It has been estimated that following the synaptic release of glutamate, the concentration in the synaptic cleft rises rapidly to around 1 mM and then decays with a time constant of about 1 ms (Clements, Lester, Tong, Jahr & Westbrook, 1992). Brief (1-5 ms) pulses of 1 mM glutamate to outside-out membrane patches containing N-methyl-D-aspartate (NMDA) channels yield currents that decay over a few hundred milliseconds, which resemble the NMDA component of an excitatory postsynaptic current (EPSC) (Lester, Clements, Westbrook & Jahr, 1990; Edmonds & Colquhoun, 1992). Given the short duration of time that glutamate is available for binding to postsynaptic receptors following release, and the slow rise and fall of currents through NMDA receptor channels, it is unlikely that an individual NMDA receptor would be activated more than once during a single synaptic event. The nature of these individual channel ‘activations’ is therefore considered to underlie the shape of the synaptic current (Lester et al. 1990; Gibb & Colquhoun, 1992; Lester & Jahr, 1992). The term ‘activation’ defines the sequence of events that takes place between the first opening that follows binding of agonist and the last opening before complete dissociation of agonist; the experimental estimate of this event is a ‘super-cluster’ (see Discussion for more rigorous definitions). The fact that two different molecules, a co-agonist and an agonist, are necessary to open the NMDA channel suggests that glycine as well as glutamate may determine the duration of NMDA single-channel activations. In all the experiments described here, however, glycine was used at a saturating concentration. Thus we expect the binding and unbinding of glutamate to the NMDA receptor complex (together with subsequent conformation changes) to determine the duration of receptor activations, as is presumably the case during synaptic transmission. Even under these conditions, however, the nature of such activations is complex, as illustrated by the equilibrium recordings of Gibb & Colquhoun (1991, 1992). They are far less well understood than, for instance, those of the nicotinic receptor at the neuromuscular junction (Colquhoun & Sakmann, 1985; Edmonds, Gibb & Colquhoun, 1995a,b), and no kinetic scheme is available that describes them adequately.
In this paper, we investigate the macroscopic response of NMDA receptors elicited by short (1 ms) pulses of glutamate and the structure of the super-clusters of channel activity elicited by very low glutamate concentrations, and explore the relationship between these two sorts of measurement. We have examined these issues for two recombinant NMDA receptors, namely NR1a/NR2A channels and NR1a/NR2D channels. In recordings from transfected human embryonic kidney (HEK) 293 cells, these two subunit combinations display the fastest and slowest offset rates, respectively, following the termination of a 300 ms application of 100 μM glutamate (Monyer, Burnashev, Laurie, Sakmann & Seeburg, 1994), or following brief pulses of glutamate (Vicini et al. 1998).
We find that the macroscopic current decay following a 1 ms concentration jump on NR1a/NR2D channels decays much more slowly than that mediated by NR1a/NR2A receptors. We reach the same conclusion by comparing NR1a/NR2A and NR1a/NR2D super-cluster durations recorded in the steady-state experiments employing very low agonist concentrations.