Early biophysics of the NMDA receptor channel

Abstract

The voltage-dependent block of NMDA receptor channels by external Mg2+ (Mgo) was first reported in 1984 in two papers (Mayer et al. 1984; Nowak et al. 1984) that have had a great influence on our understanding of the functional role of NMDA receptors. However, the observations also offered a new way to explain the ‘gating’ of a voltage-dependent conductance, and our paper of 1988 published in The Journal of Physiology aimed primarily at a biophysical description of the Mgo block based on single channel records (Ascher & Nowak, 1988). To characterize the Mgo block we followed the approach used by Neher & Steinbach (1978) to study how QX222 transforms into ‘bursts’ the single openings of nicotinic ACh receptor channels. Assuming that Mgo enters the NMDA receptor channel, binds to a blocking site situated deep in the membrane and can only leave to the outside after unbinding, we evaluated the rates of Mg2+ binding and unbinding in various Mg2+ concentrations and at various potentials. We then deduced from the voltage dependence of these rates the depth of the blocking site in the membrane. This depth was evaluated by a coefficient ‘δ’ that could vary between 0 and 1. Our value of δ was close to 1, suggesting that the blocking site was actually very close to the inner limit of the membrane. This value was somewhat higher than the values obtained by analysis of the I–V relations of whole cell currents by Mayer & Westbrook (1985). We also characterized at the single channel level the Ca2+ permeability of the NMDA receptor channel. We measured the shifts of the reversal potential in different external Ca2+ concentrations and deduced the ratio of the permeabilities of Ca2+ and monovalent cations from the Goldman–Hodgkin–Katz voltage equation. Our results agreed with the values obtained by Mayer & Westbrook (1987) using I–V relations for whole cell current. We also observed that an increase in external Ca2+ reduced the single channel conductance, indicating that Ca2+ permeates the channel more slowly than monovalent cations. Our evaluation of the depth of the Mgo blocking site was soon put in doubt by the observation that the value of δ we deduced for Mgo block was not easily reconciled with the voltage dependence of the block by internal Mg2+ (Mgi) (Johnson & Ascher, 1990). The ‘crossing of the deltas’ paradox was solved by Jon Johnson and his collaborators, who showed that access of Mg2+ to the channel is prevented when permeant ions bind at the outer surface of the membrane. In the model of Antonov & Johnson (1999) the δ for Mgo is now equal to 0.5. We should also acknowledge that our single channel recordings made us miss the ‘slow Mgo unblock’ which was later described by Spruston et al. (1995) on whole cell current relaxations following voltage jumps, modelled by Vargas-Caballero & Robinson (2004) and by Kampa et al. (2004), and shown to be NR2 subunit dependent by Clarke & Johnson (2006). Recently the same authors (Clarke & Johnson, 2008) have shown that the slow block is the consequence of a voltage dependent gating which does not require Mgo. From 1991 onward, the cloning of the NMDA receptor subunits radically renewed the study of Mg2+ block and Ca2+ permeability. It rapidly led to the identification of the key amino acids involved in the binding of Mg2+ (reviewed by Dingledine et al. 1999), and it also revealed the heterogeneity of Mg block among NMDA receptors subtypes (not yet well understood). Concerning the Ca2+ permeability, the use of calcium indicators has allowed direct comparison of the Ca2+ influx and the total current and thus evaluation of the ‘fractional Ca2+ current’ (Pf). When appropriate corrections are made, the value of Pf agrees very well with the predictions of the GHK equation (Schneggenburger 1996). The molecular structures responsible for the Ca2+ permeability have been partially identified and comprise both a deep site, the N site of the NR1 subunit, and a superficial site at the entrance of the channel, the DRPEER motif, also specific to the NR1 subunit (Watanabe et al. 2002). Despite all these advances, one cannot yet say that either Mg2+ block or Ca2+ permeation are understood at the molecular level. We still lack a structural model of the NMDA receptor channel, but it may not be too far away.