Anchors aweigh: NMDA receptors set sail from their presynaptic port

Abstract

NMDA receptors are unevenly expressed in neuronal cell membranes, reaching their highest concentration in the postsynaptic density (PSD), in register with presynaptic active zones. This confined distribution partly arises from biochemical interactions that the large cytosolic tail of NMDA receptors (> 600 amino acids) establishes with subsynaptic scaffold proteins of the PSD (PSD-95, SAP102), which ultimately anchor them to the intracellular cytoskeleton. In part, it is also due to neighbouring proteins and lipids that, through weak molecular interactions and steric hindrance, prevent NMDA receptors from drifting away from the PSD. All these interactions are transient and reversible in nature, and most NMDA receptors do not spend their entire lifetime docked at the postsynaptic harbour (Choquet & Triller, 2003). Occasionally, they remove the mooring and start sailing across extrasynaptic seas until they reach a neighbouring harbour where to spend a few seconds of their life. In cultured hippocampal neurons, at room temperature, optical tracking of Cy3-conjugated antibodies against the NR1 subunit shows that lateral diffusion of synaptic NMDA receptors is around 0.021 μm2 s−1, slower than that of extrasynaptic ones (0.037 μm2 s−1) (Groc et al. 2004). These estimates have recently been re-evaluated in view of the subunit composition of NMDA receptors with different subcellular locations (Groc et al. 2006). The proportion of immobile NMDA receptors is greater for those containing the NR2A subunit (Groc et al. 2006), although the biochemical mechanisms underlying this behaviour remain a matter of current investigation. It also remains to be established to what extent this correlates with different geometries of spine heads and necks or with developmental changes in the properties of the molecular interactions between NMDA receptors and neighbouring proteins. Since sailing needs a breeze, something over the cell membrane has to aid the NMDA receptor boat. Indeed, proteins of the extracellular matrix, like reelin, appear to favour NMDA receptor navigation (Groc et al. 2007). If postsynaptic NMDA receptors move, there is little reason, at least in principle, why they should not do so even at presynaptic sites. In the cortex, presynaptic NMDA receptors have been proposed to exist in specific layers of the entorhinal, visual and somatosensory region (Woodhall et al. 2001; Brasier & Feldman, 2008). This has been generally inferred from electrophysiological recordings where pharmacological block of NMDA receptors is associated with increase in the paired-pulse ratio of evoked EPSCs or decrease in miniature EPSC frequency. Whether these effects are directly consequent to prevention of Ca2+ influx through NMDA receptors or indirectly due to inhibition of a depolarization-induced recruitment of voltage-gated Ca2+ channels, they do indicate that at certain synapses NMDA receptor activation enhances the probability of neurotransmitter release. By using a similar approach in acute slices from adult rats, in layer V of the entorhinal cortex, Yang et al. (2008) (this issue of The Journal of Physiology) have recently caught sight of NMDA receptors cruising amidst the waves of presynaptic terminals. The conclusion is based on the observation that if postsynaptic NMDA receptors are blocked, frequency-dependent facilitation of evoked EPSCs and frequency of spontaneous EPSCs recover from the irreversible, use-dependent block by the NMDA receptor antagonist MK-801. A definitive proof of this interpretation would inevitably require a multidisciplinary effort, in which functional electrophysiological approaches are combined with anatomical evidence for presynaptic NMDA receptors and live imaging of their trajectories. But in the absence of yet-unidentified presynaptic anchoring proteins, this remains a feasible hypothesis provided the interactions with neighbouring proteins inside and underneath the cell membrane are not strong enough to constantly drive NMDA receptors aground. One would then speculate that lateral diffusion takes place between active zones, where neurotransmitter vesicles fusing with the presynaptic membrane continuously push receptors away, and the sites of exo-endocytosis, where receptor trafficking occurs. What implications on synaptic transmission would such a dynamic behaviour of presynaptic NMDA receptors have? It could work as a homeostatic mechanism to maintain constant the baseline level of release probability. For this to happen, lateral diffusion should occur within the time frame of NMDA receptor deactivation (∼100 ms) and a more quantitative analysis in the future will help to elucidate this scenario. In general though, it could provide neurons with an additional tool to regulate synaptic strength. It would be interesting to know whether NMDA receptor activation per se alters the mobility of receptors nearby: Ca2+ influx through NMDA receptors could lead to protein kinase C activation, which has been shown to increase their own mobility (Groc et al. 2004). Finally, I wonder whether glutamate escape from active terminals would also alter the dynamic behaviour of receptors at adjacent synapses. The notion of receptor mobility has definitely changed our understanding of synaptic function. No doubt the extraordinary adventures of NMDA receptors will soon lead us to better shores, with clearer views of the physiological processes studied so far.