- Email: [email protected]
Morphofunctional evidence for mature synaptic contacts on the Mauthner cell of 52-hour-old zebrafish larvae
Pergamon
PII:
Neuroscience Vol. 80, No. 1, pp. 133–145, 1997 Copyright ? 1997 IBRO. Published by Elsevier Science Ltd Printed in Great Britain. All rights reserved 0306–4522/97 $17.00+0.00 S0306-4522(97)00092-4
MORPHOFUNCTIONAL EVIDENCE FOR MATURE SYNAPTIC CONTACTS ON THE MAUTHNER CELL OF 52-HOUR-OLD ZEBRAFISH LARVAE A. TRILLER,* P. ROSTAING,* H. KORN† and P. LEGENDRE†‡§ *CJF 94-10 INSERM, Ecole Normale Supe´rieure, 46 rue d’Ulm, 75005, Paris, France †Laboratoire de Biologie Cellulaire et Mole´culaire (INSERM U 261), De´partement des Biotechnologies, Institut Pasteur, 25 rue du Dr. Roux, 75015, Paris, France Abstract––In a previous study, miniature inhibitory synaptic events recorded in the Mauthner cell of the 52-hour-old zebrafish larvae (Brachydanio rerio) were found to be mainly glycinergic. Their amplitude distribution was not Gaussian and it was proposed that their large amplitude variation might reflect the activation of immature synapses. However, ultrastructural studies of the synaptic contacts over the M-cell soma of 52 h larvae described here, revealed that numerous synaptic contacts on this neuron are already mature at this developmental stage and that most of them already contain a single active zone. As in the adult goldfish, immunohistochemistry indicates the presence of both glycine- and GABA-immunoreactive boutons which establish synaptic contacts. We also found that, in addition to the predominant glycinergic postsynaptic inhibitory currents, some postsynaptic currents are also GABAergic since they are specifically inhibited by bicuculline (20 µM). GABAergic miniature events (time to peak close to 0.8 ms and decay time-constant close to 4–5 ms) were only detected in the presence of 11.5 mM [KCl]o. Their amplitude distributions were well fitted by one, or at most two, Gaussian curves. Outside-out recordings showed one class of GABA receptors with a main conductance state of 23 pS. This indicates that the smallest GABAergic miniature inhibitory synaptic events correspond to the opening of 14–20 chloride channels Pre- and postsynaptic factors which contribute to the predominance of glycinergic synaptic currents over GABAergic ones in untreated preparations and to the striking differences between their frequencies and their respective amplitude distribution histograms are discussed with reference to the morphological characteristics of the mature synaptic endings impinging on this still developing neuron. ? 1997 IBRO. Published by Elsevier Science Ltd. Key words: inhibitory synapses, GABAA receptors, glycine receptors, miniature synaptic events, electron microcopy, patch-clamp.
Excitatory and inhibitory synaptic activity has been extensively studied in many neurons27,51 including the Mauthner cell (M-cell) of teleosts.29 In this neuron, excitatory and inhibitory synaptic contacts have been morphologically characterized and their respective locations are stereotyped in the goldfish (Carassius auratus),28,38 as well as in the zebrafish (Brachydanio rerio).21,22 Interestingly, the spontaneous synaptic events recorded from the M-cells of both adult goldfish and zebrafish larvae are almost completely inhibitory. Miniature inhibitory postsynaptic currents (mIPSCs), which are mainly glycin§To whom correspondence should be addressed. ‡Present address: Institut des Neurosciences, CNRS URA 1488, Universite´ Pierre et Marie Curie, 7 Quai Saint Bernard, 75005 Paris, France. Abbreviations: AC, axon-cap; EGTA, ethyleneglycolbis (aminoethylether)tetra-acetate; EPTA, ethanolic phosphotungstic acid; HEPES, N-2-hydroxyethylpiperazineN*-2-ethanesulphonic acid; IR, immunoreactive; M-cell, Mauthner cell; mIPSC, miniature inhibitory postsynaptic current; PBS, phosphate-buffered saline; PDP, presynaptic dense projection; TTX, tetrodotoxin; ttp, time-topeak.
ergic, have similar kinetics in both species (time to peak <1 ms and average decay time-constant ranging from 4 to 7 ms;26,34). In the goldfish, the histograms of miniature events could be described by a single Gaussian curve.24. In contrast, in 52-hour-old zebrafish larvae the amplitudes of miniature events ranged from 20 to 500 pA and their distribution consisted of the sum of Gaussian curves, which were not equally spaced.34 This type of distribution is similar to that reported for both adult and developing mammal brains, where amplitude distribution histograms of inhibitory and excitatory miniature events are often skewed5,8,12,34,36,48,55 and, in some structures, clearly multimodal.8,34,37,55 It is notable that in most cases, these amplitude distributions cannot result from cable filtering properties of the cell membrane.4 It has been shown in the goldfish M-cell that glycinergic boutons operate in an all-or-none manner.29 In this system, presynaptic boutons localized on the M-cell soma contain a single grid58 and physiological data suggest that the content of no more than one vesicle can be released at each active
133
134
A. Triller et al.
zone29 since, for example, the amplitude histograms of miniature inhibitory events are unimodal. In contrast, amplitude histograms were multimodal in the presence of tetrodotoxin (TTX) when synaptic events were recorded in the lateral dendrite,23 where synaptic boutons have a variable number (one to four) of active zones.54 Thus, since the hindbrain is not completely developed in 52-hour-old zebrafish larvae,21,22 the widespread amplitude histograms of glycinergic mIPSCs recorded in the M-cell34 could result, at this stage, from the presence of immature synaptic contacts and/or the presence of a variable number of active zones per bouton which may lead to multivesicular release.23 In the present study, we used ultrastructural and physiological methods to characterize further the general properties of the inhibitory system which controls M-cell excitability in the zebrafish larvae. Our ultrastructural studies suggest that a large number of synapses are already structurally mature in 52-hour-old zebrafish larvae and have already a single active zone per bouton. Furthermore, immunohistochemistry studies indicated that in addition to glycinergic terminals, immunoreactive boutons apposed to the M-cell soma can be GABAergic. This finding prompted us to analyse the properties of the miniature synaptic events they generate in the M-cell. EXPERIMENTAL PROCEDURES
Morphological methods Fifty-two-hour-old zebrafish larvae were fixed by immersion in 4% glutaraldehyde in phosphate-buffered saline (PBS, O.12 M, pH 7.4) for 30 min at room temperature, and were kept overnight in the same fixative at 4)C. After three rinses in PBS, whole larvae were dipped in OsO4 (2%) for 2 h at 4)C, washed, and stained en bloc in uranyl acetate. They were then dehydrated and embedded in Araldite. Ultrathin (pale yellow) sections were stained with uranyl and lead citrate. For the identification of presynaptic grids the larvae were fixed as above, washed in PBS and kept for 2 h in ethanolic phosphotungstic acid (EPTA) at 60)C as described elsewhere.58 Semithin (0.5 µm) sections were directly observed under the electron microscope without further staining. With this method, the extracellular space was weakly electron-dense, while the cytoplasm was almost transparent to electrons, thus allowing the contours of presynaptic boutons to be clearly distinguished. One of the main advantages of this approach is that it provides en face views of semithin sections so both presynaptic grids and postsynaptic differentiations can be visualized. Moreover, the number of presynaptic dense projections (PDPs) per grid can be determined. PDPs are not considered as a structure in themselves and they may, instead, correspond to a collapse of the cytoskeleton and of the proteins involved in release between vesicles adjacent to the presynaptic releasing membrane.14,33 For immunocytochemistry, animals were fixed in a mixture of 2% paraformaldehyde and 2% glutaraldehyde for 1 h at room temperature and then kept overnight at 4)C. Following osmification and dehydration, they were included in Durcupan. Semithin (1 µm) and ultrathin (pale yellow) sections were immunostained to detect glycine (antiglycine; Neosystem Lab. Strasbourg, France) and GABA (antiGABA; Immunotech SA, Marseille France). Briefly (for more details see Ref. 60), sections etched for 7 min with 1% orthoperiodic acid in distilled water were incubated
with primary antibodies (anti-glycine 1:200 or anti-GABA 1:100). The antiglycine was mixed with a conjugate of GABA and glutaraldehyde (GABA-G) and the anti-GABA with a conjugate of glycine and glutaraldehyde (Gly-G). Each amino acid conjugate was used at 50 µM. For glycine and GABA, the corresponding immunoreactivity was abolished by preadsorption of the antibody with the corresponding amino acid conjugated with glutaraldehyde. Controls without primary antibodies gave no specific staining. Antibody binding sites were detected with an ABC–HRP kit (Elite Vectastain*, Vector Laboratories, Burlingame, CA) for semithin sections and with goat anti-rabbit immunoglobulins coupled to 15 nm colloidal gold particles (1:50, BioCell, UK) for the electron microscope. Physiological methods Isolated brain preparation. Isolated zebrafish brain was prepared as described before.34 Briefly, the brains of newly hatched larva were dissected out under binocular observation and then glued to a coverslip using a plasma thrombin embedding procedure.10 Before experiments, brain preparations were kept for 15–30 min in an oxygenated (95%O2, 5%CO2) bathing solution containing (in mM) NaCl 145; KCl 1.5; CaCl2 2; MgCl2 1; NaHCO3 26; NaH2PO4 1.25; glucose 10, with osmolarity adjusted to 330 mOsm. Outside-out and whole-cell recordings. Standard whole-cell and outside-out recordings13 were made on the M-cells of mounted brains placed on a microscope stage (Nikon Optiphot). The preparation was continuously perfused at room temperature (20)C) with the oxygenated bathing solution (2 ml/min) in the recording chamber (0.5 ml). Patch-clamp electrodes were pulled either from thin-wall (whole-cell recordings) or thick-wall (outside-out recordings) borosilicate glass. They were fire-polished and filled with (in mM): CsCl 135; MgCl2 2; NaATP 4; EGTA 10; HEPES 10; pH 7.2. The osmolarity was adjusted to 290 mOsm. Whole-cell and outside-out recordings were made using 1–3 MÙ and 10–15 MÙ electrodes, respectively. Outsideout patches were obtained by slowly pulling the pipettes out of the brain. Currents were recorded using an Axopatch 1D (Axon instruments), filtered at 10 KHz, and stored using a digital tape recorder (DAT DTR 1201, SONY). During whole-cell recordings, the series-resistance (4–10 MÙ) was monitored by applying 2 mV hyperpolarizing pulses and 50–70% compensated. To ensure cell dialysis, measurements were made on data obtained at least 3–5 min after the whole-cell configuration was established.34 Drug delivery. Drugs were applied to the preparation via an array of four flowpipes of 400 µm diameter positioned above the brain. TTX (1 µM), strychnine (1 µM) and bicuculline methiodide (20 µM) were dissolved in a perfusate containing (in mM) NaCl 135; KCl 11.5; CaCl2 2; MgCl2 1; glucose 10 and HEPES 10; pH 7.2; osmolarity 330 mOsm. A high KCl solution (11.5 mM) was used to enhance mIPSC frequency in the presence of TTX. All solutions were prepared daily. Bicuculline methiodide (Sigma) was always freshly prepared as this compound is rapidly degraded in ionic solutions.17 Analysis of whole-cell currents. Ongoing synaptic activity was digitized off-line with a Macintosh IICi computer at 24 KHz using MMII software (GW instruments). Synaptic events were automatically detected, as previously described,3,34,35 with home-made software written in Labview II.2 (National Instruments). Briefly, the signals were filtered at 2 KHz, using a second-order Bessel digital filter. The peak of each response was measured as the maximal amplitude deviation from the baseline. The decay time-constant
Mature inhibitory synapses in zebrafish larvae hindbrain
Fig. 1. Cytological features of the 52-hour-old zebrafish Mauthner cell and axon-cap. A) Transverse section through the soma (S) of the neuron at the level of its nucleus (n) with the characteristic ventral and lateral dendrites (VD and LD, respectively). The axon initial segment (IS) is oriented toward the midline and is surrounded by a specialised neuropil, the axon-cap (AC). B) Higher power micrograph showing that the neuron contains numerous polyribosomes (encircled), an extensive rough endoplasmic reticulum (arrows) and numerous Golgi apparati (crossed arrows). C) Micrograph of the axon cap with profiles containing dense core vesicles (arrowhead) or small clear vesicles (arrow). D) Another example showing the glomerular structure of the neuropil within the axon-cap, with small clear vesicle-containing profiles engaged in symmetrical axoaxonic synapse (arrows), and dense core vesicles (arrowhead). These neuronal elements are surrounded by glial processes (asterisks). E) Micrograph illustrating the presence of numerous septate-like junctions (arrows) at the border of the axon-cap. F) Higher power micrograph showing the regularly spaced membranes and the electron-dense structure of the extracellular space. Scale bars, A=5 µm; B–F=0.5 µm.
135
136
A. Triller et al.
Mature inhibitory synapses in zebrafish larvae hindbrain of inhibitory currents was obtained by automatically fitting their decay phase with a single exponential. Assuming that, as in the adult goldfish, all inhibitory presynaptic boutons impinging on the soma contain one active zone,58 and that the amplitude distribution of quantal events is Gaussian,24,26 the amplitude histograms of mIPSCs were treated as the sum of an unknown number of overlapping Gaussian curves34 (and see Ref. 23 for detailed algorithm). Fits with Gaussian curves were then compared to fits with a continuous function of the form (a#(1"exp("x/b))^c)# (exp("x/d)), using a simplex algorithm (Axograph 3, Axon Instruments). In all cases the mIPSC amplitude histograms were better fitted with the sum of Gaussian curves (Standard Square Error gaussians
Identification of the Mauthner cell The M-cell was identified on semithin sections by its localization in front of the otic vesicle, just under the dorsal germinative epithelium, and by its characteristic shape. As seen with low-power electron microscopy, the M-cell contained at 52 h a clear round nucleus and was crescent shaped (Fig. 1A), with a lateral and a ventral dendrite (LD and VD, respectively) which were not fully developed at this stage. The initial segment of the axon arising from the convex side of the cell was oriented toward the midline. The clear cytoplasm (Fig. 1B) contained the classical organelles, Golgi apparatus, smooth and rough reticulum, mitochondria and a large number of free polyribosomes. At this age, the initial segment was already surrounded by a dense neuropil, the axon-cap (Fig. 1A), a structure characteristic of the M-cell. This region displayed numerous axonal profiles containing small clear round vesicles sometimes associated with dense
137
core vesicles (Fig. 1C). As observed around the initial segment of the M-cell of other mature teleosts,39 these profiles are axoaxonic synapses, with presynaptic differentiations in front of each other. These connections are surrounded by thin glial processes (Fig. 1D) which give them a glomerular structure. Within the axon cap, the extracellular space is of variable size, ranging from 10 to 30 nm, and the cleft is relatively transparent to electrons. However, in many instances, the distance between the membranes of adjacent glial cells, or glial cells and axons, remains constant with a width of 20–25 nm (Fig. 1B–C). It is then filled with uneven electron-dense and pseudoperiodic matrix. These ‘‘differentiated’’ appositions vary in length, from 0.1 to 1.5 µm, and they are similar to the septate-like junctions found in the goldfish axon-cap.57 Synaptic boutons over the Mauthner cell soma We investigated the synaptic investment in the soma and in the initial portion of the dendrites. Vesicles containing profiles with a more or less mature synaptic structure were observed all around the M-cell, although the dorsal side of this neuron was less densely innervated. Outside the axon-cap, boutons was covered by a layer of glial processes, forming a synaptic bed comparable to that observed around the goldfish M-cell,45 fish lateral vestibular nucleus30 and rat lateral vestibular neurons.50 Various types of profiles could be discriminated according to their vesicular content. They were found at any location around the cell body, except for the axoaxonic synapses (see above) which were only present within the axon-cap. Some profiles adjacent to the M-cell contained clear multimorphic vesicles (40– 80 nm) intermingled with large (90–100 nm) dense core vesicles (Fig. 2A) while others contained smaller (60–80 nm) dense core vesicles alone. These profiles were just adjacent to the M-cell surface and did not display synaptic differentiations. They may correspond to the so-called ‘‘immature’’ synapses previously described.22 However, in the absence of serial sections, we could not be sure whether or not any morphologically differentiated structure existed between the two elements.
Fig. 2. Ultrastructural characteristics of axonic profiles contacting the M-cell (M). All illustrated types of terminals were found over the entire neuronal surface, without regional preference. A) Ending with ‘‘multimorphic’’ large clear vesicles (arrow) and dense core vesicles (arrowhead). B) Axonic profile with numerous dense core (arrowheads) small clear (arrows) vesicles. C) Example of a well-defined synaptic complex (between bars); note the presynaptic accumulation of small clear spherical vesicles (arrows), the presynaptic dense projections (arrowhead), and the thin postsynaptic density. D) Another example of a morphologically mature synapse with a pleiomorphic population of vesicles in which the presynaptic dense projections are well visible (same symbols as in C). E–F) EPTA-stained semithin sections demonstrating the structural maturity of synaptic complexes (between bars in E), and an en face view showing that the boutons contain a single active zone (arrow in F). In E and F, the contours of the boutons are indicated by arrowheads. G–H) The M-cell surface is also reached by profiles bearing gap junctions (crossed arrows) and occasionally, small round vesicles (arrows) are also apparent. Note the fuzzy electron density (arrowheads) adjacent to the postsynaptic plasma membrane. Scale bars=0.2 µm.
138
A. Triller et al.
Fig. 3. Histogram of the number of presynaptic dense projections contained in individual presynaptic grids.
M-cells were also in relation to boutons containing spherical (Fig. 2C) or pleiomorphic small (30–40 nm) clear vesicles (Fig. 2D). They established synaptic contacts with the characteristic features of synaptic complexes including pre- and postsynaptic differentiations and an electron-dense material within the cleft. These junctions were specifically stained with EPTA impregnation, which allows visualization and assessment of the maturity of synaptic complexes.6,64 This procedure revealed that, at identified synapses, the presynaptic grids and postsynaptic differentiations were already structurally mature (Fig. 2E). Using semithin sections, ‘‘en face’’ views of the synaptic complex could be obtained and individual presynaptic dense projections in a grid could be seen. In most cases the whole synaptic grid area could be visualized in these 0.5 mm thick sections.58,64 However, in the 52-hour-old zebrafish M-cell, the extracellular space is not as electron-opaque as in the adult goldfish. As a consequence, it was not as easy to delineate the contour of the boutons as in the goldfish. None the less, it could be determined that the mean number of dense projections was 10.27&3.62 (n=105) per grid, and that their overall distribution was skewed (Fig. 3). To avoid any ambiguities, we considered that a grid is formed when it contains five or more dense projections compacted with a distance of less than 80 nm between each PDPs.54 We found that the presynaptic grids, on the zebrafish M-cell soma, were compact and not perforated as observed on the soma of adult goldfish M-cell.58 Of 96 boutons in which the contour could be clearly identified (see Fig. 2F), 89 contained a single active zone. Three boutons contained two active zones and the remaining four probably had two active zones, but the limits of the latter were uncertain. Profiles establishing gap junctions with the M-cell were also detected at the level of the axon hillock (Fig. 2G–H). Occasionally, a few round vesicles were present next to the presynaptic membrane suggesting that such boutons are involved in mixed, electrical and chemical transmission and may therefore belong to spiral fibre endings.38
Immunocytochemical studies on semithin sections revealed that boutons enriched in GABA or glycine are present over the M-cell soma and on the initial segment (Fig. 4A1, B1). This distribution contrasts with that observed on the adult M-cell of goldfish, where the inhibitory endings in the axon-cap were exclusively glycinergic.60 Electron microscopic analysis indicated that the synaptic density was 0.83 boutons per µm of membrane (165 boutons on 198 µm of membrane, from two sections separated by 5 µm). The proportions of glycine- and GABAimmunoreactive (IR) boutons, evaluated on these two non-serial sections, were similar and corresponded to 11% and 12%, respectively. In addition, glycine-IR (Fig. 4A2–3) and GABA-IR (Fig. 4B2–3) endings established mature synaptic complexes. Interestingly, some of the GABA-IR boutons were over spine-like protrusions (Fig. 4B2–3). Inhibitory postsynaptic currents Previously we showed that mIPSCs recorded in 1 µM TTX+10 mM Mg2+ were 95% blocked by 1 µM strychnine34 while bicuculline had no action on mIPSCs. One possible explanation is that spontaneous GABA events have a low probability of occurrence. If this proposition is true, one would expect the likelihood of detecting GABAergic mIPSCs to be enhanced by a procedure known to increase mIPSCs frequency. To test this hypothesis, we recorded synaptic activity in the presence of 1 µM TTX and 1.3 mM CaCl2 and applied, by bath perfusion, a solution containing 11.5 mM KCl. As expected, in the presence of this high [KCl]o external solution, the overall frequency of the synaptic events increased and ranged from 44 to 182 Hz, compared with 4 to 16 Hz in the controls (1.3 mM [KCl]o). Under these experimental conditions and in the presence of 1 µM strychnine, GABAergic events were isolated by their sensitivity to 20 µM bicuculline (a competitive GABAA antagonist). The application of 1 µM strychnine rapidly decreased the amplitude and frequency of mIPSCs and a stable level was reached within 2–3 min of application. Specifically, strychnine inhibited 90.1&8.4% (n=6 cells) of spontaneously occurring synaptic events (Fig. 5A). Subsequent addition of 20 µM bicuculline blocked 90% of the residual events, suggesting that they result from the activation of GABAA receptors (Fig. 5B). The few remaining events observed in the presence of strychnine and bicuculline were not analysed. Although the frequency of GABAergic mIPSCs is lower than that of glycinergic mIPSCs, basal activation of GABAA receptors might occur by continuous release of the neurotransmitter. Such activation of GABAA receptors would contribute to the fluctuations of background noise. Basal noise fluctuations were measured using point-per-point amplitude
Mature inhibitory synapses in zebrafish larvae hindbrain
139
Fig. 4. Morphological evidence that GABAergic and glycinergic terminals establish morphologically differentiated synaptic contact with the 52-hour-old M-cell. Glycine (A1–3) and GABA (B1–3) immunoreactivity on semithin (A1; B1) and ultrathin (A2–3; B2–3) sections. A1; B1) Examples of immunoreactive profiles (arrows) adjacent to the soma (s) and initial segment (double arrowheads). A2–3) Examples of glycine-IR profiles establishing differentiated synaptic complexes (between arrows). B2–3) Examples of GABA-IR profiles with synaptic contacts (between arrows), here over spine-like M-cell protrusions (crossed arrows). Scale bars, A1–B1:=10 µm; A2–3, B2–3=0.25 µm.
histograms34 during non-active periods. The standard deviation (SDm) of basal noise ranged from 2.9 to 5.6 in the presence of TTX plus strychnine (n=6). Adding bicuculline to the bath solution changed SDm values by at most "8 to 17%, depending on the cell tested (n=6). This result suggests that there is no significant basal activation of GABAA receptors on the zebrafish M-cell (P<0.05 paired t-test).
General properties of GABAergic events The bicuculline-sensitive synaptic events recorded in the presence of strychnine and TTX were considered as GABAergic miniature postsynaptic currents (mIPSCs). Their amplitude distributions were obtained from data sets of 180 to 822 events. In all cells tested (n=6), histograms could be well resolved into
140
A. Triller et al.
Fig. 5. Whole-cell recording of glycinergic and GABAergic miniature synaptic events. (A) Cumulative amplitude histogram of mIPSCs collected at a holding potential of "50 mV in the presence of a solution containing 1 µM TTX and 11.5 mM [KCl]o (m=4010/42 s) and of a solution containing 1 µM TTX, 11.5 mM [KCl]o and 1 µM strychnine (m=768/42 s). Note that strychnine strongly reduced the proportion of the largest events. Insert: four superimposed epochs (1 s each) of spontaneous postsynaptic activity in the presence of TTX and 11.5 mM [KCl]o before (upper panel) and during (lower panel) strychnine applications. B) Amplitude histogram of strychnine-resistant mIPSCs before (m=768/42 s) and during application of 20 µM bicuculline (m=14/42 s) (bin=2 pA). The amplitude histogram was well fitted by the sum of two Gaussian curves with means and S.Ds of 15.9&4.6 pA and 28.9&10.6 pA, respectively. The insert shows four superimposed epochs before (upper panel) and during (lower panel) the perfusion of bicuculline which dramatically reduced the number of residual strychnine-resistant mIPSCs. A and B were obtained from the same cell.
one or two Gaussians (see Experimental Procedures). The second class was detected only in histograms including large data sets (n=3). As illustrated in Fig. 5B, these two classes were fitted by the sum of two
Gaussian curves. The mean amplitude of the first class was close to 18 pA (17.9&3.3 pA; n=6; Vh="50 mV) and it was close to 34 pA for the second class (33.9&6.8 pA; n=3).
Mature inhibitory synapses in zebrafish larvae hindbrain
Fig. 6. (A) Example of the distribution of time to peak of strychnine resistant mIPSCs recorded at "50 mV (60 s duration: bin=0.1 ms, n=420). The distribution was well fitted by a single Gaussian curve with a mean and S.D. of 0.71&0.2 ms. The insert shows that measurements were made between the onset and the maximum of the mIPSCs. B) Distribution of the decay time-constants (same cell as in A) fitted with a single exponential. In this example the mean decay time-constant was 4.73&1.17 ms. Insert: 10 superimposed strychnine-resistant mIPSCs (upper traces) and their average (lower trace).
As for glycinergic mIPSCs,34 GABAergic mIPSCs had fast kinetics. Time to peak (ttp) distributions were well fitted by a single Gaussian curve (Fig. 6A) giving a mean ttp value of 0.62&0.12 ms (n=6 cells). The value of the decay time-constant (ôd) of the strychnine-resistant mIPSCs was obtained by fitting the first 10 ms of the IPSCs repolarizing phase fitted by a single exponential curve. ôd ranged from 2 to 30 ms (Fig. 6B). Although the decay time-constant histogram is slightly skewed, one predominant class clearly emerged (Fig. 6B). This class ranged from 2 to
141
Fig. 7. Outside-out recordings of single-channel currents evoked by GABA and glycine applications (A) Effect of 20 µM GABA and 5 µM glycine applications to the same patch during outside-out recordings (Vh="50 mV; filter 1 KHz). Note that glycine evoked larger opening states than GABA. (B) Point-per-point amplitude histogram of a data segment (shown in the insert) during 20 µM GABA application (bin 0.01 pA). The mean amplitude of GABA-evoked opening states was obtained by fitting the amplitude distribution with Gaussian curves. In this cell the mean amplitude of GABA-evoked opening states was 1.15&0.38 pA, indicating a main conductance state of 21 pS (reversal potential=0 mV).
8 ms and was well fitted by a Gaussian curve, with a mean of 3.67&0.23 ms (n=6 cells). There was no correlation between ttp, ôd and mIPSCs amplitude in any cell tested (P<0.15). GABA-evoked single-channel openings in outside-out recordings The size of mIPSPs depends on the amplitude of the main conductance state of the activated channel and the number of receptors activated after an exocytosis. In order to study the conductance properties of GABAA receptor-channels, openings were induced by the application of 20 µM GABA to outside-out patches excised from the M-cell.
142
A. Triller et al.
At a holding potential of "50 mV, GABA applications evoked single-channel openings in all patches tested. Applications of 5 µM glycine to the same patches (n=5) evoked single-channel events with an amplitude larger than that produced by GABA (Fig. 7A). No kinetic analysis was performed on these events. The GABA-activated channels had a single conductance state which was measured using point-per-point histograms (Fig. 7B). At "50 mV, this conductance state had an averaged value of 22.7&2.2 pS (n=5 patches) for an observed reversal potential close to 0 mV. This value is two to four times smaller than that previously described34 for glycine receptor-channels (41–43 pS and 81–86 pS). It allowed us to calculate the minimal number of GABA-gated channels simultaneously activated after exocytosis, assuming that synaptic and extrasynaptic GABAA receptors are functionally identical.8 The amplitude histograms of mIPSCs suggest that it ranges from 14 to 20 channels, which is close to that obtained for GABA synapses in other species8,32,63 and for glycine synapses on zebrafish larvae M-cells.34 DISCUSSION
Identification of the Mauthner cell One of the elements which allows unambiguous identification of the M-cell is the axon-cap (AC). It is a specialized neuropil which surrounds the initial segment of the M-cell of teleosts, and which has been characterized by both light and electron microscopy.38,39,57 It has been shown that the AC is directly involved in the generation of an electrical field effect, that is of the electrical inhibition which results from the channelling across the membrane of presynaptic terminals of M-cell action potential currents due to the high electrical resistance of the extracellular tissue in the AC.25 Three structural features are specific to this neuropil: i) it is surrounded by glial elements, the cap cells; ii) axons spiral around the initial segment and establish axoaxonic symmetrical synapses between themselves with a pseudo-glomerular structure; iii) in the peripheral part of the AC, axons and glial cells are tightly linked by septate-like junctions in goldfish and in tench.57 In 52-hour-old zebrafish larvae we could identify the latter two features unambiguously but a large number of free polyribosomes were also observed, this last feature being commonly observed in immature neurons.43 Such observations confirm that the M-cell system is not completely mature at this stage.22 Morphological analysis of synaptic contacts on 52hour-old zebrafish Mauthner cells Our morphological studies show that the M-cell soma and initial portions of dendrites contact synaptic endings with the structural characteristics of
mature synapses: the presynaptic elements contain small clear, round or pleiomorphic vesicles and presynaptic differentiations and are adjacent to postsynaptic differentiations. Also, and as in the adult, other profiles containing round vesicles establish gap junctions with the M-cell and are found on the initial segment as well as on the soma. We could not identify in the zebrafish, as in goldfish,58 all the types of endings which were stained, but it was possible to study their active zones. The mean number of PDPs for all synaptic boutons analysed was 10, that is slightly less than that reported (n=13) at the glycinergic unmyelinated club-endings of the goldfish.58 This number should parallel the maximum number of vesicles surrounding the PDPs and adjacent to the plasma membrane which was estimated to be 40 to 60 in the goldfish. The spiral fibres and the unmyelinated club endings described by Nakajima38 in the axon cap are already present in the teleost larvae. Outside this region, precise identification of the terminals involved was more difficult, mainly because axons are not myelinated at this stage of development. However, two types of boutons could be distinguished according to the shape of their synaptic vesicles i.e. round and flat/pleiomorphic. They could correspond to ‘‘large vesicle boutons’’ and ‘‘small myelinated club endings’’ for the first group and to the ‘‘small vesicle boutons’’ for the second. Other profiles adjacent to the zebrafish M-cell membrane were less easily correlated with classes of endings known to contact the adult goldfish M-cell. Some boutons were filled with medium-sized dense core vesicles, and others with large clear vesicles of variable sizes and with large dense core vesicles. Neither of the two classes established differentiated synaptic complexes. The medium dense core vesicles were comparable in size to those generally associated with catecholamine.15 The large dense core vesicles were similar to those present in somatostatincontaining fibres that have been described around the adult M-cell53 or to those found in rat dorsal horn and which contain neurotensin.56 Whether or not the large multimorphic vesicles correspond to immature contacts remains to be determined. Although they could also result from fixation artefacts it should be pointed out that at this stage the M-cell measures about 100 µm from the tip of its ventral dendrite to that of its lateral one, with a diameter of about 6–10 µm. Therefore, the M-cell has not yet reached its adult size (300 µm in length and diameter 30– 40 µm). Since the covering ratio of the adult zebrafish M-cell appears to be relatively high,21,22 it is likely that new synapses are still being formed at this stage of maturation. GABAergic miniature inhibitory postsynaptic currents With immunocytochemistry, we found that a similar proportion of glycine- and GABA-IR profiles are
Mature inhibitory synapses in zebrafish larvae hindbrain
present on the soma, outside and within the AC. This contrasts with the goldfish where glycine-containing profiles predominate in the AC and where GABAergic profiles are only observed outside this region.52 Whole-cell recordings indicated that these GABAergic synaptic boutons are functional. But as in goldfish, M-cell inhibitory synaptic events are mainly glycinergic34 due to a lower probability of release and/or to a still incomplete maturation of the release machinery at GABAergic terminals (see below). The low frequency of GABAergic events observed even in the presence of a high K+ extracellular solution can therefore explain why they have not been detected, so far, in adult goldfish and in previous studies of zebrafish larvae.34,35 Most GABAergic mIPSCs (95%) recorded from the zebrafish M-cell have a short decay phase (4– 5 ms) which can be fitted by a single exponential curve. It is unlikely that the few slower mIPSCs (ôd=15–25 ms) reflect space-clamp error since there is no correlation between their ttp, decay time-constant or amplitude.44 These GABAergic IPSCs have kinetics similar to those observed in other species and on adult neurons. GABAergic IPSCs have a fast or slow decay phase depending on the cell tested8,19,32,41,42,49,63 and these differences might result from the activation of different GABA receptorsubtypes expressed by the same neuron.42,49 Further, slow events might reflect the activation of particular GABAA receptors with desensitized states prolonging GABAA channel responses to neurotransmitter release.18 The single conductance state value (23 pS) of the GABA-gated channels is similar to that observed in other species.8,16,32,40 This value may vary according to the subunit combination of the receptor channels.2,61,62 Although many GABAA receptor subunits have been cloned, single-channel properties of only a few of the many possible combinations have been analysed in heterologous expression systems.2,61,62 Moreover, GABAA receptor subunits from zebrafish brain have not yet been cloned. Therefore, it is not possible to define which combination of subunits is expressed by the M-cell. Miniature inhibitory postsynaptic currents activity Although we did not make detailed measurements on the proportion of glycine- or GABA-IR boutons, visual inspection suggested that their numbers do not differ greatly. In contrast, the frequency of GABAergic mIPSCs was significantly lower than that of glycinergic events. This might depend on differences in the post- and/or presynaptic mechanisms. A first hypothesis would be that some synapses with boutons that contain GABA either do not express postsynaptic receptors or express them at such a low density that synaptic events cannot be generated. This is probably not the case since GABA-
143
gated channel activity was detected in all patches tested during outside-out recordings. In mammals, even immature synapses possess active receptors11 and postsynaptic maturation corresponds essentially to a change in receptor subtypes expressed7 rather than to changes in their density.31 Alternatively, the low frequency of GABAergic mIPSCs might result from differences in the presynaptic mechanism leading to vesicle exocytosis. Synaptic boutons with different release probabilities have been described in mammals.46 Even when neurotransmitters are co-localized, as for GABA and glycine in synapses on cerebellar granule cells,20 synaptic boutons appear to release only one type of molecule (GABA on granule cells). Differences in release probability between GABAergic and glycinergic boutons could also reflect differential expression of calcium-channel subtypes between synapses during development, as recently proposed in hippocampal neurons.20 In the 52-hour-old zebrafish M-cell, GABA mIPSCs have less complex amplitude distributions than the multimodal histograms of glycinergic events.34 The presence, in the vicinity of the M-cell, of growth cone-like profiles containing large vesicles could account for the amplitude fluctuation of glycinergic mIPSCs. In fact, neurotransmitters are released from growth cone terminals,1,65 but the release probability and the number of neurotransmitter molecules issued at immature synapses are very low.1,65 Since one active zone can release at most one vesicle at mature synapses,29,58 the number of presynaptic active zones per bouton might, in principle, have accounted for the difference between the amplitude distribution histograms of glycinergic and GABAergic mIPSCs recorded in our preparation. mIPSC amplitude distributions recorded in the soma of the adult goldfish M-cell are unimodal while they are multimodal for mIPSCs collected from the lateral dendrite.23 Accordingly, 96% of the unmyelinated endings and small vesicle boutons contain a single active zone in the soma and in the axon-cap of the adult goldfish, while 30% of the small vesicle boutons contain two or more presynaptic grids on the distal half of the lateral dendrite.54 However, in the zebrafish M-cell 92% of the endings adjacent to the soma or within the axon-cap also contain a single active zone, independently of their classification or transmitter content. Therefore, the difference between the shape of the amplitude distributions of glycinergic and GABAergic mIPSCs is not related to differences in the number of active zones per terminal. On the other hand, a multivesicular release occurring at these newly formed synapses cannot be completely excluded since the mean value of the second class of the GABAergic mIPSCs is twice that of the first one. The skewness of the GABAergic mIPSCs amplitude histogram may also result from a large variation in the number of neurotransmitter
144
A. Triller et al.
molecules released per exocytosis, as recently proposed for GABA mIPSCs in retinal cells.9 CONCLUSION
In conclusion, differences in the amplitude distributions of glycinergic and GABAergic mIPSCs are most likely due to postsynaptic factors. As previously suggested for GABA mIPSCs in mammals,8 the multimodal amplitude histograms of glycinergic mIPSCs could result from the presence of variable
proportions of different glycine receptor types from one synapse to another while the number and the identity of postsynaptic GABAA receptors could be less variable. To be tested, this hypothesis will require investigations using quantified confocal fluorimetric measurements as developed for the adult goldfish M-cell.47,59
Acknowledgements—This work was supported by INSERM and Associaton Franc¸aise pour la Myopathie.
REFERENCES
1. 2. 3. 4. 5. 6. 7. 8. 9. 10. 11. 12. 13. 14. 15. 16. 17. 18. 19. 20. 21. 22. 23. 24. 25. 26. 27. 28. 29.
Allen T. G. J. and Brown D. A. (1996) Detection and modulation of acetylcholine release from neurites of rat basal forebrain cells in culture. J. Physiol., Lond. 492, 453–466. Angelotti T. P. and MacDonald R. L. (1993) Assembly of GABAA receptor subunits: á1â1 and á1â1ã2s subunits produce unique ion channels with dissimilar single-channel properties. J. Neurosci. 13, 1429–1440. Ankri N., Legendre P., Faber D. S. and Korn H. (1994) Automatic detection of spontaneous synaptic responses in central neurons. J. Neurosci. Meth. 52, 87–100. Bekkers J. M., Richerson G. B. and Stevens C. F. (1990) Origin of variability in quantal size in cultured hippocampal neurons and hippocampal slices. Proc. natn. Acad. Sci. U.S.A. 87, 5359–5362. Bekkers J. M. and Stevens C. F. (1990) Presynaptic mechanism for long-term potentiation in the hippocampus. Nature 346, 724–729. Bloom F. E. and Aghajanian G. K. (1968) Fine structural and cytochemical analysis of the staining of synaptic junctions with phosphotungstic acid. J. ultrastruct. Res. 22, 361–375. Dunn S. M. J., Bateson A. N. and Martin I. L. (1994) Molecular neurobiology of the GABAA receptor. Int. Rev. Neurobiol. 36, 51–96. Edwards F. A., Konnerth A. and Sakmann B. (1990) Quantal analysis of inhibitory synaptic transmission in the dentate gyrus of rat hippocampal slices: a patch clamp study. J. Physiol., Lond. 430, 213–249. Frerking M., Borges S. and Wilson M. (1995) Variation in GABA mini amplitude is the consequence of variation in transmitter concentration. Neuron 15, 885–895. Ga¨hwiler B. H. and Brown D. A. (1985) GABAB-receptor-activated K+ currents in voltage-clamped CA3 pyramidal cells in hippocampal cultures. Proc. natn. Acad. Sci. U.S.A. 8, 1558–1562. Grantyn R., Kraszewski K., Melnick I., Taschenberger H. and Warton S. S. (1995) In vitro development of vertebrate central synapses. Perspect. dev. Biol. 2, 387–397. Gulyas A. I., Miles R., Sik A., Toth K., Tamamaki N. and Freund T. F. (1993) Hippocampal pyramidal cells excite inhibitory neurons through a single release site. Nature 366, 683–687. Hamill O. P., Marty A., Neher E., Sakmann B. and Sigworth F. J. (1981) Improved patch clamp techniques for high-resolution current recordings from cells and cell free patches. Pflu¨gers Arch. 391, 85–100. Horokawa N. K., Sobue K., Kanda K., Harada A. and Yorifuji H. (1989) The cytoskeletal architecture of the presynaptic terminal and molecular structure of synapsin I. J. Cell Biol. 108, 111–126. Ho¨kfelt T. (1991) Neuropeptides in perspective: the last ten years. Neuron 7, 867–879. Isaacson J. S., Solis J. M. and Nicoll R. A. (1993) Local and diffuse synaptic actions of GABA in the hippocampus. Neuron 10, 165–175. Johnston G. A. R. (1991) GABAA antagonists. Semin. Neurosci. 3, 205–210. Jones M. V. and Westbrook G. L. (1995) Desensitized states prolong GABAA channel responses to brief agonist pulses. Neuron 15, 181–191. Kaneda M., Farrant M. and Cull Candy S. G. (1995) Whole cell and single-channel currents activated by GABA and glycine in granule cells of the rat cerebellum. J. Physiol., Lond. 485, 419–435. Kenneth P., Sholz P. and Miller R. J. (1995) Developmental changes in presynaptic calcium channels coupled to glutamate release in cultured rat hippocampal neurons. J. Neurosci. 15, 4612–4617. Kimmel C. B. and Model P. G. (1978) Developmental study of the Mauthner cell. In Neurobiology of the Mauthner Cell (eds Faber D. S. and Korn H.), pp. 183–220. Raven Press, New York. Kimmel C. B., Sessions S. K. and Kimmel R. J. (1981) Morphogenesis and synaptogenesis of the zebrafish Mauthner neuron. J. comp. Neurol. 198, 101–120. Korn H., Bausela F., Charpier S. and Faber D. S. (1993) Synaptic noise and multiquantal release at dendritic synapses. J. Neurophysiol. 70, 1249–1254. Korn H., Burnod Y. and Faber D. S. (1987) Spontaneous quantal currents in a central neuron match predictions from binomial analysis of evoked responses. Proc. natn. Acad. Sci. U.S.A. 84, 5981–5985. Korn H. and Faber D. S. (1975) An electrically mediated inhibition in goldfish medulla. J. Neurophysiol. 38, 452–471. Korn H. and Faber D. S. (1990) Transmission at a central inhibitory synapse: IV. Quantal structure of synaptic noise. J. Neurophysiol. 63, 198–222. Korn H. and Faber D. S. (1991) Quantal analysis and synaptic efficacy in the CNS. Trends Neurosci. 14, 439–445. Korn H., Faber D. S. and Triller A. (1990) Convergence of morphological, physiological, and immunocytochemical techniques for the study of single Mauthner cells. In Handbook of Chemical Neuroanatomy. (eds Bjo¨rklund A., Ho¨kfelt T., Wooterlood F. G. and van der Pol A. N.), Vol. 8, pp. 403–480. Elsevier Science Publishers, Amsterdam. Korn H., Mallet A., Triller A. and Faber D. S. (1982) Transmission at a central inhibitory synapse. II. Quantal description of release, with a physical correlate for binomial n. J. Neurophysiol. 48, 679–707.
Mature inhibitory synapses in zebrafish larvae hindbrain 30. 31. 32. 33. 34. 35. 36. 37. 38. 39. 40. 41. 42. 43. 44. 45. 46. 47. 48. 49. 50. 51. 52. 53. 54. 55. 56. 57. 58. 59. 60. 61. 62. 63. 64. 65.
145
Korn H., Sotelo C. and Bennet M. V. L. (1977) The lateral vestibular nucleus of the toadfish Opsanus tau: ultrastructural and electrophysiological observations with special references to electrotonic transmission. Neuroscience 2, 851–884. Kraszewski K. and Grantyn R. (1992) Development of Gabaergic connections in vitro: increasing efficacy of synaptic transmission is not accompanied by changes in miniature currents. J. Neurobiol. 23, 766–781. Kraszewski K. and Grantyn R. (1992) Unitary, quantal and miniature GABA-activated synaptic chloride currents in cultured neurons from the rat superior colliculus. Neuroscience 47, 555–570. Landis D. M. D., Hall K., Weinstein L. A. and Reese T. S. (1988) The organisation of the cytoplasm at the presynaptic active zone of a central nervous system synapse. Neuron 1, 201–209. Legendre P. and Korn H. (1994) Glycinergic inhibitory synaptic currents and related receptor-channel in the zebrafish brain. Eur. J. Neurosci. 6, 1544–1557. Legendre P. and Korn H. (1995) Voltage-dependence of conductance changes evoked by glycine release in the zebrafish brain. J. Neurophysiol. 6, 2404–2412. Manabe T., Renner P. and Nicoll R. A. (1992) Postsynaptic contribution to long-term potentiation revealed by the analysis of miniature synaptic currents. Nature 355, 50–55. Min M. Y. and Appenteng K. (1996) Multimodal distribution of amplitudes of miniature and spontaneous EPSCs recorded in rat trigeminal motoneurones. J. Physiol., Lond. 494, 171–182. Nakajima Y. (1974) Fine structure of the synaptic endings on the Mauthner cell of goldfish. J. comp. Neurol. 156, 375–402. Nakajima Y. and Kohno K. (1978) Fine structure of the Mauthner cell: synaptic topography and comparative study. In Neurobiology of the Mauthner cell. (eds Faber D. S. and Korn H.), pp. 133–166, Raven Press, New York. Newland C. F., Colquhoun D. and Cull-Candy S. G. (1991) Single channels activated by high concentrations of GABA in superior cervical ganglion neurones of the rat. J. Physiol., Lond. 432, 203–233. Otis T. S. and Mody I. (1992) Modulation of decay kinetics and frequency of GABAA receptor-mediated spontaneous inhibitory postsynaptic currents in hippocampal neurons. Neuroscience 49, 13–32. Pearce R. A., Grunder S. D. and Faucher L. D. (1995) Different mechanisms for use-dependent depression of two GABAA-mediated IPSCs in rat hippocampus. J. Physiol., Lond. 484, 425–435. Peters A. and Palay S. L. and de Webster F. H. (1991) The fine structure of the nervous system. Neurons and their Supporting Cells. 3rd edn. Oxford University Press, New York and Oxford. Rall W. (1969) Time constants and electrotonic length of membrane cylinders and neurons. Biophys. J. 9, 1483–1508. Robertson J. D., Bodenheimer T. S. and Stage D. E. (1963) The ultrastructure of Mauthner cell synapses and nodes in the goldfish brain. J. Cell Biol. 19, 159–199. Rosenmund C., Clements J. D. and Westbrook G. L. (1993) Nonuniform probability of glutamate release at a hippocampal synapse. Science 262, 754–757. Seitanidou T., Triller A. and Korn H. (1992) Partial glycinergic innervation induces transient changes in the distribution of a glycine receptor-associated protein in a central neuron. J. Neurosci. 12, 116–131. Silver R. A., Traynelis S. F. and Cull-Candy S. G. (1992) Rapid time-course miniature and evoked excitatory currents at cerebellar synapses in situ. Nature 355, 163–166. Soltesz I., Smetters D. K. and Mody I. (1995) Tonic inhibition originates from synapses close to the soma. Neuron 14, 1273–1283. Sotelo C. and Palay S. L. (1972) The fine structure of the vestibular nucleus in the rat. II Synaptic organisation. Brain Res. 18, 93–115. Stevens C. F. (1993) Quantal release of neurotransmitter and long term potentiation. Cell 72/Neuron 10 (Suppl.), 55–63. Sur C., McKernan R. and Triller A. (1995) GABAA receptor-like immunoreactivity in the goldfish brainstem with emphasis on the Mauthner cell. Neuroscience 66, 697–706. Sur C., Triller A. and Korn H. (1994) Colocalisation of somatostatin with GABA or glutamate in distinct afferent terminals presynaptic to the Mauthner cell. J. Neurosci. 14, 576–589. Sur C., Triller A. and Korn H. (1995) Morphology of the release site of inhibitory synapses on the soma and dendrite of an identified neuron. J. comp. Neurol. 351, 247–260. Tang C. M., Margulis M., Shi Q. Y. and Fielding A. (1994) Saturation of postsynaptic glutamate receptors after quantal release of transmitter. Neuron 13, 1385–1393. Todd A. J., Spike R. C., Price R. F. and Neilson M. (1994) Immunocytochemical evidence that neurotensin is present in glutamatergic neurons in the superficial dorsal horn of the rat. J. Neurosci. 14, 774–784. Triller A. and Korn H. (1980) Glio-axonic junctional like complexes at the Mauthner cell’s axon cap of teleosts: a possible morphological basis for field effect inhibitions. Neurosci. Lett. 18, 275–281. Triller A. and Korn H. (1982) Transmission at a central inhibitory synapse. III Ultrastructure of physiologically identified and stained terminals. J. Neurophysiol. 48, 708–736. Triller A., Seitanidou T., Frankson O. and Korn H. (1990) Size and shape of glycine receptor clusters in a central neuron exhibit a somato-dendritic gradient. The New Biologist 2, 637–641. Triller A., Sur C. and Korn H. (1993) Heterogeneous distribution of glycinergic and GABAergic afferents on an identified central neuron. J. comp. Neurol. 339, 83–96. Verdoorn T. A., Draguhn A., Ymer S., Seeburg P. H. and Sakmann B. (1990) Functional properties of recombinant GABAA receptors depend on subunit composition. Neuron 4, 919–928. Verdoorn T. A. (1994) Formation of heteromeric ã-aminobutyric acid type A receptors containing two different a subunits. Molec. Pharmac. 45, 475–480. Virginio C., Martina M. and Cherubini E. (1995) Spontaneous GABA-mediated synaptic currents in cerebellar granule cells in culture. NeuroReport 6, 1285–1289. Vrensen G., Nunes Cardozo J., Muller L. and Van Der Want J. J. L. (1980) The presynaptic grid: a new approach. Brain Res. 184, 23–40. Young S. H. and Poo M. M. (1983) Spontaneous release of transmitter from growth cones of embryonic neurones. Nature 305, 634–637. (Accepted 13 February 1997)
PII:
Neuroscience Vol. 80, No. 1, pp. 133–145, 1997 Copyright ? 1997 IBRO. Published by Elsevier Science Ltd Printed in Great Britain. All rights reserved 0306–4522/97 $17.00+0.00 S0306-4522(97)00092-4
MORPHOFUNCTIONAL EVIDENCE FOR MATURE SYNAPTIC CONTACTS ON THE MAUTHNER CELL OF 52-HOUR-OLD ZEBRAFISH LARVAE A. TRILLER,* P. ROSTAING,* H. KORN† and P. LEGENDRE†‡§ *CJF 94-10 INSERM, Ecole Normale Supe´rieure, 46 rue d’Ulm, 75005, Paris, France †Laboratoire de Biologie Cellulaire et Mole´culaire (INSERM U 261), De´partement des Biotechnologies, Institut Pasteur, 25 rue du Dr. Roux, 75015, Paris, France Abstract––In a previous study, miniature inhibitory synaptic events recorded in the Mauthner cell of the 52-hour-old zebrafish larvae (Brachydanio rerio) were found to be mainly glycinergic. Their amplitude distribution was not Gaussian and it was proposed that their large amplitude variation might reflect the activation of immature synapses. However, ultrastructural studies of the synaptic contacts over the M-cell soma of 52 h larvae described here, revealed that numerous synaptic contacts on this neuron are already mature at this developmental stage and that most of them already contain a single active zone. As in the adult goldfish, immunohistochemistry indicates the presence of both glycine- and GABA-immunoreactive boutons which establish synaptic contacts. We also found that, in addition to the predominant glycinergic postsynaptic inhibitory currents, some postsynaptic currents are also GABAergic since they are specifically inhibited by bicuculline (20 µM). GABAergic miniature events (time to peak close to 0.8 ms and decay time-constant close to 4–5 ms) were only detected in the presence of 11.5 mM [KCl]o. Their amplitude distributions were well fitted by one, or at most two, Gaussian curves. Outside-out recordings showed one class of GABA receptors with a main conductance state of 23 pS. This indicates that the smallest GABAergic miniature inhibitory synaptic events correspond to the opening of 14–20 chloride channels Pre- and postsynaptic factors which contribute to the predominance of glycinergic synaptic currents over GABAergic ones in untreated preparations and to the striking differences between their frequencies and their respective amplitude distribution histograms are discussed with reference to the morphological characteristics of the mature synaptic endings impinging on this still developing neuron. ? 1997 IBRO. Published by Elsevier Science Ltd. Key words: inhibitory synapses, GABAA receptors, glycine receptors, miniature synaptic events, electron microcopy, patch-clamp.
Excitatory and inhibitory synaptic activity has been extensively studied in many neurons27,51 including the Mauthner cell (M-cell) of teleosts.29 In this neuron, excitatory and inhibitory synaptic contacts have been morphologically characterized and their respective locations are stereotyped in the goldfish (Carassius auratus),28,38 as well as in the zebrafish (Brachydanio rerio).21,22 Interestingly, the spontaneous synaptic events recorded from the M-cells of both adult goldfish and zebrafish larvae are almost completely inhibitory. Miniature inhibitory postsynaptic currents (mIPSCs), which are mainly glycin§To whom correspondence should be addressed. ‡Present address: Institut des Neurosciences, CNRS URA 1488, Universite´ Pierre et Marie Curie, 7 Quai Saint Bernard, 75005 Paris, France. Abbreviations: AC, axon-cap; EGTA, ethyleneglycolbis (aminoethylether)tetra-acetate; EPTA, ethanolic phosphotungstic acid; HEPES, N-2-hydroxyethylpiperazineN*-2-ethanesulphonic acid; IR, immunoreactive; M-cell, Mauthner cell; mIPSC, miniature inhibitory postsynaptic current; PBS, phosphate-buffered saline; PDP, presynaptic dense projection; TTX, tetrodotoxin; ttp, time-topeak.
ergic, have similar kinetics in both species (time to peak <1 ms and average decay time-constant ranging from 4 to 7 ms;26,34). In the goldfish, the histograms of miniature events could be described by a single Gaussian curve.24. In contrast, in 52-hour-old zebrafish larvae the amplitudes of miniature events ranged from 20 to 500 pA and their distribution consisted of the sum of Gaussian curves, which were not equally spaced.34 This type of distribution is similar to that reported for both adult and developing mammal brains, where amplitude distribution histograms of inhibitory and excitatory miniature events are often skewed5,8,12,34,36,48,55 and, in some structures, clearly multimodal.8,34,37,55 It is notable that in most cases, these amplitude distributions cannot result from cable filtering properties of the cell membrane.4 It has been shown in the goldfish M-cell that glycinergic boutons operate in an all-or-none manner.29 In this system, presynaptic boutons localized on the M-cell soma contain a single grid58 and physiological data suggest that the content of no more than one vesicle can be released at each active
133
134
A. Triller et al.
zone29 since, for example, the amplitude histograms of miniature inhibitory events are unimodal. In contrast, amplitude histograms were multimodal in the presence of tetrodotoxin (TTX) when synaptic events were recorded in the lateral dendrite,23 where synaptic boutons have a variable number (one to four) of active zones.54 Thus, since the hindbrain is not completely developed in 52-hour-old zebrafish larvae,21,22 the widespread amplitude histograms of glycinergic mIPSCs recorded in the M-cell34 could result, at this stage, from the presence of immature synaptic contacts and/or the presence of a variable number of active zones per bouton which may lead to multivesicular release.23 In the present study, we used ultrastructural and physiological methods to characterize further the general properties of the inhibitory system which controls M-cell excitability in the zebrafish larvae. Our ultrastructural studies suggest that a large number of synapses are already structurally mature in 52-hour-old zebrafish larvae and have already a single active zone per bouton. Furthermore, immunohistochemistry studies indicated that in addition to glycinergic terminals, immunoreactive boutons apposed to the M-cell soma can be GABAergic. This finding prompted us to analyse the properties of the miniature synaptic events they generate in the M-cell. EXPERIMENTAL PROCEDURES
Morphological methods Fifty-two-hour-old zebrafish larvae were fixed by immersion in 4% glutaraldehyde in phosphate-buffered saline (PBS, O.12 M, pH 7.4) for 30 min at room temperature, and were kept overnight in the same fixative at 4)C. After three rinses in PBS, whole larvae were dipped in OsO4 (2%) for 2 h at 4)C, washed, and stained en bloc in uranyl acetate. They were then dehydrated and embedded in Araldite. Ultrathin (pale yellow) sections were stained with uranyl and lead citrate. For the identification of presynaptic grids the larvae were fixed as above, washed in PBS and kept for 2 h in ethanolic phosphotungstic acid (EPTA) at 60)C as described elsewhere.58 Semithin (0.5 µm) sections were directly observed under the electron microscope without further staining. With this method, the extracellular space was weakly electron-dense, while the cytoplasm was almost transparent to electrons, thus allowing the contours of presynaptic boutons to be clearly distinguished. One of the main advantages of this approach is that it provides en face views of semithin sections so both presynaptic grids and postsynaptic differentiations can be visualized. Moreover, the number of presynaptic dense projections (PDPs) per grid can be determined. PDPs are not considered as a structure in themselves and they may, instead, correspond to a collapse of the cytoskeleton and of the proteins involved in release between vesicles adjacent to the presynaptic releasing membrane.14,33 For immunocytochemistry, animals were fixed in a mixture of 2% paraformaldehyde and 2% glutaraldehyde for 1 h at room temperature and then kept overnight at 4)C. Following osmification and dehydration, they were included in Durcupan. Semithin (1 µm) and ultrathin (pale yellow) sections were immunostained to detect glycine (antiglycine; Neosystem Lab. Strasbourg, France) and GABA (antiGABA; Immunotech SA, Marseille France). Briefly (for more details see Ref. 60), sections etched for 7 min with 1% orthoperiodic acid in distilled water were incubated
with primary antibodies (anti-glycine 1:200 or anti-GABA 1:100). The antiglycine was mixed with a conjugate of GABA and glutaraldehyde (GABA-G) and the anti-GABA with a conjugate of glycine and glutaraldehyde (Gly-G). Each amino acid conjugate was used at 50 µM. For glycine and GABA, the corresponding immunoreactivity was abolished by preadsorption of the antibody with the corresponding amino acid conjugated with glutaraldehyde. Controls without primary antibodies gave no specific staining. Antibody binding sites were detected with an ABC–HRP kit (Elite Vectastain*, Vector Laboratories, Burlingame, CA) for semithin sections and with goat anti-rabbit immunoglobulins coupled to 15 nm colloidal gold particles (1:50, BioCell, UK) for the electron microscope. Physiological methods Isolated brain preparation. Isolated zebrafish brain was prepared as described before.34 Briefly, the brains of newly hatched larva were dissected out under binocular observation and then glued to a coverslip using a plasma thrombin embedding procedure.10 Before experiments, brain preparations were kept for 15–30 min in an oxygenated (95%O2, 5%CO2) bathing solution containing (in mM) NaCl 145; KCl 1.5; CaCl2 2; MgCl2 1; NaHCO3 26; NaH2PO4 1.25; glucose 10, with osmolarity adjusted to 330 mOsm. Outside-out and whole-cell recordings. Standard whole-cell and outside-out recordings13 were made on the M-cells of mounted brains placed on a microscope stage (Nikon Optiphot). The preparation was continuously perfused at room temperature (20)C) with the oxygenated bathing solution (2 ml/min) in the recording chamber (0.5 ml). Patch-clamp electrodes were pulled either from thin-wall (whole-cell recordings) or thick-wall (outside-out recordings) borosilicate glass. They were fire-polished and filled with (in mM): CsCl 135; MgCl2 2; NaATP 4; EGTA 10; HEPES 10; pH 7.2. The osmolarity was adjusted to 290 mOsm. Whole-cell and outside-out recordings were made using 1–3 MÙ and 10–15 MÙ electrodes, respectively. Outsideout patches were obtained by slowly pulling the pipettes out of the brain. Currents were recorded using an Axopatch 1D (Axon instruments), filtered at 10 KHz, and stored using a digital tape recorder (DAT DTR 1201, SONY). During whole-cell recordings, the series-resistance (4–10 MÙ) was monitored by applying 2 mV hyperpolarizing pulses and 50–70% compensated. To ensure cell dialysis, measurements were made on data obtained at least 3–5 min after the whole-cell configuration was established.34 Drug delivery. Drugs were applied to the preparation via an array of four flowpipes of 400 µm diameter positioned above the brain. TTX (1 µM), strychnine (1 µM) and bicuculline methiodide (20 µM) were dissolved in a perfusate containing (in mM) NaCl 135; KCl 11.5; CaCl2 2; MgCl2 1; glucose 10 and HEPES 10; pH 7.2; osmolarity 330 mOsm. A high KCl solution (11.5 mM) was used to enhance mIPSC frequency in the presence of TTX. All solutions were prepared daily. Bicuculline methiodide (Sigma) was always freshly prepared as this compound is rapidly degraded in ionic solutions.17 Analysis of whole-cell currents. Ongoing synaptic activity was digitized off-line with a Macintosh IICi computer at 24 KHz using MMII software (GW instruments). Synaptic events were automatically detected, as previously described,3,34,35 with home-made software written in Labview II.2 (National Instruments). Briefly, the signals were filtered at 2 KHz, using a second-order Bessel digital filter. The peak of each response was measured as the maximal amplitude deviation from the baseline. The decay time-constant
Mature inhibitory synapses in zebrafish larvae hindbrain
Fig. 1. Cytological features of the 52-hour-old zebrafish Mauthner cell and axon-cap. A) Transverse section through the soma (S) of the neuron at the level of its nucleus (n) with the characteristic ventral and lateral dendrites (VD and LD, respectively). The axon initial segment (IS) is oriented toward the midline and is surrounded by a specialised neuropil, the axon-cap (AC). B) Higher power micrograph showing that the neuron contains numerous polyribosomes (encircled), an extensive rough endoplasmic reticulum (arrows) and numerous Golgi apparati (crossed arrows). C) Micrograph of the axon cap with profiles containing dense core vesicles (arrowhead) or small clear vesicles (arrow). D) Another example showing the glomerular structure of the neuropil within the axon-cap, with small clear vesicle-containing profiles engaged in symmetrical axoaxonic synapse (arrows), and dense core vesicles (arrowhead). These neuronal elements are surrounded by glial processes (asterisks). E) Micrograph illustrating the presence of numerous septate-like junctions (arrows) at the border of the axon-cap. F) Higher power micrograph showing the regularly spaced membranes and the electron-dense structure of the extracellular space. Scale bars, A=5 µm; B–F=0.5 µm.
135
136
A. Triller et al.
Mature inhibitory synapses in zebrafish larvae hindbrain of inhibitory currents was obtained by automatically fitting their decay phase with a single exponential. Assuming that, as in the adult goldfish, all inhibitory presynaptic boutons impinging on the soma contain one active zone,58 and that the amplitude distribution of quantal events is Gaussian,24,26 the amplitude histograms of mIPSCs were treated as the sum of an unknown number of overlapping Gaussian curves34 (and see Ref. 23 for detailed algorithm). Fits with Gaussian curves were then compared to fits with a continuous function of the form (a#(1"exp("x/b))^c)# (exp("x/d)), using a simplex algorithm (Axograph 3, Axon Instruments). In all cases the mIPSC amplitude histograms were better fitted with the sum of Gaussian curves (Standard Square Error gaussians
Identification of the Mauthner cell The M-cell was identified on semithin sections by its localization in front of the otic vesicle, just under the dorsal germinative epithelium, and by its characteristic shape. As seen with low-power electron microscopy, the M-cell contained at 52 h a clear round nucleus and was crescent shaped (Fig. 1A), with a lateral and a ventral dendrite (LD and VD, respectively) which were not fully developed at this stage. The initial segment of the axon arising from the convex side of the cell was oriented toward the midline. The clear cytoplasm (Fig. 1B) contained the classical organelles, Golgi apparatus, smooth and rough reticulum, mitochondria and a large number of free polyribosomes. At this age, the initial segment was already surrounded by a dense neuropil, the axon-cap (Fig. 1A), a structure characteristic of the M-cell. This region displayed numerous axonal profiles containing small clear round vesicles sometimes associated with dense
137
core vesicles (Fig. 1C). As observed around the initial segment of the M-cell of other mature teleosts,39 these profiles are axoaxonic synapses, with presynaptic differentiations in front of each other. These connections are surrounded by thin glial processes (Fig. 1D) which give them a glomerular structure. Within the axon cap, the extracellular space is of variable size, ranging from 10 to 30 nm, and the cleft is relatively transparent to electrons. However, in many instances, the distance between the membranes of adjacent glial cells, or glial cells and axons, remains constant with a width of 20–25 nm (Fig. 1B–C). It is then filled with uneven electron-dense and pseudoperiodic matrix. These ‘‘differentiated’’ appositions vary in length, from 0.1 to 1.5 µm, and they are similar to the septate-like junctions found in the goldfish axon-cap.57 Synaptic boutons over the Mauthner cell soma We investigated the synaptic investment in the soma and in the initial portion of the dendrites. Vesicles containing profiles with a more or less mature synaptic structure were observed all around the M-cell, although the dorsal side of this neuron was less densely innervated. Outside the axon-cap, boutons was covered by a layer of glial processes, forming a synaptic bed comparable to that observed around the goldfish M-cell,45 fish lateral vestibular nucleus30 and rat lateral vestibular neurons.50 Various types of profiles could be discriminated according to their vesicular content. They were found at any location around the cell body, except for the axoaxonic synapses (see above) which were only present within the axon-cap. Some profiles adjacent to the M-cell contained clear multimorphic vesicles (40– 80 nm) intermingled with large (90–100 nm) dense core vesicles (Fig. 2A) while others contained smaller (60–80 nm) dense core vesicles alone. These profiles were just adjacent to the M-cell surface and did not display synaptic differentiations. They may correspond to the so-called ‘‘immature’’ synapses previously described.22 However, in the absence of serial sections, we could not be sure whether or not any morphologically differentiated structure existed between the two elements.
Fig. 2. Ultrastructural characteristics of axonic profiles contacting the M-cell (M). All illustrated types of terminals were found over the entire neuronal surface, without regional preference. A) Ending with ‘‘multimorphic’’ large clear vesicles (arrow) and dense core vesicles (arrowhead). B) Axonic profile with numerous dense core (arrowheads) small clear (arrows) vesicles. C) Example of a well-defined synaptic complex (between bars); note the presynaptic accumulation of small clear spherical vesicles (arrows), the presynaptic dense projections (arrowhead), and the thin postsynaptic density. D) Another example of a morphologically mature synapse with a pleiomorphic population of vesicles in which the presynaptic dense projections are well visible (same symbols as in C). E–F) EPTA-stained semithin sections demonstrating the structural maturity of synaptic complexes (between bars in E), and an en face view showing that the boutons contain a single active zone (arrow in F). In E and F, the contours of the boutons are indicated by arrowheads. G–H) The M-cell surface is also reached by profiles bearing gap junctions (crossed arrows) and occasionally, small round vesicles (arrows) are also apparent. Note the fuzzy electron density (arrowheads) adjacent to the postsynaptic plasma membrane. Scale bars=0.2 µm.
138
A. Triller et al.
Fig. 3. Histogram of the number of presynaptic dense projections contained in individual presynaptic grids.
M-cells were also in relation to boutons containing spherical (Fig. 2C) or pleiomorphic small (30–40 nm) clear vesicles (Fig. 2D). They established synaptic contacts with the characteristic features of synaptic complexes including pre- and postsynaptic differentiations and an electron-dense material within the cleft. These junctions were specifically stained with EPTA impregnation, which allows visualization and assessment of the maturity of synaptic complexes.6,64 This procedure revealed that, at identified synapses, the presynaptic grids and postsynaptic differentiations were already structurally mature (Fig. 2E). Using semithin sections, ‘‘en face’’ views of the synaptic complex could be obtained and individual presynaptic dense projections in a grid could be seen. In most cases the whole synaptic grid area could be visualized in these 0.5 mm thick sections.58,64 However, in the 52-hour-old zebrafish M-cell, the extracellular space is not as electron-opaque as in the adult goldfish. As a consequence, it was not as easy to delineate the contour of the boutons as in the goldfish. None the less, it could be determined that the mean number of dense projections was 10.27&3.62 (n=105) per grid, and that their overall distribution was skewed (Fig. 3). To avoid any ambiguities, we considered that a grid is formed when it contains five or more dense projections compacted with a distance of less than 80 nm between each PDPs.54 We found that the presynaptic grids, on the zebrafish M-cell soma, were compact and not perforated as observed on the soma of adult goldfish M-cell.58 Of 96 boutons in which the contour could be clearly identified (see Fig. 2F), 89 contained a single active zone. Three boutons contained two active zones and the remaining four probably had two active zones, but the limits of the latter were uncertain. Profiles establishing gap junctions with the M-cell were also detected at the level of the axon hillock (Fig. 2G–H). Occasionally, a few round vesicles were present next to the presynaptic membrane suggesting that such boutons are involved in mixed, electrical and chemical transmission and may therefore belong to spiral fibre endings.38
Immunocytochemical studies on semithin sections revealed that boutons enriched in GABA or glycine are present over the M-cell soma and on the initial segment (Fig. 4A1, B1). This distribution contrasts with that observed on the adult M-cell of goldfish, where the inhibitory endings in the axon-cap were exclusively glycinergic.60 Electron microscopic analysis indicated that the synaptic density was 0.83 boutons per µm of membrane (165 boutons on 198 µm of membrane, from two sections separated by 5 µm). The proportions of glycine- and GABAimmunoreactive (IR) boutons, evaluated on these two non-serial sections, were similar and corresponded to 11% and 12%, respectively. In addition, glycine-IR (Fig. 4A2–3) and GABA-IR (Fig. 4B2–3) endings established mature synaptic complexes. Interestingly, some of the GABA-IR boutons were over spine-like protrusions (Fig. 4B2–3). Inhibitory postsynaptic currents Previously we showed that mIPSCs recorded in 1 µM TTX+10 mM Mg2+ were 95% blocked by 1 µM strychnine34 while bicuculline had no action on mIPSCs. One possible explanation is that spontaneous GABA events have a low probability of occurrence. If this proposition is true, one would expect the likelihood of detecting GABAergic mIPSCs to be enhanced by a procedure known to increase mIPSCs frequency. To test this hypothesis, we recorded synaptic activity in the presence of 1 µM TTX and 1.3 mM CaCl2 and applied, by bath perfusion, a solution containing 11.5 mM KCl. As expected, in the presence of this high [KCl]o external solution, the overall frequency of the synaptic events increased and ranged from 44 to 182 Hz, compared with 4 to 16 Hz in the controls (1.3 mM [KCl]o). Under these experimental conditions and in the presence of 1 µM strychnine, GABAergic events were isolated by their sensitivity to 20 µM bicuculline (a competitive GABAA antagonist). The application of 1 µM strychnine rapidly decreased the amplitude and frequency of mIPSCs and a stable level was reached within 2–3 min of application. Specifically, strychnine inhibited 90.1&8.4% (n=6 cells) of spontaneously occurring synaptic events (Fig. 5A). Subsequent addition of 20 µM bicuculline blocked 90% of the residual events, suggesting that they result from the activation of GABAA receptors (Fig. 5B). The few remaining events observed in the presence of strychnine and bicuculline were not analysed. Although the frequency of GABAergic mIPSCs is lower than that of glycinergic mIPSCs, basal activation of GABAA receptors might occur by continuous release of the neurotransmitter. Such activation of GABAA receptors would contribute to the fluctuations of background noise. Basal noise fluctuations were measured using point-per-point amplitude
Mature inhibitory synapses in zebrafish larvae hindbrain
139
Fig. 4. Morphological evidence that GABAergic and glycinergic terminals establish morphologically differentiated synaptic contact with the 52-hour-old M-cell. Glycine (A1–3) and GABA (B1–3) immunoreactivity on semithin (A1; B1) and ultrathin (A2–3; B2–3) sections. A1; B1) Examples of immunoreactive profiles (arrows) adjacent to the soma (s) and initial segment (double arrowheads). A2–3) Examples of glycine-IR profiles establishing differentiated synaptic complexes (between arrows). B2–3) Examples of GABA-IR profiles with synaptic contacts (between arrows), here over spine-like M-cell protrusions (crossed arrows). Scale bars, A1–B1:=10 µm; A2–3, B2–3=0.25 µm.
histograms34 during non-active periods. The standard deviation (SDm) of basal noise ranged from 2.9 to 5.6 in the presence of TTX plus strychnine (n=6). Adding bicuculline to the bath solution changed SDm values by at most "8 to 17%, depending on the cell tested (n=6). This result suggests that there is no significant basal activation of GABAA receptors on the zebrafish M-cell (P<0.05 paired t-test).
General properties of GABAergic events The bicuculline-sensitive synaptic events recorded in the presence of strychnine and TTX were considered as GABAergic miniature postsynaptic currents (mIPSCs). Their amplitude distributions were obtained from data sets of 180 to 822 events. In all cells tested (n=6), histograms could be well resolved into
140
A. Triller et al.
Fig. 5. Whole-cell recording of glycinergic and GABAergic miniature synaptic events. (A) Cumulative amplitude histogram of mIPSCs collected at a holding potential of "50 mV in the presence of a solution containing 1 µM TTX and 11.5 mM [KCl]o (m=4010/42 s) and of a solution containing 1 µM TTX, 11.5 mM [KCl]o and 1 µM strychnine (m=768/42 s). Note that strychnine strongly reduced the proportion of the largest events. Insert: four superimposed epochs (1 s each) of spontaneous postsynaptic activity in the presence of TTX and 11.5 mM [KCl]o before (upper panel) and during (lower panel) strychnine applications. B) Amplitude histogram of strychnine-resistant mIPSCs before (m=768/42 s) and during application of 20 µM bicuculline (m=14/42 s) (bin=2 pA). The amplitude histogram was well fitted by the sum of two Gaussian curves with means and S.Ds of 15.9&4.6 pA and 28.9&10.6 pA, respectively. The insert shows four superimposed epochs before (upper panel) and during (lower panel) the perfusion of bicuculline which dramatically reduced the number of residual strychnine-resistant mIPSCs. A and B were obtained from the same cell.
one or two Gaussians (see Experimental Procedures). The second class was detected only in histograms including large data sets (n=3). As illustrated in Fig. 5B, these two classes were fitted by the sum of two
Gaussian curves. The mean amplitude of the first class was close to 18 pA (17.9&3.3 pA; n=6; Vh="50 mV) and it was close to 34 pA for the second class (33.9&6.8 pA; n=3).
Mature inhibitory synapses in zebrafish larvae hindbrain
Fig. 6. (A) Example of the distribution of time to peak of strychnine resistant mIPSCs recorded at "50 mV (60 s duration: bin=0.1 ms, n=420). The distribution was well fitted by a single Gaussian curve with a mean and S.D. of 0.71&0.2 ms. The insert shows that measurements were made between the onset and the maximum of the mIPSCs. B) Distribution of the decay time-constants (same cell as in A) fitted with a single exponential. In this example the mean decay time-constant was 4.73&1.17 ms. Insert: 10 superimposed strychnine-resistant mIPSCs (upper traces) and their average (lower trace).
As for glycinergic mIPSCs,34 GABAergic mIPSCs had fast kinetics. Time to peak (ttp) distributions were well fitted by a single Gaussian curve (Fig. 6A) giving a mean ttp value of 0.62&0.12 ms (n=6 cells). The value of the decay time-constant (ôd) of the strychnine-resistant mIPSCs was obtained by fitting the first 10 ms of the IPSCs repolarizing phase fitted by a single exponential curve. ôd ranged from 2 to 30 ms (Fig. 6B). Although the decay time-constant histogram is slightly skewed, one predominant class clearly emerged (Fig. 6B). This class ranged from 2 to
141
Fig. 7. Outside-out recordings of single-channel currents evoked by GABA and glycine applications (A) Effect of 20 µM GABA and 5 µM glycine applications to the same patch during outside-out recordings (Vh="50 mV; filter 1 KHz). Note that glycine evoked larger opening states than GABA. (B) Point-per-point amplitude histogram of a data segment (shown in the insert) during 20 µM GABA application (bin 0.01 pA). The mean amplitude of GABA-evoked opening states was obtained by fitting the amplitude distribution with Gaussian curves. In this cell the mean amplitude of GABA-evoked opening states was 1.15&0.38 pA, indicating a main conductance state of 21 pS (reversal potential=0 mV).
8 ms and was well fitted by a Gaussian curve, with a mean of 3.67&0.23 ms (n=6 cells). There was no correlation between ttp, ôd and mIPSCs amplitude in any cell tested (P<0.15). GABA-evoked single-channel openings in outside-out recordings The size of mIPSPs depends on the amplitude of the main conductance state of the activated channel and the number of receptors activated after an exocytosis. In order to study the conductance properties of GABAA receptor-channels, openings were induced by the application of 20 µM GABA to outside-out patches excised from the M-cell.
142
A. Triller et al.
At a holding potential of "50 mV, GABA applications evoked single-channel openings in all patches tested. Applications of 5 µM glycine to the same patches (n=5) evoked single-channel events with an amplitude larger than that produced by GABA (Fig. 7A). No kinetic analysis was performed on these events. The GABA-activated channels had a single conductance state which was measured using point-per-point histograms (Fig. 7B). At "50 mV, this conductance state had an averaged value of 22.7&2.2 pS (n=5 patches) for an observed reversal potential close to 0 mV. This value is two to four times smaller than that previously described34 for glycine receptor-channels (41–43 pS and 81–86 pS). It allowed us to calculate the minimal number of GABA-gated channels simultaneously activated after exocytosis, assuming that synaptic and extrasynaptic GABAA receptors are functionally identical.8 The amplitude histograms of mIPSCs suggest that it ranges from 14 to 20 channels, which is close to that obtained for GABA synapses in other species8,32,63 and for glycine synapses on zebrafish larvae M-cells.34 DISCUSSION
Identification of the Mauthner cell One of the elements which allows unambiguous identification of the M-cell is the axon-cap (AC). It is a specialized neuropil which surrounds the initial segment of the M-cell of teleosts, and which has been characterized by both light and electron microscopy.38,39,57 It has been shown that the AC is directly involved in the generation of an electrical field effect, that is of the electrical inhibition which results from the channelling across the membrane of presynaptic terminals of M-cell action potential currents due to the high electrical resistance of the extracellular tissue in the AC.25 Three structural features are specific to this neuropil: i) it is surrounded by glial elements, the cap cells; ii) axons spiral around the initial segment and establish axoaxonic symmetrical synapses between themselves with a pseudo-glomerular structure; iii) in the peripheral part of the AC, axons and glial cells are tightly linked by septate-like junctions in goldfish and in tench.57 In 52-hour-old zebrafish larvae we could identify the latter two features unambiguously but a large number of free polyribosomes were also observed, this last feature being commonly observed in immature neurons.43 Such observations confirm that the M-cell system is not completely mature at this stage.22 Morphological analysis of synaptic contacts on 52hour-old zebrafish Mauthner cells Our morphological studies show that the M-cell soma and initial portions of dendrites contact synaptic endings with the structural characteristics of
mature synapses: the presynaptic elements contain small clear, round or pleiomorphic vesicles and presynaptic differentiations and are adjacent to postsynaptic differentiations. Also, and as in the adult, other profiles containing round vesicles establish gap junctions with the M-cell and are found on the initial segment as well as on the soma. We could not identify in the zebrafish, as in goldfish,58 all the types of endings which were stained, but it was possible to study their active zones. The mean number of PDPs for all synaptic boutons analysed was 10, that is slightly less than that reported (n=13) at the glycinergic unmyelinated club-endings of the goldfish.58 This number should parallel the maximum number of vesicles surrounding the PDPs and adjacent to the plasma membrane which was estimated to be 40 to 60 in the goldfish. The spiral fibres and the unmyelinated club endings described by Nakajima38 in the axon cap are already present in the teleost larvae. Outside this region, precise identification of the terminals involved was more difficult, mainly because axons are not myelinated at this stage of development. However, two types of boutons could be distinguished according to the shape of their synaptic vesicles i.e. round and flat/pleiomorphic. They could correspond to ‘‘large vesicle boutons’’ and ‘‘small myelinated club endings’’ for the first group and to the ‘‘small vesicle boutons’’ for the second. Other profiles adjacent to the zebrafish M-cell membrane were less easily correlated with classes of endings known to contact the adult goldfish M-cell. Some boutons were filled with medium-sized dense core vesicles, and others with large clear vesicles of variable sizes and with large dense core vesicles. Neither of the two classes established differentiated synaptic complexes. The medium dense core vesicles were comparable in size to those generally associated with catecholamine.15 The large dense core vesicles were similar to those present in somatostatincontaining fibres that have been described around the adult M-cell53 or to those found in rat dorsal horn and which contain neurotensin.56 Whether or not the large multimorphic vesicles correspond to immature contacts remains to be determined. Although they could also result from fixation artefacts it should be pointed out that at this stage the M-cell measures about 100 µm from the tip of its ventral dendrite to that of its lateral one, with a diameter of about 6–10 µm. Therefore, the M-cell has not yet reached its adult size (300 µm in length and diameter 30– 40 µm). Since the covering ratio of the adult zebrafish M-cell appears to be relatively high,21,22 it is likely that new synapses are still being formed at this stage of maturation. GABAergic miniature inhibitory postsynaptic currents With immunocytochemistry, we found that a similar proportion of glycine- and GABA-IR profiles are
Mature inhibitory synapses in zebrafish larvae hindbrain
present on the soma, outside and within the AC. This contrasts with the goldfish where glycine-containing profiles predominate in the AC and where GABAergic profiles are only observed outside this region.52 Whole-cell recordings indicated that these GABAergic synaptic boutons are functional. But as in goldfish, M-cell inhibitory synaptic events are mainly glycinergic34 due to a lower probability of release and/or to a still incomplete maturation of the release machinery at GABAergic terminals (see below). The low frequency of GABAergic events observed even in the presence of a high K+ extracellular solution can therefore explain why they have not been detected, so far, in adult goldfish and in previous studies of zebrafish larvae.34,35 Most GABAergic mIPSCs (95%) recorded from the zebrafish M-cell have a short decay phase (4– 5 ms) which can be fitted by a single exponential curve. It is unlikely that the few slower mIPSCs (ôd=15–25 ms) reflect space-clamp error since there is no correlation between their ttp, decay time-constant or amplitude.44 These GABAergic IPSCs have kinetics similar to those observed in other species and on adult neurons. GABAergic IPSCs have a fast or slow decay phase depending on the cell tested8,19,32,41,42,49,63 and these differences might result from the activation of different GABA receptorsubtypes expressed by the same neuron.42,49 Further, slow events might reflect the activation of particular GABAA receptors with desensitized states prolonging GABAA channel responses to neurotransmitter release.18 The single conductance state value (23 pS) of the GABA-gated channels is similar to that observed in other species.8,16,32,40 This value may vary according to the subunit combination of the receptor channels.2,61,62 Although many GABAA receptor subunits have been cloned, single-channel properties of only a few of the many possible combinations have been analysed in heterologous expression systems.2,61,62 Moreover, GABAA receptor subunits from zebrafish brain have not yet been cloned. Therefore, it is not possible to define which combination of subunits is expressed by the M-cell. Miniature inhibitory postsynaptic currents activity Although we did not make detailed measurements on the proportion of glycine- or GABA-IR boutons, visual inspection suggested that their numbers do not differ greatly. In contrast, the frequency of GABAergic mIPSCs was significantly lower than that of glycinergic events. This might depend on differences in the post- and/or presynaptic mechanisms. A first hypothesis would be that some synapses with boutons that contain GABA either do not express postsynaptic receptors or express them at such a low density that synaptic events cannot be generated. This is probably not the case since GABA-
143
gated channel activity was detected in all patches tested during outside-out recordings. In mammals, even immature synapses possess active receptors11 and postsynaptic maturation corresponds essentially to a change in receptor subtypes expressed7 rather than to changes in their density.31 Alternatively, the low frequency of GABAergic mIPSCs might result from differences in the presynaptic mechanism leading to vesicle exocytosis. Synaptic boutons with different release probabilities have been described in mammals.46 Even when neurotransmitters are co-localized, as for GABA and glycine in synapses on cerebellar granule cells,20 synaptic boutons appear to release only one type of molecule (GABA on granule cells). Differences in release probability between GABAergic and glycinergic boutons could also reflect differential expression of calcium-channel subtypes between synapses during development, as recently proposed in hippocampal neurons.20 In the 52-hour-old zebrafish M-cell, GABA mIPSCs have less complex amplitude distributions than the multimodal histograms of glycinergic events.34 The presence, in the vicinity of the M-cell, of growth cone-like profiles containing large vesicles could account for the amplitude fluctuation of glycinergic mIPSCs. In fact, neurotransmitters are released from growth cone terminals,1,65 but the release probability and the number of neurotransmitter molecules issued at immature synapses are very low.1,65 Since one active zone can release at most one vesicle at mature synapses,29,58 the number of presynaptic active zones per bouton might, in principle, have accounted for the difference between the amplitude distribution histograms of glycinergic and GABAergic mIPSCs recorded in our preparation. mIPSC amplitude distributions recorded in the soma of the adult goldfish M-cell are unimodal while they are multimodal for mIPSCs collected from the lateral dendrite.23 Accordingly, 96% of the unmyelinated endings and small vesicle boutons contain a single active zone in the soma and in the axon-cap of the adult goldfish, while 30% of the small vesicle boutons contain two or more presynaptic grids on the distal half of the lateral dendrite.54 However, in the zebrafish M-cell 92% of the endings adjacent to the soma or within the axon-cap also contain a single active zone, independently of their classification or transmitter content. Therefore, the difference between the shape of the amplitude distributions of glycinergic and GABAergic mIPSCs is not related to differences in the number of active zones per terminal. On the other hand, a multivesicular release occurring at these newly formed synapses cannot be completely excluded since the mean value of the second class of the GABAergic mIPSCs is twice that of the first one. The skewness of the GABAergic mIPSCs amplitude histogram may also result from a large variation in the number of neurotransmitter
144
A. Triller et al.
molecules released per exocytosis, as recently proposed for GABA mIPSCs in retinal cells.9 CONCLUSION
In conclusion, differences in the amplitude distributions of glycinergic and GABAergic mIPSCs are most likely due to postsynaptic factors. As previously suggested for GABA mIPSCs in mammals,8 the multimodal amplitude histograms of glycinergic mIPSCs could result from the presence of variable
proportions of different glycine receptor types from one synapse to another while the number and the identity of postsynaptic GABAA receptors could be less variable. To be tested, this hypothesis will require investigations using quantified confocal fluorimetric measurements as developed for the adult goldfish M-cell.47,59
Acknowledgements—This work was supported by INSERM and Associaton Franc¸aise pour la Myopathie.
REFERENCES
1. 2. 3. 4. 5. 6. 7. 8. 9. 10. 11. 12. 13. 14. 15. 16. 17. 18. 19. 20. 21. 22. 23. 24. 25. 26. 27. 28. 29.
Allen T. G. J. and Brown D. A. (1996) Detection and modulation of acetylcholine release from neurites of rat basal forebrain cells in culture. J. Physiol., Lond. 492, 453–466. Angelotti T. P. and MacDonald R. L. (1993) Assembly of GABAA receptor subunits: á1â1 and á1â1ã2s subunits produce unique ion channels with dissimilar single-channel properties. J. Neurosci. 13, 1429–1440. Ankri N., Legendre P., Faber D. S. and Korn H. (1994) Automatic detection of spontaneous synaptic responses in central neurons. J. Neurosci. Meth. 52, 87–100. Bekkers J. M., Richerson G. B. and Stevens C. F. (1990) Origin of variability in quantal size in cultured hippocampal neurons and hippocampal slices. Proc. natn. Acad. Sci. U.S.A. 87, 5359–5362. Bekkers J. M. and Stevens C. F. (1990) Presynaptic mechanism for long-term potentiation in the hippocampus. Nature 346, 724–729. Bloom F. E. and Aghajanian G. K. (1968) Fine structural and cytochemical analysis of the staining of synaptic junctions with phosphotungstic acid. J. ultrastruct. Res. 22, 361–375. Dunn S. M. J., Bateson A. N. and Martin I. L. (1994) Molecular neurobiology of the GABAA receptor. Int. Rev. Neurobiol. 36, 51–96. Edwards F. A., Konnerth A. and Sakmann B. (1990) Quantal analysis of inhibitory synaptic transmission in the dentate gyrus of rat hippocampal slices: a patch clamp study. J. Physiol., Lond. 430, 213–249. Frerking M., Borges S. and Wilson M. (1995) Variation in GABA mini amplitude is the consequence of variation in transmitter concentration. Neuron 15, 885–895. Ga¨hwiler B. H. and Brown D. A. (1985) GABAB-receptor-activated K+ currents in voltage-clamped CA3 pyramidal cells in hippocampal cultures. Proc. natn. Acad. Sci. U.S.A. 8, 1558–1562. Grantyn R., Kraszewski K., Melnick I., Taschenberger H. and Warton S. S. (1995) In vitro development of vertebrate central synapses. Perspect. dev. Biol. 2, 387–397. Gulyas A. I., Miles R., Sik A., Toth K., Tamamaki N. and Freund T. F. (1993) Hippocampal pyramidal cells excite inhibitory neurons through a single release site. Nature 366, 683–687. Hamill O. P., Marty A., Neher E., Sakmann B. and Sigworth F. J. (1981) Improved patch clamp techniques for high-resolution current recordings from cells and cell free patches. Pflu¨gers Arch. 391, 85–100. Horokawa N. K., Sobue K., Kanda K., Harada A. and Yorifuji H. (1989) The cytoskeletal architecture of the presynaptic terminal and molecular structure of synapsin I. J. Cell Biol. 108, 111–126. Ho¨kfelt T. (1991) Neuropeptides in perspective: the last ten years. Neuron 7, 867–879. Isaacson J. S., Solis J. M. and Nicoll R. A. (1993) Local and diffuse synaptic actions of GABA in the hippocampus. Neuron 10, 165–175. Johnston G. A. R. (1991) GABAA antagonists. Semin. Neurosci. 3, 205–210. Jones M. V. and Westbrook G. L. (1995) Desensitized states prolong GABAA channel responses to brief agonist pulses. Neuron 15, 181–191. Kaneda M., Farrant M. and Cull Candy S. G. (1995) Whole cell and single-channel currents activated by GABA and glycine in granule cells of the rat cerebellum. J. Physiol., Lond. 485, 419–435. Kenneth P., Sholz P. and Miller R. J. (1995) Developmental changes in presynaptic calcium channels coupled to glutamate release in cultured rat hippocampal neurons. J. Neurosci. 15, 4612–4617. Kimmel C. B. and Model P. G. (1978) Developmental study of the Mauthner cell. In Neurobiology of the Mauthner Cell (eds Faber D. S. and Korn H.), pp. 183–220. Raven Press, New York. Kimmel C. B., Sessions S. K. and Kimmel R. J. (1981) Morphogenesis and synaptogenesis of the zebrafish Mauthner neuron. J. comp. Neurol. 198, 101–120. Korn H., Bausela F., Charpier S. and Faber D. S. (1993) Synaptic noise and multiquantal release at dendritic synapses. J. Neurophysiol. 70, 1249–1254. Korn H., Burnod Y. and Faber D. S. (1987) Spontaneous quantal currents in a central neuron match predictions from binomial analysis of evoked responses. Proc. natn. Acad. Sci. U.S.A. 84, 5981–5985. Korn H. and Faber D. S. (1975) An electrically mediated inhibition in goldfish medulla. J. Neurophysiol. 38, 452–471. Korn H. and Faber D. S. (1990) Transmission at a central inhibitory synapse: IV. Quantal structure of synaptic noise. J. Neurophysiol. 63, 198–222. Korn H. and Faber D. S. (1991) Quantal analysis and synaptic efficacy in the CNS. Trends Neurosci. 14, 439–445. Korn H., Faber D. S. and Triller A. (1990) Convergence of morphological, physiological, and immunocytochemical techniques for the study of single Mauthner cells. In Handbook of Chemical Neuroanatomy. (eds Bjo¨rklund A., Ho¨kfelt T., Wooterlood F. G. and van der Pol A. N.), Vol. 8, pp. 403–480. Elsevier Science Publishers, Amsterdam. Korn H., Mallet A., Triller A. and Faber D. S. (1982) Transmission at a central inhibitory synapse. II. Quantal description of release, with a physical correlate for binomial n. J. Neurophysiol. 48, 679–707.
Mature inhibitory synapses in zebrafish larvae hindbrain 30. 31. 32. 33. 34. 35. 36. 37. 38. 39. 40. 41. 42. 43. 44. 45. 46. 47. 48. 49. 50. 51. 52. 53. 54. 55. 56. 57. 58. 59. 60. 61. 62. 63. 64. 65.
145
Korn H., Sotelo C. and Bennet M. V. L. (1977) The lateral vestibular nucleus of the toadfish Opsanus tau: ultrastructural and electrophysiological observations with special references to electrotonic transmission. Neuroscience 2, 851–884. Kraszewski K. and Grantyn R. (1992) Development of Gabaergic connections in vitro: increasing efficacy of synaptic transmission is not accompanied by changes in miniature currents. J. Neurobiol. 23, 766–781. Kraszewski K. and Grantyn R. (1992) Unitary, quantal and miniature GABA-activated synaptic chloride currents in cultured neurons from the rat superior colliculus. Neuroscience 47, 555–570. Landis D. M. D., Hall K., Weinstein L. A. and Reese T. S. (1988) The organisation of the cytoplasm at the presynaptic active zone of a central nervous system synapse. Neuron 1, 201–209. Legendre P. and Korn H. (1994) Glycinergic inhibitory synaptic currents and related receptor-channel in the zebrafish brain. Eur. J. Neurosci. 6, 1544–1557. Legendre P. and Korn H. (1995) Voltage-dependence of conductance changes evoked by glycine release in the zebrafish brain. J. Neurophysiol. 6, 2404–2412. Manabe T., Renner P. and Nicoll R. A. (1992) Postsynaptic contribution to long-term potentiation revealed by the analysis of miniature synaptic currents. Nature 355, 50–55. Min M. Y. and Appenteng K. (1996) Multimodal distribution of amplitudes of miniature and spontaneous EPSCs recorded in rat trigeminal motoneurones. J. Physiol., Lond. 494, 171–182. Nakajima Y. (1974) Fine structure of the synaptic endings on the Mauthner cell of goldfish. J. comp. Neurol. 156, 375–402. Nakajima Y. and Kohno K. (1978) Fine structure of the Mauthner cell: synaptic topography and comparative study. In Neurobiology of the Mauthner cell. (eds Faber D. S. and Korn H.), pp. 133–166, Raven Press, New York. Newland C. F., Colquhoun D. and Cull-Candy S. G. (1991) Single channels activated by high concentrations of GABA in superior cervical ganglion neurones of the rat. J. Physiol., Lond. 432, 203–233. Otis T. S. and Mody I. (1992) Modulation of decay kinetics and frequency of GABAA receptor-mediated spontaneous inhibitory postsynaptic currents in hippocampal neurons. Neuroscience 49, 13–32. Pearce R. A., Grunder S. D. and Faucher L. D. (1995) Different mechanisms for use-dependent depression of two GABAA-mediated IPSCs in rat hippocampus. J. Physiol., Lond. 484, 425–435. Peters A. and Palay S. L. and de Webster F. H. (1991) The fine structure of the nervous system. Neurons and their Supporting Cells. 3rd edn. Oxford University Press, New York and Oxford. Rall W. (1969) Time constants and electrotonic length of membrane cylinders and neurons. Biophys. J. 9, 1483–1508. Robertson J. D., Bodenheimer T. S. and Stage D. E. (1963) The ultrastructure of Mauthner cell synapses and nodes in the goldfish brain. J. Cell Biol. 19, 159–199. Rosenmund C., Clements J. D. and Westbrook G. L. (1993) Nonuniform probability of glutamate release at a hippocampal synapse. Science 262, 754–757. Seitanidou T., Triller A. and Korn H. (1992) Partial glycinergic innervation induces transient changes in the distribution of a glycine receptor-associated protein in a central neuron. J. Neurosci. 12, 116–131. Silver R. A., Traynelis S. F. and Cull-Candy S. G. (1992) Rapid time-course miniature and evoked excitatory currents at cerebellar synapses in situ. Nature 355, 163–166. Soltesz I., Smetters D. K. and Mody I. (1995) Tonic inhibition originates from synapses close to the soma. Neuron 14, 1273–1283. Sotelo C. and Palay S. L. (1972) The fine structure of the vestibular nucleus in the rat. II Synaptic organisation. Brain Res. 18, 93–115. Stevens C. F. (1993) Quantal release of neurotransmitter and long term potentiation. Cell 72/Neuron 10 (Suppl.), 55–63. Sur C., McKernan R. and Triller A. (1995) GABAA receptor-like immunoreactivity in the goldfish brainstem with emphasis on the Mauthner cell. Neuroscience 66, 697–706. Sur C., Triller A. and Korn H. (1994) Colocalisation of somatostatin with GABA or glutamate in distinct afferent terminals presynaptic to the Mauthner cell. J. Neurosci. 14, 576–589. Sur C., Triller A. and Korn H. (1995) Morphology of the release site of inhibitory synapses on the soma and dendrite of an identified neuron. J. comp. Neurol. 351, 247–260. Tang C. M., Margulis M., Shi Q. Y. and Fielding A. (1994) Saturation of postsynaptic glutamate receptors after quantal release of transmitter. Neuron 13, 1385–1393. Todd A. J., Spike R. C., Price R. F. and Neilson M. (1994) Immunocytochemical evidence that neurotensin is present in glutamatergic neurons in the superficial dorsal horn of the rat. J. Neurosci. 14, 774–784. Triller A. and Korn H. (1980) Glio-axonic junctional like complexes at the Mauthner cell’s axon cap of teleosts: a possible morphological basis for field effect inhibitions. Neurosci. Lett. 18, 275–281. Triller A. and Korn H. (1982) Transmission at a central inhibitory synapse. III Ultrastructure of physiologically identified and stained terminals. J. Neurophysiol. 48, 708–736. Triller A., Seitanidou T., Frankson O. and Korn H. (1990) Size and shape of glycine receptor clusters in a central neuron exhibit a somato-dendritic gradient. The New Biologist 2, 637–641. Triller A., Sur C. and Korn H. (1993) Heterogeneous distribution of glycinergic and GABAergic afferents on an identified central neuron. J. comp. Neurol. 339, 83–96. Verdoorn T. A., Draguhn A., Ymer S., Seeburg P. H. and Sakmann B. (1990) Functional properties of recombinant GABAA receptors depend on subunit composition. Neuron 4, 919–928. Verdoorn T. A. (1994) Formation of heteromeric ã-aminobutyric acid type A receptors containing two different a subunits. Molec. Pharmac. 45, 475–480. Virginio C., Martina M. and Cherubini E. (1995) Spontaneous GABA-mediated synaptic currents in cerebellar granule cells in culture. NeuroReport 6, 1285–1289. Vrensen G., Nunes Cardozo J., Muller L. and Van Der Want J. J. L. (1980) The presynaptic grid: a new approach. Brain Res. 184, 23–40. Young S. H. and Poo M. M. (1983) Spontaneous release of transmitter from growth cones of embryonic neurones. Nature 305, 634–637. (Accepted 13 February 1997)