Signaling Mechanisms

We now turn our attention to discussing the mechanisms by which L-LTP-associated changes in gene expression are achieved. How does the L-LTP-triggering calcium elevation at the synaptic spine get a signal to the nucleus? In the following section, I will summarize a variety of studies investigating this question. Leading groups in this area are the laboratories of Dan Storm, Dick Tsien, Eric Kandel, Susumu Tonegawa, Alcino Silva, Tim Bliss, and Jocelyn Caboche—the following summary draws extensively from the findings, thoughts, and models from these various groups as well as work from my own lab.

By way of introduction let me point out that what we are talking about for the remainder of this chapter is a complex multistage process. Of course, the core signal transduction component is calcium activation of kinase cascades that phos-phorylate transcription factors and thereby regulate gene expression. However, many other signals get integrated into this basic process. There are pathways controlling nuclear translocation of transcription factors and kinases. "Transcription factors" should really be thought of as multiprotein complexes that integrate a variety of signals in order to compute whether gene expression should be altered. A single gene can integrate the output of multiple transcrip-tional regulatory elements. After the gene is transcribed, mRNA processing and trafficking mechanisms are subject to additional layers of regulation. Translation of the mRNA into protein is subject to regulation as we discussed in the last chapter. In L-LTP, there are likely multiple waves of altered gene expression, in part because one category of regulated genes codes for transcription factors, which when expressed can lead to secondary alterations in gene expression. There are mysterious but clearly extant mechanisms for temporal and spatial integration of signals in the nucleus. There are mechanisms for controlling gene expression that involve relief of inhibitory constraints, such as histone acetylation. Transcriptional repressors also impinge upon mechanisms for expression of target genes, and these are themselves subject to regulation.

Clearly a detailed description of all these processes is beyond the scope of a single book chapter. What I will do here is distill the essential processes into a working model of how gene expression is regulated in L-LTP, limiting myself to those processes where there are direct experimental data linking them to activity-dependent synaptic plasticity in area CA1 and the dentate gyrus.

It may be helpful to break down L-LTP-related mechanisms for regulating gene expression into the following basic components of L-LTP to help organize your thinking:

1. A core signal transduction cascade linking calcium to the transcription factor CREB

2. Modulatory influences that impinge upon this cascade

3. Additional transcription factors besides CREB that may be involved

4. Genes targeted in L-LTP

5. mRNA targeting and transport

Of course, the altered mRNA expression must be converted into a physiologic read-out by some mechanism, giving us a more hypothetical sixth category:

6. Effects of the gene products on synaptic structure

I will organize the remainder of the chapter around these six topics.

A. A Core Signal Transduction Cascade Linking Calcium to the Transcription Factor CREB

Introduction to Qene Transcription

Genes are stretches of DNA that have the capacity to code for a functional protein. Transcription of the DNA into a protein-encoding mRNA begins with a transcriptional promoter complex binding to a TATA sequence (TATA box) in the DNA sequence, which promotes association of the RNA polymerase II complex with the DNA and transcription at the transcription start site (see references 15 and 16 and Figure 2). This transcription machinery is regulated by transcription factors that bind to upstream regulatory elements (REs), that is, DNA sequences that the transcription factors specifically recognize. The transcription factors and their associated co-activators bind to their REs and enable gene transcription of the downstream target gene. Transcription factor activity is regulated by a variety of post-translational processes including phosphorylation, redox state, ubiquitination, and degradation of associated inhibitory proteins. Co-activators whose activity is typically necessary for the transcription factor per se to be active usually bind to the promoter complex and/or acetylate histones locally to free up DNA for transcription.

A diagram of the basic structure of the CREB transcriptional complex is given in Figure 2 for reference purposes. A more realistic picture of CREB bound to the DNA double helix is given in Figure 3. The CRE (5'-TGACGTCA-3') is the DNA sequence identified by CREB. The activity of CREB is regulated by phosphorylation at Ser 133, which can be phosphorylated by PKA, CaMKII and CaMKIV, and RSK2 (among many others). This phosphorylation event,

Ponse Immunitaire Theileria Annulata

FIGURE 2 The CREB/CRE gene regulation system. Phosphorylated CREB recruits a supramolecular complex to the CyclicAMP response element in DNA, triggering increased expression of downstream target genes. CBP, an accessory to CREB, facilitates gene expression by modulation of RNA polymerase activity and histone acetylation. Adapted from Shaywitz and Greenberg, (16).

FIGURE 2 The CREB/CRE gene regulation system. Phosphorylated CREB recruits a supramolecular complex to the CyclicAMP response element in DNA, triggering increased expression of downstream target genes. CBP, an accessory to CREB, facilitates gene expression by modulation of RNA polymerase activity and histone acetylation. Adapted from Shaywitz and Greenberg, (16).

Creb Structure
FIGURE 3 Crystal structure of CREB bound to the CRE. Data obtained from the Brookhaven National Protein Structure Data Bank and rendered with Rasmol. Figure courtesy of Jennifer Gatchel.

however, is not sufficient for transcriptional activation; binding of CREB Binding Protein (CBP) is also necessary. CBP binding and activation is itself regulated by phospho-rylation; in particular, phosphorylation and activation by CaMKIV is relevant in the present context. CBP does a number of things—it binds to phosphorylated CREB, it helps bridge CREB to the promoter complex structurally, and it is a Histone Acetyl Transferase (HAT) enzyme that acetylates histone lysine residues to free up bound DNA.

I use CREB as an example because, as described earlier, it, along with the Serum Response Element (SRE)-recognizing transcription factor elk-1, has been directly implicated in L-LTP. How is it that the calcium trigger for L-LTP signals to CREB and its associated proteins? The relevant pathways are summarized in Figure 4. Most of the details of the relevant signal transduction cascades were presented in the last chapter, so I will not reiterate them here.

L-LTP-inducing calcium elevation post-synaptically activates Adenlyl Cyclase I/ VIII and via B-raf activates MEK (reviewed in reference 17). In addition, as described in the last chapter, ras-coupled receptors and

PKC-coupled receptors also can feed into this pathway. The product of MEK activation, dually phosphorylated ERK (ppERK) is then translocated into the nucleus. This nuclear translocation of ppERK is a regulated process—both BDNF and PKA can control this translocation by mechanisms that are still being investigated. Active ERK in the nucleus is most likely coupled to CREB phosphorylation via activation of a member of the pp90rsk family of S6 kinases, RSK2. Ser133 of CREB is not a substrate for ERK; ERK's effect is indirect through activating RSK2. Phosphorylation of Ser133 by RSK2 recruits the CREB binding protein, CBP, to the initiator complex and thereby promotes transcription. CBP activation is itself regulated by CaMKIV phosphorylation, and most likely both CBP activation by CaMKIV and ERK regulation of RSK2/CREB are necessary events for L-LTP induction (see 7,18).

This model is consistent with a wide variety of evidence demonstrating that AC knockouts, inhibition of PKA and ERK, and loss of CaMKIV function all lead to loss of L-LTP (see reference 19 and Table 1). Also, the various positive data demonstrating CREB phosphorylation and CRE-mediated gene expression with L-LTP, described in Table 1, motivate this model.

However, one might wonder what happened to CaMKII and PKA phospho-rylation of CREB, since these kinases are perfectly capable of phosphorylating Ser133 in CREB. The short answer is that available data do not indicate that PKA and CaMKII have access to CREB, at least in rat hippocampus and dentate gyrus. A variety of experiments have shown that CaMKII inhibitors do not affect synaptic activity-dependent CREB phosphorylation in these systems. In fact, if anything CaMKII is an inhibitor of CREB through phosphorylation at a site other than Ser133.

Also, Eric Roberson and others in my lab showed that PKA cannot cause elevated CREB phosphorylation without going

Camkii Creb
FIGURE 4 Activity-dependent regulation of gene expression in neurons. See text for details and discussion.

through the ERK cascade (20). Activation of PKA by application of forskolin to hip-pocampal slices results in ERK activation in area CA1, and this manipulation also elicits increased CREB phosphorylation. Eric and his colleagues determined the effects of MEK inhibition on forskolin stimulation of CREB phosphorylation in area CA1, and, surprisingly, the MEK inhibitor U0126 completely blocked CREB phosphorylation in response to forskolin application. While this was unexpected, this effect has been confirmed by Lu, Kandel, and Hawkins (21) and the data are consistent with additional work by Impey et al. (12). Thus, these data demonstrate that activation of MAPK results in increased CREB phosphorylation in area CA1; interestingly, these data also indicate that the cAMP pathway utilizes the MAPK cascade as an obligatory intermediate in regulating CREB phosphorylation in area CA1.

B. Modulatory Influences That Impinge Upon This Cascade

One of the most important co-regulators of this cascade is CaMKIV. Mice (or hip-pocampal slices) deficient in CaMKIV signaling have deficits in L-LTP, hippocampal CREB phosphorylation, and CREB-mediated gene expression (see Figure 1 and references 7, 8, and 22). As mentioned previously, CaMKIV acts by phosphorylating Ser301 in CBP, co-regulating the CREB/CBP complex (7). Regulation of CaMKIV likely depends on calmodulin translocation to the nucleus, an interesting attribute that may confer a capacity for temporal integration onto the nuclear read-out of cellular calcium signals (23). The pathways for nuclear calcium signaling are still being worked out, but they may be initiated by local calcium flux at the cell body coupled with ancillary signals from the synapse (24-26). The bottom line of all this is that it is important to remember that increased CREB phospho-rylation at Ser133 is not sufficient to give altered gene expression—an additional CaMKIV-mediated signal through the CREB co-activator CBP is necessary for altered gene expression as well (7). Thus, the ERK and CaMKIV pathways act in concert to trigger L-LTP-associated altered gene expression.

The growth factor BDNF also triggers modulatory mechanisms that feed into the CREB regulation cascade. BDNF is released with L-LTP inducing stimulation (27, 28) and likely contributes to L-LTP induction by two means (see Figure 4). First, it may act via ras to help to activate the MEK/ERK pathway directly, a mechanism for augmenting ERK-dependent gene expression (29). In addition, BDNF by mechanisms still being worked out controls ERK translocation into the nucleus, providing a gate-keeping role for triggering L-LTP (30). This BDNF-dependent regulation of ppERK translocation to the nucleus is necessary for L-LTP to be induced, conferring an additional signal integration mechanism onto the system. This mechanism also likely contributes to BDNF-induced plasticity per se (31).

Additional modulatory signal integration mechanisms also can operate at the synaptic level by augmenting activation of the ERK/CREB pathway. Many possibilities along these lines have already been discussed in earlier chapters when we talked about the numerous complexities of regulating the level of calcium achieved postsynaptically with LTP-inducing stimulation. One specific example that has received attention experimentally is regulation of L-LTP induction by DA (32). Dopamine coapplication during theta-frequency synaptic activity augments the induction of L-LTP. In the mouse hippocampus this pathway appears to be particularly important for generating CREB-dependent L-LTP (5). DA may augment NMD A receptor activation by way of the cAMP gate as we discussed in the last chapter, or may enhance ERK activation by elevating cAMP levels, or both.

In addition, nitric oxide acting through the cGMP-dependent protein kinase may also modulate L-LTP induction by impinging upon CREB phosphorylation (21). However, this pathway appears to operate in parallel to the pathway outlined in Figure 4, as opposed to influencing it directly or indirectly.

Finally, it is important to note that a static diagram such as Figure 4 cannot adequately convey some of the kinetic complexities that are known to exist in this system. Work from Gang-yi Wu, Karl Dieisseroth, and others in Dick Tsien's lab has shown that there are important temporal integration mechanisms that play a role in activity-dependent nuclear signaling in hippocampal neurons (23; reviewed along with other interesting aspects of this system in references 25 and 33). Specifically, repeated stimuli lead to a prolonged activation of the ERK/CREB pathway, a mechanism that is likely relevant in the system computing whether to trigger altered gene expression and L-LTP. There also exists in pyramidal neurons a "fast" CaMKIV-dependent signaling mechanism direct to CREB, which depends on activation of L-Type calcium channels (24). The role of this mechanism in L-LTP induction is still under investigation. It is possible that these two pathways act sequentially to give the biphasic CREB phosphorylation that has been observed with L-LTP-inducing stimulation (34). Regardless, it is clear that a number of temporal factors impinge upon the CREB transcriptional regulation pathway with L-LTP-inducing stimulation; these factors likely are controlled by distinct activity-dependent signaling mechanism such as L-channel activation and NMDA receptor-dependent events (26).

C. Additional Transcription Factors Besides CREB That May Be Involved in L-LTP Induction

The activation and nuclear translocation of ERK can lead to the activation of several transcription factors besides CREB, such as Elk-1 and c-Myc. Historically prominent among the transcription factors regulated by ERK is Elk-1, which when phospho-rylated at multiple sites by ERK cooperates with serum response factor (SRF) to drive transcription of serum response element (SRE)-controlled genes. Elk-1 phosphory-lation increases with L-LTP-inducing stimulation, a mechanism that likely triggers SRE-mediated changes in gene expression (35). One specific candidate target for this pathway is the transcription factor zif268, whose expression is mediated by Elk-1/ SRE-dependent processes. In addition, the target gene Arc/Arg3.1 may also be regulated by this pathway independently of CREB (36).

A final pathway potentially contributing to L-LTP induction is the NFkB pathway (15, 37; see Figure 4). Nuclear Factor kappa B (NFkB) is a transcription factor that normally resides in the cytoplasm, bound to an inhibitory partner IkB (Inhibitor of kappa B). NFkB is activated when the kinase IkB Kinase (IKK) phosphorylates IkB, which leads to loss of IkB by ubiquitin-mediated proteolysis. The free, active NFkB can then translocate to the nucleus and effect transcription of its target genes. Inhibition of NFkB activity leads to a reduction of L-LTP (38), and Ari Routtenberg's group has demonstrated NFkB activation with LTP-inducing stimulation (39). The mechanism of NFkB activation in LTP is unclear at this time, however interesting possible mechanisms include activation by reactive oxygen species like superoxide, direct phosphorylation of IkB by PKC, or indirect activation of IKK by PKC.

D. Gene Targets in L-LTP

This is one of the most fascinating aspects of L-LTP investigation, and an area where we have just started to scratch the surface in uncovering the relevant players. Identifying the gene targets of activity-dependent transcriptional regulation in L-LTP specifically, and in long-term memory generally, will be a watershed event in our understanding the molecular basis of memory. Clearly, identifying the genes whose expression changes with LTP (and learning) will give us needed clues concerning how very long-lasting changes in synaptic function are achieved in the CNS.

I already have listed the known gene targets in L-LTP as part of the measure-related experiments supporting a role for altered gene expression in L-LTP (see Table 1). In this section, I will reorganize the list along some functional lines. The take-home message from this section is threefold. First, we clearly are at an early stage with these studies because not many candidates have been identified. Second, where target genes have been identified the functions of their gene products suggest that there are many complex post-gene-read-out layers of signal transduction involved in L-LTP. Third, even with the paucity of targets that

Was this article helpful?

0 0
The Chakra Checklist

The Chakra Checklist

The chakras are described as being aligned in an ascending column from the base of the back to the top of the head. New Age practices frequently associate each chakra with a particular color.

Get My Free Ebook

Post a comment