Action of ammonium on a chloride pump. Removal of hyperpolarizing inhibition in an isolated neuron

Chloride distribution in Aplysia neurones

1. The intracellular Cl(-) concentration (Cl(i)) and the membrane potential (E(m)) were measured in the medial pleural neurones of Aplysia under various experimental conditions designed to determine the Cl(-) conductance of the neurones and investigate the possibility of an active Cl(-) transport.2. The magnitude of the Cl(-) conductance of the cell depends on the experimental conditions.3. In normal sea water, large changes of E(m) produced by passing current across the cell membrane caused no change of Cl(i), suggesting that the Cl(-) conductance was low. Similarly, moderate changes of E(Cl) produced by decreasing Cl(o) or increasing Cl(i) had little or no effect on E(m).4. A high Cl(-) conductance was observed in high K(o) or very low Cl(o). It was greatly reduced if the external Ca(2+) was replaced by Co(2+), or in the presence of tubocurarine, or if the experiment was performed on an isolated cell soma. The high Cl(-) conductance is therefore attributed to the release of ACh and perhaps other transmitters from synaptic terminals.5. High concentrations of tetraethylammonium ions or procaine induced a depolarization of the cell, but a decrease of Cl(i). The rate of fall of Cl(i) was increased by lowering external K(+) or raising external Ca(2+), and was decreased by replacing external Ca(2+) by Co(2+).6. NH(4) (+) ions applied externally had effects similar to those of K(+) ions. In situations in which intracellular NH(4) (+) might be increased a fall in Cl(i) was observed.7. The changes of Cl(i) caused by TEA, procaine, or internal NH(4) (+) occur against the driving force for passive Cl(-) movements. They are still observed in isolated cell bodies, and cannot be attributed to the activation of synaptic channels.8. Some interpretations of these anomalous Cl(-) movements are discussed which could also account for the difference between E(Cl) and E(m) observed under normal conditions.

Cation transport by the neuronal K(+)-Cl(-) cotransporter KCC2: thermodynamics and kinetics of alternate transport modes

Both Cs(+) and NH(4)(+) alter neuronal Cl(-) homeostasis, yet the mechanisms have not been clearly elucidated. We hypothesized that these two cations altered the operation of the neuronal K(+)-Cl(-) cotransporter (KCC2). Using exogenously expressed KCC2 protein, we first examined the interaction of cations at the transport site of KCC2 by monitoring furosemide-sensitive (86)Rb(+) influx as a function of external Rb(+) concentration at different fixed external cation concentrations (Na(+), Li(+), K(+), Cs(+), and NH(4)(+)). Neither Na(+) nor Li(+) affected furosemide-sensitive (86)Rb(+) influx, indicating their inability to interact at the cation translocation site of KCC2. As expected for an enzyme that accepts Rb(+) and K(+) as alternate substrates, K(+) was a competitive inhibitor of Rb(+) transport by KCC2. Like K(+), both Cs(+) and NH(4)(+) behaved as competitive inhibitors of Rb(+) transport by KCC2, indicating their potential as transport substrates. Using ion chromatography to measure unidirectional Rb(+) and Cs(+) influxes, we determined that although KCC2 was capable of transporting Cs(+), it did so with a lower apparent affinity and maximal velocity compared with Rb(+). To assess NH(4)(+) transport by KCC2, we monitored intracellular pH (pH(i)) with a pH-sensitive fluorescent dye after an NH(4)(+)-induced alkaline load. Cells expressing KCC2 protein recovered pH(i) much more rapidly than untransfected cells, indicating that KCC2 can mediate net NH(4)(+) uptake. Consistent with KCC2-mediated NH(4)(+) transport, pH(i) recovery in KCC2-expressing cells could be inhibited by furosemide (200 microM) or removal of external [Cl(-)]. Thermodynamic and kinetic considerations of KCC2 operating in alternate transport modes can explain altered neuronal Cl(-) homeostasis in the presence of Cs(+) and NH(4)(+).

Methionine sulfoximine, a glutamine synthetase inhibitor, attenuates increased extracellular potassium activity during acute hyperammonemia

Hyperammonemia causes glutamine accumulation and astrocyte swelling. Inhibition of glutamine synthesis reduces ammonia-induced edema formation and watery swelling in astrocyte processes. Ordinarily, astrocytes tightly control extracellular K+ activity [K+]e. We tested the hypothesis that acute hyperammonemia interferes with this tight regulation such that [K+]e increases and that inhibition of glutamine synthetase reduces this increase in [K+]e. Ion-sensitive microelectrodes were used to measure [K+]e in parietal cortex continuously over a 6-h period in anesthetized rats. After i.v. sodium acetate infusion in eight control rats, plasma ammonia concentration was 33 +/- 26 mumol/L (+/- SD) and [K+]e remained stable at 4.3 +/- 1.6 mmol/L. During ammonium acetate infusion in nine rats, plasma ammonia increased to 594 +/- 124 mumol/L at 2 h and to 628 +/- 135 mumol/L at 6 h. There was a gradual increase in [K+]e from 3.9 +/- 0.7 to 6.8 +/- 2.7 mmol/L at 2 h and 11.8 +/- 6.7 mmol/L at 6 h. In eight rats, L-methionine-D,L-sulfoximine (150 mg/kg) was infused 3 h before ammonium acetate infusion to inhibit glutamine synthetase. At 2 and 6 h of ammonium acetate infusion, plasma ammonia concentration was 727 +/- 228 and 845 +/- 326 mumol/L, and [K+]e was 4.5 +/- 1.9 and 6.1 +/- 3.8 mmol/L, respectively. The [K+]e value at 6 h was significantly less than that obtained with ammonium acetate infusion alone but was not different from that obtained with sodium acetate infusion. We conclude that acute hyperammonemia impairs astrocytic control of [K+]e and that this impairment is linked to glutamine accumulation rather than ammonium ions per se.

Ammonia: key factor in the pathogenesis of hepatic encephalopathy

There is substantial clinical and experimental evidence to suggest that ammonia toxicity is a major factor in the pathogenesis of hepatic encephalopathy associated with subacute and chronic liver disease. Ammonia levels in patients with severe liver disease are frequently found to be elevated both in blood and cerebrospinal fluid (csf). Hepatic encephalopathy results in neuropathological damage of a similar nature (Alzheimer type II astrocytosis) to that found in patients with congenital hyperammonemia resulting from inherited defects of urea cycle enzymes. Following portocaval anastomosis in the rat, blood ammonia concentration is increased 2-fold, and brain ammonia is found to be increased 2-3-fold. Administration of ammonia salts or resins to rats with a portocaval anastomosis results in coma and in Alzheimer type II astrocytosis. Since the CNS is devoid of effective urea cycle activity, ammonia removal by brain relies on glutamine formation. Cerebrospinal fluid and brain glutamine are found to be significantly elevated in cirrhotic patients with encephalopathy and in rats following portocaval anastomosis. In both cases, glutamine is found to be elevated in a region-dependent manner. Several mechanisms have been proposed to explain the neurotoxic action of ammonia. Such mechanisms include: Modification of blood-brain barrier transport; alterations of cerebral energy metabolism; direct actions on the neuronal membrane; and decreased synthesis of releasable glutamate, resulting in impaired glutamatergic neurotransmission.

Effects of ammonium salts on synaptic transmission to hippocampal CA1 and CA3 pyramidal cells in vivo

The effects of ammonium acetate or chloride, perfused through the lateral ventricle, were studied on the hippocampal formation of the rat. During perfusion with ammonia, the population spikes, evoked by stimuli delivered to the fimbria, were first increased and then reduced. On the other hand, the late positive wave gradually decreased throughout the application of ammonia. The inhibition, studied by the paired-pulse test, was found to be reduced when the population spike was transiently enhanced, indicating that disinhibition could be responsible for the enhancement of synaptically evoked responses. Neither antidromically evoked population spikes nor the typical effects of iontophoretically applied glutamate, aspartate or gamma-aminobutyrate were changed by ammonia. These findings can be accounted for by a single action of ammonia, a depression of excitatory synaptic transmission, the excitatory synapses on inhibitory interneurons being more readily depressed than those on the pyramidal cells. Both effects, early hyperexcitability and late depression, are probably due to a reduction in the release of the excitatory neurotransmitter, glutamate and/or aspartate. We tentatively suggest that these mechanisms are responsible for some of the symptoms observed during the development of hyperammonemic encephalopathies.

The ionic mechanism of postsynaptic inhibition in motoneurones of the frog spinal cord

The ionic mechanism of postsynaptic inhibition in frog spinal motoneurones was studied with conventional and with ion-sensitive microelectrodes. In these neurones the inhibitory postsynaptic potential was depolarizing, its reversal potential being 15 mV less negative than the resting membrane potential. During the inhibitory postsynaptic potential the input resistance of the motoneurones was reduced to 20% of the resting value, indicating a strong increase of membrane conductance. The Cl- equilibrium potential calculated from intra- and extracellular Cl- activity measurements coincided with the reversal potential of the inhibitory postsynaptic potential to within a few millivolts. During repetitive inhibitory postsynaptic activity the intracellular Cl- activity decreased markedly, while the extracellular Cl- activity increased slightly. These changes of intra- and extracellular Cl- activities were no longer found after suppression of the inhibitory postsynaptic potential by strychnine. Blockade of an active, inward-going Cl- transport system in motoneurones by NH+4 led to a shift of the Cl- equilibrium potential and the reversal potential of the inhibitory postsynaptic potential towards the resting membrane potential. After prolonged action of NH+4, the Cl- equilibrium potential approached the membrane potential to within 5 mV, while the reversal potential of the inhibitory postsynaptic potential and resting membrane potential coincided. The difference between Cl- equilibrium potential and membrane potential after blockade of the Cl- pump is traced back to interfering intracellular ions, such as HCO-3 or SO42-, leading to an overestimation of intracellular Cl- activity and to the calculation of an erroneous Cl- equilibrium potential. Inhibitory amino acids like gamma-aminobutyrate or beta-alanine evoked depolarizations with reversal potentials similar to that of the inhibitory postsynaptic potential. These depolarizations were associated with a marked decrease of neuronal input resistance during inhibition. During the actions of these compounds a decrease of intracellular and a small increase of extracellular Cl- activity were found. The activities of other ions (K+, Ca2+ and Na+) did not change significantly, with the exception of extracellular K+ activity, which was slightly increased. Evidence is presented that the inhibitory postsynaptic potential, as well as the depolarizing action of inhibitory amino acids in motoneurones, is the result of an increase in membrane Cl- permeability and an efflux of Cl- from these cells, while other ions do not seem to be involved.

Ammonia, postsynaptic inhibition and CNS-energy state

Ammonia intoxication decreases the hyperpolarizing action of postsynaptic inhibition. This study examines the metabolic state of the spinal cord during this effect of ammonia intoxication on spinal motoneurons. ATP, ADP, AMP, the adenylate energy charge, glucose, PCr, pyruvate, alpha-ketoglutarate and glutamate were unchanged during the effect of ammonia on the hyperpolarizing action of postsynaptic inhibition. NH4+, glutamine and lactate were increased. Ammonia intoxication affected postsynaptic inhibition without changes of the resting membrane potential, the neuron input resistance, the action potential and EPSPs. The encephalopathy caused by ammonia intoxication is known to occur without an alteration of the tissue energy state. The effect of ammonia intoxication on postsynaptic inhibition can be considered as a cause of the encephalopathy because postsynaptic inhibition is altered without a change of the tissue energy state, the resting membrane potential, the whole neuron resistance, the action potential and EPSPs.

The effects of temperature, pH and Cl-pump inhibitors on GABA responses recorded from cat dorsal root ganglia

GABA applied by iontophoresis produced GABA-induced currents (GCs) and GABA-induced depolarizations (GDs) which were recorded intracellularly from cat dorsal root ganglia (DRG). Lowering the temperature (37 to 27 degrees C) of the preparation depressed the amplitude of GCs while prolonging their rise-time and decay time. This depressant action was mainly due to a hyperpolarizing shift in the GABA equilibrium potential (EGABA). GABA responses could also be depressed by alkalinization of the superfusion solution or addition of putative chloride pump inhibitors, e.g. SITS, furosemide or bumetanide. However, the mechanism by which these latter procedures depressed GABA responses was not due to a shift in EGABA as occurred with lowered temperature. Instead we suggest that alkalinization or the putative chloride pump inhibitors affect the chloride channel or some other site associated with the GABA receptor complex and cause the depression we observed. GABA responses could be facilitated by lowering the pH of the superfusion solution or by injecting ammonium ion into a DRG. These results suggest that a temperature-sensitive, inwardly directed chloride pump that is resistant to SITS, furosemide or bumetanide, operates in cat DRG.

Effectiveness of some anions in sustaining the efferent inhibition in the frog labyrinth

Efferent inhibition in the frog labyrinth is sustained by the release of acetylcholine (ACh) which opens a Cl(-)-channel in the hair cell membrane. To investigate more closely the nature of the permeability change underlying the ACh reaction, the external Cl(-) was replaced by anions of increasing hydrated size, and to test the possible role of a Cl(-)-pump in the sensory cells, drugs were applied which are known to block active cl(-) pumping in other systems. Experiments indicate that the ACh-operated inhibitory channel of the hair cell is larger than at other inhibitory synapses (or approximately 0.7 nm), while pharmacological treatments (DNP, NaN3, acetazolamide, ammonium acetate, DIDS) fail to demonstrate any active distribution of Cl(-) across the hair cell membrane.

The role of intracellular chloride in hyperpolarizing post-synaptic inhibition of crayfish stretch receptor neurones

1. The intracellular Cl(-) activity (a(Cl) (i)) of isolated crayfish stretch receptor neurones was measured using liquid ion exchanger Cl(-)-selective micro-electrodes. The potential developed due to the difference between the normal extracellular Cl(-) activity (a(Cl) (o)) and a(Cl) (i) (V(Cl)) was compared with the simultaneously measured reversal potential of the inhibitory post-synaptic potential (E(i.p.s.p.)) to further clarify the ionic basis of the i.p.s.p..2. In normal Ringer solution, V(Cl) (63.3 +/- 2.3 mV) was found to be close to the resting membrane potential (E(m), 62.6 +/- 3.9 mV) while E(i.p.s.p.) (74.5 +/- 1.9 mV) was more negative than either. The V(Cl) value corresponds to an apparent a(Cl) (i) of 12.7 +/- 1.3 mM, which is about 4 mM more than required for a Cl(-) governed E(i.p.s.p.) of 74.5 mV.3. Reducing a(Cl) (o) caused smaller changes in V(Cl) than predicted for passive Cl(-) re-distributions. On complete removal of extracellular Cl(-) (Cl(o) (-)), V(Cl) increased to 84.6 +/- 2.7 mV, equivalent to an apparent a(Cl) (i) of about 5 mM-Cl(-). This value can be used as an estimate of the level of intracellular interference on the Cl(-)-selective micro-electrode.4. Increasing extracellular K(+) (K(0) (+)) decreased both V(Cl) and E(i.p.s.p.). Decreasing K(o) (+) had the converse effect. The time course of the changes in V(Cl) and E(i.p.s.p.) was much the same. The difference between V(Cl) and E(i.p.s.p.) decreased to about 3 mV in high K(o) (+), and increased to about 30 mV in low K(o) (+). This variation in the difference between E(i.p.s.p.) and V(Cl) is consistent with the assumption that anions other than Cl(-) contribute to the recorded V(Cl) rather than another ion contributes to the inhibitory current.5. Application of 5 mM-NH(4) (+) or of frusemide (6 x 10(-4) M) decreased V(Cl) and E(i.p.s.p.). The difference between V(Cl) and E(i.p.s.p.) was also decreased.6. We conclude that a(Cl) (i) is lower than predicted from a passive distribution and thus the chloride equilibrium potential (E(Cl)) is more negative than E(m). If a constant intracellular interference equivalent to about 4 mM-Cl(-) is assumed to contribute to the recorded V(Cl), E(Cl) was approximately equal to E(i.p.s.p.) in all the experimental conditions. Therefore we suggest that the i.p.s.p. is solely generated by Cl(-) ions.

Ammonium action on post-synaptic inhibition in crayfish neurones: implications for the mechanism of chloride extrusion

1. The reversal potential of the Cl(-)-dependent, inhibitory post-synaptic potential (E(i.p.s.p.)) was measured in the isolated crayfish stretch receptor neurone using two intracellular micro-electrodes. The difference between E(i.p.s.p.) and the resting membrane potential (E(m)), the i.p.s.p. driving force, was reversibly decreased by addition of NH(3)/NH(4) (+), and the mechanism of this decrease was investigated.2. The NH(3)/NH(4) (+)-induced decrease in i.p.s.p. driving force was dose-dependent with an onset at about 0.2 mM. E(i.p.s.p.) always remained more negative than E(m) or, when the neurone was spontaneously firing, the threshold potential. E(m) and resting membrane resistance (R(m)) also decreased in a dose-dependent fashion. Synaptic conductance (g(s)) increased with low doses, but decreased on application of 20 mM-NH(3)/NH(4) (+). All the effects were fully reversible on return to normal Ringer solution.3. Intracellular acidification (substitution of 50% Cl(-) by acetate compared with isethionate) considerably reduced the i.p.s.p. driving force. Simultaneous application of NH(3)/NH(4) (+) and acetate-substituted Ringer solution caused a similar decrease in the driving force to application of the same concentration of NH(3)/NH(4) (+) under normal conditions. Increasing the extracellular pH at which a given concentration of NH(3)/NH(4) (+) was applied caused a smaller decline in the i.p.s.p. driving force. These results suggest that intracellular acidification decreases the i.p.s.p. driving force and that the NH(3)/NH(4) (+)-induced decline is caused by an action of the ammonium ion.4. Elevation of extracellular K(+) (K(+) (0)) decreased the i.p.s.p. driving force, E(m) and R(m), and increased g(s). Reduction of K(+) (0) had the converse effects on all parameters.5. Application of Rb(+) or Cs(+) mimicked the effects of NH(3)/NH(4) (+). Substitution of K(+) (0) by Rb(+), Cs(+) or NH(3)/NH(4) (+) opposed or even reversed the increase in i.p.s.p. driving force when Na(+) was used as the substitute. The effectiveness of the various cations in decreasing the driving force was in the following order: Rb(+) > NH(4) (+) > K(+) > Cs(+).6. Inhibition of the Na pump by ouabain or K(+)-free Ringer solution caused a gradual reduction in the i.p.s.p. driving force. Since the driving force also decreased when the Na(+) gradient probably was increased (elevated K(+) (0)), this suggests a dependence on the K(+) gradient rather than the Na(+) gradient or the Na pump itself.7. Frusemide (6 x 10(-4) M) reversibly decreased the i.p.s.p. driving force and E(m), and increased g(s). R(m) was not significantly affected. Application of frusemide in the presence of 5 mM-Rb(+) and vice versa, caused a further reduction in the driving force. The recovery of the driving force on removal of either agent was slowed by the presence of the other.8. Application of 4,4-diisothiocyanostilbene-2,2-disulphonic acid (DIDS; 10(-4) M) caused spontaneous firing and reduced E(i.p.s.p.) to the threshold potential. R(m) and g(s) increased. The effects were slowly reversible on removal of the drug.9. It is proposed that the i.p.s.p. driving force is maintained by a K(+)-Cl(-) co-transport mechanism, driven by the K(+) gradient. The K(+) site exhibits the binding selectivity: Rb(+) > NH(4) (+) > K(+) > Cs(+) and the mechanism is inhibited partially by frusemide and completely by DIDS.

Effects of furosemide on neural mechanisms in Aplysia

The effects of furosemide on action potentials and responses to several neurotransmitters have been studied in the neurons of Aplysia. Furosemide (10(-7) and 10(-3) M) does not visibly affect the normal action potential in R15 neurons. However, when TTX (30 microM) is used to block the sodium component in R15, the remaining spike (presumably the calcium component) is increased in amplitude in the presence of furosemide. Furosemide also alters transmitter-induced conductances. Furosemide greatly reduces the amplitude and shifts, in a depolarizing direction, the reversal potential of chloride-dependent responses to gamma-aminobutyric acid (GABA) and acetylcholine (ACh). This suggests that furosemide both blocks the chloride channel and inhibits a chloride pump. ACh-induced sodium responses were also reduced by furosemide but to a lesser extent than chloride responses. The potassium response to ACh and a voltage-dependent calcium response to serotonin were not altered. These results indicate that furosemide could alter synaptic responses both presynaptically by enhancement of calcium flux during the action potential and postsynaptically by blockade of chloride and sodium conductances.

Ammonia and disinhibition in cat motor cortex by ammonium acetate, monofluoroacetate and insulin-induced hypoglycemia

Ammonia intoxication abolished the suppression of action potential generation by cortical postsynaptic inhibition due to the inactivation of neuronal Cl- extrusion. The disinhibition by ammonia intoxication occurred when ammonia concentrations in the cerebral cortex were increased to 320% of normal. Fluoroacetate poisoning and insulin-induced hypoglycemia, which are known to increase ammonia concentrations in the CNS and previously have been shown to inactivate Cl- extrusion in spinal motoneurons, abolished the suppression of action potential generation by cortical postsynaptic inhibition like ammonia intoxication. This disinhibition occurred at unchanged cerebral ammonia concentrations. The effect of fluoroacetate and insulin induced hypoglycemia on cortical postsynaptic inhibition is either due to a direct, i.e. not ammonia mediated, inactivation of neuronal Cl- extrusion or due to a disturbance of the synaptic mechanisms mediated by the transmitter of cortical inhibition, GABA. Toxic-metabolic encephalopathies which increase cerebral ammonia concentrations beyond 320% of normal may produce a dysfunction of the CNS due to inactivation of neuronal Cl- extrusion leading to ineffective cortical inhibition. However, in fluoroacetate poisoning and insulin-induced hypoglycemia increased ammonia concentrations in the CNS have only a secondary role in initiating a dysfunction of the CNS since disinhibition occurs before ammonia concentrations increase.

Crayfish stretch receptor: an investigation with voltage-clamp and ion-sensitive electrodes

1. The membrane characteristics of the slowly adapting stretch receptor from the crayfish, Astacus fluviatilis, were examined with electrophysiological techniques consisting of membrane potential recording, voltage clamp and ion-sensitive microelectrodes. 2. The passive membrane current (Ip) following step changes of the membrane potential to levels above 0 mV required more than a minute to decay to a steady-state level. 3. The stretch-induced current (SIC, where SIC = Itotal--Ipassive) was not fully developed until the Ip had decayed to a steady state. 4. With Ip at the steady state and the stretch-induced current at the O-current potential, a slow stretch-induced inward current was isolated. The latter reaches a maximum after 1 sec of stretch and declines even more slowly after stretch. The I-V relation of the slow current had a negative slope and reversed sign near the resting potential. It is suggested that this current is due to a Cl- conductance change. 5. The stretch-induced current, consisting of a rapid transient phase and a steady component can be isolated from the slow stretch-induced current at a holding potential corresponding to the resting potential. 6. The SIC-Em relation is non-linear and reverses sign at about +15 mV. 7. In a given cell, the reversal potential of the stretch-induced potential change obtained with current clamp coincided with the 0-current potential of the stretch-induced current obtained by voltage clamp. The average value from twenty-six cells was +13 +/- 6.5 mV; cell to cell variability seemed to be correlated with dendrite length. 8. Tris (mol. wt. 121) or arginine (mol. wt. 174) susbstituted for Na+ reduces but does not abolish the stretch-induced current. 9. The permeability ratios of Tris:Na and arginine:Na were estimated from changes in the 0-current potential as these cations replaced Na+ in the external medium. The PTris:PNa was somewhat higher (0.31) than the Parginine:PNa ratio (0.25). 10. Changes in the external Ca2+ concentration had no effect on the 0-current potential in Na or Tris saline. However, reducing Ca2+ did augment the stretch-induced current in either saline. A tenfold reduction of Ca2+ increased the conductance (at the 0-current level) about twofold. 11. Intracellular K+ and Cl- activities were obtained with ion sensitive electrodes. The average values from six cells were aiK = 133 +/- 34 mM and aiCl = 15.2 +/- 1.8 mM S.D.). EK was about 20 mV more negative than Em and ECl was about 10 mV more positive than Em. 12. aik and resting Em undergo large changes in K+-free solutions. After 60 min, ak was reduced eightfold and Em was reduced from -67 to -40 mV. Reduced Ca2+ in K+-free augments the rate of these changes. Receptor potential amplitude was also reduced in K+-free solution but could be restored upon polarizing the membrane to the pre-existing resting level.

The blockade of GABA mediated responses in the frog spinal cord by ammonium ions and furosemide

1. A variety of compounds which are known to block chloride transport in a variety of systems have been examined for their effects on amino acid and synaptic responses in the frog spinal cord in vitro. 2. A number of monocarboxylic aromatic acids, copper sulphate, and acetazolamide had no effect on any of the responses. 3. Ammonium ions blocked the motoneurone hyperpolarizing responses to all the neutral amino acids. In addition it selectively blocked dorsal root potentials and the action of GABA and beta-alanine on primary afferents. 5. Intracellular recording from dorsal root ganglion cells demonstrated that furosemide had little effect on the reversal potential for the GABA response. These results suggest that furosemide acts primarily by blocking the conductance increase elicited by GABA. 6. The results with furosemide provide indirect evidence that chloride ions are involved in generating the GABA depolarizations of primary afferent terminals and dorsal root potentials.