1. The whole-cell patch clamp recording technique was employed to investigate the intracellular regulation of N-methyl-D-aspartate (NMDA) receptors in cultured murine hippocampal neurones. Excitatory amino acids were repeatedly applied at regular intervals during intracellular dialysis with solutions of various composition. 2. Currents evoked by L-aspartate, an agonist of NMDA receptors, gradually 'washed out' to approximately 50% of their initial amplitude during dialysis with an intracellular solution containing CsCl and EGTA as a calcium buffer. In contrast, responses to kainate did not wash out. The wash-out of L-aspartate currents followed an exponential time course with a time constant of about 150 s. Wash-out did not appear to be related to desensitization of NMDA receptors. 3. Following wash-out, L-aspartate responses were blocked by Mg2+, ketamine or D-2-amino-5-phosphonovalerate indicating that these responses were still mediated by NMDA receptors. Furthermore, responses to NMDA itself showed wash-out to the same extent and with a time course similar to that for L-aspartate responses. 4. Neither the time course nor the extent of the wash-out of responses to L-aspartate was affected when the Ca2+ concentration of the dialysate was varied from zero to 1.5 x 10(-5) M. In addition, wash-out was unaffected by substitution of BAPTA for EGTA, indicating that wash-out was not a consequence of changes in intracellular pH related to the binding of Ca2+ to the buffer or to the kinetics of this binding. Therefore, the wash-out of NMDA currents could not be attributed to a gradual elevation of the concentration of intracellular Ca2+. 5. The extent of the wash-out of L-aspartate currents was similar for cells held at +40 versus -60 mV although the rate of wash-out was slower at the depolarized potential. In addition, the reversal potential of these currents was not altered, demonstrating that a change in driving force did not account for a component of the wash-out. 6. Inclusion of an ATP regeneration solution (Forscher & Oxford, 1985) in the dialysate prevented the wash-out of L-aspartate currents. ATP alone was less effective in preventing wash-out whereas phosphocreatine and creatine phosphokinase were ineffective by themselves. Wash-out also occurred when ATP was replaced with the non-hydrolysable analogue, beta, gamma-methyleneATP, or with GTP. In cells where wash-out of L-aspartate currents had been established, subsequent dialysis with the ATP regenerating solution partially reversed this wash-out.(ABSTRACT TRUNCATED AT 400 WORDS)