Contributions of space-clamp errors to apparent time-dependent loss of Mg2+ block induced by NMDA

N-methyl-d-aspartate receptors (NMDARs) govern synaptic plasticity, development, and neuronal response to insult. Prolonged activation of NMDARs such as during an insult may activate secondary currents or modulate Mg 2+ sensitivity, but the conditions under which these occur are not fully defined. We reexamined the effect of prolonged NMDAR activation in juvenile mouse hippocampal slices. NMDA (10 μM) elicited current with the expected negative-slope conductance in the presence of 1.2 mM Mg 2+ . However, several minutes of continued NMDA exposure elicited additional inward current at −70 mV. A higher concentration of NMDA (100 µM) elicited the current more rapidly. The additional current was not dependent on Ca 2+ , network activity, or metabotropic NMDAR function and did not persist on agonist removal. Voltage ramps revealed no alteration of either reversal potential or NMDA-elicited conductance between −30 mV and +50 mV. The result was a more linear NMDA current-voltage relationship. The current linearization was also induced in interneurons and in mature dentate granule neurons but not immature dentate granule cells, dissociated cultured hippocampal neurons, or nucleated patches excised from CA1 pyramidal neurons. Comparative simulations of NMDA application to a CA1 pyramidal neuron and to a cultured neuron revealed that linearization can be explained by space-clamp errors arising from gradual recruitment of distal dendritic NMDARs. We conclude that persistent secondary currents do not strongly contribute to NMDAR responses in juvenile mouse hippocampus and careful discernment is needed to exclude contributions of clamp artifacts to apparent secondary currents. NEW & NOTEWORTHY We report that upon sustained activation of NMDARs in juvenile mouse hippocampal neurons there is apparent loss of Mg 2+ block at negative membrane potentials. However, the phenomenon is explained by loss of dendritic voltage clamp, leading to a linear current-voltage relationship. Our results give a specific example of how spatial voltage errors in voltage-clamp recordings can readily be misinterpreted as biological modulation.