Abstract
The NMDA receptor is a Ca(2+)-permeant glutamate receptor which plays key roles in health and disease. Canonical NMDARs contain two GluN2 subunits, of which 2A and 2B are predominant in the forebrain. Moreover, the relative contribution of 2A vs. 2B is controlled both developmentally and in an activity-dependent manner. The GluN2 subtype influences the biophysical properties of the receptor through difference in their N-terminal extracellular domain and transmembrane regions, but they also have large cytoplasmic Carboxyl (C)-terminal domains (CTDs) which have diverged substantially during evolution. While the CTD identity does not influence NMDAR subunit specific channel properties, it determines the nature of CTD-associated signalling molecules and has been implicated in mediating the control of subunit composition (2A vs. 2B) at the synapse. Historically, much of the research into the differential function of GluN2 CTDs has been conducted in vitro by over-expressing mutant subunits, but more recently, the generation of knock-in (KI) mouse models have allowed CTD function to be probed in vivo and in ex vivo systems without heterologous expression of GluN2 mutants. In some instances, findings involving KI mice have been in disagreement with models that were proposed based on earlier approaches. This review will examine the current research with the aim of addressing these controversies and how methodology may contribute to differences between studies. We will also discuss the outstanding questions regarding the role of GluN2 CTD sequences in regulating NMDAR subunit composition, as well as their relevance to neurodegenerative disease and neurodevelopmental disorders.