AMPA receptors predominantly containing the GluR4-flop subunit mediate fast excitatory synaptic transmission at the calyx of Held, a giant synapse in the auditory pathway. Whole-cell patch-clamp recordings were made to study transmission during trains of stimuli at this synapse. Lister Hooded rats (10-12 days old) were killed by decapitation and 250 µm thick slices of auditory brainstem made containing the medial nucleus of the trapezoid body (MNTB). MNTB neurons were loaded with 7 µM fura-2 AM for 5 min and imaged at 380 nm in order to locate cells with intact synaptic connections. EPSC trains (10-100 Hz) lasting 1 s were evoked via a bipolar platinum electrode placed on the midline and delivered five times at 30 s intervals. Evoked glutamatergic EPSCs underwent significant depression during the stimulus train, with more depression occurring at the higher stimulus frequencies. EPSC amplitudes at the end of a train of 10 or 100 Hz were 38 ± 4.3 and 12 ± 1.2 % of the initial EPSC amplitude, respectively (mean ± S.E.M.; n = 12 cells).

Bath application of the low-affinity, non-specific glutamate receptor antagonist γ-D-glutamylglycine (γ-DGG; 4 mM) reduced the amplitude of the first EPSC by 70 % at all stimulus frequencies. Synaptic depression observed during the train was reduced in γ-DGG. The amplitude of the last EPSC of a 50 Hz train was 50 ± 8.7 % of initial EPSC amplitude compared with 22 ± 3.8 % in the absence of γ-DGG, while at 10 Hz this was 43 ± 5.1 and 39 ± 4.9 %, respectively (n = 3 cells). In addition, facilitation of the early EPSCs was revealed for stimulus frequencies above 20 Hz by γ-DGG. The amplitude of the second EPSC at 50 Hz was 52 % larger in the presence of γ-DGG compared with that in the absence of γ-DGG, but at 10 Hz, the amplitude of the second EPSC was 85 % of the first in both conditions.

Application of 50 µM cyclothiazide (in the presence or absence of γ-DGG) slowed the decay time of each individual EPSC in the train but did not affect the EPSC amplitude or the rate of EPSC depression during the train at all stimulus frequencies (n = 4 cells). Our findings indicate that blockade of AMPA receptors by γ-DGG reveals a cyclothiazide-insensitive component to receptor desensitisation at physiological stimulus rates. This was confirmed using a computational model of transmission at this synapse, and suggests that studies of transmission may underestimate release probability during periods of high release in the presence of cyclothiazide.

This work was funded by The Wellcome Trust.