A slow NMDA channel: in search of a role

Recombinant NMDA receptors comprising NR1/NR2D subunits have unique properties when compared with other subtypes. These include low channel conductance with a characteristic pattern of transition between subconductance states and reduced sensitivity to Mg²⁺ blockade. Native NMDA receptors matching the low conductance and Mg²⁺ sensitivity of NR1/NR2D recombinant channels have been found in a few neuronal populations, in particular in developing Purkinje neurons (Momiyama et al. 1996). Ultra-rapid agonist application by means of a piezoelectric translator to excised outside-out membrane patches allows concentration jumps in the milliseconds range similar to those occurring during vesicular synaptic release. The high affinity of glutamate for NR1/NR2D receptors produces an extremely slow channel deactivation in response to brief glutamate applications (Monyer et al. 1994). This property has now been identified for the first time by Misra et al. (this issue of The The Journal of Physiology, Misra et al. 2000b) for native NMDA channels in mammalian neurons.

Excitatory synaptic transmission to cerebellar Purkinje neurons, unlike most other neurons in the brain, does not involve NMDA receptors. Therefore, the presence of NMDA receptors in these neurons suggests that they are extrasynaptic. Indeed, NMDA receptors with properties of NR1/NR2D recombinant receptors have been indirectly demonstrated, in cerebellar interneurons, to be localized at sites outside the synapse. These studies extend the growing evidence that the composition of synaptic receptors differs from that of extrasynaptic receptors in the cerebellum (Rumbaugh & Vicini, 1999; Misra et al. 2000a) and elsewhere. Additionally, the report by Misra et al. (2000b) implies that the presence of these extrasynaptic NMDA receptors is developmentally regulated because they are not found in the adult. Developmentally regulated extrasynaptic NMDA receptors have been recently reported in cerebellar granule neurons (Rumbaugh & Vicini, 1999). The specific regulation of synaptic and extrasynaptic receptors with distinct subunit composition during development is well established for nicotinic receptors at the neuromuscular junction and possibly similar mechanisms may take place at central synapses.

What is the role of NR1/NR2D receptors in cerebellar neurons? The two speculative interpretations raised by the authors are equally exciting. The first possibility is that extrasynaptic receptors perform a specific physiological role. Indeed NMDA receptors have been proposed to be essential in neurotrophic actions as well as in neuronal survival during development. Additionally, Ca²⁺ permeability and the negative slope conductance generated by Mg²⁺ blockade make these receptor-channels an essential component of the rhythmic oscillatory behaviour of membrane potential in several neurons. Whether extrasynaptic NR1/NR2D receptors in Purkinje neurons fulfil any of these roles has yet to be determined and will be the challenge of futures studies.

The second possibility raised by the authors is equally intriguing. In cerebellar neurons or elsewhere in the brain, synaptic channels may be slowed down by the expression of the NR2D subunit together with other NR2 subunits. Slowly decaying synaptic currents will increase the time window for activity-dependent synaptic reinforcement favouring synapse strengthening. Furthermore, the unique long-lasting channel activity may achieve a constant background current simply as a consequence of the spontaneous release of neurotransmitter occurring with miniature synaptic currents. While most studies show a developmental disappearance of the NR2D subunit, if it remains in combination with other subunits exclusively in synaptic receptors, it may not be easily identified with mRNA or protein analysis.

The presence of NR2D subunits in native receptors could possibly confer current deactivation lasting several seconds. In support of this, Misra et al. (2000b) show channel openings occurring as late as 30 s after brief glutamate applications. The characteristic pattern of transition between sub-conductance states convincingly demonstrates the presence of the NR2D subunit. Can slow deactivation alone be taken as a hallmark of the presence of NR2D subunits in native NMDA receptors? In dendritic patches from hippocampal neurons NMDA channel currents lasting seconds after brief glutamate applications have been described (Spruston et al. 1995). While it is possible that this channel activity arises from the presence of the NR2D subunit in the assembly of the receptor, the large conductance of these channels from dendritic recordings points to other interpretations. In fact, the conductance resulting from the combination of NR2D with other NR2 subunits results in a unique channel conductance unlike that of any other known subunit combination (Cheffings & Colquhoun, 1999). This, however, has yet to be demonstrated in native receptors. Clearly what is needed is a convincing demonstration of the formation of trimeric NMDA receptors including the NR2D subunit along with the study of their specific biophysical properties, in particular current deactivation in response to brief glutamate pulses.

In conclusion, the presence of slow deactivating currents in native neurons can be taken as evidence for the presence of the NR2D subunit if supported by single-channel conductance data. This work also lends support to the proposal that the subunit composition of the NMDA receptor subtype determines a sort of ‘molecular kinetic fingerprinting’ as has been proposed for GABA_A receptors (McClellan & Twyman, 1999).