Protectedsite Phosphorylation In

figure shows several renderings of our model of the PKC catalytic core and C-terminal phosphorylation sites.

Our modeling suggests two potential types of protected phosphorylation sites in the C-terminal autophosphorylation domain of PKC. The first type is exemplified by the T641 phosphorylation site. The phosphate at T641 is immersed in a cleft in the upper lobe of the catalytic core, pointed inward toward the center of the molecule (right-hand panels, white arrows). In addition, T641 sits on the interior of a pronounced angle in the peptide backbone of the adjacent residues, limiting accessibility of the phosphate from the exterior of the molecule. This conformation suggests that the phosphate at T641 is normally well-protected by the catalytic core on one face and also by the adjacent peptide backbone.

Interestingly, the other autophophory-lation site, T634, presents another type of configuration. In our model of the catalytic core of PKC, the phosphate at T634 appears to point outward and be at the surface of the catalytic core. Thus, protection of this type of site would necessitate that other regions of PKC beyond the catalytic core extend over or near the site to limit phosphatase accessibility. Although at present no direct structural information is available, domains potentially involved include both the Ca+2 and lipid-binding domains of PKC (see Figures 2 and 7). In the future, it will be interesting to determine the configuration of these domains relative to the PKC catalytic core.

BOX 2 Dephosphorylation of autophosphorylated PKC by PP1 is accelerated in the presence of PKC activators. (Left-hand panel A) Purified PKC PII (100 ng per lane) was autophosphorylated in vitro, then dephosphorylated with PP1 for the indicated time in the presence of either 210 pM/6 pM PS/DAG (97 mol% and 3 mol%, respectively) and 400 pM Ca+2 (activating conditions, right half), or Triton X-100 (2 mM) mixed micelles containing 35 pM/1 pM PS/DAG (2 mol% and 0.05 mol%, respectively) with 100 pM Ca+2 (nonactivating conditions, left half). (Left-hand panel B) Quantitation of the data. Immuno-reactivity refers to binding of antibody 96160 (see Figure 10) to these samples (blots not shown). Aliquots were removed into SDS-PAGE sample buffer and electrophoresed using 7% SDS-PAGE, transferred to nitrocellulose, and 32P quantitated by phosphoimaging. The right-hand panels are space-filling models of the PKC catalytic domain and associated autophosphorylation sites. Reproduced from Sweatt et al. (46).

BOX 2 Dephosphorylation of autophosphorylated PKC by PP1 is accelerated in the presence of PKC activators. (Left-hand panel A) Purified PKC PII (100 ng per lane) was autophosphorylated in vitro, then dephosphorylated with PP1 for the indicated time in the presence of either 210 pM/6 pM PS/DAG (97 mol% and 3 mol%, respectively) and 400 pM Ca+2 (activating conditions, right half), or Triton X-100 (2 mM) mixed micelles containing 35 pM/1 pM PS/DAG (2 mol% and 0.05 mol%, respectively) with 100 pM Ca+2 (nonactivating conditions, left half). (Left-hand panel B) Quantitation of the data. Immuno-reactivity refers to binding of antibody 96160 (see Figure 10) to these samples (blots not shown). Aliquots were removed into SDS-PAGE sample buffer and electrophoresed using 7% SDS-PAGE, transferred to nitrocellulose, and 32P quantitated by phosphoimaging. The right-hand panels are space-filling models of the PKC catalytic domain and associated autophosphorylation sites. Reproduced from Sweatt et al. (46).

Our data suggest that the protection from phosphatase activity is conformation-dependent. For example, in the presence of its normal activators PKC is readily dephosphorylated (see Box 2 and references 30 and 48). This observation suggests that PKC phosphorylation is reversible under specific conditions, such as might occur when the cell receives a depoten-tiating signal. Specifically, in thinking about regulation of PKC autophosphorylation in depotentiation, we proposed that upon entry of low levels of Ca2+, protein phosphatase 2B is activated, leading to dephosphorylation of protein phosphatase inhibitor 1, thereby causing activation of protein phosphatase 1. The depotentiation-associated Ca2+ signal (in conjunction with other activators, for example transient DAG production) also changes the conformation of PKC, exposing the phosphate at the "protected" site. Protein phosphatase 1 can then dephosphorylate PKC, returning the enzyme to its original state. Alternatively, protein phosphatase 2B could directly dephosphorylate PKC.

Finally, although we have proposed the "protected site" model based on our studies of PKC autophosphorylation in LTP, phosphorylation of protected sites might be a general biochemical mechanism for the generation of stable, long-lasting physiologic changes. This idea is appealing because such a mechanism confers three attributes upon a change in an enzymatic system: stability, constancy, and regulated reversibility. The change is stable in the sense that it is long-lasting in the cell. The magnitude of the change is constant because a constant fraction of the enzyme stays phosphorylated, unless additional stimulation occurs. Finally, although the change can be long-lasting, upon receiving a specific signal the change can be readily reversed, restoring the system to its original state. It will be interesting in the future to determine if protected site phosphorylation is used in other enzyme systems in the generation of persisting cellular signals.

Calpain and PKMg

PKC was originally discovered not as second-messenger-regulated enzyme, but rather as a kinase activated secondary to proteolytic cleavage. Various proteases like trypsin and the calcium-activated protease calpain can clip PKC in its central "hinge" region (see Figure 7), releasing the N-terminal inhibitory domain and liberating the free, active C-terminal catalytic domain. This active fragment of PKC is referred to as PKM.

Of course, this mechanism of proteolytic activation of PKC has great appeal as a potential mechanism for generating a long-lasting signal in LTP. Making a long story short, it turns out that this acute proteolysis of PKC is not a dominant mechanism in NMDA receptor-dependent LTP in area CA1, although it does appear to be involved in NMDA receptor-independent LTP in this same region (34, 36). (We will return to an additional potential target of calpain proteolysis, the cytoskeleton, in the next section.)

However, there is a role for a consti-tutively active PKM as a persisting signal in E-LTP, it is just that the mechanism of its generation is not proteolysis. Todd Sacktor's research group has spent many years tracking down the basis for generation of this persistent signal in LTP, and I will briefly summarize their findings.

Todd's group has shown that a con-stitutively active PKC isoform, the PKM zeta (PKMQ isoform, is synthesized de novo after LTP induction (see reference 54 and Figure 1). PKMZ is a second-messenger-independent, constitutively active fragment of PKCZ that lacks the regulatory domain. This fragment is synthesized from a unique mRNA that codes for the truncated form of the enzyme. An LTP-associated increase in the amount of PKMZ protein lasts at least 2 hours after LTP-inducing tetanus, an effect that is NMDA receptor-dependent (34, 54). Also, inhibitors of PKMZ applied after LTP-inducing stimulation reverse the expression of LTP (55). Interestingly, a decrease of PKMZ is seen after LTD induction in area CA1, which suggests that bi-directional regulation of PKMZ may contribute to potentiation and depression of synaptic transmission in area CA1 (56).

Thus, increased synthesis of a consti-tutively active PKC fragment represents a fourth category of persisting signal in LTP. It is interesting that the seemingly more straightforward mechanism of direct proteolysis of pre-existing PKCZ is not used—perhaps this mechanism has been reserved by evolution for use in other forms of synaptic plasticity (36). Finally, I note that because the formation of PKMZ in LTP is protein-synthesis-dependent, this is not strictly speaking a mechanism for E-LTP as I have defined it. However, I included the example in this section because it is an example of a mechanism for generating an autonomously active kinase, like the other examples presented, and the time course of PKMZ formation is compatible with a role in early stages of LTP.

C. A Final Potential Target of Calcium—Phospholipases

There are a number of calcium-activated phospholipases in neurons that are potential targets of the LTP-inducing calcium signal. These include phospholipases C, D, and A2, which cleave off various parts of membrane phospholipids (summarized in Figure 11). We have already talked about PLC in the context of PKC activation in LTP

FIGURE 11 Sites of cleavage of membrane phospholipids by phospholipases. The left-hand panel illustrates bonds that are hydrolyzed by phospholipases A1, A2, C, and D. Note that each cleavage is a hydrolysis reaction, leaving free hydroxyl (OH) groups at the cleavage site for each of the two products. PLA1 and PLA2 liberate a free fatty acid (FA, see lower right panel) and a lyso-phospholipid. PLC liberates diacyl glycerol (DAG) and a phosphorylated head group (see upper right panel). PLD liberates phosphatidic acid (PA) and the free, hydroxy-lated head group.

FIGURE 11 Sites of cleavage of membrane phospholipids by phospholipases. The left-hand panel illustrates bonds that are hydrolyzed by phospholipases A1, A2, C, and D. Note that each cleavage is a hydrolysis reaction, leaving free hydroxyl (OH) groups at the cleavage site for each of the two products. PLA1 and PLA2 liberate a free fatty acid (FA, see lower right panel) and a lyso-phospholipid. PLC liberates diacyl glycerol (DAG) and a phosphorylated head group (see upper right panel). PLD liberates phosphatidic acid (PA) and the free, hydroxy-lated head group.

induction, but what about phospholipases in LTP maintenance? Phospholipases are noteworthy in this context because they can generate membrane-permeant compounds like arachidonic acid (AA) and DAG, which if persistently produced could serve as retrograde signaling compounds in order to maintain changes in the presynaptic compartment.

One specific mechanism that has been proposed in this context is the persistent generation of AA by postsynaptic PLA2 (57). Tim Bliss's group has published evidence that there is a persisting increase in AA after LTP-inducing stimulation in vivo and proposed that this might serve as a persisting potentiating signal in E-LTP. Potential targets of AA are numerous. For example, AA can activate PKC, and this could serve as a presynaptic facilitation mechanism. However, AA can also be converted to a wide variety of active metabolic products by the cyclo-oxygenase pathway (which produces prostaglandins and associated compounds) and the lipoxygenase pathway (which produces active hydroxy peroxy eicosa fetraenoic acid (HPETE) metabolites). Any of a number of these compounds could serve as potentiating signals by binding to cell surface receptors pre- or postsynaptically.

The potential mechanisms for generating a persistent increase in AA or other phospholipase-derived messengers in E-LTP are unknown at present. In addition, some disagreement exists in the literature concerning whether this mechanism is necessary for E-LTP (58). Thus, at present, this mechanism remains more in the category of interesting possibility versus established mechanism.

D. Section Summary: Mechanisms for Generating Persisting Signals in E-LTP

As was emphasized at the beginning of the chapter, the capacity to generate a persisting signal in response to a transient stimulus is the biochemical sine qua non of memory formation. In this section, we have seen several examples of these types of processes that have been proposed to be involved in the maintenance of early stages of LTP (see Table 2). These are the four best-characterized solutions to the problem of making a lasting signal in E-LTP, and these reactions serve as general prototypes for solutions to the problem of neuronal activity-dependent generation of molecular memory traces.

It is a useful exercise to perform a compare-and-contrast concerning these four documented mechanisms for molecular information storage (see Table 2). In the case of CaMKII, current models propose that Ca/CaM stimulation of the enzyme results in self-perpetuating inter-subunit autophosphorylation, which coupled with low phosphatase activity leads to a persisting level of active enzyme in the PSD. The read-out of this persisting signal potentially includes structural changes in

TABLE 2 Proposed Mechanisms for Generating Persisting Signals in E-LTP

Molecule Mechanism Role

CaMKII Self-perpetuating autophosphorylation coupled with Effector phosphorylation, low phosphatase activity Structural changes

Various PKCs Direct, irreversible covalent modification by reactive Effector phosphorylation oxygen species

PKC a/piI Protected-site autophosphorylation resistant to Protection from down-regulation, phosphatase activity Subcellular localization

PKMZ De novo synthesis of a constitutively active kinase Effector phosphorylation the PSD and increased phosphorylation of target effectors. A second example is that various PKCs can undergo direct, irreversible covalent modification by reactive oxygen species. Oxidation of zinc finger domains in the enzyme leads essentially to irreversible activation of the enzyme, which is then free to phosphorylate target effectors. The third example may act in concert with this mechanism. Persistently activated PKC in cells is typically rapidly down-regulated as a homeostatic mechanism. However, PKC autophosphorylation leads to protection from down-regulation. When stimulated by calcium and DAG, PKC a/PII can undergo protected-site autophosphorylation at a site resistant to phosphatases, leading to a persisting signal that can help maintain the constitutively active kinase in the cell. Finally, in the case of PKMZ, de novo synthesis of a con-stitutively active kinase lacking the normal autoinhibitory domain leads to the presence of a persistently activated kinase in the cell. Overall, these are four unique and elegant biochemical solutions to the problem of making a lasting signal in response to a transient signal. In the next section, we will discuss the targets of these persisting signals that lead to enhanced synaptic transmission— the translation of the persisting signal into persisting effects at the synapse.

Brain Battalion

Brain Battalion

Get All The Support And Guidance You Need To Be A Success At Beefing Up Your Brain. This Book Is One Of The Most Valuable Resources In The World When It Comes To Understanding Your Brain to Come Up With A Battle Plan For Intelligence.

Get My Free Ebook


Post a comment