Rundown of N-methyl-D-aspartate channels during whole-cell recording in rat hippocampal neurons: role of Ca2+ and ATP.

1. N-methyl-D-aspartate (NMDA) channel activity was studied on cultured rat hippocampal neurons in whole-cell voltage-clamp mode. NMDA responses were evoked by rapid application of NMDA and the cytosol was modified using pipette dialysis and intracellular perfusion. 2. In the presence of 2 mM [Ca2+]o with 2.4 mM BAPTA (1,2-bis(O-aminophenoxy)ethane-N,N,N',N'-tetraacetic acid) and 0.4 mM Ca2+ in the whole-cell pipette, the response evoked by regular applications of 10 microM NMDA gradually decreased during prolonged whole-cell recording. After 25 min the peak current was reduced to 56 +/- 1.6% of control. Channel 'rundown' could be prevented by inclusion of an ATP regenerating solution in the pipette. 3. Rundown did not occur in Ca(2+)-free medium even in the absence of added ATP regenerating solution. Rundown was also prevented by increasing [BAPTA]i to 10 mM whereas raising [Ca2+]i by inhibiting the Na(+)-Ca2+ exchanger or by perfusing the patch pipette with high [Ca2+]i (15-1000 microM) reversibly inhibited the NMDA current. By contrast, the rundown of kainate responses was Ca(2+)-independent. 4. The rate and reversibility of rundown was use-dependent. Rundown did not occur with infrequent NMDA applications (0.2/min). Following channel rundown in Ca(2+)-containing medium, a 5 min pause in agonist applications or adding ATP regenerating solution by intracellular perfusion resulted in complete recovery. However, rundown did not recover following large currents evoked by 300 microM NMDA or when 10 mM EGTA was used as the intracellular buffer. Protease inhibitors did not prevent irreversible rundown. 5. ATP-gamma-S (4 mM) was less effective than the ATP regenerating solution in preventing rundown. Likewise, intracellular dialysis with alkaline phosphatase, phosphatase 1 or calcineurin did not induce rundown and addition of phosphatase inhibitors also did not block rundown. Thus receptor dephosphorylation did not appear to be primarily responsible for channel rundown. 6. The mean open time and unitary conductance of the NMDA channel were unaffected by rundown as estimated by fluctuation analysis. The conductance was 42.8 +/- 2.9 nS before and 43.7 +/- 2.8 nS after rundown. The mean open times were 17.3 and 4.0 ms before and 15.9 and 4.0 ms after rundown. However the open probability was reduced following rundown as determined by the onset of MK-801 block of steady-state NMDA currents. 7. Our results suggest that an increase in intracellular calcium leads to channel rundown during whole-cell recording by reducing the open probability of the NMDA channel.(ABSTRACT TRUNCATED AT 400 WORDS)