Address correspondence and reprint requests to Dr Freesia L. Huang, Building 49, Room 6A36, NIH, 49 Convent Drive, MSC 4510, Bethesda, Maryland 20892–4510. E-mail: email@example.com
Neurogranin/RC3 (Ng) is a postsynaptic protein kinase C (PKC) substrate and calmodulin (CaM)-binding protein whose CaM-binding affinity is modulated by Ca2+, phosphorylation and oxidation. Ng has been implicated in the modulation of postsynaptic signal transduction pathways and synaptic plasticity. Previously, we showed a severe deficit of spatial memory in Ng knockout (KO) mice. Activation of the NMDA receptor and its downstream signaling molecules are known to be involved in long-term memory formation. In the present study, using mouse hippocampal slices, we demonstrated that NMDA induced a rapid and transient phosphorylation and oxidation of Ng. NMDA also caused activation of PKC as evidenced by their phosphorylations, whereas, such activations were greatly reduced in the KO mice. A higher degree of phosphorylation of Ca2+/CaM-dependent kinase II and activation of cyclic AMP-dependent protein kinase were also evident in the WT compared to those of the KO mice. Phosphorylation of downstream targets, including mitogen-activated protein kinases and cAMP response element-binding protein, were significantly attenuated in the KO mice. These results suggest that by its Ca2+-sensitive CaM-binding feature, and through its phosphorylation and oxidation, Ng regulates the Ca2+- and Ca2+/CaM-dependent signaling pathways subsequent to the stimulation of NMDA receptor. These findings support the hypothesis that the derangement of hippocampal signal transduction cascades in Ng KO mice causes the deficits in synaptic plasticity, learning and memory that occur in these mice.
Neurogranin/RC3 (Ng) is a 78-amino acid, neuron-specific, postsynaptic protein kinase C (PKC) substrate, expressed in the neuronal cell bodies and dendrites in hippocampus, neocortex, amygdala, and striatum (for review, see Gerendasy and Sutcliffe 1997). Ng binds calmodulin (CaM) in a Ca2+-sensitive manner and its CaM-binding affinity is attenuated by phosphorylation with PKC and by oxidation with nitric oxide (NO) (Huang et al. 2000). These unique biochemical properties have endowed Ng its function to modulate neuronal Ca2+ and Ca2+/CaM levels. To investigate the role of Ng in neuronal function, we have generated a strain of Ng knockout mouse (KO) (Pak et al. 2000). Previously, we showed that these Ng KO mice exhibited deficits in long-term potentiation (LTP), autophosphorylation of Ca2+/CaM-dependent kinase II (CaMKII), and spatial memory (Pak et al. 2000; Miyakawa et al. 2001). Using phorbol ester and forskolin, we also showed that the activations of PKC and cyclic AMP-dependent protein kinase (PKA), respectively, and their downstream components were attenuated in these Ng KO mice (Wu et al. 2002).
Activation of the N-methyl-D-aspartate (NMDA) receptor is critical for several forms of synaptic plasticity that underlies learning and memory. Several reports indicated that NMDA receptor activation is positively coupled to stimulation of PKC, CaMKII, adenylyl cyclase (AC) 1 and 8 and it is the activation of these Ca2+- and Ca2+/CaM-dependent enzymes and their downstream targets that contribute to NMDA receptor-dependent LTP (Wong et al. 1999; MacDonald et al. 2001; Bauer et al. 2002). Stimulation of PKC causes phosphorylation of Ng and intracellular application of antibodies, capable of binding to the Ng phosphorylation site domain, has been shown to prevent the induction of LTP (Fedorov et al. 1995). CaMKII, like Ng, is enriched at postsynaptic site and activated by Ca2+/CaM via Ca2+ influx through NMDA receptors (Lisman et al. 1997). The kinase not only phosphorylates its target substrates but also autophosphorylates rapidly at Thr286 of α-CaMKII (Soderling et al. 2001). Mice lacking αCaMKII show deficits in LTP and spatial learning (Giese et al. 1998). Similarly, AC1 and 8 are Ca2+/CaM-dependent enzymes, the activation of which causes an increase in cAMP and stimulation of PKA. Mice devoid of AC1 and 8 also exhibit deficits in LTP and learning and memory (Wong et al. 1999). It is well established that stimulation of NMDA receptor activates various pathways leading to the phosphorylation and activation of mitogen-activated protein kinase (MAP kinase) cascade and cAMP response element-binding protein (CREB) (Sweatt 2001; Kida et al. 2002; Kornhauser et al. 2002). Numerous other reports all indicated that these NMDA-mediated and Ca2+- and Ca2+/CaM-stimulated signaling reactions were important in the permanent memory formation.
In order to further dissect the involvement of Ng in the process of learning and memory, we compared the differences between WT and KO mice in Ca2+- and Ca2+/CaM-dependent signaling pathways mediated by NMDA. We showed that NMDA induced a rapid and transient phosphorylation and oxidation of Ng in the hippocampal slices of WT mice. In these mice, NMDA also caused activation of various PKC isozymes and PKA. Importantly, the activation was to a greater degree in the WT than that in the KO mice. Furthermore, Ng KO mice exhibited a lesser degree of activation of CaMKII, and other downstream signaling components, including MAP kinase, and CREB as compared to those of WT mice. These data suggest that Ng is involved and may act as an upstream modulator in the NMDA-mediated signaling pathway that are important for learning and memory.
Materials and Methods
NMDA was purchased from Calbiochem (San Diego, CA, USA). Tetrodotoxin (TTX), 2-amino-5-phosphonovaleric acid (APV), 2-(2-amino-3-methoxyphenyl)-4H-1-benzopyran-4-one (PD98059), chelerythrine and cAMP were obtained from Sigma (St Louis, MO, USA). Kemptide was purchased from Bachem (Torrance, CA, USA) and PKA inhibitor peptide (residues 5–22) from SynPep Corp. (Dublin, CA, USA). [γ-32P]ATP and Enhanced Chemiluminescence detection kit were obtained from Perkin Elmer Life Science (Boston, MA, USA). AG 1-X8 resin, protein assay reagent, and horseradish peroxidase-conjugated goat antirabbit IgG and goat antimouse IgG were from Bio-rad Laboratories (Hercules, USA).
Preparation of mouse hippocampal slices and stimulation with NMDA
After decapitation, the brain from adult mouse (3–5 months old) was removed immediately and placed into ice-cold artificial cerebrospinal fluid (ACSF, in mM: NaCl 124, KCl 4.9, CaCl2 2.0, NaHCO3 26, KH2PO4 1.20, MgSO4 2.0, glucose 10) and bubbled with 95% O2/5% CO2. Both hippocampi were removed and cut into 400 µm transverse slices using a McIlwain tissue chopper. The slices were incubated in ACSF in a chamber saturated with 95% O2/5% CO2 at room temperature for 3 h. TTX (5 µm) was added during the last 2 h of incubation to reduce endogenous synaptic activity before stimulation. Hippocampal slices from either WT or KO mice were treated with 100 µm NMDA in Mg2+ free-ACSF with added 10 µm glycine for the indicated times. To examine the effects of NMDA antagonist and PKC inhibitor, 500 µm APV or 20 µm chelerythrine, respectively, was added 15 min before the addition of NMDA. After the indicated incubation times, ACSF was removed, and the slices were kept frozen at −80°C until processing. Each sample containing 4–6 slices was sonicated in homogenization buffer (HB: 50 mm Tris-HCl, pH 7.5, 2 mm DTT, 2 mm EDTA, 1 mm EGTA, 50 µm AEBSF, 50 mm KF, 5 mm sodium pyrophosphate, 50 nm okadaic acid, 10 µg/mL leupeptin, and 5 µg/mL each of aprotinin and pepstatin A) containing 1% SDS, and the clarified homogenate was used for protein determination and immunoblot analysis. For control, hippocampal slices were either untreated or mock-treated for the duration of the experiments after pre-incubation. These samples, when immunoblot analyzed with various antibodies, did not produce any changes in the states of phosphorylation over the period of incubation.
Determination of Ng oxidation
Hippocampal slices from WT mice, after incubation with 100 µm NMDA for the indicated times, were homogenized in HB containing 0.5% NP-40 and 20 mm iodoacetamide without SDS and DTT. For detection of the oxidized form of Ng, the sample (containing 20 µg of protein/lane) was treated with SDS-gel sample buffer without mercaptoethanol and tracking dye and applied to a 10–20% gradient SDS-gel, which had been pre-run for 15 min. The oxidized Ng, has two intramolecular disulfide bonds, and therefore migrated faster than the reduced species, but both can be recognized by the same antibody. The extent of Ng oxidation was calculated by dividing the intensity of the oxidized form with the sum of the oxidized and reduced forms.
Frozen hippocampal slices were rapidly thawed and briefly sonicated on ice, in 100 µL of HB without SDS, but containing 0.1% NP-40. Homogenates were centrifuged at 20 800 × g for 5 min at 4°C, and the supernatant was used for protein determination and PKA assay. PKA assay was measured at 30°C for 5 min in a 25-µL mixture containing 30 mm Tris-HCl, pH 7.5, 6 mm MgCl2, 120 µm[γ-32P]ATP, 1.0 mg/mL BSA, 40 µm Kemptide, with or without 10 µm cAMP, and 1 µg protein of tissue extract. Reactions were stopped with 100 µL of ice-cold 20% trichloroacetic acid containing 20 mm ATP. After standing in ice for 10 min, the mixture was centrifuged, and the supernatant passed through a mini column (0.5 mL) of AG1-X8 resin in acetate form. The column was washed twice each with 1 mL of 30% acetic acid, 32P-labeled peptide substrates in the eluate were measured in a scintillation counter. All reactions were run in duplicate. PKA activities were expressed as pmol 32P incorporated into peptide substrate per µg protein per 5 min. Activities measured with cAMP were considered as total PKA activity, which remained constant throughout the incubation period. Activities measured without cAMP (cAMP-independent activity) following NMDA treatment were calculated as a percent of total activity. When protein extract was pre-treated with synthetic PKA inhibitor peptide (PKI, residues 5–22) before assay, the cAMP-activated activity and the increase in the cAMP-independent activity resulting from NMDA treatment were completely abolished.
Proteins of 20–30 µg from each sample were electrophoresed on a 10% SDS-gel (10–20% gradient gel for Ng) and were transferred onto nitrocellulose membrane overnight at 4°C. The membranes were washed for 10 min with TTBS (20 mm Tris-HCl, pH 7.5, containing 0.15 m NaCl, and 0.05% Tween 20), blocked with 5% non-fat dried milk in TTBS for 1 h before incubation with the primary antibodies for 3–4 h and subsequently with horseradish peroxidase (HRP)-conjugated secondary antibodies for 1 h at room temperature. The following primary antibodies were used at 1 : 1000 dilution: antiphospho-αCaMKII (Thr286) (Promega, Madison, WI, USA), antiphospho-MEK (New England Biolab), antiphospho-MAP kinase (Cell Signaling Technology, Beverly, MA, USA), antiphospho-CREB (Cell Signaling Technology), antiphospho-PKCα (Ser657) (Upstate, Charlottesville, VA), antiphospho-PKCε (Ser729) (Upstate), antiphospho-PKCpan(βII, Ser660) (Cell Signaling Technology), and antiphospho-PKCα/βII (Thr638/641) (Cell Signaling Technology). Use of antibodies #3615 and #270 for the detection of, respectively, phospho- and total Ng were as described previously (Pak et al. 2000; Wu et al. 2002). Signals were revealed via enhanced chemiluminescence reagent (ECL) and all blots were developed in the linear range suitable for densitometry. PKC, αCaMKII, MEK, MAP kinase, and CREB immunoreactivities were probed on those membranes previously blotted with phosphorylation-dependent antiobodies and have been stripped using buffer containing 62.5 mm Tris-HCl, pH 6.7, 100 mm 2-mercaptoethanol, and 2% SDS for 30–45 min at 55°C with constant shaking.
The data were expressed as mean ± SEM of at least three independent experiments each using slices from at least 6 of the WT or KO mice. Quantitative analysis for immunoblot was done by scanning the X-ray film and using the Fotodyne Gel-Pro Analyzer program (Media Cybernetics, Silver Spring, MD, USA). Statistical analysis was conducted by one-way anova followed by paired comparisons using Student's t-test.
NMDA mediated both phosphorylation and oxidation of Ng in hippocampal slices of WT mice
Exposure of WT mice hippocampal slices to 100 µm NMDA resulted in a time-dependent phosphorylation of Ng (Fig. 1a) as demonstrated by immunoblot analysis with #3615 antibody, which is specific for Ser36 phosphorylated Ng (Pak et al. 2000). NMDA-induced phosphorylation of Ng, at this concentration of NMDA, was transient, it reached a maximal level within 2–5 min and declined to basal level after 10 min. Degrees of Ng phosphorylation at 2 and 5 min were, respectively, 152 ± 4.6 and 139 ± 6.6% of basal level (n = 11, p < 0.001). After 20 min of incubation, phosphorylation of Ng slightly decreased below the level prior to treatment. Total Ng levels were unchanged by the above treatments throughout the incubation period as shown in the immunoblot analyses with #270 antibody, which is independent of phosphorylation state of Ng. To investigate whether the increased phosphorylation of Ng was mediated by NMDA receptor activation, we pre-incubated hippocampal slices with NMDA receptor antagonist APV (500 µm) for 15 min before NMDA treatment. Figure 1(a, inset) showed that APV blocked the NMDA-mediated phosphorylation of Ng.
Previous in vitro experiments have indicated that oxidation of Ng resulted in intramolecular disulfide formation and caused an attenuation of its binding affinity for CaM (Sheu et al. 1996; Huang et al. 2000). To investigate the physiological relevance of oxidized Ng, we tested the oxidation of this protein in mouse hippocampal slices with NMDA stimulation. Here, we showed that 100 µm NMDA, the same concentration used in the previous Ng phosphorylation experiment, caused a time-dependent oxidation of Ng. The oxidized Ng, having two intramolecular disulfide bonds was more compact and migrated faster than the reduced Ng as evidenced by the standard oxidized Ng run parallel with the samples (Fig. 1b). The same set of samples, when run with sample buffer containing tracking dye and mercaptoethanol, migrated as a single band as the reduced Ng (data not shown). Prior to treatment, the oxidized form of Ng was 6.3% of total Ng. A 2-fold increase in the oxidized form of this protein was observed following exposure to 100 µm NMDA for 2–5 min Unlike phosphorylation reaction, which was very brief, the oxidation reaction declined slowly. At 20–30 min, oxidized Ng was still at 7–8% of total Ng. These results demonstrate that NMDA receptor activation leads to both phosphorylation and oxidation of Ng in WT mice hippocampal slices. Needless to say, such reaction does not take place in the KO mice, as Ng is completely devoid in these mice.
Deletion of Ng attenuated NMDA-mediated phosphorylation of PKC isozymes
Ng has been shown to be phosphorylated in hippocampal slices under conditions where PKC was activated, such as during high frequency stimulation or phorbol ester treatment (Chen et al. 1997; Wu et al. 2002). We assayed PKC activities in the cytosolic and particulate membrane fractions of hippocampal slices post-NMDA treatment. Previously under the same experimental conditions we successfully demonstrated the translocation of PKC post-PMA treatment (Wu et al. 2002). However, translocation of PKC activities from the cytosolic to the particulate fractions could not be demonstrated following NMDA treatment. In addition, immunoblot analysis of these cytosolic and membrane samples with a polyclonal antibody that recognize conventional PKC did not detect any translocation either (data not shown). It has been reported that the activation of PKC family members are regulated by phosphorylation at three highly conserved sites in the catalytic domain (Keranen et al. 1995). One of these sites is located in the activation loop of PKC and is phosphorylated by PDK1, the other two are autophosphorylation sites, located in the turn motif and C-terminal hydrophobic motif (Edwards et al. 1999; Cenni et al. 2002). To investigate whether PKC activation is involved in NMDA receptor-mediated phosphorylation of Ng, we analyzed the activation of PKC in hippocampal slices after NMDA stimulation with various phosphorylation site-specific antibodies against PKC. We used antiphospho-PKC antibodies to detect phosphorylation specifically at Ser657 of PKCα, and Ser729 of PKCε, and PKCpan antibodies to detect phosphorylation mainly of Ser660 of PKCβΙΙ, but also Ser729 of PKCε and Ser657 of PKCα; these are the C-terminal auto-phosphorylation sites of hydrophobic motif. We also used antibodies specifically detecting phosphorylation of Thr638 of PKCα and the corresponding Thr641 of PKCβΙΙ, these are the penultimate C-terminal auto-phosphorylate sites of turn motif. We found that NMDA caused time-dependent phosphorylation of PKCα (Ser657), PKCε (Ser729) and PKCαII (Ser660) in hippocampal slices of WT mice, whereas, such activations were much reduced in KO mice (Fig. 2a). However, neither in WT nor KO mice, there was detectable change in phosphorylation of PKCα/βII(Thr638/641) during NMDA treatment. Furthermore, preincubation of hippocampal slices with chelerythrine 20 µm for 15 min to block PKC activation before NMDA stimulation resulted in a blockade of the NMDA-mediated phosphorylation of Ng (Fig. 2b). Apparently, failure to detect PKC translocation following NMDA treatment does not necessarily mean non-success for PKC activation, as PKC autophosphorylation at C-terminal domain important for PKC activation (Edwards et al. 1999; Cenni et al. 2002) and Ng phosphorylation was verified. These results show that NMDA caused autophosphorylation and activation of PKC and, in turn, phosphorylation of Ng; deletion of Ng results in the retardation of PKC isozymes phosphorylation/activation.
Ng KO mice have a defective mechanism for the autophosphorylation of CaMKII
We have shown previously that hippocampal slices of Ng KO mice displayed a reduced ability to generate activated CaMKII after stimulation of protein phosphorylation and oxidation by treatment with phosphatase inhibitor, okadaic acid, and NO donor, sodium nitroprusside, respectively (Pak et al. 2000). To test whether CaMKII autophosphorylation in the KO mice is also defective in response to neurotransmitter, hippocampal slices were treated with NMDA. As shown in Fig. 3, application of 100 µm NMDA to hippocampal slices of WT mice resulted in a robust increase in CaMKII phosphorylation detected by αCaMKII-Thr286-PO4-specific antibody. It peaked within 2–5 min but remained elevated for up to 30 min. In contrast, the same concentration of NMDA didn't produce such an increase of CaMKII autophosphorylation in Ng KO mice. The total hippocampal CaMKII immunoactivity in KO mice was about 10–15% higher than that of the WT, whereas the basal level of autophosphorylated CaMKII in the KO mice was found to be only about 50% of that in the WT mice. These results were consistent with our previous data determined by CaMKII activity assay (Pak et al. 2000). Together, these results indicate that the Ng KO mice are less able than the WT mice in the autophosphorylation of αCaMKII.
NMDA produced different kinetics of PKA activation between WT and KO mice
NMDA receptor stimulation leads to activation of PKA upon the induction of LTP (Roberson and Sweatt 1996). To address the question whether the deletion of Ng gene in mice alters the NMDA-mediated PKA activation, we measured PKA activity with and without cAMP in hippocampal extracts following NMDA stimulation (Fig. 4a). We took the increase in cAMP-independent activity as an index for PKA activation. Under control conditions, cAMP-independent activities determined in hippocampal protein extracts of both WT and KO mice were, respectively, 12.7 ± 0.70 (WT, n = 6) and 10.9 ± 0.95 (KO, n = 6) pmol per µg protein per 5 min, which corresponded to 18.8 ± 0.71% (WT, n = 6) and 18.3 ± 0.86% (KO, n = 6), respectively, of total (cAMP-dependent) activities that were 67.5 ± 3.95 (WT, n = 6) and 59.5 ± 1.92 (KO, n = 6) pmol per µg protein per 5 min (Fig. 4a). NMDA (100 µm) produced a time-dependent increase in cAMP-independent PKA activities while total activity in the presence of cAMP was not affected (Fig. 4a). After 2, 5, 10, and 20 min of incubations, cAMP-independent activities of PKA in WT mice were averaged 1.2-fold greater than those of the KO mice (percent of cAMP-independent activities were, WT (n = 6): 22.7 ± 0.73%, 2 min; 23.0 ± 0.75%, 5 min; 24.0 ± 1.45%, 10 min; and 23.8 ± 1.39%, 20 min; vs. KO (n = 6): 18.9 ± 0.92%, 2 min; 19.2 ± 1.15%, 5 min; 19.0 ± 1.10%, 10 min; 18.8 ± 0.57%, 20 min; t-test, p < 0.05–0.01) (Fig. 4). These results indicate that after deletion of Ng gene, NMDA does not produce an efficient activation of PKA.
Ng KO mice exhibited less activations of MAP kinase cascade and CREB following NMDA treatment
As the activation of NMDA receptor and its downstream signaling are required in LTP and learning and memory, we then analyzed the differences in phosphorylation of the downstream targets in the WT and KO mice responding to NMDA treatment. We focused on MAP kinase and CREB, because it was demonstrated that both the PKC and PKA systems were upstream regulators of MAP kinase cascades in hippocampal slices. In the present experiment, we found that 100 µm NMDA application to hippocampal slices of either WT or KO mice elicited phosphorylation of MAP kinase kinase (MEK) (Fig. 5a), however, the activation of MEK in KO mice was one half that of the WT. We also determined the activation of MAP kinase by immunoblotting with phosphorylation-dependent antibodies. Figure 5(b) showed that NMDA treatment results in rapid activations of both p44 and p42 MAP kinases within 10 min in WT mice, and the phosphorylation declined to near basal level after 30 min of incubation. In contrast, the same concentration of NMDA didn't promote such an increase in MAP kinase phosphorylation in Ng KO mice, as a matter of fact, slight dephosphorylation occurred after 20 min of incubation. Total MAP kinase levels, however, remained unchanged by the above treatments in both WT and KO mice. The enhanced phosphorylation of MAP kinases by NMDA is receptor-mediated, as it was completely blocked by the NMDA receptor antagonist APV at 500 µm (Fig. 5b inset). Preincubation of hippocampal slices with 50 µm PD98059, a specific inhibitor of MEK, for 40 min, before NMDA treatment, completely inhibited MAP kinase activation (data not shown).
CREB is a transcription factor that is phosphorylated in response to Ca2+ influx; it is also a downstream target of MAP kinases in hippocampus (Roberson et al. 1999). We analyzed the phosphorylation of CREB at Ser133 during NMDA treatment. As shown in Fig. 6, incubation of hippocampal slices of WT mice with NMDA led to a robust increase in CREB phosphorylation (percent of 0 min: 2 min, 196 ± 9.3%; 5 min, 164 ± 8.4%; 10 min, 142 ± 10.4%; 20 min, 124 ± 14.4%; 30 min, 100 ± 5.0%; n = 5). In contrast, Ng KO mice exhibited a much reduced increase of phosphorylation on CREB. Like MAP kinases, after 20 min of incubation, dephosphorylation of CREB was already taking place in these Ng KO slices. The total CREB level remained unchanged throughout the whole period of incubation in both WT and KO mice.
It is widely accepted that NMDA-type glutamate receptor constitutes a major source of Ca2+ influx into neurons following tetanic stimulation that induces LTP. This Ca2+ influx has been shown to activate Ca2+- dependent protein kinases such as PKCs and Ca2+/CaM-dependent enzymes such as CaMKII (Giese et al. 1998), AC (Wong et al. 1999), and NO synthase (Bon and Garthwaite 2003). These enzymes have prominent roles in the synaptic plasticity underlying learning and memory (Sweatt 2001). Ng is a postsynaptic protein believed to serve as a CaM buffer to regulate neuronal intracellular Ca2+ and Ca2+/CaM levels (Gerendasy and Sutcliff 1997; Huang et al. 2000). The present data demonstrated that NMDA caused both phosphorylation and oxidation of Ng in hippocampal slices of WT mice. As both phosphorylated and oxidized Ng have reduced CaM binding affinities, phosphorylation and oxidation of Ng would be an efficient mechanism for prolonging the availability of Ca2+/CaM for such activator-requiring enzymes in the post-NMDA receptor signaling pathways. Thus, Ng enhances several NMDA receptor-mediated signaling reactions involved in the long term memory formation. Lacking Ng, hippocampal slices of Ng KO mice displayed a reduced ability to generate autophosphorylated and activated CaMKII and PKC isozymes, as well as PKA activation. Subsequently, phosphorylations of the downstream targets, including MEK, MAP kinases, and CREB were also significantly attenuated in KO mice following treatment with NMDA. These results suggest that Ng is an upstream modulator of the multisignaling pathways involved in learning and memory, and the attenuation of hippocampus signal transduction cascades in the Ng null mice might be the culprit for their severe deficits in learning and memory. Alternatively, we addressed the possibility that the NgKO mice may have defect in their NMDA receptor expression. Immunoblot analyzes for NR2A and NR2B subunits of NMDA receptor did not indicate any difference between WT and KO mice in their levels of receptor expression.
Electrophysiological and biochemical approaches have identified activation of PKC as one of the key reactions in many different cellular processes including the early and late-phase of LTP, learning and short- and long-term memories (Kleschevnikov and Routtenberg 2001; Sweatt 2001). Early studies showed that Ng was a prominent PKC substrate. On the other hand, level of Ng may also determine the extent of PKC activation and this is attributed by the buffer capacity of Ng for CaM that prevents the total binding of Ca2+ to CaM. In KO, all the calcium entry will be complexed with CaM and leave few free calcium for PKC activation. Thus, Ng can be considered as both upstream and downstream of PKC activation.
It is generally thought that PKC translocation from cytosol to the membrane is a measure of enzyme activation. However, results from different studies revealed that activation of some neurotransmitter receptors, including NMDA receptors, muscarinic acetycholine receptors, and metabotropic glutamate receptors, did not cause a detectable translocation of PKC isozymes (Angenstein et al. 1997). Keranen et al. (1995) reported that PKC was post-translationally processed by three distinct phosphorylations, two of them, namely Thr641 and Ser660 in the sequence of PKCβΙΙ are autophosphorylation sites located at the C-terminal turn motif and hydrophobic motif, respectively, of PKC. These autophosphorylations took place after first trans-phosphorylated at Thr500 of activation loop by a phosphoinositide-dependent kinase. While all phosphorylations are important for PKC activation, they were considered to occur constitutively post-translation, and agonist-mediated phosphorylation was not reported. In transfected HEK293 cells or NIH3T3 fibroblasts, Cenni et al. (2002) showed PDGF-dependent phosphorylation of Thr566 of activation loop and Ser729 of hydrophobic motif in PKCε. Using antibody specific for both phosphorylated Thr634 and Thr641 of PKCβII, Sweatt et al. (1998) and Atkins et al. (1998) demonstrated an increase in immunoreactivity in hippocampus CA1 post-high frequency stimulation and fear conditioning, respectively. Our present experiment (Fig. 2), however, showed a strong NMDA-mediated increase in immunoreactivities of antibodies PKCpan, which detects mainly phosphorylated PKCβΙΙ (Ser660) as well as phosphorylated PKCε (Ser729) and PKCα (Ser657). The post-NMDA-mediated increase of phosphorylation in the latter two PKCs were also evidenced by using isozyme specific antibodies, namely, antiphospho-PKCα (Ser657) and antiphospho-PKCε (Ser729). On the other hand, immunoreactivity of antiphospho-PKCα(Thr638)/PKCβΙΙ(Thr641) remained unaltered post-NMDA treatment. In addition, this antibody detected purified rat brain PKCβ and incubating the enzyme with ATP/Ca2+/PS/DG did not further increase the immunoreactivity (data not shown). All in all, these results point to the possibility that, NMDA, via Ca2+ influx, promotes PKC activation/phosphorylation strongly at Ser660 of PKCβII, Ser729 of PKCε and Ser657 of PKCα, the C-terminal hydrpphobic motif autophosphorylation site, but not that at Thr641 of PKCβII or Thr638 of PKCα, the penultimate C-terminal turn motif autophosphorylation site, which is likely already phosphorylated in unstimulated neuron. Sweatt et al. (1998) also detected phosphorylation in their purified rat brain PKCβ preparation and suggested that Thr641 of PKCβII was already phosphorylated and the increase in immunoreactivity post-high frequency stimulation could be due to phosphorylation of Thr 634 as their antibody detected both Thr634 and Thr641. Whether Thr634 phosphorylation can be stimulated by NMDA awaits to be determined by the development of a site-specific antibody.
The direct result of PKC activation after NMDA stimulation is the increased phosphorylation of target substrates such as Ng. This NMDA-mediated PKC activation mechanism is obviously down-regulated in Ng KO mice. Previously, with PMA (Wu et al. 2002), we also observed a retarded translocation of PKC in Ng KO mice. In light of the importance of PKC in synaptic plasticity, the defect of PKC activation in these Ng KO mice may explain, at least partially, their inferior behavioral performance (Pak et al. 2000).
We also found that Ng in mice hippocampal slices was oxidized by the addition of NMDA. It has been shown that activation of NMDA receptors and the subsequent increase in intracellular Ca2+ and Ca2+/CaM levels activate NO synthase. We previously showed that NO synthase inhibitors were effective in blocking the NMDA-mediated oxidation of Ng, suggesting that Ng oxidation serves as a downstream target of the NO signaling (Li et al. 1999). It is readily understood that after NMDA receptor stimulation, the influxed Ca2+ will replace Ng from the CaM/Ng complex to form Ca2+/CaM. It is this availability of Ca2+/CaM that activates the CaMKII at the postsynaptic sites and other Ca2+/CaM-stimulated enzymes including AC1 and 8, and NOS. Activation of PKC and NOS cause phosphorylation and oxidation, respectively, of Ng, and these modified Ng were unable to bind CaM, thus, prolonging the availability of Ca2+/CaM. Without Ng to buffer CaM, as in the KO mice, the modulations of postsynaptic free Ca2+ and Ca2+/CaM would be askew and so would the downstream physiological responses.
Phosphorylation and activation of α-amino-3-hydroxy-5-methylisoxazole-4-propionic acid (AMPA)-type glutamate receptor by CaMKII has been reported as an important step during LTP (Derkach et al. 1999; Lee et al. 2000). Therefore, a deficit in the activation of CaMKII as in the Ng KO mice could inhibit hippocampal synaptic plasticity and hippocampus-dependent learning and memory. It has been well documented that PKA is critically involved not only in the induction of LTP and short-term memory (Blitzer et al. 1995; Roberson et al. 1996), but also in the late phase of LTP and in hippocampus-based long-term memory (Abel et al. 1997). It is also known that application of NMDA in the incubation medium of hippocampal slices is sufficient to activate PKA, as we have shown in this study. Very likely, PKA is activated by elevated cAMP resulting from activation of two Ca2+/CaM-dependent forms of adenylyl cyclases, namely AC1 and AC8, that are also enriched in the rodent hippocampus. In addition, stimulated PKC after Ca2+ influx will also activate AC 2 and 7, which could further increase the cAMP level and lead to a greater increase in PKA activity (Tang and Hurley 1998; Wong et al. 1999). Needless to say, deletion of Ng gene, as in the Ng KO mice, has greatly impeded the activation of PKA, either by NMDA (this study) or by forskolin (Wu et al. 2002), and the subsequent signaling involved in hippocampus-dependent learning and memory.
It is well recognized now that both PKA and PKC signaling pathways link NMDA receptor activation to MAP kinases activation and the later is a critical regulator of CREB phosphorylation (Roberson et al. 1999). Understandably, the insufficient PKA and PKC activation as occurring in the Ng KO mice would have a decisive impact on the downstream signaling of phosphorylation of MAP kinase and the transcriptional factor, CREB. Indeed, these Ng null mice displayed a reduced ability to generate phosphorylated and activated forms of MAP kinases and CREB following treatment with NMDA in hippocampal slices. In addition, they did not show optimal activation of MAP kinase cascade and CREB when hippocampal slices were directly treated with PMA for PKC, and forskolin for PKA activation (Wu et al. 2002). There is strong evidence to indicate that MAP kinases play critical roles in triggering the late phase of LTP through regulating CREB phosphorylation (Sweatt 2001), and CREB-dependent transcription is required for the cellular events underlying long-term memory (Josselyn et al. 2001; Nguyen 2001). Impey et al. (1996) directly demonstrated CREB activation in hippocampal CA1 after LTP-inducing stimulation. Thus, the derangement of MAP kinase cascades and CREB in the Ng KO mice, reduces gene expression and protein synthesis for memory registration, and may eventually lead to their deficits in synaptic plasticity and learning and memory as we previously reported (Pak et al. 2000).
In summary, using a Ng KO model, this study demonstrates that Ng, with its unique CaM binding property, plays a first hand regulatory role in the availability and maintanance of Ca2+ and Ca2+/CaM in the hippocampus in response to NMDA receptor activation. Meanwhile, phosphorylation and oxidation of Ng finely tunes the activation of Ca2+- and Ca2+/CaM-requiring events involved in neuronal signaling of learning and memory. These findings provide the likely explanation at molecular level for the impairment in learning and memory of these Ng KO mice and advance the possibility that Ng functions as an upstream modulator in signaling pathways in these processes.