|Home | About | Journals | Submit | Contact Us | Français|
Expression of the NMDA receptor (NMDAR) subunit NR3A reaches its highest level in layer V of the developing rodent cortex during the second postnatal week, a peak period of synaptogenesis. Incorporation of NR3A leads to the formation of non-canonical, Mg2+ insensitive NMDARs; but it is not known whether they participate in synaptic transmission and maturation. Here we show that in the second postnatal week, layer V pyramidal neurons in the somatosensory cortex of wild type (WT) mice exhibited evoked excitatory postsynaptic currents (eEPSCs) with 3 to 6 fold lower Mg2+ sensitivity than NR3A knockout (KO) mice and their reversal potential was ~2 mV more negative compared to KO mice consistent with decreased PCa of NMDARs. Surprisingly, ablation of NR3A also led to a 20 fold reduction of the ratio of AMPAR- to NMDAR-mediated eEPSC amplitudes in KO mice. Insertion of AMPARs at the synapses of layer V pyramidal neurons appears to be facilitated by the expression of Mg2+ insensitive NMDARs. The data indicate that NR3A plays a significant role in the development of excitatory synapses in layer V of the developing neocortex.
N-methyl-D-aspartate receptors (NMDARs) belong to the family of ionotropic glutamate receptors and are critical for normal and abnormal development and function of the brain (Dingledine et al., 1999). NMDARs play important roles in neuronal differentiation, migration, and synapse formation in the developing brain and continue to mediate neuronal signalling and plasticity in the adult central nervous system. Pathological NMDAR activity contributes to the pathogenesis of neurological and psychiatric diseases (Rakhade and Jensen, 2009). NMDARs are tetrameric complexes composed of two NR1 and two identical or different NR2 subunits (Cull-Candy and Leszkiewicz, 2004). These “canonical” NMDARs are activated upon concomitant binding of the agonist glutamate and the co-agonist glycine, and are characterized by their high permeability for Ca2+ ions (PCa) and their voltage-dependent block by physiological concentrations of Mg2+. Glutamate binds to NR2 subunits while glycine or D-serine, a second naturally occurring co-agonist, bind to NR1 subunits (Wolosker, 2007). The exquisite Mg2+ sensitivity and the high PCa distinguish NMDARs from other “non-NMDA” type ionotropic glutamate receptors, such as the α-amino-3-hydroxy-5-methylisoxazole-4-propionate receptors (AMPARs) and underpin their pleiotropic roles in brain development and function. However, recent results indicate that two additional subunits, NR3A (Ciabarra et al., 1995; Das et al., 1998; Sucher et al., 1995; Wong et al., 2002) and NR3B (Chatterton et al., 2002; Matsuda et al., 2002; Nishi et al., 2001), with equal homology to NR1 and NR2 subunits, can combine with NR1 to form Mg2+ insensitive excitatory glycine receptors (Chatterton et al., 2002; Smothers and Woodward, 2007) or with both NR1 and NR2 to form non-canonical, largely Mg2+ insensitive NMDARs with reduced PCa (Das et al., 1998; Matsuda et al., 2003; Pérez-Otaño et al., 2001; Sasaki et al., 2002). While evidence for excitatory glycine receptors in the brain is lacking to date (Chatterton et al., 2002; Matsuda et al., 2003), developmental and regional differences (Binshtok et al., 2006; Bowe and Nadler, 1990; Chahal et al., 1998; Gonzales, 1992; Kato and Yoshimura, 1993; Kirson et al., 1999; Morrisett et al., 1990; Sutor et al., 1987) as well as injury and disease related changes in the Mg2+ sensitivity of NMDAR-mediated currents have been reported (Andre et al., 2004; Furukawa et al., 2000; Starling et al., 2005). The molecular basis underpinning these differences remains to be established in most cases with some notable exceptions (Binshtok et al., 2006; Momiyama et al., 1996; Takahashi et al., 1996), where reduced Mg2+ sensitivity of NMDARs has been linked to expression of NR2C and/or NR2D subunits (Monyer et al., 1994). Recently, Mg2+-insensitive NMDAR-mediated currents and expression of NR3 have been described in oligodendrocytes (Karadottir et al., 2005; Manning et al., 2008; Verkhratsky and Kirchhoff, 2007) and NMDAR-gated channels with reduced Mg2+ sensitivity have been observed in cortical neurons of WT but not NR3A KO mice (Sasaki et al., 2002). Consistent with these findings and as predicted from co-expression studies of NR3A with NR1 and NR2 in human embryonic kidney (HEK) 293 cells (Sasaki et al., 2002), NMDAR-mediated whole-cell currents in hippocampal neurons of NR3A transgenic mice also exhibited reduced Mg2+ sensitivity (Tong et al., 2008). The observation that NR3A KO mice showed increased dendritic sprouting (Das et al., 1998) suggested that Mg2+ insensitive NMDARs might play a role in synaptogenesis and/or synaptic maturation (Sucher et al., 1995; Wong et al., 2002) consistent with the reported role of NMDARs but not AMPARs in the suppression of neurite elongation and sprouting in cultured neurons in vitro (Lin and Constantine-Paton, 1998). However, it is not known to date whether or not these non-canonical NMDARs actually participate in synaptic transmission. We undertook the present study to investigate this question.
We previously found that NR3A protein expression reached its highest expression levels during the second postnatal week but decreased gradually thereafter to low levels in adulthood (Wong et al., 2002). Consistent with these findings, Sasaki et al (2002) found that NMDA receptor mediated currents in cortical neurons from WT mice continuously increased between P5 and P16, while in NR3A knock-out mice, the NMDA current density reached levels comparable to adult levels after P8. Together, these data indicated that the maximum level of expression of functional NMDA receptors containing NR3A was likely to be observed after P8 but before P13, coincident with the period of peak synaptogenesis in the cortex. Hence, we chose the time window from P9 to P12 for our studies. We chose to examine NMDAR-mediated eEPSCs in layer V neocortical pyramidal neurons because NR3A but not NR3B mRNA and protein are highly expressed (Sasaki et al., 2002; Wong et al., 2002) in these cells during this time window. We recorded NMDAR-mediated spontaneous (s) and eEPSCs in the presence of physiological extracellular Mg2+ and fitted the data with the Woodhull model to quantitatively compare the Mg2+ sensitivity in WT and KO mice (Mayer et al., 1984; Nowak et al., 1984; Wollmuth et al., 1998).
In this study, we examined NMDAR-mediated eEPSCs in layer V pyramidal neurons in the somatosensory cortex of 9 to 13 day old WT (n = 32) and NR3A KO mice (n = 21), a developmental period of elevated cortical NR3A protein expression (Wong et al., 2002). We pharmacologically isolated and evoked NMDAR-mediated EPSCs by intracortical electrical stimulation applied via a bipolar electrode placed at the interface between layers I and II (Kumar and Huguenard, 2001, 2003). We chose intracortical stimulation because NMDAR-mediated EPSCs of layer V pyramidal neurons evoked by this route were shown previously to be less Mg2+ sensitive than those evoked by callosal or white matter stimulation (Kumar and Huguenard, 2001, 2003). In most WT neurons, eEPSCs consisted of both an early component (Fig. 1, top left panel, black arrow) and a late component (Fig. 1, top left panel, grey arrow, see also supplementary data, Fig. 1) (Agmon et al., 1996). Both WT and KO neurons exhibited eEPSCs (Fig. 1, bottom left panel, black arrow) with an early component mediated by monosynaptic inputs with similar synaptic latency (supplementary data, Fig. 2) as observed previously at this postnatal age (Salami et al., 2003) but only WT neurons exhibited a late component, which was reported previously to originate from intracortical polysynaptic responses (Agmon et al., 1996; Luhmann and Prince, 1990). The weighted mean decay time constants (τw) of the NMDAR-mediated eEPSCs exhibiting only the early component in both WT (183. 4 ± 23.5, n = 5, p < 0.05; Fig. 2A) and KO mice (149.9 ± 26.3 ms, n = 9, p < 0.05; Fig. 2A) were significantly faster than the NMDAR-mediated eEPSCs with late component in WT animals (287.1 ± 22.3 ms, n = 15; Fig. 2B). All currents were completely blocked by addition of the competitive NMDAR antagonist DL-APV (100 µM; not illustrated).
The early component of the eEPSCs in WT and KO mice and the late component of the eEPSCs in WT exhibited a non-linear current (I)/voltage (V) relationship with a negative slope at holding potentials (Vh) between approximately −60 and −20 mV, which is characteristic of the voltage-dependent blockade of NMDARs by extracellular Mg2+ ions (Mayer et al., 1984; Nowak et al., 1984; Wollmuth et al., 1998). The shape of this I/V is well described by the Woodhull model as demonstrated by the excellent correspondence between experimental and model-derived values of the Mg2+ dissociation constants (Kd) (Wollmuth et al., 1998). To quantitate the differences in Mg2+sensitivity of NMDAR-mediated eEPSCs in WT and NR3A KO mice (Kirson and Yaari, 1996), we fitted this model to the normalized average I/V curves of WT (n = 20) and KO (n = 9) neurons for which we obtained data over the entire voltage range at physiological Mg2+ concentration (Fig. 3). In a second approach to assess Mg2+ sensitivity, we fitted each cell’s I/V data separately with the Woodhull model and subjected the parameters to statistical analysis comparing WT and KO (Table 1). Both approaches yielded parameters indicating decreased Mg2+ sensitivity in WT compared to KO neurons. The Kd of Mg2+ at 0 mV of the early component of the eEPSCs was significantly higher in WT (17.73 ± 5.38 mM) than in KO animals (5.34 ± 2.13 mM; p < 0.05). While neither the maximal conductance (gmax) nor the voltage dependence (δ) differed significantly between WT (gmax = 7.78 ± 0.76 nS, δ = 0.95 ± 0.03) and KO (gmax = 5.02 ± 1.12 nS, = 0.75 ± 0.10), the reversal potential (Vrev), was significantly smaller in WT mice (1.69 ± 0.62 mV) than in KO (3.60 ± 0.74 mV, p < 0.05). Interestingly, the observed negative shift of Vrev in WT is predicted by the decreased Ca2+ permeability that has been previously observed in HEK293 cells expressing triheteromeric NMDARs containing the NR3 subunit compared with diheteromeric channels (Pérez-Otaño et al., 2001; Sasaki et al., 2002).
While most WT and KO neurons exhibited eEPSCs as illustrated in Fig. 1, there were notable exceptions (Fig. 4). We found that two WT neurons exhibited significant inward currents even at very hyperpolarized Vh (Fig. 4). It was unlikely that space clamp errors or dendritic filtering contaminated these recordings (Fig. 4, insert in the left panel), as there was no significant correlation between the decay time constant and the peak amplitude (Kirson and Yaari, 1996). These observations raise the question as to the molecular basis underpinning this unusual I/V relationship without negative slope conductance despite the presence of extracellular Mg2+. Along these lines, we asked whether these data were consistent with a model in which the average decrease in the Mg2+ sensitivity could be explained by the presence of parallel, Mg2+ sensitive and Mg2+ insensitive conductances resulting from the simultaneous expression of canonical, Mg2+ sensitive and non-canonical, Mg2+ insensitive NMDARs. Consistent with this idea, the data could be fitted equally well by the sum of a linear, Mg2+ insensitive conductance as observed previously upon expression of NR1, NR3A and NR3B in HEK293 cells (Smothers and Woodward, 2007) and a Mg2+ sensitive conductance described by the Woodhull model (Fig. 4B). Interestingly, this parallel-conductance model also fitted the I/V curves of NMDAR-mediated whole-cell currents previously recorded in the presence of 1 and 10 mM extracellular Mg2+ in HEK293 cells that had been transfected with NR1, NR2B and NR3A (Sasaki et al., 2002). Specifically, the model indicates that the fraction of Mg2+ insensitive NMDARs in those cells was as high as 50% (not illustrated), which underpins the clearly noticeable outward rectification of the I/V curves observed in those experiments (Sasaki et al., 2002).
Canonical NMDARs exhibit a larger single channel conductance than non-canonical, NR3A containing NMDARs (Pérez-Otaño et al., 2001; Sasaki et al., 2002) and NMDAR-mediated whole cell currents in cortical neurons from NR3A KO mice were increased (Das et al., 1998; Sasaki et al., 2002). Consistent with these previously reported data, we found that the average amplitude of the NMDAR-mediated sEPSCs was smaller in WT (35.71 ± 0.83 pA) than in NR3A KO neurons (40.74 ± 1.28 pA; Fig 5, A and B). Although this observation is consistent with the notion that expression of the NR3A subunit decreased the NMDAR-mediated currents in WT, it is possible that a difference in action potential firing between WT and NR3A KO neurons may have contributed to an increase in the amplitudes of NMDAR-mediated sEPSCs. Therefore, we also performed non-stationary fluctuation analysis (Traynelis et al., 1993) of sEPSC events to derive computationally the average number and single-channel conductance of NMDARs mediating sEPSCs in WT and KO mice (supplementary data, Fig. 3). The data revealed a mean single-channel conductance of 53.4 pS in WT and 79.6 pS KO animals (Supplementary data, Fig 3.), indicating that expression of NR3A subunit in WT mice decreased NMDAR single-channel conductance. At the same time, WT neurons exhibited ~ 33% more NMDAR channels (n = 32, supplementary data, Fig. 3.) at their synaptic sites than NR3A KO mice (n = 24, supplementary data, Fig. 3.).
Multiple lines of evidence indicate that synaptic NMDARs are expressed earlier than AMPARs during development (Petralia et al., 1999; Wu et al., 1996) and mediate not only sprouting of neurites in the absence of AMPARs but the activity-dependent insertion of AMPARs into these synapses, which are therefore “silent” at negative membrane potentials (Liao et al., 1995; Shi et al., 1999). Notably, this is precisely the period during which NR3A protein is expressed at high levels during development (Wong et al., 2002). We wondered whether the decreased Mg2+ sensitivity of NMDARs might be accompanied by increased AMPAR-mediated responses as previously reported in retinal ganglion cells of WT compared to NR3A KO mice (Sucher et al., 2003). Therefore, we compared AMPAR and NMDAR-mediated eEPSCs close to the cells’ resting potential at −60 mV in WT and KO mice (Fig. 6A), which is more relevant to the physiological activity of NR3A containing NMDARs than at positive holding potentials, where both NR3-containing and NR3-lacking NMDARs are Mg2+ insensitive. The AMPAR-mediated eEPSCs in both WT and KO animals exhibited only one component. The average amplitude of the AMPAR-mediated eEPSCs was significantly larger in WT compared to KO mice (123.5 ± 53.9 vs. 16.6 ± 16.6 pA, n = 5) and the ratio of AMPAR- to NMDAR-mediated eEPSC amplitudes was more than 20 fold higher in WT than in KO neurons (Fig. 6B).
We undertook the present study in order to test the hypothesis that NR3A containing NMDARs with greatly reduced Mg2+ sensitivity participate in synaptic transmission during early postnatal cortical development. To this end, we examined the Mg2+ sensitivity of synaptic NMDARs in neocortical pyramidal neurons in layer V of WT and NR3A KO mice. We hypothesized that the Mg2+ sensitivity should be reduced in WT mice in the second postnatal week when these cells express high levels of NR3 protein. Consistent with this hypothesis, whole cell patch-clamp recordings demonstrated reduced Mg2+ sensitivity of NMDARs in WT compared to KO mice. Furthermore, we found that the Mg2+ sensitivity of the early component of the eEPSCs was higher than that of the late component, the latter of which appeared to be absent in KO animals. The early component of the eEPSCs represents NMDAR-mediated monosynaptic inputs of layer V pyramidal neurons in both WT and KO (Agmon et al., 1996). The origin of the late component of the eEPSCs is less clear and might result from stimulation of synapses mediating polysynaptic inputs (Agmon et al., 1996) or be due to activation of extrasynaptic NMDARs (Fellin et al., 2004). The apparent absence of the late component in KO animals suggests that NR3A plays a role in the maturation of synapses in layer V pyramidal neurons at this developmental stage.
The Mg2+ sensitivity of the NMDAR-mediated eEPSCs in KO mice reported here was similar to that previously observed in CA1 pyramidal cells in the developing mouse hippocampus (Kirson and Yaari, 1996) and recombinant receptors composed of NR1 and NR2A or NR2B subunits (Monyer et al., 1994). The decay kinetics of the early component of the eEPSCs in layer V pyramidal neurons in both WT and KO mice were comparable to those in recombinant NMDARs composed of NR1 and NR2 subunits (Vicini et al., 1998). However, both NMDAR-mediated eEPSC components in WT neurons exhibited a significantly increased Kd, while Vrev was shifted to more negative values. In addition, non-stationary fluctuation analysis revealed that the average singe-channel conductance in WT animals was smaller than that in KO, which was similar to the maximum single channel conductance in the spinal cord at similar developmental age (Green and Gibb, 2001; Palecek et al., 1999). These findings are consistent with a model suggesting that layer V pyramidal neurons in WT mice at this developmental age express both canonical NMDARs with high single channel conductance and non-canonical, NR3A-containing NMDARs with lower single channel conductance (Pérez-Otaño et al., 2001; Sasaki et al., 2002). The data suggest that on average a relatively small number of channels were activated at WT synapses under the conditions of our experiments. The mean gmax values measured in our experiments indicate that ~70 canonical, diheteromeric NMDAR channels contributed to the early component of the eEPSCs in both WT and KO mice, while WT mice expressed in addition ~50 non-canonical, NR3A-containing triheteromeric channels with half the single channel conductance of diheteromeric channels (Pérez-Otaño et al., 2001; Sasaki et al., 2002). This model is supported by the data derived from the non-stationary fluctuation analysis of sEPSCs showing that WT neurons expressed ~ 33% more NMDAR channels than NR3A KO neurons (supplementary data, Fig. 3). The data are consistent with earlier findings indicating that the genetic ablation of NR3A did not lead to a compensatory increase in the expression of the NR3B subunit in these mice (Chatterton et al., 2002; Wong et al., 2002).
EPSCs with similarly decreased Mg2+ sensitivity and fast decay kinetics have been observed previously at cerebellar granule cell synapses in the developing cerebellum of rats. Interestingly, the much slower decay kinetics characteristic of NR2C containing NMDARs were not observed in these neurons until the sixth postnatal week (Cathala et al., 2000). While this observation was interpreted to suggest gradual incorporation of the NR2C subunit (Cathala et al., 2000), it is interesting to note that cerebellar granule cells were found to express high levels of the NR3A subunit during this developmental period (Wong et al., 2002). Thus, together with our data, the results reported by Cathala and colleagues (Cathala et al., 2000) are equally consistent with the notion that cerebellar granule cell synapses expressed non-canonical triheteromeric NMDARs containing the NR3A subunit in addition to or preceding a switch to diheteromeric NMDARs containing the NR2C subunit.
eEPSCs in WT but not KO mice exhibited a late component with further reduced Mg2+ sensitivity. A similar late component of eEPSCs has been seen in previous work where it was thought to either result from stimulation of synapses mediating polysynaptic inputs (Agmon et al., 1996; Luhmann and Prince, 1990) or be due to activation of extrasynaptic NMDARs (Fellin et al., 2004). The apparent absence of the late component in KO animals suggests that deletion of NR3A changes the functional organization, of synapses in layer V pyramidal neurons at this developmental stage. The observed decrease of functional AMPARs in the KO mice is intriguing but it remains to be established whether it is a direct consequence of diminished excitatory drive due to the absence of NR3A containing Mg2+ insensitive NMDARs in KO mice or an indirect effect linked to changes in NR3A associated proteins, for example (Chan and Sucher, 2001; Pérez-Otaño et al., 2006).
The second postnatal week is characterized by rapid synaptogenesis and intense synaptic plasticity (Perkins and Teyler, 1988). At the same time, pyramidal neurons of layer V undergo rapid development (Zhang, 2004) and start to differentiate at the end of the first postnatal week into two major classes (Bannister, 2005; Kasper et al., 1994; Molnar and Cheung, 2006). As it has been reported that application of NMDAR antagonists delayed the differentiation of the two classes (Molnar and Cheung, 2006), it is possible that Mg2+-insensitive NMDARs might play a role in this process in addition to their role in shaping synaptic connections, although the neurons recorded in the present study in both WT and KO animals exhibited morphological signs of type 1 neurons (Bannister, 2005; Molnar and Cheung, 2006).
Layer V pyramidal neurons in the prefrontal cortex of patients with schizophrenia were reported to have smaller dendritic field sizes than non-schizophrenic controls. NR3A mRNA was reported to be selectively increased in dorsolateral prefrontal cortex in schizophrenics (Mueller and Meador-Woodruff, 2004) and genetic variations of the NR3A subunit have been linked to the modulation of prefrontal cerebral activity in healthy humans (Gallinat et al., 2007). Together, these data suggest that NR3 containing NMDARs may play a critical role in the development of a subset of cortical neurons and their synaptic connections with consequences that may persist into adult life. As low levels of NR3A are detectable in the adult brain (Wong et al., 2002), however, it is possible that this subunit may also play a role during synaptogenesis and/or synaptic remodelling later in life.
Experiments were performed on coronal cortical slices (300 µm thickness) from the somatosensory cortex obtained from 9 to 13 day old C57Bl/6 wild type and NR3A knockout mice (Das et al., 1998). Mouse pup brains were rapidly dissected from the skull and placed in ice-cold cutting solution (containing: 210 mM sucrose, 2.5 mM KCl, 1.02 mM NaH2PO4, 0.5 mM CaCl2, 10 mM MgSO4, 26.19 mM NaHCO3, and 10 mM D-glucose, pH 7.4) bubbled with 95% O2/5% CO2 at 4°C. Slices were prepared with a vibratome in cold cutting solution, incubated in a chamber at 32°C for 30 min with continuously oxygenated artificial cerebrospinal fluid (ACSF) containing the following (in mM): 124 NaCl, 5 KCl, 1.25 NaH2PO4, 2 CaCl2, 1.2 MgSO4, 26 NaHCO3, 10 D-glucose, pH 7.4. Slices were kept in oxygenated ACSF at room temperature for at least 1 hour before electrophysiological recordings.
Whole-cell patch clamp recordings were made from pyramidal neurons in layer V of the agranular somatosensory cortex (Kasper et al., 1994; Voelker et al., 2004). Using infrared/differential interference contrast microscopy, neurons were visually identified by their large soma, large dendritic trunk and in some experiments by the observation of their characteristic tufted apical dendrites ending in layer I in biocytin-filled neurons (not illustrated; (Bannister, 2005; Kasper et al., 1994)). The brain slices were continuously superfused with ACSF bubbled with 95% O2 and 5% CO2 at 1–1.5 ml/min. The internal solution of the patch pipette consisted of the following (in mM): 110 Cs-sulfonate, 10 TEA-Cl, 5 QX-314, 4 NaCl, 2 MgCl2, 10 EGTA, 10 HEPES, 7 phosphocreatine, creatine phosphokinase (17 U/ml), 4 ATP-Mg, and 0.3 GTP, pH 7.25. Filled electrodes had resistances of 4–6 MΩ. The liquid junction potential with these solutions was +8.7 mV. All data shown have been corrected for this value. The seal resistance was ≥ 700 MΩ and the series resistance, which was monitored using hyperpolarizing voltage steps (5 mV) during the recordings, was typically less than 20 MΩ. Experiments were terminated if the series resistance changed by more than 10%. All recordings were performed at room temperature (22–24°C).
Inhibitory GABA (γ-aminobutyric acid) and glycine receptors were blocked with picrotoxin (60 µM) and strychnine (2 µM), respectively. Excitatory AMPA/kainate receptors were blocked with 2,3-dihydroxy-6-nitro-7-sulfamoyl-benzo[f]quinoxaline-2,3-dione (NBQX, 10 or 20 µM). The identity of NMDAR-mediated eEPSCs was pharmacologically confirmed by their sensitivity to the specific NMDAR-antagonist DL-(±)-2-amino-5-phosponovalerate (DL-APV, 100 µM). Drugs were added to the ACSF supplemented with NMDAR co-agonist glycine (10 µM) and applied through a gravity superfusion system. Nifedipine (5 µM added to the external solution) or (±)-methoxyverapamil-hydrochloride (D600; 0.1 mM added to the patch-pipette internal solution) was used to block L-type calcium channels. To evoke EPSCs, cells were clamped at voltages from −80 to +50 mV in 10 mV increments. A bipolar tungsten electrode was placed at the interface between layers I and II in the same cortical column as the recorded neurons (Kumar and Huguenard, 2001, 2003). Stimuli were delivered at low frequency (0.033 Hz and 0.3 ms in duration) through a bipolar tungsten electrode. Stimulus intensity/response curves were obtained (60 – 350 µA), and the stimulus intensity was set to 1.5 times the threshold. The threshold was defined as the stimulation intensity that evoked monosynaptic responses with constant latency (supplementary data, Fig. 1 and Fig. 2) at all holding potentials tested (Dobrunz and Stevens, 1997; Kumar and Huguenard, 2003) and with amplitudes larger than 2.5 times root-mean-square baseline noise (1.5 pA) (Wyllie et al., 1994). In most WT neurons, polysynaptic eEPSCs (Luhmann and Prince, 1990) were recorded at this intensity without any indication of space clamp errors. Linear regression was used to fit the rising phase of monosynaptic eEPSCs (from 20% to 80% of the peak amplitude). The extrapolated intersection of the regression line with the baseline was taken as the eEPSC starting point. The latency was defined as the time difference between this point and the center of the stimulus artifact (Jonas et al., 1993). The latency data in both WT and KO resembled a Gaussian distribution with similar mean and standard deviation (supplementary data, Fig. 2), suggesting that the monosynaptically stimulated synapses across experiments in both WT and KO mice were similar. The decay kinetics of the NMDAR-mediated eEPSCs were similar to those reported previously (Flint et al., 1997; Kirson and Yaari, 1996, 2000; Kumar and Huguenard, 2003). We tested for correlations between the amplitude and the decay time constants of the polysynaptic eEPSCs using Pearson linear correlation. However, no correlation (r2 ≤ 0.02) was found indicating that potential space-clamp errors were minimal in our recordings (Kirson and Yaari, 1996). Most NMDARs in layer V pyramidal neurons have been localized to proximal dendrites at a distance of between ~200 to 300 µm form the soma (Dodt et al., 1998; Frick et al., 2001). Together, the data indicate that it can be reasonably assumed that whole-cell recordings at the soma of the NMDAR-mediated EPSCs in this study were not significantly distorted by space-clamp errors (Williams and Mitchell, 2008). Spontaneous (s) EPSCs were continuously recorded for 20–30 minutes while cells were held at + 40 mV.
Data were collected using an Axopatch 200A amplifier (Molecular Devices, Sunnyvale, CA) and Clampex 9.2 software (Molecular Devices, Sunnyvale, CA) with compensation for series resistance (70%) and cell capacitance, filtered at 2 kHz, and digitized at 20 kHz using a Digidata 1320A (Molecular Devices, Sunnyvale, CA). Spontaneous EPSCs were analyzed using Clampfit 9.2 in template search mode. The template was constructed from an episode of original traces containing sEPSC events with slow rising and exponential decay phase, features characteristic of NMDAR-mediated EPSCs. All automatically detected sEPSC events were inspected and confirmed manually.
In equation 1, I is the peak current amplitude in pA, V is the holding potential in mV, [Mg2+]o is 1.2 mM in the ACSF and zF/RT = 0.07874 at 24°C. The values of the maximal conductance (gmax) and reversal potential (Vrev) were determined by linear regression of the current at Vh between +30 and +50mV and −20 and +10 mV, respectively. The non-linear least-squares estimates of the dissociation constant (Kd) of Mg2+ binding to NMDARs at 0 mV and the fraction (δ) of the membrane electric field sensed by the blocking site were determined using Clampfit or GraphPad Prism 4.0 (GraphPad Software Inc, La Jolla, CA). Group data is expressed as mean ± standard error mean (S.E.M.), and n = number of cells tested.
Decay kinetics in WT and KO were fitted with a two-component equation
as described previously (Jonas et al., 1993; Kirson and Yaari, 1996; Vicini et al., 1998). In equation 2, I is the peak current amplitude in pA, Af and As are the amplitudes of the fast and slow decay components, and τf and τs are the fast and slow decay time constants, t is the time in ms. In some cells As values were close to zero (< 0.01), suggesting that one component was sufficient to fit the current decay. For the comparison of the decay kinetics between WT and KO mice, we used a weighted mean decay time constant
Statistical significance was assessed using Student’s t-test or the Mann-Whitney t-test in case of unequal variance. A significance level of p < 0.05 (α = 0.05) was applied to all tests. Pearson linear correlation was used in order to test for correlations between pairs of variables.
We wish to thank Dr Nobuki Nakanishi for providing the NR3A KO mice. This work was supported by NIH-NS 31718 (FEJ), the Tuberous Sclerosis Alliance (FEJ, NJS) and Citizens United for Research in Epilepsy (NJS).
Publisher's Disclaimer: This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final citable form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.