* Your assessment is very important for improving the workof artificial intelligence, which forms the content of this project
Download physiological and chemical architecture of a lobster ganglion with
Survey
Document related concepts
Endomembrane system wikipedia , lookup
Chemical synapse wikipedia , lookup
Extracellular matrix wikipedia , lookup
Tissue engineering wikipedia , lookup
Cytokinesis wikipedia , lookup
Cell encapsulation wikipedia , lookup
Programmed cell death wikipedia , lookup
Cell growth wikipedia , lookup
Cellular differentiation wikipedia , lookup
Cell culture wikipedia , lookup
Transcript
PHYSIOLOGICAL OF A LOBSTER REFERENCE AND CHEMICAL ARCHITECTURE GANGLION WITH PARTICULAR TO GAMMA-AMINOBUTYRATE AND GLUTAMATE1 M. OTSUKA,2 E. A. KRAVITZ, Department of Neurobiology, Harvard Medical (Received for publication AND D. D. POTTER School, Boston, Massachusetts November 18, 1966) IS LITTLE DOUBT that gamma-aminobutyric acid (GABA) is the inhibitory transmitter compound at the lobster neuromuscular junction (see 20). Special interest, therefore, is attached to the chemistry of this compound. GABA is about 100 times more concentrated in inhibitory than fibers in excitatory axons; the concentration is about 0.1 M in inhibitory if GABA is freely dissolved in the axoplasm (13,15). The origin of the GABA difference appears to lie in an asymmetric distribution of the enzyme that synthesizes GABA from glutamate, glutamic decarboxylase, the pathway for subsequent degradation being about equally active in the two types of axons (14). The mechanisms regulating the levels of the enzymes, on the other hand, are unknown; it is possible that differential enzyme synthesis or destruction accounts for the decarboxylase difference. Since the neuronal cell body is presumably the site of enzyme synthesis, the aim of the present work was to develop methods for finding and isolating the cell bodies of efferent excitatory and inhibitory neurons within the lobster central nervous system. As a first step toward understanding enzyme regulation in the GABA pathway, the contents of the substrates, GABA and glutamate, were measured. During the course of this work observations were made on the physiological organization of a ganglion, and these are also presented. The present study, like earlier ones on peripheral axons, was aided by the favorable anatomy of the lobster. The total number of efferent neurons is small, usually between two and six per muscle, at least one of which is inhibitory (cf. 12, 31); this eases the problem of locating any particular cell body. Like the axons, the cell bodies are large enough to be isolated with hand-held instruments. In contrast to the cell bodies of most vertebrate neurons, they are apparently free of synapses (1); thus interpretation of the chemical analysis is simplified. Finally, the ganglia, like the peripheral nerves, show a striking constancy of organization; many of the efferent cell bodies can be recognized by their size and position alone. A brief abstract of this work has been published (21). THERE by Public Health Service Grants 5 K3 NB 7833-05, 6 K3 and NB 02253-07. Fellow. Present address: Dept. of Pharmacology, Faculty and Dental University, Tokyo, Japan. Downloaded from http://jn.physiology.org/ by 10.220.33.5 on June 18, 2017 1 This work was supported 5899-OIAI, NB 02X3-06, 2 Rockefeller Foundation of Medicine, Tokyo Medical HD 726 OTSUKA, KRAVITZ, AND POTTER Animals. Lobsters (Homarus americanus) were obtained from local dealers and kept in moist seaweed in a cold room at 3 C, sometimes for several days. They weighed 0.5-1.5 kg. Preparation. A portion of the abdomen containing the first through the fourth ganglia was isolated (Fig. 1A). If the extensor muscles (cf. Fig. 12B) were used, the main branch of the second ganglionic root was exposed byremoving the overlying shell, skin, and muscle. The extensor muscles were isolated, remaining connected to the rest of the abdomen by to permit the extensor muscles to be of the abdomen is placed ventral side up the nerve bundle. The nerve is sufficiently long placed with their deep face exposed when the rest (Fig. 1A). If the extensor muscles were not needed, they were cut away with the shell to which they are attached. Further dissection was done in oxygenated physiological solution (hereafter called saline, see below for composition), with the temperature maintained between 7 and 14 C by thermoelectric cooling units upon which the chamber rested. The ventral surface of the ganglionic chain was exposed by removing midline portions of the overlying ribs, the joint membranes, and the underlying pigmented skin (Fig. 1A). To permit transmitted light to reach the ganglia, the flexor musculature dorsal to the ganglionic chain was divided in the midline, care being taken to minimize damage to the third roots. The only Downloaded from http://jn.physiology.org/ by 10.220.33.5 on June 18, 2017 FIG. 1. A: an abdominal segment of the lobster, viewed from the ventral side. The ganglionic chain was exposed by removing ribs and skin in the midline. Flexor muscles were divided at the midline. At the left, the extensor muscles are attached to the preparation by the nerve supply of one segment (the main branch of the second root). B : ventral surface of the nerve cord, showing the second abdominal ganglion. The cell bodies were exposed by removing the connective tissue sheath and perineural tissue. The cell bodies are shown at higher magnification in Fig. 3. Rl, R2, and R3; first, second, and third ganglionic roots. GABA AND GLUTAMATE IN CELL BODIES The ventral surface, ganglion consists of with toluidine blue. bodies were exposed 12 and 13. structure joining the two sides was now the ganglionic chain with its roots. The two halves were fixed with forceps or pins so that slight tension was placed on the roots, holding the ganglia taut (see Fig. 1, A and B). Light was reflected obliquely from below to give a dark-field effect. To expose the cell bodies, the outer connective tissue sheath (Fig. 2A) was cut along a lateral margin, reflected to one side, and removed. The mass of soft perineural tissue (Fig. 2A) was washed away with a gentle stream of saline from a fine-tipped pipette (see Fig. 2B). The washing process was stopped as soon as the cells were clearly recognized. Under the dissecting microscope the ganglion now had the appearance shown in Figs. 1B and 3. Cold, oxygenated saline was usually directed into the neighborhood of the ganglion throughout the experiment. An outlet attached to a suction line maintained a fluid level barely covering the tissues. In some experiments the bath was oxygenated directly without perfusion. The methods for identifying the cell bodies of efferent excitatory and inhibitory neurons are described in RESULTS section. Downloaded from http://jn.physiology.org/ by 10.220.33.5 on June 18, 2017 FIG. 2. A: transverse section of the third abdominal ganglion. on which the cell bodies lie, is uppermost. The main mass of the neuropil in which synapses are found. The ganglion was stained B : transverse section of a second abdominal ganglion in which the cell before fixation. The highly vacuolated cytoplasm is characteristic of 727 728 OTSUKA, KRAVITZ, AND POTTER Isolation and transfer procedures. Two procedures for isolating the cells were used. In method 1, the cell bodies were isolated by free-hand dissection using forceps under a dissecting microscope at a magnification of 40 X. The sharpened points of stainless forceps were polished with rouge to reduce the tendency of the isolated cells to stick to the forceps. The cell body was first separated from its neighbors by softening and separating the surrounding tissue with a fine stream of saline or with forceps. When the cell body had been freed, the axon was grasped with forceps as far from the cell body as possible and severed (Fig. 4A). Remaining tissue was removed from the isolated cell as thoroughly as possible to reduce the chance of contamination by small cell bodies that lie beneath the large cells (Fig. 2). A histological section of such an isolated cell body is shown in Fig. 4C; some contaminating tissue remains, but most of the sample consists of the nerve cell body. under the abdominal dissecting ganglion, microscope. showing the appearance of the living This is the same ganglion as Fig. cell bodies 14B. Resting potentials of isolated cells were usually between 15 and 40 mv, with none higher than 45 mv; in most cases, the resting potential was not measured. The isolated cell was then drawn into a pipette with a constricted neck (Fig. 4B) and transferred to a 0.2-ml conical tube with about 1 ~1 of saline. Two microliters of 0.2 N HCl were added to extract GABA and glutamate and to denature eneymes. The sample was then stored at - 20 C until analyzed. Isolation and extraction usually took less than 4 mm. Method 2 for isolating the cell bodies resulted in more consistent chemical measurements. The whole ganglion was quickly removed and placed in isopentane kept near its freezing point with liquid nitrogen. The preparation was dried in vacua (about 20 ,.L Hg) for a week at -30 C. The dry ganglion was placed in xylene to make it more transparent. The desired cells were removed with forceps and transferred and extracted as above. Xylene was chosen as a clearing agent because GABA and glutamate are virtually insoluble in it. In one series of experiments, cells were physiologically identified before the ganglia were frozen; neighboring cells were stained by dye injection (see below) so that the identified cells could be recognized in the dried ganglia. In another series of experiments (reported Downloaded from http://jn.physiology.org/ by 10.220.33.5 on June 18, 2017 FIG. 3. The third GABA AND GLUTAMATE IN CELL BODIES Downloaded from http://jn.physiology.org/ by 10.220.33.5 on June 18, 2017 FIG. 4. Isolation and transfer of single cell bodies. A : the cell body at the lower right was separated from its neighbors by free-hand dissection. Its axon was severed, and the cell lies free on the ganglion. B: the cell body has been drawn into a transfer pipette. The constricted neck of the pipette prevents the cell from being lost, and by slowing fluid movement improves control. C: section of an isolated cell body, which was fixed in 1 y0 OsOa, embedded in Epon and stained with toluidine blue. Adhering tissue is present at the upper right and lower left. 730 OTSUKA, KRAVITZ, AND POTTER Downloaded from http://jn.physiology.org/ by 10.220.33.5 on June 18, 2017 in the bottom two columns of Table 3) an attempt was made to minimize disturbance to the ganglia before freezing. Instead of isolating a segment of the abdomen in the usual way, the brain was crushed and the whole animal immobilized. The second abdominal ganglion was exposed and the connective tissue sheath and perineural tissue over the ventral surface removed. Circulating blood was cleared from the ganglion during this operation with a jet of cooled, oxygenated saline. The ganglion was removed and frozen, the operation taking about 10 min in all. Less than 1 min elapsed from cutting the roots and connectives to immersing in the isopentane. The cell bodies were presumably frozen instantly since they were directly exposed to the freezing solution (cf. Fig. ZB). The cells were identified in the dried ganglion by visual inspection. GABA clnd glutamate assays. The microtubes containing the cell bodies were lyophilized. GABA was assayed by a modification of the method of Jakoby and Scott (7). The details of this assay have already been published (15). With large inhibitory cell bodies the sample could be divided, so that a control tube containing aminooxyacetic acid could be run to give tissue blank levels of fluorescence. Glutamate was determined by dividing the sample in half and treating one portion with a bacterial decarboxylase which converts glutamate to GABA. Both portions were then assayed for GABA; the difference between them represents the glutamate in the sample. Glutamine will also react in this assay but independent checks have demonstrated that glutamine levels are low in these cells. The details of the glutamate assay were as follows: The lyophilized sample was dissolved in 5 ,~l of pyridine-hydrochloride buffer, with NaCl (25). Samples of 2 ~1 were transferred to two pH 4.6 adjusted to 0.1 M chloride O.l-ml test tubes. Two microliters of decarboxylase were added to the first tube (containing sufficient enzyme to convert all the glutamate present to GABA within 15 min) and 2 ~1 of heated decarboxylase to the second tube. The tubes were capped with parafilm and incubated for 15 min at room temperature. They were heated at 95 C for 1 min, 1 pl of 1 M Tris, pH 7.9 was added and a standard GABA assay was run, as previously described. The bacterial decarboxylase was prepared from dried Escherichia coli cells purchased from Worthington Biochemical Corp. The enzyme was extracted by homogenization in a French pressure cell, precipitation of contaminating proteins at pH 4.6, fractionation with solid (NH&SO* (O-60 (y(, saturation) and dialysis against 1,000 volumes of 0.01 M potassium phosphate, pH 6.5. This purification procedure is a modification of the method of Shukuya and Schwert (25). Estimation of cell body volume. Only rough measurements of cell volume could be made on the living cell with the dissecting microscope. It was found that the manipulation required to measure the diameter of living cells under the compound microscope (cell transfers, temperature changes, etc.) resulted in reduced GABA contents. The method used, therefore, was to derive an average volume for each cell type by reconstruction from serial sections of fixed, plastic-embedded ganglia. These ganglia were from lobsters of t,he same size and condition as those used for chemical analyses. In two ganglia, certain cells were physiologically identified in the usual way and their positions in the ganglion sketched. The ganglia were then fixed in ice-cold acrolein-glutaraldehyde fixative modified after Sandborn et al. (23) and postfixed in phosphate-buffered 0~04. The ganglia were dehydrated in methanol, stained with toluidine blue, and embedded in Epon 9. They were serially sectioned with a steel knife (J. Alvarez and E. J. Furshpan, unpublished data) at 4 ,u. Shrinkage of the ganglia during these procedures was found to be negligible. The cells could be easily identified by comparing a reconstructed ventral view from the serial sections with the sketch of the living ganglion. The volume of a cell was derived by tracing each section at known magnification on paper, weighing the tracings to determine their area, and multiplying the area by the known thickness of the sections. It was found that the large inhibitory cells could be confidently identified on histological grounds alone; they contained a higher proportion of large vacuoles (5-50 ,U in diameter; see Fig. 2B) than neighboring excitatory cells. Moreover, certain excitatory cells could be identified from their positions in the ganglion. The second abdominal ganglia from five other animals were fixed without physiological manipulation. They were treated as described above, except that they were sectioned at 11 or 13 ,u. The proportion of the cell volume occupied by vacuoles was estimated in 13 cells. PhysiologicaL solution. The solution had the following composition (millimoles) : NaCl, 462; KCl, 15.6; CaC12, 25.9; MgS04, 8.3; Tris(hydroxymethyl)aminomethane, 10.0; GABA AND GLUTAMATE IN CELL BODIES 731 FIG. 5. Electrode used for stimulating the peripheral nerve in the middle of its course. One lead (dotted line) is surrounded by a polyethylene tube and is exposed to the bath only at the bottom of the groove in which the nerve lies. The other lead (solid black line) is wound around the polyethylene t*ube. was exposed to the saline only in the groove. the tube. Stimulating voltages were as small assembly was carried on a ball-joint manipulator When the nerve in the groove was stimulated, centrifugal directions at the same threshold. A second lead was wound on the outside of as those used with suction electrodes. The and could easily be bent into any position. impulses spread in both centripetal and RESULTS PART I Physiological architecture of the ganglion The abdominal ganglia were used because the cell bodies lie nearly in one plane (Figs. 2A, 3), and are among the largest in the lobster nervous system, several being over 200 ,U in diameter. Usually the second and third abdominal ganglia were chosen. Cells of exposed ganglia were impaled under visual control. The resting potentials recorded with our low-resistance microelectrodes were usually 40-70 mv, and with few exceptions the cell bodies showed action potentials of less than 25 mv, suggesting that the membrane of the cell body Downloaded from http://jn.physiology.org/ by 10.220.33.5 on June 18, 2017 maleic acid, 10.0; NaOH, about 10 to bring the pH to 7.5. Just before using, glucose (2 mg/ml) was added. The saline was cooled and oxygenated. Electrodes. Glass micropipettes were used for recording and passing current. They were filled with 3 M KC1 or 2 M K- or Na-citrate and had resistances of 2-20 megohms when measured in 3 M KCl. The electrodes could be connected via a manual switch either to a high impedance preamplifier or to a pulse generator. For marking cell bodies, organic dyes were injected electrophoretically from the microelectrodes. Electrodes with tips about 1 ,U in diameter were filled with dye solutions, about one-quarter saturated, by pushing the dye down the shank with a fine glass filament (Alvarez and Furshpan, unpublished). The dyes used were fluorescein sodium, chromotrope 2R, light green S F yellowish, and fast green FCF. They are listed in order of decreasing rate of diffusion within the cells. It was frequently necessary to stimulate a peripheral nerve in the middle of its course without disturbing its central and peripheral connections. A convenient electrode for doing this was made by enclosing a platinum wire with a flattened end in a polyethylene tube about 4 cm long (Fig. 5). The end of the polyethylene tube was sealed and a groove for supporting t,he nerve and confining the current was made with a hot needle. The wire 732 OTSUKA, KRAVITZ, AND POTTER FIG. 6. Simultaneous intracellular recordings of spontaneous activity in 11 of the second ganglion and in the superficial flexor muscle innervated by this neuron. A and B are different preparations. Arrows: EPSP’s. This figure illustrates the correspondence between impulses in 11 and IJP’s in the opposite superficial flexor muscle. In B, the first IJP was superimposed on the decline of a spontaneous EJP. The time course of the action potential is shorter than in A; such variation was frequently seen in our preparations. In B, 2-mv scale refers to upper trace; 4-mv scale to lower trace. Downloaded from http://jn.physiology.org/ by 10.220.33.5 on June 18, 2017 and its slender process behaved passively under our experimental conditions (cf. 26). At the outset our problem was to find the efferent neurons among all the cells in the ganglion (Fig. 3) and to determine the function of each. Efferent excitatory cells were found relatively easily because a visible muscle contraction usually appeared when they were stimulated intracellularly with a single shock. The possibility that the cell was an interneuron could then be virtually eliminated by showing that antidromic impulses reached the cell body. The first efferent inhibitory cell body was located by trial and error, after a prolonged search. Following the demonstration that this cell was rich in GABA, chemical analysis provided a rapid means for discovering others. Our progress was entirely dependent on the fact that each cell has a characteristic size and position, so that once identified it could readily be found from animal to animal. Identification of efferent inhibitory cell bodies. The first inhibitory cell body identified innervates the superficial flexor muscle illustrated in Fig. 12% muscle 1. Advantage was taken of the fact that this muscle receives a spon GABA AND GLUTAMATE IN CELL BODIES 733 taneous barrage of excitatory and inhibitory impulses, as can be seen in Figs. 6 and 7 (cf. 12). The inhibitory junctional potentials (IJP’s) in the muscle disappeared if the ganglion of the same segment was cut away. Thereupon, in a series of such ganglia, cell bodies were impaled systematically in search of one whose spontaneous activity was related to the IJP’s in the muscle. Eventually such a cell was found on the contralateral side; this cell is shown, labeled 11, in Figs. l2A and 14. Simultaneous recordings in the cell body and the muscle are shown in Fig. 6; an action potential an the cell body invariably preceded the IJP’s in the muscle by about 20 msec. The one-toone relation between the action potential in the cell body and an IJP in the muscle demonstrated that the neuron was at least associated with the inhibitory pathway to the muscle fiber. Two further observations indicated that the neuron lay directly in the inhibitory pathway; subthreshold activity in the cell body (arrows, Fig. 6) were not followed by IJP’s in the muscle, and when the cell body was stimulated with injected current, an IJP appeared in the muscle with about the same delay as with spontaneous firing (Fig. 7). To test whether the neuron was directly connected to the muscle, the nerve leading to the muscle was stimulated in midcourse with gradually increasing intensity. At a sharply defined threshold an IJP appeared in the muscle fiber and an antidromic action potential was observed in the cell body (Fig. 8B); this observation greatly reduced the chances that we were dealing with an interneuron. The action potential in the cell body of this neuron was characteristically less than 10 mv in size; nevertheless, it could be distinguished from excitatory postsynaptic potentials (EPSP’s) by its faster time course and conspicuous afterpositivity (cf. Figs. 6, 9). When the 11 cell bodies (one on either side of the ganglion) were impaled Downloaded from http://jn.physiology.org/ by 10.220.33.5 on June 18, 2017 FIG. 7. Effect of stimulating II on the membrane potential of the superficial flexor muscle innervated by this cell. Short current pulses were delivered to II of the second ganglion through an intracellular electrode. The muscle received a high-frequency barrage of spontaneous EJP’s upon which the evoked IJP’s were superimposed at nearly constant latency (dotted line). with microelectrodes, synchrony of action potentials was often observed (Fig. 9A; see description of electrical coupling, below). The procedures discussed above would, therefore, not have distinguished which of the two cells was attached to a particular muscle. However, as the experiment proceeded, the action potentials did not remain synchronous, an impulse in one cell being accompanied only by an EPSP in the other (Fig. 9AZ). An IJP in the superficial flexor muscle was invariably associated with an impulse in the FIG. 9. The upper part of the figure shows simultaneous recording of spontaneous activity in both 11 cell bodies of the second abdominal ganglion. In Al, early in the experiment, action potentials in the two cells are synchronous. Later in the experiment, A2, the synchrony is no longer perfect. Arrow indicates an EPSP. Part B is from a different preparation, in the same condition as A2. It shows that IJP’s in the superficial flexor muscle are not invariably associated with impulses in the homolateral Il. (B3; compare with Fig. 6.) All records are intracellular. Downloaded from http://jn.physiology.org/ by 10.220.33.5 on June 18, 2017 FIG. 8. Stimulation of the nerve to the superficial flexor muscle while recording intracellularly from the muscle and from the opposite II of the third ganglion. In A, the stimulus intensity was subthreshold. In B, activity appeared in the muscle fiber and neuron at the same stimulus strength. l-mv scale refers to upper trace. GABA AND GLUTAMATE IN CELL BODIES 735 FIG. IO. Stimulating and recording electrodes were inserted into the cell body of 13, while recording from the extensor muscle it innervates, M. dorsalis Prof. In A, the current pulse was just subthreshold for 13; in B just suprathreshold. An IJP appeared in the muscle at the same stimulus strength as the impulse in 13. 0.4-mv scale refers to the upper beam. GABA analysis was then used to locate two more pairs of inhibitory cell bodies, 12’s and 13’s. The characteristic position of these high GABA cells is shown in Figs. 12A and 14. The physiological procedures for demonstrating an inhibitory function were similar to those just described for Il. One-to-one correspondence between the neuronal action potential and the muscle was established for 12 and 13 with stimulation in the cell body and in the The experiment with conduction in the orthodromic direction is illustrated for 13 at a sharply defined threshold an all-or-none action potential was produced body and an IJP in the muscle about 30 msec later. IJP in the periphery. in Fig. IO; in the cell Identification of efferent excitatory cell bodies. The cells that produced visible contraction in the periphery are labeled Ml through Ml4 in Fig. 12A. In the smaller cell bodies (M8’s, Mll) trains of stimuli were occasionally required to produce contraction; in the others a single suprathreshold pulse was sufficient. For six cells, M2, M5, M6, M9, MlO, and M14, a one-to-one correspondence between an antidromic action potential in the cell body and Downloaded from http://jn.physiology.org/ by 10.220.33.5 on June 18, 2017 contralateral cell body, but this correspondence did not hold for the ipsilateral cell body (Fig. 9B). From this consistent finding we conclude that the axons are crossed. The presence of bidirectional one-to-one synapses in crustacean ganglia (described below; see also 27) still leaves some doubt whether the 11’s are efferent cells or interneurons. However, if they are interneurons there should be other cells in the ganglion, the true efferent neurons, with the same behavior; thorough exploration in many preparations never revealed such cells. Moreover, the cells in question consistently have a high GABA content (PART II) not seen in excitatory neurons. From all this evidence we conclude that the 11’s are the efferent inhibitory neurons innervating the superficial flexor muscles. 736 OTSUKA, KRAVITZ, AND POTTER FIG. 11. Stimulation of the third root while recording intracellularly from M9 of the second ganglion and the muscle it innervates, M. obliquus ant. 4. (The experimental arrangement is thus similar to that of Fig. 8.) The stimulus strength was subthreshold in A. Activity appeared in both cell body and muscle at the same threshold (B). The lo-mv scale applies to the upper beam. Functional architecture of the abdominal ganglia. There are three conspicuous features of the ganglionic organization: 1) Each cell has a functional partner with the same size and position on the opposite side of the ganglion; 2) each cell lies in close association with particular other cells, forming a characteristic grouping; and 3) the groups are quite constant in location, but within each group relative positions of cells are not necessarily constant. The positions and certain features of the identified cells are described below. The description applies to the second and third abdominal ganglia; all but the sixth ganglion appear to be similar, but the others have not been studied in detail. No consistent differences were observed between 0.5-kg and 1.5-kg lobsters except that the cells were somewhat smaller in the former. The identified cells are shown in six representative ganglia in Figs. 12, 13, and 14, and the muscles they control are listed in Table 1. We did not attempt to find the entire field of innervation of any neuron. Most cells were traced to a single muscle only, but it is highly likely that the innervations of domen. The numbered muscle 4 it has been V = ventral; D = dorsal. muscles are named in Table cut to reveal 8 and 9 more Gl = first abdominal ganglion. 1. Muscle 7 corresponds clearly. A =anterior; to 10; like P =posterior; Downloaded from http://jn.physiology.org/ by 10.220.33.5 on June 18, 2017 an excitatory junctional potential (EJP) in the muscle leaves little doubt that the cells are connected directly to the muscles. Such an experiment is shown for M9 in Fig. Il. There is little doubt that M4 is the giant motoneuron described by Johnson (8) and Wiersma (28) since its axon is crossed and leaves through the third root of the same segment, and since it was the only cell that regularly produced contraction in more than one flexor muscle. The other M cells are probably directly connected to muscles since a single shock or low frequency of stimulation produced EJP’s in muscle fibers. The ganglion contains a pair of interneurons that form segments of the lateral giant fiber system. These cells drive the M4 axons and other efferent axons on a one-to-one basis (32) and ought to produce a pronounced contraction when stimulated, but we did not succeed in finding their cell bodies. Downloaded from http://jn.physiology.org/ by 10.220.33.5 on June 18, 2017 B ANTERIOR POSTERIOR D. FIG. 12. Identified excitatory and inhibitory neurons of the second ganglion and the muscles they innervate. A : a composite physiological map derived from two ganglia from different animals. Physiologically identified cell bodies are traced with black lines. The cell bodies are named as described in the text. The peripheral field of innervation of each cell is given in Table 1. Not all of the M8 cell bodies were identified in this experiment (compare with the right side of Fig. 14A). B : drawing of the musculature of the lobster ab- 738 OTSUKA, KRAVITZ, AND POTTER FIG. 13. Physiological maps of the anterior flexor excitatory neurons in three ganglia. In B and C the midline of the ganglion falls approximately at the right border of the photograph. This figure illustrates the variability of the relative positions of Ml, M2, and M3 (compare also Fig. 12A) but it can be seen that the relative sizes of the cells and their position as a group are constant. Medial flexor excitatory cells, M4 and M5. These two large cell bodies are always adjacent and in the center of the ganglion near the midline, M4 being usually the larger (Fig. 12A). M4 lies anterior or anteromedial to M5. As indicated in Fig. 14, their axons cross and leave through the third root of the same segment. These cells are so constant in size and position that they can be recognized by visual cues alone. Lateral flexor excitatory cells, M6 and M7. These are two cell bodies of about the same size, usually in contact, and, with 12, the largest in the ganglion (Fig. 12). Their axons emerge, uncrossed, through the third root of the same segment (Fig. 14, A and Bl). As a pair they are conspicuous, lying over the base of the first root in relation to a cluster of small cells (MS’s) and to 12; however, their relative positions are variable (Figs. 12A; 14, A and Bl). These neurons often show a burst of action potentials, usually overshooting, after penetration with a microelectrode. It is sometimes difficult to produce contraction by intracellular stimulation of these cells. Downloaded from http://jn.physiology.org/ by 10.220.33.5 on June 18, 2017 12,13, and M4 are broader than is indicated in Table 1. The functions of the cells are indicated in Fig. 14 and the destinations of the axons are shown with arrows. Anterior flexor excitatory cells, Ml, M2, and M3. These three cells lie at the anterior pole of the ganglion; their relative sizes are constant, Ml being the smallest and M3 the largest (Figs. 12,13, and 14). They innervate flexor muscles and are labeled with an F in Fig. 14. Their axons cross the midline and leave through the third root of the ganglion just anterior. While the groups are in corresponding positions on the two sides of the ganglion, the arrangement of the cells within the cluster appears to be random (Figs. 12, 13, and 14). Since other cell bodies of similar size are found in this region, these three neurons must be physiologically identified in each preparation GABA AND GLUTAMATE IN CELL BODIES 739 Downloaded from http://jn.physiology.org/ by 10.220.33.5 on June 18, 2017 Swimmeret excitatory cells, M8’s. These are part of a compa .ct cluster of relatively small cell bodies lying at the base of the first root and usually bordered by 12 and the lateral and posterior flexor excitatory cells. They are labeled “s” in Fig. 14, A and Bl. Their axons leave uncrossed through the first root. As many as five cell bodies have been found in this group (e.g., stimulation. Fig. 14A). They must be identified by intracellular Posterior flexor excitatory cells M9, MlO, and Ml 1. These are two cell bodies (M9, MlO) of similar size and a smaller one (Mll, labeled with an “f” in Fig. 14) located at the posterior-lateral part of the ganglion. All send axons through the ipsilateral third root of the same segment. Since we have not distinguished M9 and Ml0 functionally, we arbitrarily call the cell adjoining the small cell cluster M9. The positions of M9 and Ml0 are variable, and at least one of the two often needs physiological identification. They probably innervate different parts of M. obliquus ant. The small cell, Mll, is usually located medial or posterior to the larger ones, but in many preparations it was difficult to find, possibly because it lies deeper. Ml1 can be recognized only by stimulation. Extensor excitatory cells M12, M13, and M14. These three neurons lie posterior to the inhibitory group and medial to the posterior flexor excitatory cells (Figs. 12A and 14). Their axons run through the ipsilateral second root of the same segment (Fig. 14). Their relative positions are variable, and usually they can be recognized only by intracellular stimulation. Inhibitory cells II, 12, and 13. These cells are adjacent to each other in spite of the fact that 11 innervates a weak tonic flexor, 12 innervates the main twitch flexor and 13 extensor muscles. This is in contrast to the groups of excitatory cells which always innervate synergist muscles. The inhibitory axons decussate while the axons of neighboring excitatory cells do not (Fig. 14). Therefore, inhibitory cells are not associated with excitatory cells which innervate the same muscles (Fig. 17), nor is there a consistent relation between inhibitory and excitatory cells which innervate synergistic or antagonistic muscles (Fig. 12, Fig. 14). For these reasons, 11, 12, and 13 appear to form a natural group. 12 is the largest cell in the ganglion. 13 lies posterior or posteromedial to 12. 11 is usually in contact with 12, but its position is variable. 12 and 13 could be located with reasonable confidence by visual inspection, but 11 required physiological identification. 11 and 12 send their axons through the third root, 13 through the second. Confirmation of the identity of 12 and 13 was easy since no other large cells near 12 and 13 send their axons to the contralateral roots. When the superficial or main branch of the third root was stimulated with low intensity, the only large cell in the posterior and contralateral half of the ganglion responding with an antidromic action potential was 12. Similarly, on stimulation of the second root, 13 was the only large cell responding with antidromic action potentials. It is likely that 12 corresponds to the common flexor inhibitor described in the crayfish abdomen by Kennedy and Takeda (ll), since it sends branches into all major subdivisions of the third root. 740 OTSUKA, KRAVITZ, 14A. POTTER (See opposite page.) General remarks. We have seen no exception to the observations by Wiersma (29) that efferent axons to the swimmeret, extensor, and flexor muscles run through the first, second, and third roots, respectively. In Fig. 14, A and B, it can be seen that the excitatory cell bodies in the anterior part of the ganglion send axons to the opposite side, while the posterior groups of excitatory cells give rise to uncrossed axons. The three pairs of identified inhibitory axons on each side are crossed. About 80% of the large cells in the ganglion have been identified: this constitutes about 10% of the total cell population (30). Electrical coupling between cell bodies. Electrical synapses were sought between pairs of inhibitory cells by the methods used in earlier studies (e.g., 5, 27); the membrane potential of one cell was altered by passing hyperpolarizing pulses while recording from both cells. The potential changes in pairs of 11’s and pairs of 12’s had the characteristics of passive spread of current, and hyperpolarizations spread equally well in either direction. These results show that these cells are linked to each other by electrical synapses. We have no convincing demonstration of electrical coupling between 13’s. The attenuation of the potential spreading from one 12 cell body to the other was usually about 60 : 1. In the case of 11’s, the attenuation was about 50 : 1. Evidence from histology and from experiments with injections of dyes that diffuse rapidly within the cells indicates that the axons of the inhibitory neurons decussate near the dorsal surface of the ganglion. If this is the place Downloaded from http://jn.physiology.org/ by 10.220.33.5 on June 18, 2017 Fig. AND GABA AND GLUTAMATE IN CELL BODIES Downloaded from http://jn.physiology.org/ by 10.220.33.5 on June 18, 2017 FIG. 14. Physiological and chemical maps of two ganglia. Part A shows the second ganglion of a l&kg lobster, and B the third ganglion of a 0.5-kg lobster. In A and Bl, physiologically identified inhibitors have been marked with white and excitors with black. B2 is a chemical map of the same ganglion as Bl; cells containing more than 2 X lo-l1 moles of GABA were marked white; others containing no detectable GABA were marked black. F =excitatory cells innervating flexor muscles; f =excitatory cells innervating the superficial flexor muscles (i.e., Mll). E =excitatory cells of the extensor musculature. s =excitatory cells of the swimmeret muscles (i.e., M8). 11, 12, and 13 =inhibitory cell bodies. Arrows indicate whether the axons are crossed or uncrossed; a diagonal arrow indicates that the axon leaves through the third root. All but the most anterior cells send axons through roots of the same ganglion. Not all cells were identified in both ganglia, e.g., the swimmeret neurons on the left side of photograph A. This figure may be compared with Fig. 12A. OTSUKA, 742 KRAVITZ, AND POTTER of electrical coupling, it is not surprising that the interaction appears to be weak when recorded far away in the cell bodies on the ventral surface. As was mentioned above, the spontaneous activity of the 11’s was often synchronous (Fig. 9A). The experiment illustrated in Fig. 15 suggests that this synchrony is not produced by common interneurons, but is achieved directly. When one of the Table _____I~~-~~ 1. Cells Axons 13 Axons M4 M5 11 12 abdowlinal ganglion and the muscles they control* crossed, emerging through 3rd root of lstl ganglion M. obliquus posterior 2 (3), medial part M. obliquus anterior 2 (2) M. t,ransversus abdominis 2, ventral part (9) rossed, emerging through M. do rsalis profundus, crossed, emerging M. obliquus M. obliquus M. obliquus M. kansversus M. obliquus M. ventralis M. obliquus 2nd root of 2nd 1ateral part ( 13) tlhrough 3rd root of posterior 3 (5), lateral anterior 2, caudal part anterior 4 (6) abdominis 3, dorsal anterior 3 (4) superficialis abdominis anterior 2, caudal part ganglion 2nd ganglion part (10) part-/ 2, 3 (I) (10) Axons M8 uncrossed, emerging through M. remotor II pedis spuris 1st root of 2nd ganglion (swimmeret muscle) Axons Ml2 Ml3 Ml4 uncrossed, emerging through 2nd root of 2nd M. dorsalis profundus, medial part (11) M. dorsalis profundus, middle part (12) M. dorsalis profundus, lateral part (13) ganglion Axons M6 M7 M9 Ml0 Ml1 uncrossed, emerging through 3rd root M. transversus abdominis 3, dorsal M. obliquus anterior 2, caudal part M. obliquus anterior 4(6) M. obliquus anterior 4 (6) M. ventralis superficialis abdominis ganglion of 2nd partt (10) 2, 3 (l), lateral part * Posit,ions of the cells and muscles are shown in Fig. 12. Numbers in parentheses the labels on tlhe muscles in Fig. 121% The nomenclature is that, of Schmidt (24). corresponding to 8 in segment posterior to second rib. refer to 7 Muscle 11’s was internally stimulated, the muscle innervated by the other consistently showed IJP’s. It is probable that the electrical synapse is the mechanism for coupling the activities of the two inhibitory neurons. Weak electrical interaction was also observed between neighboring large excitatory cells within the anterior, medial, and lateral flexor groups. This coupling persisted after the cell bodies were separated by removing the tissue between them, demonstrating that the site of the current spread is not at the cell body. Downloaded from http://jn.physiology.org/ by 10.220.33.5 on June 18, 2017 Axons Ml M’t M3 of the second GABA AND GLUTAMATE IN CELL BODIES 743 PART II Chemical architecture of the ganglion GABA content of excitatory and inhibitory cell bodies. A correlation between GABA content and physiological function was established in experiments in which many cells were removed by free-hand dissection from living ganglia, following physiological identification. The results of such assays are shown in Table 2. Individual values are given to illustrate the variability of the results. GABA analyses were made on 36 excitatory cell bodies. In 32, no GABA was detected (less than lo-l1 moles per cell). Four of the cells appeared to contain l-l.7 x lo-l1 moles, but this is below the limit of accurate measureI’abLe 2. GABA contents of excitatory and inhibitory cell bodies from fresh ganglia -Cell Type GABA, lo-11 moles Average, 10-l” moles Excitors Anterior flexor Medial flexor Lateral flexor Pas terior flexor Extensor Inhibitors 11 12 13 2.1, 3.5, 2.6 2.7 7.2, 12.5, 5.1, 7.1, 10.0, 6.3, 7.1, 7.3, 7.8, 7.3, 7.9, 7.3, 8.0, 9.0 6.9, 2.6, 1.9, 6.8, 2.7, 5.7 B&h 0.5- and 1.5~kg lobst.ers were was 2 X lo-l1 moles. Semicolons separate same ganglion or from different ganglia. used. The limit values obtained of accurate from cells 7.9 4.4 measurement on opposite of GABA sides of the Downloaded from http://jn.physiology.org/ by 10.220.33.5 on June 18, 2017 FIG. 15. Effect of stimulating homolateral 11 of the third ganglion on the membrane potential of the superficial flexor muscle 3, 4. Short current pulses were delivered to 11 cell body through an intracellular electrode. Note that latency is variable from stimulation to stimulation. Arrow indicates the onset of the IJP. 744 OTSUKA, KRAVITZ, AND POTTER Downloaded from http://jn.physiology.org/ by 10.220.33.5 on June 18, 2017 ment (2 X lo-l1 moles). In these cases the possibility of contamination of samples with small inhibitory cell bodies must be considered. In an attempt to determine more accurately the GABA content within excitatory cell bodies, two samples containing six and eight cells were analyzed. The GABA contents of these samples were less than 1.3 x lo-l1 moles, again below the limit of accurate measurement; the average GABA content per cell was therefore less than 2x lo-l2 moles. The conclusion wa s that GABA if present in excitatory cell bodies is in amounts too small to be reliably measured by our methods. This stands in sharp contrast to the results with physiologically identified inhibitory cells from fresh ganglia (Table 2); in all but one of the 23 cells more than 2 X lo-l1 moles of GABA was present. The differences in average GABA content are due mainly to differences in cell volume. 11 is small and difficult to isolate with hand-held instruments; therefore few analyses were performed on this cell. Figure J4B illustrates the correlation between inhibitory function and a high GABA content, and excitatory function and a low GABA content. All the marked cells in Fig. 14Bl were physiologically identified and then the ganglion was photographed. Cells with more than 2 X lo-l1 moles of GABA are presented in white in Fig. 1442; those without detectable GABA in black. Twenty-three of the cells were assayed for GABA including many of the identified cells. The GABA contents of eight identified cells were assumed from the data of Table 2, and these cells have been marked black or white accordingly. While the procedures described above established a clear difference between excitatory and inhibitory cells, within each type of inhibitory cell there was considerable variation in the GABA content. This could be due to 1) differences in cell volume from animal to animal (see below), 2) leakage or changed GABA metabolism during the impalement and stimulation required for physiological identification, and 3) damage during isolation of the living cell with forceps. The results of three experiments designed to test these points are shown in Table 3. In the first experiment the inhibitory cells were identified by visual inspection alone and were removed quickly from the living ganglia with forceps; there were no failures to detect GABA and the average GABA contents were somewhat higher than those reported in Table 2. In the second experiment the cells were physiologically identified, but they were removed after the ganglia had been freeze-dried, as described in METHODS. Again, the GABA contents were higher, especially in the 13’s. Finally, the two lower columns of Table 3 show the results obtained when ganglia were rapidly removed from the whole animal and frozen, as described in METHODS. These cells were visually identified. The average GABA contents were highest of all. These values are the most reliable because there was the least chance for damage to the cells, and because it is unlikely that the GABA contents were affected significantly by altered metabolism, assuming that the enzyme activities in the cell bodies are similar to the optimal ac- GABA Table --__----___ 3. GAHA AND contents 1-1- GLUTAMATE of excitatory and IN inhibitory cell GABA Content, Visually from 12 13 bodies BODIES isolated lo-11 molgnt 745 in several Average, ways* lo-11 molest identified inhibitors fresh ganglia 8.3, 5.0, Visually from 12 13 11.4, 6.5, 8.6 5.7, 5.2, 7.8, 4.1, 8.7 5.4 4.5 identified indried ganglia 7.6, 9.0, 8.4 6.6, 8.4, 6.0, 8.3 6.5 5.0 identified inhibitors dried ganglia Visually from M4 M6 M9 Ml identified excitors dried ganglia and M5 and M7 and Ml0 and M2 * For the methods of identifying 0.5-kg lobsters only. 7 The 11.0, 6.5, 9.7, 9.8 6.1, 7.5, 5.9, 7.8 9.2 6.8 4 0, 0, 2 pooled 0, 0, 1.3 0.5 pooled 0.5 cells 0.5, 0 cells and limit isolating of accurate the cells, see text of PART II. Data measurement was 2 X lo+ moles. tivities measured in peripheral axons (14). (These ganglia were immersed in circulating blood, and the neurons were intact, until less than 1 min before freezing. In 1 min, a volume of homogenized axoplasm equal to the 12 cell body (see Table 5) is able to synthesize or degrade only about 0.3y0 of the amount of GABA in the 12 cell body.) It should be noted that the highest GABA contents per cell found with the various isolation procedures were similar (Tables 2 and 3); we conclude that the major causes of variation in the results were damage during isolation and differences in cell volume from ganglion to ganglion (see below). Glutamate content of excitatory and inhibitory cell bodies. Glutamate analyses were performed on 12, 13, M6, and M7 cell bodies (Table 4). The limit of accurate measurement in all cases was 4 X UP moles. In contrast to Table 4. Glutamate contents of excitatory Cell and inhibitory Glutamate, cell bodies from lo-11 moles fresh Average, ganglia” 10-u moles - 12 13 15.0, 8.2, 5.9, 13.0, 11.0, 11.4, 5.2, 9.9, 6.5 8.6, 15.0, 7.2, 9.5, 7.8 M (6 or 7) 9.6, * Data are from 0.5-kg 6.5, 12.7, lobsters 6.5, only. 7.3 6.8, 14.0, 16.5, 10.3 9.6 8.5 Downloaded from http://jn.physiology.org/ by 10.220.33.5 on June 18, 2017 physiologically hibitors from 12 13 from CELL OTSUKA, 746 Table KRAVITZ, 5. Volumes AND of cell bodies POTTER of efferent neurons* ~~ Cell Total Volume, 12 M6 and M4 M5 M7 Cytoplasm, 10-g liter? 6.74-6.48 7.2-7.0 8.42-8.07 6.65-6.25 8.93-6.79 5.35-4.9s 3.34-3.42 3.85 3.57 3.4 4.51-4.63 3.64-3.5 4.97-4.95 5.4 -5.4 4.21 6.29-6.73 3.45-3.35 2.24 2.32 3.32 2.5 4.4-4.5 4.16-4.23-3.79-3.81 4.95-4.4 -5.13 6.05-4.14 3.81 5.32 4.47 4.38 3.32 3.75 5.33 3.05 Average, 10-g litert Total volume: Cytoplasm: 6.89 3.75 Total volume: Cytoplasm: 4.67 2.97 Total volume: 4.64 Total volume: 4.49 Total volume: 3.86 * All cells are from second abdominal ganglia of 0.5-kg lobsters treated in METHODS, Measurements (5) were on physiologically identified cells; all visually identified. Hyphens separate measurements on the same ganglion. sents the total volume of the cell body less the volume of cytoplasmic vacuoles $ When two values were available for a ganglion, they were averaged and the computing the value in this column. as described others were t This repre(cf. Fig. 2B). mean used in the result with GABA, excitatory and inhibitory cells contained similar amounts of glutamate. This is similar to the findings in isolated axons (14). It provides an internal control for the GABA analyses, demonstrating that the excitatory cells were not selectively damaged by the isolation procedures. Glutamate is a fairly uniform constituent of tissues, and the uncertainty introduced by contamination with glia and connective tissue is greater than with the GABA analyses. An isolated cell with a small amount of adhering tissue is shown in section in Fig. 4C. This qualification must be borne in mind in considering the results obtained with these methods. Concentration of GABA and glutamate in excitatory and inhibitory cell bodies. Our calculations of the concentrations of these substances depend on volume measurements made from sections of fixed ganglia (see METHODS). These volumes are given in Table 5. Values obtained from cells of the same ganglion are separated by hyphens; such pairs of cells generally had similar volumes. A conspicuous feature of the cell bodies was vacuoles of an uncer- Downloaded from http://jn.physiology.org/ by 10.220.33.5 on June 18, 2017 13 10-g liter GABA AND GLUTAMATE IN CELL BODIES 747 Table 6. GAHA and glutawzate concentrations in excitatory and GABA, 10-S molesJiter inhibitory Glutamate, cell bodies* 10-S moles/liter Cell 12 I3 M (6 or 7) M (4 or 5) Total cell volume Cytoplasm Total cell volume Cytoplasm 13.4 14.6 24.5 22.9 14.9 20.6 18.3 27.5 32.3 (1 *a * These values were obtained by dividing the average GABA and glutamate contents (Table 3, bott,om categories and Table 4) by the average volumes (Table 5). It should be noted that the concentration of GABA in M (4 and 5) is uncertain because the GABA content, is below the limit of accurate measurement. GABA contents shown in the lower columns tents shown in Table 4. of Table 3 and glutamate con- DISCUSSION Architecture of the abdominal ganglia. A high degree of constancy in the size, position, and connections of individual neurons appears to be general in invertebrate nervous systems (e.g., 1). Indeed, from available physiological and anatomical information there is no reason to doubt that this is also true of chordate nervous systems. This constancy was essential to the present study, as it was in studies on other forms (e.g., 4, 9, 10, 19). With regard to the way cells are arranged in the ganglion, it is found that neurons controlling certain muscles do not lie together, or even at corresponding points in successive ganglia. For example, at least seven excitatory cells innervate M. obliquus ant. 3 (Fig. 16); these lie in four cell groups in three ganglia, both sides being represented. Figure 17 illustrates that the identified inhibitory cells are not found adjacent to the excitatory cells with which their axons run in the periphery. (It should be noted that the course of axons through the ganglion is not known in detail; Figs. 16 and 17 are diagrammatic in this regard.) That cells lying at widely separate points may send axons to the same target is familiar Downloaded from http://jn.physiology.org/ by 10.220.33.5 on June 18, 2017 tain nature (Fig. 2B). In 12 and 13 these vacuoles consistently formed a large portion of the cell body, about 40%. Usually less than 15% of the cytoplasm of M6, and M7, and other excitatory cells was composed of the vacuoles. This difference between the large excitatory and inhibitory cells was so pronounced that it could be used to locate 12 and 13 rapidly in the sections. In 13 cells an independent measure of the vacuoles was available; this was subtracted from the total volume to give the “cytoplasmic volume.” The nucleus was found to comprise about 1 y0 of the total, but the contributions of mitochondria, endoplasmic reticulum, and other cellular components are not yet known. In Table 6 the concentrations of GABA and glutamate are shown, computed both for total volume and cytoplasmic volume, using the average 748 OTSUKA, KRAVITZ, AND POTTER FIG. 16. Excitatory of M. obliquus ant. tion, see text. innervation 3. For descrip- reasonable to suppose that the same Figs. 16 and 17. Similarly, we may suppose that the excitatory cells are grouped according to their reflex connections. In some cases adjacent excitatory cells send axons to the same muscle (e.g., M8’s; M9 and MlO), in other cases synergistic muscles are innervated (e.g., Ml-M3; M4 and M5; M6 and M7). One is reminded that in the mammalian spinal cord, motoneurons innervating the same muscle are clustered and may also be closely associated with motoneurons innervating synergistic muscles (e.g., 22). The invariable clustering of the inhibitory cells 11, 12, and I3 is more surprising. These cells innervate muscles of different or contrary function prof. lateral part M.vent. superficialis 922 R3 caudal FIG. 17. Locations of excitatory For description, see text. R2 and ganglionic roots, respectively. part R3 and inhibitory indicate that cell bodies the axons innervating lie in the three muscles. second and third Downloaded from http://jn.physiology.org/ by 10.220.33.5 on June 18, 2017 in neuroanatomy; an example from primates is the convergence of spinal interneurons and cortical cells on spinal motoneurons. In such instances the converging neurons have different inputs. Kennedy and Takeda (12) have demonstrated that axons innervating the superficial flexor muscle of the crayfish abdomen subserve different reflexes, and it is is true of the cells illustrated in GABA AND GLUTAMATE IN CELL BODIES 749 Downloaded from http://jn.physiology.org/ by 10.220.33.5 on June 18, 2017 and doubtless receive quite different inputs. It is possible that they do have certain connections in common or that they interact in a way that we did not detect; still another possibility is that 11, 12, and 13 have a common embryological origin. Perhaps cells destined to have inhibitory chemistry differentiate at special points in the developing ganglia and then acquire distinct and appropriate incoming and outgoing connections. In this regard it may be noted that another pair of high-GABA cells of unknown function lie in the midline in the immediate vicinity of this cluster. It must be borne in mind that so far only three inhibitory neurons have been identified; of course, clustering of synergistic inhibitory neurons may also occur in the lobster CNS. The present point is simply that in this instance neurons of antagonistic function are consistently found together. The grouping of neurons with common transmitter chemistry is familiar in the mammalian nervous system. Motoneurons are clustered, as are cell bodies of preganglionic neurons of the autonomic nervous system. In the cerebellum it has been reported that the layers of cells are discrete with respect to function, e.g., granule cells being excitatory and Purkinje cells inhibitory (3,6). Another point of interest concerns the constant relative sizes of the cell bodies. Alm .ost certainly this is related to the number of muscle fibers innervated, as in the case of mammalian motoneurons (18) . The largest cell body is 12 which innervates the massive caudal part of the anterior oblique muscle, and doubtless other flexor bundles as well. In contrast, one of the smallest of the identified cells, 11, innervates only the thin superficial flexor muscle. The smallest excitatory neurons we identified innervate the diminutive swimmeret muscles. The possible role of electrical coupling in coordinating the activities of cells on the two sides of the ganglion was discussed above. We find that neurons with chemical inhibitory synapses in muscle also form excitatory synapses in the CNS that are electrical Thus a cell can be ex .citatory and inhibitory without the need for a second transmit ter compou .nd or an altered postsynaptic chemistry. Another case of a neuron that forms both excitatory and inhibitory contacts has recently been reported in the snail Aplysia by Kandel et al. (9). In this case, a single transmitter, acetylcholine, and two different postsynaptic chemistries are apparently involved. GABA and glutamate in the cell bodies of efferent neurons. Our long-term interest in the cell bodies is to discover the mechanism of regulation of the enzymes of the GABA pathway in excitatory and inhibitory neurons. As yet we have no information about the enzymes; the present paper provides information only about the substrates, GABA and glutamate. There is no doubt that the GABA content of inhibitory cell bodies is considerably higher than that of excitatory cell bodies and that glutamate is present in about equal amounts in the two types of cells. We have confidence in the values for GABA contents of inhibitory cells because similar high values were obtained with several methods of identification and isolation 750 OTSUKA, KRAVITZ, AND POTTER Downloaded from http://jn.physiology.org/ by 10.220.33.5 on June 18, 2017 (Tables 2 and 3). The glutamate contents are more tentative because of inevitable contamination with small amounts of other tissues. Our calculations of concentration within the cells are less certain for several reasons. First, we do not know how these amino acids are distributed in the cytoplasm, Since there is no evidence in lobsters that either the substrates or enzymes are bound to cellular constituents in axons or cell bodies (cf. 14), it would be reasonable to assign GABA and glutamate contents to cytoplasm exclusive of organelles. Light microscopy provides a rough estimate of the volume occupied by vacuoles and nucleus (Table 5), but the volumes of mitochondria and other formed elements are not Yet known. In view of these uncertainties, the only safe conclusion is that the concentrations calcula ted from total cell volumes represent lower limits. Another uncertainty regarding GABA and glutamate concentrations arises from the fact that we could not measure the volumes of the cells that were chemically analyzed. By dividing average substrate contents by average volumes we may be obscuring differences from animal to animal. It would be interesting to know why the volumes of 12 and 13 vary over about a twofold range (Table 5). It is unlikely that there are great differences in GABA concentration between cell bodies and axons of inhibitory neurons. The average concentrations of GABA in the abdominal cell bodies (Table 6) are about one-seventh (total volume) or one-fourth (cytoplasmic volume) those previously found in inhibitory axons of the walking legs (15). Since we have not measured substrate contents in two parts of the same neuron, however, this comparison between cell body and axon is tentative. If the con .centrat ion of GABA in excitatory cell bodies is similar to that in lim b axons (lo-” M or less), it is not surprising that the amounts we observed were below the limit of accurate measurement; a cell body with a volume of 4.5 x lO-g liter would contain 4.5 x lo-l2 moles or less (cf. Table 3). The rather high concentrations of the physiologically active compounds glutamate and GABA throughout the lobster neurons may be compared with the distribution of acetylcholine and catecholamines in mammalian nerve cells. By histochemical means it has been demonstrated that catecholamines are found throughout the cells (2), but the concentration gradient from cell body to nerve terminal is reported to be steep (1: 100 to 1: 1000). Acetylcholine is found in axons (17) at a considerable distance from the terminals, in axoplasm is not known. The relatively high concenbut the concentration tration of GABA and catecholamines in cell bodies permits a conclusion about transmitter chemistry from anal yses of this part of the cell. It is not yet known whether this is also true of cholinergic neurons. SO far, in the lobster nervous system a high content of GABA has labeled a cell as inhibitory. If this is true in other nervous systems it could be valuable in mapping certain inhibitory pathways. Recent work on the cerebellum has indicated that Purkinje cells are inhibitory. The layer in which they lie contains considerably more GABA than other cerebellar layers (16). Analyses of single cells could show whether GABA is an index GABA AND GLUTAMATE IN CELL BODIES 751 of inhibitory function in the mammalian CNS. Neither acetylcholine nor noradrenaline analyses can be used in this way to indicate neuronal function, SUMMARY ACKNOWLEDGMENTS We gratefully acknowledge the histological assistance of Mrs. Florence Miss Karen Fischer. Robert Bosler gave us unfailing assistance throughout We received much helpful advice from Professors Kuffler, Furshpan, Hubel, Foster and the work. and Wiesel. REFERENCES I. 2. 3. 4. 5. 6. 7. 8. 9. 10. 11. BULLOCK, T.H. ANDHORRIDGE,G. A. Structure and Function in the Nervous Systems of Invertebrates. San Francisco: Freeman, ~01s. I, and II, 1965. DAHLSTR~M, A., FUXE, K., AND HILLARP, N.-A. Site of action of reserpine. Acta Pharmacol. Toxicol. 22: 277-292, 1965. ECCLES, J. C., LLIN&, R., AND SASAKI, K. Intracellularly recorded responses of the cerebellar Purkinje cells. ExptZ. Brain Res. 1: 161-183, 1966. FURSHPAN, E.J. AND FURUKAWA, T. Intracellular and extracellular responses of the several regions of the Mauthner cell of the goldfish. J. Neurophysiol. 25: 732-771,1962. FURSHPAN, E. J. AND POTTER, D. D. Transmission at the giant motor synapses of the crayfish. J. Physiol., London, 145: 289-325, 1959. ITO, M. The origin of cerebellar inhibition on Deiters’ and intracellebellar nuclei. In: Studies in Physiology, edited by D. R. Curtis and A. K. McIntyre, Berlin: Springer, 1965, p. 100-106. JAKOBY, W. B. AND SCOTT, E. M. Aldehyde oxidation. III. Succinic semialdehyde dehydrogenase. J. BioZ. Chem. 234: 937-940, 1959. JOHNSON,G. E. Giant nerve fibers in crustaceans with special reference to Cambarus and PaZaemonetes. J. Comp. NeuroZ. 36: 323-373, 1924. KANDEL, E.R., FRAZIER, W.T., AND COGGESHALL, R.E. Opposite synaptic actions mediated by different branches of an identifiable interneuron in ApZysia. Science 155: 346-349, 1967. KANDEL, E. R. AND TAUC, L. Heterosynaptic facilitation in neurones of the abdominal ganglion of ApZysia depilans. J. PhysioZ., London 181: l-27, 1965. KENNEDY, D. AND TAKEDA, K. Reflex control of abdominal flexor muscles in the crayfish. I. The twitch system. J. ExptZ. BioZ. 43: 211-227, 1965. Downloaded from http://jn.physiology.org/ by 10.220.33.5 on June 18, 2017 1. The physiological organization of lobster ganglia has been correlated with certain chemical properties of excitatory and inhibitory neurons. Methods were developed for physiologically identifying and then isolating the cell bodies of efferent neurons. Gamma-aminobutyric acid (GABA) and glutamate contents of the cell bodies were assayed by enzymic methods. 2. The physiological functions of 21 pairs of neurons were determined. The cell bodies occur in clusters which are constant in location and composition. Adjoining excitatory cells innervate the same or synergistic muscles. In contrast, the identified inhibitory cells are grouped together although they innervate muscles with different or opposing action. Certain cell bodies are so constant in size and position that they can be recognized with confidence by visual inspection alone. 3. In conformity with previous studies on axons, GABA analysis revealed two populations of cell bodies; inhibitory cells with a high GABA content (the lower limit being about 0.014 M) and excitatory cells with a low GABA content (below the limit of accurate measurement with our assay). Glutamate is about equally concentrated in the two types of cell bodies. 752 12. 13. 14. 15. 16. 18. 19. 20. 21. 22. 23. 24. 25. 26. 27. 28. 29. 30. 31. 32. KRAVITZ, AND POTTER KENNEDY, D. AND TAKEDA, K. Reflex control of abdominal flexor muscles in the crayfish. II. The t.onic system. J. Ex@. Bid. 43: 229-246, 1965. KRAVITZ, E. A., KUFFLER, S. W., AND POTTER, D. D. Gamma-aminobutyric acid and other blocking compounds in Crustacea. III. Their relative concentrations in separated motor and inhibitory axons. J. Neurophysiol. 26: 739-751, 1963. KRAVITZ, E. A., MOLINOFF, P. B., AND HALL, Z. W. A comparison of the enzymes and substrates of gamma-aminobutyric acid metabolism in lobster excitatory and inhibitory axons. Proc. Natl. Acad. Sci. U. S. 54: 778-782, 1965. A further study of the distribution of y-aminoKRAVITZ, E. A. AND POTTER, D. D. butyric acid between excitatory and inhibitory axons of the lobster. J. Neurochem. 12: 323-328, 1965. KURIYAMA, K., HABER, B., SISKEN, B., AND ROBERTS, E. The y-aminobutyric acid system in rabbit cerebellum. Proc. Natl. Acad. Sci. U.S. 55: 846-852, 1966. LOEWI, 0. AND HELLAUER, H. uber das Acetylcholin in peripheren Nerven. Arch. Ges. PhysioZ. 240: 760-775, 1938. MCPHEDRAN, A. M., WUERKER, R. B., AND HENNEMAN, E. Properties of motor units in a homogeneous red muscle (soleus) of the cat. J. Neurophysiol. 28: 71-84, 1965. NICHOLLS, J. G. AND KUFFLER, S. W. Extracellular space as a pathway for exchange between blood and neurons in the central nervous system of the leech: ionic composition of glial cells and neurons. J. Neurophysiol. 27: 645-671, 1964. OTSUKA, M., IVERSEN, L. L., HALL, Z. W., AND KRAVITZ, E. A. Release of gammaaminobutyric acid from inhibitory nerves of lobster. Proc. N&Z. Acad. Sci. U. S. 56: 1110-1115, 1966. The r-aminobutyric acid (GABA) OTSUKA, M., KRAVITZ, E. A., AND POTTER, D. D. content of cell bodies of excitatory and inhibitory neurons of the lobster. Federation Proc. 24: 399, 1965. ROMANES, G. J. The motor pools of the spinal cord. In: Progress in Brain Research: Organization of the SpinaZ Cord, edited by J. C. Eccles and J. P. Schade. Amsterdam: Elsevier, 1964, vol. 11, p. 93-119. SANDBORN, E., KOEN, P. F., MCNABB, J. D., AND MOORE, G. Cytoplasmic microtubules in mammalian cells. J. Ultrastruct. Res. 11: 123-138, 1964. SCHMIDT, W. Die Muskulatur von Astacus fZuviatiZis (Potamobius astacus L.) Ein Beitrag zur Morphologie der Decapoden. 2. Wiss. ZooZ. 113: 165-251, 1915. SHUKUYA, R. AND SCHWERT, G. W. Glutamic acid decarboxylase. I. Isolation procedures and properties of the enzyme. J. BioZ. Chem. 235: 1649-1661, 1957. TAKEDA, K. AND KENNEDY, D. Soma potentials and modes of activation of crayfish motoneurons. J. CeZZ. Comp. PhysioZ. 64: 165-182, 1964. WATANABE, A. AND GRUNDFEST, H. Impulse propagation at the septal and commissural junctions of crayfish lateral giant axons. J. Gen. PhysioZ. 45: 267-308, 1961. Giant nerve fiber system of the crayfish. A cont.ribution to WIERSMA, C. A. G. comparative physiology of synapse. J. NeurophysioZ. 10: 23-38, 1947. WIERSMA, C. A. G. On the motor nerve supply of some segmented muscles of the crayfish. Arch. NeerZ. Sci. 28: 413-422, 1947. WIERSMA, C. A. G. On the number of nerve cells in a crustacean central nervous system. Actcz PhysioZ. Pharmacol. NeerZ. 6: 135-142, 1957. WIERSMA, C. A. G. AND RIPLEY, S. H. Innervation patterns of crustacean limbs. Physiol. Comparata et Oecol. 2: 391-405, 1952. WIERSMA, C. A. G. AND SCHALLEK, W. Potentials from motor roots of the crustacean central nervous system. J. NeurophysioZ. 10: 323-329, 1947. Downloaded from http://jn.physiology.org/ by 10.220.33.5 on June 18, 2017 17. OTSUKA,