Morphological correlates of synaptic transmission in lamprey spinal cord.

JOURNAL OF NEUROPHYSIOLOGY Vol. 39, No. 2, March 1976. Printed

in U.S.A.

Morphological Correlates of Synaptic Transmission in Lamprey Spinal Cord BURGESS

N. CHRISTENSEN

Division of Biological and Medical Providence, Rhode Island 02912

Sciences,

Brown

PROCESS BY WHICH transmitter is released at chemical synaptic junctions has been elucidated principally through electrophysiological analysesof the events at neuromuscularjunctions. The quanta1hypothesis, first formulated by Del Castillo and Katz (15), quantitatively describes the release of chemical transmitters from presynaptic terminals. This hypothesis has been tested at various other synaptic junctions in mammalian and invertebrate preparations (6-8, 17, 18, 28, 30, 32, 33). If, as a corollary of this hypothesis proposes, synaptic vesicles in presynaptic terminals represent the quanta1 package, a direct relationship should exist between releaseof transmitter and morphological alterations in synaptic vesicles. This relationship has been the subject of several morpholog. ical studieswhich describe the dynamics of vesicle movement and formation at both the neuromuscular junction and central nervous system (CNS) synapses. Andres (1) first suggested that vesicles might be formed by microinvagination of presynaptic membrane, and Westrum (49) extended this idea to include the possibility that the general source of vesicle membranemight be th$ entire presynaptic surface. Several authors have described ultrastructural changesin vesicles after prolonged electrical activity in presynaptic axons (11, 12, 19, 21, 23, 25, 26, 34, 36-38, 45). Heuser and Reese (22), in a study of vesicle formation associated with electrical activity in motor terminals at the frog neuromuscularjunction, showed that vesicular membrane is recycled. Their findings suggestthat after discharging its contents into the synaptic cleft, the vesicular membrane coalesces with the plasma membrane. This membraneis retrieved by coated vesicles pinching off from regions outside the area of synaptic contact. These vesicles then lose their coats and coalesce to form synaptic cisternae from which new vesicles are formed. THE

Received for publication

April

24, 1975.

University,

Previous studies have used the neuromuscular junction and autonomic ganglia to investigate the morphological changesassociatedwith transmitter release. The CNS poses special problemsfor the ultrastructural characterization of physiologically active synapses because of the complex synaptic organization and the difficulty in identifying active synaptic junctions. In the present experiments the ultrastructure of physiologically identified synaptic contacts made between a giant axon and a giant interneuron in the lamprey’s spinal cord were studied by combining intracellular dye injection with light and electron microscopy. Physiological experiments indicate that synaptic transmission occurs by a combined electrical and chemical mechanism. The morphological analysis of these same synaptic junctions confirmed this result. Ultrastructural differences in these same synaptic junctions are compared with the synaptic ultrastructure of unstimulated axons. The total number of synaptic vesicles in the presynaptic terminations of the stimulated axon was counted and compared with estimates of the fraction of the total vesicle pool available for release. METHODS

Preparation Sea lampreys (Pe tromyzon marinus) were caught in New England streams,transported to the laboratory, and kept in aerated aquaria at 9OC. Both larvae and newly metamorphosed adults were used in experiments. At the time of an experiment the animal was decapitated and the spinal cord removed and placed in a chamber where it was continuously perfused with oxygenated physiological solution at lo15OC.The composition of the perfusate was: 91 mM NaCl, 2.1 mM KCl, 2.6 mM CaC12,1.8 mM MgC12, 20 mM NaHCO,, and 4 mM glucose (40) Figure 1 (top) shows a diagrammatic representation of a lateral view of the sealamprey. 197

Downloaded from www.physiology.org/journal/jn by ${individualUser.givenNames} ${individualUser.surname} (129.186.138.035) on January 10, 2019

198

B. N. CHRISTENSEN

i rent

giant

ax

FIG. 1. Top: diagram of a lateral view of lamprey with the general position of the reticulospinal and giant interneurons sketched on the surface. Bottom: expanded view of reticulospinal cell axon-giant interneuron relationship within the spinal cord. Electrical apparatus and electrode position describe the experimental paradigm. Electrical stimulation, iontophoresis, or recording can be achieved through either electrode.

Giant interneurons are located in the lateral gray columns in the caudal one-third of the spinal cord. These are large neurons which may be

easily visualized during the experimental procedure by transillumination of the spinal cord. According to Rovainen (43) these interneurons are higher order sensory neurons. The axon of the interneuron has been traced by Rovainen et al. (44) to the level of the brain stem. Giant axons, whose cell bodies are located in the brain stem, descend the spinal cord making en passant synaptic contacts with several cell types including giant interneurons (4 l-43).

Electrophysiology Figure 1 (bottom) shows a diagrammatic view of the segment of spinal cord containing the cell illustrated in the top of the figure and includes the electrodes and recording apparatus. Beveled (2, 9, 27) microelectrodes filled with Procion brown (13) 4% w/v were used to penetrate a giant interneuron (Fig. 1 bottom). After penetration of the interneuron, a second electrode was used to search for a giant axon coupled monosynaptically to the interneuron. Penetrated axons were stimulated by passing depolarizing or hyperpolarizing current of sufficient duration and intensity through the electrode. Recording of a short-latency PSP

FIG. 2. Synaptic response (upper trace) to a single stimulation of the presynaptic axon (lower trace). Stimulation was achieved by passing hyperpolarizing current through the intracellular electrode. The action potential was produced by the rapid depolarization of the membrane when the current was turned off (anodal break). An electrotonic synaptic response (single arrow) and a chemical synaptic response (double arrow) are indicated. Scale 2.5 mV upper trace, 25 mV lower trace, 5 ms. from the interneuron (Fig. 2) confirmed monosynaptic coupling between the axon and interneuron. Rovainen reports that the EPSP produced on stimulation of a monosynaptically coupled Muller axon consists of both an electrical and a chemical synaptic response. The response in Fig. 2 resembles the synaptic action of a ventrally located Muller axon synapsing on a giant interneuron as described by Rovainen (4 1). The later, chemically mediated response (double arrow), was shown by Rovainen to disappear in the presence of high Mg2+ and low Ca2+, leaving only the electrotonic potential (single arrow). Not all giant axon-interneuron interactions show this mixed response; some consist only of a chemically mediated response (unpublished observations). Synaptic delay was calculated to determine if neurons were monosynaptically connected, The conduction velocity for the giant axon was estimated as the distance from axon stimulation site to giant interneuron (measured with an eyepiece micrometer) divided by the time from the onset of the action potential in the giant axon to the appearance of the electrotonic EPSP (assuming no synaptic delay for this response). In the preparation of Fig. 2 the conduction velocity was 3 m/s and the synaptic delay for the chemically mediated EPSP, 1.75 ms. In several experiments conduction velocities were measured in giant axons by inserting two microelectrodes at a known distance in the same axon. The injection of current through one electrode was followed by an action potential in the axon. The time of arrival of this action potential at the second electrode was measured and the conduction velocity calculated. The conduction velocities measured ranged from 2.5 to 4 m/s.


SYNAPTIC

TRANSMISSION

Morphology After recording physiological interactions, the giant axon and interneuron were both injected with the dye Procion brown until each could be seen in the spinal cord. This was accomplished by passing hyperpolarizing pulses (500 ms, l/s) for from 10 to 30 min. The spinal cord was then removed from the recording chamber and processed for light and electron microscopy (EM). During the iontophoretic injection of the dye the presynaptic axon was excited repeatedly by anodal break stimulation (cf. Fig. 2). Since the dye is toxic, as evidenced by loss of membrane potential during injection, the excitation cannot have occurred throughout the entire injection period, but its exact duration is uncertai:]. The tissue processing procedure described by Smith et al. (46) was used except for a slight modification in fixation times. The tissue was fixed in 2.5% glutaraldehyde for 30 min and postfixed in 2% osmium tetroxide for 3 min. Dehydration was carried out in an alcohol series followed by propylene oxide and the tissue embedded in a mixture of Epon and Araldite. After curing, sections 5 pm thick were cut using an LKB Ultratome III with a dry glass knife, removed from the knife, and floated on water drops on a clean glass slide. The injected giant axon was located and traced in serial sections to the level of the giant interneuron. Those 5-pm sections in which an injected dendrite appeared in close apposition to the injected giant axon were photographed at 240~ with bright field illumination. These light micrographs facilitated orientation and location of areas of interest when viewing in the EM. The sections were then removed from the glass slide and remounted on a newly faced blank plastic block (10) on which a drop of chrom-alum gelatin had been placed. The block was then placed in a 40°C oven until dry (5 min). Chrom-alum gelatin was used as the glue instead of epoxy cement or liquid embedding plastic because the section remains flat after drying. This facilitates subsequent alignment of the block face for ultrathin sectioning. This is an important point in technique since recovery of as many ultrathin sections as possible is vital for serial reconstruction of synaptic junctions and subsequent counting of synaptic vesicles. The remounted section was carefully trimmed with a razor blade, leaving the area containing the possible synaptic contact. This 5-pm section was then resectioned for electron microscopy on a LKB Ultratome III with a diamond knife. In the experiment in which all synaptic junctions were

IN

SPINAL

199

CORD

located, a total of 301 sections cut from five 5-pm sections were examined. The average section thickness was about 80 mpm, as determined from interference colors (silver to silver-gold). Sections were collected on Formvar-coated 2 x 1 mm slot grids. Occasionally thin sections were lost during sectioning or mounting on grids. The gaps in the series were noted so that if they occurred within a synaptic contact they could be accounted for during the serial reconstruction. Sections were examined with a Jeolco 100 B electron microscope at 100 kV. An estimate of the total number of synaptic vesicles was made by counting the number of vesicles in each section through a synaptic contact. These counts were made from 8 x 10 prints (final magnification 30,000~ ). All vesicles were counted irrespective of their ultrastructural characteristics (see below). No attempt was made to correct for counting a vesicle twice if it appeared in two adjacent sections. Each section from a contact was counted 5 times and the average calculated. In cases where a missing section occurred, the number of vesicles in the sections bordering the missing one were averaged and this number used for the missing section.

Three-dimensional

reconstruction

Three-dimensional reconstruction of synaptic contacts from the EM material was done using facilities in the Department of Biology at the University of California at San Diego and in the Department of Biology at Columbia University, New York (29). RESULTS

Morphology Figure 3A-E shows light micrographs of serial 5-pm cross sections of the spinal cord at the level of the giant interneuron. The giant axon of Fig. 3A (marked a), located in the dorsolateral spinal cord, was penetrated and injected with dye about 4 mm rostra1 to the interneuron and traced in 5-pm serial sections from the injection site to the level of the interneuron. Dye does not appear in the axon in these sections since the spinal cord was removed and fixed before the dye could diffuse to this level and possibly obscure presynaptic morphology. The unstained length of the injected axon was readily followed in serial sections to the level of the giant interneuron. The injected cell and dendritic processes appeared red in the light microscope and were easily distinguished in otherwise unstained material. The giant axon was


B. N. CHRISTENSEN

FIG. 3. A-E: light micrographs of five adjacent sections through the injected neuron, each 5 w thick. Three of the Procion brown-filled dendrites (1, 2, 3) can be followed in serial sections from cell (GI) to identified axon (a). F: low-power EM micrograph of identified axon (ax) and injected dendrite (1) from the section shown in A. Scale in A, 20 m for LM and 5 fun for EM.

found to make 22 discrete synaptic contacts on three different dendrites, labeled I, 2, 3 in the sections shown in Fig. 3A-E. These dendrites could be traced in the serial sections to the cell body and appeared in close apposition to the labeled axon. This technique left no doubt that an injected dendrite appearing in isolation at one level did in fact originate from the cell body at a different level. In sections on either side of those shown in the figure, no dye-filled processes were observed to approach this axon. The distance from closest synaptic contact to cell body for the three dendrites was 135, 110, and 1.50 pm, respectively. Nine contacts were made on dendrite 1, nine on 2, and four on 3. Figure 3F is a low-power electron micrograph of an ultrathin section taken from the 5-q section shown in Fig. 3A. Note the density of the dendritic process due to the Procion brown and the absence of the dye from the giant axon (ax) at this level. The arrow indicates the location of a synaptic contact illustrated at higher power in Fig. 4.

Several features of synaptic morphology were common to all 22 contacts made by the stimulated axon on the three dendrites of the giant interneuron. These are summarized briefly here, and illustrated in detail below. Every contact was associated with an extensive outpouching of the axoplasmic membrane. Outpouching of this degree was not generally seen at contacts made between the more ventrally located unstimulated axon (Fig. 3A arrow) on these same dendrites, nor has it been described in other studies of synapses in lamprey spinal cord (42, 43, 46). Some of the contacts were made on dendritic spines and some were made directly on the smooth shaft of the dendritic process. Fewer vesicles per contact were seen in the stimulated giant axon compared with contacts made between the unstimulated ventral axon and the dendrites of the giant intemeuron. Further, at unstimulated contacts the vesicles were generally arranged in a regular geometrical array (cf. Fig. 9C and Smith et al. (46)), whereas all of the 22 contacts made by the


SYNAPTIC

FIG.

4.

A-I:‘serial

TRANSMISSION

IN SPINAL

CORD

201

sections through one synaptic contact. See text for details. Scale 1.5 /.u%


202

B. N. CHRISTENSEN

stimulated axon showed features attributed to active synapses, including a disorganization from a regular geometrical array of synaptic vesicles, an increase in the number of complex vesicles, and an increase in synaptic structures considered to be synaptic cisternae (22). Figure 4A-I presents several serial sections taken through the contact indicated in Fig. 3F (arrow). These electron micrographs illustrate the general synaptic organization seen in all 22 contacts. The number of synaptic vesicles per

4

FIG. 5. A-C: stereo pairs of a computer reconstruction of the synaptic contact shown in Fig. 4. A: the arrow points directly into the socket formed by the outpouching of the presynaptic giant axon. The postsynaptic spine would fit into this socket. The sheet of membrane (axolemma) formed by the giant axon appears in stereo as disappearing into the page. B: postsynaptic dendrite and spine. The spine process is facing away (arrow) and ends of the dendrite have been truncated (open arrows). C: pre- and postsynaptic elements coupled. Spacing between ultrathin sections is exaggerated. These stereo pairs may be viewed with a simple stereo viewer. This reconstruction was done by J. Miller in the Dept. of Biology, University of California, San Diego.

section is surprisingly low and the vesicles are not arranged in any recognizable geometrical array (cf. Fig. SC). Since serial sections were taken through the entire synaptic region, it is unlikely that any associated synaptic vesicles were missed. The large arrows in Fig. 4A and G clearly show vesicles in intimate association with the axoplasmic membrane. These vesicles are identical to those described by Gray and Willis (21) as complex vesicles and by Kaneseki and Kadota (25) as coated vesicles. Several formed complex vesicles are unattached to the synaptic membrane (Fig. 4C, F, G, H, small arrows). Attachment of the vesicles to membrane outside of the specific synaptic contact area supports the view of Westrum (49) and Heuser and Reese (22) that vesicles may be formed from membrane of the presynaptic surface.. The extent of outpouching of the membrane which forms this en passant synapse is evident in Fig. 4A-I. Much of the presynaptic membrane appears to engulf the dendritic spine, but only a small region contributes to the functional synaptic contact, as defined by dense presynaptic membrane differentiation. This functional membrane is indicated in Fig. 4E for the chemical synapse (cs) and electrotonic synapse (arrow). The electrotonic synapse is shown in Fig. 6C at higher power. Figure 5A-C are stereo pairs of the computer reconstruction of the synapse of Fig. 4. This computer synthesis illustrates the large amount of membrane which is used to form the presynaptic outpouching. Relatively little of this membrane appears to participate in the functional contact as judged by the region of presynaptic membrane differentiation (cf. Fig. 4E). For example, the electrotonic synaptic contact occurs only over four or five serial sections. In this example, if the sections are assumed to be about 0.08 pm thick, then the electrotonic synapse occupies a patch of membrane about 0.2 q2. Three additional electrotonic contacts made by this giant axon occupied about the same area. Figure 6A shows the section of Fig-. 4E at higher power and Fig. 623, from a different contact, shows an electrical and chemical synaptic contact (es and cs) made on the smooth shaft of the dendrite. These two micrographs show another feature of the synaptic contacts, the synaptic cisternae, which are well developed in Fig. 6B (arrows). Coated vesicles can be seen attached to the end of the cisterna of Fig. 6B. This is consistent with the idea that vesicles may coalesce to form synaptic cisternae as described by Heuser and Reese (22) at the frog neuromuscular junction. Figure 6C is a highpower micrograph of the section which lies be-


SYNAPTIC

TRANSMISSION

IN SPINAL

CORD

203

FIG. 6. A: higher power micrograph of section shown in Fig. 4E. Both chemical and electrotonic synapses are made on the bulbous ends of a dendritic spine. B: a different synaptic contact made on dendrite (1) but located on the smooth shaft of the dendrite. C: high-power microgmph showing details of electronic junction and complex vesicles. See text for details. Scale in B is 0.5 m for A and B, 0.12 m for C.

section E and F of Fig. 4 and showsthe structure of the electrotonic junction (between the arrows) as well as the structure of the complex vesicles (asterisks). The preservation of the membranesforming the electrotonic junction shown in Fig. 6C is evidence that the dye injection does not disrupt intercellular contacts. The unequivocal presenceof synaptic cisternae in several. synaptic contacts made by the stimulated axon contrasts sharply with the contacts made on dendrites by the unstimulated tween

ventral giant axon. This sameobservation was made in other preparations where giant axons were stimulated by current pulsesapplied to the whole spinal cord with platinum-wire electrodes. In these experiments synapsesmade by giant axons on giant interneurons as well as unidentified postsynaptic elementswere examined for morphological changes attributed to the stimulation procedure. According to the hypothesis of Heuser and Reese (22), these cisternae are formed by coated vesicles coalescing


204

B. N.

CHRISTENSEN


SYNAPTIC

TRANSMISSION

within the presynaptic terminal. Another possibility, however, is that cisternae are formed directly from the axoplasmic membrane. In stimulated synaptic junctions, fine processes of axoplasmic membrane extend into the presynaptic terminal and often have coated synaptic vesicles attached to them. Figure 7A, B shows an example of these axoplasmic invaginations from an experiment in which only giant interneurons were injected with Procion brown. In this experiment, the rostra1 end of the isolated spinal cord was stimulated at 2/s for one-half hour and the tissue immediately fixed. Threshold stimulus strength was determined for several of the giant axons and during the stimulation period, the intensity was increased to 5x threshold. Postsynaptic activity was recorded from one giant interneuron during the stimulation period. The contacts in Fig. 7A, B are made on the dendritic process from one interneuron injected with Procion brown. Sites of synaptic contact are indicated by asterisks in the postsynaptic dendrite. In Fig. 7A, B, coated vesicles can be seen attached to the axoplasmic invaginations (large arrows). It is possible that these invaginations pinch off and form the cisternae. In Fig. 7B a cisterna (open arrow) with no attached vesicles is shown to be entirely separated from the axoplasmic membrane. It must either have been formed within the terminal, or (as suggested by the axoplasmic invaginations) pinched off from the axoplasmic membrane as an already-formed cisterna. Axoplasmic invaginations are sometimes seen to surround what appears to be either a dendrite or dendritic spine pushing up into the axon. Figure 7C, D, from the same preparation, shows a synaptic area containing several vesicles. This junction involves neither an injected pre- or postsynaptic element. The arrows indicate in Fig. 7C a band of axoplasmic membrane invaginating the axon and surrounding what may be a dendrite or dendritic spine (asterisk). In Fig. 70, two coated vesicles can be seen (arrows) attached to this membrane. This same configuration was seen in one of the 22 synaptic contacts from the preparation in which the complete reconstruction was done. Figure 8A-E shows this complex arrangement between the giant axon and an injected dendrite from the interneuron and what may be a dendrite or dendritic spine from another uninjetted cell. In Fig. 8A the arrow indicates a vesicle attached to membrane within the axon ter-

IN

SPINAL

CORD

205

minal. This membrane is not continuous with the axoplasmic membrane in this section, but may have been in a missing adjacent section. The configuration thus may resemble that illustrated in Fig. 7C, D. In Fig. 8B this membrane has formed an enclosed ring with a branch directed toward a portion of the injected dendrite. Also in these sections, several vesicles can be seen attached to the presynaptic membrane (arrows). This is probably a double contact made between the giant axon and the giant interneuron and an uninjected process from another neuron. Figure 8B shows what appears to be synaptic membrane differentiation with synaptic vesicles clustered near the membrane of the uninjected process. Figure 8C-E shows three more adjacent serial sections through this double contact. Vesicles are attached (arrows) at both synaptic junctions. The injection of dye into the postsynaptic neuron does not appear to alter the morphology of the presynaptic element. Synaptic contacts on uninjected postsynaptic neurons were indistinguishable from contacts made on the injected giant cell. Figure 9A, B, C, compares the morphology of several synaptic contacts. Figure 9A is a contact made between the injected giant axon and the interneuron, and 9B is between the ventral giant axon and the inter-neuron. The contact in Fig. 9A shows an electrotonic synapse (es) at this level as well as several coated synaptic vesicles (arrows). At a different level through the synapse, membrane differentiation indicating a chemical contact was seen, which is consistent with the presence of the synaptic vesicles. Figure 9B shows the one contact of the many examined between the unstimulated ventral axon and the interneuron which shows some resemblance to the contacts of the stimulated axon. Coated vesicles (arrow) and some out-

pouching of the presynaptic membrane can be seen here. This contact should be compared with others made between the ventral axon and the interneuron shown in Fig. 9C (SC). These contacts show a regular geometrical arrangement of the synaptic vesicles next to the presynaptic membrane,larger numbersof vesicles, and much less outpouching of presynaptic membrane. The contact made between the stimulated axon and interneuron in Fig. 9C (filled arrow) showed almost complete disorganization. A few synaptic vesicles (small arrows) were present, but for the most part the


B. N. CHRISTENSEN

206

FIG.

8. A-E:

serial sections through a double contact. See text for details. Scale 0.25 pm.

internal structure of the synapse consists of what appears to be disorganized membrane. It is difficult to distinguishthe boundary (open arrows) between the synapse of the giant axon and of an adjacent terminal. The differentiation is based on the observation that the adjacent

terminal contains flat vesicles. Examination of all 22 synapses indicated the vesicles to be spherical, suggesting that they might contain excitatory transmitter (48). This is confirmed by the fact that stimulation of this axon produced an EPSP in the interneuron.


SYNAPTIC

TRANSMISSION

IN SPINAL

CORD

201

I

: , \ .~ ‘, j( , :.,;,“*:, _ _) I’, \ :\ ___ “” ,“,:~ ^. ,.‘“*

\. _,

..I.

Lo,,,

4: * I

.I

‘. ‘_

*

FIG. 9. A-C: morphology of synaptic contacts between dendrite and injected dorsal and uninjected ventral axon. See text for details. Scale 0.5 pm.


208

B. N. CHRISTENSEN

ELECI’ROPHYSIOLOGY The morphological data obtained in this study permit one to compare the total number of vesicles in all synaptic contacts with the size of the transmitter pool available for release. If the quanta1 hypothesis is applicable at synapses in lamprey spinal cord and if the quanta1 package is a synaptic vesicle, l then the number of vesicles available for release, n, is given by: n = m/p

(1)

where m is the mean quantum content (i.e., the number of vesicles releasing their contents per presynaptic action potential) and p is the average probability of release. In this experiment, m was not directly determined for this particular giant axon-interneuron synapse but an average value of 50 was used. This value was determined from a series of experiments designed to estimate the mean quantum content at synaptic junctions in lamprey spinal cord (unpublished observations). The EPSPs in Fig. 10 were used to estimate the probability of release. The presynaptic axon was stimulated by short hyperpolarizing pulses applied to the intracellular electrode. The response in Fig. 10 was produced by a series of four stimuli with an interval of 1 s between each stimulus. It is assumed that the amount of reduction in EPSP amplitudes produced by a series of stimuli to the presynaptic axon is due to depletion of transmitter (47). If 100% of the transmitter is available before the first stimulus, then the second response is reduced by the fraction of transmitter released during the first response. For the example shown in Fig. 10, the amplitude of the first response is 4.5 mV and the second 3.9 mV. During the l-s interval between stimuli, some recovery of the depleted transmitter presumably takes place. Here, a recovery of 5% is assumed (14); in other words, the second response is assumed to be 5% larger than it would have been had no recovery occurred. Thus the second response, in the absence of recovery, would have been 3.7 mV. The probability of release is then the fractional amount released by the first stimulus which is the percent decrease of the second response compared to the first (3, 14). In this case, the fraction of the pool released by the first response is 18% and, hence, p is 0.18. It has been assumed here for simplicity that p does not change with n as suggested by Christensen and Martin (14). l It is generally assumed, as a corollary of the quanta1 release hypothesis, that a single synaptic vesicle represents one quantum. This assumption is based more on its simplicity than on any direct evidence.

5 mV 1Oms

FIG. 10. Synaptic response to train of four stimuli. Note depression of chemical response but no change in electrical response (arrow).

Using equation I the total releasable pool can be estimated and, in this case, is about 300 vesicles. Table 1 lists the vesicles counted per synaptic contact for all contacts. The total of about 5,000 probably underestimates the number of vesicles present in the resting terminal because transmitter is released by stimulation during the injection procedure. In any case, there remains at least an order of magnitude difference between the releasable transmitter pool and total transmitter store. This may be compared with results from the neuromuscular junction, where about 1,000 vesicles make up the releasable pool from a total pool size which ranges between 200,000 and 700,000 (31). DISCUSSION Several morphological features characterize synapses made by stimulated giant axons on interneurons in lamprey spinal cord. These include a low number of synaptic vesicles per synapse and the formation of complex vesicles and synaptic cisternae. Furthermore, there appears to be a large amount of membrane outpouching in the vicinity of the stimulated synaptic contacts. These features are not usually seen in uns timulated giant axons. There are several possible explanations for these observations. They may simply represent typical characteristics of the synaptic contacts made between this giant axon, and be unrelated to activity. The ideal control for this study would have been to compare the same synaptic contacts before and after stimulation. Because it is impossible in practice to study dynamic changes with electron microscopy, the comparison actually made was between a stimulated axon and a different unstimulated axon in the same preparation. In addition, synaptic contacts made by giant axons from other, unstimulated preparations were examined as controls. The morphology of unstimulated synaptic junctions described here corroborates the earlier descriptions of Smith et al. (46). Furthermore, in the experiments where all synaptic contacts


SYNAPTIC

TRANSMISSION

1. Vesicles counted per synaptic contact for all contacts

TABLE

No. I

Distance, w 135

Synapse

A B C D E F G H I II 110 L M N 0 P v R S T III 150 U V W X Total 22 Avg 241 vesicles per contact

No. of Vesicles 168 126 638 327 231 299 241 76 375 440 422 355 145 70 219 320 loo 120 75 100 160 300 5,307

Electrical Contact

X X X

X

The table lists dendrite number, the distance from the closest synaptic contact to cell body, the number of synaptic contacts made on the dendrite, the number of synaptic vesicles counted in each contact, and if an electrotonic junction was established in conjunction with the chemical contact. were examined each one showed the morphological characteristics attributed to synaptic activity. It seems highly likely, therefore, that these features are associated with active synapses. The unusual synaptic morphology of the stimulated axon could represent an artifactual response to injection of Procion brown into the presynaptic axon In this experiment the dye was not allowed to diffuse from the injection site to the synaptic contacts. This was confirmed in both light and electron micrographs. It is possible that there was an unsuspected effect at the synapses due to some highly diffusable but invisible byproduct of the dye. This seems unlikely since identical ultrastructural changes are observed in presynaptic axons stimulated by large electrodes applied to the whole spinal cord, but which have not been injected with Procion brown (cf. Fig. 7). The likeliest possibility is that the morphological features characterizing the synapses made by the giant axon represent structural changes due to activity (36, 37).

IN SPINAL

CORD

209

The morphological analysis of serial sections through synaptic contacts in lamprey spinal cord are consistent with the concept of vesicle recycling as described by Heuser and Reese (22). In fact, all of the morphological correlates of active synapses were observed and it is concluded that the process of exocytosis occurs at these identified synapses in much the same way as at the neuromuscular junction. It is further suggested that synaptic cisternae may be formed by invaginations of axoplasmic membrane as well as by the coalescence of coated vesicles. Another aspect of the contact sites on different dendrites is the observation that the same presynaptic axon makes contacts on spines and directly on the smooth shaft of the dendrites from the same cell (cf. Fig. 6A-E). This is of interest in view of the proposed function of the spine as an area of electrical isolation from the parent dendrite, as hypothesized by Diamond (16) or as a structure which might influence the final output by the electrical resistance of its stem, as suggested by Rall (39). One might expect the same presynaptic input to make all contacts either on spines or directly on the dendrites. The coexistence of electrical and chemical contacts as described by Rovainen (43) has been confirmed morphologically and physiologically in this study. Rovainen describes this dual mechanism for the contact between a ventral giant axon arising from a Muller cell in the brain stem and a giant interneuron. The present study demonstrates that similar interaction occurs between a dorsolateral giant axon and the giant interneuron. Recently, Ffenninger and Rovainen (35) have shown evidence from freeze-cleave preparations of lamprey spinal cord that these two types of contacts are located together at many of the same en passant contacts. Rall (39) and Jack and Redman (24) have described methods for assessing electrotonic conduction in dendritic trees and for estimating distances from soma to synaptic input location on dendrites from membrane potential changes produced at the soma. The present results indicate a maximum contact-soma distance difference of 110 versus 150 pm between contacts made on dendrite 2 and those made on dendrite 3 of the cell. It might be expected that this differential would result in a time dispersion of the synaptic effect at the soma. This would be reflected by distinguishable peaks on the synaptic response recorded in the soma. However, there is no indication from the electrophysiology that the response is produced by activity in contacts at different distances along the dendrites. Two peaks are seen; the first presuma-


210

B. N. CHRISTENSEN

bly corresponds to transmission across electrotonic junctions and the second to transmission across chemical junctions (cf. Fig. 2). It is possible that the synaptic contacts are so close to the cell body that synaptic action at the different sites cannot be resolved in the electrical record. Also, the dendrites may compensate for the differences in distance by differences in their cable properties. This interpretation seems unnecessarily cqmplex since inputs to the interneuron, if they occurred on different dendrites, would have to be geometrically arranged so that the distance effect could be adequately compensated. A third possibility, also unlikely, is that synaptic contacts on only one dendrite are active at a given time. It seems most probable that all 22 contacts are made within the same electrotonic distance from the soma. Dendrites from interneurons have been seen to extend to the contralateral spinal cord by Rovainen (44) and in this laboratory. It is possible that contacts on these longer dendrites may be several space constants from the cell body. The effects of a single presynaptic input could be investigated using this method and results compared with the models described by Rall (39) and Jack and Redman (24). It is estimated that 6% of the total transmitter pool (n) is available for release. This estimate is somewhat larger than estimates at the neuromuscular junction but much smaller than the 20% estimated at autonomic ganglia (5). The observation that so few synaptic vesicles occurred at synaptic junctions raised the question whether transmitter might be released from the cytosol but still in a quanta1 fashion, as suggested by Ginsborg (20) and Birks (4). The estimate of n in lamprey gives only an approximate value for several reasons. First, no correction for counting a vesicle twice was made. Second, the stimulation and injection procedure produced an apparent reduction in the vesicle population of the resting synapse, and for this reason the figure of 6% representing the releasable fraction of the total pool is probably too large. The findings show only that the pool size is adequate to account for the amount of transmitter released. The resting pool is probably larger than that calculated here, and there is no evidence that it is inadequate. The possibility that transmitter release occurs from the cytosol still awaits evidence for release in the absence of morphologically identified vesicles e SUMMARY

The dye Procion brown was used to identify in the light and electron microscope, synaptic contacts made between monosynaptically coupled neurons in the lamprey spinal cord whose synaptic interaction had been recorded.

Synaptic contacts were made on different dendrites of the postsynaptic cell at different distances from the soma. Some of the contacts were made on dendritic spines and some on the smooth shaft of the dendrites. Serial sections through synaptic contacts made on dendritic processes of the postsynaptic cells were used for three-dimensional reconstruction of the synapses using computer graphics techniques. The computer reconstructions and detailed examination of the serial EM micrographs revealed the large proliferation of membrane involved in making these en passant synapses as well as the morphological changes due to stimulation of the presynaptic axon. These changes include depletion of synaptic vesicles and formation of complex vesicles and synaptic cisternae. Besides chemical synaptic contacts, four electrotonic contacts were located, confirming the mixed electrochemical synaptic response recorded from the postsynaptic cell. The mean quantum content was estimated and compared with the estimate of the available transmitter pool, assuming the quanta1 release hypothesis applies at these synapses. The total transmitter pool was estimated by counting all synaptic vesicles in all synaptic contacts. It was estimated that about 6% of the total transmitter pool is available for release at these synapses. This compares with less than 1% at the neuromuscular junction and about 20% at sympathetic synapses. These results support the hypothesis that synaptic vesicles may be recycled as described by Heuser and Reese (22) at the neuromuscular junction. Ongoing studies are investigating the effect on a variety of synaptic junctions to stimulation for different periods of time of presynaptic axons. The methods described in this study can also be used to test the models of synaptic interaction on dendritic trees described by Rall (39) and Jack and Redman (24). ACKNOWLEDGMENTS

I thank Dr. F. E. Ebner and Dr. J. T. McIlwain for helpful comments on the manuscript and Dr. T. Tobias for many helpful discussions and assistance during a part of this work. I thank John Miller for computer assistance on some of the three-dimensional reconstructions and C. Critz for photographic assistance. Facilities for three-dimensional reconstruct ion of synaptic contacts from the EM material were kindly made available by Professor Allen Selverston, in the Dept. of Biology at the University of California at San Diego, and by Professor Cyrus Levinthal in the Dept. of Biology at Columbia University, New York. Requests for details of these reconstruction methods should be addressed to these laboratories. This work was supported in part by National Institutes of Health Grant NS 09632 and National Science Foundation Grant BMS 502822.


SYNAPTIC

TRANSMISSION

IN SPINAL

CORD

211

REFERENCES im Zentral1. ANDRES, K. H. Mikropinozytose nervensystem. Histochemie 64: 63-73, 1964. K. Fluores2. BARRETT, J. N. AND GRAUBARD, cent staining of cat motoneurons in vivo with beveled micropipettes. Brain .Res. 18: 556-568, 3.

cervical 20.

1970. BETZ,

W. J. Depression of transmitter release at the neuromuscular junction of the frog. J.

ganglion

of the cat. Experientia

27:

285-287, 1971. GINSBORG, B.

L. The vesicle hypothesis for the release of acetylcholine. In: Excitatory Synaptic Mechanisms, edited by P. Anderson and J. K. S. Jansen. Oslo: Universitetsforlaget, 1970, p.

21.

77-82. GRAY,

Physiol., JACK,

4.

Physiol., BIRKS,

6.

Can. J. Biochem. Physiol. 39: 787-827, 1961. BLACKMAN, J. G., GINSBORG, B. L., AND RAY,

E. G. AND WILLIS, R. A. On synaptic vesicles, complex vesicles, and dense projections. Brain Res. 24: 149-168, 1970. 22. HEUSER, J. E. AND REESE, T. S. Evidence for recycling of synaptic vesicle membrane during transmitter release at the frog neuromuscular 1973. junction. J. Cell Biol. 57: 315-344, 23. HUBBARD, J. I. AND KWANBUNBUMPEN, S. Evidence for the vesicle hypothesis. J.

C. Synaptic transmission in the sympathetic ganglion of the frog. J. Physiol-, London 167:

24.

London

206:

629-644,

1970.

R. I. The relationship of transmitter release and storage to fine structure in a sympathetic ganglion. J. Neurocytol. 3: 133-160, 1974. 5. BIRKS, R. I. AND MACINTOSH, F. C. Acetylcholine metabolism of a sympathetic ganglion.

7.

8.

9.

10.

11.

12.

13.

14.

15.

355-373, 1963. BLACKMAN, J.

G., GINSBORG, B. L., AND RAY, C. On the quanta1 release of the transmitter at a sympathetic synapse, J. Physiol., London 167:

25.

402-414, BOYD,

26.

London BROWN,

A. R. The end-plate muscle. J. Physiol.,

132: 74-91, 1956. K. T. AND FLAMING,

D. G. Beveling of fine micropipette electrodes by a rapid precision method. Science 185: 693-695, 1974. CAMPBELL, R. D. AND HERMANS, C. 0. A rapid method for resectioning 0.5-4 rnp epoxy sections for electron microscopy. Stain Technol. 47: 115-118, 1972. CECCARELLI, B. AND HURLBUT, W. P. The effects of prolonged repetitive stimulation in hemicholinium on the frog neuromuscular junction. J. Physiol., London 247: 163-188, 1975. CECCARELLI, B., HURLBUT, W. P., AND MAURO, A. Depletion of vesicles from frog neuromuscular junctions by prolonged tetanic stimulation. J. Cell Biol. 54: 30-38, 1972. CHRISTENSEN, B. N. Procion brown: an intracellular dye for light and electron microscopy. Science 182: CHIUSTENSEN,

1255-1256,

Physiol., London DEL CASTILLO,

210:

933-945,

29.

J. Physiol., LEVINTHAL,

16.

London 124: 560-574, DIAMOND, J., GRAY,

19.

505-529, FRIESEN,

postsynaptic potentials in spinal motoneurones.

1954.

E. G., AND YASARGIL, G. M. The function of the dendritic spine: an hypothesis. J. Physiol., London 202: 116P, 1969. 17. DUDEL, J. AND KUFFLER, S. W. The quanta1 nature of transmission and spontaneous miniature potentials at the crayfish neuromuscular junction. J. Physiol., London 155: 514-529, 1961. 18. ELMQUIST, D. AND QUASTEL, D. M. J. A quantitative study of end-plate potentials in isolated human muscle. J. Physiol., London 178: 1965.

A. J. D. AND KHATTER, J. C. Effect of stimulation of svnaptic vesicles in the superior

London 175: 81-99, 1964. C., MACAGNO, E., AND TOUNTAS,

C. Computer-aided reconstruction from serial sections. Federation Proc. 33: 2336-2340, 1974. A. W. The quanta1 components of the 30. LILEY, mammalian end-plate potential. J. Physiol., 31. 32.

33.

1970.

J. AND KATZ, B. Quanta1 components of the end-plate potential. J. Physiol.,

1972.

28.

27.

1974.

B. N. AND MARTIN, A. R. Estimates of probability of transmitter release at the mammalian neuromuscular junction. ,J.

194: 407-420,

Physiol., London 215: 321-352, 1971. KANASEKI, T. AND KADQTA, K. The “‘vesicle in a basket.” J. Cell Biol. 42: 202-220, 1969. KORNELIUSSEN, H. Ultrastructure of normal and stimulated motor endplates. Histochemie 130: 28-57, 1972. KRIPKE, B. R. AND OGDEN, T. E. A technique for beveling fine micropipettes. Electroencephalog. Clin. Neurophysiol. 36: 323-326, 1974. KUNO, M. Q uan ta 1 components of excitatory

1963.

I. A. AND MARTIN, potential in mammalian

London

J. J. B. AND REDMAN, S. J. An electrical description of the motoneurone, and its application to the analysis of synaptic potentials. J.

London MARTIN,

133: 571-587,

1956.

A. R. Quanta1 nature of synaptic transmission. Physiol. Rev. 46: 51-66, 1965. MARTIN, A. R. AND PILAR, G. Quanta1 components of the synaptic potential in the ciliary ganglion of the chick. J. Physiol., London 175: 1-16, 1964. MILEDI, R. Spontaneous synaptic potentials and quanta1 release of transmitter in the stellate ganglion of the squid. J. Physiol., London 192: 379-406, PARDUCZ,

1967.

A. AND FEHER, 0. Fine structural alterations of presynaptic endings in the superior cervical ganglion of the cat af&r exhausting preganglionic stimulation. Experiktia 26: 629-630, 1970. 35. PFENNINGER, K. H. AND ROVAINEN, C. M. Stimulation- and calcium dependence of vesicle attachment rates in the presynaptic membrane; a freeze-cleave study on the lamprey spinal cord. Brain Res. 72: l-23, 1974. 36. PYSH, J. J. AND WILEY, R. G. Morphologic alteration of synapses in electrically stimulated superior cervical ganglia of the cat. Science 176: 191-199, 1972. R. G. Synaptic vesi37. PYSH, J. J. AND WILEY, cle depletion and recovery in cat sympathetic 34.


212

B. N.

ganglionelectrically

CHRISTENSEN

stimulated in vivo. J. CeZI

43.

Biol.

38.

39.

40.

41.

42.

60: 365-374, 1974. QUILLIAM, J. P. AND TAMARJND,

D. L. Some effects of preganglionic nerve stimulation of synaptic vesicle populations in the rat cervical ganglion. J. Physiol., London 235: 3 17-332, 1973. RALL, W. Cable properties of dendrites and effects of synaptic location. In: Excitatory Synaptic Mechanisms, edited by P. Anderson and J. K. S. Jansen. Oslo: Universitetsforlaget, 1970, p. 175-l 88. ROVAINEN, C. M. Physiological and anatomical studies on large neurons of central nervous system of the sea lamprey (Petromyzon marinus). I. Muller and Mauthner cells. J. Neurophysiol. 30: 1000-1023, 1967. ROVAINEN, C. M. Physiological and anatomical studies on large neurons of central nervous system of the sea lamprey (Petromyzon maeinus). II. Dorsal cells and giant interneurons. J. Neurophysiol. 30: 1024-1042, 1967. ROVAINEN, C. M. Synaptic interactions of reticulospinal neurons and nerve cells in the spinal cord of the sea lamprey. J. Comp. Neurol. 154: 207-224, 1974.

44.

45.

46.

47.

48* 49.

ROVAINEN, C. M, Synaptic interactions of identified nerve cells in the spinal cord of the sea lamprey. J. Camp* Neurol. 154: 189-206, 1974. ROVAINEN, C. M., JOHNSON, E. A., ROACH, E. A., AND MANKOVSKY, J. A. Projections of individual axons in lamprey spinal cord determined by tracings through serial sections. J. Comp. Neural. 149: 193-201, 1973. SHERMAN, R. G. AND ATWOOD, H. L. Correlated electrophysiological and ultrastructural studies of a crustacean motor unit. J. Gen. Physiol. 59: 586-615, 1972. SMITH, D. S., JARLFORS, U., AND BERANEK, R. Organization of synaptic axoplasm in the lamprey central nervous system. J. CeZl Biol. 46: 199-219, 1970. TAKEUCHI, A. On the long lasting depression in neuromuscular transmission of frog. Japan. J. Physiol. 8: 102-113, 1958. UCHIZONO, K. Characteristics of excitatory and inhibitory synapses in central nervous system of cat. Nature 207: 642-643. 1965. WESTRUM, L. E. On the origin of synaptic vesicles in cerebral cortex. J. Physiol., London 179: 4-6P, 1965.


CNQX and DNQX block non-NMDA synaptic transmission but not NMDA-evoked locomotion in lamprey spinal cord.

Effects of ketamine on synaptic transmission in cat spinal cord.

Axonal regeneration in the adult lamprey spinal cord.

Estimates of cable parameters in lamprey spinal cord neurones.

Subcellular distribution of serotonin in the lamprey spinal cord.

Effects of guanidine on synaptic transmission in the spinal cord of the frog.

Primary receptor for inhibitory transmitters in lamprey spinal cord neurons.

Hypoxic failure of synaptic transmission in the isolated spinal cord, and the effects of divalent cations.

Hydrogen peroxide modulates synaptic transmission in ventral horn neurons of the rat spinal cord.

Effects of cyclobenzaprine on spinal synaptic transmission.

Localization of synaptic input on dendrites of a lamprey spinal cord neurone from physiological measurements of membrane properties.

Enhancement in Ia-motoneuron synaptic transmission caudal to chronic spinal cord transection.

Synaptic transmission between rat spinal cord explants and dissociated superior cervical ganglion neurons in tissue culture.

Immediate increase in Ia-motoneuron synaptic transmission caudal to spinal cord transection.

Preferential regeneration of spinal axons through the scar in hemisected lamprey spinal cord.

Interferon alpha inhibits spinal cord synaptic and nociceptive transmission via neuronal-glial interactions.

Serotonergic transmission after spinal cord injury.

GABAergic control of rhythmic activity in the presence of strychnine in the lamprey spinal cord.

Localization and electrical characteristics of a giant synapse in the spinal cord of the lamprey.

VEGF modulates synaptic activity in the developing spinal cord.

The effect of centrally acting myorelaxants on NMDA receptor-mediated synaptic transmission in the immature rat spinal cord in vitro.

Laterally projecting cerebrospinal fluid-contacting cells in the lamprey spinal cord are of two distinct types.

Synaptic modulation of excitatory synaptic transmission by nicotinic acetylcholine receptors in spinal ventral horn neurons.

Comparative study of the monoamine neuron system in the spinal cord of the lamprey and hagfish.