European Journal of Pharmacology, 179 (1990) 111-118
111
Elsevier EJP 51257
The effect of GABA on the frog optic rectum is sensitive to ammonium and to penicillin G e n i e Y. M a z d a , A n d r e a Nistri a n d L u c i a Sivilotti Department of Pharmacology, St. Bartholomew's Hospital Medical College, University of London, London ECI M 6BQ, U.K.
Received 19 October 1989, revised MS received 10 January 1990, accepted 23 January 1990
Excitatory postsynaptic potentials (termed U1 and U2) were extracellularly recorded from the frog optic tectum in vitro following electrical stimulation of the contralateral optic nerve. ~'-Aminobutyric acid (GABA) and glycine elicited a large and sustained enhancement of these synaptic waves. In the presence of the CI- transport inhibitor ammonium (NH~) the effects of GABA or glycine were progressively reduced to about 50% of their initial action without changes in the control synaptic waves. In 50% CI- media the depression of GABA and glycine responses by NH~ was more intense. Other CI- transport inhibitors such as bumetanide, piretanide and 4-acetamido-4'-isothiocyanatostilbene-2,2'-disulphonate (SITS) were inactive against responses to GABA or glycine. Penicillin, a C1channel blocker, antagonized the action of GABA and glycine, while increasing the amplitude of the U2 waveform. The present results provide pharmacological evidence in support of the CI- dependence of the unusual action of G A B A or glycine in facilitating excitatory synaptic transmission in the optic tectum. GABA (~,-arninobutyric acid); Glycine; Optic tectum; Ammonium; Penicillin; CI- transport; Excitatory synaptic transmission
1. Introduction While
the
action
of
3,-aminobutyric
acid
(GABA) on central neurones is usually inhibitory and mediated via G A B A A or GABA B receptors (Bormann, 1988), a major difference is found in the optic tectum of the frog. In this tissue, in fact, G A B A produces a large enhancement of excitatory synaptic transmission (Nistri and Sivilotti, 1985). A similar action of G A B A is also found in the guinea pig superior colliculus, the mammalian region homologous to the amphibian optic tectum (Arakawa and Okada, 1988). In both of these preparations, the concentration of bath-applied G A B A needed to evoke a half-maximal response Correspondence to: A. Nistri, Pharmacology Department, St. Bartholornew's Hospital Medical College, CharterhouseSquare, London EC1M 6BQ, U.K.
(EDs0) is around 100 /~M. While this value may seem higher than that reported in studies carried out in neuronal cultures (Segal and Barker, 1984), it is necessary to mention that in the majority of the latter investigations G A B A and its agonists were usually applied by pressure ejection from micropipettes, a method which does not easily allow an estimate of the drug concentration at the level of the neuronal membrane. The receptor pharmacology of the action of G A B A on the optic tectum is unusual since the effect of G A B A is insensitive to bicuculline or benzodiazepines, blocked by picrotoxin and not mimicked by baclofen (Sivilotti and Nistri, 1989). In essence, these atypical properties raise the possibility that in the visual system there is a novel population of G A B A receptors hitherto unknown. The question of how GABA can elicit these responses impinges upon the ionic mechanisms
0014-2999/90/$03.50 © 1990 Elsevier Science Publishers B.V. (Biomedical Division)
112 underlying GABA-mediated receptor activation. Although the effect of GABA on the optic tectum is dependent on the concentration of external C1(Sivilotti and Nistri, 1989), these experiments were conducted with extracellular microelectrode recordings which made difficult the precise identification of ionic species involved. While the option of prolonged intracellular recording during ionsubstitution experiments is presently precluded owing to the fragility and small size of tectal neurones (Antal et al., 1986), further proof of the role of CI- in the response to GABA may be obtained with two pharmacological approaches. The first one stems from the reliance of central neurones on membrane pumps to maintain an asymmetrical CI- distribution across their membrane (Thompson et al., 1988a). This active transport system can be specifically blocked by ammonium (NH~ ; Lux, 1971; Raabe and Gumnit, 1975) or loop diuretics such as bumetanide (Thompson et al., 1988b) and piretanide (Wojtowicz and Nicoll, 1982) which are more selective than their parent compound furosernide. These blockers were therefore tested on the electrophysiological response of tectal neurones to GABA. The second approach is based on the use of penicillin which is known to block C1- channels activated by GABA (Chow and Mathers, 1986); it was therefore deemed of interest to study whether this substance could alter the effect of GABA on the optic tectum.
2. Materials and methods
2.1. Electrophysiological recordings The experimental techniques have been described in detail by Nistri and Sivilotti (1985) and Sivilotti and Nistri (1989). In brief, a glass microelectrode (filled with 2 M NaC1 or Na acetate) was placed at a depth of 100 ~m in the in vitro optic tectum of the frog (R. temporaria) to record excitatory monosynaptic potentials (U1 and U 2 waves) elicited by electrical stimulation of the contralateral optic nerve. The brain tissue was constantly superfused with oxygenated Ringer solution of the following composition (mM): NaCl
111; KCI 2.5; CaCI2 2.5; NaHCO3 17; NaH2PO 4 • 2 H 2 0 0.1; glucose 4, at 6-7°C and pH 7.4. All test substances were applied via the superfusing solution. Electrophysiological responses were amplified, digitized and played back onto a linear pen recorder. Data are presented as means + S.E. with statistical analysis performed with the Student's t-test.
2.2. Materials The following test substances were used: GABA (BDH), glycine, ammonium chloride (NH4CI), penicillin (Na salt) and 4-acetamido-4'-isothiocyanatostilbene-2,2'-disulphonate (SITS)(Sigma). We gratefully acknowledge a gift of bumetanide from Leo Laboratories and piretanide from Hoechst. Stock solutions (50 mM) of bumetanide or piretanide were prepared in 96% ethanol. Small aliquots were then diluted in Ringer solution to the final required concentration. Equivalent amounts of ethanol (giving a final 0.1% concentration) were added to control solutions and were found to have no effect on synaptic transmission or amino acid responses. Piretanide solutions were always protected from light.
3. Results The database of this study comprises 24 preparations from which synaptic responses were successfully recorded for several hours. As previously reported (Sivilotti and Nistri, 1989), at a depth of 100/,tm from the pial surface the largest negative component of the synaptic field potential elicited by optic nerve stimulation is the U 1 wave (see fig. 1). This response represents monosynaptic synchronous activity triggered in a population of tectal neurones by a distinct class of unmyelinated axons of the optic nerve (Chung et al., 1974). The U, wave is followed by the shallower and prolonged U 2 wave (fig. 1) generated via a similar mechanism by a different group of optic nerve fibres. It must be noted that, at the low temperature used in the present study, the conduction velocity of U~ and U 2 fibres is 0.12 and 0.09 m / s (Sivilotti, 1988). Conduction along the axons of the optic nerve and
113
NR
GABA (1raM)
!
3os
9os
30s
905
WASH t
tJ 2
NH4(lmMI
J 0.5mV 40ms
Fig. 1. Effect of NH 2 on GABA-elicited responses of the frog optic tectum. Digitized chart records (DC coupled; negativity downwards) taken 30 or 90 s after application of GABA display the sustained enhancement of the U 1 and U2 synaptic waves (indicated by arrows). Top row shows records obtained in control Ringer solution (NR). Bottom row shows data after 50 rain in the presence of ammonium. Note that the effect of GABA is brief and fading occurs (see response at 90 s).
o p t i c tract (4 m m total length) a c c o u n t s for m u c h (33 a n d 45 ms, r e s p e c t i v e l y ) o f the t i m e - t o - p e a k for the U 1 a n d U 2 waves. A m a x i m a l l y e f f e c t i v e c o n c e n t r a t i o n of G A B A (1 r a M ) elicited, w i t h i n 30 s f r o m its a p p l i c a t i o n , an a v e r a g e i n c r e a s e o f 129 :t: 18 a n d 123 _+ 25% in the a m p l i t u d e o f t h e U t a n d U 2 waves, r e s p e c t i v e l y . U n l e s s o t h e r w i s e indic a t e d , results are p r e s e n t e d for the c h a n g e s in U]
N o t e also t h a t N H 2 d i d n o t a l t e r the a m p l i t u d e o f the U t a n d U 2 w a v e f o r m s in the a b s e n c e o f G A B A . In the p r e s e n c e o f NH~- t h e s t e a d y state r e s p o n s e s to 1 m M G A B A ( a f t e r 90 s e x p o s u r e ) w e r e signific a n t l y r e d u c e d to 58% o f their c o n t r o l s ( t a b l e 1). S u b m a x i m a l r e s p o n s e s to a l o w e r c o n c e n t r a t i o n o f
a m p l i t u d e . T h e U 2 w a v e closely f o l l o w e d a n y varia t i o n in U 1 size a l t h o u g h in a b s o l u t e t e r m s is was a much smaller synaptic response,
Effect of CI- transport inhibitors and low CI- media on the enhancement of the U 1 wave amplitude induced by GABA (1 mM) or glycine (0.3 mM). Results are expressed as fractional increases in U 1 wave amplitude normalized with respect to the action of 1 mM GABA in each experiment, n ~ number of experiments, a p < 0.005; b p ~< 0.05 calculated from the actual increases in U l wave observed in each group of experiments. In seven experiments, in which recordings were performed with 2 M Na acetate-filled microelectrodes to exclude the possibility that leakage of CI- from the electrode tip might influence the responses, results were the same as those obtained with NaCIfilled microelectrodes. Separate tests (not entered in table 1) showed that 1 h preincubation of the tissue with bumetanide or piretanide did not alter the decrease in GABA responsiveness brought about by 50% CI media.
3.1. Effects of CI- pump inhibitors In a n u m b e r o f c e n t r a l n e u r o n e s NH~- b l o c k s an active E l - t r a n s p o r t , thus d e p r e s s i n g s y n a p t i c inhibition and GABA-mediated responses (Lux, 1971; R a a b e a n d G u m n i t , 1975; H o t s o n a n d Prince, 1980). In the frog o p t i c t e c t u m a p p l i c a t i o n o f NH~- p r o d u c e d a d e c r e a s e in the e n h a n c i n g a c t i o n of G A B A (fig. 1). In p a r t i c u l a r , s u c h a d e c r e a s e was o b s e r v e d as an i n a b i l i t y by the tissue tO m a i n t a i n its r e s p o n s e to G A B A , f o l l o w i n g the initial b r i e f e n h a n c e m e n t o f the s y n a p t i c w a v e s ( c o m p a r e the e f f e c t o f G A B A at 30 a n d 90 s). H e n c e , by using NH,~ it was p o s s i b l e to i n d u c e a s t r o n g f a d e o f the a c t i o n o f G A B A , a p h e n o m e non normally absent under control conditions (top r o w of fig. 1; see also N i s t r i a n d Sivilotti, 1985).
TABLE l
Treatment
GABA
Glucine
n
Control solution NH,~ (1 mM) 50%c1 NH~- plus 50% C1Bumetanide (0.1 mM) Piretanide (0.5 raM) SITS(0.5 mM)
1 0.58±0.08 a 0.4]+0.20 b 0.21 ± 0.14 h 1.05 ±0.06 0.83 0.72±0.18
0.47 + 0.06 0.12±0.04 ~' 0.]5+0.04 0.08 +0.03 b 0.59 ±0.16 0.30 0.38:1:0.16
19 4 4 4 6 2 3
114 NH ÷ 4
GABA response
50% C l -
1
v
v
O.
O.
0.1 0
50
100
150
rain Fig. 2. Time course of decline in GABA-induced response following application of 1 m M N H ~ . Abscissa: elapsed time. Ordinate: log response evoked by 1 m M G A B A (1 = effect in control Ringer solution). The response decay from zero to 50 min is fitted by linear regression with a correlation coefficient r = - 0 . 9 5 6 8 and an estimated time constant of 72.9 rain. Data points are from four experiments (S.E. bars omitted when smaller than symbols). Note that the response decline saturated after 50 rain. Further reduction was achieved by using a low C1- solution.
G A B A (0.3 mM) were similarly decreased to 57 + 15% of their controls, indicating that the degree of response antagonism was indepedent of the agonist concentration. In order to check for a direct block of GABA receptors by N H ~ , it was necessary to test another putative transmitter, namely glycine (0.3 mM), which acts via its own specific receptors to enhance synaptic responses in a fashion similar to GABA (Sivilotti and Nistri, 1986). Table 1 shows that responses to glycine were also strongly reduced by N H ~ . Figure 2 shows the time course of the decline in the response to 1 mM G A B A following application of N H ~ . The decay of this response could be approximately fitted by a single exponential down to a value corresponding to half of its original size which was attained after 50 min superfusion with N H ~ . Longer exposures did not produce further depression of these responses. Exposure of the optic tectum to a low C1solution (in which 50% C I - was replaced by the presumably impermeant anion isethionate) elicited a depression of responses to GABA without altering the amplitude of control synaptic waves. The time course of the low C1- effect was similar to the one found in the presence of N H ~ (fig. 3) and reached an apparent equilibrium after 30 min without further decline with longer exposures
(Sivilotti and Nistri, 1989). At 30 rain in low C1medium the effect of G A B A was reduced to 41 + 2% of its initial value (table 1). Combination experiments were then performed to examine whether switching to a low C I - solution in the presence of N H ~ (after this drug had produced its GABA response
50%CI-
1.0 0.5
0.2
O. 1 0
10
min
20
30
Fig. 3. Time course of decline in GABA-induced response following exposure to 50% C1- solution. Abscissa: elapsed time. Ordinate: log response evoked by 1 mM GABA. The response decay is fitted by linear regression with a correlation coefficientr = -0.937 and an estimated time constant of 55.5 rain. Data are from four experiments. No further decline of the
response was found after 30 rain.
115
full effect) could completely block the response to GABA. Figure 2 shows that NH~ and low external C1- elicited a further slow depression of the GABA response down to 19 + 13% of its initial value. Similar data were obtained for the effect of glycine in the presence of NH~ a n d / o r 50% CImedia (table 1). Table 1 shows results obtained with other C1- transport inhibitors. Bumetanide (0.1 m M ) o r piretanide (0.5 m M ) h a d little effect on the tissue responses to GABA or glycine even after 1 h exposure to these diuretics. The compound SITS (0.5 mM), which is an inhibitor of anion exchange processes (Wieth, 1979; Thompson et al., 1988b), only slightly reduced the action of GABA. This reduction was, however, accompanied by a 33% decline in the amplitude of the control synaptic waves, suggestive of a non-specific depression of neuronal activity,
is known to produce half-maximal responses (Sivilotti and Nistri, 1989). Although in five experiments penicillin did not change the amplitude of the U 1 wave (0.76 + 0.07 mV in penicillin solution vs. 0.80 + 0.13 in control Ringer solution), a significant (P < 0.01) increase in the U2 wave was observed (0.78 + 0.18 vs. 0.45 + 0.13 mV in control conditions). Nevertheless, unlike the antagonism of GABA or glycine responses which returned to 60% of their initial value after 45 min washout, this augmentation of the U2 waveform persisted in spite of 1 h washout.
4. Discussion The principal finding of the present study is the sensitivity of GABA- or glycine-induced responses of the frog optic tectum to ammonium and penicillin, employed as pharmacological tools to test the CI- dependence of the action of these amino acids. The main effect of GABA on tectal neurones is to enhance their excitatory synaptic transmission via an unconventional receptor system (Nistri and Sivilotti, 1985; Sivilotti and Nistri, 1989). Glycine possesses a similar effect exerted via a distinct receptor type, although high concentrations of
3.2. Effects of penicillin This substance was applied at a concentration of 2 mM. Figure 4 shows that after 20 min exposure to penicillin a significant antagonism of the responses to GABA and glycine was present, In the case of GABA the degree of antagonism was similar (about 40%) against the maximally effective dose (1 mM) and the 0.1 mM dose which
AVN1
l
l
GABA
lmM
GABA
0.1mM
•
glycine 0.3 mM
Fig. 4. Effect of penicillin (2 mM) on tectal responses induced by GABA or glycine. Ordinate shows increases in U 1 amplitude (A VN) normalized with respect to the action of 1 mM GABA in each one of the five experiments. * P < 0.05. Note reduction of GABA and glycine effectiveness. El Control solution; [] penicillin solution.
116
glycine ( > 0.3 mM) elicit complex responses cornprising pre- and postsynaptic components (Sivilotti and Nistri, 1986). The ionic mechanism(s) underlying responses to GABA (or glycine) remains unclear and can only be solved by intracellular recording from tectal neurones, technically difficult to impale because of their small size (Antal et al., 1987; Sivilotti, 1988). The present study was therefore designed to seek pharmacological evidence in favour of C1- mediating the action of GABA, although it could not identify the direction of CI- flux via GABA receptoractivated channels, Vertebrate central neurones use active transport processes to achieve an unequal distribution of CI- across their cell membrane (Lux, 1971; Thompson et al., 1988a). Two types of CI- transport are known: one which relies on the co-transport of CI- with a cation (to preserve electroneutrality) and is blocked by N H g and loop diuretics (Lux, 1971; Thompson et al., 1988a,b). The other one operates via the exchange of CI- with anions and is blocked by SITS (Wieth, 1979; Thompson et al., 1988b). Since NH~ depressed the action of GABA on tectal neurones, the most likely explanation was that the reduced activity of GABA was due to NH~-induced disruption of the CItransport mechanism of these cells, with a consequent alteration in their transmembrane C1gradient. A similar phenomenon has previously been observed on mammalian and frog neurones (Lux, 1971; Nicoll, 1978). It seemed of interest that GABA-induced responses displayed strong fading in the presence of N H ~ . Previous experiments have shown that the action of GABA can lead to a rapid fall of the CI- gradient in central neurones (Segai and Barker, 1984). Presumably, tectal cells relied on a NH~-sensitive CI- pump to establish a large CI- gradient which was not strongly altered during the application of GABA. Only when such a pump was inhibited by N H ~ , a noticeable fading of GABA responses became apparent. It appears unlikely that NH~ was producing its effect via GABA receptor block because responses mediated by glycine receptor activation were also reduced. Nor was there evidence of any general neurodepressant or toxic action of NH~" since normal synaptic transmission was unaffected
by this compound. Furthermore, there is no experimental support for N H g as a direct blocker of C1- channels opened by GABA in mammalian (Lux, 1971) or frog (Nicoll, 1978) neurones; a channel-blocking effect of this compound would also be inconsistent with its positive charge which may be expected to be repelled by permeability channels which allow passage of anions. It was noteworthy that NH~ did not fully block the action of GABA, even if a more substantial block developed in low C1- medium. One possibility is that, in addition to CI-, other yet unidentified ions contributed to generate the response to GABA. Moreover, one can postulate that glial cells surrounding tectal neurones were buffering changes in external C1- in the confined space of synaptic junctions (Bormann and Kettenmann, 1988). Alternatively, other CI- transport mechanisms might have been implicated in the action of GABA. Nevertheless, various CI- transport blockers such as the diuretic agents bumetanide and piretanide, or SITS were little effective in blocking the action of GABA; these findings accord with the resistance of some central neurones to these substances (Wojtowicz and Nicoll, 1982; Thompson et al., 1988b) and indicate that in the brain the CI- carrier molecules possess transport sites pharmacologically distinct from those found in the kidney. Higher concentrations of loop diuretics were avoided to exclude their non-specific actions (Li and Kau, 1988) including block of C1- channels (lnomata et al., 1988). In the frog optic tectum the CI- transport mechanism was presumably still operating at the low ambient temperature used for recording: in keeping with this notion Nicoll (1978) reported that, in the frog spinal cord, cold temperature (unlike NH~" treatment) had comparatively less effect on the C1- gradient needed for GABA responses. High doses of penicillin are known to produce convulsions (Mandell and Sande, 1985). This effect is thought to involve antagonism of GABAmediated responses (Davidoff, 1972) via a block of CI- channels (Chow and Mathers, 1986). In the present study penicillin antagonized GABA and glycine responses, a result consistent with a CIchannel block by this antibiotic since it appears
117
that GABA and glycine act via activation of a
References
similar C1- channel (Bormann et al., 1987). The similarity between the GABA- and the glycine-operated channels does not extend to the entire structure of the receptor complex. In fact, in the frog optic t e c t u m a s in other preparations (Bor-
Antal, M., N. Matsumoto and G. Szekely, 1986, Tectal neurons of the frog: intracellular recording and labelling with cobalt electrodes, J. Comp. Neurol. 246, 238. Arakawa,T. and Y. Okada, 1988, Excitatory and inhibitory action of GABA on synaptic transmission in the guinea-pig superior colliculus slice, European J. Pharmacol. 158, 217. Bormann, J., 1988, Electrophysiology of GABA~, and GABA B receptor subtypes, Trends Neurosci. 11, 112. Bormann, .I., O.P. Hamill and B. Sakmann, 1987, Mechanism of anion permeation through channels gated by glycine and "r-aminobutyric acid in mouse cultured spinal neurones, J. Physiol. 385, 243. Bormann, J. and H. Kcnenmann, 1988, Patch-clamp study of y-aminobutyric acid receptor C1 channels in cultured astrcx:ytes, Pr~x:. Natl. Acad. Sci. U.S.A. 85, 9336. Chow, P. and D. Mathers, 1986, Convulsant doses of penicillin shorten the lifetime of GABA-induced channels in cultured central neurones, Br. J. Pharmacol. 88, 541. Chung, S.H., T.V.P. Bliss and M.J. Keating, 1974, The synaptic organization of optic afferent in the amphibian tectum,
mann, 1988) responses to GABA or glycine are selectively sensitive to the antagonists picrotoxin and strychnine, respectively (Nistri and Sivilotti, 1985; Sivilotti and Nistri, 1986). The reduction by penicillin of the maximal tissue response to 1 mM GABA also accords with the non-competitive n a l u r e o f its antagonism (Pickles and Simmonds, 1980). It was not unexpected to observe an increase in the U2 wave amplitude in the presence of penicillin. A similar phenomenon has been noted before (Sivilotti and Nistri, 1989) with picrotoxin or bicuculline (the latter being a very weak GABA antagonist in this tissue) and it seems unrelated to G A B A response blockade. It is therefore c o n ceivable that penicillin was enhancing the U2 waveform via other effects including increased excitability of some presynaptic nerve terminals (Prince, 1978). In conclusion, the present results suggest that GABA and glycine operate through a similar CImechanism responsible for their effects on tectal synaptic transmission. While the two amino acids seem to activate such a mechanism through distinct receptor types, it is tempting to speculate that in the optic tectum the C1- dependence of these responses implies homologous C1- channels. This hypothesis is consistent with data from m o u s e cultured neurones in which GABA and glycine appear to open similar C I - channels (Bormann et al., 1987). Patch clamp recording from frog tectal n e u r o n e s should help to address this issue directly.
Acknowledgements This work was supported by a grant from the Joint Research Board of St. Bartholomew's Hospital. L.S. was a fellow of ARIN. G.Y.M. was supported by the MRC and by the Medical College of St. Bartholomew's Hospital.
Proc. R. Soc. London Set. B 187, 421. Davidoff, R.A., 1972, Penicillin and inhibition in the cat spinal cord, Brain Res. 45, 636. Hotson, J.R. and D.A. Prince, 1980, A calcium-activated hyperpolarization follows repetitive firing in hippocampal neurons, J. Neurophysiol. 43, 409. Inomata, N., T. lshihara and N. Akaike, 1988, Effects of diuretics on GABA-gated chloride current in frog isolated sensory neurones, Br. J. Pharmacol. 93, 679. Li, J.H.-Y. and S.T. Kau, 1988, Bumetanide stimulation of sodium permeability of the apical membrane of toad urinary bladder. J. Pharmacol. Exp. Ther. 246, 980. Lux. H.D., 1971, Ammonium and chloride extrusion: hyperpolarizing synaptic inhibition in spinal motoneurons, Sci-
ence 173. 555.
Mandell, G.L. and M.A. Sande, 1985, Antimicrobial Agents, in: The Pharmacological Basis of Therapeutics, ed. A. Goodman Gilman, L.S. Goodman, T.W. Rail and F. Murad (Macmillan, New York) p. 1115. Nicoll, R.A., 1978, The blockade of GABA mediated responses in the frog spinal cord by ammonium ions and furosemide, J. Physiol. 283, 121. Nistri, A. and L. Sivilotti, 1985, An unusual effect of ~'aminobutyric acid on synaptic transmission of frog tectal neurones in vitro, Br. J. Pharmacol. 85, 917. Pickles, H.G. and M.A. Simmonds, 1980, Antagonism by penicillin of y-aminobutyric acid depolarizations at presynaptic sites in rat olfactory cortex and cuneate nucleus in vitro, Neuropharmacology 19, 35. Prince, D.A., 1978, Neurophysiology of epilepsy, Ann. Rev. Neurosci. 1,395. Raabe, W. and R.J. Gumnit, 1975, Disinhibition in cat motor cortex by ammonia, J. Neurophysiol. 38, 347. Segal, M. and J.L. Barker, 1984, Rat hippocampal neurons in culture: properties of GABA-activated C I ion conductance, J. Neurophysiol. 51,500.
118 Sivilotti, L.G., 1988, Pharmacological actions of neutral amino acids on synaptic transmission in the frog optic tectum (Ph.D. Thesis, University of London). Sivilotti, L. and A. Nistri, 1986, Biphasic effects of glycine on synaptic responses of the frog optic tectum in vitro, Neurosci. Lett. 66, 25. Sivilotti, L. and A. Nistri, 1989, Pharmacology of a novel effect of 7-aminobutyric acid on the frog optic tectum in vitro, European J. Pharmacol. 164, 205. Thompson, S.M., R.A. Deisz and D.A. Prince, 1988a, Relative contributions of passive equilibrium and active transport to
the distribution of chloride in mammalian cortical neurons, J. Neurophysiol. 60, 105. Thompson, S.M., R.A. Deisz and D.A. Prince, 1988b, Outward chloride/cation transport in mammalian cortical neurons, Neurosci. Lett. 89, 49. Wieth, J.O., 1979, Bicarbonate exchange through the human red blood cell membrane determined with [14C]bicarbonate, J. Physiol. 294, 521. Wojtowicz, J.M. and R.A. Nicoll, 1982, Selective action of piretanide on primary afferent GABA responses in the frog spinal cord, Brain Res. 236, 173.
111
Elsevier EJP 51257
The effect of GABA on the frog optic rectum is sensitive to ammonium and to penicillin G e n i e Y. M a z d a , A n d r e a Nistri a n d L u c i a Sivilotti Department of Pharmacology, St. Bartholomew's Hospital Medical College, University of London, London ECI M 6BQ, U.K.
Received 19 October 1989, revised MS received 10 January 1990, accepted 23 January 1990
Excitatory postsynaptic potentials (termed U1 and U2) were extracellularly recorded from the frog optic tectum in vitro following electrical stimulation of the contralateral optic nerve. ~'-Aminobutyric acid (GABA) and glycine elicited a large and sustained enhancement of these synaptic waves. In the presence of the CI- transport inhibitor ammonium (NH~) the effects of GABA or glycine were progressively reduced to about 50% of their initial action without changes in the control synaptic waves. In 50% CI- media the depression of GABA and glycine responses by NH~ was more intense. Other CI- transport inhibitors such as bumetanide, piretanide and 4-acetamido-4'-isothiocyanatostilbene-2,2'-disulphonate (SITS) were inactive against responses to GABA or glycine. Penicillin, a C1channel blocker, antagonized the action of GABA and glycine, while increasing the amplitude of the U2 waveform. The present results provide pharmacological evidence in support of the CI- dependence of the unusual action of G A B A or glycine in facilitating excitatory synaptic transmission in the optic tectum. GABA (~,-arninobutyric acid); Glycine; Optic tectum; Ammonium; Penicillin; CI- transport; Excitatory synaptic transmission
1. Introduction While
the
action
of
3,-aminobutyric
acid
(GABA) on central neurones is usually inhibitory and mediated via G A B A A or GABA B receptors (Bormann, 1988), a major difference is found in the optic tectum of the frog. In this tissue, in fact, G A B A produces a large enhancement of excitatory synaptic transmission (Nistri and Sivilotti, 1985). A similar action of G A B A is also found in the guinea pig superior colliculus, the mammalian region homologous to the amphibian optic tectum (Arakawa and Okada, 1988). In both of these preparations, the concentration of bath-applied G A B A needed to evoke a half-maximal response Correspondence to: A. Nistri, Pharmacology Department, St. Bartholornew's Hospital Medical College, CharterhouseSquare, London EC1M 6BQ, U.K.
(EDs0) is around 100 /~M. While this value may seem higher than that reported in studies carried out in neuronal cultures (Segal and Barker, 1984), it is necessary to mention that in the majority of the latter investigations G A B A and its agonists were usually applied by pressure ejection from micropipettes, a method which does not easily allow an estimate of the drug concentration at the level of the neuronal membrane. The receptor pharmacology of the action of G A B A on the optic tectum is unusual since the effect of G A B A is insensitive to bicuculline or benzodiazepines, blocked by picrotoxin and not mimicked by baclofen (Sivilotti and Nistri, 1989). In essence, these atypical properties raise the possibility that in the visual system there is a novel population of G A B A receptors hitherto unknown. The question of how GABA can elicit these responses impinges upon the ionic mechanisms
0014-2999/90/$03.50 © 1990 Elsevier Science Publishers B.V. (Biomedical Division)
112 underlying GABA-mediated receptor activation. Although the effect of GABA on the optic tectum is dependent on the concentration of external C1(Sivilotti and Nistri, 1989), these experiments were conducted with extracellular microelectrode recordings which made difficult the precise identification of ionic species involved. While the option of prolonged intracellular recording during ionsubstitution experiments is presently precluded owing to the fragility and small size of tectal neurones (Antal et al., 1986), further proof of the role of CI- in the response to GABA may be obtained with two pharmacological approaches. The first one stems from the reliance of central neurones on membrane pumps to maintain an asymmetrical CI- distribution across their membrane (Thompson et al., 1988a). This active transport system can be specifically blocked by ammonium (NH~ ; Lux, 1971; Raabe and Gumnit, 1975) or loop diuretics such as bumetanide (Thompson et al., 1988b) and piretanide (Wojtowicz and Nicoll, 1982) which are more selective than their parent compound furosernide. These blockers were therefore tested on the electrophysiological response of tectal neurones to GABA. The second approach is based on the use of penicillin which is known to block C1- channels activated by GABA (Chow and Mathers, 1986); it was therefore deemed of interest to study whether this substance could alter the effect of GABA on the optic tectum.
2. Materials and methods
2.1. Electrophysiological recordings The experimental techniques have been described in detail by Nistri and Sivilotti (1985) and Sivilotti and Nistri (1989). In brief, a glass microelectrode (filled with 2 M NaC1 or Na acetate) was placed at a depth of 100 ~m in the in vitro optic tectum of the frog (R. temporaria) to record excitatory monosynaptic potentials (U1 and U 2 waves) elicited by electrical stimulation of the contralateral optic nerve. The brain tissue was constantly superfused with oxygenated Ringer solution of the following composition (mM): NaCl
111; KCI 2.5; CaCI2 2.5; NaHCO3 17; NaH2PO 4 • 2 H 2 0 0.1; glucose 4, at 6-7°C and pH 7.4. All test substances were applied via the superfusing solution. Electrophysiological responses were amplified, digitized and played back onto a linear pen recorder. Data are presented as means + S.E. with statistical analysis performed with the Student's t-test.
2.2. Materials The following test substances were used: GABA (BDH), glycine, ammonium chloride (NH4CI), penicillin (Na salt) and 4-acetamido-4'-isothiocyanatostilbene-2,2'-disulphonate (SITS)(Sigma). We gratefully acknowledge a gift of bumetanide from Leo Laboratories and piretanide from Hoechst. Stock solutions (50 mM) of bumetanide or piretanide were prepared in 96% ethanol. Small aliquots were then diluted in Ringer solution to the final required concentration. Equivalent amounts of ethanol (giving a final 0.1% concentration) were added to control solutions and were found to have no effect on synaptic transmission or amino acid responses. Piretanide solutions were always protected from light.
3. Results The database of this study comprises 24 preparations from which synaptic responses were successfully recorded for several hours. As previously reported (Sivilotti and Nistri, 1989), at a depth of 100/,tm from the pial surface the largest negative component of the synaptic field potential elicited by optic nerve stimulation is the U 1 wave (see fig. 1). This response represents monosynaptic synchronous activity triggered in a population of tectal neurones by a distinct class of unmyelinated axons of the optic nerve (Chung et al., 1974). The U, wave is followed by the shallower and prolonged U 2 wave (fig. 1) generated via a similar mechanism by a different group of optic nerve fibres. It must be noted that, at the low temperature used in the present study, the conduction velocity of U~ and U 2 fibres is 0.12 and 0.09 m / s (Sivilotti, 1988). Conduction along the axons of the optic nerve and
113
NR
GABA (1raM)
!
3os
9os
30s
905
WASH t
tJ 2
NH4(lmMI
J 0.5mV 40ms
Fig. 1. Effect of NH 2 on GABA-elicited responses of the frog optic tectum. Digitized chart records (DC coupled; negativity downwards) taken 30 or 90 s after application of GABA display the sustained enhancement of the U 1 and U2 synaptic waves (indicated by arrows). Top row shows records obtained in control Ringer solution (NR). Bottom row shows data after 50 rain in the presence of ammonium. Note that the effect of GABA is brief and fading occurs (see response at 90 s).
o p t i c tract (4 m m total length) a c c o u n t s for m u c h (33 a n d 45 ms, r e s p e c t i v e l y ) o f the t i m e - t o - p e a k for the U 1 a n d U 2 waves. A m a x i m a l l y e f f e c t i v e c o n c e n t r a t i o n of G A B A (1 r a M ) elicited, w i t h i n 30 s f r o m its a p p l i c a t i o n , an a v e r a g e i n c r e a s e o f 129 :t: 18 a n d 123 _+ 25% in the a m p l i t u d e o f t h e U t a n d U 2 waves, r e s p e c t i v e l y . U n l e s s o t h e r w i s e indic a t e d , results are p r e s e n t e d for the c h a n g e s in U]
N o t e also t h a t N H 2 d i d n o t a l t e r the a m p l i t u d e o f the U t a n d U 2 w a v e f o r m s in the a b s e n c e o f G A B A . In the p r e s e n c e o f NH~- t h e s t e a d y state r e s p o n s e s to 1 m M G A B A ( a f t e r 90 s e x p o s u r e ) w e r e signific a n t l y r e d u c e d to 58% o f their c o n t r o l s ( t a b l e 1). S u b m a x i m a l r e s p o n s e s to a l o w e r c o n c e n t r a t i o n o f
a m p l i t u d e . T h e U 2 w a v e closely f o l l o w e d a n y varia t i o n in U 1 size a l t h o u g h in a b s o l u t e t e r m s is was a much smaller synaptic response,
Effect of CI- transport inhibitors and low CI- media on the enhancement of the U 1 wave amplitude induced by GABA (1 mM) or glycine (0.3 mM). Results are expressed as fractional increases in U 1 wave amplitude normalized with respect to the action of 1 mM GABA in each experiment, n ~ number of experiments, a p < 0.005; b p ~< 0.05 calculated from the actual increases in U l wave observed in each group of experiments. In seven experiments, in which recordings were performed with 2 M Na acetate-filled microelectrodes to exclude the possibility that leakage of CI- from the electrode tip might influence the responses, results were the same as those obtained with NaCIfilled microelectrodes. Separate tests (not entered in table 1) showed that 1 h preincubation of the tissue with bumetanide or piretanide did not alter the decrease in GABA responsiveness brought about by 50% CI media.
3.1. Effects of CI- pump inhibitors In a n u m b e r o f c e n t r a l n e u r o n e s NH~- b l o c k s an active E l - t r a n s p o r t , thus d e p r e s s i n g s y n a p t i c inhibition and GABA-mediated responses (Lux, 1971; R a a b e a n d G u m n i t , 1975; H o t s o n a n d Prince, 1980). In the frog o p t i c t e c t u m a p p l i c a t i o n o f NH~- p r o d u c e d a d e c r e a s e in the e n h a n c i n g a c t i o n of G A B A (fig. 1). In p a r t i c u l a r , s u c h a d e c r e a s e was o b s e r v e d as an i n a b i l i t y by the tissue tO m a i n t a i n its r e s p o n s e to G A B A , f o l l o w i n g the initial b r i e f e n h a n c e m e n t o f the s y n a p t i c w a v e s ( c o m p a r e the e f f e c t o f G A B A at 30 a n d 90 s). H e n c e , by using NH,~ it was p o s s i b l e to i n d u c e a s t r o n g f a d e o f the a c t i o n o f G A B A , a p h e n o m e non normally absent under control conditions (top r o w of fig. 1; see also N i s t r i a n d Sivilotti, 1985).
TABLE l
Treatment
GABA
Glucine
n
Control solution NH,~ (1 mM) 50%c1 NH~- plus 50% C1Bumetanide (0.1 mM) Piretanide (0.5 raM) SITS(0.5 mM)
1 0.58±0.08 a 0.4]+0.20 b 0.21 ± 0.14 h 1.05 ±0.06 0.83 0.72±0.18
0.47 + 0.06 0.12±0.04 ~' 0.]5+0.04 0.08 +0.03 b 0.59 ±0.16 0.30 0.38:1:0.16
19 4 4 4 6 2 3
114 NH ÷ 4
GABA response
50% C l -
1
v
v
O.
O.
0.1 0
50
100
150
rain Fig. 2. Time course of decline in GABA-induced response following application of 1 m M N H ~ . Abscissa: elapsed time. Ordinate: log response evoked by 1 m M G A B A (1 = effect in control Ringer solution). The response decay from zero to 50 min is fitted by linear regression with a correlation coefficient r = - 0 . 9 5 6 8 and an estimated time constant of 72.9 rain. Data points are from four experiments (S.E. bars omitted when smaller than symbols). Note that the response decline saturated after 50 rain. Further reduction was achieved by using a low C1- solution.
G A B A (0.3 mM) were similarly decreased to 57 + 15% of their controls, indicating that the degree of response antagonism was indepedent of the agonist concentration. In order to check for a direct block of GABA receptors by N H ~ , it was necessary to test another putative transmitter, namely glycine (0.3 mM), which acts via its own specific receptors to enhance synaptic responses in a fashion similar to GABA (Sivilotti and Nistri, 1986). Table 1 shows that responses to glycine were also strongly reduced by N H ~ . Figure 2 shows the time course of the decline in the response to 1 mM G A B A following application of N H ~ . The decay of this response could be approximately fitted by a single exponential down to a value corresponding to half of its original size which was attained after 50 min superfusion with N H ~ . Longer exposures did not produce further depression of these responses. Exposure of the optic tectum to a low C1solution (in which 50% C I - was replaced by the presumably impermeant anion isethionate) elicited a depression of responses to GABA without altering the amplitude of control synaptic waves. The time course of the low C1- effect was similar to the one found in the presence of N H ~ (fig. 3) and reached an apparent equilibrium after 30 min without further decline with longer exposures
(Sivilotti and Nistri, 1989). At 30 rain in low C1medium the effect of G A B A was reduced to 41 + 2% of its initial value (table 1). Combination experiments were then performed to examine whether switching to a low C I - solution in the presence of N H ~ (after this drug had produced its GABA response
50%CI-
1.0 0.5
0.2
O. 1 0
10
min
20
30
Fig. 3. Time course of decline in GABA-induced response following exposure to 50% C1- solution. Abscissa: elapsed time. Ordinate: log response evoked by 1 mM GABA. The response decay is fitted by linear regression with a correlation coefficientr = -0.937 and an estimated time constant of 55.5 rain. Data are from four experiments. No further decline of the
response was found after 30 rain.
115
full effect) could completely block the response to GABA. Figure 2 shows that NH~ and low external C1- elicited a further slow depression of the GABA response down to 19 + 13% of its initial value. Similar data were obtained for the effect of glycine in the presence of NH~ a n d / o r 50% CImedia (table 1). Table 1 shows results obtained with other C1- transport inhibitors. Bumetanide (0.1 m M ) o r piretanide (0.5 m M ) h a d little effect on the tissue responses to GABA or glycine even after 1 h exposure to these diuretics. The compound SITS (0.5 mM), which is an inhibitor of anion exchange processes (Wieth, 1979; Thompson et al., 1988b), only slightly reduced the action of GABA. This reduction was, however, accompanied by a 33% decline in the amplitude of the control synaptic waves, suggestive of a non-specific depression of neuronal activity,
is known to produce half-maximal responses (Sivilotti and Nistri, 1989). Although in five experiments penicillin did not change the amplitude of the U 1 wave (0.76 + 0.07 mV in penicillin solution vs. 0.80 + 0.13 in control Ringer solution), a significant (P < 0.01) increase in the U2 wave was observed (0.78 + 0.18 vs. 0.45 + 0.13 mV in control conditions). Nevertheless, unlike the antagonism of GABA or glycine responses which returned to 60% of their initial value after 45 min washout, this augmentation of the U2 waveform persisted in spite of 1 h washout.
4. Discussion The principal finding of the present study is the sensitivity of GABA- or glycine-induced responses of the frog optic tectum to ammonium and penicillin, employed as pharmacological tools to test the CI- dependence of the action of these amino acids. The main effect of GABA on tectal neurones is to enhance their excitatory synaptic transmission via an unconventional receptor system (Nistri and Sivilotti, 1985; Sivilotti and Nistri, 1989). Glycine possesses a similar effect exerted via a distinct receptor type, although high concentrations of
3.2. Effects of penicillin This substance was applied at a concentration of 2 mM. Figure 4 shows that after 20 min exposure to penicillin a significant antagonism of the responses to GABA and glycine was present, In the case of GABA the degree of antagonism was similar (about 40%) against the maximally effective dose (1 mM) and the 0.1 mM dose which
AVN1
l
l
GABA
lmM
GABA
0.1mM
•
glycine 0.3 mM
Fig. 4. Effect of penicillin (2 mM) on tectal responses induced by GABA or glycine. Ordinate shows increases in U 1 amplitude (A VN) normalized with respect to the action of 1 mM GABA in each one of the five experiments. * P < 0.05. Note reduction of GABA and glycine effectiveness. El Control solution; [] penicillin solution.
116
glycine ( > 0.3 mM) elicit complex responses cornprising pre- and postsynaptic components (Sivilotti and Nistri, 1986). The ionic mechanism(s) underlying responses to GABA (or glycine) remains unclear and can only be solved by intracellular recording from tectal neurones, technically difficult to impale because of their small size (Antal et al., 1987; Sivilotti, 1988). The present study was therefore designed to seek pharmacological evidence in favour of C1- mediating the action of GABA, although it could not identify the direction of CI- flux via GABA receptoractivated channels, Vertebrate central neurones use active transport processes to achieve an unequal distribution of CI- across their cell membrane (Lux, 1971; Thompson et al., 1988a). Two types of CI- transport are known: one which relies on the co-transport of CI- with a cation (to preserve electroneutrality) and is blocked by N H g and loop diuretics (Lux, 1971; Thompson et al., 1988a,b). The other one operates via the exchange of CI- with anions and is blocked by SITS (Wieth, 1979; Thompson et al., 1988b). Since NH~ depressed the action of GABA on tectal neurones, the most likely explanation was that the reduced activity of GABA was due to NH~-induced disruption of the CItransport mechanism of these cells, with a consequent alteration in their transmembrane C1gradient. A similar phenomenon has previously been observed on mammalian and frog neurones (Lux, 1971; Nicoll, 1978). It seemed of interest that GABA-induced responses displayed strong fading in the presence of N H ~ . Previous experiments have shown that the action of GABA can lead to a rapid fall of the CI- gradient in central neurones (Segai and Barker, 1984). Presumably, tectal cells relied on a NH~-sensitive CI- pump to establish a large CI- gradient which was not strongly altered during the application of GABA. Only when such a pump was inhibited by N H ~ , a noticeable fading of GABA responses became apparent. It appears unlikely that NH~ was producing its effect via GABA receptor block because responses mediated by glycine receptor activation were also reduced. Nor was there evidence of any general neurodepressant or toxic action of NH~" since normal synaptic transmission was unaffected
by this compound. Furthermore, there is no experimental support for N H g as a direct blocker of C1- channels opened by GABA in mammalian (Lux, 1971) or frog (Nicoll, 1978) neurones; a channel-blocking effect of this compound would also be inconsistent with its positive charge which may be expected to be repelled by permeability channels which allow passage of anions. It was noteworthy that NH~ did not fully block the action of GABA, even if a more substantial block developed in low C1- medium. One possibility is that, in addition to CI-, other yet unidentified ions contributed to generate the response to GABA. Moreover, one can postulate that glial cells surrounding tectal neurones were buffering changes in external C1- in the confined space of synaptic junctions (Bormann and Kettenmann, 1988). Alternatively, other CI- transport mechanisms might have been implicated in the action of GABA. Nevertheless, various CI- transport blockers such as the diuretic agents bumetanide and piretanide, or SITS were little effective in blocking the action of GABA; these findings accord with the resistance of some central neurones to these substances (Wojtowicz and Nicoll, 1982; Thompson et al., 1988b) and indicate that in the brain the CI- carrier molecules possess transport sites pharmacologically distinct from those found in the kidney. Higher concentrations of loop diuretics were avoided to exclude their non-specific actions (Li and Kau, 1988) including block of C1- channels (lnomata et al., 1988). In the frog optic tectum the CI- transport mechanism was presumably still operating at the low ambient temperature used for recording: in keeping with this notion Nicoll (1978) reported that, in the frog spinal cord, cold temperature (unlike NH~" treatment) had comparatively less effect on the C1- gradient needed for GABA responses. High doses of penicillin are known to produce convulsions (Mandell and Sande, 1985). This effect is thought to involve antagonism of GABAmediated responses (Davidoff, 1972) via a block of CI- channels (Chow and Mathers, 1986). In the present study penicillin antagonized GABA and glycine responses, a result consistent with a CIchannel block by this antibiotic since it appears
117
that GABA and glycine act via activation of a
References
similar C1- channel (Bormann et al., 1987). The similarity between the GABA- and the glycine-operated channels does not extend to the entire structure of the receptor complex. In fact, in the frog optic t e c t u m a s in other preparations (Bor-
Antal, M., N. Matsumoto and G. Szekely, 1986, Tectal neurons of the frog: intracellular recording and labelling with cobalt electrodes, J. Comp. Neurol. 246, 238. Arakawa,T. and Y. Okada, 1988, Excitatory and inhibitory action of GABA on synaptic transmission in the guinea-pig superior colliculus slice, European J. Pharmacol. 158, 217. Bormann, J., 1988, Electrophysiology of GABA~, and GABA B receptor subtypes, Trends Neurosci. 11, 112. Bormann, .I., O.P. Hamill and B. Sakmann, 1987, Mechanism of anion permeation through channels gated by glycine and "r-aminobutyric acid in mouse cultured spinal neurones, J. Physiol. 385, 243. Bormann, J. and H. Kcnenmann, 1988, Patch-clamp study of y-aminobutyric acid receptor C1 channels in cultured astrcx:ytes, Pr~x:. Natl. Acad. Sci. U.S.A. 85, 9336. Chow, P. and D. Mathers, 1986, Convulsant doses of penicillin shorten the lifetime of GABA-induced channels in cultured central neurones, Br. J. Pharmacol. 88, 541. Chung, S.H., T.V.P. Bliss and M.J. Keating, 1974, The synaptic organization of optic afferent in the amphibian tectum,
mann, 1988) responses to GABA or glycine are selectively sensitive to the antagonists picrotoxin and strychnine, respectively (Nistri and Sivilotti, 1985; Sivilotti and Nistri, 1986). The reduction by penicillin of the maximal tissue response to 1 mM GABA also accords with the non-competitive n a l u r e o f its antagonism (Pickles and Simmonds, 1980). It was not unexpected to observe an increase in the U2 wave amplitude in the presence of penicillin. A similar phenomenon has been noted before (Sivilotti and Nistri, 1989) with picrotoxin or bicuculline (the latter being a very weak GABA antagonist in this tissue) and it seems unrelated to G A B A response blockade. It is therefore c o n ceivable that penicillin was enhancing the U2 waveform via other effects including increased excitability of some presynaptic nerve terminals (Prince, 1978). In conclusion, the present results suggest that GABA and glycine operate through a similar CImechanism responsible for their effects on tectal synaptic transmission. While the two amino acids seem to activate such a mechanism through distinct receptor types, it is tempting to speculate that in the optic tectum the C1- dependence of these responses implies homologous C1- channels. This hypothesis is consistent with data from m o u s e cultured neurones in which GABA and glycine appear to open similar C I - channels (Bormann et al., 1987). Patch clamp recording from frog tectal n e u r o n e s should help to address this issue directly.
Acknowledgements This work was supported by a grant from the Joint Research Board of St. Bartholomew's Hospital. L.S. was a fellow of ARIN. G.Y.M. was supported by the MRC and by the Medical College of St. Bartholomew's Hospital.
Proc. R. Soc. London Set. B 187, 421. Davidoff, R.A., 1972, Penicillin and inhibition in the cat spinal cord, Brain Res. 45, 636. Hotson, J.R. and D.A. Prince, 1980, A calcium-activated hyperpolarization follows repetitive firing in hippocampal neurons, J. Neurophysiol. 43, 409. Inomata, N., T. lshihara and N. Akaike, 1988, Effects of diuretics on GABA-gated chloride current in frog isolated sensory neurones, Br. J. Pharmacol. 93, 679. Li, J.H.-Y. and S.T. Kau, 1988, Bumetanide stimulation of sodium permeability of the apical membrane of toad urinary bladder. J. Pharmacol. Exp. Ther. 246, 980. Lux. H.D., 1971, Ammonium and chloride extrusion: hyperpolarizing synaptic inhibition in spinal motoneurons, Sci-
ence 173. 555.
Mandell, G.L. and M.A. Sande, 1985, Antimicrobial Agents, in: The Pharmacological Basis of Therapeutics, ed. A. Goodman Gilman, L.S. Goodman, T.W. Rail and F. Murad (Macmillan, New York) p. 1115. Nicoll, R.A., 1978, The blockade of GABA mediated responses in the frog spinal cord by ammonium ions and furosemide, J. Physiol. 283, 121. Nistri, A. and L. Sivilotti, 1985, An unusual effect of ~'aminobutyric acid on synaptic transmission of frog tectal neurones in vitro, Br. J. Pharmacol. 85, 917. Pickles, H.G. and M.A. Simmonds, 1980, Antagonism by penicillin of y-aminobutyric acid depolarizations at presynaptic sites in rat olfactory cortex and cuneate nucleus in vitro, Neuropharmacology 19, 35. Prince, D.A., 1978, Neurophysiology of epilepsy, Ann. Rev. Neurosci. 1,395. Raabe, W. and R.J. Gumnit, 1975, Disinhibition in cat motor cortex by ammonia, J. Neurophysiol. 38, 347. Segal, M. and J.L. Barker, 1984, Rat hippocampal neurons in culture: properties of GABA-activated C I ion conductance, J. Neurophysiol. 51,500.
118 Sivilotti, L.G., 1988, Pharmacological actions of neutral amino acids on synaptic transmission in the frog optic tectum (Ph.D. Thesis, University of London). Sivilotti, L. and A. Nistri, 1986, Biphasic effects of glycine on synaptic responses of the frog optic tectum in vitro, Neurosci. Lett. 66, 25. Sivilotti, L. and A. Nistri, 1989, Pharmacology of a novel effect of 7-aminobutyric acid on the frog optic tectum in vitro, European J. Pharmacol. 164, 205. Thompson, S.M., R.A. Deisz and D.A. Prince, 1988a, Relative contributions of passive equilibrium and active transport to
the distribution of chloride in mammalian cortical neurons, J. Neurophysiol. 60, 105. Thompson, S.M., R.A. Deisz and D.A. Prince, 1988b, Outward chloride/cation transport in mammalian cortical neurons, Neurosci. Lett. 89, 49. Wieth, J.O., 1979, Bicarbonate exchange through the human red blood cell membrane determined with [14C]bicarbonate, J. Physiol. 294, 521. Wojtowicz, J.M. and R.A. Nicoll, 1982, Selective action of piretanide on primary afferent GABA responses in the frog spinal cord, Brain Res. 236, 173.
