The elusive roles of NMDA receptor amino-terminal domains

The influx of ions into the neuron forms the requisite step for the passage of information from one neuron to the other neuron. The functional units of the neuron – ion channels – carry out this essential step efficiently and in a highly coordinated manner. Generally, a neuron expresses a variety of ion channels that permit the movement of specific ions. However, the fidelity for the flow of specific ions through a particular ion channel is often defined by the biophysical properties of that ion channel such as activation and inactivation, etc. Interestingly, many of the biophysical properties of an ion channel are acquired by protein domains of that ion channel. This form of dependence on multiple intrinsic factors for given biophysical properties gets complicated if that ion channel has numerous protein domains and each protein domain functions as a semiautonomous entity. The N-methyl-d-aspartate receptors (NMDARs) are one such particular ion channel that has distinct profiles of biophysical properties due to the presence of several different semiautonomous protein domains (Traynelis et al. 2010).

NMDARs are glutamate-gated cation channels that are ubiquitously expressed in various regions of the brain and are involved in synaptic transmission. Structurally, NMDARs are hetero-tetramers of subunits GluN1, GluN2 and GluN3. In order to form a functional NMDAR ion channel complex, assemblies of two GluN1 subunits together with either two GluN2 subunits or a combination of GluN2 and GluN3 subunits are required. However, the presence of isoforms in each NMDAR subunit results in the existence of a large repertoire of NMDAR subunit combinations giving rise to functionally distinct NMDARs. The GluN2 family includes four members (GluN2A–D) encoded by four different genes, while the GluN3 family includes two members (GluN3A and B) encoded by two separate genes. In contrast to GluN2 and GluN3 subunits, the GluN1 subunit is encoded by a single gene but due to alternative mRNA splicing of three exons (5, 21 and 22) leads to eight possible isoforms for GluN1 (GluN1–1a to 4a and GluN1–1b to 4b). The isoforms ‘b’ has an additional stretch of 21 amino acids in the amino-terminal region of the GluN1 subunit that is encoded by exon 5. The exon 21 encodes a 37-amino-acid segment at the C-terminus while the exon 22 encodes a 38-amino-acid segment also at the C-terminus of the GluN1 subunit. The distribution and expression patterns of all NMDAR isoforms vary at different brain regions and throughout development. Each NMDAR subunit is a modular protein that contains four distinct semiautonomous domains: the amino-terminal domain (ATD; roughly divided into two halves, R1 and R2), the ligand-binding domain (LBD; formed from amino-acid segments S1 and S2), the transmembrane domain (TMD) and the C-terminal domain (CTD). The ATD and LBD are extracellular while the CTD is intracellular (Traynelis et al. 2010). Studies performed by employing electrophysiological, pharmacological and molecular biological strategies on native tissues and on heterologously expressed NMDAR subunits (different isoform combinations that are either wild type, chimeric, mutant and truncated versions) have provided the essential roles of these protein domains in conferring NMDARs with unique functional and heterogeneous biophysical properties. These properties can be broadly grouped into two main categories (Paoletti, 2011): (a) permeation properties, which are mainly acquired due to the TMD, include single-channel conductances and their block by extracellular Mg²⁺ ions, and (b) gating properties, which include sensitivity to various agonists and antagonists, activation and deactivation kinetics, channel mean open time and maximal open probability, and channel shut times, etc. Initially, the LBD and TMD were believed to be solely responsible for most of the gating properties, and therefore both exogenous and endogenous compounds would modulate the function of NMDARs by binding to these domains. However, more recent studies have shown that the ATD is equally if not more important than the LBD and TMD in influencing gating properties of NMDARs.

How can the ATD of NMDAR subunits influence various biophysical properties of NMDARs? The detailed mechanisms of the ATD's role in NMDAR function are less well understood, but modelling, functional and crystallographic data support the hypothesis that the ATD has indirect effects on the LBD and TMD of the NMDAR complex. An oversimplified and generalized model is that R1 and R2 of the ATD oscillate between open, closed and various intermediate conformational states at rest as well as under conditions during and after binding to allosteric modulators, consequently influencing the subsequent downstream protein domains, i.e. the LBD and TMD (Gielen et al. 2009). But there exists many differences in interactions, orientations and structural rearrangements under different conditions within the ATD of NMDAR subunits resulting in observing heterogeneity in ATD modulation of NMDAR function. Taken together, the ATDs of NMDAR subunits have significant role on biophysical and pharmacological properties of NMDARs. Briefly, electrophysiological studies performed on truncated, chimeric and mutated ATDs of GluN2-containing NMDARs have shown alterations in channel maximal open probability, and deactivation and desensitization kinetics, at least, at GluN2A- or GluN2B-containing NMDARs. Furthermore, many studies have shown the role of the GluN1–1a (the most widely used GluN1 variant) ATD in influencing the biophysical properties of the GluN2-containing NMDAR subunits. For example, the GluN1–1a subunit expressed with GluN2D-containing NMDARs in a heterologous system exhibited low open probability, variations in deactivation and desensitization kinetics, etc (Gielen et al. 2009; Yuan et al. 2009). Despite the knowledge available on the ATD of NMDAR subunits, it still remains unclear how GluN1 splice variants influence biophysical and pharmacological properties of various GluN2 subunit-containing NMDARs, in particular, GluN2D-containing NMDARs.

However, in a recent study in The Journal of Physiology, Vance et al. (2012), by employing the heterologous expression approach, have dissected the role of GluN1 splice variants, specifically the GluN1–1b isoform, on the biophysical and pharmacological properties of GluN1/GluN2D-containing NMDARs. They found that inclusion of a 21-amino-acid stretch in the ATD of the GluN1 subunit (i.e. the GluN1–1b isoform) increases open probability and deactivation kinetics, and decreases agonist potencies of GluN1/GluN2D-containing NMDARs as opposed to GluN1–1a isoform-containing GluN1/GluN2D receptors. Moreover, they extended their approach in characterizing and comparing the effects of other GluN1 splice variants (i.e. GluN1–1a to 4a and GluN1–1b to 4b) on biophysical and pharmacological properties of GluN1/GluN2D-containing NMDARs. Interestingly, they found that Lys211 encoded in exon 5 mediates most of the effects on agonist potencies as well as deactivation kinetics of GluN1/GluN2D-containing NMDARs. Finally, they proposed a model that fits with the single channel and macroscopic properties of GluN1/GluN2D receptors. Although Vance et al. (2012) have yet again demonstrated the exceptional role of the ATD in influencing NMDAR function, it is still not well understood how a 21-amino-acid stretch in the ATD can induce such major effects on biophysical and pharmacological properties of GluN1/GluN2D receptors. This could be due in part to the lack of structural data on the ATD in full-length NMDA receptors. However, Vance et al. (2012) have predicted that Lys211 in the ATD of GluN1–1b subunits plays an important role in intra-subunit interaction with GluN2D receptors, thus observing lower agonist potencies and rapid deactivation time course in their study.

Nevertheless, in spite of such significant progress in understanding the role(s) of the ATD in modulating NMDAR function, several questions remain unanswered (Hansen et al. 2010). For instance, how does the ATD of NMDAR subunits exerts its role in assembly, targeting and trafficking of NMDARs? Second, how does the modification of NMDAR subunit ATD consensus sites for N-glycosylation impact biophysical properties of NMDARs? Third, an increasing amount of evidence suggests that the in vivo composition of NMDARs is not always a binary assembly of GluN1/GluN2 (A–D) or GluN1/GluN3 (A–B), but that NMDARs can also exist as triheteromeric subunit assemblies (GluN1/GluN2 (A–D)/GluN3 (A–B)). Therefore, what is the role of the ATD in influencing the biophysical properties of triheteromeric NMDARs? Finally, it has been demonstrated that the ATD of NMDAR subunits interacts with a variety of extracellular synaptic proteins. However, it is unclear if these interactions have any influence on ATD modulation of NMDAR function (Hansen et al. 2010).

It is necessary to understand the role(s) of the ATD as well as other protein domains in proper functioning of NMDARs because these receptors play a critical role in several physiological and pathological processes. NMDARs are highly permeable to Ca²⁺ ions and this property allows NMDARs to operate as efficient integrators of synaptic activity and the activation of a multitude of intracellular signalling cascades. The influx of Ca²⁺ ions via NMDARs into a postsynaptic neuron has a tremendous impact on that neuron because the amplitude and time course of the calcium transient (which mainly depends on the biophysical properties of NMDARs) results in altering intracellular signalling events that are central for inducing the process of long-term potentiation (Traynelis et al. 2010). Unfortunately, the excess influx of calcium ions through NMDARs leads to deleterious effects on the postsynaptic neuron under conditions such as ischaemic stroke where overstimulation of NMDARs occurs. Furthermore, inappropriate activation or function of NMDARs protein domains can lead to certain chronic pathologies such as epilepsy and schizophrenia (Lau & Zukin, 2007; Traynelis et al. 2010). Also, given the ambiguous roles for various protein domains of NMDARs in a multitude of neurodegenerative disorders, studies such as that of Vance et al. (2012) significantly contribute to our understanding of the previously unappreciated role of the ATD in NMDAR physiology. In conclusion, a better understanding of the molecular determinants underling the proper function of NMDARs creates the potential for therapeutic intervention while revealing new insights into multiple brain pathologies that involve predominantly NMDAR dysfunction.