no structure equivalent to the transverse tubules of vertebrate skeletal muscle.
APLYSIA NEUROMUSCULAR JUNCTIONS
537
Fig. 5. An electron micrograph of longitudinally sectioned Aplysia muscle. Both thick and thin filaments are present going in and out of the plane of section and not arranged in transversely repeating units. In some places (e.g., a t the arrow) an axial periodicity is visible along the thick filaments. An electron-dense patch, designated by d, can be seen to be associated with thin filaments. An example of the tubular sarcoplasmic reticulum is seen a t s.r. Bar = 1 fi.
538
ORKAND AND ORKAND
Fig. 6 . An electron micrograph of 2 adjacent Aplysia muscle cells. The apposing surfaces of the 2 adjacent cells are specialized by 2 sets of complementary cytoplasmic densities (fascia adhaerentes). Subsarcolemmal cisternae of the sarcoplasmic reticulum are also shown in this micrograph (arrows). Bar = l w.
Intermuscular junctions. Light microscopy of transverse sections shows that most of the muscle fibers appear distinctly separate from the surrounding ones. However, there are small areas where adjacent fibers come into close contact. By electron microscopy, we see that these junctional areas are most obviously specialized by complementary accumulations of electrondense granular material, similar to the dense bodies found within the cells applied to the plasma membranes of the apposing cells (Fig. 6). These are simi-
APLYSIA NEUROMUSCULAR JUNCTIONS
539
Fig. 7. (a) This neuromuscular junction has, in addition to “clear” and “filled” synaptic vesicles, a mitochondrion. It has no postjunctional specialization. Bar = 1 F . (b) A neuromuscular junction in Aplysia. The axon terminal is filled with electron-lucent vesicles and a few dense ones. The “top” of the terminal lies naked in the endomysial connective tissue space, uncovered by glial cell. There is no postsynaptic specialization except for the subsarcolemmal cistern beneath the synapse (arrow). Bar = 1p.
lar to the types of attachment plaques often referred to as fascia adhaerentes (Farquhar and Palade, 1963). Although more complicated interdigitations between muscle cells have been seen, it has thus far been difficult to follow the membranes, and no closer junctions have yet been seen. Neuromuscular junctions. Neuromuscular synapses are easily found in this tissue. It is likely that each muscle fiber receives several such contacts. Most frequently, the axon terminal lies in a shallow groove on the fiber surface. The synaptic gap is about 200 A and is not intruded by the basal lami-
540
ORKAND AND ORKAND
na of the muscle surface. Subsarcolemmal cisternae of endoplasmic reticulum are sometimes present under the synapse (Fig. 7b) but do not appear to be more prevalent than elsewhere. Seventeen presynaptic axon terminals were examined and appeared morphologically similar. They are filled with vesicles, most of which are circular and electron-lucent. Several larger vesicles with an electron-dense content are also present (Figs. 7a, b). Figures 7a and b show neuromuscular junctions whose axon terminals are unaccompanied by glial cells. Although this is a common observation, occasionally a terminal is covered by processes presumably belonging to glial cells. Electrophysiology
Synctial properties of muscle fibers. While advancing a recording microelectrode into and through a bundle of muscle fibers, one records a series of transient negative potentials of up to 80 mV. If the electrode was withdrawn slightly while in the middle of such a bundle, a resting potential was recorded of -60 to -80 mV which remained stable for many minutes. The recording of these large stable potentials suggests that the electrode produces relatively little damage to the muscle fibers. When a separate current passing electrode is inserted into the same muscle bundle, 200-500 p away from the recording electrode either in the longitudinal or transverse direction, one
Fig. 8. Electrical coupling between muscle fibers. (A) Voltage recording and current passing electrodes inserted in the same muscle bundle but separated by about 500 p longitudinally and 200 p transversely. (B) Current electrode withdrawn from muscle bundle. Appearance of electrotonic potentials in (A) indicates electrical continuity between separate fibers in the same muscle bundle.
APLYSIA NEUROMUSCULAR JUNCTIONS
541
records an electrotonic potential. Figure 8 shows records obtained in one such experiment. When the recording electrode was advanced slightly and the resting potential lost, the electrotonic potentials disappeared. Subsequent experiments (Taraskevich and Orkand, unpublished observations) using bridge circuits permitting simultaneous recording of the membrane potential from both current passing and voltage recording electrodes indicated that such potentials were only observed when both electrodes were intracellular, i.e., recording a resting membrane potential. If now the recording electrode is withdrawn from the bundle and inserted into an adjacent bundle a few hundred microns distant or further, no such potentials are recorded. Thus, the numerous small fibers which make up a muscle bundle, usually 100 to 500 p in diameter, constitute an electrical syncytium. Occasionally, these large muscle bundles branch. Under these conditions, both branches appeared electrically connected to the point of branching. Innervation pattern. Some general features of the pattern of innervation were determined by varying the strength and frequency of stimuli applied to the nerve roots and observing the contraction of the muscle through the dissecting microscope. Maximal stimulation of the nerve root leading to the left side of the muscle produced contraction of the left side with trains of stimuli a t low frequency (5Isec for 1 sec). As the frequency of stimuli in the train was increased, the left margin of the muscle contracted more strongly and the contraction spread across the muscle. With stimulation a t 20/sec for 1 sec,
Fig. 9. Pattern of innervation. Junctional potentials produced by maximal stimulation of the left (1) and right (r) roots with the recording microelectrode inserted in various regions of protractor muscle. Recording positions as indicated: left, middle of fiber on left edge of muscle; right, middle of fiber on right edge of muscle; middle, three recordings from same muscle bundle in middle of muscle; top, 2 mm from top of muscle; middle, middle of muscle; bottom, 2 mm from bottom of muscle. Muscle bundle was 10 mm long. All resting potentials -70 to -80 mV.
542
ORKAND AND ORKAND
the whole muscle contracted, stronger on the left side than on the right. At a given frequency of stimulation, a decrease in the strength of stimulation produced a weaker contraction. An analogous result was obtained with stimulation of the right root, but now contraction was stronger on the right side of the muscle. These results suggest that each root innervated the entire muscle, densely on the side which it entered and more sparsely as it progressed across the muscle. The decrease in contraction with a decrease in stimulus strength suggests there was more than one excitatory axon in each root. The validity of these conclusions was confirmed with intracellular recording. In Figure 9 it can be seen that a single maximal nerve volley produced excitatory junctional potentials (ejp’s) of decreasing amplitude as one moved the intracellular electrode away from the side of nerve entry. Near the middle of the muscle, fibers were found where stimulation of either root produced. about the same depolarization. As the electrode was moved longitudinally
Fig. 10. Effect of decreasing strength of stimulation to nerve root on recordings from three muscle fibers, a, b, and c. In each case, decreasing the strength of stimulation decreased the amplitude of the ejp, indicating innervation of the muscle bundle by two excitatory axons. Stimulation frequency 1/2 sec.
along the muscle, as shown in the center of Figure 9, the absolute and relative ejp amplitudes varied to a much less extent than with horizontal displacements. Excitatory junctional potentials were recorded with intracellular recording anywhere in the muscle. Evidence for innervation of muscle bundles by more than one excitatory axon from a single root is shown in Figure 10. In most cases tested, each root sent two excitatory axons to each muscle bundle. However, in a few cases (about lo%), the muscle bundle appeared to be innervated either by one or three axons from a single root. We could not rule out the possibility that in these cases either an axon had been damaged or an additional axon was stimulated due to spread of the stimulating current. These results clearly indicate that the muscle bundles are innervated by more than one excitatory axon (polyneuronal) and at many points along their length (multiterminal). No evidence of inhibitory innervation was found with root stimulation. In order to reveal the inhibitory innervation, it is necessary to stimulate the inhibitory cell body directly (Banks 1975). The threshold for the inhibitory axon is higher than that of the excitatory axons,
APLYSIA NEUROMUSCULAR JUNCTIONS
543
2
-
2
m
1
0
M
m
-0
.-
n
0.5
*
0)
0.2 0
0.2
0.4
0.6
Time (sec) Fig. 11. Decay of facilitation following a single maximal stimulus to a nerve root. Inset shows recordings for two stimuli a t three different intervals. When potentials summed, amplitude of second ejp was measured from extrapolated decay of first ejp. Each point is the average of 4 to 8 trials. Bars a t *SE.
and under conditions of root stimulation, the small inhibitory junctional potentials are masked by the much larger ejp’s. Summation and facilitation. When the interval between single stimuli to the two roots was short, the ejp’s summed. When a single root was stimulated twice the potentials not only summed, but also the second’ejp was facilitated, i.e., increased in amplitude. For our analysis, we will consider the amount of facilitation, ( f ) ,to be the fractional increase in depolarization of the test ejp, (un), compared with the initial ejp, (uo), and is given by the equation f = u, - uo/uo (Mallart and Martin, 1967). In the experiment illustrated in Figure 11,it can be seen that the facilitation decreases as the interval between stimuli is increased. The time for decay of facilitation to one-half in this experiment is about 100 msec. The timecourse of the decay of facilitation is not a simple exponential and possibly consists of a fast and slow phase, as is the case a t the frog neuromuscular junction (Mallart and Martin, 1967). With a train of nerve volleys, the facilitation increases progressively (Fig. 12). If one assumes that the facilitation resulting from each volley has the same magnitude and timecourse and that the process sums in a linear manner, it
ORKAND AND ORKAND
544
28
1.5
1
I
m/
0.5
0 Time (sec) Fig. 12. Growth of facilitation with train of stimuli. Same cell as Fig. 11. Top: records of ejp at 5 and 10/sec. Graph shows growth of facilitation for the two trains. w lO/sec. A 5/sec. Points are averages for three trains; bars *SE. Solid lines are predicted facilitation assuming each stimulus produces identical facilitation which sums linearly with that remaining from earlier stimuli. Data for decay of facilitation following each stimulus from Fig. 11.
should be possible to predict the growth of facilitation solely from knowledge of the decay curve. This possibility has been tested, as shown in Figure 12, and the agreement between the observed growth of facilitation and that predicted from the decay of facilitation is reasonably good. The timecourse of facilitation varied greatly from one preparation to the next. In some of the preparations there was facilitation with stimulation at 2 sec intervals, whereas in others, measurable facilitation was not observed until the interval was reduced to 0.5 sec. The largest facilitation was observed in an experiment conducted at 18OC; the smallest facilitation was at
APLYSIA NEUROMUSCULAR JUNCTIONS
545
24OC, suggesting that the decay of facilitation is highly temperature dependent. Increasing the frequency of stimulation always led to an increased ejp amplitude and eventual contraction of the muscle with subsequent loss of the intracellular recording. In a few instances it was possible to maintain the intracellular recording during contraction, and in these cases it was clear that contraction resulted from the depolarization produced by summation and facilitation of ejp’s and not regenerative muscle spikes. Such results were obtained both from muscles tested within one hour after dissection and those tested a t longer periods at either room temperature or cooled to 10°C (Banks, unpublished observations). However, since the depolarization of the membrane necessary to produce contraction through summation and facilitation of ejp’s did not exceed 25 mV, we have not ruled out the possibility that greater depolarizations might lead to regenerative muscle spikes. DISCUSSION
These results, along with those of Banks (1975), establish a number of the basic features of a neuromuscular system in the buccal mass of Aplysia. The structure of the muscle fibers differs from the structure of the best studied molluscan muscle, the anterior byssus retractor muscle of Mytilis (Hanson and Lowy, 1959, 1960; Philpott, Kahlbrock, and Szent-Gyorgyi, 1969), most prominently in that it lacks the large diameter (up to 1000 A), thick filaments which contain paramyosin. The dimensions of its elements more closely resemble the obliquely striated muscles of Ascaris (Rosenbluth, 1965), Glycera (Rosenbluth, 1968), and earthworms (Rosenbluth, 1972). However, those elements are not nearly so well ordered in Aplysia buccal mass. If this muscle is obliquely striated, the angle of its striation pattern is so small that it is not easily discerned. The Aplysia muscle does appear to be similar in structure to another molluscan muscle, the adductor of the glochidian larva of Anodanta s p . (Zs.Nagy and Labos, 1969; Kidokoro, Hagiwara, and Henkart, 1974). This neuromuscular system has a number of properties in common with neuromuscular systems in crustacea (Furshpan, 1959). The functional muscle unit, in this case the 50-600 p diameter bundle of muscle fibers, is innervated at many points along its length, and by both excitatory and inhibitory axons. Although we have not ruled out the possibility that some kind of regenerative action potential might occur with large depolarizations, contractions can readily be produced as a result of summation and facilitation of ejp’s. We have also established that the individual 4-12 p diameter fibers are formed into an electrical synctium. The low resistance pathways probably correspond to the intermuscular junctional complexes apparent in the electron-micrographs, although the presence of “gap” junctions has not been established. It is possible that the entire muscle constitutes an electrical syncytium through branching of the muscle bundles, as in the case of the frog heart ventricle (Weidmann, 1966), the somatic muscle cells in Ascaris (De
-
546
ORKAND AND ORKAND
Bell, del Castillo, and Sanchez, 1963), and the mammalian smooth mus& (Tomita, 1966). By selectively stimulating one root, one side of the muscle contracts more strongly than the other. However, we do not know if during normal feeding behavior this ability is utilized (Huxley, 1853). It is possible that axons in separate roots originate from the same cell body, as has been found in Pleurobranchea (Siegler, Mpitsos, and Davis, personal communication). Hoyle and Smyth (1963) have found that giant barnacle muscle fibers are innervated by a number of axons, some of which travel in separate nerve roots. Due to facilitation, it is possible to achieve greater depolarization of the muscle membrane with multiple impulses over the same excitatory pathway than with the same number of impulses in more than one motorneuron. The importance of this facilitation awaits an understanding of the functional relations of the motorneurons within the buccal ganglion. In this connection it is interesting that neurons within the buccal ganglion are capable of responding to excitatory stimuli at 50 impulses per second (Gardner, 1971); this is well in excess of the five to ten impulses per second necessary to cause contraction of the muscle through summation and facilitation of ejp’s. These results are in striking contrast to those obtained by Kater et al. (1971) for three muscles in the gastropod mollusc Helisoma, where only excitatory neuromuscular transmission similar to that of the frog and mammalian “twitch” systems, was found. The excitatory innervation of the extrinsic protractor is possibly cholinergic. Relatively low concentrations of acetylcholine ( lop5 M ) cause sustained contractures (unpublished observations), as is the case for the radula protractor of Busycon (Hill, 1970). Much higher concentrations M ) of glutamate, dopamine, epinephrine, histamine, or 5-hydroxytryptamine do not produce visible contractures. Until the site of acetylcholine action to produce contracture is delineated (Florey and Kreibel, 1969) and further pharmacological studies are made, it is best to defer conclusions regarding the possible cholinergic nature of transmission. Acetylcholine, as well as dopamine and glutamate produce contractures of a small nearby muscle in the buccal mass (Taraskevich, Gibbs, Schmued, and Orkand, unpublished observations). The 17 presynaptic terminals surveyed in the electron microscope are similar in appearance to cholinergic junctions in vertebrates (e.g., Birks, Huxley, and Katz, 1960), since they contain primarily electron-lucent vesicles with a few electron-dense vesicles. They differ from vertebrate neuromuscular junctions in not having an elaborate subneural apparatus or basement lamella in the synaptic cleft. In these respects, they more closely resemble neuromuscular junctions in crustacea (Atwood and Morin, 1970) and other molluscs (Heyer et al., 1973). Since all of the terminals we observed had a similar morphology, we do not know if they were all excitatory, inhibitory, or a mixture of both types. This study was supported by U S . Public Health Service Grants NS-08346, NS-09521, and PL480. The authors thank Djordje Konjevic and Linda Macaulay for necessary technical assis-
APLYSIA NEUROMUSCULAR JUNCTIONS
547
tance and Professors J. D. French and L. Rakic, whose mutual interest and effort in fostering international cooperation in research made the study possible. REFERENCES
ATWOOD, H. L. and MORIN,W. A. (1970). Neuromuscular and axo-axonal synapses in the crayfish opener muscle. J . Ultrastruct. Res. 32: 351-369. RANKS,F. (1975). Inhibitory transmission a t a molluscan neuromuscular junction, J. Neurobiol. 6: 429-433. BIRKS,R., HUXLEY,H. E. and B. KATZ (1960). The fine structure of the neuromuscular junction of the frog. J . Physiol. 150: 134-144. BIIRKF,, W. (1957). Spontaneous potentials in slow muscle fibres of the frog. J . Physiol. 135: 511-521. CAREW,T. J., PINSKER, H., RUBINSON,K. and E. R. KANDEL(1974). Physiological and biochemical properties of neuromuscular transmission between identified motoneurons and gill muscle in Aplysia. J . Neurophysiol. 27: 1020-1040. DAVIS,W. J . and MPITSOS, G. J. (1971). Behavioral choice and habituation in the marine mollusk Pleurobranchaca californica. 2. Vergl. Physiol. 75: 207-232. DE BELL,
Neuromuscular Junctions in the Buccal Mass of Aplysia: Fine Structure and Electrophysiology of Excitatory Transmission P. M. DRKAND* and R. K. ORKANDt Marine Biology Laboratory, Kotor, Yugoslavia, and Departments of Anatomy and Biology, Uniuersity of California, Los Angeles, California
SUMMARY
The lower extrinsic protractor muscle in the buccal mass of Aplysia consists of bundles of muscle fibers 4-12 in diameter, containing thick and thin filaments that are not arranged in a transversely striated pattern. Individual fibers come close to one another and form specialized junctional regions. Electrophysiological evidence indicates that the muscle fibers form an electrical syncytium. Muscle bundles are innervated by more than one excitatory axon a t a number of points along their length. The presynaptic terminals contain spherical electron-lucent vesicles and a few larger electron-dense vesicles. There are no obvious structural postsynaptic specializations. Graded contraction can result from summation of excitatory junctional potentials in separate axons or from summation and facilitation of junctional potentials from a single axon. The buildup of facilitation during a train of stimuli results from the linear summation of facilitation remaining from preceding impulses. INTRODUCTION
The central nervous systems of gastropod molluscs, especially Aplysia, have been the subject of intensive anatomical and physiological study. By contrast, little is known about the neuromuscular systems in these animals (for reviews see Gerschenfeld, 1973; Heyer, Kater, and Karlsson, 1973). Despite the small size of the muscle fibers and tough connective tissue (Hoyle, 1964), Kater, Heyer, and Hegmann (1971) and Carew, Pinsker, Rubinson, and Kandel (1974) were able to obtain intracellular records of electrical responses of gastropod muscle fibers to the stimulation of motor axons. Kater et al. (1971) studied three muscles in the snail (posterior jugalis, buccal retractor, and columellar), and reported that they are comparable to the vertebrate “twitch” system; i.e., they are innervated by a single excitatory motor-
* Present address: Department of Anatomy, University of Pennsylvania, School of Medicine, Philadelphia, Pennsylvania 19174. + Present address: Department of Physiology and Pharmacology, University of Pennsylvania, School of Dental Medicine,. Philadelphia, Pennsylvania 19104. 531 01975 by John Wiley & Sons, Inc.
532
ORKAND AND ORKAND
neuron and produce overshooting action potentials. Carew et al. (1974) found summation and potentiation of excitatory junctional potentials (ejp’s) with repetitive stimulation of an appropriate motorneuron innervating the gill muscles in Aplysia. The present paper reports the results of our study of some of the main anatomical features and of excitatory neuromuscular transmission in the buccal mass of Aplysia. In a separate communication, preliminary evidence is presented for peripheral inhibitory transmission in the same muscle. Banks (1975) has found neurons in the buccal ganglion which give rise to hyperpolarizing junctional potentials in the muscle fibers, and decrease the amplitude and hasten the decay of the ejp’s. This neuromuscular system is of particular interest because much is already known about the neuronal organization and electrophysiological properties of neurons in the buccal ganglion in Aplysia (Gardner, 1971) and the closely related Nuuanax (Levitan, Tauc, and Segundo, 1970; Spira and Bennett, 1972) and Pleurobranchea (Davis and Mpitsos, 1971). METHODS Aplysia depilans 8 to 15 cm long were collected from the Adriatic Sea near Budva (Yugoslavia) and maintained in tanks of running seawater obtained from the Bay of Kotor. The animal was pinned out while immersed in seawater and opened by a ventral incision. The buccal mass and associated buccal ganglia were removed and placed in a petri dish half-filled with paraffin wax. Two longitudinal cuts in the midsagittal plane of the buccal mass permitted the removal of the radula and separated the two symmetrical preparations of buccal musculature each with its associated buccal ganglion. We used the muscle described by Huxley (1853, p. 58) as the lower extrinsic protractor. The spontaneously contracting esophogeal muscle was cut away and the
I P
Fig. 1. Diagram of the muscle preparation from one-half of the buccal mass of Aplysia. Muscle studied is the lower extrinsic protractor muscle. I t is usually about 1 cm2. Two nerve roots leave the buccal ganglion to innervate the muscle. b.g., Buccal ganglion; c.e.r., circumesophageal ring; e, esophagus; l.e.p.m., lower extrinsic protractor muscle; m, mouth; n.r., nerve roots.
APLYSIA NEUROMUSCULAR JUNCTIONS
533
preparation pinned out flat in the dish as shown in Figure 1. The second half of the buccal mass was stored in a beaker of seawater in the refrigerator to be used a few hours later or even the next morning without obvious deterioration. Some of the observations were repeated in Los Angeles with the buccal mass from Aplysia californica. We did not find any difference in the preparations from the two species. Hist o 1ogy After pinning out the buccal mass, and immobilizing the muscle and the two nerve roots entering it, the seawater was replaced by a fixative containing 1%glutaraldehyde, 4% paraformaldehyde in 0.1 M phosphate buffer and 3% NaCI, 0.2% CaC12. This fluid was replaced several times during the first hour of fixation and then allowed t o remain overnight. After a rinse in buffer, thin slices of the surface fan-shaped protractor musculature and pieces of the nerve root were postosmicated, dehydrated, and embedded in EPON 812. In several cases, the tissue was stained en bloc in aqueous 4% uranyl acetate before dehydration. Thick (about 1 p ) sections of the plastic embedded blocks were made with a Porter-Blum MT-2 ultramicrotome and stained with toluidine blue for light microscopic observation. Thin sections, stained with lead citrate and uranyl acetate, were examined in the JEOL 7T electron microscope. Electrophysiology Intracellular recording was accomplished with conventional 3 M KCl microelectrodes (resistance 10-30 MQ) connected to a high impedance solid state preamplifier with negative capacitance. The output of the amplifier was displayed on a Tektronix 565 oscilloscope and photographed with a Grass camera. A second, independently manipulated 3 M KCl electrode was occasionally inserted into a cell to pass current. The Ag/AgCl indifferent electrode in the bath was connected to virtual ground a t the summing junction of an operational amplifier used t o measure the amount of current passed through the second microelectrode. The buccal ganglion was cut off, and each of the two nerves leading to the muscle was mounted on a separate pair of platinum stimulating electrodes placed just below the surface of the natural seawater bathing solution. A frequent problem was the spontaneous movement of the preparation. One Ource of this mechanical activity was the small pieces of esophogus attached to the buccal mass; a second source was the muscle itself. Intracellular recording revealed two types of spontaneous electrical activity. The first was small spontaneous ejp’s (0-2 mV in amplitude), which on gross inspection appeared similar to spontaneous epp’s in frog slow muscle fibers (Burke, 1957). Th at is, they appeared to occur a t random intervals and to vary greatly in size and timecourse; the smallest ones disappearing into the noise level. The second kind of potential was larger, up to the size of neurally evoked ejp’s and occurred in rhythmic bursts. Since the buccal ganglion was removed, and we found no evidence for more peripheral nerve cells, we suspect they arose from spontaneous action potentials in motor nerve axons. The possibility that peripheral cell bodies exist has not been ruled out. Often these latter potentials disappeared within a couple of hours after dissection and it was possible to work with mechanically quiescent preparations. RESULTS
His to logy Light microscopy. Upon gross examination or under low magnification, the lower extrinsic protractor musculature of the buccal mass appears to consist of rather broad bands (50-600 p in width) which seem to be large muscle fibers. However, transverse sections have revealed that these bands are actually bundles of much smaller muscle fibers surrounded by perimysial connective tissue. Those in Figure 2a are around 8 p in diameter, but in larger animals the fiber diameters are greater. The muscle cells have centrally placed nuclei (arrow in Fig. 2a), usually accompanied by a central core of clear cyto-
534
ORKAND AND ORKAND
Fig. 2. (a) A light micrograph of transversely sectioned Aplysia muscle from the buccal mass. Small muscle fibers (here only 5-10 p in diameter) are grouped into fascicles about 50-500 thick. In some fibers, the section passes through a centrally placed nucleus (filled arrow). A nerve bundle containing 2 axons is also in this field (open arrow). Bar = 10 p. (b) A light micrograph of a transverse section through one of the nerve roots which originate in a buccal ganglion and end in the buccal mass. It has several large axons (up to about 20 p ) (e.g., a t arrow) and many hundreds of tiny axons not resolvable by light microscopy. Bar = 10 p.
APLYSIA NEUROMUSCULAR JUNCTIONS
535
plasm. Other fibers have a dense core. Large axons can often be identified in the perimysium (Fig. 2a). The two nerve roots which enter the muscle on each side arise from one of the buccal ganglia. Light microscopy shows that they contain several large
Fig. 3. An electron micrograph of transversely sectioned Aplysia muscle. The three fibers shown here have central cores containing mitochondria surrounded by contractile filaments. Notice the array of dense bodies within the fibers and applied to their cell membranes. The cells are surrounded by a rather dense concentration of collagen fibrils. Bar = 1k .
536
ORKAND AND ORKAND
Fig. 4. An electron micrograph of transversely sectioned Aplysia muscle. In several areas (e.g., circled) an ordered array of thin filaments surround a thick one. Tubules of sarcoplasinic reticulum can also be seen. Bar = 1fi.
axons (arrows in Fig. 2b). The rest of the root is occupied by a myriad of tiny axons (about 0.1 p ) not resolvable by light microscopy. Electron microscopy: Morphology of the muscle cells. Electron micrographs of transverse sections show that the muscle fibers are bound within their bundles by a rather dense collagenous connective tissue (Fig. 3). Thley are irregularly polygonal in shape, and invariably display a thick outer rind of contractile material surrounding a central core containing mitochondria, as well as nuclei. The contractile material consists of thick (-300 A) and thin (-70 A) filaments in an interdigitating array. In muscle fibers that are relatively stretched, thin filaments can often be seen in rather orderly orbits around thick ones (Fig. 4). The thick (Fig. 5) filaments display an axial periodicity of 150-200 A. They are not grouped into distinct myofibrils; rather there are irregular, branching spaces between clusters of filaments. Electron-dense, granular bodies, usually circular in cross section and about 0.1-0.2 p in diameter, are scattered among the clusters of filaments, and are also found heneath the cell membrane at irregular intervals (Fig. 3). In longitudinal sections one can demonstrate that they are attachment sites for thin myofilaments (Fig. 5). Longitudinal sections also reveal that the thick and thin filaments and the dense bodies are not arranged in any repeating units; i.e., the muscle is not transversely striated (Fig. 5). The sarcoplasmic reticulum of these muscle fibers consists of caveolae and flattened cisternae lying immediately beneath the plasma membrane, and of tubules which run, more or less, longitudinally throughout the rest of the fiber (Figs. 3, 5, and 6). We have found no structure equivalent to the transverse tubules of vertebrate skeletal muscle.
APLYSIA NEUROMUSCULAR JUNCTIONS
537
Fig. 5. An electron micrograph of longitudinally sectioned Aplysia muscle. Both thick and thin filaments are present going in and out of the plane of section and not arranged in transversely repeating units. In some places (e.g., a t the arrow) an axial periodicity is visible along the thick filaments. An electron-dense patch, designated by d, can be seen to be associated with thin filaments. An example of the tubular sarcoplasmic reticulum is seen a t s.r. Bar = 1 fi.
538
ORKAND AND ORKAND
Fig. 6 . An electron micrograph of 2 adjacent Aplysia muscle cells. The apposing surfaces of the 2 adjacent cells are specialized by 2 sets of complementary cytoplasmic densities (fascia adhaerentes). Subsarcolemmal cisternae of the sarcoplasmic reticulum are also shown in this micrograph (arrows). Bar = l w.
Intermuscular junctions. Light microscopy of transverse sections shows that most of the muscle fibers appear distinctly separate from the surrounding ones. However, there are small areas where adjacent fibers come into close contact. By electron microscopy, we see that these junctional areas are most obviously specialized by complementary accumulations of electrondense granular material, similar to the dense bodies found within the cells applied to the plasma membranes of the apposing cells (Fig. 6). These are simi-
APLYSIA NEUROMUSCULAR JUNCTIONS
539
Fig. 7. (a) This neuromuscular junction has, in addition to “clear” and “filled” synaptic vesicles, a mitochondrion. It has no postjunctional specialization. Bar = 1 F . (b) A neuromuscular junction in Aplysia. The axon terminal is filled with electron-lucent vesicles and a few dense ones. The “top” of the terminal lies naked in the endomysial connective tissue space, uncovered by glial cell. There is no postsynaptic specialization except for the subsarcolemmal cistern beneath the synapse (arrow). Bar = 1p.
lar to the types of attachment plaques often referred to as fascia adhaerentes (Farquhar and Palade, 1963). Although more complicated interdigitations between muscle cells have been seen, it has thus far been difficult to follow the membranes, and no closer junctions have yet been seen. Neuromuscular junctions. Neuromuscular synapses are easily found in this tissue. It is likely that each muscle fiber receives several such contacts. Most frequently, the axon terminal lies in a shallow groove on the fiber surface. The synaptic gap is about 200 A and is not intruded by the basal lami-
540
ORKAND AND ORKAND
na of the muscle surface. Subsarcolemmal cisternae of endoplasmic reticulum are sometimes present under the synapse (Fig. 7b) but do not appear to be more prevalent than elsewhere. Seventeen presynaptic axon terminals were examined and appeared morphologically similar. They are filled with vesicles, most of which are circular and electron-lucent. Several larger vesicles with an electron-dense content are also present (Figs. 7a, b). Figures 7a and b show neuromuscular junctions whose axon terminals are unaccompanied by glial cells. Although this is a common observation, occasionally a terminal is covered by processes presumably belonging to glial cells. Electrophysiology
Synctial properties of muscle fibers. While advancing a recording microelectrode into and through a bundle of muscle fibers, one records a series of transient negative potentials of up to 80 mV. If the electrode was withdrawn slightly while in the middle of such a bundle, a resting potential was recorded of -60 to -80 mV which remained stable for many minutes. The recording of these large stable potentials suggests that the electrode produces relatively little damage to the muscle fibers. When a separate current passing electrode is inserted into the same muscle bundle, 200-500 p away from the recording electrode either in the longitudinal or transverse direction, one
Fig. 8. Electrical coupling between muscle fibers. (A) Voltage recording and current passing electrodes inserted in the same muscle bundle but separated by about 500 p longitudinally and 200 p transversely. (B) Current electrode withdrawn from muscle bundle. Appearance of electrotonic potentials in (A) indicates electrical continuity between separate fibers in the same muscle bundle.
APLYSIA NEUROMUSCULAR JUNCTIONS
541
records an electrotonic potential. Figure 8 shows records obtained in one such experiment. When the recording electrode was advanced slightly and the resting potential lost, the electrotonic potentials disappeared. Subsequent experiments (Taraskevich and Orkand, unpublished observations) using bridge circuits permitting simultaneous recording of the membrane potential from both current passing and voltage recording electrodes indicated that such potentials were only observed when both electrodes were intracellular, i.e., recording a resting membrane potential. If now the recording electrode is withdrawn from the bundle and inserted into an adjacent bundle a few hundred microns distant or further, no such potentials are recorded. Thus, the numerous small fibers which make up a muscle bundle, usually 100 to 500 p in diameter, constitute an electrical syncytium. Occasionally, these large muscle bundles branch. Under these conditions, both branches appeared electrically connected to the point of branching. Innervation pattern. Some general features of the pattern of innervation were determined by varying the strength and frequency of stimuli applied to the nerve roots and observing the contraction of the muscle through the dissecting microscope. Maximal stimulation of the nerve root leading to the left side of the muscle produced contraction of the left side with trains of stimuli a t low frequency (5Isec for 1 sec). As the frequency of stimuli in the train was increased, the left margin of the muscle contracted more strongly and the contraction spread across the muscle. With stimulation a t 20/sec for 1 sec,
Fig. 9. Pattern of innervation. Junctional potentials produced by maximal stimulation of the left (1) and right (r) roots with the recording microelectrode inserted in various regions of protractor muscle. Recording positions as indicated: left, middle of fiber on left edge of muscle; right, middle of fiber on right edge of muscle; middle, three recordings from same muscle bundle in middle of muscle; top, 2 mm from top of muscle; middle, middle of muscle; bottom, 2 mm from bottom of muscle. Muscle bundle was 10 mm long. All resting potentials -70 to -80 mV.
542
ORKAND AND ORKAND
the whole muscle contracted, stronger on the left side than on the right. At a given frequency of stimulation, a decrease in the strength of stimulation produced a weaker contraction. An analogous result was obtained with stimulation of the right root, but now contraction was stronger on the right side of the muscle. These results suggest that each root innervated the entire muscle, densely on the side which it entered and more sparsely as it progressed across the muscle. The decrease in contraction with a decrease in stimulus strength suggests there was more than one excitatory axon in each root. The validity of these conclusions was confirmed with intracellular recording. In Figure 9 it can be seen that a single maximal nerve volley produced excitatory junctional potentials (ejp’s) of decreasing amplitude as one moved the intracellular electrode away from the side of nerve entry. Near the middle of the muscle, fibers were found where stimulation of either root produced. about the same depolarization. As the electrode was moved longitudinally
Fig. 10. Effect of decreasing strength of stimulation to nerve root on recordings from three muscle fibers, a, b, and c. In each case, decreasing the strength of stimulation decreased the amplitude of the ejp, indicating innervation of the muscle bundle by two excitatory axons. Stimulation frequency 1/2 sec.
along the muscle, as shown in the center of Figure 9, the absolute and relative ejp amplitudes varied to a much less extent than with horizontal displacements. Excitatory junctional potentials were recorded with intracellular recording anywhere in the muscle. Evidence for innervation of muscle bundles by more than one excitatory axon from a single root is shown in Figure 10. In most cases tested, each root sent two excitatory axons to each muscle bundle. However, in a few cases (about lo%), the muscle bundle appeared to be innervated either by one or three axons from a single root. We could not rule out the possibility that in these cases either an axon had been damaged or an additional axon was stimulated due to spread of the stimulating current. These results clearly indicate that the muscle bundles are innervated by more than one excitatory axon (polyneuronal) and at many points along their length (multiterminal). No evidence of inhibitory innervation was found with root stimulation. In order to reveal the inhibitory innervation, it is necessary to stimulate the inhibitory cell body directly (Banks 1975). The threshold for the inhibitory axon is higher than that of the excitatory axons,
APLYSIA NEUROMUSCULAR JUNCTIONS
543
2
-
2
m
1
0
M
m
-0
.-
n
0.5
*
0)
0.2 0
0.2
0.4
0.6
Time (sec) Fig. 11. Decay of facilitation following a single maximal stimulus to a nerve root. Inset shows recordings for two stimuli a t three different intervals. When potentials summed, amplitude of second ejp was measured from extrapolated decay of first ejp. Each point is the average of 4 to 8 trials. Bars a t *SE.
and under conditions of root stimulation, the small inhibitory junctional potentials are masked by the much larger ejp’s. Summation and facilitation. When the interval between single stimuli to the two roots was short, the ejp’s summed. When a single root was stimulated twice the potentials not only summed, but also the second’ejp was facilitated, i.e., increased in amplitude. For our analysis, we will consider the amount of facilitation, ( f ) ,to be the fractional increase in depolarization of the test ejp, (un), compared with the initial ejp, (uo), and is given by the equation f = u, - uo/uo (Mallart and Martin, 1967). In the experiment illustrated in Figure 11,it can be seen that the facilitation decreases as the interval between stimuli is increased. The time for decay of facilitation to one-half in this experiment is about 100 msec. The timecourse of the decay of facilitation is not a simple exponential and possibly consists of a fast and slow phase, as is the case a t the frog neuromuscular junction (Mallart and Martin, 1967). With a train of nerve volleys, the facilitation increases progressively (Fig. 12). If one assumes that the facilitation resulting from each volley has the same magnitude and timecourse and that the process sums in a linear manner, it
ORKAND AND ORKAND
544
28
1.5
1
I
m/
0.5
0 Time (sec) Fig. 12. Growth of facilitation with train of stimuli. Same cell as Fig. 11. Top: records of ejp at 5 and 10/sec. Graph shows growth of facilitation for the two trains. w lO/sec. A 5/sec. Points are averages for three trains; bars *SE. Solid lines are predicted facilitation assuming each stimulus produces identical facilitation which sums linearly with that remaining from earlier stimuli. Data for decay of facilitation following each stimulus from Fig. 11.
should be possible to predict the growth of facilitation solely from knowledge of the decay curve. This possibility has been tested, as shown in Figure 12, and the agreement between the observed growth of facilitation and that predicted from the decay of facilitation is reasonably good. The timecourse of facilitation varied greatly from one preparation to the next. In some of the preparations there was facilitation with stimulation at 2 sec intervals, whereas in others, measurable facilitation was not observed until the interval was reduced to 0.5 sec. The largest facilitation was observed in an experiment conducted at 18OC; the smallest facilitation was at
APLYSIA NEUROMUSCULAR JUNCTIONS
545
24OC, suggesting that the decay of facilitation is highly temperature dependent. Increasing the frequency of stimulation always led to an increased ejp amplitude and eventual contraction of the muscle with subsequent loss of the intracellular recording. In a few instances it was possible to maintain the intracellular recording during contraction, and in these cases it was clear that contraction resulted from the depolarization produced by summation and facilitation of ejp’s and not regenerative muscle spikes. Such results were obtained both from muscles tested within one hour after dissection and those tested a t longer periods at either room temperature or cooled to 10°C (Banks, unpublished observations). However, since the depolarization of the membrane necessary to produce contraction through summation and facilitation of ejp’s did not exceed 25 mV, we have not ruled out the possibility that greater depolarizations might lead to regenerative muscle spikes. DISCUSSION
These results, along with those of Banks (1975), establish a number of the basic features of a neuromuscular system in the buccal mass of Aplysia. The structure of the muscle fibers differs from the structure of the best studied molluscan muscle, the anterior byssus retractor muscle of Mytilis (Hanson and Lowy, 1959, 1960; Philpott, Kahlbrock, and Szent-Gyorgyi, 1969), most prominently in that it lacks the large diameter (up to 1000 A), thick filaments which contain paramyosin. The dimensions of its elements more closely resemble the obliquely striated muscles of Ascaris (Rosenbluth, 1965), Glycera (Rosenbluth, 1968), and earthworms (Rosenbluth, 1972). However, those elements are not nearly so well ordered in Aplysia buccal mass. If this muscle is obliquely striated, the angle of its striation pattern is so small that it is not easily discerned. The Aplysia muscle does appear to be similar in structure to another molluscan muscle, the adductor of the glochidian larva of Anodanta s p . (Zs.Nagy and Labos, 1969; Kidokoro, Hagiwara, and Henkart, 1974). This neuromuscular system has a number of properties in common with neuromuscular systems in crustacea (Furshpan, 1959). The functional muscle unit, in this case the 50-600 p diameter bundle of muscle fibers, is innervated at many points along its length, and by both excitatory and inhibitory axons. Although we have not ruled out the possibility that some kind of regenerative action potential might occur with large depolarizations, contractions can readily be produced as a result of summation and facilitation of ejp’s. We have also established that the individual 4-12 p diameter fibers are formed into an electrical synctium. The low resistance pathways probably correspond to the intermuscular junctional complexes apparent in the electron-micrographs, although the presence of “gap” junctions has not been established. It is possible that the entire muscle constitutes an electrical syncytium through branching of the muscle bundles, as in the case of the frog heart ventricle (Weidmann, 1966), the somatic muscle cells in Ascaris (De
-
546
ORKAND AND ORKAND
Bell, del Castillo, and Sanchez, 1963), and the mammalian smooth mus& (Tomita, 1966). By selectively stimulating one root, one side of the muscle contracts more strongly than the other. However, we do not know if during normal feeding behavior this ability is utilized (Huxley, 1853). It is possible that axons in separate roots originate from the same cell body, as has been found in Pleurobranchea (Siegler, Mpitsos, and Davis, personal communication). Hoyle and Smyth (1963) have found that giant barnacle muscle fibers are innervated by a number of axons, some of which travel in separate nerve roots. Due to facilitation, it is possible to achieve greater depolarization of the muscle membrane with multiple impulses over the same excitatory pathway than with the same number of impulses in more than one motorneuron. The importance of this facilitation awaits an understanding of the functional relations of the motorneurons within the buccal ganglion. In this connection it is interesting that neurons within the buccal ganglion are capable of responding to excitatory stimuli at 50 impulses per second (Gardner, 1971); this is well in excess of the five to ten impulses per second necessary to cause contraction of the muscle through summation and facilitation of ejp’s. These results are in striking contrast to those obtained by Kater et al. (1971) for three muscles in the gastropod mollusc Helisoma, where only excitatory neuromuscular transmission similar to that of the frog and mammalian “twitch” systems, was found. The excitatory innervation of the extrinsic protractor is possibly cholinergic. Relatively low concentrations of acetylcholine ( lop5 M ) cause sustained contractures (unpublished observations), as is the case for the radula protractor of Busycon (Hill, 1970). Much higher concentrations M ) of glutamate, dopamine, epinephrine, histamine, or 5-hydroxytryptamine do not produce visible contractures. Until the site of acetylcholine action to produce contracture is delineated (Florey and Kreibel, 1969) and further pharmacological studies are made, it is best to defer conclusions regarding the possible cholinergic nature of transmission. Acetylcholine, as well as dopamine and glutamate produce contractures of a small nearby muscle in the buccal mass (Taraskevich, Gibbs, Schmued, and Orkand, unpublished observations). The 17 presynaptic terminals surveyed in the electron microscope are similar in appearance to cholinergic junctions in vertebrates (e.g., Birks, Huxley, and Katz, 1960), since they contain primarily electron-lucent vesicles with a few electron-dense vesicles. They differ from vertebrate neuromuscular junctions in not having an elaborate subneural apparatus or basement lamella in the synaptic cleft. In these respects, they more closely resemble neuromuscular junctions in crustacea (Atwood and Morin, 1970) and other molluscs (Heyer et al., 1973). Since all of the terminals we observed had a similar morphology, we do not know if they were all excitatory, inhibitory, or a mixture of both types. This study was supported by U S . Public Health Service Grants NS-08346, NS-09521, and PL480. The authors thank Djordje Konjevic and Linda Macaulay for necessary technical assis-
APLYSIA NEUROMUSCULAR JUNCTIONS
547
tance and Professors J. D. French and L. Rakic, whose mutual interest and effort in fostering international cooperation in research made the study possible. REFERENCES
ATWOOD, H. L. and MORIN,W. A. (1970). Neuromuscular and axo-axonal synapses in the crayfish opener muscle. J . Ultrastruct. Res. 32: 351-369. RANKS,F. (1975). Inhibitory transmission a t a molluscan neuromuscular junction, J. Neurobiol. 6: 429-433. BIRKS,R., HUXLEY,H. E. and B. KATZ (1960). The fine structure of the neuromuscular junction of the frog. J . Physiol. 150: 134-144. BIIRKF,, W. (1957). Spontaneous potentials in slow muscle fibres of the frog. J . Physiol. 135: 511-521. CAREW,T. J., PINSKER, H., RUBINSON,K. and E. R. KANDEL(1974). Physiological and biochemical properties of neuromuscular transmission between identified motoneurons and gill muscle in Aplysia. J . Neurophysiol. 27: 1020-1040. DAVIS,W. J . and MPITSOS, G. J. (1971). Behavioral choice and habituation in the marine mollusk Pleurobranchaca californica. 2. Vergl. Physiol. 75: 207-232. DE BELL,
