There is controversy over the extent to which glutamate released at one synapse can diffuse to activate postsynaptic receptors at another nearby synapse. Uptake of glutamate helps to prevent such cross-talk, which could otherwise compromise the independence of signal transmission and information storage at individual synapses. We investigated the effect of preventing uptake by the glial transporters GLAST (knocked out in transgenic mice; Watase et al. 1998) and GLT-1 (blocked with 200 µM dihydrokainate), on parallel fibre EPSCs evoked at -70 mV in Purkinje cells from P14- to 21-day-old mice (killed humanely according to Home Office regulations).
If parallel fibre synapses operate independently then the EPSC waveform, and the effect on the waveform of blocking uptake, should be independent of the number of fibres stimulated. Experimentally, however, we found that when a single stimulus was applied in the molecular layer, the EPSC duration increased by 20 % as the number of fibres stimulated was increased from ~7 to ~90, and removal of GLAST (or block of GLT-1 when GLAST was knocked out) prolonged the EPSC much more when more fibres were stimulated. A similar effect of blocking GLT-1, dependent on the number of fibres stimulated, was seen both at 27 and at 36 °C. Thus there is cross-talk between parallel fibre synapses, even at a physiological temperature, and GLAST and GLT-1 curtail the EPSC by preventing glutamate diffusing between synapses.
When parallel fibre EPSCs were evoked by single stimuli to the granule cell layer, with the aim of producing a more physiological pattern of excitation (unlike for molecular layer stimulation which will activate many nearby fibres), the EPSC duration was shorter than for molecular layer stimulation, and was independent of the number of fibres stimulated, presumably because the activated fibres were separated by a distance too great for cross-talk to occur.
When trains of action potentials were evoked by molecular layer stimulation (200 Hz, 10 pulses), blocking uptake by GLAST (or blocking GLT-1 when GLAST was knocked out) dramatically prolonged the EPSC, even when only a small number of fibres were stimulated, presumably because the large amount of glutamate release saturated uptake by the remaining transporters present.
These results show that glial cell glutamate transporters increase the information processing capacity of the cerebellum by allowing neighbouring synapses to operate more independently.
We thank Professor K. Tanaka for the knock-out mice, and Stephen McGuiness for genotyping. This work was supported by a Marie-Curie Fellowship, the EU, The Wellcome Trust and a Wolfson-Royal Society Award.