39
40
structural change in the LBD of the GluN1 subunit. As for the N786 residue in the GluN3B subunit, it is currently unknown whether this site is glycosylated in the recombinant GluN3B subunit, however, since we observed changes in the "tail" current amplitude of GluN1/GluN3B receptors with the N786 residue mutated, we suggest that glycosylation at this residue alters the activation of GluN1/GluN3B receptors by altering the LBD in the GluN3B subunit, rather than by affecting desensitization kinetics via the LBD of the GluN1 subunit.
We have previously shown a strong association of GluN3A-containing NMDARs with numerous lectins such as ConA, WGA and AAL mainly selective for mannose residues, N- acetylglucosamine residues, and fucose residues, respectively. Moreover, our previous biochemical data indicate that GluN3A-containing NMDARs contain a wide variety of glycan structures (Kaniakova et al., 2016). Here, we complemented these findings by showing that lectins modulate the functional properties of GluN1/GluN3 receptors.
Considering that 14 out of 19 tested lectins significantly slowed τw of desensitization and altered peak amplitude, we hypothesized that this modulating effect is mediated by a decrease in desensitization kinetics. Indeed, after introducing mutations into the GluN1 subunit rendering it insensitive to glycine up to 30 mM (Kvist Trine, Greenwood Jeremy, 2013), we observed no changes after lectin application. These results confirmed our hypothesis that lectins modulate receptor desensitization, and are consistent with previous findings that ConA reduces desensitization at AMPARs and KARs, though it has no pronounced effect on GluN1/GluN2 receptors (Everts et al., 1997; M. L. Mayer and Vyklicky, 1989).
Interestingly, we observed the strongest modulating effect on GluN1/GluN3A and GluN1/GluN3B receptors with AAL, however, receptors, containing a replacement at the N565 site of GluN3A and the N465 site of GluN3B showed similar sensitivity to AAL as WT GluN1/GluN3 to ConA and WGA. This observation suggests that the N565 residue in the GluN3A subunit and its homologous residue N465 in the GluN3B subunit are involved in potentiating the modulatory effect of AAL, which could be explained by a higher specificity or affinity of AAL for different glycans compared to WGA or ConA (Dam and Fred Brewer, 2009; Moremen et al., 2014). Furthermore, we observed that the application of ConA, WGA or AAL in the continuous presence of glycine had virtually no effect on the GluN1/GluN3A or GluN1/GluN3B receptor desensitization; this finding is consistent with
41
a previous report that ConA-induced reduction of KAR desensitization is a conformation- dependent process (Everts et al., 1999). Taken together, our findings reveal new information on the role of N-glycans in regulating the functional properties of GluN1/GluN3 receptors in mammalian cells and provide new clues about lectins as a tool for modulating non- conventional NMDARs.
6.2. Three pathogenic mutations in the M3 domain of the GluN1 subunit regulate the surface delivery and pharmacological sensitivity of NMDARs
In Kolcheva et. al, 2021, we focused on three pathogenic mutations in the M3 domain of the GluN1 subunit – M641I, A645S, and Y647S – associated with intellectual disability, movement disorders, and seizures (Lemke et al., 2016). We showed that these mutations affect the surface delivery of NMDARs in HEK293 cells in a subunit-dependent manner;
the GluN1-M641I mutation significantly reduces the surface expression of GluN1/GluN2A receptors, but not GluN1/GluN2B or GluN1/GluN3A receptors, the GluN1-Y647S mutation reduces the surface expression of GluN1/GluN2A and GluN1/GluN2B receptors, but not GluN1/GluN3A receptors, and the GluN1-A645S mutation does not affect the level of surface expression whether it is co-expressed with GluN2A, GluN2B or GluN3A subunits.
Moreover, our microscopy data obtained from hippocampal neurons (DIV14) demonstrated the same changes in surface expression as observed in HEK293 cells when mutated GluN1 subunits were co-expressed with GluN2A. These results are consistent with data showing that hippocampal neurons have low expression of endogenous GluN3 subunits and high expression of GluN2A (Hansen et al., 2021; Henson et al., 2010; Paoletti et al., 2013; Vieira et al., 2020). Similar to our findings, it has been shown that a de novo mutation GluN1- G620R, associated with developmental delay and behavioural abnormalities, reduces the surface delivery of GluN1/GluN2B receptors, but not GluN1/GluN2A receptors (Chen et al., 2017).
Our group has previously shown that specific amino acid residues in the M3 domain of GluN1 and GluN2 (GluN1-W636, GluN2A-W634, GluN2B-W635) contribute to the regulation of NMDAR trafficking to the cell surface (Kaniakova et al., 2012). Furthermore, pathogenic mutations GluN2B-W607C and GluN2B-S628F in the TMD, associated with intellectual disability and developmental delay, reduce NMDAR surface delivery, as well as the sensitivity of NMDARs to glutamate, glycine, and to Mg2+ block (Vyklicky et al., 2018).
42
In our further experiments we also observed changes in receptor functional properties, specifically, we found that the EC50 value for glycine was increased for GluN1- M641I/GluN3A receptors compared to WT GluN1/GluN3A, and we found that GluN1- A645S/GluN3A receptors had a slower τw of desensitization. The changes in desensitization are consistent with evidence that mutations in the TMD, including the SYTANLAAF motif, alter the current kinetics of GluN1/GluN2A receptors (Hu and Zheng, 2005). Although we found that the GluN1-Y647S mutation does not alter the surface expression of GluN1/GluN3A receptors, our electrophysiological experiments revealed no glycine- induced current unless CGP-78608 was applied (Grand et al., 2018). Therefore, we hypothesise that the GluN1-Y647S mutation presumably changes specific functional properties of GluN1/GluN3A receptors, such as the Po. However, we cannot test this assumption because, unlike the open-channel blocker of GluN1/GluN2 receptors MK-801, no specific open-channel blocker of GluN1/GluN3A receptors exists at present (Chatterton et al., 2002).
Reduction in agonist and/or antagonist sensitivity was reported for other pathogenic mutations, such as GluN2B-N615I located at the beginning of the M2–M3 linker, and GluN2B-V618G, located in the M2–M3 linker, both associated with West syndrome. These mutations reduce the NMDAR sensitivity to Mg2+ and memantine (Fedele et al., 2018). The GluN1-G620R mutation, mentioned above, also reduces the sensitivity to agonist and the Mg2+ block (Chen et al., 2017). Moreover, single residue S632 in the M3 domain of the GluN2A subunit, and residues at the corresponding position in other GluN2 subunits have been reported to underlie the subunit-specific differences in Mg2+ block between GluN1/GluN2A-B and GluN1/GluN2C-D receptors (Retchless et al., 2012). However, in our study, we did not observe changes in the Mg2+ sensitivity of GluN1/GluN2 receptors with any of the examined mutations. Nevertheless, we observed that the GluN1-A645S mutation reduces the sensitivity of GluN1/GluN2 receptors to memantine in the absence of extracellular Mg2+, which is consistent with previous results obtained using Xenopus oocytes expressing GluN1/GluN2B receptors (Kashiwagi et al., 2002). Interestingly, we found that in the presence of physiological concentration of Mg2+, the IC50 value for memantine was decreased for the GluN1-M641I/GluN2A receptors but increased for the GluN1- A645S/GluN2A receptors, when compared to WT GluN1/GluN2A receptors. Similar results were obtained when we studied these GluN1 subunits expressed in hippocampal neurons.
43
Our analysis of NMDA-induced excitotoxicity in hippocampal neurons also supported this finding. Specifically, memantine reduced excitotoxicity in neurons expressing the GluN1- M641I subunit but was significantly less neuroprotective in neurons expressing the GluN1- A645S subunit in comparison to neurons expressing the WT GluN1 subunit.
In addition, we found that the mutations M641I-GluN1 and A645S-GluN1 differentially affect the onset (τon) and offset (τoff) kinetics of inhibition by memantine, as well as memantine’s ability to induce SSI at NMDARs. Specifically, we found that the τoff
of memantine inhibition was slower for GluN1-M641I/GluN2A receptors and faster for GluN1-A645S/GluN2A receptors, both in the absence and presence of Mg2+, when compared to WT GluN1/GluN2A receptors. Previous studies found that the slow component of τoff (τslow) represents memantine unbinding from its second low-affinity binding site at the GluN1/GluN2 receptor (Glasgow et al., 2018). Thus, when we focused on τslow, we found increased values for GluN1-M641I/hGluN2A receptors and decreased values for GluN1- A645S/GluN2A receptors, when compared to WT GluN1/GluN2A receptors under all experimental conditions tested. These findings indicate that these two pathogenic mutations in the M3 domain of the GluN1 subunit differentially affect the binding and unbinding kinetics of memantine, and may affect the SSI of memantine or, based on recent studies, may affect the membrane-to-channel transfer of memantine (Wilcox et al., 2022). Taken together, our data indicate that specific pathogenic mutations in the M3 domain of the GluN1 subunit differentially affect the surface delivery and pharmacological properties of NMDARs.
6.3. Effect of glycine-binding site structural features of GluN1 and GluN3 on the surface expression of NMDARs
In Skrenkova et al., 2019, we studied the role of the glycine-binding sites of GluN1 and GluN3A subunits in the regulation of NMDARsurface expression. Although many studies suggest that the CTDs of GluN subunits play an important role in the regulation of trafficking of NMDARs to the cell surface, due to the variety of CTD ER retention motifs and binding partners (Horak and Wenthold, 2009; Standley et al., 2000), the extracellular part of the NMDAR is also involved in this process (Chen et al., 2017; Horak et al., 2008;
Kaniakova, Lichnerova, et al., 2012; Vyklicky et al., 2018). In particular, it has been reported that reduced GluN2B subunit affinity to glutamate and GluN1 subunit affinity to glycine,
44
associated with mutations in these subunits, correlate with reduced surface expression of GluN1/GluN2B receptors, and this process is likely regulated during the ligand binding checkpoint in the ER (Kenny et al., 2009; She et al., 2012). To investigate whether the integrity of the glycine-binding sites of the GluN1 and GluN3A subunits is required for the correct delivery of NMDARs to the cell surface, we have employed several mutations in the GluN1 LBD that were previously shown to reduce the affinity of GluN1/GluN2 receptors to glycine (Kvist et al, 2013). Specifically, we introduced the following mutations in the GluN1 subunit: A714L, F484A, T518L, D732A, F484A+T518L, and homologous mutants in the GluN3A subunit: T825L, Y605A, S633L, D845A, Y605A+S633L. Using fluorescence microscopy, we found that mutations decreased the receptor surface expression to the following extent: GluN1/GluN3A = GluN1-A714L/GluN3A > GluN1- F484A/GluN3A > GluN1-T518L/GluN3A > GluN1-D732A/GluN3A > GluN1- F484A+T518L/GluN3A. Interestingly, the order of surface reduction correlates with the order of the glycine affinity reduction in GluN1/GluN2 receptors. However, GluN1/GluN3A receptors have four binding sites for glycine with much higher binding affinity in the GluN3A subunit (Yao and Mayer, 2006), therefore, these mutations can affect glycine sensitivity of GluN1/GluN3A receptors differently compared to GluN1/GluN2 receptors.
Thus, we first examined how the described mutations affect the functional properties of GluN1/GluN3A receptors expressed in HEK293 cells. For this purpose, we used the whole- cell patch-clamp technique with rapid solution exchange system, which is important for detecting the peak glycine-induced current prior to the fast receptor desensitization. We observed ~5-fold decrease in glycine sensitivity for GluN1-F484A/GluN3A receptors compared to WT GluN1/GluN3A receptors, moreover, cells expressing GluN1- D732A/GluN3A receptors had no detectable current, in agreement with our microscopy data and a previous report that this mutation reduces the surface expression of GluN1/GluN2 receptors (Kenny et al., 2009). GluN1-F484A+T518L/GluN3A receptors did not desensitize at any of the glycine concentrations, since it has been reported that these mutations render the GluN1 subunit insensitive to glycine up to 30 mM (Kvist Trine, Greenwood Jeremy, 2013). Thus, we were unable to perform detailed functional analyses of GluN1- D732A/GluN3A and GluN1-F484A+T518L/GluN3A receptors.
Concerning mutations in the GluN3A subunit, we observed a similar rank order of NMDAR surface expression of as was observed with the homologous mutants in the GluN1
45
subunit. Using the patch-clamp technique, we found that only GluN1/GluN3A-T825L and GluN1/GluN3A-Y605A receptors carried detectable currents, whereas receptors with the other three mutations did not produce measurable currents. Furthermore, we did not observe differences in the τw of desensitization between mutated and WT GluN1/GluN3A receptors, in agreement with the idea that the desensitization of GluN1/GluN3A receptors is regulated by the glycine-binding site in the GluN1 subunit. Peak current analysis showed 9-fold and 29-fold decrease in glycine affinity for GluN1/GluN3A-T825L and GluN1/GluN3A-Y605A receptors, respectively, in comparison with WT GluN1/GluN3A receptors, and this strongly correlated with their level of surface expression.
Since we could not detect current with GluN1/GluN3A-S633L, GluN1/GluN3A- D845A, and GluN1/GluN3A-Y605A+S633L receptors, we hypothesized that this may be not only due to a decrease in surface expression but also due to a change in desensitization and/or a change in glycine affinity of these receptors. Thus, we employed the GluN1-F484A mutation, which slows down the kinetics of receptor desensitization without affecting the surface expression of GluN1/GluN3A receptor, and co-transfected it with mutated GluN3A subunits. We observed detectable current with GluN1-F484A/GluN3A-S633L receptors elicitd by 1, 3, or 10 mM glycine, but not with the rest of the mutated GluN1/GluN3A receptors. In addition, we employed CGP-78608 that has been shown to potentiate GluN1/GluN3A receptor currents by abolishing the GluN1-mediated inhibition (Grand et al., 2018), in order to unmask undetectable current. However, we still did not observe responses from HEK293 cells expressing GluN1/GluN3A-S633L, GluN1/GluN3A-D845A, and GluN1/GluN3A-Y605A+S633L receptors. Taken together, our results reveal new information about the role of glycine-binding sites in the GluN1 and GluN3A subunits in the regulation of function and delivery of GluN1/GluN3A receptors.
6.4. The pathogenic S688Y mutation in the ligand‑binding domain of the GluN1 subunit regulates the properties of NMDARs
In Skrenkova et al., 2020, we investigated how the pathogenic missense mutation S688Y in the S2 lobe of the GluN1 LBD, associated with severe early infantile encephalopathy, affects the function and trafficking of NMDARs. We performed in-silico modelling and found that the S688Y mutation reduces the apparent affinity of the GluN1 LBD for glycine and D-serine. To confirm these findings, we conducted electrophysiological
46
experiments in HEK293 cells expressing either WT or mutated GluN1 subunit together with GluN2A, GluN2B, or GluN3A subunits and measured the EC50 values for glycine and D- serine. We observed a significant decrease in receptor affinity for both glycine and D-serine, with a more pronounced effect on glycine potency, in line with previous findings showing that GluN1-S688A mutation leads to a reduction of glycine potency (Kuryatov et al., 1994), and the suggestion that the S688 residue in the GluN1 subunit plays a key role in ligand recognition (H. Furukawa and Gouaux, 2003). Moreover, we measured the EC50 values for glycine in hippocampal neurons expressing YFP-GluN1 or YFP-GluN1-S688Y subunit and we observed significant differences in the EC50 values between WT and mutated NMDARs, specifically, receptors with the GluN1-S688Y mutation had the EC50 of 220 µM compared to 0.2 µM for neurons expressing the WT subunit. In addition, we found that GluN1- S688Y/GluN3A receptors had increased τw of desensitization compared to WT GluN1/GluN3A receptors, which is in agreement with previous reports that structural changes in the GluN1 LBD of GluN1alter the desensitizing properties of GluN1/GluN3A receptors (Awobuluyi et al., 2007; Kvist et al., 2013). Taken together, our electrophysiological data are consistent with our in-silico modelling and show that the S688Y mutation in the LBD of the GluN1 subunit alters agonist binding.
Our microscopy data revealed that the GluN1-S688Y mutation significantly reduces the delivery of GluN3A-containing NMDARs to the cell surface, in line with our previous finding that the LBD sensitivity to glycine plays a crucial role in the regulation of NMDAR surface delivery. However, the GluN1-S688Y mutation did not affect the surface delivery of receptors containing GluN2A or GluN2B subunits, in contrast to the previous finding that the GluN1-D732A mutation affects the trafficking of GluN1/GluN2A receptors. This variation may be explained by differences between the GluN1-D732A and GluN1-S688Y mutations with respect to changing the conformation of functional NMDAR heterotetramers, thereby affecting the surface delivery of GluN1/GluN2A receptors. It has been shown that there is not always a correlation between agonist potency and surface expression, thus, for example, a GluN2A-V734L mutation in the LBD that slightly reduces glutamate sensitivity, does not affect NMDAR surface delivery, suggesting that additional mechanisms are involved in the regulation of receptor expression (Swanger et al., 2016). In other cases, however, NMDARs carrying different pathogenic mutations in the LBDs of GluN2A and GluN2B do exhibit a clear correlation between the EC50 for L-glutamate and surface
47
expression. For example, the pathogenic E413G mutation in the GluN2B subunit profoundly reduced the surface delivery of NMDARs. Overall, our results provide additional information about the role of the LBD in the regulation of surface delivery and function of NMDARs.