Akap

PKA, PP2B

Kinase and phosphatase localization

CaMKII

Signal transduction

Regulate likelihood of LTP induction

This consideration also serves as an important caveat for interpreting results from NMDA receptor knockout mice. This apparently "clean" experimental manipulation, wherein the NMDA receptor is entirely lost, likely results in a large number of secondary effects on molecules normally associated with the NMDA receptor postsynaptically. In fact, experiments using various deletion mutants missing the cytoplasmic anchoring domains of the NMDA receptor have allowed dissection of the role of the NMDA receptor as a scaffolding protein versus its role as a ligand-gated ion channel (50). Deletion of the intracellular domain of the NMDA receptor appears to be sufficient to account for essentially all the physiologic and behavioral deficits observed in NMDA receptor knockout mice—the upshot is that the role of the NMDA receptor as a component of the PSD infrastructure is just as important as its role as a ligand-gated ion channel.

Additional Direct Interactions with the NMDA Receptor

The NMDA receptor NR1 and NR2 sub-units also bind spectrin, the actin-binding protein. This may serve as an additional cytoskeleton anchoring site postsynap-tically. Moreover, this interaction is subject to regulation by phosphorylation— tyrosine phosphorylation of NR2B leads to decreased interactions of spectrin with the receptor, and NR1 interaction with spectrin is modulated by serine/threonine phos-phorylation. However, the role of these effects in synaptic plasticity is not clear at this point.

Finally, as described earlier in the section on direct modulation of NMDA receptors, the scaffolding protein RACK1 promotes formation of a fyn/RACK1/NR2B complex that actually inhibits fyn phosphorylation of the NMDA receptor and diminishes current through the receptor (see Figure 2). Also, PDS-95 modulates src phosphory-lation of NMDARs, and src potentiation of

NMDAR currents appears to require the presence of PSD-95.

Consideration of the complicated structure and regulation of the postsynaptic density complex highlights the importance of thinking of the entire postsynaptic domain as a large functional unit. The NMDA receptor is embedded in a dynamic multiprotein complex that it regulates and in turn that regulates it. While many details of the structural components of the PSD complex are still being worked out and their roles in LTP induction are being actively investigated, it is clear that disrupting one or more of the cogs in this machine can lead to disruption of the proper function of the NMDA receptor.

AMPA Receptors

AMPA receptors, of course, provide the initial depolarization, either locally or distally in the neuron, that ultimately results in NMDA receptor activation. As such, alterations in the AMPA receptor protein or its associated interacting proteins can lead to loss of proper regulation of NMDA receptor activation. However, in this section we will focus on the AMPA receptor as a structural component of the synapse.

AMPA receptors, like NMDA receptor, also reside postsynaptically but are in much more of a state of flux than NMDA receptors. In fact, the average half-life for an AMPA receptor in the postsynaptic membrane is 15 minutes. We will return to some implications of this in the next chapter. Also, as we have discussed in previous chapters, AMPA receptor membrane insertion can be activity-dependent. Thus, the AMPA receptor should probably not be thought of like the NMDA receptor—the NMDA receptor likely serves a frankly structural role in addition to its function as a ligand-gated ion channel while the AMPA receptor is more peripherally associated with the PSD (see reference 51).

The AMPA receptor binds at least two "structural" proteins—Protein Interacting with C Kinase-1 (PICK-1), which binds PKC, and Glutamate Receptor Interacting Protein (GRIP) (see reference 47). GRIP is a multidomain scaffolding protein that likely functions in AMPA receptor trafficking. GRIP also binds to GRIP Associated Protein-1 (GRASP1), a Guanine nucleotide Exchange Factor (GEF) for ras (see references 52 and 53)—the functional role of GRASP1 at the synapse is unclear at present. AMPA receptors can also bind N-Ethyl maleimide Sensitive Factor (NSF), a vesicle-associated protein that may also be involved in receptor membrane insertion is in a fashion reminiscent of its role presynap-tically in neurotransmitter vesicle fusion.

AMPA receptors also bind the A kinase anchoring protein AKAP79, an interaction that appears to be mediated by the PSD-95 homologue SAP-97 (54-56). As the name implies, AKAPs bind and localize PKA by interacting with the regulatory subunits of the kinase. The general role of AKAPs is to help localize PKA near relevant targets such as the AMPA receptor postsynapti-cally. The story is actually more complicated than that, because AKAP79 in the hippocampus also binds and localizes a protein phosphatase, PP2B (aka Calcineurin). As a first approximation, it is useful to think of proteins such as AKAPs as serving a role to increase the signal-to-noise ratio for signal transduction—localizing kinases close to their substrates to increase the efficacy of phosphorylation, but also localizing phosphatases to those same substrates in order to keep their basal phosphory-lation low and to allow for rapid reversal of phosphorylation events once the kinase activation is over (56). AKAP79 may also serve specifically to localize the calcium-sensitive phosphatase PP2B to the AMPA receptor in order to facilitate calcium-dependent AMPA receptor dephosphorylation and down-regulation in LTD (57).

AKAP79 also serves an additional scaffolding protein function. It binds to cyclase-coupled receptors such as the beta-adrenergic receptor, localizing recep tor, effector, kinase, substrate, and phos-phatase all together in a supramolecular complex. In the context of the hippocampal pyramidal neuron synapse, this might allow for enhanced beta adrenergic receptor modulation of AMPA receptor function, enhancing AMPA receptor function via PKA-dependent phosphory-lation. This might serve in the induction of LTP as a mechanism whereby cyclase-coupled receptors can augment AMPA receptor-mediated membrane depolarization and indirectly augment NMDA receptor activation.

AKAP79 also binds to PKC, again localizing this kinase near its substrate, the AMPA receptor. By analogy to the scenario outlined earlier for AKAP/PKA, this scaffolding activity might facilitate PKC-coupled receptor augmentation of AMPA receptor function during the induction of LTP as well. As we will discuss in the next chapter, PKC phosphorylation of AMPA receptors also contributes to E-LTP expression, and of course AKAP79 localization of PKC near AMPA receptors would help facilitate this mechanism as well.

CaMKII

CaMKII is highly enriched at the post-synaptic density complex. This enrichment in part occurs through CaMKII binding to the actin cytoskeleton, and the anchor for the cytoskeleton is the NMDA receptor as we have discussed extensively. Thus, one purpose of the NMDA receptor/PSD-95/ cytoskeleton complex is to help localize CaMKII to the PSD domain. This keeps a critical effector of the NMDA receptor, CaMKII, tightly bound and localized for effective responsiveness to NMDA receptor activation. Interaction of CaMKII with the PSD also can be regulated by CaMKII autophosphorylation—John Lisman has proposed this as a mechanism contributing to the maintenance of E-LTP, a model that we will return to in the next chapter.

While the various scaffolding proteins— PSD-95 and the like—that we discussed ealier are involved upstream of the NMDA receptor, regulating its function, CaMKII is downstream of the NMDA receptor. However, I list it as a component of the synaptic infrastructure necessary for proper NMDA receptor function because it is such an important and direct target of the NMDA receptor—in essence loss of CaMKII function may functionally translate as equivalent to loss of NMDA receptor function. In addition, CaMKII binding to the PSD complex may play a structural role in concert with the actin cytoskeleton to serve as part of the infrastructure necessary for the NMDA receptor to function appropriately.

V. LTP INDUCTION COMPONENT 4—FEED-FORWARD AND FEEDBACK MECHANISMS THAT REGULATE THE LEVEL OF CALCIUM ATTAINED

There is a clear consensus that elevation of postsynaptic calcium is necessary for LTP induction, so clearly any process that modulates the postsynaptic calcium level has the capacity to affect LTP induction. Unfortunately that's about where the clarity ends. In this section, we will deal with some of the known processes whereby calcium levels in the postsynaptic spine are regulated. (See table for summary.)

In one sense, dendritic spines are specialized calcium-handling compartments (see Sabatini, Oertner, and Svoboda

(58) for a very nice treatment of this idea). They contain many molecules dedicated to calcium handling that affect the kinetics of calcium elevation, kinetics which are a critical determinant for (a) whether synaptic strength is changed and (b) whether LTP or LTD is induced. A few generalizations can be made in this context based on published work in this area. One, calcium elevation in a spine is largely compartmentalized to that spine—the narrow spine neck greatly limits calcium diffusion out of the spine in response to NMDA receptor activation, for example. Two, the level of calcium attained determines whether synaptic strength goes up or down—modest levels yield synaptic depression (LTD or depotentiation) and higher levels yield LTP. Three, the kinetics of calcium entry secondary to NMDA receptor activation versus back-propagating action potential/VDCC-dependent calcium influx are different—NMDA receptor activation gives a much longer-lasting elevation of spine calcium than does activation of VDCCs. Four, a seminal finding from Rob Malenka's lab made clear that a relatively prolonged (> 2 seconds) elevation of postsynaptic calcium is necessary for LTP induction in CA1 pyramidal neurons. The basic idea is that this prolonged calcium elevation is necessary to trigger the biochemical processes subserving LTP maintenance (which we will discuss in the next chapter). With these four principles in mind it becomes quite clear that regulating the kinetics of calcium handling and the steady-state level of

Table 4 Calcium Feedback and Feed-Forward Mechanisms

Molecule/Organelle Role Modulator/Regulator

VDCCs Augment NMDA receptor-dependent PKA

Ca influx Ca influx due to bpAPs Regulate ERK activation

Endoplasmic reticulum Ca efflux from ER, limit LTP? PLC-coupled receptors

(Ca-ATPase/IP3R/RyR)

Presynaptic mitochondria Regulate presynaptic Ca levels Unknown calcium achieved locally in the dendritic spine is going to be a critical component of LTP induction. What is not clear are the exact molecular processes that impinge upon these variables.

In this section, I will briefly overview two postsynaptic processes and one presy-naptic process that are involved in synaptic calcium handling and describe some of the available literature investigating the role of these processes in LTP induction. The specific systems I will describe are postsy-naptic Voltage-dependent Calcium Channels (VDCCs), the postsynaptic endoplasmic reticulum (ER), and presynaptic mitochondria. The precise roles of these three systems/processes in LTP induction are quite murky at present, and many more years of work are likely to be necessary before a clear picture emerges concerning exactly what is happening with these molecules and organelles during LTP induction. However, a number of studies using inhibitors and knockouts of various components of these systems have been published, and a brief review of these studies is appropriate in order to set the stage for thinking about this category of molecular mechanisms.

A. VDCCs

The role of VDCCs in LTP induction is difficult to study because VDCCs are necessary for the synapse to function at all—presynaptic VDCCs are what allow the calcium influx necessary for neurotrans-mitter release. However, the presynaptic channels involved in release are largely "N" and "P"-type VDCCs (59), and "L"-type (aka high-voltage activated) VDCCs appear to be the principal dendritic VDCCs, at least as far as LTP induction is concerned. Thus, one can use L-type VDCC antagonists such as nifedipine and nitrendipine to try to dissect out the contributions of VDCCs versus NMDA receptors in triggering LTP induction.

A wide variety of studies have made it clear that L-type VDCCs are necessary for NMDA receptor-independent LTP at CA1

synapses—LTP induced, for example, by 200-Hz stimulation or K channel blockade. This necessity for VDCC function for these forms of LTP helped establish that this type of LTP is indeed distinct from NMDA receptor-dependent LTP (60, 61). Also, based on these studies, it is clear that calcium influx through VDCCs is sufficient to cause synaptic potentiation. However, are VDCCs necessary for NMDA receptor-dependent LTP as well? Might calcium influx through VDCCs augment the calcium influx that occurs via NMDA receptors? This is clearly a possibility given that Ito et al. (62) have found that theta-stimulation-induced LTP is significantly attenuated by blockade of VDCCs. (100 Hz HFS-induced LTP is generally held to be independent of VDCCs.) In addition, Dudek and Fields (63) also found that VDCC activation occurs with theta-type stimulation, and that downstream activation of the ERK cascade is dependent upon calcium influx through VDCCs.

One potential mechanism for this effect is activation of VDCCs by back-propagating action potentials (63, 64). Thus, calcium influx due to back-propagating action potentials might sum with calcium coming in via NMDA receptors, augmenting the postsynaptic calcium signal. Additional considerations involving K channel inacti-vation also apply in this scenario (see reference 64), which also serves to augment membrane depolarization and calcium influx.

This process might also be subject to neuromodulation as well. An early collaborative study by Chetkovich, Gray, Johnston, and Sweatt (65) showed that postsynaptic VDCCs are directly up-regulated by the cAMP/PKA cascade. This would allow the opportunity for adenylyl cyclase-coupled receptors, such as beta-adrenergic receptors, to modulate VDCCs by this mechanism as well. As we discussed in the last chapter, coincidence detection mechanisms such as this appear to be particularly important for theta-type LTP induction in area CA1.

B. The Spine Apparatus

The postsynaptic spine has a specialized form of endoplasmic reticulum referred to as the spine apparatus. The spine apparatus performs like the ER in other parts of the cell, as an intracellular calcium store. The spine apparatus has at least three types of calcium channels in it. The first is the Ca ATPase that pumps calcium out of the cytoplasm and into the spine apparatus. The second is the ryanodine receptor (RyR), which is a calcium-gated calcium channel that allows calcium out of the ER. The RyR functions in calcium-induced calcium release (CICR). The third channel is the inositol tris-phosphate (IP3)-gated calcium channel that responds to the second messenger IP3 to cause calcium efflux. IP3 receptors and RyR generally act in concert—IP3 triggers local calcium release, which activates RyR (66). You also may recall that we have seen the IP3 receptor already in this chapter—it is one of the proteins that interacts with the mGluR anchoring protein HOMER, localizing these receptors near the ER and vice versa. It also is clear that calcium influx through NMDA receptors can trigger secondary CICR from the spine apparatus (67).

The literature on the role of the spine apparatus in LTP could at best be considered "messy." Depletion of spine apparatus calcium with the Ca ATPase inhibitor thapsigargin blocks the induction of LTP, although this effect may be limited to modest tetanic stimulation protocols (68, 69). This suggests that mobilization of intracellular calcium from the spine apparatus is necessary for LTP induction under some conditions. However, additional studies with inhibitors/knockouts of the RyR and the IP3 receptor have yielded ambiguous results. Various groups have reported that loss of RyR3 receptor function leads to an attenuation of LTP (70, 71), no effect on LTP (72), or an augmentation of LTP (73). Genetic deletion of IP3 receptors leads to a complex phenotype—generally not affecting LTP induced with HFS, but augmenting lower-frequency stimulation such that stimuli normally causing LTD now elicit LTP (72, 74). In general, it is difficult to come up with a model for the role of the spine apparatus in LTP induction at this point. It seems as if the spine apparatus may somehow serve to limit LTP induction, perhaps by biasing the synapse toward depotentiating mechanisms. However, investigations of these mechanisms is clearly at a very early stage, and the studies are made more difficult by the complex interacting machinery of the spine apparatus.

C. Mitochondrial Calcium-Handling

Mitochondria are generally not found in the postsynaptic spine but rather are restricted in distribution to the presynaptic terminal and postsynaptic dendritic shaft. One of the many roles of mitochondria is in presynaptic calcium handling—they can serve as calcium buffers by taking up cytoplasmic calcium. Michael Levy in my laboratory, in collaboration with Bill Craigen, has been investigating the potential role of mitochondria in regulating calcium handling, and thus synaptic plasticity at Schaffer-collateral synapses (75). While these studies are at an early stage, Michael would kill me if I didn't take the opportunity in this section of the book to present a synopsis of his interesting findings in this area (76).

The mitochondrial outer membrane has within it one of three isoforms of a family of porin proteins, known as the voltage-dependent anion channels (VDACs). Mitochondrial porins conduct small molecules and constitute one component of the permeability transition pore that opens in response to mitochondrial membrane depolarization, such as occurs with cyto-plasmic calcium elevation. Because mito-chondrial porins have significant roles in diverse cellular processes including regulation of mitochondrial ATP and calcium flux, my colleagues Michael Levy, Bill Craigen, and Ed Weeber sought to determine their importance in learning and synaptic plasticity using knockout mice. They found that fear conditioning and Morris water maze spatial learning are disrupted in VDAC1- and VDAC3-deficient mice. They also found that LTP induction was similarly blocked in these mice, and moreover that acute inhibition of the mitochondrial permeability transition pore by cyclosporin A in wild-type hippocampal slices reproduces the electrophysiological phenotype of VDAC-deficient mice. All these effects occurred in the absence of gross disruptions of baseline synaptic transmission. These results demonstrate a dynamic functional role for mitochondrial porins and the permeability transition pore in learning and synaptic plasticity. Our current working hypothesis is that presynaptic mitochondria contribute to regulating acute and baseline presynaptic calcium levels, selectively affecting calcium handling and pre-synaptic neurotransmitter release during periods of high-frequency synaptic activity.

VI. LTP INDUCTION COMPONENT 5—EXTRINSIC SIGNALS THAT REGULATE THE RESPONSE TO THE CALCIUM INFLUX

This is one of the most fascinating areas of investigation into the biochemistry of LTP induction. Rich in associative mechanisms, this category is more widely studied than some of the other areas we have been discussing. We will be discussing two signal transduction cassettes that can serve to "gate" LTP induction. The first example will draw from PKA regulation of protein phosphatase activity—a system referred to as the "cAMP Gate" for LTP induction. Strong evidence exists in the literature that this is an important mechanism for regulating the likelihood of LTP induction at Schaffer collateral synapses. The second example is the PKC/Neurogranin system— a postsynaptic system for regulating the level of free calmodulin, and, hence, for regulating calmodulin-responsive enzymes. The role for this system in regulating the likelihood of LTP induction is slightly more speculative than the cAMP gate, but it still is based on a substantial body of literature.

A. The cAMP Gate for LTP Induction

Bob Blitzer, Ravi Iyengar, and Manny Landau were among the first scientists to truly appreciate the biochemical complexity of LTP induction. They have proposed and investigated a model for regulating the likelihood of LTP induction that they have termed the cAMP gate (77). I find this neologism very appealing—it succinctly captures the essential function of a somewhat complex mechanism for augmenting protein kinase activation in response to the initial elevations of calcium with LTP-inducing stimulation.

When considering protein phospho-rylation, most neuroscientists think first of protein kinases, the "on" switches, relegating protein phosphatases to the subordinate role of turning enzymes back "off"

Table 5 Extrinsic Signals Modulating the Calcium Response

Regulatory System Molecules Involved Role

The cAMP gate PKA/PP1/I1/PP2B Phosphatase inhibition

Augmented kinase signaling

The PKC/neurogranin system PLC/PKC/neurogranin/CaM Augmenting CaMKII activation

Augmenting Ca-sensitive cyclase after their job is completed. However, protein phosphatases play a dynamic and important role in regulating synaptic plasticity and triggering memory formation (see reference 78). A new understanding is emerging of how protein phosphatases are quite active participants in regulating neuronal function, and that a dynamic interplay occurs between phosphatases and kinases in order to set thresholds determining whether a given neuronal input will be able to trigger a long-lasting neuronal change. In the next two paragraphs, I will illustrate this concept with a brief description of the components of the cAMP gate—it likely will be helpful to refer to Figure 5 in envisioning how the cAMP gate works.

As has been described in more detail by Blitzer et al. (77), there is an interesting interplay of the calcium-sensitive phos-phatase, calcineurin, and PKA in controlling the activity of another protein phosphatase, PP1 (protein phosphatase 1). The activity of PP1 is regulated by an inhibitory protein, inhibitor 1 (I1); and only the phosphorylated version of I1 is effective at inhibiting PP1. Thus, the capacity of I1 to block PP1 activity is regulated by PKA-dependent phosphorylation of I1-PKA through phosphorylation of I1 leads to PP1 inhibition. Calcineurin-dependent dephos-phorylation of I1 leads to increased PP1 activity by relieving the I1-dependent inhibition. Thus, calcineurin counteracts the effects of PKA, indirectly activating PP1 through I1 dephosphorylation. The net result is that elevation of cAMP levels and activation of PKA leads to phosphatase inhibition—a mechanism for the PKA

FIGURE 5 Model for the cAMP gate. The model illustrates an interaction of calcineurin and the cAMP-dependent protein kinase (PKA) in regulating the phosphorylation of proteins important for setting a threshold for triggering long-lasting synaptic effects. Blocking calcineurin leads to protein phosphatase 1 (PP1) inhibition indirectly through increasing inhibitor 1 (I1) phosphorylation. Loss of calcineurin and PP1 activity leads to a shift in the balance of substrate protein phosphorylation (receptors, channels, etc.). Increased phosphorylation of these effectors changes synaptic function in such a way that there is an increased likelihood of reaching a threshold for triggering lasting effects. Reproduced from Sweatt (93).

FIGURE 5 Model for the cAMP gate. The model illustrates an interaction of calcineurin and the cAMP-dependent protein kinase (PKA) in regulating the phosphorylation of proteins important for setting a threshold for triggering long-lasting synaptic effects. Blocking calcineurin leads to protein phosphatase 1 (PP1) inhibition indirectly through increasing inhibitor 1 (I1) phosphorylation. Loss of calcineurin and PP1 activity leads to a shift in the balance of substrate protein phosphorylation (receptors, channels, etc.). Increased phosphorylation of these effectors changes synaptic function in such a way that there is an increased likelihood of reaching a threshold for triggering lasting effects. Reproduced from Sweatt (93).

cascade to amplify the activity of any protein kinase that is activated with LTP-inducing stimulation.

These known regulatory mechanisms lead to a model for how PKA controls protein dephosphorylation and the triggering of synaptic plasticity (see Figure 5). PKA, in part, regulates the induction of long-lasting changes by inhibiting activity, enhancing the phosphorylation of key enzymes whose activity is necessary for triggering and maintaining synaptic poten-tiation. In essence, in this model calcineurin and PP1 act like a brake on the formation of synaptic potentiation, and PKA relieves this breaking mechanism. Moreover, the mechanism can serve as the basis for a biochemical coincidence detector—simultaneous activation of the PKA cascade with other protein kinases can lead to signal amplification and the triggering of unique phosphorylation-dependent events.

Bob Blitzer and his colleagues have tested key aspects of this model for regulating LTP induction and have developed a substantial body of evidence in accordance with their hypothesis (see 1, 77, 79). Isabel Mansuy, Danny Winder, and Eric Kandel have also studied certain aspects of the role or protein phosphate regulation in synaptic plasticity in vivo and in learned behavior in the intact animal (80, 81). These latter investigators have found that genetically engineered, calcineurin-inhibited animals exhibit an increased likelihood of triggering robust long-term potentiation in several hip-pocampal subregions including area CA1— a test of a key prediction of the cAMP gate model. In an interesting additional experiment, Malleret et al. (81) showed that calcineurin blockade-dependent augmentation of long-term potentiation in hip-pocampal slices was blocked by blocking the cAMP-dependent protein kinase, again in concordance with the cAMP gate model.

They also found that calcineurin inhibition enhanced the duration of object recognition in learning tasks in vivo—in effect calcineurin inhibition led to memory improvement. They found significant effects for tasks involving memories of minutes-to-hours duration, and also effects in a variation of the procedure that measures longer-term memory that lasts for one week. These data extend the relevance of the cAMP gate into the behaving animal.

Of course, a key question concerns the identity of the substrate proteins whose phosphorylation is augmented by the cAMP gate. Blitzer et al. (77), have focused on CaMKII as a target of the gate, and this certainly is an important target as we will discuss at the end of this chapter and in the next chapter. Additional candidates at this point include protein kinase C, voltage-dependent potassium channels, glutamate receptors and their associated proteins, and any of a number of components of the ERK MAP kinase cascade as we discussed in earlier sections of this chapter. In short, any phosphorylation event involved in LTP induction is subject to regulation by this mechanism, making it a pluripotent system for controlling synaptic plasticity.

B. The PLC/PKC/Neurogranin System

The capacity of extrinsic synaptic signals to enhance Ca-triggered events postsynap-tically is not limited to the PKA system. PKC, acting through the calmodulin (CaM)-binding protein neurogranin (NG), can also lead to enhancement of the activation of Ca/CaM responsive enzymes. (Neurogranin was originally identified by subtractive cloning and termed RC3, and it is also referred to as P17. I use "neurogranin" because that is the name used most frequently in the literature; (see reference 82). Relevant targets of this system include both CaMKII and the Ca/CaM sensitive forms of adenylyl cyclase that are known to be important for LTP induction. The concept is the same as the cAMP gate— amplification of calcium signals through regulation of its target effectors. However, the PKC/Neurogranin system resides upstream of Ca/CaM-sensitive enzymes, increasing their activation by increasing the level of postsynaptic free CaM. This is in contrast to the cAMP gate, which amplifies downstream effects by enhancing substrate phosphorylation.

How does the neurogranin gate work? Neurogranin is a low-molecular-weight member of the calpacitin family of proteins. NG is localized to the postsynaptic cell and in fact is one of those proteins, like CaMKII, whose mRNA is selectively targeted to the dendritic region. NG functions as a calmodulin-binding protein and binds calmodulin in the absence of calcium (see Figure 6). The notable attribute of NG is that it is a PKC substrate, and when phosphorylated by PKC, it is unable to bind CaM. Thus, PKC phosphorylation of NG regulates the postsynaptic level of free calmodulin. PKC activation leads to inhibition of CaM binding to its NG localization scaffold, increasing the concentration of free CaM postsynaptically. This free CaM is then able to respond to calcium signals— PKC acting via NG can amplify Ca/CaM signaling postsynaptically.

Principal workers in figuring all this out have been Dan Gerendasy and Gregor Sutcliffe, Pierre DeGraan, Freesia and Kuo-Ping Huang, Dan Storm, and Eric Klann and Shu-Jen Chen from my lab. Various experiments by these investigators and others have demonstrated that PKC phosphorylation of NG occurs with LTP-inducing stimulation (83, 84), that PKC phosphorylates NG and regulates CaM binding and CaM level (85), that inhibition of NG function affects LTP induction (81, 86, 87), and that PKC/CaMKII crosstalk occurs in LTP induction (88). I refer the reader to an excellent comprehensive review by Dan Gerendasy and Gregor Sutcliffe for further details and references (82).

The upshot of all this is that the PKC/ NG system can serve as a modulator of LTP induction. Moreover, the system in concert with a calcium signal can serve as a coincidence detector—a modest calcium signal coupled with PKC activation and

Metabotropic

Metabotropic

FIGURE 6 PKC phosphorylation of neurogranin regulates the level of free calmodulin. Neurogranin binds calmodulin in the absence of calcium. However, when PKC phosphorylates neurogranin, it is unable to bind calmodulin, freeing it to respond to calcium signals. See further explanation in text.

enhanced levels of free CaM will elicit a response not attainable by either signal in isolation.

Finally, it is very important to note that the cAMP gate and the PKC/neurogranin system do not necessarily operate in isolation. As shown in Figure 7, it is quite straightforward to interdigitate these two systems in a mutually reinforcing cascade (86). The figure shows that an initial calcium signal through the NMDA receptor can be reinforced by a PKC signal (the PKC neurogranin gate) and a cAMP signal (the cAMP gate) to give an augmented CaMKII stimulation. In essence, the CaMKII activation is amplified on the front end by the PKC/NG gate and on the output side by the cAMP gate. The cAMP Gate and the PKC/NG gate also can interact with each other in a mutually reinforcing fashion.

Cooperative, amplified systems such as this are the hallmark of biochemical cascades set up to serve as step-functions— allowing for an essentially all-or-none response after the appropriate preconditions

FIGURE 7 PKC/neurogranin system operates in concert with the cAMP gate. This diagram presents a schematic of the effects of the PKC/neurogranin system and the cAMP gate when considered together. CamKII activity is robustly activated by the simultaneous action of a cAMP coupled neurotransmitter, a DAG coupled neurotransmitter and an active NMDA receptor. See further discussion in text.

FIGURE 7 PKC/neurogranin system operates in concert with the cAMP gate. This diagram presents a schematic of the effects of the PKC/neurogranin system and the cAMP gate when considered together. CamKII activity is robustly activated by the simultaneous action of a cAMP coupled neurotransmitter, a DAG coupled neurotransmitter and an active NMDA receptor. See further discussion in text.

are met. In the case of the example shown in Figure 7, the simultaneous presence of a cAMP-coupled neurotransmitter, a DAG-coupled neurotransmitter, and an active NMDA receptor (depolarization plus glutamate) would give a robust output in terms of CaMKII activation. This, of course, is four-way coincidence detection— precisely the type of biochemical information processing capacity that is necessary for CA1 pyramidal neurons to trigger long-lasting changes in synaptic strength in response to multimodal sensory inputs as we discussed in Chapters 3 and 5. We will proceed to other examples of this type of multiple-input coincidence detection later.

With respect to Figure 7, keep in mind that while it is a static figure, the system may operate as a temporal integrator as well. An initial calcium signal may activate the cAMP gate and the PKC/NG system, allowing for enhanced responsiveness to a subsequent calcium signal when it arrives.

In fact, data from the Huang's laboratories investigating neurogranin knockout mice suggest directly that this is a role for this system—the augmentation of potentiating responses with repetitive stimulation (87, 89).

VII. LTP INDUCTION COMPONENT 6—THE

Was this article helpful?

0 0
How To Win Your War Against Anxiety Disorders

How To Win Your War Against Anxiety Disorders

Tips And Tricks For Relieving Anxiety... Fast Everyone feels anxious sometimes. Whether work is getting to us or we're simply having hard time managing all that we have to do, we can feel overwhelmed and worried that we might not be able to manage it all. When these feelings hit, we don't have to suffer. By taking some simple steps, you can begin to create a calmer attitude, one that not only helps you feel better, but one that allows you the chance to make better decisions about what you need to do next.

Get My Free Ebook


Post a comment