An Electron Microscope Study of Synaptic Contacts in the Abdominal Ganglion of Aplysia californica ‘

J. P. TREMBLAY,’ M. COLONNIER AND H. MCLENNAN Laboratoires de Neurobiologie, H6pital de 1’Enfant-Jesus, Pauillon Notre-Dame, 2075 aue d e Vitre, Quebec, Quebec, G I J 5 B 3

ABSTRACT The fine structure of the abdominal ganglion of Aplysia californica has been studied in preparations fixed by immersion in aldehydes, either directly or after a survival of a few hours in artificial sea water. The central core of neuropil is surrounded by a rind of neuronal cell bodies floating in a subcapsular space containing a loose meshwork of neuronal and glial processes, separated by wide extracellular spaces. Large primary processes with deeply infolded membranes leave the neuronal perikarya and enter the neuropil where they branch into smaller processes containing either neurofilaments, neurotubules or both. Some have the appearance of initial segments. The neuropil is not a homogeneous structure. Rather, four types of zones can be distinguished: (1) zones of fibers of passage coursing together in the neuropil and making few synaptic contacts; (2) zones of neurosecretory fibers containing large granules and dense-core vesicles, again making few synaptic contacts; (3) zones with a great variety of synaptic contacts between medium size and small profiles; and (4) glomerular zones. The differentiated membranes of the synapses are characterized by a slight increase in density and by being regularly parallel to each other. Presynaptic densities are sometimes quite prominent but specialized dense cytoplasmic opacities have never been seen bordering the postsynaptic membranes, i.e., all synapses are of the symmetrical type. Interlemma1 opacities vary considerably in density. In zone 3, the synaptic vesicles are of several sizes, are round, oval or flat, and are either clear or filled with different types of dense material. The population of vesicles within a single profile may consist either of a homogeneous group of similar vesicles, or of various mixtures of two or three kinds of vesicles. In profiles with mixtures of clear and large dense-core vesicles, it is often only the clear vesicles which agglomerate towards the differentiated membranes. In such cases the large dense-core vesicles lie as a peripheral halo around the clear vesicles. Here, and especially in other large neuronal profiles not forming contact in the plane of section, they can be seen to associate specifically with mitochondria and glycogen. It is proposed that they do not contain neurotransmitters but are related to mitochondrial activities such as the storage of ATP or the movement of calcium ions. In profiles with mixtures of clear and small dense-core vesicles, both types of vesicles often touch the presynaptic membrane, suggesting the release of two transmitters or of a modulator or neurohormone with a transmitter, by a single terminal. Serial synapses are present in this zone. The glomerular zones contain small profiles forming many synaptic contacts, some of which are arranged in such a way as to suggest the existence of “reciprocal” serial synapses. The abdominal ganglion Of APlYsia californica has several individually identifiable neurophysiologiant neurons‘ Many gists have taken advantage of this characterJ. COMP. NEUR. (1979)188: 367-390.

1

Supported by the Medical Research Council of Canada. Departement d’anatomie, Faculte de Medecine, Universite

La:&lam Senior Fellow on leave from the Department of Phymology, University of British Columbia, Vancouver.

367

368

J. P. TREMBLAY, M. COLONNIER AND H. MCLENNAN

istic to study synaptic transmission (Arvanitaki and Chalazonitis, '49; Kandel and Tauc, '65a,b), the bursting activity of certain cells (Arvanitaki and Chalazonitis, '64), the initiation and propagation of spikes (Tauc and Hughes, '62), pharmacological properties (Gerschenfeld, '73; Kehoe, '72) and especially the relations between behavior and electrophysiology (Kupfermann and Kandel, '69; Kandel, '76). The nervous system of Aplysia has also attracted the attention of morphologists. In 1961, Bullock published a light microscopic study of the abdominal ganglion. The first electron microscopic study was made by Rosenbluth in 1963. He described the capsule, the peripheral zone, the subcapsular sinus, the granular content of the neuronal cell bodies and their invagination by glial cells. He noted that in the neuropil most synaptic vesicles have osmiophilic centers. Gerschenfeld ('63) studied the synapses in the abdominal ganglia of other gastropod molluscs. He distinguished between clear, dense-core and neurosecretory vesicles, but reported that synaptic contacts were infrequent and interpreted them to be of the axo-axonal type. In 1967, Coggeshall published a light and electron microscope study of the abdominal ganglion ofAplysia. He demonstrated that the fibrous capsule contains blood vessels surrounded by myoendothelial cells and muscle fibers, and that granule-filled processes end within the sheath without morphological evidence for specialized contacts. Synapses were again described as being rare. The glial cells were shown to contain glycogen particles and to form desmosome-like contacts. Frazier et al. ('67) published a classical paper describing the morphological and functional properties of identified neurons. Morphologically, they emphasized the presence of dense granules in some neurons. They presented a few photomicrographs of synapses but once more considered them rare. Weinreich e t al. ('73) described vesicles with an eccentric opaque core as characteristic of 5HT neurons in Aplysia and Tritonia. Gillette and Pomeranz ('75) filled neuron L10 with cobalt and showed that this cholinergic cell contains both clear and dense-core vesicles. They considered synapses to be abundant on this neuron but could not demonstrate its processes to be presynaptic to other profiles. The synapses on L10 and R2 were also described by Thompson et al. ('76) following intracellular injections of radioactive fucose. More recently Graubard ('78) has reported the existence of serial synapses.

All the studies of the general organization of the whole ganglion have been done after primary osmium fixation. Some of the more recent studies have used aldehyde fixation, but they have focussed their attention on the morphology of particular neurons. We therefore decided to study synaptic types and their overall distribution in the ganglion, taking advantage of the wide areas of well fixed material that can be obtained after aldehyde immersion. We were surprised to find that a t least four types of zones could be distinguished in the neuropil of the abdominal ganglion on the basis of the number and type of vesicle-containing profiles, and that a t least one of these zones contained many more typical synaptic contacts than one would imagine from reading the literature. We have also found evidence of synaptic contacts on glial cells. The present paper deals with interneuronal junctions. A second (Colonnier et al., '79) presents the evidence for the existence of neuroglial contacts. MATERIALS AND METHODS

The abdominal ganglia were dissected from five healthy Aplysia californica (supplied by Pacific Bio Marine, Venice, California) weighing 100 to 200 g. They were fixed by immersion immediately after dissection or after 2 to 3 hours in artificial sea water. The fixative consisted of 2%formaldehyde and 1%glutaraldehyde in a 0.1 M PO, buffer a t pH 7.4, containing 3%NaCl and 4%sucrose. After 4 hours in the fixative, the ganglia were cut in four quarters with a razor blade and fixed for an additional 8 hours. They were kept overnight in the 0.1 M PO, buffer a t pH 7.4, with 3% NaCl and 3% sucrose. They were then postfixed in 2%osmium in a 0.1 M PO, buffer a t pH 7.4, with 3% NaCl but no sucrose, for 90 minutes. The samples were dehydrated in ethanol, embedded in epon, sectioned and stained on the grid with uranyl acetate and lead citrate for electron microscopy. Sections for orientation in light microscopy were stained with toluidine blue. RESULTS

Light microscopy Ganglia of invertebrates are typically organized in the form of a fibrous core surrounded by a so-called nuclear rind, itself covered by a connective tissue capsule (fig. 1A). The nuclea r rind consists of large neuronal cell bodies and of small glial cells. Each neuron sends a single primary process into the darkly stain-

SYNAPTIC CONTACTS I N APLYSIA

369

Fig. 1A Cross section through one quarter of an abdominal ganglion of Aplysia californica. cap, capsule; g, glomeruli; gn, glial nuclei; ss, subcapsular sinus; ssp, subcapsular space. B Primary process of neuron coursing through the subcapsular space to the neuropil. gn, glial nuclei. C Neuronal cell body within the neuropil. gn,glial nuclei.

ing core to form a dense neuropil. Notice that there is a light staining peripheral zone (ssp) around the dense central core, separating it from the cell bodies. The primary process must pass through the loosely arranged glial meshwork of this area (see electron microscopic description below) to reach the neuropil. A large irregular clear space can be seen reaching up to the capsule in the upper right corner of

figure 1A (ss): this is the subcapsular sinus (Rosenbluth, '63). The large neuronal cell bodies are characterized by the punctate arrangement of their nuclear chromatin and by the segregation of their cytoplasm into a dark internal endoplasm and a lighter staining external ectoplasm (fig. 1B). This segregation is typical of large neurons in gastropods (Bullock and Hor-

370

J. P. TREMBLAY, M. COLONNIER AND H. MCLENNAN

ridge, ’65).Their surface membrane is irregu- small neuronal processes in the bundles also larly folded into many small processes. The lie close to each other and the whole bundle is large primary process emerging from the neu- surrounded by one or two layers of glial procron often immediately breaks up into numer- esses. Glycogen-containing profiles of glial ous parallel branches, many of which fuse origin also lie in intimate contact with neuback into larger trunks (fig. 1B). ronal cell bodies, which they invaginate (fig. Though most neuronal cell bodies are found 2: g). The folds of the neuronal cytoplasmic in the rind, a few rare individuals are embed- membrane seen in light microscopy are the reded in the neuropil, where they are easily dis- ciprocal image of these invaginations. tinguished from the small surrounding glial The small neuronal profiles coursing as buncells (fig. 1C). dles in the peripheral zone are often seen to Even in light microscopy, it is obvious in to- contain dense-core and clear vesicles. Typical luidine blue stained sections, that the neu- synaptic contacts are formed between them. ropil does not consist of homogeneous tissue. One of these, though not obvious a t this low Some areas are quite dark; others are paler. In power, is present in figure 2 (s) and another is figure lA, three small pale areas (g) are par- illustrated a t a higher power in the companion tially surrounded by the processes of a glial paper (Colonnier et al., ’79). Another contact cell. These encapsulated zones will be seen in is present on a process outside of the bundles, electron microscopy to contain a synaptic neu- a t the right of figure 2 (arrow). Such contacts ropil different from that present in other parts have been interpreted as axoglial synapses of the ganglion and for this reason will be (Colonnier et al., ’79). called “glomeruli.” Neuronal processes Electron microscopy The complex invaginations of the memNeuronal cell bodies and peripheral zone branes of large neuronal processes coursing to Part of a neuronal cell body and a segment and entering the neuropil are illustrated in of the light staining peripheral zone surround- figure 3. In transverse section (fig. 3B) they ing the fibrous dense core can be seen in fig- often acquire strange forms, the cytoplasm giving the appearance of cut-out figures. ure 2. Within the cytoplasm of the cell body, the These figures probably extend for long disdistinction between endoplasm and ectoplasm tances, constantly changing their configurais evident. It is due to the tendency of ribo- tion. Indeed in longitudinal sections, the large somes to agglomerate towards the nucleus, processes (4-5 Fm) seem to break up into many and for neurotubules and neurofilaments to parallel branches which rejoin as larger occupy a larger portion of the cytoplasm close branches, or possibly as a complete trunk furto the cytoplasmic membrane. ther along the process (fig. 3A). The infolded Within the peripheral zone, a glial cell is neuronal membranes are usually closely apidentified by its small size and by the promi- posed together, though intervening glial pronent perinuclear chromatin of its nucleus files are sometimes present. These large neu(gn).It gives rise to a long process which ex- ronal profiles typically contain many neutends to partially enclose a bundle of small rotubules and few neurofilaments. Within the neuronal processes filled with microtubules neuropil, some large profiles are also seen with and coursing through the region. The glial a predominance of neurofilaments but their process contains small discrete grains seen a t membranes do not show complex infoldings. higher power to consist of glycogen-like parti- Medium size and small processes may contain cles. Several other glycogen-containing pro- mainly neurotubules or mainly neurofilafiles can be seen, largely separated from other ments. These do not have membranous folds, processes by wide extracellular spaces except but some medium-size, tubule-filled processes where they surround neuronal processes. Neu- have odd shapes and surround other small ronal processes, on the other hand, do not neuronal profiles (fig. 6).A dense material is usually lie free within the large extracellular sometimes seen bordering the membranes of spaces of the peripheral zone. The only excep- some profiles containing neurotubules and tions are a few isolated profiles filled with neurofilaments (fig. 3C) in a manner reminislarge dense-core vesicles. Most are surrounded cent of that described for initial segments of by closely apposed glial processes, the inter- axons in vertebrates (Palay et al., ’68; Peters vening space measuring about 20 nm. The et al., ’68). Within these profiles, the tubules

SYNAPTIC CONTACTS IN APLYSIA

371

Fig. 2 Subcapsular space of abdominal ganglion. g, glial process invading the neuronal cell body; gn, nucleus of a glial cell; n, nucleus of neuronal cell body; nb, neuronal bundles; 8 , synaptic contact in a neuronal bundle; arrow, aynaptic contact interpreted to be on a glial cell process.

sometimes appear to converge into bundles as also described for initial segments, but this is never as distinct as in vertebrate material. Zones of neuropil Four types of loosely interwoven zones can be distinguished within the neuropil on the basis of the number and type of synaptic contacts present. A first type of zone consists mainly of neuronal processes filled with typical, evenly dispersed microtubules and neurofilaments, and rarely forming synaptic contacts with adjoining processes. The processes are undoubtedly

groups of fibers of passage and would correspond to some of the paler zones seen in light microscopy. In some instances, as in figure 4, many of the smaller neuronal processes appear to converge upon one or two vesicle-containing profiles. A second type of zone consists of large numbers of contiguous profiles, containing large dense granules and large dense-core vesicles but never forming synaptic contacts with each other (fig. 5A). Most of the dense organelles in figure 5B are what we have called granules; measure from 50 to 200 nm; are round, oval or flat and vary from light gray to black. Most do

372

J. P. TREMBLAY, M. COLONNIER A N D H. MCLENNAN

Fig. 3A Longitudinal section through primary process of neuron, in the neuropil. B Cross section through primary process of neuron in the neuropil. C Initial segment of axon in the neuropil.

SYNAPTIC CONTACTS IN APLYSIA

373

Fig. 4 Zone of fibers of passage.

not appear to have a membrane. This is particularly convincing when the granule is light gray. When they are very dense, a membrane may not be visible on some of them because the opaque material extends up to and hides the membranes. The large dense-core vesicles measure up to 130 nm. The dense-core is separated from an obvious enveloping membrane by a light staining peripheral halo (fig. 5B: arrow). Though the profiles in this zone do not form synaptic contacts with each other, they are sometimes postsynaptic to profiles containing other types of vesicles and lying a t the periphery of the zone. A suggestion of this is indicated by the arrow in figure 5A. A third type of zone more closely resembles vertebrate neuropil. I t contains many medium size to small processes filled with neurotubules and neurofilaments and vesicle-containing profiles forming many typical synaptic contacts. Figure 6 illustrates such a zone. A large bundle of processes can be seen to the right. One of the processes has an odd shape and surrounds several of the others. Two synaptic contacts are formed between the processes of the bundle (9). To the left, four vesicle-

containing profiles can be seen synapsing in the plane of section. A synaptic contact is said to be present between two apposed profiles when the following criteria are met. At least one of the profiles contains a group of vesicles which agglomerate towards a specific portion of the two apposed membranes. These portions of both membranes are differentiated, i.e., they are slightly denser and more rigidly parallel to each other than most other parts of the membranes. A more or less structured opacity can be seen between them. Small dense opacities are sometimes present between the vesicles on the presynaptic membrane. Note that the density of the differentiated membranes in Aplysia varies considerably from one contact to another. In some, they are only slightly denser than most adjoining membranes (fig. 7A). In others, they are very dense and thick (fig. 8E). These differences cannot usefully serve to define different types of synapses on the basis of membrane differentiation because there is a continuous gradation of densities from one to the other. Moreover, we have never seen a well defined opacity in the cyto-

374

J. P. TREMBLAY, M. COLONNIER AND H. MCLENNAN

Fig. 5A Zone of neurosecretory fibers. arrow, synaptic contact on neurwecretory fiber. B Profile containing mainly large “granules.” arrow, large dense-core vesicle. For explanation, see text

SYNAPTIC CONTACTS IN APLYSIA

375

Fig. 6 Zone of mixed synaptic contacts. 8, synaptic contacts.

plasm immediately bordering the postsynaptic membrane. All synapses therefore appear to be of Gray’s type 2 (Gray, ’59) or of the socalled “symmetrical” variety (Colonnier, ’68). Like Gray’s type 2 many of them show little or no increase in size of the extracellular space between the differentiated membranes. The closest approximation to an “asymmetrical” contact are the very prominent membrane differentiations illustrated in figure 9C. From the point of view of vesicle morphology however, there are many different kinds of synapses. There is a wide variety of vesicles in terms of size, shape and content, and there are many different mixtures of these different types of vesicles in individual profiles. Some terminals contain mainly clear, round

vesicles measuring approximately 50-90 nm and agglomerating towards differentiated membranes (fig. 7). Careful examination, however, shows that the density inside the vesicles varies considerably and that small punctate opacities are seen in some of them. Larger dense-core vesicles are often seen to the side of the vesicle agglomeration away from the differentiated membranes. Other terminals contain nearly exclusively dense-core vesicles or granules (fig. 8).Figure 8A illustrates a profile with large round to oval granules many of which do not appear to be enclosed in a membrane. They are of the same type and size as those described in the zone of neurosecretory fibers. Very rarely, as here, the granules abut onto differentiated

376

J. P. TREMBLAY, M. COLONNIER A N D H. MCLENNAN

Fig. 7 A-C. Presynaptic profiles with predominantly clear vesicles. a-e and arrow: for explanation, see text.

membranes, and the profile may thus be considered a s forming a synaptic contact. In most cases, however, the presynaptic profiles contain dense-core vesicles with obvious enclosing membranes. In some profiles, they are round to oval, and measure up to 130 nm (fig.

8B). In other profiles, vesicles of t h e same size tend to be mainly oval in shape and are probably best considered as egg shaped (fig. 8 0 . Another group of terminals contains uniformly small dense-core vesicles measuring 50-80 nm. The vesicles often tend to form rather dis-

SYNAPTIC CONTACTS IN APLYSIA

Fig. 8A Presynaptic profile with large dense “granules.” B Presynaptic profile with large dense-core vesicles. C Presynaptic profile with ovoid large dense-core vesicles. D Presynaptic profile with small dense-core vesicles. E Presynaptic profile with small dense-core vesicles. F Presynaptic profile with small and large dense-core vesicles. G Presynaptic profiles with small and large dense-core vesicles.

377

378

J. P. TREMBLAY, M. COLONNIER AND H. MCLENNAN

Fig. 9 A-D. Presynaptic profiles with mixtures of clear and large dense-core vesicles in which the two types are segregated and only the clear vesicles touch the presynaptic n'smbrane. gly, glycogen-like particles; m, mitochondria.

tinct rows along the differentiated membranes. Figures 8D and 8E illustrate such contacts with one and three rows of vesicles, respectively. Finally some terminals contain mixtures of dense-core vesicles varying in size from 50-120 nm, and in shape from round to flat (figs. 8F,G). In these cases it is nearly always the smaller vesicles which agglomerate toward the differentiated membranes. The larger vesicles tend to lie away from the contacting surface. Apart from these two diametrically opposed groups of vesicle populations, i.e., those with

mainly clear and those with nearly exclusively dense-core vesicles, there is a wide variety of vesicle populations made up of mixtures of clear and dense-core vesicles. These can be subdivided into two groups. In the first group, the clear vesicles are remarkably well segregated from the dense-core vesicles. In such cases, it is the clear vesicles which agglomerate towards the differentiated membranes, the dense-core vesicles lying behind the agglomeration, away from the differentiation (fig. 9). In all cases, the clear vesicles represent a relatively homogeneous round popula-

SYNAPTIC CONTACTS IN APLYSIA

379

Fig. 10 A,B. Association of mitochondria, glycogen and mitochondria in large neuronal profiles. For explanation, see text.

tion measuring 50-80 nm. The dense-core vesicles, however, vary from one terminal to the next. The one illustrated in figure 9A has small and large, round dense-core vesicles measuring up to 130 nm. The one in figure 9B has a mixture of round to flat, and small to large dense-core vesicles again measuring up to 130 nm. A few dense granules are also present. The terminal in figure 9C has large densecore vesicles measuring up to 200 nm, but here the dense core consists of finely dispersed, discrete small particles. That in figure 9D is similar but some of the vesicles contain the more usual homogeneous dense cores. Some of the dense vesicles in this terminal seem to surround a mitochondrion (m) and there is a patch of glycogen close by. A mitochondrion can also be seen amongst the dense-core vesicles in figure 9C and a patch of glycogen in 9B. (Note also the patch of glycogen associated with the few dense-core vesicles in figure 7A.l It is interesting in this respect that in numerous large and medium size processes, granules and dense-core vesicles are clearly associated

with mitochondria and patches of glycogen (fig. 10). The relation is particularly clear in one group of figure 10A, where the outer membrane of the mitochondrion extends to enclose a sac filled with glycogen particles. In the second group of mixed populations, clear and dense-core vesicles are uniformly intermixed and both abut onto the differentiated membranes. An often encountered example is seen in figure 11A. Here the size of the dense-core vesicles, 50-90 nm, is similar to that of the clear vesicles (50-80 nm). A more unusual example is seen in figure 11B. A mixture of small clear and dense-core vesicles are seen close to the membrane. Uniformly dense vesicles, some of which are elongate, are present as a second tier behind them. Another unusual type is illustrated in figure 11C. Here a few of the vesicles measure up to 100 nm, but the striking feature is the long stream of closely packed vesicles extending from the membrane differentiation. As many as 550 are present in this one cross section. Not all of the foregoing types of vesicle pop-

380

J. P. TREMBLAY, M. COLONNIER AND H. McLENNAN

ulations can be interpreted as discrete, well separated groups. There are many intermediate types. The synapse of figure 7A, for example, is a borderline case which was presented as containing mainly clear vesicles but could easily have been classified with those containing a mixture of clear and dense-core vesicles, where only the clear vesicles agglomerate towards differentiated membranes; again, in some terminals which appear to have segregated clear and dense-core vesicles, one or two small dense-core vesicles may abut onto the differentiated membranes. The large number of possibilities and the many intermediate types have defied all our attempts a t a rigorous classification.

There is, moreover, another unusual population of vesicles t h a t warrants description. I t is composed of vesicles measuring 25-140 nm and characterized by unusually prominent membranes (fig. 12). In some preparations, there seems to be a granular material attached to the membrane, giving the appearance of a dense periphery rather than a dense core. Most profiles containing such vesicles do not form synaptic contacts. In the rare cases when they do, the vesicles bordering the presynaptic membrane are of the more usual small clear variety (fig. 12). Presynaptic elements may contain predominantly neurofilaments (figs. 8C,E,F) or microtubules (fig. 11C), approximately equal num-

Fig. 11 A-C. Presynaptic profiles with evenly intermixed clear and small denee-core vesicles, both abutting on the presynaptic membrane. For explanation, see text.

SYNAPTIC CONTACTS IN APLYSIA

bers of both types of organelles, or relatively few of either type (fig. 7C: profiles c, d). It is obvious from the last figure that some of the clear presynaptic processes such as those of

Fig. 12 Presynaptic profile containing small to large vesicles. The membranes of the vesicles are bordered by an intravesicular opacity.

38 1

profiles c and d are continuous with others containing mainly neurotubules (profile e). Similarly, postsynaptic processes may also contain mainly neurofilaments (fig. 7B: profile c) or mainly neurotubules (fig. 8E) or they may be relatively free of either organelle. Those without filaments and tubules are either small spine-like protrusions from larger profiles (fig. 8G) or processes containing groups of vesicles (fig. 7C: profile c). It is thus not surprising that serial synapses are quite frequently encountered. One is illustrated in figure 7B where profile “a” contacts profile “b,” itself presynaptic to profile “c.” Another is seen in figure 7C where profile “a” contacts profile “c” which forms a tangentially-cut synapse on the tubule-filled profile “b.” Profile “c” itself is continuous with swelling “d” containing a clump of glycogen surrounded by dense-core vesicles. At the arrow, a group of clear vesicles points to a small dense triangle of membranes, suggesting a contact with two adjoining tubule-filled profiles. In several instances, there is a group of vesicles to each side of differentiated membranes (fig. 13). One of these groups usually abuts on the differentiated membranes (fig. 13A: a) and is best considered as presynaptic. In such cases the presence of vesicles close to but not touching the differentiated membrane on the postsynaptic side may be fortuitous’ In Other Cases, however, both €TOUPS of vesicles are equidistant from the differentiated mem-

Fig. 13 A,B. Arrangements suggestive of reciprocal synapses. a-c and arrow: for explanation, see text.

382

J. P. TREMBLAY, M. COLONNIER AND H. MCLENNAN

Fig, 14 Two contacts on the stem of a spine-like protrusion.

branes (fig. 13B: between profiles a and b) and one is tempted to interpret these as reciprocal contacts. Figure 13B also serves to illustrate a n oddity which is sometimes seen in the extracellular space, when three profiles meet to form triangular shaped spaces close to synaptic contacts, i.e., the presence of a small circular membranous profile in the extracellular space itself (arrow). Another regularly though infrequently encountered postsynaptic element is the narrow stem of spine-like protrusions extending from large tubule-filled processes (fig. 14). It usually receives two contacts and is darker than either the parent profile or the spine spherule. Often one of the terminals contains mainly clear vesicles and the other, a mixture of evenly intermixed, clear and dense-core vesicles. The fourth type of zone contains many very small tubule-filled profiles measuring about 0.25-0.50 p m in cross section. They often form large glomeruli surrounded by glial processes (fig. 15).The most surprising feature of these glomeruli, when compared with the third type of zone described above, is t h a t nearly all presynaptic profiles have virtually the same type of vesicle population. These consist of mixtures of clear and dense-core vesicles measuring up to 120 nm (fig. 16). When they form contacts in the plane of section, the postsynaptic membrane differentiation is often minimal but the presynaptic membrane is

studded with prominent presynaptic densities in between the vesicles. Some of them synapse on small to medium size processes containing mainly neurofilaments (fig. 16A). Many congregate to form many contacts on single, very small clear spine-like protrusions measuring only 250-450 nm (figs. 16C,D). In others the vesicles aggregate a t an angle which is apposed to two other profiles (fig. 16B: c). The presynaptic densities lie a t that junction as if both adjoining profiles are being contacted. The postsynaptic elements contain either the same type of vesicles as the terminal or are filled with microtubules. Sometimes two adjoining terminals, both containing vesicles aggregated towards presynaptic densities, meet a t the same angle (fig. 16B: b,c) where they seem to contact each other as well as tubule-filled processes (a). Finally, more rarely, the postsynaptic element may contain dense-core vesicles or granules (fig. 16E). Though large groups of such terminals are found in the glomeruli, smaller groups of the same type exist in the rest of the neuropil, merging freely with the other zones. The cells of origin of the processes found in this glomerular zone are not known. DISCUSSION

Types of neuronal processes The usual description of invertebrate ganglia is of a ring of cell bodies surrounding a central core of neuropil. The cell bodies are said to send a large axon into the neuropil, where it subdivides into many branches. According to this scheme all contacts between processes in the neuropil would be axo-axonal. In electron microscopy, the large primary process given off by the cell body is filled with the evenly dispersed microtubules typical of dendrites i n vertebrate tissue. Further, within the neuropil, there are profiles containing mainly microtubules, which therefore would also be dendritic. Others contain mainly neurofilaments and would correspond to axons. Still others have the ultrastructural characteristics of initial segments. All these facts suggest that the cell bodies of many neurons should be considered as giving rise to a large primary dendrite which enters the neuropil where i t gives rise both to the initial segment of t h e axon and to secondary dendritic branches. This is consistent with the studies of Tauc (Tauc, '62a,b; Tauc and Hughes, '631, who showed in Aplysia that axonal spikes are initiated a t some distance from the cell body

SYNAPTIC CONTACTS I N APLYSIA

383

Fig. 15 Glomerular zone, encapsulated by glial processes.

and t h a t in between the trigger zone and t h e soma there is a n intermediate neuronal process which is of lower excitability. These processes of lower excitability have been estimated to extend a s far as 200-300 Fm, which places the trigger zone, presumably the initial segments, squarely in the neuropil. These physiological characteristics also suggest t h a t the primary process which arises from the soma would be more usefully considered a s a dendrite than as a n axon. Axons given off by dendrites rather than cell bodies are common in the vertebrate CNS. The whole matter of the definition of axons and dendrites in this context has been reviewed by Bodian ('62) who has presented a useful generalized scheme for vertebrate neurons. Calling t h e primary process a dendrite would unify the vertebrate and invertebrate schemes. Tauc's work also suggests however t h a t a n invertebrate neuron may give off several axons. This was con-

cluded from the fact t h a t t h e spike initiated in one axonal branch does not necessarily invade either other branches or the soma itself. In morphological terms this implies t h a t t h e primary process in Aplysia would give rise to several initial segments. Trophospongiurn a n d dendritic infolding One of t h e most striking features o f A p l y s i a neurons and of their primary dendrite coursing to the neuropil is the large infolding of their membranes. Around the cell body t h e infolding of the membrane is accompanied by a complementary invasion of glial processes. In other invertebrate neurons t h e glial cell bodies themselves invade the neuron. This arrangement is referred to as Holmgren's trophospongium, the name reflecting its presumed trophic function (see review, Bullock and Horridge, '65). I t has been described in Aplysia neurons by Bullock ('61) and Cog-

384

J. P. TREMBLAY. M. COLONNIER AND H. MCLENNAN

Fig. 16 A-E. Synaptic contacts in glomerular zone. a-d: for explanation, see text.

SYNAPTIC CONTACTS IN APLYSIA

geshall (’67). The infolding of the membrane of the primary process is often considered as belonging to the trophospongium and indeed in this situation the folds are also often accompanied by a glial invasion. However this is not always the case. The membrane sometimes seems simply to fold into complex patterns, while in others the large process seems to break down into several branches which run in parallel and fuse back together distally, with virtually no glial contribution. This should clearly not be considered as part of a “trophospongium.” I t has been suggested that the large folds explain that processes can be stretched by a factor of four without affecting conduction time (Batham, ’61).The membrane would simply unfold when the pulling pressure is applied. I t is difficult to conceive that this would be the case in those instances when the infolding results in a series of parallel branches which join back to form a large process. It has been suggested that the increase in membrane surface would lead to a decrease in conduction time and a lengthening of the time course of the action potential as it invades the soma. This is indeed what happens to axonal spikes when they are conducted antidromically towards the cell body (Tauc, ’62a). It is noteworthy that no synaptic contacts have been seen on neuronal cell bodies nor on the few large primary dendrites seen leaving the cell bodies in our sections.

385

account for the reports of axons coursing among the cell bodies found in several Golgi studies (Havet, 1899a,b; Bullock and Horridge, ’65: p. 183) in spite of the conspicuous absence of synapses on the soma. It must be pointed out however that synapses are sometimes seen on small free processes in the subcapsular space (arrow in fig. 2) which could be interpreted as somatic tendrils arising from the trophospongium. We rather believe that these contacts are on glial processes (Colonnier et al., ’79).

Zones of the neuropil The presence of different types of synaptic zones in the neuropil of the abdominal ganglion of Aplysia is surprising. This neuropil is usually considered a s undifferentiated and as such to represent a lower level of development (Bullock, ’61; Bullock and Horridge, ’65: p. 54) than that found in other invertebrates in which, for example, the neuropil is segregated into glomeruli. Zones with few or no contacts are obviously the equivalent of fiber tracts, although some fields, such as the one illustrated in figure 4, suggest an orderly arrangement of the processes around synaptic terminals. Islands of contiguous profiles filled with large dense granules and dense-core vesicles, never forming contacts with each other, suggest that these profiles are not synaptic terminals containing classical neurotransmitters. Subcapsular sinus and subcapsular space They might be interpreted as neuromodulaAround the cell bodies, and between them tory or neurosecretory. All typical synaptic contacts in the abdomand the densely stained central core of neuropil, there is a peripheral zone of loosely inal ganglion of Aplysia lack a well-defined arranged glial cells. This zone was identified specialized cytoplasmic opacity bordering the by Rosenbluth (’63) who demonstrated that post -synaptic membrane. The best examples its wide extracellular space is continuous with suggesting such an opacity in our material a still wider space immediately underneath would be a t best border-line in vertebrate mathe fibrous capsule, the subcapsular sinus. terial. The distinction between symmetrical Since electron microscopy clearly reveals the and asymmetrical synapses (Colonnier, ’68) is continuity of the subcapsular sinus with the therefore useless in the present context. Corspaces of the zone and since the cell bodies relations have been made between excitatory really lie within it, it seems logical to refer to and asymmetrical contacts and between inthe space containing this loose meshwork as hibitory and symmetrical contacts in vertethe subcapsular space. Apart from the neu- brate material (see review, Colonnier, ’74). ronal cell bodies and their primary dendrites, This concept has recently received strong supthis space contains bundles of small neuronal port by the demonstration that in cortex, gluprocesses surrounded by glial fibers. They are tamic acid decarboxylase is localized in termithus most reasonably interpreted as small nals of the symmetrical type, most of which processes derived from several primary proc- end on pyramidal cell somata and that their esses, coursing together and synapsing, within cells of origin are the aspiny stellates (Riback, the subcapsular sinus. The significance of this ’78). This system is believed to mediate perisomatic network is unclear, but would GABA-ergic inhibition. If the material re-

386

J. P. TREMBLAY. M. COLONNIER AND H. MCLENNAN

sponsible for the specialized opacity defining an asymmetrical contact is usually associated with excitatory synapses in vertebrate CNS, it is obvious that i t is at least not present in the same form or in sufficient quantities in Aplysia to be easily revealed by the staining procedure we have used and no functional correlation is thus possible. The zone with many synapses is interesting because of the wide variety of contacts it contains in terms of vesicle morphology. A distinction was made between terminals containing mainly clear vesicles, mainly dense-core vesicles and mixtures of both. Those containing mixtures were further divided into subgroups. We cannot sufficiently emphasize that these should not be considered as discrete types. There is a wide variety of intermediate types, representing a continuous spectrum. Role of dense-core vesicles Terminals containing exclusively clear vesicles probably do not exist in Aplysia except as an artefact of section. Even in terminals where a large group of apparently clear vesicles agglomerate towards differentiated membranes, there are usually a few granules or dense-core vesicles on the outskirts of the agglomeration. This group fuses imperceptibly with that of terminals containing mixtures of clear and dense-core vesicles in which the clear vesicles abut on the differentiated membranes and are backed by a large population of dense-core vesicles. Also, those vesicles which we have classified as clear often contain small particles and densities making them quite similar t o evenly intermixed clear and dense-core vesicle populations. The vesicles in figure 7B illustrating one group are quite similar to those of figure 11C illustrating the other. Dense granules and dense-core vesicles thus seem to be omnipresent in Aplysia synaptic contacts. The extremes of their different distribution in different terminals suggest intriguing ideas as to their function. We have called granules the large (up t o 200 nm) round to oval electron opaque organelles which do not have a peripheral halo. These granules are present in several nerve processes. Large round dense-core vesicles (50-140 nm) are also abundant. They differ from the granules by having a well defined membrane or a peripheral halo. These two populations of organelles are similar to the dense-core granules described by Coggeshall ('671, Frazier et

al. ('67) and Chase and Goodman ('77) as belonging to the rostra1 "white" cells R3-Rl4. They would contain peptidergic material. Frazier et al. ('67) have shown that these white cells send some of their processes into the sheath (which would be a neurohemal organ) and other processes through the neuropil to the branchial nerve. The group of profiles containing granules and large dense-core vesicles described in figure 5A is possibly formed of such processes either ending here or on their way to the branchial nerve. We have mentioned that this group of fibers receives few synaptic contacts. This is in agreement with the physiological observation by Frazier e t al. ('67) that the "white" cells R3-Rl4 are relatively insensitive to neural inputs. Profiles containing similar granules and large dense-core vesicles do make rare synaptic contacts. They are illustrated in figures 8A and 8B. This suggests t h a t presumed neurosecretory cells can make specific synaptic contacts with other neurons. Is it thus possible that peptides might act as neurotransmitters in Aplysia as substance P has been suggested to do in mammals? Other hormones, such as the thyrotropin-releasing hormone (TRH) have been found in specific tracts in the spinal cord (Hokfelt et al., '75). TRH inhibits supraspinal neurons (Renaud and Martin, '75) and stimulates motoneurons in the frog (Nicoll, '77). It is therefore conceivable that processes from neurosecretory cells of Aplysia could form synaptic contacts and exert localized effects. Other types of dense-core vesicles are also found in synaptic contacts. Each vesicle type could contain a different transmitter. Densecore vesicles have been associated with monoamines (see review, Peters e t al., '76). Vesicles measuring 100 nm with an eccentric electron opaque material have been identified in 5 HT containing neurons of Aplysia (Weinreich et al., '73). Dense-core vesicles are also found, however, in the identified cholinergic neuron L10 and in its processes (Gillette and Pomeranz, '75; Thompson et al., '76) and purified dense-core vesicles from mammalian brain have been shown to contain ACh (Schubert and Klier, '77). Dense-core vesicles are frequently found in our study, in the same profiles as clear vesicles. The large dense-core vesicles are sometimes segregated and form a distinct group behind the clear vesicles. The small dense-core vesicles are often evenly intermixed with the clear vesicles, both touch-

SYNAPTIC CONTACTS IN APLYSIA

387

ing the presynaptic membrane. Several dif- brates (Peters et al., ’76) and of molluscs ferent hypotheses can be formulated t o ex(Orkand and Orkand, ’75). Yet their possible role in these terminals is rarely discussed. plain the presence of dense and clear vesicles in the same synaptic contact. A given synaptic Frequency of synaptic contacts profile may of course contain more than one transmitter. The presence of two neurotransAlthough previous studies (Gerschenfeld, mitters has been reported in the same cell in ‘63; Coggeshall, ’67; Frazier, ’67) using primolluscs: 5HT and DA (Kerkut et al., ’67); mary osmium fixation have reported that syn5HT and ACh (Cottrell, ’76); 5HT, octopamine aptic contacts with definite membrane differand ACh (Brownstein et al., ’74). These re- entiation are rare in the neuropil of the ports, however, have been contested (Osborne, abdominal ganglion of Aplysia, our study es’77) and attributed t o contamination of the tablishes that synapses are abundant in some sample. A second possibility is that the dense- regions of the neuropil. In non-glomerular core and clear vesicles represent different zones, they range from 0 to 42 contacts per 100 forms of packaging of the same transmitter. p m Z(average 10) and in the glomerular zones, As mentioned above ACh is sometimes associ- form 5 t o 40 contacts per 100 pm’ (average ated with dense-core vesicles. Clear vesicles 20). This is comparable to the molecular layer have also been considered as dense-core vesi- of turtle cortex where the density can be calcles emptied of their dense content (Tranzer culated from the data of Ebner and Colonnier and Thoenen, ’67). A third line of interpreta- (‘78) to average about ten synapses per 100 tion is that the dense-core vesicles found in pm’. the same terminals as clear vesicles do not Serial and reciprocal synapses contain a “classical” neurotransmitter but Serial synapses may be implicated in prerather a modulatory substance or hormone which would modify the action of the trans- synaptic inhibition and presynaptic facilitamitter contained in the clear vesicles. The tion. These phenomena have been studied in modulator could for example select the kind of Aplysia (Tauc, ’65; Kandel and Tauc, ’65a,b; receptor with which the transmitter will in- Tauc and Epstein, ’67; Castelluci and Kandel, teract. This hypothesis could account for the ’76; Shimahara and Tauc, ’75). Graubard (‘78) change from excitatory to inhibitory re- has reported that 4% of the synapses are sponses to the same transmitter observed at serial. an identified cholinergic synapse in Aplysia Synapses with a Y shaped arrangement of (Watchtel and Kandel, ’71). Finally these the differentiated membranes as illustrated dense-core vesicles could contain some other in figure 16B are very abundant in the glomersubstance liberated during synaptic stimula- ular zones. They might be considered as a tion. I t is well known that ATP is liberated special class of serial synapses. In figure 16B, with ACh during synaptic stimulation, but profiles b and c both synapse on profile a. ProWhittaker et al. (’72) have pointed out that files b and c might also be considered as conATP and ACh are not necessarily in the same tacting each other reciprocally. The three prosynaptic vesicles. An interesting observation files might thus be considered as forming and in this respect is that the large dense-core ves- two s e r i a l sequences, i.e., b-c-a icles are often associated with mitochondria c - L a . We might thus speak of “reciprocal and glycogen. In synapses of the ciliary gan- serial synapses.” Apart from those contacts arranged serially glion, vesicles have been described budding from mitochondria (Hamori and Dyachkova, in the plane of section about one third of syn’64). Could the large dense-core vesicles con- apses are on profiles which contain vesicles. It tain some metabolite of mitochondria as for is more than likely that many of these vesicleexample ATP? It is also known that mito- containing postsynaptic profiles also form chondria accumulate calcium ions and are an synaptic contacts out of the plane of section important calcium buffer (Alnaes and Ra- and that the number of serial contacts far exhaminoff, ’75). Is i t possible that the dense- ceeds the reported percentages. Synapses on tubule-containing profiles repcore vesicles have a role t o play in the sequestering of calcium ions? Large dense- resent 23%of the total. In mammals, such procore vesicles are present in most synaptic ter- files are usually considered as belonging to minals containing mainly small clear vesicles, dendrites most of which would probably conand in the neuromuscular junction of verte- duct decrementally. Some of these are the in-

388

J. P. TREMBLAY, M. COLONNIER AND H. MCLENNAN

termediate elements in the serial synapses described above. Profiles containing predominantly neurofilaments also receive synaptic contacts. They are considered as axonal in mammalian tissue. If this correlation holds, i t would be the first demonstration of synaptic contacts along the course of an axon, since all axo-axonal synapses so far described have been either in the region of the initial segment or on an axonal terminal. Synapses along the course of an axon could modulate the propagation of the action potential. Number of vesicles and size of reserve pool I t is not possible to quantify the number of synaptic vesicles in a terminal without serial sections. Even on individual sections however it is obvious that there are large variations in the number of vesicles found in the different synaptic profiles. The synaptic contact in figure 8D for example has only one row of 11 synaptic vesicles close to the differentiated membrane. The cross section of the terminal of figure 8B contains 214 vesicles and that of figure 11C contains 550 vesicles. It is not unreasonable to assume that synapses of the types found in the last two examples would be able to sustain a constant rate of transmitter release during a long train of presynaptic stimulation while synapses similar to those illustrated in figure 8D might produce decremental release with sustained stimulation. Though the number of vesicles is surely not the only factor responsible for facilitation or depression during stimulation, i t may nevertheless be an important contributing factor. Our study of the general organization of the abdominal ganglion ofAplysia emphasizes the complexity of its synaptic organization. The neuropil is composed of four different types of zones and of many different types of synaptic contacts, while a complex perisomatic network exists in the subcapsular space. Moreover as presented in the next paper (Colonnier et al., '79) synaptic contacts are found on glial cells both in the neuropil and in the subcapsular space. ACKNOWLEDGMENTS

The authors thank Professor Eric Kandel for his helpful review of the manuscript. LITERATURE CITED Alnaes, E., and R. Rahaminoff 1975 On the role of mito. chondria in transmitter release from motor nerve termi. nals. J. Physiol. (London), 248: 285-306.

Arvanitaki, A., and N. Chalazonitis 1949 Prototypes d'interactions neuroniques e t transmission synaptique. DonnCs bioblectriques de preparations cellulaires. Arch. Sci. Physiol., 3: 303-338. 1964 Processus d'excitation d'un neurone autoBiol., 98: 1119. actif sur ondes lentes. C.R. SOC. Batham, E. J. 1961 Infoldings of the nerve fiber membranes in the opisthobranch mollusc Aplysia californica. J. biophys. biochem. Cytol., 9: 490-491. Bodian, D. 1962 The generalized vertebrate neuron. Science, 137: 323-326. Brownstein, M. J.,J. M. Saavedra, J. Axelrod, G. H. &man and D. 0. Carpenter 1974 Coexistence of several putative neurotransmitters in single identified neurons of Aplysia. Proc. Natl. Aead. Sci. (USA.), 71: 4662-4665. Bullock, T. H. 1961 On the anatomy of the giant neurons of the visceral ganglion of Aplysia. In: Nervous Inhibition. E. Florey, ed. Pergamon Press, London, pp. 233-240. Bullock, T. H., and G. A. Horridge 1965 Structure and Function in t h e Nervous Systems of Invertebrates. W. A. Freeman and Co.,San Francisco. Castellucci, V., and E. R. Kandel 1976 Presynaptic facilitation 88 a mechanism for behavioral sensitization in ApZysia. Science, 194: 1176-1178. Chase, R., and H. E. Goodman 1977 Homologous neurosecretory cell groups in the land snail Achatina fulica and the sea slug Aplysia californica. Cell Tissue Res., 176: 109-120.

Coggeshall, R. E. 1967 A light and electron microscope study of t h e abdominal ganglion ofAplysia californica. J. Neurophysiol., 30: 1264-1287. Colonnier, M. 1968 Synaptic patterns on different cell types in the different laminae of the cat visual cortex. An electron microscope study. Brain Res., 9: 268-287. 1974 Spatial interrelationships 88 physiological mechanism in t h e central nervous system. In: Essays on the Nervous System. R. Bellairs and E. G. Gray, eds. Clarendon Press, Oxford, pp. 344-366. Colonnier, M., J. P. Tremblay and H. McLennan 1979 Synaptic contacts on glial cells in the abdominal ganglion of Aplysia californica. J. Comp. Neur., 188: 391-400. Cottrell, G. A. 1976 Does t h e giant cerebral neurone of Helix release two transmitters, acetylcholine and serotonin? J. Physiol. (London), 259: 44-45P. Ebner, F. F.,and M. Colonnier 1978 A quantitative study of synaptic patterns in turtle visual cortex. J. Comp. Neur.,

-

179: 263-276.

Frazier, W. T., E. R. Kandel, I. Kupfermann, R. Waziri and R. E. Coggeshall 1967 Morphological and functional properties of identified neurons in the abdominal ganglion ofAplysia californica J. Neurophysiol., 30: 1288-1351. Gerschenfeld, H.M. 1963 Observations on the ultrastructure of synapses in some pulmonate molluscs. 2. Zellforsch., 60: 258-275. 1973 Chemical transmission in invertebrate central nervous systems and neuromuscular junctions. Physiol. Rev., 53: 1-119. Gillette, R.,and B. Pomeranz 1975 Ultrastructural correlates of interneuronal function in t h e abdominal ganglion of Aplysia californica. J. Neurohiol., 6: 463-474. Graubard, K. 1978 Serial synapses in Aplysia. J. Neurobiol., 9: 325-328. Gray, E. G. 1959 Axosomatic and axodendritic synapses of the cerebral cortex. J. Anat. (London), 93: 420-433. Hamori, J., and L.N. Dyachkova 1964 Electron microscope studies on developmental differentiation of ciliary ganglion synapsea in the chick. Acta Biol. (Hungary), 15:

213-230. Havet, J. 1899a Note preliminaire sur le systeme nerveux des L i m a (methode de Golgi). Anat. Anz., 16:

241-248.

SYNAPTIC CONTACTS IN APLYSZA

1899b L’etat moliniforme des neurones chez les invertebris. Trab. Lab. Investig. biol. Univ. Madr., 14: 35-85. Hokfelt, T., K. Fuxe, 0. Johansson, S. Seffcoate and N. White 1975 Distribution of thyrotropin-releasing-hormone (TRH) in the central nervous system a s revealed with immunohistochemistry. Eur. J. Pharmacol., 34: 389-392. Kandel, E. R. 1976 Cellular Bases of Behavior. W. H. Freeman and Co.,San Francisco. Kandel, E. R., and J. Kupferman 1970 The functional organization of invertebrate ganglia. Annu. Rev. Physiol., 32: 193-258. Kandel, E. R., and L. Tauc 1965a Heterosynaptic facilitation in neurons of the abdominal ganglion of Aplysia depilans. J. Physiol. (London), 181: 1-27. 1965b Mechanism of heterosynaptic facilitation in the giant cell of t h e abdominal ganglion of Aplysia depilans. J. Physiol. (London), 181: 28-47. Kehoe, J. 1972 Three acetylcholine receptors in Aplysia neurones. J. Physiol. (London), 225: 115. Kerkut, G. A., C. B. Sedden and R. J. Walker 1967 Uptake of DOPA and 5-hydroxytryptophane by monoamine-forming neurones in t h e brain of Helix aspersa. Comp. Biochem. Physiol., 23: 159-162. Kupfermann, I.,and E. Kandel 1969 Neuronal controls of a behavior response mediated by the abdominal ganglion of Aplysia. Science, 164: 847-850. Nicoll, R. 1977 Excitatory action of TRH on spinal motoneurons. Nature, 265: 242-243. Orkand, P. M., and R. K. Orkand 1975 Neuromuscular junctions in the buccal mass of Aplysia: fine structure and electrophysiology of excitatory transmission. J. Neurobiol., 6: 531-548. Osborne, N. N. 1977 Do snail neurones contain more than one neurotransmitter? Nature, 270: 622-623. Palay, S. L., C. Sotelo, A. Peters and P. M. Orkand 1968 The axon hillock and the initial segment. J. Cell Biol., 38: 193-201. Peters, A., S. L. Palay and H. F. Webater 1976 The Fine Structure of t h e Nervous System: The Neurons and Supporting Cells. W. B. Saunders &., Philadelphia. Peters, A., C. C. Proskauer and Kaiserman-Abramof 1968 The small pyramidal neuron of the rat cerebral cortex. The axon hillock and initial segment. J. Cell Biol., 39: 604-619.

389

Renaud, L. P., and J. P. Martin 1975 Thyrotropin releasing hormone (TRH): depressant action on central neuronal activity. Brain Res., 86: 150-154. Riback, C. E. 1978 Aspinous and sparselyspinous stellate neurons in the visual cortex of rats contain glutamic acid decarboxylase. J. Neurocytol., I: 461-478. Rosenbluth, J. 1963 The visceral ganglion of Aplysia californica. Z. Zellforsch., 60: 213-236. Schubert, D., and F. G. Klier 1977 Storage and release of acetvlcholine bv a clonal cell line. Proc. Natl Acad. Sci. cu.6A.i 74: 5184.5188. Shimahara, T., and L. Tauc 1975 Heterosynaptic facilitation in the giant cell ofdplysia. J. Physiol. (London),247: 321-341. T a w , L. 1962a Site of origin and propagation of spike in the giant neuron of Aplysia. J. Gen. Physiol., 45: 1077-1097. 1962b Identification of active membrane area in the giant neuron of Aplysia. J. Gen. Physiol., 45: 1099-11 15. 1965 Presynaptic inhibition in the abdominal ganglion of Aplysia. J. Physiol. (London), 181: 282.307. T a w , L.,and R. Epstein 1967 Heterosynaptic facilitation aa a distinct mechanism in Aplysia. Nature.214: 724-725. Tauc, L., and G. M. Hughes 1962 Modes of initiation and propagation of spikes in the branching axon8 of molluscan central neurons. J. Gen. Physiol., 46: 533-549. Thompson, E. B., J. H. Schwartz and E. R. Kandel 1976 A radioautographic analysis in the light and electron microscope of identified Aplysia neurons and their processes after intrasomatic injection of [W-L-fucose.Brain Res., 112: 251-281. Tranzer, J. P., and H. Thoenen 1967 Significance of “empty vesicles” in postganglionic sympathetic nerve terminals. Experientia, 23: 123-124. Watchtel, H., and E. R. Kandel 1971 Conversion of synaptic excitation to inhibition a t a dual synapse. J. Neurophysiol., 34: 56-68. Weinreich, D., M. W. McCaman, R. E. McCaman and J. E. Vaughn 1973 Chemical, enzymatic and ultrastructural characterization of 5-hydroxytryptamine-containing neurons from the ganglia of Aplysia californica and Tritonia diomedia. J. Neurochem., 20: 969-976. Whittaker, V. P., M. J. Dowdall and A. F. Boyne 1972 The storage and release of acetylcholine by cholinergic nerve terminals: recent results with non-mammalian preparaSymp., 36: 49-68. tions. Biochem. SOC. ~

~~~