Sites of antagonist action on N-methyl-D-aspartic acid receptors studied using fluctuation analysis and a rapid perfusion technique.

1. Mouse hippocampal neurons in dissociated culture were grown at low density on previously plated hippocampal glial cell cultures and voltage clamped using the tight seal whole-cell patch-clamp technique. Flow pipes were used to rapidly exchange the extracellular solution, and to apply N-methyl-D-aspartic acid (NMDA) and some NMDA antagonists. Fluctuation analysis was used to estimate changes in the behavior of NMDA-activated ion channels during application of antagonists. In the presence of NMDA control spectra were well fit by single Lorentzian functions consistent with mean open times of 5-6 ms. 2. Two antagonists thought to act at the NMDA receptor agonist recognition site, 2-amino-5-phosphonovaleric acid (AP5) and kynurenic acid, did not produce changes in the mean open time or single channel conductance, consistent with their action as competitive antagonists. Onset of antagonism and recovery from the action of both AP5 and kynurenic acid was rapid and complete within 1 s. However, raising the extra-cellular glycine concentration, from 1 microM to 1 mM, reduced the potency of 100 microM kynurenic acid as an NMDA antagonist, suggesting that kynurenate has an additional action as a competitive antagonist at the glycine modulatory site on NMDA receptor channels. 3. In the presence of 150 microM magnesium NMDA spectra recorded at -60 mV were fit by double Lorentzian functions, consistent with single-channel events consisting of bursts of openings lasting 3.3 ms in duration, interrupted by blocking and unblocking events of average duration 0.18 ms. The onset and recovery from magnesium antagonism was rapid, and complete within 1 s, but was highly voltage dependent and at +40 mV magnesium (150 microM) failed to produce NMDA antagonism. These results are consistent with a voltage-dependent channel block of NMDA receptor channels produced by binding of magnesium to a site within the ion channel. 4. Zinc (30 microM) was a potent NMDA antagonist at both -60 and +40 mV, and at either potential appeared to reduce the mean open time of NMDA-activated ion channels from about 5 ms to approximately 3 ms. Over the frequency range measured, 1-1,000 Hz, NMDA spectra were well fit by single Lorentzians during zinc antagonism, in contrast to results obtained with magnesium. The mean single channel conductance also decreased in the presence of zinc to approximately 75% of control. Onset of antagonism and recovery from the action of zinc was rapid and complete within 1 s.(ABSTRACT TRUNCATED AT 400 WORDS)