The postnatal development of spontaneous GABAergic transmission between cerebellar Golgi cells and granule cells was investigated with voltage-clamp recording from rat cerebellar slices, in symmetrical Cl- conditions. Between postnatal days 7 and 14 (P7-14), bicuculline- and TTX (tetrodotoxin)-sensitive spontaneous inhibitory postsynaptic currents (sIPSCs), occurred at high frequency in 56% of granule cells. Between P10 and P14, sIPSCs were superimposed on a tonic current of -12 +/- 1.8 pA at -70 mV, that was accompanied by noise with a variance of 17 +/- 3 pA2. Both the current and noise were inhibited by bicuculline. TTX blocked the bicuculline-sensitive current and noise by approximately 60%. Between P18 and P25, sIPSCs were less frequent; all cells showed tonic, bicuculline-sensitive currents, but these were partially inhibited by TTX (approximately 35%). Between P40 and P53, sIPSCs were rare; tonic, bicuculline-sensitive currents and noise were greater in amplitude, with mean values of -17 pA and 22 pA2 at -70 mV, they were present in all cells but they were not inhibited by TTX. Glycine receptor channels that were expressed in immature, but not adult cells, did not mediate spontaneous currents. Our results indicate that spontaneous transmission onto cerebellar granule cells in immature animals consists primarily of action potential-dependent, phasic release of vesicular GABA. This generates GABAA receptor-mediated sIPSCs. The effects of GABA transporter blockers suggest that it also produces the TTX-sensitive current-noise, as GABA spills out of synapses to activate extrasynaptic receptors or receptors in neighbouring synapses. In older animals, action potential-independent release of transmitter is predominant and results in tonic activation of GABAA receptors. This does not appear to be spontaneous vesicular release of GABA. Neither does it appear to be reversed uptake of GABA, although further work is required to rule out these possibilities.

No abstract

In the mammalian central nervous system amino acids such as L-glutamate and L-aspartate are thought to act as fast synaptic transmitters. It has been suggested that at least three pharmacologically-distinguishable types of glutamate receptor occur in central neurons and that these are selectively activated by the glutamate analogues N-methyl-D-aspartate (NMDA), quisqualate and kainate. These three receptor types would be expected to open ion channels with different conductances. Hence if agonists produce similar channel conductances this would suggest they are acting on the same receptor. Another possibility is suggested by experiments on spinal neurons, where GABA (gamma-amino butyric acid) and glycine appear to open different sub-conductance levels of one class of channel while acting on different receptors. By analogy, several types of glutamate receptor could also be linked to a single type of channel with several sub-conductance states. We have examined these possibilities in cerebellar neurons by analysing the single-channel currents activated by L-glutamate, L-aspartate, NMDA, quisqualate and kainate in excised membrane patches. All of these agonists are capable of opening channels with at least five different conductance levels, the largest being about 45-50 pS. NMDA predominantly activated conductance levels above 30 pS while quisqualate and kainate mainly activated ones below 20 pS. The presence of clear transitions between levels favours the idea that the five main levels are all sub-states of the same type of channel.

L-GLUTAMATE and L-aspartate are thought to have a widespread function as synaptic transmitters in the mammalian central nervous system and there are at least three types of neuronal glutamate receptors, which can be activated by the selective agonists N-methyl-D-aspartate (NMDA), quisqualate and kainate. Recent experiments indicate that glutamate receptors also occur in astrocytes. We have used patch-clamp methods to determine whether one type of macroglial cell, the type-2 astrocyte, possesses glutamate receptors, as previously proposed from neurochemical studies. We find that glutamate and related amino acids can evoke whole-cell and single-channel currents in type-2 astrocytes from rat cerebellum. Although these cells are found mainly in white matter, where neurotransmission does not occur, their processes are closely associated with axons at nodes of Ranvier, suggesting that such receptors are involved in neuronal-glial signalling at the node. Our experiments show that glial cells possess quisqualate- and kainate-receptor channels but lack receptors for NMDA. Interestingly, these glutamate channels exhibit multiple conductance levels that are similar in amplitude to the neuronal glutamate channels.

SUMMARY1. Two clones of rat phaeochromocytoma PC12 cells have been used to study the expression of Ca2+ channels and their possible involvement in neuronal differentiation. One clone differentiated morphologically when exposed to nerve growth factor (NGF) for 4 days (PC12 cells), while the other clone was insensitive to NGF, but differentiated morphologically in the presence of ouabain (041 mM

In many studies of central synaptic transmission, the quantal properties of miniature synaptic events do not match those derived from synaptic events evoked by action potentials. Here we show that at mossy fiber-granule cell (MF-gc) synapses of mature cerebellum, evoked excitatory postsynaptic currents (EPSCs) are multiquantal, and their amplitudes vary in discrete steps, whereas miniature (m)EPSCs are monoquantal or multiquantal with quantal parameters identical to those of the EPSCs. In contrast, at immature MF-gc synapses, EPSCs are multiquantal, but their amplitudes do not vary in discrete steps, whereas most mEPSCs seem to be monoquantal with a broad and skewed amplitude distribution. The results demonstrate that quantal variance decreases during synaptic development. They also directly confirm the quantal hypothesis of neurotransmission at a mature brain synapse.

Several factors contribute to the shape of excitatory postsynaptic currents (EPSCs) in CNS neurons, among them the kinetics of presynaptic release, transmitter clearance, and the properties and distribution of postsynaptic receptors. The decays of AMPA receptor-mediated EPSCs at rat cerebellar mossy fibre-granule cell (MF-gc) synapses follow a bi-exponential time-course. The fast component dominates the decay, accounting for 84-94% of the peak amplitude. Here we show that both components of decay, and also the risetimes, became faster during postnatal maturation. At adult, but not immature, synapses, the risetimes and decays of evoked multiquantal EPSCs were similar to those of monoquantal miniature (m)EPSCs. The faster risetimes at mature synapses reflected increased synchrony of multivesicular release, whereas the faster decays appeared to reflect changes in the properties of postsynaptic receptors. Inhibition of glutamate uptake was without effect on evoked EPSCs at both ages. Furthermore, after slowing receptor desensitization with cyclothiazide, the EPSCs at mature synapses decayed as slowly as EPSCs at immature synapses, suggesting that faster glutamate clearance does not account for the developmental speeding of EPSC decay. Our results support previous conclusions that glutamate clearance and receptor deactivation are important determinants of the fast decay component at immature synapses. Desensitization becomes increasingly important during development and plays a major role in shaping EPSC decay at mature synapses.

3. From the relative areas under current amplitude histograms it was estimated that the percentage of openings with conductances greater than 30 pS was about 83 % with NMDA, 79 % with glutamate and 78 % with aspartate. In some patches, the majority of > 30 pS events evoked by these agonists were to the maximum conductance of 48 pS, whereas in other patches there were more 38 pS openings than 48 pS openings. Only 27 % of quisqualate openings, and about 10 % of kainate openings, were > 30 pS.4. Of the small amplitude (< 20 pS) events, 93 % of quisqualate openings were to the 8 pS level whereas -87 % of < 20 pS currents produced by NMDA, glutamate and aspartate were to the 18 pS level (the remainder being 8 pS). Direct transitions could occur between certain levels (including events above and below 30 pS) suggesting that these are sublevels of multiple-conductance channels. The most frequently occurring transitions were between the 48 and 38 pS levels, and the 38 and 18 pS levels.5. Channel openings occurred in bursts, within which individual openings were separated either by brief closures (gaps), or by direct transitions between the multiple conductance levels. The briefest of these gaps (< 200-400 7. XVhen applied to isolated patches, kainate (and to a lesser extent quisqualate) gave a noise increase, accompanied in some patches by discrete single-channel openings (similar in conductance to those produced by other agonists). Analysis of kainate currents in outside-out patches gave two component spectra with time constants of 0 9 + 0d and 16 4 + 2-4 ms (E -=70 mV). The mean single-channel conductance estimated from the spectra was y(kainate) = [03 +02 pS. The fast component contributed most power to spectra of kainate and quisqualate singlechannel currents.8. Our results may be explained by assuming that NMDA, quisqualate and kainate receptors are coupled to separate channels, and that some, or all, of these receptor channels possess multiple conductance levels. All the channels appear to show two kinetically distinct open states, and two (or three) shut states. Differences in the properties of glutamate channels present in large cerebellar neurones and in certain other neurones are discussed.