GluK2 Q/R editing regulates kainate receptor signaling and long-term potentiation of AMPA receptors

Summary Q/R editing of the kainate receptor (KAR) subunit GluK2 radically alters recombinant KAR properties, but the effects on endogenous KARs in vivo remain largely unexplored. Here, we compared GluK2 editing-deficient mice that express ∼95% unedited GluK2(Q) to wild-type counterparts that express ∼85% edited GluK2(R). At mossy fiber-CA3 (MF-CA3) synapses GluK2(Q) mice displayed increased postsynaptic KAR function and KAR-mediated presynaptic facilitation, demonstrating enhanced ionotropic function. Conversely, GluK2(Q) mice exhibited reduced metabotropic KAR function, assessed by KAR-mediated inhibition of slow after-hyperpolarization currents (ISAHP). GluK2(Q) mice also had fewer GluA1-and GluA3-containing AMPA receptors (AMPARs) and reduced postsynaptic AMPAR currents at both MF-CA3 and CA1-Schaffer collateral synapses. Moreover, long-term potentiation of AMPAR-mediated transmission at CA1-Schaffer collateral synapses was reduced in GluK2(Q) mice. These findings suggest that GluK2 Q/R editing influences ionotropic/metabotropic balance of KAR signaling to regulate synaptic expression of AMPARs and plasticity.

In recombinant systems KARs containing unedited GluK2(Q) have a higher conductance than edited GluK2(R). 43,47Moreover, disrupting ADAR2-mediated GluK2 Q/R editing enhances the surface expression, single channel conductance, and Ca 2+ permeability of postsynaptic KARs in cultured neurons. 36,43,48Therefore, we measured basal KAR-mediated postsynaptic responses at MF-CA3 synapses in WT and GluK2(Q) mice.Because both the glutamate receptor complement and presynaptic facilitation at MF-CA3 synapses are stable after the 2 nd postnatal week 49,50 we used P14-P21 mice and confirmed the stability of AMPAR and KAR expression within this age range.
2][53] Postsynaptic responses were recorded from CA3 pyramidal neurons in whole-cell voltage clamp configuration in the presence of picrotoxin (50mM) and D-APV (50mM) to block GABA A Rs and NMDARs, respectively.The stimulation intensity applied to presynaptic axons was increased in small increments until a postsynaptic response was observed (Figure 1B).Interestingly, the percentage of trials that evoked a synaptic response (success rate) was decreased in the GluK2(Q) mice (Figure 1C) (WT = 43 G 5%, GluK2(Q) = 27 G 1%; unpaired t-test, p = 0.017), suggesting either a decrease in the number of release sites or reduced probability of glutamate release from the same number of sites.Furthermore, the average AMPAR-EPSC amplitude for successful trials was reduced in GluK2(Q) mice (Figure 1D) (WT = 94.5 G 8.8pA, GluK2(Q) = 59.1 G 3.5pA; unpaired t-test, p = 0.0044).

Enhanced presynaptic facilitation in GluK2(Q) mice
6][57] To investigate how GluK2 Q/R editing affects presynaptic KAR function at MF-CA3 synapses, we measured short-term facilitation of presynaptic release in the presence of picrotoxin (50mM) across a range of timescales.By measuring Enhanced postsynaptic KAR and reduced AMPAR currents at MF-CA3 synapses in GluK2(Q) mice (A) The intronic region between exon 12 (M2) and 13 (M3) in the grik2 gene contains an editing complementary site (ECS), located $1900nt downstream of exon 12.In the WT mice, this region is intact.However, in homozygous GluK2(Q) mice, a 600bp region is deleted from the ECS site.This prevents ADAR2 binding and subsequent editing of GluK2 pre-mRNA at the Q/R site, leading to the translation of >95% unedited GluK2(Q) subunits.EPSC AMPA , we examined paired-pulse facilitation (PPF), observed at 50ms stimulation intervals, and accumulation of frequency facilitation (FF), observed by changing the frequency of stimulation from 50ms to 1s intervals, in acute slices from WT and GluK2(Q) mice using previously described protocols. 15,18,19,58 In all experiments the purity of mossy fiber input was determined by addition of the group II metabotropic glutamate receptor agonist DCG-IV (2mM) 59 and recordings were excluded if there was <70% inhibition of synaptic responses.Furthermore, to exclude the possibility that differences in the purity of mossy fiber inputs contributed to the enhanced short-term facilitation in GluK2(Q) mice we analyzed the degree of DCG-IV inhibition.No differences were observed (Figure 2C) (WT = 91.5 G 2.1%, GluK2(Q) = 93.2G 1.2%; unpaired t-test, p = 0.4879) and there was no correlation between degree of inhibition by DCG-IV and PPR (Figures S1A and S1B) (WT, r = 0.277, R 2 = 0.0769, p = 0.383; GluK2(Q), r = À0.280,R 2 = 0.0784, p = 0.378; 95% confidence interval).There was also no correlation between EPSC amplitude and PPR (Figures S1C and S1D) (WT, r = 0.127, R 2 = 0.0163, p = 0.692; GluK2(Q), r = 0.0383, R 2 = 0.00147, p = 0.906; 95% confidence interval).EPSC initial amplitudes in response to the first stimulus were set to be similar between genotypes (WT = 156.6G 14.4pA, GluK2(Q) = 143.9G 17.8pA; unpaired t-test, p = 0.5830).
Taken together with the reduced success rate observed in the minimal stimulation experiments (Figure 1C), these data suggest that basal release probability is reduced, and presynaptic facilitation is enhanced, at MF-CA3 synapses in GluK2(Q) mice.

Metabotropic KAR function is impaired in GluK2(Q) mice
Because the GluK2 Q/R editing site is within the channel pore region it is unsurprising that editing impacts on KAR ionotropic signaling but the effects of Q/R editing on metabotropic signaling have not been investigated.We therefore assessed if GluK2 Q/R editing alters metabotropic function by measuring KAR inhibition of the slow afterhyperpolarization current (I sAHP ) in acute hippocampal slices. 16,25,26,60I sAHP currents were evoked in whole-cell voltage clamped CA3 pyramidal cells by depolarizing the membrane potential to 0mV from À50mV for 200ms in the presence of 50mM picrotoxin, 50mM D-APV, 40mM GYKI53655 and 1mM CGP55845 to inhibit GABA A Rs, NMDARs, AMPARs and GABA B Rs, respectively. 26Robust and stable I sAHP currents were obtained in both WT and GluK2(Q) mice, with no difference in baseline amplitudes between genotypes (Figure 3A) (WT = 67.0G 8.0pA, GluK2(Q) = 63.1 G 6.6pA; unpaired t-test, p = 0.712).Activation of synaptic KARs by MF stimulation (10 stimuli at 25Hz every 20 s for 10 mins), as previously described, 26,61 produced a consistent depression of I sAHP in WT mice but the extent of depression was decreased in GluK2(Q) mice (Figure 3B) (WT = 45.1 G 3.7%, GluK2(Q) = 26.7 G 2.7%; unpaired t-test, p = 0006).These data indicate that KAR metabotropic signaling is impaired in GluK2(Q) mice.

Impaired LTP at Schaffer collateral-CA1 synapses in GluK2(Q) mice
Since we observed a global reduction in the levels of AMPARs in the GluK2(Q) mice, we next characterized CA1 synapses by stimulation of a single presynaptic Schaffer-collateral axon at 1Hz in acutely prepared hippocampal slices. 49Postsynaptic responses were recorded from CA1 pyramidal neurons in whole-cell voltage clamp configuration in the presence of picrotoxin (50mM) and D-APV (50mM) to block GABA A Rs and NMDARs.As described for MF-CA3 synapses (Figure 1), stimulation intensity of presynaptic axons was increased until a postsynaptic response was observed (Figure 6A).The percentage of trials that failed to evoke a synaptic response (failure rate) was the same for both WT and GluK2(Q) mice (Figure 6B) (WT = 0.60 G 0.05%, GluK2(Q) = 0.69 G 0.04%; unpaired t-test, p = 0.2).However, consistent with MF-CA3 synapses and the loss of AMPARs from the synaptosomes, the average AMPAR-EPSC amplitude for successful trials was reduced in GluK2(Q) mice (Figure 6C) (WT = 28.8G 1.7pA, GluK2(Q) = 19.2G 1.8pA; unpaired t-test, p = 0.002).
These results further confirm the role of GluK2 Q/R editing in maintaining AMPAR-mediated synaptic transmission and indicate that this phenomenon occurs at different hippocampal synapses.
Since GluK2 editing prevents Ca 2+ entry through the KAR ion channel we wondered if predominant expression of unedited GluK2(Q), which traffic more efficiently and gate Ca 2+ , would reveal EPSC KAR at Schaffer collateral synapses.However, no EPSC KAR were detected following bursts of 3 stimuli at 167Hz given to Schaffer collaterals to drive multiple axons (Figure S2).
We next tested if the reduced synaptic expression of GluA1-and GluA3-containing AMPARs observed in GluK2(Q) mice impacted on the expression of long-term potentiation (LTP).We have previously shown that sharp-wave/ripple like (RL) patterns of activity induce LTP at Schaffer collateral-CA1 synapses that is expressed by increased AMPAR at the synapse, and therefore chose this synapse for these experiments. 24Extracellular field potential recordings from acute hippocampal slices revealed that high frequency stimulation that replicates the in vivo patterns of hippocampal RL activity 24,64 induced robust LTP of AMPAR-mediated EPSPs in WT mice but LTP was significantly reduced in GluK2(Q) mice (Figure 6D) (WT: 172.8 G 10.4% in test pathway vs. 97.7 G 3.8% in control pathway; unpaired t-test, p = 0.0002; GluK2(Q): 133.8 G 3.6% in test pathway vs. 97.2G 4.9% in control pathway; unpaired t-test; p = 0.0001; comparison between test pathways, unpaired t-test, p = 0.0087).The paired-pulse ratio remained unchanged after induction of LTP in both WT and GluK2(Q) mice (Figure 6E) (WT = 2.04 G 0.14 baseline, 1.74 G 0.17 after LTP, p = 0.325; GluK2(Q) = 2.13 G 0.14 baseline, 1.95 G 0.11 after LTP, p = 0.65; two-way ANOVA with Sidak's multiple comparison test).These data suggest that preventing KAR Q/R editing reduces basal synaptic AMPAR expression and impairs the synaptic recruitment of AMPARs required for LTP.

DISCUSSION
GluK2 Q/R editing is developmentally 40,65 and activity-dependently regulated 35,36 to modulate and accommodate distinct synaptic and network diversity.However, how GluK2 editing impacts on KAR function and subsequent downstream AMPAR function has not been explored.To address these outstanding questions we compared WT mice, which contain <15% unedited GluK2(Q), and GluK2 Q/R editing deficient (GluK2(Q)) mice that have >95% unedited GluK2(Q). 44n recombinant expression systems KARs containing unedited GluK2(Q) gate Ca 2+ and have a much greater single channel conductance than those containing edited GluK2(R) ($150ps compared to <10ps). 43We therefore predicted that postsynaptic ionotropic KAR function would be enhanced in GluK2(Q) mice.We found this to be the case, but the increase in EPSC KA we observed (WT = 4.66 G 0.63pA, GluK2(Q) = 7.31 G 0.50pA) was markedly less than expected based on the data from heterologous expression systems.However, it should be noted that, to our knowledge there is no available information in the literature on how heteromeric KARs containing GluK2(Q) or GluK2(R) subunits behave when combined with GluK4/GluK5 in invivo conditions.Moreover, recombinant KARs in heterologous cell lines generally lack auxiliary subunits and/or interacting proteins (e.g., Netos, KRIP6) that play crucial roles in regulating KAR localization, gating, and function. 1Thus, it is likely that cell-line based recombinant KAR systems exhibit different properties to endogenous KARs in primary cultured neurons or cultured ex vivo slices.We also anticipated that the increased conductance and Ca 2+ permeability of GluK2(Q)-containing KARs should boost presynaptic KAR function, resulting in enhanced short-term facilitation at both 50ms (PPF) and 1s (FF) timescales.This is indeed what we found in the mice, but we also observed an increase in failure rate in the minimal stimulation experiments.These data suggest an additional factor of reduced basal probability of release, or number of release sites, within the large presynaptic mossy fiber boutons, which on its own is predicted to increase presynaptic facilitation.Minimal stimulation at 3Hz is sufficient to engage KAR-mediated FF and would therefore be expected to produce a lower failure rate in GluK2(Q) mice.We observed the opposite effect, suggesting that the reduction in basal probability of release is substantial.Our data cannot distinguish between reduced probability of release and reduced number of release sites so further anatomical investigation would be necessary to address this.Nonetheless, overall, our data indicate that GluK2(Q) mice exhibit enhanced pre-and postsynaptic KAR function.
In stark contrast to their enhanced ionotropic KAR function, GluK2(Q) mice show reduced metabotropic function measured by inhibition of I sAHP at MF-CA3 synapses. 16,25,26,60Since I sAHP controls the excitability of CA3 pyramidal neurons and response to synaptic stimulation, 26 GluK2(Q) is predicted to cause a reduction in the regulation of synaptic integration by KARs.Possible explanations for the diminished metabotropic KAR signaling in GluK2(Q) mice include a reduction in expression of GluK1 and GluK2 KAR subunits and/or Q/R editing statedependent conformational changes which regulate metabotropic signaling.It remains unclear and controversial which, and how, specific KAR subunits contribute to KAR metabotropic signaling.Indeed, it has been proposed by different groups that GluK1, GluK2, or GluK5 are required for G protein coupling and metabotropic effects. 16,21,24,34,66Thus, although we cannot draw definitive mechanistic conclusions, our data demonstrate that GluK2(Q) mice show reduced metabotropic KAR signaling and, at the same time, enhanced ionotropic function.
Pre-and postsynaptic KARs induce bidirectional plasticity. 6,8,24,38,67Presynaptic KARs at MF-CA3 synapses induce LTD that is sensitive to Ca 2+ levels suggesting a direct activation of ionotropic KARs or activation of voltage-gated Ca 2+ channels through metabotropic KAR signaling. 68A critical role for postsynaptic KAR signaling is activity-dependent regulation of both KAR and AMPAR surface expression.For example, depending on the extent of activation, KARs can enhance or reduce AMPAR surface expression to evoke LTP or LTD via metabotropic or ionotropic signaling, respectively. 24,38Our data also show that editing impacts on NMDAR-induced LTP at Schaffer collateral synapses in CA1, with GluK2(Q) mice exhibiting reduced LTP.GluK2(Q) mice also had reduced AMPAR-EPSCs at mossy fiber and Schaffer collateral synapses and lower GluA1 and GluA3 levels in synaptosomal fractions, suggesting that KARs not only regulate activity-dependent AMPAR trafficking but may also maintain basal synaptic levels of Ca 2+ -permeable AMPARs.
It should be noted that activity-dependent regulation of KAR alters the synaptic expression of GluA1 and GluA2 AMPARs, while GluK2 editing specifically alters the basal synaptic expression of GluA1 and GluA3 containing AMPARs. 24,38Importantly, these changes were specific to AMPARs, since synaptic levels of the NMDAR subunits GluN1 and GluN2A were unchanged, indicating the loss of AMPARs is not due to wholescale changes in synaptic composition in the editing-deficient mice.
These findings demonstrate that KAR activity mediates the 'tone' of synaptic AMPARs at MF-CA3 and Schaffer collateral synapses and suggests a wider role where KAR signaling may set the tone for synaptic AMPAR composition more broadly.We speculate that this may be a homeostatic mechanism in which the presence of high-conductance, Ca 2+ -permeable GluK2(Q)-containing KARs causes a compensatory decrease in Ca 2+ -permeable GluA1/GluA3-containing AMPARs to balance synaptic responsiveness.
Unlike recombinant systems or manipulated primary neuronal cultures, the transgenic mice we use have a germline mutation.Thus, the overwhelming majority of GluK2 is unedited throughout development.Therefore, the downregulation of KAR subunit synaptic expression could be a homeostatic mechanism to minimize the excitotoxic effects of Ca 2+ entry through the Ca 2+ -permeable receptors.While we did not assess how AMPAR subunit expression and activity is altered through development in GluK2(Q) compared to WT mice, we suggest that this may correlate with the observed effects of KAR signaling on expression of AMPARs during the development of synaptic circuits, 50,69 and our results indicate that this developmental regulation may extend into adulthood.Furthermore, although we have not investigated this directly, it remains possible that altered KAR signaling and reduction in GluA1 and GluA3 AMPAR subunits could impact neuronal development and network formation.
In healthy adult brain GluK2-containing KARs predominantly comprise edited GluK2(R). 70Using mice that almost exclusively express only GluK2(Q) we show that the ionotropic/metabotropic balance of KAR signaling is radically altered by a lack of GluK2 editing.Based on these results, we propose that unedited GluK2(Q)-containing KARs primarily or exclusively function as ion channels with enhanced conductance for both mono and/or di-valent cations, whereas the edited GluK2(R)-containing KARs act as metabotropic receptors to regulate and maintain network activity.
These findings are important because GluK2 Q/R editing is subject to both developmental and activity-dependent control. 36Moreover, it has been reported that the proportion of edited GluK2(R) is increased to 85% in patients with a pharmaco-resistant temporal lobe epilepsy (TLE), 70 raising the possibility that increased inhibition of I sAHP , and thereby hyperexcitability, could underpin seizure generation.Conversely, reduction in KAR-mediated inhibition of I sAHP in GluK2(Q) mice might render protection against spontaneous seizures.Taken together our data indicate that physiologically and pathologically relevant alterations in GluK2 editing may dynamically regulate KAR function, signaling mode, maintenance of functional neuronal networks, and set the threshold for the induction of plasticity.Thus, in conclusion, our results highlight that GluK2 Q/R editing acts as a previously unsuspected molecular switch that regulates the enigmatic dual-mode capability of KARs to operate via either ionotropic or metabotropic signaling, to initiate distinct and diverse downstream pathways.

Limitations of the study
We are mindful that this study is mainly electrophysiological and therefore does not address the molecular pathways and mechanisms underlying altered KAR signaling in the absence of GluK2 Q/R editing, and the subsequent reduction in synaptic expression of the GluA1 and GluA3 AMPAR subunits.In addition, although we show altered synaptic expression of KAR subunits, further investigations are required to understand how Q/R editing modulates the composition of functional receptors.We further reveal a reduction in GluA1 and GluA3 subunits, likely indicating that KAR-meditated signaling specifically regulates Ca 2+ permeable AMPARs.However, further investigations will be required to address these questions directly.
We believe the data presented provide compelling evidence that GluK2 Q/R editing significantly regulates KAR signaling and synaptic functions at various synapses in the hippocampus.This work, however, is mainly restricted to young animals (P14-P21), and we have yet to identify how a lack of GluK2 Q/R editing affects synaptic function in adults.Thus, a detailed investigation of the molecular pathways, receptor composition, developmental regulation, and modulation of neuronal excitability by I SAHP provide exciting avenues for future research to understand the role of KAR editing in various neurological and neurodevelopmental disorders, and its potential for future therapeutic intervention.The cells were held in voltage clamp mode and evoked EPSCs were obtained by stimulating the mossy fiber pathway with a bipolar stimulating electrode placed in the dentate gyrus hilus layer (or glass monopolar electrode for minimal stimulation experiments).Picrotoxin (50mM) was included in the aCSF to inhibit GABA A receptors (except for CA1 field recordings).Cells with series resistance above 30 MU or where series resistance changed by >20% were excluded from analysis.To confirm the purity of mossy fiber inputs, the group-II mGluR agonist DCG-IV (2mM) was bath applied for 5-10 min at the end of experiments with mossy fiber stimulation. 26Recordings were only included in analysis if DCG-IV reduced EPSCs by >70%.

STAR+METHODS KEY RESOURCES
Minimal stimulation CA3 pyramidal cells.Cells were voltage clamped at À60mV and MF-EPSCs were evoked by moving a mono-polar stimulating electrode filled with caesium-based whole-cell solution (in mM: NaCl, 8; CsMeSO 4 , 130; HEPES, 10; EGTA, 0.5; MgATP, 4; NaGTP, 0.3; QX314.Cl, 5; Spermine, 0.1) around the inner border of dentate gyrus granule cells until a response was observed.Stimulation intensity was adjusted just above the threshold for activation of a synaptic response.Consecutive traces were recorded at a frequency of 3Hz.No prominent polysynaptic activation was observed using this low intensity stimulation.AMPAR-EPSCs were measured for 5-10 min in the presence of D-APV (50mM) for a minimum of 150 trials.Subsequently, GYKI53655 (40mM) was applied to block AMPAR responses and KAR-EPSCs were measured after 15 min of GYKI53655 application and for 5-10 min and a minimum of 150 trials.
CA1 pyramidal cells.Cells were voltage clamped at À60mV and Schaffer collaterals were evoked by moving a mono-polar stimulating electrode filled with caesium-based whole-cell solution (in mM: NaCl, 8; CsMeSO 4 , 130; HEPES, 10; EGTA, 0.5; MgATP, 4; NaGTP, 0.3; QX314.Cl, 5; Spermine, 0.1) around the stratum radiatum until a response was observed.Stimulation intensity was adjusted just above the threshold for activation of a synaptic response.Consecutive traces were recorded at a frequency of 1Hz.No prominent polysynaptic activation was observed using this low intensity stimulation.AMPAR-EPSCs were measured for 5-10 min in the presence of D-APV (50mM) for a minimum of 150 trials.

KAR/AMPAR ratio
CA3 pyramidal neurons were voltage clamped at À60mV in the presence of D-APV (50mM).Mossy fibers were stimulated with a burst of 3 stimuli at 167Hz every 20s to evoke AMPAR/KAR-EPSCs.Stable AMPAR-EPSCs were recorded for 20 min and then KAR-EPSCs were recorded for 20 min in the presence of the AMPAR antagonist GYKI53655 (40mM).

Paired-pulse and frequency facilitation
CA3 neurons in acute hippocampal slices were voltage clamped at À70mV in the presence of picrotoxin (50mM).To measure PPF, EPSCs were evoked by pairs of stimuli to mossy fibers at an inter-stimulus interval of 50ms, every 20s.For FF experiments, single stimuli were given at 0.05Hz for 10 min before stimulation frequency was increased to 1Hz for 1 min.After this, stimulation frequency was returned to 0.05Hz.DCG-IV (2mM) was then applied for at least 5 min to assess purity of MF input.

Field potential recordings
Extracellular field potentials (fEPSPs) were recorded from stratum radiatum in CA1 using a 3-5 MU glass pipette filled with aCSF.Two stimulation electrodes (bipolar) were positioned on opposite sides of the recording electrodes equidistant from the pyramidal layer to evoke two independent inputs (Stim1 and Stim2).LTP induction protocol was delivered only to Stim1 and was alternately positioned closer to the CA3 region or to subiculum in different recordings.Paired stimuli (50ms inter-stimulus interval) were given every 10s to each pathway, alternating between the control and test pathway (Stim1 and Stim2).LTP consisted of 20 bursts of 20 stimuli at 200Hz given every 5s.The recordings were performed in the absence of picrotoxin.

Figure 3 .
Figure 3. Impaired metabotropic KAR function in GluK2(Q) mice (A) Average baseline amplitude of I sAHP currents in WT (red) and GluK2(Q) (blue) mice.(B) Sample traces from WT before (red) and after synaptic KAR stimulation (green) (top left) and GluK2(Q) mice (blue) before and after synaptic KAR stimulation (top middle) and EPSC KA following synaptic stimulation (top right).Timeline showing inhibition of I sAHP following synaptic KAR activation (bottom left).Quantification of percentage inhibition of I sAHP following synaptic KAR activation in WT and GluK2(Q) mice (bottom right).N = 4 animals, n = 13 cells, ns p > 0.05, **p < 0.0001; un-paired t-test with Welch's correction.

Figure 4 .
Figure 4. Altered synaptic KAR subunit expression in GluK2(Q) mice (A) Representative Western blots of total and synaptosomal fraction samples from a single cerebral hemisphere of WT or GluK2(Q) mice for KAR subunits (left).Quantification of proteins expressed as percentage of WT protein after normalizing to b -actin (right).b-actin was used as a loading control.N = 7 animals; ns p > 0.05, *p < 0.05, **p < 0.01; unpaired t-test with Welch's correction.(B) Representative Western blots of total and synaptosomal fraction samples from a single cerebral hemisphere of WT or GluK2(Q) mice for auxiliary KAR subunits Neto1 and Neto2 (left) Quantification of proteins expressed as percentage of WT protein after normalizing to b-actin (right).b-actin was used as a loading control N = 7; ns p > 0.05, *p < 0.05, **p < 0.01; unpaired t-test with Welch's correction.

Figure 5 .
Figure 5. Reduced synaptic AMPAR expression in GluK2(Q) mice (A) Representative Western blots of total and synaptosomal fraction samples from a single cerebral hemisphere of WT or GluK2(Q) mice for the AMPAR subunits GluA1-3 (left).Quantification of proteins expressed as percentage of WT protein after normalizing to b-actin (right).b-actin was used as a loading control.N = 7 animals; ns p > 0.05, *p < 0.05, **p < 0.01; unpaired t-test with Welch's correction.(B) Representative Western blots of total and synaptosomal fraction samples from a single cerebral hemisphere of WT or GluK2(Q) mice for the NMDAR subunits GluN1 and GluN2A (left).Quantification of proteins expressed as percentage of WT protein after normalizing to b-actin (right).b-actin was used as a loading control N = 6 animals (B); ns p > 0.05, *p < 0.05, **p < 0.01; unpaired t-test with Welch's correction.