7.2.1. Protein Kinase A Regulation of NMDA Receptors

Protein kinase A (PKA) has been shown to increase NMDAR currents as indicated through elevation of PKA activity by forskolin or cAMP analogs or direct intracellular administration of PKA [16]. Moreover, NMDAR currents are increased by activation of PKA through stimulating G-protein coupled receptors (GPCRs) including β-adrenergic receptors with agonists norepinephrine and isoproterenol [129]. The increase in NMDAR current appears to occur through increased gating as indicated by increased channel open probability (P_o) and currents evoked by exogenously administrated NMDAR agonists. NMDAR-mediated excitatory postsynaptic currents (EPSCs) are increased by PKA. The enhancements may be due in part to PKA-mediated suppression of the desensitization of synaptic NMDARs [98].

PKA is held in association with NMDAR complexes through binding to two main scaffolding proteins or A kinase anchoring proteins (AKAPs). AKAP 79/150 [54] interacts indirectly with NMDARs via PSD-95 [23] and yotiao binds directly to the NR1 subunit [68,109,145]. A critical feature of these scaffolds is that they anchor both PKA and the phosphoprotein phosphatases—calcineurin and protein phosphatase 1 (PP1), respectively—that oppose the action of PKA on NMDAR function (see below). This allows highly localized, tightly balanced, and complex regulation of NMDAR function by the interplay of PKA and these phosphatases. Under basal conditions in one prominent model, constitutively active PP1 may keep NMDARs in a state of dephosphorylation and low activity. Upon activation by cAMP, PKA may phosphorylate PP1, decreasing its activity and thereby, with direct phosphorylation of the channel, lead to a shift in the balance of NMDARs to a higher phosphorylation state and thus a higher activity state [82,145].

While most studies focused on NMDAR currents, recent evidence suggests that, in addition to changes in channel gating, permeability of the channel to Ca²⁺ may be regulated by PKA phosphorylation [112]. Most compellingly, Ca²⁺ entry through the channels, as assayed by an indicator dye overload technique, was suppressed by inhibitors of PKA to a much greater extent than would be predicted by the decrease in NMDAR current. Consistent with this, PKA blockers reduced the P_Ca/P_monovalent ratio, as determined from reversal potential shifts of NMDAR currents. It was also found that NMDAR-mediated Ca²⁺ increases in dendritic spines were suppressed by these blockers, with little or no decrease in NMDAR EPSCs. Notably this effect was more prominent in neurons from young than in those from mature animals. While this work relied exclusively on blockers of PKA, it raises the possibility that phosphorylation may regulate channel permeability in addition to changes in channel gating and receptor trafficking, and that this regulation may be developmentally controlled.

MacDonald and colleagues [75] described an additional PKA-mediated regulatory mechanism by which PKA may conversely decrease NMDAR currents. They found that activation of GPCRs such as the PDGF receptor leads to activation of PKA and subsequent phosphorylation of the Csk tyrosine kinase. Csk is a major negative regulator of Src family kinases. PKA-mediated phosphorylation of Csk activates this kinase, inhibiting Src kinases, and thereby suppressing NMDAR currents.

7.2.2. Protein Kinase C Upregulation of NMDA Receptors

The serine/threonine PKC family is ubiquitously expressed and involved in multiple neuronal functions including neurotransmitter release, receptor regulation, and synaptic remodeling. Based on its structure and selective sensitivity to second messenger activators Ca²⁺ and diacylglycerol (DAG), PKC can be classified into three major groups. The conventional or calcium-dependent cPKCs (α, βI, βII, and γ) are activated by Ca²⁺ and DAG. The novel or calcium-independent nPKCs (δ, θ, η, and ɛ) lack a Ca²⁺ binding domain but are still activated by DAG. Finally, the atypical, aPKCs (ζ and λ/ι), are both Ca²⁺- and DAG-independent but sensitive to other phospholipids [4]. PKC activation is associated with its translocation from the cytosol to the different intracellular compartments including plasma and nuclear membranes, where it is held close to the pertinent substrates by interacting with PKC anchoring proteins or RACKs (receptors for activated C kinase).

Each PKC isozyme may have a specific RACK or anchoring protein that directs the relocation of PKC after its activation and in part mediates isozyme-specific function [108]. PKC isozymes (β, γ, and ɛ) were found in the NMDAR complex in the PSD [44]. Thus, upon activation, translocated membrane-bound PKC may phosphorylate NMDARs and other proteins in the PSD.

Evidence supporting a role for PKC in regulating NMDAR function is abundant. Electrophysiological studies showed that activation of PKC by application of 4β-phorbol 12–myristate 13-acetate (4β-PMA) enhanced peak NMDA-evoked currents recorded from isolated CA1 hippocampal neurons. This potentiation is prevented by PKC inhibitors chelerythrine or calphostin C, confirming the role of endogenous PKC in response to 4β-PMA [72]. The constitutively active fragment of PKC (PKM) also potentiates peak NMDA currents in hippocampal neurons [148]. PKC modulation of peak NMDA currents depends primarily on the NR2 subunits expressed. Enhancement by PKC is pronounced for receptors containing the NR2A or NR2B subunits, but absent for receptors containing the NR2C or NR2D subunits [135,47]. Instead, activation of PKC inhibits NR2C and NR2D responses [31]. PKC also enhances the functions of synaptically located NMDARs as indicated by the increased NMDAR-mediated components of spontaneous miniature currents by PKM [72].

In Xenopus oocytes expressing native NMDARs, stimulation of PKC using 12-O-tetradecanoyl phorbol-13-acetate (TPA) potentiates NMDA channel activity, with no change in single-channel conductance, reversal potential, or mean open time [69]. Activation of GPCRs including phosphoinositol-coupled metabotropic glutamate receptors [113], muscarinic receptors, and lysophosphatidic acid receptors [72] also potentiates NMDAR currents via activation of PKC.

PKC potentiation of NMDAR function may be mediated through direct phosphorylation of NMDARs. Biochemical studies show that NR1, NR2A, and NR2B subunits may be phosphorylated by PKC both in vitro and in vivo [62,127]. Phosphorylation sites have been identified for NR1 (Ser-890 and Ser-896) [127], NR2A (Ser-1416) [28], and NR2B (Ser-1303 and Ser-1323) [65]. Using specific inhibitors of PKC isoforms and antibodies recognizing specifically phosphorylated serine, PKC sites of NR1 have been found to be phosphorylated by different PKC isoforms, with Ser-896 phosphorylated by PKC α and Ser-890 phosphorylated by PKCγ [104].

Earlier studies have shown PKC-induced potentiation of NR1/NR2A receptor currents in mutant receptors lacking the entire intracellular domains of both NR1 and NR2A [152], which contain consensus PKC phosphorylation sites. It is therefore unlikely that the potentiation is caused by direct phosphorylation of NMDAR subunits. Instead, PKC could modify associated proteins, involved in signaling and/or trafficking of NMDARs.

Evidence supports the role of direct phosphorylation in PKC-mediated potentiation of NMDAR currents. In Xenopus oocytes expressing NR1/NR2B receptors, direct phosphorylation of NR2B (Ser-1303 and Ser-1323) is involved in the PKC-mediated potentiation of NMDAR currents, as mutation of either of these residues severely reduces PKC potentiation [65]. A comparable effect was seen for insulin that potentiates NMDAR currents through PKC. Mutation of Ser-1303 and Ser-1323 of NR2B significantly reduces the potentiation effect of insulin on NR1/NR2B receptors. Similarly, mutating of homologous sites in NR2A (Ser-1291 and Ser-1312) abolishes the insulin potentiation of NR1/NR2A receptors [48].

Evidence indicates that PKC potentiation of NMDARs may involve cross-talk with tyrosine kinase pathways. Studies of GPCRs have shown that the Src tyrosine kinase is involved in the PKC-induced potentiation and functions as the downstream of PKC. Activation of muscarinic receptors and lysophosphatidic acid receptors potentiates NMDARs current in isolated CA1 hippocampal neurons via activation of PKC. This potentiation is blocked by tyrosine kinase inhibitors genistein or lavendustin A, Src unique domain peptide fragments (40–58), and the anti-cst1 anti-body that selectively inhibits the Src family of kinases [72]. The intermediary between PKC and Src is focal adhesion kinase cell adhesion kinase-β/proline-rich tyrosine kinase 2 (CAKβ/Pyk2) [40]. PKC activates CAKβ/Pyk2 in hippocampal neurons [63,40] that in turn binds to and activates Src kinase [40]. Thus, PKC potentiation may also be mediated via the sequential activation of CAKβ/Pyk2 and the nonreceptor tyrosine kinase Src (see section below).

In addition to increasing NMDAR currents per se in neurons in the trigeminal nucleus caudalis, activating PKC causes relief of the voltage-dependent blockade of NMDARs by Mg²⁺ [18]. While this relief is prominent in these neurons, the degree of relief of Mg²⁺ blockade is markedly less in neurons in the hippocampus [139]. Thus, this mechanism for enhancing NMDAR currents in physiological conditions may have limited relevance to the functions of NMDARs in the orofacial processing region of the brain stem.

PKC thus regulates multiple properties of NMDARs. In addition to potentiating peak NMDAR current, PKC also enhances NMDAR desensitization. In CA1 pyra-midal neurons of the hippocampus, activation of PKC potentiates peak NMDAR currents and also enhances inactivation of steady-state NMDAR currents. This Ca²⁺- dependent inactivation is mediated via the competitive binding of Ca²⁺ CaM to a site located on the C terminus of the NR1 subunit that also binds α-actinin 2 [71]. Unlike potentiation of peak currents by PKC that occurs through the sequential activation of the tyrosine kinases CAKβ/Pyk2 and Src [72], enhancement of Ca²⁺-dependent inactivation of NMDAR currents is independent of CAKβ/Pyk2 and Src activity [71]. Activation of PKC also enhances glycine-insensitive desensitization of NR1/NR2A receptors expressed in HEK-293 cells, which is independent of previously identified PKC sites in NR1 and NR2A but may depend on the unidentified PKC sites [47]. Thus, PKC exerts a combined effect on NMDARs, namely enhancement of peak currents and suppression of steady-state currents. By regulating NMDAR desensitization kinetics, PKC may allow more precise control over the time course of Ca²⁺ entry following NMDAR activation, thereby preventing excessive and potentially damaging ionic influxes.

7.2.3. Other Serine and Threonine Kinases

7.2.4. Phosphatases Opposing Upregulation by Serine and Threonine Kinases

The major phospho-serine/threonine protein phosphatases 1 (PP1), 2A (PP2A), and 2B (PP2B or calcineurin) suppress the activities of NMDARs, presumably opposing the actions of the kinases described above. In excised patches from hippocampal neurons, applying exogenous PP1 or PP2A depressed open probability of NMDAR single channels [140]. Conversely, selective inhibitors of PP1 and PP2A, calyculin A and okadaic acid (at low concentrations), exerted the opposite effect, increasing NMDAR currents. This implies that these phosphatases endogenously regulate NMDAR channel activity. Subsequently, the regulation by PP1 was determined to be due to localization of this enzyme at NMDARs through anchoring to yotiao [145].

Likewise for calcineurin, by using cell-attached recordings in acutely dissociated adult rat dentate gyrus granule cells, inhibitors of this phosphatase (high concentration okadaic acid or FK-506) prolonged the duration of single NMDA channel openings, bursts, clusters, and superclusters [66].

These inhibitors were ineffective when Ca²⁺ entry through NMDA channels was prevented, indicating that calcineurin, activated by calcium entry through native NMDA channels, shortens the duration of channel openings. Subsequently, calcineurin-mediated feedback was shown to regulate synaptic NMDAR currents [129] and depend on anchoring of calcineurin to the scaffold AKAP 79/150 [54]. By measuring NMDAR currents from NR1/ NR2A expressing HEK-293 cells, Ser-900 and Ser-929 were identified as residues dephosphorylated by calcineurin [58].

The phosphatases are subject to regulation that may vary in different neuronal types leading to cell-type specific regulation. For example, in striatal neurons, stimulating D1 dopamine receptors activated adenylate cyclase, increased cAMP, and activated PKA that then phosphorylated and activated the protein phosphatase inhibitor, DARPP-32. Phosphorylated DARPP-32 is a potent inhibitor of PP1 [11]. In hippocampal neurons, PP1 is inhibited by a different protein known as inhibitor 1, which is also a substrate of PKA [110]. Thus, PKA activation leads to inhibition of PP1 and decreased dephosphorylation (i.e., enhanced phosphorylation) of downstream substrates including NMDARs [122]. In CA1 neurons, PKA-phosphorylated inhibitor 1 likely interacts with and inhibits PP1 causing enhanced phosphorylation of NMDARs [110].

7.3.1. Enhancement of NMDA Receptor Function by Src

In the mammalian CNS, five members of the Src family of nonreceptor PTKs are expressed: Src, Fyn, Yes, Lck, and Lyn. These kinases were initially thought to be involved in regulating cell proliferation and differentiation because Src, the prototype member, was initially identified as a proto-oncogene [118]. However, Src family kinases (SFKs) are expressed in neurons of the adult CNS [24,119], suggesting additional functions for SFKs since neurons are differentiated postmitotic cells.

SFKs are now known to be expressed widely throughout the CNS and involved in a range of cellular functions. One major function of SFKs in the developed CNS is regulating the activities of ion channels. In the CNS, the first type of channel found to be subject to regulation by SFKs was the NMDAR subtype of ionotropic glutamate receptor [142]. Subsequently, SFKs have been shown to regulate other types of channels in CNS neurons including voltage-gated ion channels, such as potassium [25] and calcium channels [14], as well as ionotropic neurotransmitter receptors, including GABAA (γ-aminobutyric acid type A) receptors [86,136] and nicotinic acetylcholine receptors [138].

Electrophysiological recordings from neurons showed that NMDAR currents are governed by a balance between tyrosine phosphorylation and dephosphorylation. Inhibiting endogenous PTK activity [142,143] or increasing phosphotyrosine phosphatase (PTP) activity by introducing exogenous PTP [143] leads to suppression of NMDAR currents. Conversely, inhibiting endogenous PTP activity or increasing PTK activity by introducing exogenous Src causes enhancement of NMDAR currents [142]. Exogenous Src and Fyn were found to potentiate currents mediated by recombinant NMDARs expressed in HEK-293 cells [56] and in Xenopus oocytes [20]. As shown on recordings of NMDAR single channel currents, the predominant effect of PTK activity or inhibiting PTPs was to increase NMDAR channel gating with no effect on NMDAR single channel conductance [143]. Moreover, because the effects of manipulating PTKs and PTPs were present with NMDARs in excised membrane patches, it was inferred that PTK and PTP must be intimately associated with the NMDAR complex.

While these studies also showed that exogenous SFKs are sufficient to enhance NMDAR channel gating, further work is needed to identify the endogenous PTK and PTP compounds that mediate NMDAR upregulation and downregulation, respectively. Kinases in the Src family were implicated as endogenous enzymes upregulating NMDAR activity via a phosphopeptide SFK activator (pYEEI peptide) that increased the activities of synaptic NMDAR-mediated currents in cultured neurons [150] and in CA1 pyramidal neurons in hippocampal slices [73]. When applied to the cytoplasmic aspects of inside-out membrane patches, the activating peptide produced an increase in gating of NMDARs without affecting single channel conductance. Conversely an SFK inhibitory antibody known as anti-cst1 exerted an opposite effect, depressing NMDAR channel gating.

Are all five Src family members expressed in the CNS responsible for the upregulation of NMDAR function or do specific SFKs regulate NMDARs? Src [40], Fyn [121], Lck, Lyn, and Yes [50] are found in the postsynaptic density (PSD), the main postsynaptic structural component of glutamatergic synapses. Furthermore, Src [150], Fyn, [149], Lyn, and Yes [50] were shown to be components of the NMDAR complex. Thus, Src, Fyn, Lyn, and Yes are all at appropriate locations to potentially regulate NMDAR function.

Src was implicated through the use of reagents—an inhibitory antibody (anti-src1) [100] and an inhibitory peptide (Src40–58) [150]—that selectively inhibit this kinase but not other members of the Src kinase family. Each of these Src-specific inhibitors decreases synaptic NMDAR-mediated currents and produces a decrease in NMDAR channel gating, the same changes caused by the general SFK inhibitor, the anti-cst1 antibody. Src40–58 is the immunogen for anti-src1 and corresponds to amino acids 40 to 58 within the unique domain of Src. These reagents were hypothesized to block Src-mediated upregulation of NMDAR activity by disrupting a protein–protein interaction of the Src unique domain that allows Src to interact with NMDARs to modify receptor function.

As mentioned above, exogenous Fyn was shown to increase currents mediated by recombinant NMDARs [56] but whether endogenous Fyn or other SFKs present in the CNS regulates native NMDARs remains to be tested directly because inhibitors that selectively block the activities of these SFKs have yet to be developed. However, the Src-specific inhibitors prevent the increase in channel activity produced by the SFK activating pYEEI peptide [150], implying that endogenous Src plays a critical role in the upregulation of NMDAR activity by SFKs. Src may cause upregulation of NMDAR channel gating via direct tyrosine phosphorylation of the NR2A and NR2B subunits. As discussed elsewhere [103], it is possible that phosphorylation by Src of proteins in the NMDAR complex, other than NMDAR subunits, may be responsible for Src-mediated changes in NMDAR function.

7.3.2. Anchoring Src in NMDA Receptor Complex via Unique Domain

That Src coimmunoprecipitates as part of the NMDAR complex [150] implies that it is held there by binding to an anchoring protein or proteins. The main attributes of such a protein are as follows [30]: (1) it must bind directly to the unique domain of Src through amino acids 40 to 58; (2) this binding must be prevented by a peptide with the sequence of amino acids 40 to 58 of Src (Src40–58); (3) the protein must be present at excitatory synapses and must be a component of the NMDAR complex; and (4) lack of the protein must prevent upregulation of NMDAR activity by endogenous Src. A candidate protein identified by yeast two-hybrid screening using bait constructs containing the Src unique domain was NADH dehydrogenase subunit 2 (ND2), a 347-amino acid protein known to be a subunit of the inner mitochondrial membrane enzyme NADH dehydrogenase (Complex I).

Direct interaction of the Src unique domain and ND2 was confirmed through in vitro binding assays. Results from these experiments also identified the ND2.1 region as necessary and sufficient for interacting with the Src unique domain. ND2.1 bound directly to the Src40–58 peptide and the in vitro binding of the Src unique domain to ND2.1 was prevented by Src40–58. Src and ND2 coimmunoprecipitated from tissue extracts and, importantly, from PSD preparations from brain. The co-immunoprecipitation was prevented by Src40–58, implying that the Src–ND2 interaction identified in vitro may occur in vivo.

In addition to finding ND2 in PSD protein preparations, ND2 immunoreactivity was found by immunogold electron microscopy in PSDs in the CA1 hippocampus. Coimmunoprecipitation experiments indicated that ND2 is a component of the NMDAR complex and that the Src–ND2 interaction is required for the association of Src, but not ND2, with NMDARs. Finally, we found that depleting ND2 suppresses Src association with the NMDAR complex and prevents the upregulation of NMDAR function by activating endogenous Src at excitatory synapses. These multiple and converging lines of evidence lead to the conclusion that ND2 mediates interactions between NMDARs and the unique domain of Src. Surprisingly, ND2 acts as an adaptor protein that anchors Src within the NMDAR complex, where it allows Src to upregulate NMDAR activity.

The finding that ND2 binds to Src through the unique domain establishes that, like the SH2 and SH3 domains, this part of Src is a protein–protein interaction region. ND2 binds to Src through a sequence that is not conserved among members of the Src kinase family. Because the unique domains of several Src family kinases have potential binding partners [29,134], a unifying principle for the role of this region may be in mediating protein–protein interactions. However, unlike the highly-conserved SH2 and SH3 domains that mediate interactions shared by Src family members, the weakly conserved unique domains readily allow distinct interactions for each kinase. Differences in unique domain binding partners may contribute to the functions of the various members of the Src family of kinases, including Fyn that is also known to be held within the NMDAR complex but does not interact with ND2 [30].

7.3.3. Regulation of Src within NMDA Receptor Complex

As in other systems, the activity of Src within the CNS is tightly regulated. At excitatory synapses in the adult CNS, the basal activity of Src is normally maintained in a low state but can be enhanced by upstream signaling events. Src family kinases serve as molecular hubs through which numerous signaling cascades converge to regulate NMDARs [103]. Some of the same molecules identified in other systems that regulate Src activity also play important roles in the regulation of Src within the NMDAR complex and include the activating enzymes tyrosine kinase CAKβ/ Pyk2 [40] and the protein tyrosine phosphatase PTPα [61] along with the inhibitory kinase Csk [144]. In addition to these well-characterized regulators of Src, three PSD proteins were recently identified to modulate Src within the NMDAR complex: RACK1 [149], H-Ras [126], and PSD-95 [49].

7.3.4. STEP Opposition to Src Upregulation of NMDA Receptor Function

NMDAR function is not regulated by Src alone but by the balance of the activities of Src and a PTP that depresses NMDAR gating, reversing the effects of Src. Inhibiting PTPs pharmacologically increases NMDAR channel gating in excised membrane patches [143] and PTP activity coimmunoprecipitates with NMDARs [3], indicating that endogenous PTP is intrinsic to the NMDAR complex. One family of PTPs observed at the PSD of glutamatergic synapses are the STEPs (striatal-enriched tyrosine phosphatases) [93], a family of brain-specific, nonreceptor type PTPs [9].

The STEP61 isoform has been found to be a component of the NMDAR complex in spinal cord and hippocampus [95] and therefore is located appropriately to downregulate NMDAR function. Applying recombinant STEP to the cytoplasmic aspects of inside-out membrane patches suppresses NMDAR channel gating, mimicking the effect of inhibiting Src. Similarly, recombinant STEP applied intracellularly reduces NMDAR EPSCs. Conversely, intracellular application of a function-blocking STEP antibody or a dominant-negative STEP produced an increase in NMDAR-mediated EPSCs, implying that NMDAR activity is regulated by endogenous STEP. Both the reduction of NMDAR currents produced by exogenous STEP and the increase of NMDAR currents arising from inhibiting endogenous STEP required Src since both were prevented by blocking Src activity [95]. Thus, it was concluded that STEP is the endogenous PTP that regulates the function of NMDARs in opposition to Src.

Two additional roles for STEP in the regulation of NMDARs have been elucidated. First, STEP-mediated dephosphorylation suppresses the constitutive trafficking of NMDARs, leading to a decrease in NMDAR cell surface expression [10,114]. Second, STEP has been found to dephosphorylate Fyn, reducing its activity [90] and possibly indirectly and directly suppressing tyrosine phosphorylation of NMDARs.

7.4.1. Src Upregulation of NMDA Receptors in LTP at CA1 Synapses

SFKs have been implicated from physiological and pharmacological approaches as critical for the induction of LTP in CA1 [32,73]. PTKs were first implicated on the basis that tetanus-induced LTP in CA1 neurons was prevented by broad spectrum inhibitors that did not alter preexisting potentiation, implying involvement of PTKs in induction rather than in maintenance of LTP. Upregulation of NMDAR activity by Src is required for the induction of LTP at Schaffer collateral CA1 synapses in the hippocampus [40,73,95]. Intracellular administration of Src40–58 or anti-Src1 directly into postsynaptic neurons by means of a patch pipette electrode prevents LTP induction in CA1 neurons. Because Src40–58 and anti-Src1 prevent the upregulation of NMDAR activity by endogenous Src but do not affect excitatory synaptic transmission [40,73,150], the most parsimonious explanation for the suppression of LTP induction is that these reagents disrupt the Src–ND2 interactions at synaptic NMDARs. Thus, it was inferred that the interaction between the Src unique domain and ND2 is essential for induction of LTP at CA1 synapses.

The SFK activator pYEEI peptide increased synaptic AMPAR responses, an increase prevented by Src40–58. This increase occludes that produced by tetanus, implying common signaling steps. The pYEEI-induced increase in AMPAR responses is prevented by chelating intracellular Ca²⁺, but this has no effect on pYEEI-induced increases in NMDAR currents. Because blocking NMDARs prevents the potentiation of AMPAR responses by pYEEI, the simplest model is that Src-mediated upregulation of NMDARs is necessary for tetanus-induced LTP in CA1 neurons [3]. Consistent with this model is the finding that the level of phosphorylation of Y1472 of NR2B increases following tetanic stimulation in CA1 hippocampus [89]. Tyrosine phosphorylation of NR2B increased after LTP induction in the dentate gyrus of the hippocampus [101,102].

STEP has also been implicated in the induction of LTP [95]. In hippocampal slices, inhibiting endogenous STEP activity with an inhibitory antibody delivered into CA1 neurons enhanced transmission and occluded LTP induction through a mechanism dependent on NMDARs, Ca²⁺, and Src [95]. Conversely, administering recombinant STEP into CA1 neurons prevented induction of LTP. Neither administering STEP nor inhibiting Src affected basal synaptic transmission or NMDAR currents in CA1 neurons and hence no suppression of NMDARs resulting from these experimental maneuvers might otherwise account for the blockade of LTP induction. STEP does reverse the enhancement of NMDAR currents produced by activating Src. Thus, STEP acts tonically as a brake on Src-mediated synaptic potentiation.

Consistent with the role of Src-mediated upregulation of NMDARs in LTP induction, recent studies implicated PTPα, a well-characterized activator of Src, in LTP [61,96]. Induction of LTP in hippocampal CA1 neurons was prevented by inhibiting endogenous PTPα activity through intracellular application of an inhibitory antibody [61]. LTP induction in CA1 hippocampus was impaired in mice with targeted deletions of PTPα. The impairment was associated with a reduction in phosphorylation levels of Y1472 in the NR2B C tails in the PTPα^−/− mice [96].

Induction of LTP in hippocampal CA1 neurons is prevented by blocking CAKβ using the dominant negative mutant described above [40]. Conversely, administering CAKβ into CA1 neurons produces a lasting enhancement of AMPAR synaptic responses, mimicking and occluding LTP. This CAKβ-stimulated enhancement of synaptic AMPAR responses is prevented by blocking NMDARs, chelating intracellular Ca²⁺, or blocking Src. Synaptic NMDAR currents in CA1 neurons are not tonically upregulated by CAKβ-Src signaling, but CAKβ becomes activated and recruited to Src by stimulation that produces LTP [40,59]. Thus, activation of CAKβ leading to stimulation of Src is essential for the induction of tetanus-evoked LTP.

Figure 7.2a illustrates a simple model for induction of LTP based on the work described above. It is hypothesized that tetanic stimulation rapidly activates CAKβ that associates with and thereby activates Src, allowing tonic suppression of NMDAR function by STEP to be overcome. This kinase-dependent upregulation may be further amplified by a rise in intracellular Na⁺ that occurs during high levels of activity since Src kinases increase NMDAR function and also sensitize the channels to potentiation by intracellular Na⁺ [151]. Coupled with depolarization-induced reduction of Mg²⁺ inhibition, a dramatic boost in the influx of Ca²⁺ through NMDARs sets in motion the down-stream cascade [77] that ultimately results in potentiation of synaptic AMPAR responses by recruiting new AMPARs to the synapse and/or by phosphorylating existing AMPARs.

FIGURE 7.2

(a) Model for the role of Src, STEP, CAKβ, PTPα, Csk, and ND2 in the induction of LTP at synapses, e.g., in CA1 hippocampus. Left: under basal conditions, NMDAR activity is suppressed by partial blockade of the channel by Mg²⁺ and activity (more...)

7.4.2. Kinase and Phosphatase Regulation of NMDA Receptors in Synaptic Metaplasticity

Metaplasticity is defined as regulation of the processes that underlie synaptic plasticity [2]. This can be seen in thresholds for the induction of LTP or long term depression (LTD) that may be influenced by prior activity or conditioning stimuli that alone do not alter the amplitudes of basal EPSCs or the efficacy of synaptic transmission. For NMDAR-dependent LTP or LTD, one mechanism for metaplasticity may alter the gating or expression of synaptic NMDARs. This is precisely what protein kinases and phosphatases do, as described above.

LTP and LTD can be considered in terms of the relationship of stimulation frequency and resultant change in synaptic efficacy (Figure 7.2b). From this perspective, metaplasticity is reflected by a right or left shift in the frequency–plasticity relationship. Thus, it is hypothesized that kinase-induced enhancement of NMDAR function or number results in production of LTP at lower stimulation frequencies. Conversely, when the kinase/phosphatase balance is shifted toward dephosphorylation and suppression of NMDARs, higher stimulation frequencies may be required to elicit LTP. Changes in NMDARs caused by several serine/threonine and tyrosine kinases or phosphatases have already been implicated in producing metaplasticity [75]. Because these regulatory enzymes are convergence points for many signaling pathways involving diverse upstream receptors in various CNS regions, it is anticipated that complex metaplasticity relationships existing in various neuronal types depend on the unique biochemical networks that are active at a given time.

An interesting example of the complexity by which kinase regulation of NMDARs may affect synaptic plasticity comes from recent work on Cdk5 regulation of NMDARs [39]. As noted, the conditional knockout of Cdk5 in the adult mouse brain increases NR2B expression at synapses. In hippocampal slices from these animals, thresholds for induction of LTP were lowered. Consistent with incorporation of functions of NR2B receptors, LTP induction became sensitive to the ifenprodil NR2B-selective blocker. Furthermore, the Cdk5 conditional knockouts showed improved performance in spatial learning tasks, also sensitive to ifenprodil. Thus, the control of synaptic plasticity by kinase/phosphatase regulation of NMDARs appears to depend on the ultimate effect of the particular kinase/phosphatase on the level of function or expression of synaptic NMDARs and may be highly dependent upon the enzymes, their control by cell signaling pathways, and the NMDAR subunits affected.

7.5.1. Pain

Upregulation of NMDARs appears crucial for the initiation and maintenance of the enhanced responsiveness of nociceptive neurons in the dorsal horn of the spinal cord in experimental pain models [146]. In spinal cord slices, peripheral inflammation [34] and nerve injury [45] alter NMDAR-mediated currents in superficial dorsal horn neurons. Peripheral nerve injury increases the amplitude, slows the decay phase of NMDA EPSCs [46], and produces prolonged facilitation of membrane currents and calcium transit induced by bath application of NMDA [45], thus potentiating glutamatergic transmission.

In the dorsal horn, glutamatergic transmission may be potentiated homosynaptically, as in CA1, although the predominant form of enhancement of synaptic transmission is heterosynaptic [147]. As in CA1, NMDARs in dorsal horn neurons are regulated by CAKβ Src signaling balanced by STEP activity in vitro. In vivo, tyrosine phosphorylation of NR2B in the spinal cord increases with models of inflammatory [35,36] and neuropathic pain [1]. Inhibition of SFKs in vivo delays the onset of inflammatory hyperalgesia [106] and inhibition of SFKs, PKC or group I mGluRs prevents the increase in NR2B tyrosine phosphorylation [35,36]. This indicates that a GPCR signaling cascade upstream of SFK-mediated NMDAR upregulation may be required for pain-related maladaptive changes in synaptic transmission. Peripheral nerve injury activates SFKs in lumbar spinal cord [51]. Intrathecal administration of PP2, a nonselective SFK inhibitor, suppresses mechanical hypersensitivity in nerve-injured mice [51], suggesting a role of SFK in neuropathic pain.

Studies of mice with deletions of specific SFK genes indicate that Src [70], Fyn [1], and Lyn [132] are essential for the development of neuropathic pain. Mice lacking each of these genes exhibited deficits in peripheral nerve injury-induced mechanical hypersensitivity. However, the role of these SFKs in neuropathic pain may be different. Spinal cord dorsal horn NR2B phosphorylation induced by peripheral nerve injury is reduced in both Src and Fyn mutant mice, indicating that NMDARs are downstream of Src and Fyn. However, Lyn is predominantly activated in microglia following PNI, and the upregulation of the ionotropic purinoceptor P2X4 in microglia [131] is deficient in Lyn null mutant mice.

As multiple signaling pathways converge on SFKs in synaptic transmission [103], SFK-dependent NMDAR upregulation may also serve as a convergence point in the development and maintenance of chronic pain. For example, activation of EphB in the spinal cord with ephrinB2 resulted in prolonged hyperalgesia [5], while inhibition of EphB reduced chronic inflammatory [5] and neuropathic pain [55]. EphB activation induced phosphorylation of SFKs [5], leading to phosphorylation of NR2B and amplifying NMDAR responses [125]. The convergence of multiple signaling pathways on SFKs allows both homosynaptic and heterosynaptic plasticity in the dorsal horn that is likely mediated through upregulation of NMDARs by these kinases.

Accumulating evidence indicates that PKA- and PKC-mediated NMDAR phosphorylation participates in generation of pain hypersensitivity. Increased phosphorylation of NR1 proteins in spinal dorsal horn neurons was observed in spinal cords of rats following noxious heat [12], capsaicin injection [153,155], formalin injection [53], and peripheral nerve injury [26,27]. This increase in phosphorylation was detected in both a PKC-dependent site (Ser-896) [12] and a PKA-dependent site (Ser-897) [27,53,154]. Pharmacological studies of selective protein kinase inhibitors also suggest that both PKA and PKC are involved in this increased phosphorylation of NR1 following noxious stimulation [154,155]. Thus, serine phosphorylation of NR1 subunits through PKA- and PKC-mediated pathways may contribute to both acute and persistent pain.

7.5.2. Epilepsy

Some forms of epilepsy such as that produced by kindling are also thought to depend on upregulation of NMDAR function [83]. Kindling shares physiological and pharmacological properties with LTP, so they may share common signaling pathways. Targeted disruption of Fyn in mice delayed the induction of kindling [13]. Introduction of native Fyn into Fyn^−/− mice resulted in accelerated kindling [57]. Moreover, transgenic mice expressing a constitutively active form of Fyn showed higher seizure activity [57]. Consistent with this, tyrosine phosphorylation of NR2A and NR2B increased after seizure activity in kainate-induced status epilepticus, another model of epilepsy [43,88]. In an in vitro model of epileptiform activity in the CA3 region of hippocampus, Src activity increased with epileptiform activity and the frequency of epileptiform discharge was reduced by pharmacological blockade of SFKs [105].

7.5.3. Ischemia-Induced Neuronal Cell Death

Excitotoxicity mediated by NMDARs is implicated in neuronal death in many pathological conditions including CNS ischemia, trauma, and neurodegenerative diseases. Studies of SFK signaling in models of cerebral ischemia revealed that transient ischemia induces increases in tyrosine phosphorylation of NR2A and NR2B [21,22,123,124]. This is associated with the recruitment of Src, Fyn, and CAKβ to the PSD [21,123] and with the activation of Src and CAKβ [22]. Y1472 in the NR2B C terminal tail is hyperphosphorylated in postischemic rats. Phosphorylation of Y1472 is reduced by inhibition of SFKs [22]. Furthermore, SFK inhibitors suppress NMDA-evoked excitotoxicity in vitro [38]. These results implicate a Src-mediated pathway and tyrosine phosphorylation of NMDARs in the pathophysiological mechanism of neuronal death caused by ischemia.

Upregulation of NMDARs by the serine/threonine kinase Cdk5 may also contribute to ischemia-induced neuronal cell loss. In rat hippocampal CA1 neurons, forebrain ischemia induced Cdk5-mediated phosphorylation of the NR2A subunit at Ser-1232 [137]. Inhibiting endogenous Cdk5 or perturbing interactions of Cdk5 and NR2A subunits abolished NR2A phosphorylation at Ser-1232 and protected CA1 pyramidal neurons from ischemic insult. Thus, both serine/threonine and tyrosine phosphorylation of NMDARs may be critical in deaths of neurons produced by acute ischemic injury.

7.5.4. Huntington’s Disease

NMDARs are considered to play a role in neuronal loss in several neurodegenerative conditions. The potential involvement of SFK-mediated NMDAR phosphorylation in Huntington’s disease is well investigated. This disease is a progressive neurodegenerative disorder with autosomal-dominant inheritance. The gene for Huntington’s disease encodes the huntingtin protein that has an expanded polyglutamine stretch near the 5′ end of the gene [15].

Expression of polyglutamine-expanded huntingtin in a hippocampal cell line activates Src and increases tyrosine phosphorylation of the NR2B subunit of recombinant NMDARs [115]. Expression of mutant huntingtin sensitizes NMDARs and promotes neuronal death induced by glutamate [120]. Inhibition of SFKs decreases glutamate-induced neuronal death mediated by mutant huntingtin, as does coexpression of a mutant NR2B subunit in which Y1252, Y1336, and Y1472 are substituted to phenylalanine. These results indicate involvement of an SFK-mediated signaling pathway upstream in the NMDAR-dependent degeneration of neurons in Huntington’s disease.

7.5.5. Alzheimer’s Disease

Dysregulation of NMDAR phosphorylation is also implicated in Alzheimer’s disease. A central causative factor of this disease is the accumulation of a small secreted peptide known as amyloid-β [79]. One effect of a toxic fragment of amyloid-β known as amyloid-β 1-42 is activation of the α7 nicotinic acetylcholine receptor resulting in α7-mediated Ca²⁺ influx and activation of calcineurin [114]. PP2B dephosphorylates and activates STEP, which dephosphorylates the NR2B subunit at Tyr-1472 and promotes internalization of NR2B-containing NMDARs. STEP may also depress Fyn and hence the Fyn-mediated phosphorylation of Tyr-1472 [10]. It is hypothesized that high levels of amyloid-β reduce NMDA EPSCs and inhibit synaptic plasticity [114].

7.5.6. Schizophrenia

NMDAR hypofunction is implicated in a number of the behavioral manifestations of schizophrenia—social withdrawal, increased motor stereotypy, cognitive deficits, and locomotor activity—in humans and animal models [84,85,107]. Dysregulation of NMDAR phosphorylation may thus contribute to the etiology of schizophrenia. Schizophrenia has one of the highest heritabilities among neuropsychiatric disorders. In several human association and linkage studies, ErbB4 has been identified as a key risk gene that confers susceptibility to schizophrenia [37,60,91,92], ErbB4 encodes a receptor tyrosine kinase, the ErbB4 receptor, that is expressed in the adult CNS. The cognate ligand for ErbB4 is the peptide neuregulin 1 (NRG1), also strongly linked to schizophrenia in humans [117]. In post-mortem prefrontal cortex tissues of patients with schizophrenia, marked increases of NRG1-induced activation of ErbB4 attributable to increased association of ErbB4 with PSD-95 were observed [37].

This overactivation of ErbB4 by NRG1 suppressed tyrosine phosphorylation of NR2A in human samples. In rodents, NRG1 ErbB4 signaling blocks induction of LTP at CA1 synapses [41,74], likely by suppressing Src-mediated enhancement of NMDARs. These studies lead to the hypothesis that cognitive deficits in schizophrenia may be consequences of hyperfunction of NRG1 ErbB4 signaling, leading to suppressed NMDAR-dependent synaptic plasticity [97].