Abstract: A variety of de novo and inherited missense mutations associated with neurological disorders are found in the NMDA receptor M4 transmembrane helices, which are peripheral to the pore domain in eukaryotic ionotropic glutamate receptors. Subsets of these mutations affect receptor gating with dramatic effects, including in one instance halting it, occurring at a conserved glycine near the extracellular end of M4. Functional experiments and molecular dynamic simulations of substitutions at this glycine indicate that it acts as a hinge, permitting the intracellular portion of the ion channel to laterally expand. This expansion stabilizes long-lived open states with resultant slow deactivation and high Ca2+ permeability. Our studies provide a functional and structural framework for the effect of missense mutations on NMDARs at central synapses and highlight how the M4 segment may represent a pathway for intracellular modulation of NMDA receptor function.

Abstract: Desensitization is a canonical property of ligand-gated ion channels. The process is typically examined as the loss of open channels at the onset of desensitisation and their reappearance during recovery from desensitization. We have identified an AMPAR that retains conductance following desensitisation, giving us a novel functional readout of the desensitized state. We have used this readout to probe the structure of desensitized AMPARs in their native environment.

Abstract: We developed a virtually complete 3D model of the human GluN1-GluN2B receptor based on recently solved crystal structures for frog and rat NMDARs2,3. The human NMDAR structure enabled the mutated residues to be located thereby allowing insight into their potential impact on neighbouring amino acids, including those forming the binding sites for Mg2+ and NMDAR channel blockers. The characterisation of the mutants revealed reductions in glutamate potency, increased receptor desensitisation and ablation of voltage-dependent Mg2+ block. Moreover, introducing the mutations into neurons revealed profound changes to excitatory neurotransmission that could underlie their associated neurological disorders. By attempting to reverse the gain-of-function phenotype with the channel blocker memantine revealed that the mutations significantly altered the sensitivity of the NMDAR to the blocker. Using our experimental data, we also devised a new kinetic model to understand the mechanism of action of memantine and how it is affected by endogenous levels of Mg2+. The functional effects of the missense mutations revealed gain and loss-of-function that affected NMDAR ligand binding and ion channel properties. These profiles may underpin epileptic phenotypes of patients that carry the mutations. Moreover, the study indicates that memantine could be a promising treatment for selected gain-of-function mutations in NMDA receptors.

References:
1. Burnashev N and Szepetowski P (2014) Curr Opin Pharmacol 20: 73-82
2. Lee C-H. et al. (2014) Nature 511: 191-197.
3. Karakas E and Furukawa H (2014) Science 344: 992-997.

Abstract: For each agonist-bound subunit, a rate-limiting conformational change after agonist binding and prior to channel opening is assumed. By fitting this mechanism to single channel data, rate constants are found that correctly predict macroscopic and synaptic NMDAR currents.

Abstract: NMDA receptors are Ca2+-permeable channels gated by glutamate and are essential for excitatory syn-aptic transmission during physiologic and pathologic brain processes. Ca2+-dependent inactivation (CDI) by calmodulin is a regulatory mechanism that reduces NMDA receptor gating in an activity-dependent manner. Although the obligatory GluN1 subunit directly mediates CaM binding and is required for CDI, the present literature suggests that GluN2B receptors are insensitive to CDI. We used electrophysiological recordings of recombinant NMDA receptors to examine the mechanism of CDI subtype-dependency. Whole-cell currents recorded in the presence of 2 mM external Ca2+ and overexpression of calmodulin showed that CDI is robust for GluN2A channels (CDI2A = 0.51 ± 0.05) but appeared absent for GluN2B (CDI2B = 0.04 ± 0.07). However, when Ca2+ was dialyzed intracellularly, robust CDI occurred for both GluN2A and GluN2B channels (CDI2A = 0.80 ± 0.06; CDI2B = 0.52 ± 0.06), indicating that GluN2B channels are susceptible to CDI. To investigate this apparent discrepancy, we used patch clamp fluorometry to record single-channel Na+-only currents from individual NMDA receptors and simultaneously monitor intracellular Ca2+ elevations upon ionomycin treatment. We found upon Ca2+entry, the derived kinetic models suggested that the gating kinetics of both GluN2A and GluN2B channels were perturbed by a common kinetic mechanism where by reduced activation resulted in the accumulation of channels in long-lived closed (desensitized) states. Thus, the apparent subtype-dependent susceptibility to CDI by external Ca2+ reflects intrinsic subtype differences in gating kinetics such as equilibrium open probability (Po) which determine the amount of fluxed Ca2+. To test this hypothesis, we measured CDI by external Ca2+ using whole cell recordings from GluN2B channels carrying a Lurcher (Lc) mutation, GluN1A652Y, which dramatically increased channel Po (Po, wt = 0.19 ± 0.07; Po, Lc = 0.81 ± 0.08). We found that 2 mM external Ca2+ produced robust CDI for these high-Po GluN2B channels (CDILc = 0.54 ± 0.06). Together, these results show that Ca2+ produces activity-dependent CDI for both GluN2A and GluN2B receptors. These results are consistent with CDI as a neuroprotective mechanism against excessive Ca2+ load during high-Po activation.

Abstract: N-Methyl-D-aspartate receptors (NMDARs) play essential roles in memory formation, neuronal plasticity and brain development with their dysfunction linked to a range of disorders from ischemia to schizophrenia. Zinc and pH are physiological allosteric modulators of NMDARs with GluN2A containing receptors inhibited by nanomolar concentrations of divalent zinc and by excursions to low pH. Despite the widespread importance of zinc and proton modulation of NMDARs, the molecular mechanism by which these ions modulate receptor activity has proven elusive. We used cryo-electron microscopy to elucidate the structure of the GluN1/GluN2A NMDAR in a large ensemble of conformations under a range of physiologically relevant zinc and proton concentrations. We show how zinc binding to the amino terminal domain elicits structural changes that are transduced though the ligand-binding domain and result in constriction of the ion channel gate.

Abstract: Reduced dendritic spine density has been reproducibly observed in multiple brain areas in schizophrenia (Sz), such as the auditory cortex, and is believed to underlie deficits in cortical processing. We recently reported, in a cohort of 20 Sz and 20 matched controls, that only small, presumably dynamic, spines are lost in the auditory cortex of Sz while large, likely stable, spines are preserved. Here, in an expanded cohort of 50 pairs we utilized parallel microscopy and mass spectrometry (MS) approaches identity changes in synaptic protein levels and phosphorylation (Phos) that may underlie small spine loss Sz. Fixed and Frozen auditory cortex grey matter from 50 pairs of Sz and matched control subjects was obtained from the Pitt Brain Tissue Donation Program. Fixed hemisphere was utilized in microscopy studies to assess deep layer 3 spine density and volume. The right hemisphere was utilized for 1. Targeted MS to quantify 400 proteins in homogenates; 2. Targeted MS to quantify 350 proteins in synaptosomes; 3. Differential MS to quantify >2000 Phos events homogenates from a subset (16 pairs). We found robust changes in synaptosome and Phos levels of canonical postsynaptic proteins. GRIA2 and GRIA3 were among the most significantly down regulated synaptosome proteins and these decreases were strongly correlated with small spine loss. Interestingly, we also observed altered AP2M1 phosphorylation, a documented trafficker of GRIA2 and GRIA3, at a site known to regulate its trafficking activity. These findings nominate protein trafficking and phosphorylation events that may underlie spine loss in Sz for future mechanistic investigations.

Abstract: Neurotransmitter receptor subtype and the number, the density, and their distribution relative to the location of transmitter release are likely to be the key determining factors of the properties of signal transmission. AMPA glutamate receptors containing fast kinetic GluA3 and GluA4 subunits are prominently present in subsets of neurons that are capable of firing action potentials at high frequencies, such as the auditory relay neurons on bushy cells and fusiform cells of the cochlear nucleus. We examined the number, density and organization of GluA3 and GluA4 at the synapse of the auditory nerve on bushy and fusiform cells. Using freeze-fracture immunolabeling (FRIL) we show a positive correlation between numbers of gold particles and the size of synapses for all pan AMPA, GluA3 and GluA4 subunits in the auditory nerve synapses both on bushy and on fusiform cells. These synapse types have the same number of AMPA receptors, but at auditory nerve synapses on bushy cells the gold density is higher than that on fusiform cells due to smaller postsynaptic densities. GluA3 gold labeling number and density are higher at auditory nerve synapses on bushy cells, whereas, GluA4 gold labeling number and density are higher at those on fusiform cells. The intrasynaptic distribution of gold labeling revealed that in auditory nerve synapses on bushy cells AMPA receptors -in particular GluA3, are concentrated at the center of synapse. The center concentration of AMPA receptors is absent in GluA3-knockout mice and gold particles are found evenly distributed along the auditory nerve synapse on bushy cells. GluA4 gold labeling was found homogenously distributed along both synapse types. Our findings show that GluA3 and GluA4 subunits are target-cell-dependent at auditory nerve synapses.

Abstract: Women represent a vulnerable population in the development alcohol use disorders, as they transition to dependence faster, show greater craving in response to alcohol-associated cues and stress, and exhibit higher psychiatric comorbidity than men. While relapse rates are similar across gender, women are more likely to relapse when faced with stress or negative mood states during abstinence, suggesting that these factors are important when developing gender-based treatment strategies. In parallel with this hypothesis, we recently reported that female rats show increased alcohol craving-like behavior relative to males in response to cues and stress. Thus, identifying sex-specific mechanisms mediating relapse-related behavior may improve treatment outcomes. We have begun to investigate these differences by determining sex differences in protein expression of glutamate receptor subunits in cortico-limbic circuits known to regulate reinstatement. In addition, we also tested the effect of the NMDA antagonist ketamine and the AMPA antagonist NBQX on reinstatement of alcohol seeking in males and females. Male and female rats were trained to press a lever for access to a dipper containing 10% ethanol in a 0.1% saccharin solution. Control rats responded for saccharin. Each reinforcer delivery was accompanied by a 10s light+tone cue. After >14 days of self-administration, the effects of low dose ketamine (10 mg/kg) on self-administration was determined. In addition, the effect of ketamine and NBQX at a range of doses on alcohol craving/relapse-like behavior was also determined. Finally, the amygdala was dissected from ketamine exposed and drug-naive male and female rats for analysis of glutamate receptor subunit expression by Western blot. We found that females had significantly greater expression of GluA1, GluA2/3, and a trend towards increased NR2B in the amygdala relative to males, with no differences in the PFC. Ketamine pre-treatment significantly attenuated both alcohol and saccharin self-administration in females, while only affecting saccharin self-administration in males. Ketamine also completely blocked reinstatement ("craving") of alcohol and saccharin seeking in females, while only reducing reinstatement to saccharin seeking in males. Alcohol seeking also tended to be blocked by a lower dose of ketamine (3 mg/kg) in females, that was not effective in males. Ketamine's anti-craving-like effects were not altered by prior NBQX treatment, and were transient in duration, lasting 48 hrs. NBQX alone was capable of reducing alcohol craving in both males and females. Ketamine treatment also produced changes in the membrane vs. cytosolic expression of glutamate receptor subunits in females. In conclusion, females are more sensitive to reinstate alcohol seeking in response to a combination of cues and a pharmacological stressor than males. The increased reinstatement may be related to enhanced AMPAR and NMDAR-related signaling in the amygdala. Ketamine is extremely effective in reducing reinstatement of alcohol-seeking selectively in females. Unlike the antidepressant-like effects of ketamine, the anti-relapse-like effects have a much shorter duration of action and are not mediated through AMPAR activation, and the AMPAR antagonist disrupts reinstatement on its own, suggesting that ketamine's potential beneficial effects for the two disorders are mediated by independent mechanisms. Nonetheless, these data suggest that ketamine may be an effective treatment for women with comorbid alcohol use and mood disorders.

Abstract: Agonists turn on receptors because they bind more strongly to active versus resting conformations of their target sites. We investigated the structural basis of agonist activation of endplate nicotinic acetylcholine receptors (AChRs) by using molecular dynamics simulations to identify differences between low-affinity (resting, R) and high-affinity (active, R*) neurotransmitter binding sites. Simulated and experimental binding energies agreed over an 11 kcal/mol range (4 agonists, 3 kinds of binding site, R and R*). Each binding site has a pocket delimited by 5 aromatic amino acids and covered by loop C. In the R* conformation, the pocket is smaller and twisted, the agonist is closer to the pocket center (distance dx) but further from aY190, and loop C is displaced outward to expose the pocket. In all conditions, dx was the structural parameter most correlated with binding energy and, to a first approximation, accounts for agonist affinity, efficacy and efficiency. The results suggest that the AChR binding pocket toggles between two pre-existing structures that for small ligands are essentially agonist-independent, and that loop C is a lid that opens upon receptor activation. The contraction and anticlockwise rotation of the pocket resemble at small scale to the activation conformational changes in the extracellular domain of related receptors, so the above pocket rearrangements could be the dynamic nucleus of the full receptor isomerization.

Abstract: Long-term plasticity of excitatory synaptic transmission is an important cellular mechanism underlying learning and memory. It is known that coincident of pre- and post-synaptic neuron firing induces calcium (Ca) influx through glutamatergic NMDA receptors (NMDARs). This Ca influx is transduced via Ca-binding protein, calmodulin (CaM) to engage downstream Ca/CaM-dependent signaling events that eventually lead to either long-term potentiation (LTP) or long-term depression (LTD). However, it is still not known how regulation of Ca/CaM dynamics is achieved to control the directionality of synaptic plasticity. Additionally, we know little about the molecular substrates that are critical for regulating synaptic plasticity. Using a combination of molecular approaches, electrophysiology and mass spectrometry-based quantitative phosphoproteomics, we find that neurogranin (Ng), a neuron-specific CaM binding protein, controls the threshold of spike-timing-dependent long-term potentiation (STDP-LTP) by regulating Ca/CaM-dependent protein phosphatase 2B (PP2B; Calcineurin) activity. Decreasing Ng levels blocks STDP-LTP by enhancing synaptic PP2B activity, which dephosphorylates the Grin2A subunit of NMDARs and accelerates the decay of NMDAR-mediated synaptic currents. Conversely, increasing Ng levels prolongs the STDP-LTP timing window at Schaffer Collateral-CA1 synapses by suppressing PP2B activity. Taken together, our results suggest that the dynamics of Ng levels at different behavioral and/or pathological states influence the basal phosphorylation state of neurons by tuning PP2B activity, and therefore, influence the expression of synaptic plasticity.

Abstract: Electrophysiology has been the primary means for characterizing the functions of ionotropic glutamate receptors (iGluRs) and for gaining mechanistic insight, but in recent years structures of isolated water-soluble domains and transmembrane-domain-containing constructs have provided the basis for formulating mechanistic hypotheses. Because these structures only represent sparse, often incomplete snapshots during iGluR activation, significant gaps in knowledge remain regarding structures, energetics, and dynamics of key substates along the functional processes. Some of these gaps have recently been filled by molecular dynamics simulations and theoretical modeling. In this presentation, I describe our work in the latter arena toward characterizing iGluR gating motions and developing a formalism for calculating thermodynamic and kinetic properties of stationary gating. The structures of iGluR subunits have a highly modular architecture, in which the ligand-binding domain and the transmembrane domain are well separated and connected by flexible linkers. The ligand-binding domain in turn is composed of two subdomains. During activation, agonist binding induces the closure of the intersubdomain cleft. The cleft closure leads to the outward pulling of a linker tethered to the extracellular terminus of the major pore-lining helix of the transmembrane domain, thereby opening the channel. This activation model based on molecular dynamics simulations was validated by residue-specific information from electrophysiological data on cysteine mutants. A further critical test was made through introducing glycine insertions in the linker. Molecular dynamics simulations showed that, with lengthening by glycine insertions, the linker became less effective in pulling the pore-lining helix, leading to weaker stabilization of the channel-open state. In full agreement, single-channel recordings showed that the channel open probability decreased progressively as the linker was lengthened by glycine insertions. Crystal structures of ligand-binding domains showing different degrees of cleft closure between full and partial agonists suggested a simple mechanism for AMPA receptors, but mysteries surrounded NMDA receptors, where the ligand-binding domains open to similar degrees when bound with either full or partial agonists. Our free energy simulations now suggest that broadening of the free energy basin for cleft closure is a plausible solution. A theoretical basis for these mechanistic hypotheses on partial agonisms was provided by a model for the free energy surface of a full receptor, where the stabilization by cleft closure is transmitted via the linker to the channel-open state. This model can be implemented by molecular dynamics simulations to predict thermodynamic and kinetics properties of stationary gating that are amenable to direct test by single-channel recordings. Close integration between computation and electrophysiology holds great promises in revealing the conformations of key substates in functional processes and the mechanisms of disease-associated mutations.

Abstract: A variety of de novo and inherited missense mutations associated with neurological disorders are found in the NMDA receptor M4 transmembrane helices, which are peripheral to the pore domain in eukaryotic ionotropic glutamate receptors. Subsets of these mutations affect receptor gating with dramatic effects, including in one instance halting it, occurring at a conserved glycine near the extracellular end of M4. Functional experiments and molecular dynamic simulations of substitutions at this glycine indicate that it acts as a hinge, permitting the intracellular portion of the ion channel to laterally expand. This expansion stabilizes long-lived open states with resultant slow deactivation and high Ca2+ permeability. Our studies provide a functional and structural framework for the effect of missense mutations on NMDARs at central synapses and highlight how the M4 segment may represent a pathway for intracellular modulation of NMDA receptor function.

Abstract: Systemic lupus erythematosus (SLE) is an autoimmune inflammatory disease where patients may develop perturbations in neurological and psychiatric function, leading to neuropsychiatric SLE (NPSLE). A subset of these patients produce anti-dsDNA antibodies (DNRAbs) that cross-react with the NMDA receptor. NMDA receptors are glutamate-gated ion channels that are essential for excitatory signaling in the nervous system. They are obligate heterotetramers, typically composed of two GluN1 subunits and two GluN2 subunits. DNRAbs mediate synaptic dysfunction, cognitive impairment, and excitotoxicity through NMDA receptor activity, largely in the hippocampus. The two major GluN2 subunits in the hippocampus are GluN2A and GluN2B and DNRAbs bind to both GluN2A and GluN2B-containing NMDA receptors. Nevertheless, the specific effect of DNRAbs on NMDA receptors and which NMDA receptor subunits are targeted remain elusive. Here, we show that DNRAbs exert subunit-specific effects on NMDA receptor gating. At a moderate concentration (patient titers), DNRAbs increase peak amplitude of glutamate-induced currents only in GluN2A-containing receptors but not GluN2B-containing receptors. Using high-resolution single channel recordings, we show that DNRAbs increase the mean open probability and mean open time of GluN2A-containing receptors, but not in those containing GluN2B, suggesting that DNRAbs stabilize GluN2A-containing receptors in the open state. In the hippocampus, NMDA receptors are typically tri-heteromeric, containing both GluN2A and GluN2B subunits. We find that DNRAbs potentiate GluN2A/GluN2B tri-heteromeric receptors, suggesting that GluN2A confers dominance of antibody susceptibility to NMDA receptors. Our data supports in vivo findings of GluN2A-containing NMDA receptors being the primary target of excitotoxicity from DNRAbs. Taken together, these findings suggest that targeted therapies for NPSLE patients to improve neurocognitive function should be directed towards GluN2A.

Abstract: Fine tuning of glutamatergic synaptic transmission underlies information processing in the brain and is often disrupted in neural disease and psychiatric disorders. Regulatory control of this transmission arises from subsynaptic protein organization at the nanometer scale. Our lab recently reported that presynaptic sites of evoked neurotransmitter release, indicated by clusters of active zone proteins such as RIM, are aligned across the synapse with postsynaptic nanoclusters of glutamate receptors and scaffold proteins such as PSD-95. This alignment is spatially confined to ~80 nm in diameter, much smaller than the size of the entire synapse (~300 nm). This "nanocolumn" organization is expected to increase synaptic strength by enhancing the probability of receptor activation during synaptic transmission and may reveal a modular structure important for synaptic plasticity. Crucially, the mechanism by which this trans-cellular alignment is established is unknown. One unexplored explanation with important implications for synaptic function involves the glutamate receptors themselves. AMPA and NMDA receptors are arranged in nanoclusters and are enriched within the nanocolumn. Along with their known interactions with PSD-95, the receptors have demonstrated interactions with trans-synaptic proteins - AMPARs with LRRTMs, N-Cadherin, neuroligin-neurexin, and neuronal pentraxins; NMDARs with neuroligin-neurexin as well as EphB receptors. Furthermore, the GluA2 subunit of AMPARs has been implicated in synaptogenesis and maturation of presynaptic release sites. NMDARs are also known for their involvement in synaptogenesis and silent synapses, highlighting the possibility that they may play a key role in organizing synaptic structure. Given these observations, I hypothesize that glutamate receptors mediate trans-synaptic alignment of RIM1/2 and PSD-95 nanoclusters. There are two main predictions of this hypothesis that I will test. First, synapses lacking receptors should have altered alignment. I will thus use miRNA knockdown in dissociated hippocampal neuronal culture as well as conditional knockout in hippocampal slice culture to examine trans-synaptic alignment when AMPARs and/or NMDARs have been reduced or eliminated. Second, repositioning AMPARs and NMDARs should prompt reorganization of both pre- and postsynaptic scaffolding proteins. I will test this by using optical dimerization to acutely alter receptor content at the synapse. First, I will recruit additional AMPARs and/or NMDARs to synapses and examine changes in nanoscale organization and alignment as a test of receptor sufficiency for structural changes. Next, I will change the position of the receptors within the synapse by dimerizing them to specific cell adhesion molecules with known, distinct distributions with the synapse. I predict that positioning receptors at atypical sites within the synapse will cause consequent rearrangements to both pre- and postsynaptic scaffolding. Glutamate receptors, though obviously critical for synaptic transmission, are often thought to be simply passive contributors to synapse function as they sit within the synapse or diffuse in and out. My experiments will test a dramatically different view of the receptors: that they actively shape the structure of synapses and thus control their own activation.

Abstract: We developed a virtually complete 3D model of the human GluN1-GluN2B receptor based on recently solved crystal structures for frog and rat NMDARs2,3. The human NMDAR structure enabled the mutated residues to be located thereby allowing insight into their potential impact on neighbouring amino acids, including those forming the binding sites for Mg2+ and NMDAR channel blockers. The characterisation of the mutants revealed reductions in glutamate potency, increased receptor desensitisation and ablation of voltage-dependent Mg2+ block. Moreover, introducing the mutations into neurons revealed profound changes to excitatory neurotransmission that could underlie their associated neurological disorders. By attempting to reverse the gain-of-function phenotype with the channel blocker memantine revealed that the mutations significantly altered the sensitivity of the NMDAR to the blocker. Using our experimental data, we also devised a new kinetic model to understand the mechanism of action of memantine and how it is affected by endogenous levels of Mg2+. The functional effects of the missense mutations revealed gain and loss-of-function that affected NMDAR ligand binding and ion channel properties. These profiles may underpin epileptic phenotypes of patients that carry the mutations. Moreover, the study indicates that memantine could be a promising treatment for selected gain-of-function mutations in NMDA receptors.

References:
1. Burnashev N and Szepetowski P (2014) Curr Opin Pharmacol 20: 73-82
2. Lee C-H. et al. (2014) Nature 511: 191-197.
3. Karakas E and Furukawa H (2014) Science 344: 992-997.

Abstract: NMDA receptors (NMDARs) are glutamate-gated ion channels that are central to numerous brain functions and pathologies. Key determinants of their diverse roles in synaptic physiology are the type (selectivity) and magnitude (conductance) of ions the ion channel pore allows to pass when in the open or conducting conformation. Despite the potential of targeting ion permeation in the clinic, its basis remains incompletely defined. NMDARs are composed of two obligatory GluN1 and typically two GluN2 subunits and have a layered structure with the ligand-binding domain (LBD) positioned extracellularly to the membrane-embedded transmembrane domain (TMD), forming the ion channel. The central ion permeation pathway is formed by the M2 pore loop and the M3 transmembrane segment. Surrounding this core pathway is the M1 and M4 transmembrane segments. To access the central permeation pathway, the physiological ions Na+ and Ca2+ as well as clinically relevant open channel blockers like Memantine® must pass through short polypeptide linkers that connect the LBD (S1, S2) to the TMD, the LBD-TMD linkers. To identify the specific ion permeation pathways or "portals" to the ion pore, we focused on charged amino acid side chains in these LBD-TMD linkers: S1-M1, M3-S2, and S2-M4 in the GluN1 and GluN2A subunits. The LBD-TMD linkers in GluN1/GluN2A receptors contain 16 negatively charged and 9 positively charged amino acid side chains (25 total). We assumed that if an ion used a particular pathway, a charge reversal would lead to an increase (positive-to-negative) or decrease (negative-to-positive) conductance to the ion pore. To test this hypothesis, we recorded single channel activity of wild type and individual mutant receptors containing charge-reversals to measure single channel slope conductance. Slope conductance was extracted from single channel current amplitudes recorded at different voltages (-100 mV, -75 mV, -50 mV and -25 mV). We also extracted gating properties (open probability, man open/closed times) of these receptor at -100 mV. Our experiments indicate that a major portal is formed by the S2-M4 linker in the GluN2A subunit. Understanding the interaction of the linker regions and ion flux can allude to a "permeation map" aiding in developing more specific targeting of the NMDAR ion channels than the current open channel blockers such as Memantine®.

Abstract: NMDA receptors are Ca2+-permeable channels gated by glutamate and are essential for excitatory syn-aptic transmission during physiologic and pathologic brain processes. Ca2+-dependent inactivation (CDI) by calmodulin is a regulatory mechanism that reduces NMDA receptor gating in an activity-dependent manner. Although the obligatory GluN1 subunit directly mediates CaM binding and is required for CDI, the present literature suggests that GluN2B receptors are insensitive to CDI. We used electrophysiological recordings of recombinant NMDA receptors to examine the mechanism of CDI subtype-dependency. Whole-cell currents recorded in the presence of 2 mM external Ca2+ and overexpression of calmodulin showed that CDI is robust for GluN2A channels (CDI2A = 0.51 ± 0.05) but appeared absent for GluN2B (CDI2B = 0.04 ± 0.07). However, when Ca2+ was dialyzed intracellularly, robust CDI occurred for both GluN2A and GluN2B channels (CDI2A = 0.80 ± 0.06; CDI2B = 0.52 ± 0.06), indicating that GluN2B channels are susceptible to CDI. To investigate this apparent discrepancy, we used patch clamp fluorometry to record single-channel Na+-only currents from individual NMDA receptors and simultaneously monitor intracellular Ca2+ elevations upon ionomycin treatment. We found upon Ca2+entry, the derived kinetic models suggested that the gating kinetics of both GluN2A and GluN2B channels were perturbed by a common kinetic mechanism where by reduced activation resulted in the accumulation of channels in long-lived closed (desensitized) states. Thus, the apparent subtype-dependent susceptibility to CDI by external Ca2+ reflects intrinsic subtype differences in gating kinetics such as equilibrium open probability (Po) which determine the amount of fluxed Ca2+. To test this hypothesis, we measured CDI by external Ca2+ using whole cell recordings from GluN2B channels carrying a Lurcher (Lc) mutation, GluN1A652Y, which dramatically increased channel Po (Po, wt = 0.19 ± 0.07; Po, Lc = 0.81 ± 0.08). We found that 2 mM external Ca2+ produced robust CDI for these high-Po GluN2B channels (CDILc = 0.54 ± 0.06). Together, these results show that Ca2+ produces activity-dependent CDI for both GluN2A and GluN2B receptors. These results are consistent with CDI as a neuroprotective mechanism against excessive Ca2+ load during high-Po activation.

Abstract: N-Methyl-D-aspartate receptors (NMDARs) play essential roles in memory formation, neuronal plasticity and brain development with their dysfunction linked to a range of disorders from ischemia to schizophrenia. Zinc and pH are physiological allosteric modulators of NMDARs with GluN2A containing receptors inhibited by nanomolar concentrations of divalent zinc and by excursions to low pH. Despite the widespread importance of zinc and proton modulation of NMDARs, the molecular mechanism by which these ions modulate receptor activity has proven elusive. We used cryo-electron microscopy to elucidate the structure of the GluN1/GluN2A NMDAR in a large ensemble of conformations under a range of physiologically relevant zinc and proton concentrations. We show how zinc binding to the amino terminal domain elicits structural changes that are transduced though the ligand-binding domain and result in constriction of the ion channel gate.

Abstract: Ionotropic glutamate receptors (iGluRs) are tetrameric ion channels that mediate excitatory neurotransmission. Recent structures of α-amino-3-hydroxy-5-methyl-4-isoxazole propionate (AMPA) and N-methyl-D-aspartate (NMDA) receptors permit a comparative analysis of whole-receptor dynamics for the first time. Despite substantial differences in the packing of their two-domain extracellular region, the two iGluRs share similar dynamics, elucidated by elastic network models. Motions accessible to either structure enable conformational interconversion, such as compression of the AMPA receptor toward the more tightly packed NMDA receptor conformation, which has been linked to allosteric regulation. Pivoting motions coupled to concerted rotations of the transmembrane ion channel are prominent between dimers of distal N-terminal domains in the loosely packed AMPA receptor. The occurrence and functional relevance of these motions is verified by cross-linking experiments designed to probe the computationally predicted distance changes.

Abstract: Ionotropic glutamate receptors (iGluRs) play a fundamental role in the central nervous system and their dysfunction has been implicated in many neurobiological disorders/diseases. As such they are important targets for drug discovery. We perform a systematic study of the druggability for iGluR sub-families using molecular dynamics simulations in the presence of drug-like probe molecules. First we verified the applicability of our method by confirming known agonist and modulator sites on AMPA receptors (AMPARs) and NMDA receptors. Domain closure of ligand-binding domain, which is critical to channel opening, was also observed in our simulations. Next we explored the N-terminal domain (NTD) dynamics. Our study identified a highly probable drug-binding site at the NTD dimer-dimer interface of AMPAR GluA3. This is attributed to the enhanced conformational flexibility of GluA3 that we explored further with three new crystal structures in different states.

Abstract: We discuss molecular dynamics (MD) simulations studies of AMPA receptor in open, active and desensitized states. We then focus on structural changes as the receptor traverse between these functional states. We report on the discovery of couplings between various parts of the receptor upon desensitization.

Abstract: Non-competitive inhibitors of AMPA receptors have attracted significant interest in the recent years as antiepileptic drugs. While a considerable number of small molecules have been identified and tested as AMPA receptor non-competitive antagonists, only pyridone perampanel (PMP) is available for clinical use as an antiepileptic drug. PMP, however, still causes side effects at higher doses and the need for safer and more effective drugs remains. Recently resolved X-ray crystal structures of AMPA receptor complexes of GYKI 53655 (GYKI), CP 465022 (CP), and PMP provided new structural information on non-competitive inhibitor binding to AMPA receptors. Crystal structures and previous mutagenesis studies show that while these structurally dissimilar inhibitors share a common binding site, each ligand interacts differently with individual amino acid residues. However, due to the limited resolution of the crystal structures they do not provide a detailed atomistic picture of specific protein-ligand interactions at the binding pocket. We carried out molecular docking along with molecular dynamics (MD) simulations and binding energy calculations to assess the stability of crystallographically determined binding modes and explore other potential binding modes of GYKI, PMP, and CP. Our MD studies highlight the features of the AMPA receptor non-competitive inhibitor binding site that are important in accommodating structurally different inhibitors. This information will aid in structure based design of new non-competitive inhibitors that target AMPA receptors.

Abstract: GABAA receptors are pentameric chloride channels that are composed from 19 possible subunits (alpha1-6, beta1-3, gamma1-3, delta, epsilon, theta, pi, and rho1-3). The canonical GABAA receptor is thought to contain 2 alpha subunits, 2 beta subunits, and a 5th subunit - typically gamma. However, non-canonical subunit combinations can also create functional receptors with unique pharmacology. Theta-containing combinations have been shown to create non-canonical, plasma membrane-expressed receptors with unknown pharmacology as they are not activated by GABA or known GABAA ligands. theta is expressed discretely, primarily in aminergic brain regions involved in mood and arousal. Therefore, we believe the theta subunit could prove to be a novel, druggable target for aminergic dysfunction without widespread side effects. However, to date, there are only two contradicting publications that explore the role of theta in GABAA pharmacology. Our lab has observed that theta does not assemble into pentamers when expressed in recombinant HEK293 or COS7 cells in combinations previously shown to be surface-expressed. Non-permeabilized staining of HAtheta in these combinations was nonexistent in our hands; however, HAtheta expression was confirmed intracellularly and the protein was not degraded, as shown by permeabilized staining and Western blotting. Whole-cell voltage-clamp comparisons of alpha3beta3, alpha3theta, and alpha3-only transfections reveal that alpha3theta is not different than alpha3-only with regard to current amplitude (both ~90 pA), EC50 (both ~30uM), and histamine potentiation (both ~300% potentiation). The absence of theta assembly could be due to an absence of a necessary accessory protein, the presence of a protein that inhibits assembly, or cell type and species differences. To circumvent the problem with recombinant systems, we have identified a cell line of interest (U87MG) that natively expresses the theta and alpha3 subunit mRNAs to a much greater degree than the other GABAA subunit mRNAs. If theta is assembled into the GABAA receptors in this cell line, the results of this study are expected to reveal any interacting proteins as well as expand our knowledge about the pharmacological contributions of theta to GABAA function.

Abstract: The activated state of AMPA receptors can be stabilized by small molecules that bridge the ligand binding domain dimer interface. Here, we present on the biophysical characterization of a benzothiadiazine dioxide type modulator with nanomolar potency. Small angle X-ray scattering (SAXS) experiments and NMR spectroscopy were utilized to determine the modulator's intrinsic binding affinity and binding orientation within the GluA2 ligand-binding domain dimer. Our results complement observations from calcium flux experiments and the modulator-bound X-ray structure. J. Med. Chem. 2018, 61, 251-264

Abstract: NMDA receptors (NMDARs) are important mediators of synaptic plasticity and memory. Native NMDARs consist of two GluN1 subunits with two GluN2A-D subunits, and possibly GluN3. The GRIN1 gene, which encodes GluN1 has 8 splice variants whereas the GRIN2s are unspliced. The functional roles for various GluN2 subunits in synaptic physiology have been well characterized but the biological functions of GRIN1 splice variants remain unknown in physiological contexts. Here, we examined the role of alternative splicing of GluN1 exon 5, which encodes the N1 cassette. We generated mice lacking exon 5 (GluN1a) or compulsorily expressing this exon (Glun1b). Here we show that GluN1a and GluN1b mice are viable, develop normally and have normal basal synaptic transmission at CA3-CA1 synapses of the hippocampus. The presence or absence of exon 5 did not change the input-output relationship of basal synaptic transmission. There were no significant differences between wild-type, GluN1a or GluN1b mice in their ratio of NMDAR- to AMPAR-mediated excitatory postsynaptic currents (EPSCs) nor in their paired-pulse ratio (PPR) of fEPSPs. Current-voltage relationship (I-V) of NMDARs does not change by exon 5 splicing, suggesting NMDAR channel function was not altered. However, we found that theta burst stimulation (TBS)-induced long-term potentiation (LTP) was significantly lower in hippocampal CA1 slices taken from GluN1b mice (125.48+4.60% n=10, p < 0.05 vs WT or GluN1a) compared to slices taken from either wild-type (150.39+4.61%, n=19) or GluN1a mice (147.95+5.39%). In contrast, slices from GluN1a mice had significantly stronger LTP when induced by 20Hz stimulation versus GluN1b or wild-type slices (148.03+2.86% n=6 for GluN1a, p < 0.05 vs WT or GluN1b; 130.15+1.64% n=9 for WT; 130.02+2.26% n=8 for GluN1b). Additionally, GluN1b mice had reduced learning and memory performance in the Morris-water maze test compared with wild-type mice while GluN1a mice performed better than wild-type mice. Together, these data show that splicing of exon 5 modulates LTP strength, and learning and memory behavior in mice. We have thus defined for the first time a biological function for alternative splicing of GluN1 in vivo.

Authors: Samantha Skobel, Hou-Ming Cai, Mark A. Rutherford, Maria E. Rubio

Abstract: Rapidly gating AMPA-type glutamate receptors (AMPAR; GluA2, GluA3 and GluA4 subunits) mediate synaptic transmission at the mature synapse between the inner hair cells (IHC) and the afferent fibers of the cochlear nerve (IHC synapse). However, the contribution of each type of AMPAR subunit to overall glutamatergic receptor function and afferent transmission/sensitivity in the cochlea is poorly understood. Understanding this process is important because glutamate excitotoxicity through AMPAR has been implicated in the pathogenesis of hearing loss caused by noise, ischemia, and aging. Humans show sex differences in the vulnerability to hearing loss, e.g. the male humans show greater prevalence of high-frequency hearing loss (Agrawal et al., 2008), however, the contributing biological factors are unknown. We therefore began investigating the contribution of AMPAR subunits to cochlear function, and assessed sex-specific differences in AMPAR subunits that may contribute to sex differences in excitotoxic vulnerability of IHC synapses. Our functional (Auditory Brain Stem Responses, ABRs) and ultrastructural data (Electron Microscopy) show that GluA3 AMPAR subunits have a critical role in the sexually dimorphic vulnerability to hearing loss. With confocal microscopy and quantitative immunofluorescence analysis, we show that spiral ganglion neurons and IHC synapses of GluA3 knockout (KO) mice, contain less GluA2 but more GluA4. Tone-pip ABR thresholds show that GluA3-KO mice have hearing loss across all frequencies. Intriguingly, we find that female GluA3-KO mice show more hypersensitivity to sound-induced cochlear damage than male littermates. Our goal is to define mechanistically how GluA3 contributes to the structural and molecular components of IHC synapses and to sex differences that underlie the hypersensitivity to sound-induced cochlear damage. Based on these results and by analogy to humans, we hypothesize that female humans are more resistant to high-frequency hearing loss because they have more GluA3 and thus more GluA2 and less calcium influx.

Abstract: Women represent a vulnerable population in the development alcohol use disorders, as they transition to dependence faster, show greater craving in response to alcohol-associated cues and stress, and exhibit higher psychiatric comorbidity than men. While relapse rates are similar across gender, women are more likely to relapse when faced with stress or negative mood states during abstinence, suggesting that these factors are important when developing gender-based treatment strategies. In parallel with this hypothesis, we recently reported that female rats show increased alcohol craving-like behavior relative to males in response to cues and stress. Thus, identifying sex-specific mechanisms mediating relapse-related behavior may improve treatment outcomes. We have begun to investigate these differences by determining sex differences in protein expression of glutamate receptor subunits in cortico-limbic circuits known to regulate reinstatement. In addition, we also tested the effect of the NMDA antagonist ketamine and the AMPA antagonist NBQX on reinstatement of alcohol seeking in males and females. Male and female rats were trained to press a lever for access to a dipper containing 10% ethanol in a 0.1% saccharin solution. Control rats responded for saccharin. Each reinforcer delivery was accompanied by a 10s light+tone cue. After >14 days of self-administration, the effects of low dose ketamine (10 mg/kg) on self-administration was determined. In addition, the effect of ketamine and NBQX at a range of doses on alcohol craving/relapse-like behavior was also determined. Finally, the amygdala was dissected from ketamine exposed and drug-naive male and female rats for analysis of glutamate receptor subunit expression by Western blot. We found that females had significantly greater expression of GluA1, GluA2/3, and a trend towards increased NR2B in the amygdala relative to males, with no differences in the PFC. Ketamine pre-treatment significantly attenuated both alcohol and saccharin self-administration in females, while only affecting saccharin self-administration in males. Ketamine also completely blocked reinstatement ("craving") of alcohol and saccharin seeking in females, while only reducing reinstatement to saccharin seeking in males. Alcohol seeking also tended to be blocked by a lower dose of ketamine (3 mg/kg) in females, that was not effective in males. Ketamine's anti-craving-like effects were not altered by prior NBQX treatment, and were transient in duration, lasting 48 hrs. NBQX alone was capable of reducing alcohol craving in both males and females. Ketamine treatment also produced changes in the membrane vs. cytosolic expression of glutamate receptor subunits in females. In conclusion, females are more sensitive to reinstate alcohol seeking in response to a combination of cues and a pharmacological stressor than males. The increased reinstatement may be related to enhanced AMPAR and NMDAR-related signaling in the amygdala. Ketamine is extremely effective in reducing reinstatement of alcohol-seeking selectively in females. Unlike the antidepressant-like effects of ketamine, the anti-relapse-like effects have a much shorter duration of action and are not mediated through AMPAR activation, and the AMPAR antagonist disrupts reinstatement on its own, suggesting that ketamine's potential beneficial effects for the two disorders are mediated by independent mechanisms. Nonetheless, these data suggest that ketamine may be an effective treatment for women with comorbid alcohol use and mood disorders.