Pro-death NMDA receptor signaling is promoted by the GluN2B C-terminus independently of Dapk1.

Aberrant NMDA receptor (NMDAR) activity contributes to several neurological disorders, but direct antagonism is poorly tolerated therapeutically. The GluN2B cytoplasmic C-terminal domain (CTD) represents an alternative therapeutic target since it potentiates excitotoxic signaling. The key GluN2B CTD-centred event in excitotoxicity is proposed to involve its phosphorylation at Ser-1303 by Dapk1, that is blocked by a neuroprotective cell-permeable peptide mimetic of the region. Contrary to this model, we find that excitotoxicity can proceed without increased Ser-1303 phosphorylation, and is unaffected by Dapk1 deficiency in vitro or following ischemia in vivo. Pharmacological analysis of the aforementioned neuroprotective peptide revealed that it acts in a sequence-independent manner as an open-channel NMDAR antagonist at or near the Mg2+ site, due to its high net positive charge. Thus, GluN2B-driven excitotoxic signaling can proceed independently of Dapk1 or altered Ser-1303 phosphorylation.

While multiple pathways contribute to excitotoxicity (Tymianski, 2011), the mechanism by which CTD 2B is thought to potentiate excitotoxicity is upstream of all of them (Parsons and Raymond, 2014;Tu et al., 2010;Lai et al., 2014). The mechanism is centred on Ser-1303 of CTD 2B , within a region of the CTD unique to GluN2B, and with which CaMKIIa is known to interact and phosphorylate (Bayer et al., 2001;Mao et al., 2014). It was reported that in response to ischemia or excitotoxic insults, a different kinase, Dapk1, causes Ser-1303 phosphorylation which increases NMDARdependent ionic flux (Tu et al., 2010). Consistent with this, Dapk1 -/neurons were reported to be resistant to excitotoxicity, and a cell-permeable peptide mimetic of the CTD 2B region around Ser-1303 disrupted Ser-1303 phosphorylation and was neuroprotective (Tu et al., 2010).
Given that the GluN2B-Dapk1 pathway is prominent in contemporary models of excitotoxicity (Parsons and Raymond, 2014;Lai et al., 2014) we sought to investigate this pathway further. Dapk1 has not hitherto emerged from proteomic post-synaptic density screens (Husi et al., 2000;Collins et al., 2006;Yoshimura et al., 2004;Li et al., 2004;Jordan et al., 2004;Peng et al., 2004;Cheng et al., 2006;Bayés et al., 2011), and we failed to detect it in a recent proteomic analysis of native NMDAR supercomplexes (Frank et al., 2016). Moreover, the use of cell-permeable peptides to draw wide-ranging mechanistic conclusions can be problematic without extensive controls. We investigated whether Dapk1-mediated Ser-1303 phosphorylation indeed represents the major reason why CTD 2B promotes excitotoxicity signaling better than CTD 2A , using approaches that include analysis of a new Dapk1 knockout mouse.
We confirmed that the antibody is capable of detecting changes in Ser-1303 phosphorylation in neurons: phospho-GluN2B(Ser-1303) levels in cortical neurons, as assayed by western blot using this antibody, are lowered after incubation of cortical neurons with the general kinase inhibitor staurosporine, and increased modestly by a cocktail of phosphatase inhibitors okadaic acid and FK-506 (Figure 1-figure supplement 1a,b, Figure 1-source data 1). As further evidence of specificity, we found that the antibody completely failed to react with GluN2B in which we had engineered mutations (L1298A/R1300N/S1303D) into the site for a separate study (Figure 1-figure supplement 1c,d).
We found that bath application of NMDA at excitotoxic concentrations failed to induce significant Ser-1303 phosphorylation (Figure 1a and b, Figure 1-source data 2). At the late timepoint (60 min, 50 mM NMDA) we observed a decline in Ser-1303 phosphorylation (Figure 1a Figure 1-source data 3) as well as a decline in total levels of GluN2B, consistent with observations of others who have reported partial calpain-mediated cleavage and degradation of the NMDAR CTD (Dong et al., 2006;Gascó n et al., 2008). These observations in DIV10 cortical neurons were also repeated at DIV16 (Figure 1-figure supplement 1g,h, Figure 1source data 4). We also saw no increase in Ser-1303 phosphorylation in response to oxygen-glucose deprivation (OGD) (Figure 1c,d, Figure 1-source data 5), contrary to previous reports (Tu et al., 2010).
To determine whether Dapk1 plays any role in the GluN2B Ser-1303 phosphorylation status, we obtained a Dapk1 -/mouse line, created by the International Mouse Phenotyping Consortium by targeted deletion of exon four on a C57Bl/6 background (the same strain as the Dapk1 -/mouse generated by Tu et al. (2010)). The mice had normal fertility, viability and body weight (http://www. mousephenotype.org, MGI:1916885). We confirmed that Dapk1 -/neurons expressed no Dapk1 ( Figure 1e). We compared GluN2B phospho-Ser-1303 levels in cortical neurons obtained from Dapk1 -/and Dapk1 +/+ littermates and found no difference in basal levels, nor any difference in the lowered level that we observe at longer periods of NMDA exposure (Figure 1e

Excitotoxic and ischemic neuronal death can proceed independently of Dapk1
We next addressed the more general point of the role of Dapk1 in excitotoxic neuronal death. Compared to cortical neurons cultured from their wild-type littermates, we observed no difference in NMDAR-dependent excitotoxic neuronal death in Dapk1 -/neurons at either DIV10 or DIV16 ( We then studied the influence of Dapk1 deficiency on ischemic neuronal death in vivo. We employed a model of transient global ischemia model (bilateral common carotid artery occlusion) used previously to show a protective effect of Dapk1 deficiency (Tu et al., 2010). Adult mice exposed to a transient (20 min) period of global ischemia showed characteristic selective neuronal death within the hippocampus, particularly the CA1 and CA2 regions. However, Dapk1 -/and Dapk1 +/+ mice exhibited similar levels of infarction ( Figure 2e-h, Figure 2-source data 5), contrary to previous reports (Tu et al., 2010). These observations collectively indicate that excitotoxic and ischemic neuronal death in vitro and in vivo can proceed normally in the absence of Dapk1.

TAT-NR2B CT is a direct NMDAR antagonist
In support of the Dapk1 hypothesis for CTD 2B -derived excitotoxicity, a cell-permeable (TAT-fused) peptide mimetic of the GluN2B amino acids 1292-1304 (TAT-KKNRNKLRRQHSY: TAT-NR2B CT ) was reported to prevent NMDAR-dependent GluN2B Ser-1303 phosphorylation, and excitotoxicity (Tu et al., 2010). We observed that 50 mM TAT-NR2B CT , the concentration used previously (Tu et al., 2010), was toxic to neurons (Figure 3-figure supplement 1a, Figure 3-source data 1), so we used a concentration 10 times lower (5 mM). We found that 5 mM TAT-NR2B CT completely prevents NMDA-induced excitotoxicity (Figure 3-figure supplement 1b, Figure 3-source data 2). This was surprising, given that we did not see a role for Dapk1 in excitotoxicity ( Figure 2). However, further analysis revealed the explanation: at 5 mM, TAT-NR2B CT potently inhibited NMDAR currents, acting immediately and without any need for a preincubation period (Figure 3a-d, Figure 3source data 3), and in a manner that was not readily washed out upon removal of peptide (data not shown). TAT-NR2B CT was custom synthesized for our studies by Genscript, and we found that NR2B CT (1292-1304)-TAT, a pre-made peptide sold by Merck Millipore was a similarly potent NMDAR antagonist (n = 8, Figure  . Taken together these data suggest that the net positive charge of TAT-NR2B CT is primarily responsible for its NMDAR-antagonistic properties. We hypothesized that TAT-NR2B CT may be an open channel blocker drawn partly into the pore by its high net positive charge and bind near the internal Mg 2+ binding site. Consistent with this, the presence of Mg 2+ (a pore blocker) reduced the effectiveness of TAT-NR2B CT 's antagonism (Figure 3e, Figure 3-source data 4). Another prediction of this hypothesis is that TAT-NR2B CT would be more effective at antagonising NMDARs under open-channel conditions. To test this, NMDAR currents were measured, after which neurons were incubated in TAT-NR2B CT (5 mM) for 60 s in the presence of zero glycine +100 mM AP5 ('block (1)") to ensure minimal channel opening (closed channel conditions). 60 s was chosen because under conditions of NMDAR agonism this is sufficient time to achieve maximal blockade. After 60 s, both TAT-NR2B CT and AP5 were removed from the bathing medium and NMDAR currents were subsequently re-measured (150 mM NMDA +100 mM glycine, zero Mg 2+ ). The very slow off-rate of the TAT-NR2B CT Cortical neurons at DIV10 (A) or DIV16 (B) were treated as indicated for 1 hr, with neuronal death assessed at 24 hr. The p values relate to a two-way ANOVA test of differences between WT and Dapk -/neurons (F(1,10) = 0.2676, n = 6 WT, 6 KO (DIV10); F(1,7)=0.8871, 2-way ANOVA, n = 4 WT, 5 KO (DIV16)). For each condition/genotype combination, 800-1000 cells were analysed per biological replicate. (C) OGD-induced neuronal death is independent of Dapk1. Cortical neurons at DIV10 were subjected to OGD for 120 min, before being returned to control medium. Neuronal death was assessed at 24 hr. No genotype-dependent difference was observed (F (1,12)=0.5062, p=0.490, but a strong influence of OGD was observed: F (1,12) =63.54, p<0.0001, two-way ANOVA. #p=0.0002, 0.0002 (reading left to right); Sidak's post-hoc test comparing control to OGD condition (n = 4 WT, n = 4 KO). (D) Dapk1 deficiency does not influence NMDAR currents. NMDAR currents were measured in n = 16 WT cells (from four separate cultures) and n = 25 KO cells (from six separate cultures). Currents were normalized to the mean current recorded from WT cells recorded on that precise day. p=0.411 (t = 0.831, df = 39), unpaired t-test. (E-G) Dapk1 deficiency does not influence vulnerability to ischemia in vivo. Adult male age-matched mice Figure 2 continued on next page enabled the peptide's effects on currents to be measured in the absence of the peptide in the medium. TAT-NR2B CT was then applied for a second 60 s ('block (2)') either under the same 'closed channel conditions' or under 'open channel conditions' (100 mM glycine, 150 mM NMDA). After 60 s TAT-NR2B CT was removed from the bathing medium and NMDAR currents measured for a third time. The NMDAR current remaining at the second and third measurements was calculated as a fraction of the initial current. We found that the second peptide incubation (block (2)

Discussion
Dapk1-mediated GluN2B Ser-1303 phosphorylation, and consequent enhancement of toxic Ca 2+ influx through extrasynaptic NMDARs lies at the heart of current models of excitotoxicity and of the central role of the GluN2B CTD in this process (Parsons and Raymond, 2014;Lai et al., 2014), but our study suggests that this needs to be re-appraised. Our observations regarding the (lack of) impact of Dapk1 gene deletion on neuronal vulnerability to excitotoxic and ischemic conditions is at odds with previous reports (Tu et al., 2010). The Dapk1 -/mouse that we used was generated independently of the one generated by Tu et al., although there is no a priori reason why the two lines should behave differently at this fundamental level, particularly given the very similar genetic background (C57BL/6).
The potent inhibition of NMDAR currents by TAT-NR2B CT at a concentration up to 100 times lower than that used previously (Tu et al., 2010) suggests a simple explanation for its neuroprotective effects independent of Dapk1. We are unable to explain why we observed similar effects of TAT-NR2B CT and its scrambled version, while a selective effect of TAT-NR2B CT was previously reported (Tu et al., 2010). Both scrambled versions employed had identical sequences, and the potent NMDAR antagonistic properties of our scrambled peptide are consistent with its neuroprotective properties.
The basis for CTD 2B -mediated excitotoxicity  remains incompletely understood. Exchanging the CTD of GluN2B with that of GluN2A by targeted exon exchange reduces vulnerability to excitotoxicity , without altering the proportion of NMDARs at synaptic vs. extrasynaptic sites, an important factor in excitotoxicity (Hardingham and Bading, 2010). Moreover, performing the reciprocal swap increases vulnerability (SM and GEH, unpublished observations), strongly supportive of a key role for CTD 2B . An ongoing avenue of investigation is focussed on understanding the extent to which the composition of the native NMDAR signaling complex is altered by manipulating the endogenous GluN2 CTDs in our panel of knock-in mice. We hypothesize that alterations to the complex may disturb signaling to pro-death events such as NO production, NADPH oxidase activation, oxidative stress, calpain activation and mitochondrial Ca 2+  immediately antagonize NMDAR currents upon extracellular exposure. NMDA-induced currents were recorded under whole-cell voltage clamp, with the indicated peptides (at 5 mM) applied approximately 5 s after NMDA (to allow NMDAR currents to reach steady state). Arg-rich refers to the arginine-rich positively charged peptide; Neutral refers to the neutral peptide-see main text for sequences of these as well as TAT-NR2B CT and TAT-sNR2B CT . NMDA-induced NMDAR currents were monitored for a further 45 s and the percentage drop in currents calculated, compared to no peptide at all (Con) which represents a measure of natural desensitization over this period. p<0.0001 (one-way ANOVA). *p<0.0001, Sidak's post-hoc test (n = 8 of all conditions). Example traces shown in (B) (C) and (D). Scale bar: 15 s, 500 pA. (E) NMDAR antagonism by TAT-NR2B CT is inhibited by Mg 2+ blockade. NMDAR currents were measured, after which neurons were incubated in TAT-NR2B CT (5 or 0.5 mM) for 60 s in the presence or absence of 1 mM Mg 2+ , after which NMDAR currents were measured again (in zero Mg 2+ , no peptide). p<0.0001 (one-way ANOVA, F(1,22)=47.16 (effect of [Mg 2+ ])). *p<0.0001, p=0.0007, Sidak's post-hoc test (zero Mg 2+ : n = 6 (0.5 mM), n = 7 (5 mM); 1 mM Mg 2+ : n = 6 (0.5 mM), n = 7 (5 mM)).  (Parsons and Raymond, 2014;Lai et al., 2014;Hardingham and Bading, 2010;Bano and Nicotera, 2007;Duchen, 2012;Nakamura and Lipton, 2011;Panayotis et al., 2015). Of note, we recently showed that the CTD of GluN2B (as opposed to that of GluN2A) is critically required for formation of 1.5 MDa NMDAR supercomplexes (Frank et al., 2016). Thus, regions unique to GluN2B (of which the CaMKII site is one) play a role in higher order signal complex assembly and this may underlie the key role of CTD 2B in downstream excitotoxicity .

Materials and methods
Neuronal culture, Dapk -/mice, induction of excitotoxicity and oxygenglucose deprivation Cortical mouse neurons were cultured as described (Bell et al., 2011a) at a density of between 9-13 Â 10 4 neurons per cm 2 (Lipton and Rosenberg, 1994) from E17.5 mice with Neurobasal growth medium supplemented with B27 (Invitrogen, Paisley, UK). Stimulations of cultured neurons were done in most cases after a culturing period of 9-11 days during which neurons develop a network of processes, express functional NMDA-type and AMPA/kainate-type glutamate receptors, and form synaptic contacts. Other experiments were performed at DIV 16. Dapk -/mice (colony name: H-Dapk1-B11-TM1B, MGI Allele Name: Dapk1tm1b(EUCOMM)Hmgu, RID:MGI:5756958) were generated by MRC Harwell from targeted ES cells made by The European Conditional Mouse Mutagenesis Program, as part of the International Mouse Phenotyping Program. Dapk -/genotyping reactions were performed using the following primers: A = 5 AGAGAAACTGAGGCACCTGG À3', B =, 5'-CATCCAAAGTCCACAGCCAC-3', C = 5'-CCAGTTGGTCTGGTGTCA-3' Primer pair A-B recognised the wild-type allele and amplified a product of 322 bp. Primer pair B-C recognised the mutant allele corresponding to a product of 468 bp. PCR reactions were performed using the following cycling conditions: 15 min at 95˚C; 36 cycles of 45 s at 94˚C, 45 s at 60˚C and 1 min at 72˚C; and 10 min at 72˚C.
To apply an excitotoxic insult, neurons were first placed overnight into a minimal defined medium (Baxter et al., 2011) containing 10% MEM (Invitrogen), 90% Salt-Glucose-Glycine (SGG) medium ( [Bading et al., 1993]; SGG: 114 mM NaCl, 0.219% NaHCO 3 , 5.292 mM KCl, 1 mM MgCl 2 , 2 mM CaCl 2 , 10 mM HEPES, 1 mM Glycine, 30 mM Glucose, 0.5 mM sodium pyruvate, 0.1% Phenol Red; osmolarity 325 mosm/l, [Papadia et al., 2005]). Where used, TAT-NR2B CT or TAT-sNR2B CT (5 mM) was incubated for 1 hr prior to the excitotoxic insult. Neurons were then treated with NMDA (Tocris Bioscience, Bristol, UK) at the indicated concentrations for 1 hr, after which medium was changed to NMDA-free. After a further 23 hr, neurons were fixed and subjected to DAPI staining and cell death quantified by counting (blind) the number of shrunken, pyknotic nuclei as a percentage of the total. To induce oxygen-glucose deprivation, a previously described approach was used (Bell et al., 2011b(Bell et al., , 2015. Briefly, cells were washed and incubated in glucose-free SGG (see formulation above, but with glucose replaced by mannitol) that had been previously degassed with 95% N 2 -5% C0 2 . The cells were then placed in an anoxic modular incubator chamber for 120 min (as compared to cells washed and incubated in normoxic glucose-containing SGG). For analysis of excitotoxicity, approximately 800-1000 cells were analysed per condition, per replicate (repeated across several replicates), the observer blind to genotype and experimental condition.

Electrophysiological recording and analysis
Coverslips containing cortical neurons were transferred to a recording chamber perfused (at a flow rate of 3-5 ml/min) with an external recording solution composed of (in mM): 150 NaCl, 2.8 KCl, 10 HEPES, 2 CaCl 2 , 1 MgCl 2 , 10 glucose and 0.1 glycine, pH 7.3 (320-330 mOsm). Patch-pipettes were made from thick-walled borosilicate glass (Harvard Apparatus, Kent, UK) and filled with a K-gluconate-based internal solution containing (in mM): potassium gluconate 141, NaCl 2.5, HEPES 10, EGTA 11; pH 7.3 with KOH). Electrode tips were fire-polished for a final resistance ranging between 4-8 MW. Currents were recorded at room temperature (21 ± 2˚C) using an Axopatch 200B amplifier (Molecular Devices, Union City, CA). Neurons were voltage-clamped at -60 mV, and recordings were rejected if the holding current was greater than -100 pA or if the series resistance drifted by more than 20% of its initial value (<25 MW). All NMDA currents were evoked in Mg 2+ -free external recording solution (in which MgCl 2 was substituted with 2 mM NaCl) by 150 mM NMDA +100 mM glycine except Figure 2D where 50 mM NMDA +100 mM glycine was used. Whole-cell currents were analyzed using WinEDR v3.2 software (John Dempster, University of Strathclyde, UK). The approximate number of cells to be recorded was estimated in order to detect a 25% difference in the parameter under study, powered at 80%, based on the standard deviation of data previously published by the laboratory Puddifoot et al., 2012;Hardingham et al., 2007).
To determine the the competing effect of Mg 2+ and TAT-NR2B CT , whole cell NMDA currents were recorded (as described above) followed by the inclusion of TAT-NR2B CT ± Mg 2+ in the recording solution for a blocking period of 60 s. The whole cell NMDA current was then re-assessed, and the percentage block calculated. TAT-NR2B CT was not included when NMDA currents were reassessed; this may have led to a small washout but we deemed this as negligible due to the slow-off rate of TAT-NR2B CT . To investigate the use dependency of TAT-NR2B CT , we minimized the possibility of the NMDAR channel opening by spontaneous release of glutamate by removing glycine from the ACSF and co-applying 100 mM AP5. Glycine was added back to the ACSF to measure NMDA currents and to facilitate the block of TAT-NR2B CT in the open channel configuration.

Western blotting
Western blotting was performed as described (Baxter et al., 2015). In order to minimize the chance of post-translational modifications during the harvesting process, neurons were lysed immediately after stimulation in 1.5x LDS sample buffer (NuPage, Life Technologies) and boiled at 100˚C for 10 min. Approximately 10 mg of protein was loaded onto a precast gradient gel (4-16%) and subjected to electrophoresis. Western blotting onto a PVDF membrane was then performed using the Xcell Surelock system (Invitrogen) according to the manufacturer's instructions. Following the protein transfer, the PVDF membranes were blocked for 1 hr at room temperature with 5% (w/v) non-fat dried milk in TBS with 0.1% Tween 20. The sample size was calculated based on previous experimental observations of reporting the effect and standard deviation of NMDA-induced Ser-1303 phosphorylation (Tu et al., 2010). The membranes were incubated at 4˚C overnight with the primary antibodies diluted in blocking solution: Anti phospho-(Ser-1303) GluN2B (1: 2000, Millipore), anti-Dapk1 (1:8000, Sigma), anti-GluN2B (C-terminus, 1:8000, BD Transduction Laboratories), anti-beta actin (1:200000, Abcam). For visualisation of Western blots, HRP-based secondary antibodies were used followed by chemiluminescent detection on Kodak X-Omat film. Western blots were digitally scanned and densitometric analysis was performed using Image J. All analysis of GluN2B phosphorylation was normalized to total GluN2B.

Bilateral common carotid artery occlusion
Mice were housed in individually-ventilated cages (in groups of up to five mice) under specific pathogen-free conditions and standard 12 hr light/dark cycle with unrestricted access to food and water. All experiments using live animals were conducted under the authority of UK Home Office project and personal licences and adhered to regulations specified in the Animals (Scientific Procedures) Act (1986) and Directive 2010/63/EU and were approved by both The Roslin Institute's and the University of Edinburgh's Animal Welfare and Ethics Committees. Experimental design, analysis and reporting followed the ARRIVE guidelines (https://www.nc3rs.org.uk/arrive-guidelines) where possible. The sample size was calculated based on the experimental observations of reporting the effect and standard deviation of BCCAO-induced neuronal loss in both wild-type and Dapk1 -/- (Tu et al., 2010), whose experimental observations using n = 7 per genotype we retrospectively calculated were powered at >99%. Transient bilateral common carotid artery occlusion (BCCAO) was performed in Dapk1 -/and wild-type male control mice under isoflurane anaesthesia (with O 2 and N 2 O). The operator was unaware of genotype. Core body temperature was maintained at 37 ± 0.5˚C throughout the procedure with a feedback controlled heating blanket (Harvard Apparatus, UK). Both common carotid arteries were exposed and dissected from surrounding tissues and occluded by application of an aneurysm clip for 20 min. Clips were removed, the neck wound sutured and topical local anaesthetic (lidocaine/prilocaine) was applied. Mice were recovered on a heated blanket for 4-6 hr and then returned to normal housing. After a 3 day recovery, mice were anaesthetised and perfused transcardially with saline followed by 4% paraformaldehyde. Brains were removed and rostral and caudal blocks prepared using a brain matrix (Harvard Apparatus). Blocks were post-fixed in 4% paraformaldehyde for 24 hr and processed to paraffin blocks. Sections (6 mm) were cut on a microtome (Leica) and stained with haematoxylin and eosin. Ischaemic neuronal death was quantified in the CA1 and CA2 regions of the hippocampus which are the most sensitive regions in this model. Ischaemic (dead) neurons were identified morphologically in two regions of interest (ROIs) in CA1 and the entire CA2 bilaterally. Data are expressed as the number of dead neurons as a % of total neurons in the ROI and show the mean of both hemispheres for each region. All processing and analysis was performed with the operator blind to genotype.

Statistical analysis, equipment and settings
Statistical testing involved a 2-tailed paired Student's t-test, or a one-or two-way ANOVA followed by an appropriate post-hoc test, as indicated in the legends. Cell death analyses for both in vitro and in vivo experiments were performed blind to the genotype/experimental condition. For all cell death, western blot analyses and in vitro and in vivo cell death experiments, the value of 'N' was taken as the number of independent biological replicates, defined as independently performed experiments on material derived from different animals. For western blots, we used chemiluminescent detection on Kodak X-Omat film, and linear adjustment of brightness/contrast applied (Photoshop) equally across the image, maintaining some background intensity. In any cases where lanes from non-adjacent lanes are spliced together,lanes are always from the same blot, processed in the same way, and the splicing point is clearly marked. Pictures of cells were taken on a Leica AF6000 LX imaging system, with a DFC350 FX digital camera.