Ionotropic glutamate receptors (iGluRs) are tetrameric ligand-gated ion channels that convey the majority of excitatory neurotransmission in the mammalian central nervous system. Members of the iGluR family are AMPA and kainate receptors as well as NMDA receptors. NMDA receptors are unique among the iGluRs due to the obligate heteromeric assembly and their particular roles for learning and memory formation but on the other hand NMDA receptor mediated excitotoxicity causes cell death in various pathophysiological conditions and neurodegenerative disorders. Conventional NMDA receptors are composed of two glycine binding Glun1 subunits and two glutamate binding GluN2 subunits. These receptors are also referred to as ‘coincidence detectors’ as for the activation a predepolarization of the postsynaptic membrane and ligand binding is required. Receptors composed of two glycine binding GluN1 and two glycine binding GluN3 subunits are referred to as ‘excitatory glycine receptors’. Each iGluR subunit has a modular structure and it is hypothesized that each module/domain originates from a prokaryotic ancestral protein. The extracellular N-terminal domains (NTDs) are thought to form local heterodimers between GluN1 and GluN2 or GluN3 NTDs, additionally the GluN2- or GluN3-NTDs form a homodimeric contact. The ligand-binding domains (LBDs) of GluN1 and GluN2 or GluN3 subunits are arranged in a ‘back-to-back’ conformation forming the LBD heterodimer interface, which is subdivided into three distinct sites (sites I-III). This heterodimer interface has been previously implicated in receptor desensitization, weak interface interactions increase receptor desensitization. Upon ligand-binding, the S1 and S2 subdomains of the LBDs close in a ‘venus-flytrap’ like fashion. The transmembrane domains (TMDs) of iGluRs show homology to prokaryotic K+ channels but iGluR TMDs have one transmembrane domain (TM4) more than the prokaryotic ion channel. The aim of the work presented here was to analyze intra- and intersubunit interface interactions in the NTDs, LBDs and TMDs to determine their functional implications. To achieve this, in vitro and in silico methods were combined. Molecular docking experiments and homology modelling were used to gain insight into the molecular mechanism of agonism, partial agonism and antagonism. The results indicate that for GluN1 subunits it is the size of the ligand that determines its action, i.e. agonists are small, partial agonists are slightly larger and antagonists are large molecules. However, this could not be validated for GluA2-type AMPA receptor subunits, GluN2A or GluN3A subunits. Thus, the mechanism for partial agonism seems to be not conserved but subunit specific. In order to gain insight into the role of the GluN1-GluN2 LBD heterodimer interface we analyzed the GluN1-GluN2A LBD crystal structure as well used the mutual information (MI) analysis to select amino acids for site-directed mutagenesis. The respective mutants were functionally characterized by two-electrode voltage-clamp on Xenopus oocytes. The results did not show specific effects of the mutations, hence it is not possible to imply one interaction site to one function. We conclude that the interface as a network of interactions is generally involved in receptor function. However, one mutation in site III largely decreased agonist-induced currents from NTD-lacking GluN1/GluN2 receptors. Whether this effect is caused by reduced subunit expression, receptor assembly or trafficking needs to be disclosed in further experiments. Homology model based analysis and subsequent site-directed mutagenesis and functional characterization of GluN1/GluN3A receptors led to the identification of specific residues in the GluN3A-NTD that selectively increased the receptor efficacy. Thus, the GluN3A-NTD resembles the role of an autoinhibitory domain (AID), to our knowledge no AID has been described so far for ligand-gated ion channels. In a last set of experiments we assessed interactions between the GluN1 TM4 and the TM1 and TM3 of the neighboring GluN2A subunit. Amino acid substitutions were selected based on our GluN1-GluN2A TMD model and functional characterization by TECV revealed two residues that almost completely abolished receptor function. We further used the MI analysis to probe for positions that are evolutionary interdependent and from the position of these correlations we inferred that the TM4 is mainly involved in stabilizing the TMD and thus ensuring proper ion channel function.