Proc. NatL Acad. Sci. USA Vol. 80, pp. 5126-5128, August 1983 Neurobiology

Carbachol can be released at a cholinergic ganglionic synapse as a false transmitter (acetylcholine/acetylcholinesterase/quantal release/intracellular injection) G. BAUX AND L. TAUC Laboratoire de Neurobiologie cellulaire, Centre National de la Recherche Scientifique, 91190 Gif-sur-Yvette, France Communicated by S. Hagiwara, May 3, 1983 ABSTRACT Carbachol was injected into a presynaptic cho- itory Cl--dependent postsynaptic responses in a nearby group linergic neuron in the buccal ganglion of Aplysia and the quantal of neurons (6). The pre- and postsynaptic cells were each pen- aspects of the ClP-dependent postsynaptic response to a pro- etrated by two separate low-resistance 3 M KCI-filled elec- longed stimulation were analyzed by a statistical fluctuation method. trodes (1-5 MW). The calculated amplitude of the miniature postsynaptic current Transmission was studied by using a method described by was increased with respect to control. Statistical fluctuation anal- Simonneau et aL (7). Both pre- and postsynaptic cells were si- ysis was also used to analyze the postsynaptic response obtained during ionophoretic application of acetylcholine and carbachol. multaneously voltage clamped to holding potentials of -50 mV The calculated unitary channel current was found to be greater and -80 mV, respectively. In the presence of TTX (0.1 mM), for carbachol than for acetylcholine. This increase could explain a 3-s step depolarization of the presynaptic neuron induced a the larger miniature postsynaptic current seen after intracellular postsynaptic response (LDIPSC, long depolarization-induced injection of carbachol into the presynaptic neuron if carbachol postsynaptic current), which showed fluctuations or "noise. " was released at the synapse as a false transmitter. This conclusion The amplitude of the miniature postsynaptic currents (MPSC) was supported by the observation that it was possible to restore was calculated by using Campbell's theorem, in which the size transmission at a synapse previously blocked by presynaptic in- of the unitary element or individual MPSC (i) is related to the tracellular injection of acetylcholinesterase with a presynaptic in- variance of the noise (E2) and the mean observed current change jection of carbachol. (I) by the equation: i = 2E2/I (8). Agonists were ionophoretically applied to the cell body of The specificity of the synaptic release mechanism is not ab- the postsynaptic neuron bearing AcCho receptors (9). The same solute and substances with molecular formulae relatively close equation was used to calculate the current crossing a single to the natural transmitter can be liberated. At cholinergic syn- channel opened by AcCho or carbachol, with the omission of apses, precursors like homocholine, monoethylcholine, and the factor 2, because of the pulse-like aspect of the elementary triethylcholine are taken up by cholinergic nerve terminals, current (8, 10). In all experiments, the null potential (-40 to acetylated, and released as false transmitters. This property has -55 mV) of the postsynaptic responses was continuously mon- been used as a tool to understand the mechanism of transmitter itored and the currents were expressed as conductance. release (1-4). A Fourier transform was used to calculate the power density We have investigated whether the well-known acetylcholine spectra from which the decay time constant of the MPSC and (AcCho) analogue, carbachol (carbamoylcholine chloride) can the channel open time were estimated (7). As previously shown, be released at a cholinergic synapse. Carbachol has the inter- the MPSC spectra could be approximated by single Lorentzian esting property of not being hydrolyzed by acetylcholinesterase curves (7), whereas spectra of Cl- channel noise could only be (AcChoE) and as such can be used as a pharmacological tool in fitted by double Lorentzians (11, 12). Thus, it may be hazard- the presence of large quantities of this enzyme. Carbachol is ous to give absolute values of the channel open time; the values not taken up by cholinergic terminals and having no acetyl group, reported here are indicative and concern only the slower com- it cannot be synthesized in the cytoplasm. We have overcome ponent. this difficulty by injecting carbachol into a presynaptic cholin- AcCho and carbachol were intracellularly injected by ion- ergic neuron of Aplysia buccal ganglion. We have found that it tophoresis with a constant current. Postsynaptic current re- can act as a false transmitter and can also replace AcCho at a sponses used for calculation and presented in the figures were synapse previously blocked by intracellular injection of Ac- obtained with 3-s depolarizations of the presynaptic neuron to ChoE (5). 0 mV. However, the difference in the actions of AcCho and car- bachol shown below was observed for other levels of presyn- MATERIALS AND METHODS aptic depolarization. AcChoE (Worthington) was injected as a 5% solution (wt/vol) Experiments were performed at room temperature (20-22°C) by an air pressure system on the buccal ganglion of Aplysia californica obtained from R. (5). Fay (Pacific Biomarine, Venice, CA). The ganglion was dis- sected free of connective tissue and perfused with artificial sea RESULTS water (460 mM NaCl/10 mM KCl/11 mM CaCJ2/25 mM MgCl2/ Fig. 1 illustrates the postsynaptic responses obtained when either 28 mM MgSO4/10 mM Tris-HC1, pH 7.8). Tetrodotoxin (TTX) AcCho or carbachol was injected into the presynaptic neuron. (Sigma) at 0.1 mM was added to block Na+ channels. Cells used After a delay, both drugs increase the amplitude of the mean were two cholinergic interneurons, which each induce inhib- postsynaptic current, and this increase is related to the amount The publication costs of this article were defrayed in part by page charge Abbreviations: AcCho, acetylcholine; AcChoE, acetylcholinesterase; TTX, payment. This article must therefore be hereby marked "advertise- tetrodotoxin; LDIPSC, long depolarization-induced postsynaptic cur- ment" in accordance with 18 U.S.C. §1734 solely to indicate this fact. rent; MPSC, miniature postsynaptic currents. 5126 Downloaded by guest on September 27, 2021 Neurobiology: Baux and Taue Proc. Natl. Acad. Sci. USA 80 (1983) 5127 A B to the enzymatic action of AcChoE (11). On the other hand, AcCho-activated channels may be differently affected by AcCho and carbachol. When AcCho and carbachol were ionophoresed at the same spot on the postsynaptic cell by using a double-bar- reled electrode, we found that the single channel conductances calculated for the (inhibitory) responses having the same mean current were different (Fig. 2). The mean (±SD) values ob- served in 15 experiments were 3.4 ± 0.8 pS for AcCho and 5.2 ± 1.1 pS for carbachol. In contrast, the calculated mean open time of the slower component of the Cl- channel was some- what shorter (by about 25%) with carbachol (10 ± 2 ms) than with AcCho (14 ± 2 ms). To ascertain that the difference in conductance of inhibitory channels activated by AcCho and carbachol was not an ex- -J perimental artefact, we used the same channel analysis in other Aplysia cells having excitatory receptors opening cation-selec- tive "sodium" channels (13). We found that these channels, whether opened by AcCho or carbachol, have essentially the same amplitudes but a shorter lifetime with carbachol, as al- ready shown by others (13). It should be mentioned that mean FIG. 1. Postsynaptic current response after sustained depolariza- single channel conductances and lifetimes computed from in- tions of the presynaptic neuron to 0 mV. Upper traces, DC recordings; lower traces, AC recordings at higher gain of the current fluctuations tegrated spectral densities for current fluctuations in frog end (calibrations, 4 nA for DC and 1 nA for AC recordings; time, 1 s). (A) plates were dependent on the agonists used (14). 1, test response; 2, response after a 30-min injection of carbachol with Another possible mechanism responsible for the increase in 50-nA constant current; 3, after another 30-min injection of carbachol. MPSC amplitude after carbachol injection could reside in a fa- The calculated amplitude of the MPSC was 1.09 nS in 1, 1.54 nS in 2, cilitatory action of a "leakage" or a nonquantal release of car- and 1.95 nS in 3. (B) In another preparation similar to the experiment bachol, shifting the sigmoid amplitude/concentration curve of inA, but AcCho was injected instead ofcarbachol. The amplitude ofthe MPSC remained unchanged (0.5 nS) throughout. The increase in am- released AcCho to the right (15). To test this possibility, we ana- plitude of the LDIPSC after AcCho injections is due exclusively to an lyzed the LDIPSC of noninjected preparations in the presence increase in the number of quanta released, whereas injected carbachol of carbachol (1 tkM to 0.1 mM) in the perfusing medium. In no increases both the quantal content and the size of MPSC. case was an increase in MPSC amplitude calculated. Thus, the larger unit conductance induced by carbachol on of the drug introduced into the presynaptic cell. Because in- the postsynaptic membrane explains the increase in amplitude tracellular ionophoresis increases the intracellular concentra- of the MPSC calculated after injection of carbachol into the tion of the drugs progressively, it is impossible to quantify the presynaptic neuron. After the injection of carbachol, each delay. Also, the final intracellular concentration reached after quantum released probably consists of a mixture of AcCho and injection cannot be precisely calculated, as the transfer number carbachol. The proportion of carbachol increases with the of the injection electrode is not known. It can be estimated as quantity of carbachol injected. being on the order of millimolar. The injected neuron showed The other method we used to show release of carbachol from a decrease in membrane potential, and when voltage clamped, the presynaptic neuron takes advantage of the fact that release the outward, presumably late K current was slightly decreased of AcCho is abolished by injection of AcChoE into the pre- during the depolarizing pulse. synaptic neuron. This effect was attributed to the destruction The LDIPSC was analyzed to estimate the size of the MPSC. of cytoplasmic AcCho in the terminal (5). Because carbachol is Surprisingly enough, the size of the MPSC was clearly larger not hydrolyzed by AcChoE, injection into the presynaptic neu- after injection of carbachol than after injection of AcCho. In- ron should restore neurotransmission (Fig. 3). deed, comparing LDIPSC having the same mean current am- Injection of AcChoE led to depression of the postsynaptic plitude before and after injection, the calculated mean MPSC did not change after AcCho injection but clearly increased after carbachol injection (Fig. 1). The usual increase observed in five experiments was about 35% after a 30-min injection with 50-nA A current and about 70% after a 60-min injection. On the con- trary, the calculated decay times (mean ± SD) (14 ± 2 ms) of the MPSC did not show any significant change. The increase in the MPSC amplitude after carbachol injection does not ac- count totally for the increase of the LDIPSC. The number of quanta released after carbachol injection approximately dou- bled after allowances were made for the increase in size of the MPSC and LDIPSC. Rather than considering the increase in MPSC as resulting FIG. 2. Postsynaptic current responses after ionophoresis of Ac- from an increase in the size of individual packets of AcCho, we Cho and carbachol onto somatic AcCho receptors ofthe same cell. Dou- thought that it could reflect a difference in the postsynaptic ble-barreled electrode. (A) DC response to AcCho ionophoresis (cali- effect of released carbachol. One reason for the increased MPSC bration, 10 nA, 2.5 s). (B) Background noise at high-gain AC recording amplitude may be that carbachol (not being hydrolyzed by in the absence of agonist (calibration, 0.25 nA, 1 s; also valid for C and AcChoE) would have, if released by the presynaptic cell, a more D). (C) AcCho-activated channel noise (mean channel conductance in this experiment, 2.9 pS). Same response as A. (D) Carbachol-activated potent effect on the postsynaptic receptors. However, it was channel noise recorded during a response having the same DC ampli- reported that at the same synapse in the buccal ganglion, re- tude as in C (25 nA) (mean channel conductance in this experiment, 5.9 moval of AcCho from the synaptic cleft was not closely related pS). Downloaded by guest on September 27, 2021 5128 Neurobiology: Baux and Tauc Proc. Natl. Acad. Sci. USA 80 (1983) A fact that carbachol can restore transmission after AcChoE in- 45 115 jection indicates that the release mechanism was not affected by AcChoE and that the depression was presumably due to re- moval of AcCho, as initially proposed (5). Also, the depression of the transmission by injected AcChoE was not due to the de- crease in size of the individual quanta but to the decrease in the oft AN ft) number of quanta released. Because AcChoE acted on the con- centration of cytoplasmic AcCho, this observation is contrary to the idea that AcCho is released from a cell through an AcCho 155 channel following its electrochemical gradient. An increase in cytoplasmic AcCho or carbachol concentration leads to an increase in the number of quanta released (Fig. 1B). One can ask whether this latter increase could be partly due to the depression of the late outward K+ current by the injected drugs, as reported for some quaternary ammonium ion com- pounds (20). Because some distance exists between the clamp B electrodes and the release site, the efficacy of the presynaptic voltage clamp characteristics might be thus changed. However, N- W lij1 injection of bethanechol, which differs from carbachol by one methyl group, induced an even stronger depression of the out- ward K+ current without significantly changing the amplitude of the response nor that of the MPSC. This suggests the changes in quantal release after the injection of the agonist are not re- FIG. 3. Depression ofthe postsynaptic response after intracellular lated to changes in the outward K+ current. injection of AcChoE and recovery after carbachol injection. Upper and Thus, the changes of LDIPSC amplitude after AcChoE and lower traces as in Fig. 1. (A) LDIPSC after presynaptic depolarization AcCho injection seem in favor of the idea that the intracellular to 0 mV. Time, in minutes. AcChoE was pressure injected at time 0 and concentration of cytoplasmic AcCho somehow regulates the re- the response at 45 min is identical to control. Statistical analysis in- dicated that the depression ofthe response was due to a decrease ofthe lease mechanism by acting on the number of quanta released. quantal content and that the size of individual MPSC remained un- Carbachol seems to be able to replace AcCho also in this re- changed (calibrations, upper traces, 2 nA; lower traces, 1 nA; time, 1 spect. s). (B) In another cell, 3 hr after AcChoE injection, the LDIPSC (left side) was depressed to a small value (presynaptic depolarization to + 10 We are grateful to Dr. K. Takeda for helpful comments. This work mV). Even stronger depolarizations did not increase the fluctuations. was supported in part by grants from Delegation Generale a la Re- Carbachol was then injected intothe presynaptic cell (arrow) (50 nA for cherche Scientifique et Technique and Institut National de la Sante et 20 min). At the end of injection, the reappearance ofthe LDIPSC showed de la Recherche Medicale to L.T. thatthesynaptic transmission was restored (calibrations, upper traces, 1 nA; lower traces, 0.5 nA; time, 1 s). 1. Collier, B., Boksa, P. & Lovat, S. (1979) Prog. Brain Res. 49, 107- 121. 2. Large, W. A. & Rang, H. P. (1978)J. Physiol (London) 275, 61P- response after a delay (Fig. 3A), as previously reported (5). The 62P. noise analysis showed that during the depression, the mean size 3. Large, W. A. & Rang, H. P. (1979) Prog. Brain Res. 49, 267-275. of the MPSC was not changed. Presumably, fewer quanta were 4. Whittaker, V. P. & Luqmani, Y. A. (1980) Gen. Pharmacol 11, 7- released. When the postsynaptic response, even to strongest 14. 5. Tauc, L., Hoffmann, A., Tsuji, S., Hinzen, D. H. & Faille, L. stimulation, was clearly depressed, injection of carbachol led (1974) Nature (London) 250, 496-498. to the restoration of transmission (Fig. 3B). 6. Gardner, D. (1971) Science 173, 550-553. 7. Simonneau, M., Tauc, L. & Baux, G. (1980) Proc. Natl, Acad. Sci. DISCUSSION USA 77, 1661-1665. 8. Katz, B. & Miledi, R. (1972)J. Physiol (London) 224, 665-700. Our observations indicate that carbachol can be released as a 9. Tauc, L. & Gerschenfeld, H. M. (1961) Nature (London) 192, 366- false transmitter from a cholinergic terminal. Part of the evi- 367. dence is that the amplitude of both the MPSC and postsynaptic 10. Neher, E. & Sakmann, B. (1976) Nature (London) 260, 799-802. single channel current increased when carbachol, instead of 11. Gardner, D. & Stevens, C. F. (1980)1 Physiol (London) 304, 145- AcCho, was involved. Whatever critical view is taken with re- 164. 12. Simonneau, M., Baux, G. & Tauc, L. (1980) Institut National de spect to the statistical analysis used, it is clear that following la Sante et de la Recherche Medicale Symposium, Number 13, ed. strictly identical experimental and analytical protocols, AcCho Taxi, J. (Elsevier, Amsterdam), pp. 179-189. and carbachol appear to act differently. For our purposes, it is 13. Ascher, P., Marty, A. & Neild, T. 0. (1978)J. Physiol. (London) not necessary to demonstrate that carbachol induces a channel 278, 177-206. current that is twice that induced by AcCho. It may be that this 14. Colquhoun, D., Dionne, V. E., Steinbach, J. H. & Stevens, C. difference is associated with some other receptor property, like F. (1975) Nature (London) 253, 204-206. 15. Feltz, A. & Trautmann, A. (1980) J. Physiol. (London) 299, 533- induction of bursting (16), of multiple openings (17), or of dif- 552. ferent conductance states as observed with single patch-clamped 16. Sakmann, B., Patlak, J. & Neher, E. (1980) Nature (London) 286, AcCho-activated channels (18). All of these could affect the var- 71-73. iance of the total response, resulting in an erroneous estimation 17. Colquhoun, D. & Sakmann, B. (1981) Nature (London) 294, 464- of the channel conductance. The slight decrease in the mean 466. open time for the Cl- channel activated carbachol was not 18. Hamill, 0. P. & Sakmann, B. (1981) Nature (London) 294, 462- by 464.- reflected in the decay time of the MPSC after intracellular car- 19. Katz, B. & Miledi, R. (1973)J. Physiol. (London) 231, 549-574. bachol injection. This may be related to the possibility of car- 20. Kmjevic, K. (1978) in lontophoresis and Transmitter Mechanisms bachol "rebinding" in the synaptic cleft (19). in the Mammalian Central Nervous System, eds. Ryall, R. W. & Several interesting conclusions come from this work. The Kelly, J. S. (Elsevier, Amsterdam), pp. 155-157. Downloaded by guest on September 27, 2021