Table of contents

CREB, synaptic plasticity and long term memory

Neuronal activity-dependent processes are believed to mediate the formation of synaptic connections during neocortical development, but the underlying intracellular mechanisms are not known. In the visual system, altering the pattern of visually driven neuronal activity by monocular deprivation (MD) induces cortical synaptic rearrangement during the critical period of a postnatal developmental window. Using transgenic mice carrying a CRE-lacZ reporter, it has been demonstrated that a calcium- and cAMP-regulated signaling pathway is activated following monocular deprivation. Monocular deprivation leads to an induction of CRE-mediated lacZ expression in the visual cortex preceding the onset of physiologic plasticity, and this induction is dramatically downregulated following the end of the critical period. These results suggest that CRE-dependent coordinate regulation of a network of genes may control physiologic plasticity during postnatal neocortical development (Pham, 1999).

How might CRE-mediated transcription be involved in changes in synaptic function following MD? It is envisioned that the CRE-regulated gene targets could have modulatory effects on synaptic function. Thus, a major question that arises from this work is the identity of these endogenous gene targets. A large number of relevant neuronally expressed genes have been described, many of which are known to function at synapses and which can be regulated by CREB. These include BDNF, calmodulin-dependent kinase IV, synapsin I, somatostatin (a subtype of a voltage-gated potassium channel) and the fos and jun family of immediate-early genes. The possible involvement of neurotrophins is particularly interesting since they are possible modulators of cortical synapses and are known to regulate CREB function. The fundamental importance of these findings that show induction of CRE-mediated transcription is that these results imply the activation of a network of CRE-regulated genes in response to a physiologically relevant condition. Although the changes in expression of individual genes will likely be modest, their combined transcriptional enhancement could substantially synergize to produce major physiologic effects (Pham, 1999).

Long-term facilitation of the connections between the sensory and motor neurons of the gill-withdrawal reflex in Aplysia requires five repeated pulses of serotonin (5-HT). The repeated pulses of 5-HT initiate a cascade of gene activation that leads ultimately to the growth of new synaptic connections. Several genes in this process have been identified, including the transcriptional regulators apCREB-1, apCREB-2, apC/EBP, and the cell adhesion molecule apCAM, which is thought to be involved in the formation of new synaptic connections. The transcriptional regulators apCREB-2 and apC/EBP, as well as a peptide derived from the cytoplasmic domain of apCAM, are phosphorylated in vitro by Aplysia mitogen-activated protein kinase (apMAPK). The cDNA encoding apMAPK (see Drosophila MAPK) has been cloned and apMAPK activity is increased in sensory neurons treated with repeated pulses of 5-HT and by the cAMP pathway. These results suggest that apMAPK may participate with cAMP-dependent protein kinase during long-term facilitation in sensory cells by modifying some of the key elements involved in the consolidation of short- to long-lasting changes in synaptic strength. How might the PKA and MAPK pathways converge in Aplysia sensory neurons? In the simplest case, the MAPK pathway would be downstream from the PKA pathway. In many cells, PKA negatively regulates the MAPK pathway by phosphorylating Raf-1. However, in B-Raf-containing cells, PKA activates the MAPK pathway by signaling through Rap1, a member of the Ras family of small G proteins. PKA may thus activate MAPK by phosphorylating the Aplysia homolog of Rap1, therby activating B-Raf, MEK, and MAPK (Martin, 1998).

Recent studies suggest that the CREB-CRE transcriptional pathway is pivotal in the formation of some types of long-term memory. However, it has not been demonstrated that stimuli that induce learning and memory activate CRE-mediated gene expression. To address this issue, a mouse strain transgenic for a CRE-lac Z reporter was used to examine the effects of hippocampus-dependent learning on CRE-mediated gene expression in the brain. Training for contextual conditioning or passive avoidance leads to significant increases in CRE-dependent gene expression in areas CA1 and CA3 of the hippocampus. Auditory cue fear-conditioning, which is amygdala dependent, is associated with increased CRE-mediated gene expression in the amygdala, but not the hippocampus. These data demonstrate that learning in response to behavioral conditioning activates the CRE transcriptional pathway in specific areas of brain (Impey, 1998b).

One of the hallmarks of long-term memory in both vertebrates and invertebrates is the requirement for new protein synthesis. In sensitization of the gill-withdrawal reflex in Aplysia, this requirement can be studied on the cellular level. Here, long-term but not short-term facilitation of the monosynaptic connections between the sensory and motor neurons requires new protein synthesis and is reflected in an altered level of expression of specific proteins regulated through the cAMP second-messenger pathway. Using gene transfer into individual sensory neurons of Aplysia, it can be shown that serotonin (5-HT) induces transcriptional activation of a lacZ reporter gene driven by the cAMP response element (CRE) and that this induction requires CRE-binding proteins (CREBs). The induction by 5-HT does not occur following a single pulse, but becomes progressively more effective following two or more pulses. Moreover, expression of GAL4-CREB fusion genes shows that 5-HT induction requires phosphorylation of CREB on Ser119 by protein kinase A (See PKA). These data provide direct evidence for CREB-modulated transcriptional activation with long-term facilitation (Kaang, 1993).

The switch from short- to long-term facilitation induced by behavioral sensitization in Aplysia involves CREB-like proteins, as well as the immediate-early gene ApC/EBP. Using the bZIP domain of ApC/EBP in a two-hybrid system, ApCREB2 has been cloned. It is a transcription factor constitutively expressed in sensory neurons that resembles human CREB2 and mouse ATF4. ApCREB2 represses ApCREB1-mediated transcription in F9 cells. Injection of anti-ApCREB2 antibodies into Aplysia sensory neurons causes a single pulse of serotonin (5-HT), which induces only short-term facilitation lasting minutes, to evoke facilitation lasting more than 1 day. This facilitation has the properties of long-term facilitation: it requires transcription and translation, induces the growth of new synaptic connections, and occludes further facilitation by five pulses of 5-HT (Bartsch, 1995).

The switch from short-term to long-term facilitation of the synapses between sensory and motor neurons mediating gill and tail withdrawal reflexes in Aplysia requires CREB-mediated transcription and new protein synthesis. Several downstream genes have been isolated, one of which encodes a neuron-specific ubiquitin C-terminal hydrolase. This rapidly induced gene encodes an enzyme that associates with the proteasome and increases its proteolytic activity. This regulated proteolysis is essential for long-term facilitation. Inhibiting the expression or function of the hydrolase blocks induction of long-term but not short-term facilitation. It is suggested that the enhanced proteasome activity increases degradation of substrates that normally inhibit long-term facilitation. Thus, through induction of the hydrolase and the resulting up-regulation of the ubiquitin pathway, learning recruits a regulated form of proteolysis that removes inhibitory constraints on long-term memory storage (Hedge, 1997).

While evidence has accumulated in favor of cAMP-associated genomic involvement in long-term synaptic plasticity, the mechanisms downstream of the activated nucleus that underlie these changes in neuronal function remain mostly unknown. Dendritic spines, the locus of excitatory interaction among central neurons are prime candidates for long-term synaptic modifications. Phosphorylation of the cAMP response element binding protein (CREB) is linked to formation of new spines; exposure to estradiol doubles the density of dendritic spines in cultured hippocampal neurons, and concomitantly causes a large increase in phosphorylated CREB and in CREB binding protein. Blockade of cAMP-regulated protein kinase A eliminates estradiol-evoked spine formation, as well as the CREB and CREB binding protein responses. A specific antisense oligonucleotide eliminates the phosphorylated CREB response to estradiol as well as the formation of new dendritic spines. These results indicate that CREB phosphorylation is a necessary step in the process leading to generation of new dendritic spines (Murphy, 1997).

Phosphorylation of the transcription factor CREB is thought to be important in processes underlying long-term memory. It is unclear whether CREB phosphorylation can carry information about the sign of changes in synaptic strength, whether CREB pathways are equally activated in neurons receiving or providing synaptic input, or how synapse-to-nucleus communication is mediated. Ca(2+)-dependent nuclear CREB phosphorylation was rapidly evoked by synaptic stimuli including, but not limited to, those that induced potentiation and depression of synaptic strength. In striking contrast, high frequency action potential firing alone failed to trigger CREB phosphorylation. Activation of a submembranous Ca2+ sensor, just beneath sites of Ca2+ entry, appears critical for triggering nuclear CREB phosphorylation via calmodulin and a Ca2+/calmodulin-dependent protein kinase (Deisseroth, 1996).

Gene expression regulated by the cAMP response element (CRE) has been implicated in synaptic plasticity and long-term memory. It has been proposed that CRE-mediated gene expression is stimulated by signals that induce long-term potentiation (LTP). To test this hypothesis, Mice were prepared that were transgenic for a CRE-regulated reporter construct. Long-lasting long-term potentiation (L-LTP) in the hippocampus was studied, because it depends on cAMP-dependent protein kinase activity (PKA) and de novo gene expression. CRE-mediated gene expression was markedly increased after L-LTP, but not after decremental UP (D-LTP). Furthermore, inhibitors of PKA blocked L-LTP and associated increases in CRE-mediated gene expression. These data demonstrate that the signaling required for the generation of L-LTP but not D-LTP is sufficient to stimulate CRE-mediated transcription in the hippocampus (Impey, 1996).

In conditioned taste aversion (CTA) organisms learn to avoid a taste if the first encounter with that taste is followed by transient poisoning. The neural mechanisms that subserve this robust and long-lasting association of taste and malaise have not yet been elucidated, but several brain areas have been implicated in the process, including the amygdala. An investigation was performed of the role of the amygdala in CTA learning and memory, both in general terms, and in particular, as regards the cAMP response element-binding protein (CREB). Antisense technology in vivo was combined with behavioral, molecular, and histochemical analysis. Local microinjection of phosphorothioate-modified oligodeoxynucleotides (ODNs) antisense to CREB into the rat amygdala several hours before CTA training transiently reduces the level of CREB protein during training and impairs CTA memory when tested 3-5 d later. In comparison, sense ODNs have no effect on memory. The effect of antisense is not attributable to differential tissue damage and is site-specific. CREB antisense in the amygdala has no effect on retrieval of CTA memory once it has been formed, and does not affect short-term CTA memory. It is proposed that the amygdala, specifically the central nucleus, is required for the establishment of long-term CTA memory in the behaving rat; that the process involves long-term changes, subserved by CRE-regulated gene expression, in amygdala neurons; and that the amygdala may retain some CTA-relevant information over time rather than merely modulating the gustatory trace during acquisition of CTA (Lamprecht, 1997).

The requirement for transcription during long-lasting synaptic plasticity has raised the question of whether the cellular unit of synaptic plasticity is the soma and its nucleus or the synapse. The finding that long-term memory requires alterations in gene expression and thus the nucleus - a resource shared by all the synapses within a cell - poses a cell biological paradox in the study of memory. Does the requirement of transcription for long-lasting forms of synaptic plasticity mean that long-lasting memory needs to be cell wide, or can the strength of individual synaptic connections be modified independently? To address this question, a single bifurcated Apylsia sensory neuron was cultured that made synapses with two spatially separated motor neurons. By perfusing serotonin onto the synapses made onto one motor neuron, it was found that a single axonal branch can undergo long-term branch-specific facilitation. This branch-specific facilitation depends on CREB-mediated transcription (taking place in the nucleus) and involves the growth of new synaptic connections exclusively at the treated branch (Martin, 1997).

Given that branch-specific facilitation requires the nucleus but is synapse-specific, what are the underlying mechanisms? One possibliity is that marking the synapse for synapse-specific long term facilitation requires local translation of mRNAs. Support for this possibility comes from two sets of findings: (1) many studies have shown that mRNAs, ribosomes, and components of the translation machinery are present not only in the perinuclear cell soma of neurons but also in distal neuronal processes in the postsynaptic spine and in the presynaptic terminals. (2) An intermediate form of synaptic facilitation, which is dependent on translation and not transcription, has been described in Aplysia sensory motor neurons. The protein synthesis inhibitor emertine, applied to the synaptic branch receiving serotonin, blocks branch-specific long-term facilitation. Thus, local protein synthesis appears to be required in the presynaptic sensory neuron but not in the postsynaptic motor cell. In fact, presynaptic sensory neuron axons deprived of their cell bodies are capable of protein synthesis (but not long term facilitation), and this protein synthesis is stimulated 3-fold by exposure to serotonin (Martin, 1997).

How can the products of gene expression be differentially targeted to alter synaptic strength at some synapses made by a given presynaptic neuron, but not at all the others? How can one reconcile the possibility of branch-specific modifications in synaptic connections, postulated to be characteristic of certain types of memory storage, with a genetic program (involving the nucleus) that appears to be neuron wide? One possibility is that proteins might be targeted specifically down one branch rather than another of an axon arbor. Given that a single neuron has many branches and synaptic terminals, this type of targeting mechanism has seemed unlikely. An alternative possibility is that synapses are "tagged" by synaptic stimulation to capture products of gene expression that are exported throughout the cell but are only functionally incorporated at synapses tagged by previous activity. A synaptic capture experiment was carried out by giving a single application of serotonin to one synapse, which produces only transient, short-term facilitation, immediately following five pulses of serotonin to the other synapse. This protocol produces long-lasting facilitation at both branches. Thus a single pulse of serotonin appears to mark the second set of synapses so that they can capture the products of gene expression induced by five pulses of serotonin at the first set of synapses. To determine whether synaptic capture is also dependent on local translation, emertine was perfused into the second branch following the single application of serotonin to that branch. Unlike the establishment of synapse-specific long-term facilitation at the primary site, the capture of long-term facilitation at the second site after it has been initiated or established at a primary site does not require local protein synthesis. This suggests that the retrograde signal to the nucleus from the initial site of synapse-specific long-term facilitation may itself require new protein synthesis, but that tagging of the second site does not require protein synthesis (Martin, 1997).

Hippocampal "place cells" fire selectively when an animal is in a specific location. The fine-tuning and stability of place cell firing has been compared in two types of mutant mice with different long-term potentiation (LTP) and place learning impairments. alpha calmodulin kinase II and CREBalphadelta mutants differ substantially in the severity of their defects in synaptic plasticity and learning. The alpha calmodulin kinase II mutant mice are severely impaired in spatial learning as well as in NMDAR-dependent LTP in the CA1 region of the hippocampus. In ontrast CREBalphadelta mutants have reduced LTP and mild spatial learning deficits. Place cells from both mutants showed decreased spatial selectivity. Place cell stability is also deficient in both mutants and, consistent with the severities in their LTP and spatial learning deficits, is more affected in mice with a point mutation [threonine (T) at position 286 mutated to alanine (A)] in the alpha calmodulin kinase II (alphaCaMKIIT286A) than in mice deficient for the alpha and Delta isoforms of cyclicAMP-responsive element binding proteins (CREBalphaDelta-). Thus, LTP appears to be important for the fine tuning and stabilization of place cells, and these place cell properties may be necessary for spatial learning (Cho, 1998).

The observed LTP deficits in the alphaCaMKII mutants are not due to abnormalities in GABAA inhibition, NMDAR currents, or synaptic function before and during the tetanus, suggesting that the deficits act downstream of Ca2+ influx through NMDARs. These findings are consistent with the observation that, after LTP induction, the autophosphorylated form of alpha CaMKII phosophorylates glutamate receptor subunits that may be required for LTP. In accordance with a model implicating the CaM-independent state of CaMKII in learning and memory, it was shown that the autophosphorylation of alphaCaMKII at Thr286 is required for spatial learning. Thus, the autophosphorylation of alphaCaMKII at Thr 286 is crucial for hippocampal LTP and hippocampus-dependent learning (Cho, 1998).

Studies in Aplysia, Drosophila, and mice have shown that the transcription factor CREB is involved in formation and retention of long-term memory. To analyze the impact of differential CREB levels on learning and memory, the gene dosage of CREB was analyzed in two strains of mutant mice: (1) CREBalphadelta mice, in which the alpha and delta isoforms are disrupted, but a third isoform beta is strongly up-regulated; (2) CREBcomp, a compound strain with one alphadelta allele and one CREBnull allele in which all CREB isoforms are disrupted. To minimize genetic background effects, CREB mutations were backcrossed into a C57BL/6 and a FVB/N strain, respectively, and studies were performed in F1 hybrids from these lines. CREBcomp but not CREBalphadelta F1 hybrids are impaired in water maze learning and fear conditioning, demonstrating a CREB gene dosage effect. However, analysis of the platform searching strategies in the water maze task suggests that CREBcomp mutants are impaired in behavioral flexibility rather than in spatial memory. In contrast to previous experiments using CREBalphadelta mice with different genetic background, the F1 hybrid CREBalphadelta and CREBcomp mice did not show deficits in a social transmission of food preference task nor in dentate gyrus and CA1 LTP as recorded from slice preparations. These data indicate that the hybrid vigor typical for F1 hybrids may compensate for a reduction in CREB levels in some tests. However, the persistence of clear behavioral deficits as shown by the F1 hybrid CREBcomp mice in water maze and fear conditioning indicates a robust and repeatable phenomenon that will permit further functional analysis of CREB (Gass, 1998).

Recently, it has been shown that cerebellar LTD has a late phase that may be blocked by protein synthesis inhibitors. To understand the mechanisms underlying the late phase, interference was carried out with the activation of transcription factors that might couple synaptic activation to protein synthesis. Particle-mediated transfection of cultured Purkinje neurons with an expression vector encoding a dominant inhibitory form of CREB results in a nearly complete blockade of the late phase. Kinases that activate CREB are inhibited. When LTD was assessed, it was seen that inhibition of PKA or the MAPK/RSK cascades is without effect on the late phase, while constructs designed to interfere with CaMKIV function attenuates the late phase. These results indicate that the activation of CaMKIV and CREB are necessary to establish a late phase of cerebellar LTD (Ahn, 1999).

There are at least two temporally distinct phases of memory storage: a short-term memory lasting minutes and a long-term memory lasting days or longer. These two phases differ not only in their time courses, but also in their molecular mechanisms: long-term memory, but not the short-term form, requires the synthesis of new proteins. Recent studies in Aplysia and mice have revealed that these distinct stages in behavioral memory are reflected in distinct phases of synaptic plasticity. In Aplysia, these stages have been particularly well studied in the context of sensitization, a form of learning in which an animal learns to strengthen its reflex responses to previously neutral stimuli following the presentation of an aversive stimulus. The short- and long-term memory for sensitization are mirrored by the short- and long-term facilitation of the synaptic connections between the sensory and motor neurons that mediate this reflex. This monosynaptic component can be examined not only in the intact animal, but also in a single sensory neuron cultured with its target postsynaptic motor neuron. At this cultured synapse, one pulse of 5-HT, a neurotransmitter released in vivo by interneurons activated by the sensitizing tail stimuli, produces a PKA- and PKC-mediated short-term facilitation lasting only minutes. By contrast, five spaced pulses of 5-HT elicit a long-term facilitation lasting more than 24 hr. With five pulses of 5-HT, PKA recruits MAP kinase and both translocate to the nucleus. Here, they activate CREB1 and derepress CREB2, leading to the induction of a set of immediate-early genes. Long-term facilitation is further associated with the growth of new synaptic connections (Casadio, 1999 and references therein).

A single bifurcated sensory neuron is plated in contact with two spatially separated postsynaptic motor neurons; a single synapse or group of synapses can be modified independently in a protein synthesis-dependent manner. This spatially restricted synapse specific facilitation requires the activity of CREB1 in the nucleus as well as local protein synthesis in the 5-HT-treated processes of the sensory cell. In addition to synapse-specific facilitation, a second phenomenon referred to as 'synaptic capture' has been characterized. Once synapse-specific long-term facilitation has been initiated by microperfusing five pulses of 5-HT onto one branch of the sensory neuron, a single pulse of 5-HT, which per se induces only transient facilitation, is able to recruit long-term facilitation when applied to a second branch. This synaptic capture does not require local protein synthesis (Casadio, 1999 and references therein).

As is generally the case with neurons, the sensory neuron is a highly polarized cell with several functionally distinct and spatially separate compartments, each of which is selectively innervated. In fact, in the intact animal, the endings of different serotonergic interneurons contact the sensory neurons at different sites, including its pre- and post-synaptic terminals and processes and its cell body. What then are the functional consequences of applying 5-HT selectively to the cell body? Five pulses of 5-HT applied selectively to the soma of the sensory neuron produce a cell-wide long-term facilitation (a third phenomenon) that requires CREB1-mediated transcription, but that does not persist and is not associated with growth. A similar cell-wide facilitation is obtained by injecting recombinant phospho-CREB1 protein into the sensory neuron. Cell-wide facilitation can be sustained and growth can be captured by a single pulse of 5-HT. Thus, a single pulse of 5-HT serves two functions: (1) it allows for 24 hr facilitation and growth by means of a PKA-mediated covalent modification; (2) it stabilizes growth and facilitation in a protein synthesis-dependent way to produce facilitation that persists at least 72 hr. By systematically comparing cell-wide facilitation to synapse-specific facilitation and synaptic capture, it has been found that each represents a distinct form of CREB-dependent, long-lasting synaptic plasticity. Thus, a single neuron has multiple long-term mechanisms for temporally and spatially integrating stimuli to produce transcription-dependent long-lasting changes in synaptic strength (Casadio, 1999).

In a culture system where a bifurcated Aplysia sensory neuron makes synapses with two motor neurons, repeated application of serotonin (5-HT) to one synapse produces a CREB-mediated, synapse-specific, long-term facilitation, which can be captured at the opposite synapse by a single pulse of 5-HT. Repeated pulses of 5-HT applied to the cell body of the sensory neuron produce a CREB-dependent, cell-wide facilitation, which, unlike synapse-specific facilitation, is not associated with growth and does not persist beyond 48 hr. Persistent facilitation and synapse-specific growth can be induced by a single pulse of 5-HT applied to a peripheral synapse. Thus, the short-term process initiated by a single pulse of 5-HT serves not only to produce transient facilitation, but also to mark and stabilize any synapse of the neuron for long-term facilitation by means of a covalent mark and rapamycin-sensitive local protein synthesis (Casadio, 1999).

Stimulation restricted to the cell body of a sensory neuron is sufficient to induce long-term facilitation and can do so in the complete absence of short-term facilitation at any other synapse of the neuron. Like the synapse-specific process, cell-wide facilitation requires CREB-mediated transcription. However, whereas branch specific facilitation is associated with the growth of new synaptic connections and persists beyond 72 hr, cell-wide facilitation is not associated with significant growth and decays within 48 hr. However, once the cell-wide process has been triggered by five pulses of 5-HT at the cell body, a single pulse of 5-HT applied to any one synapse can capture growth at that synapse, thereby converting a 24 hr cell-wide process into a more persistent synapse-specific form (Casadio, 1999).

These studies demonstrate that CREB-mediated long-term plasticity is not a unitary phenomenon. In fact, the same neuron can undergo four different CREB-mediated forms of long-term synaptic plasticity, depending on the site and the number of repetitions of modulatory input: synapse-specific long-term facilitation, capture of the synapse-specific form, cell-wide long-term facilitation, and capture of the cell-wide form (a fourth phenomenon). The experimental paradigm for the fourth phenomenon, capture of the cell-wide form, involves 5 pulses of 5-HT to the cell body rather than to a single branch as is carried out in the synapse specific long term paradigm. These five pulses create cell wide facilitation. In the case of this cell-wide capture, a single pulse to a branch, constituting the capture event, results in new varicosities. Although all four processes share at least one common nuclear mechanism (CREB-mediated transcription), each has distinct properties. Synapse-specific long-term facilitation initiated by five repeated pulses of 5-HT stimuli at a synapse is accompanied by growth, persists for at least 72 hr, and requires local protein synthesis for both its retrograde signal and for the stabilization of growth. Capture of the synapse-specific long-term facilitation by a marked synapse is smaller in amplitude than synapse-specific facilitation but is also accompanied by growth, is persistent, and requires local protein synthesis only for stabilization. Cell-wide long-term facilitation, generated by repeated modulatory stimuli at the cell body, is transient. This cell-wide long-term facilitation, however, can be captured by a marking signal that stimulates growth and converts the facilitation to a persistent form (referring to the fourth phenomenon of capture of the cell-wide form) (Casadio, 1999).

The finding that cell-wide, CREB-dependent cell-wide facilitation can occur in the absence of any synaptic growth was unexpected, since long-lasting increases in synaptic strength and long-term facilitation have been thought to result, at least in part, from the growth of new synaptic connections. How, then, might an increase in synaptic strength that requires transcription and translation persist for 24 hr in the absence of synaptic growth? One possibility is the recruitment of 'silent synapses'. In the intact animal, approximately 60% of the presynaptic varicosities of the sensory neurons do not have active release sites. When long-term sensitization is produced in the animal, it leads to both an increase in the number of varicosities and to an increase in the percentage of varicosities that have active zones. Whereas the increase in varicosity number persists as long as the memory, changes in the size of active zones only last a few days. Since only 50% of the varicosities in cultured sensory neurons have active zones, cell body application of 5-HT might change synaptic strength at 24 hr by increasing the incidence of active zones (Casadio, 1999).

CREB-mediated transcription appears to be necessary for the initial establishment of all four forms of synaptic plasticity, but it is not sufficient for the self-maintained stabilization of the plastic changes. To obtain persistent facilitation and the growth of new synaptic connections, one also needs a single pulse of 5-HT applied to the synapse. This single pulse by itself is only able to induce short-term facilitation. But in a cell where CREB-mediated transcription is induced by repeated stimuli, the single pulse also marks a synapse so as to allow long-term facilitation to become persistent and to be accompanied by growth of new terminals. Thus, one of the surprising functions of the short-term process is to convert a transient long-term process into a persistent one. These results illustrate again, as did earlier studies, that the short-term process serves two functions: (1) a single pulse of 5-HT alone produces a selective synapse-specific enhancement of synaptic strength that contributes to short-term memory lasting minutes; (2) in conjunction with the activation of CREB by repeated stimuli given to any other synapse or to the cell body, a single pulse of 5-HT will also act to ''mark'' that synapse for persistent synaptic facilitation and growth (Casadio, 1999).

Studies of the molecular nature of the mark necessary for synapse-specific facilitation and for capture suggest that there are two components to the marking signals: a covalent mark for the initiation of synapse-specific plasticity, mediated by PKA, and a mark for its stabilization, mediated by local protein synthesis. The initiation of synapse-specific facilitation is reflected by the facilitation (both functional and morphological) expressed at 24 hr, while the stabilization is reflected by the facilitation (both functional and morphological) expressed at 72 hr (Casadio, 1999).

Translation of messenger RNAs can be regulated in a number of ways that act on either the 5' or 3' end of the message. In response to growth factors such as cytokines and insulin, there is a 2- to 3-fold increase in protein synthesis, yet only a small subset of transcripts (10%-15%) are preferentially stimulated. The translation of this specific subset of mRNAs is enhanced by means of a signaling pathway that is blocked by rapamycin. Rapamycin acts on the kinase RAFT/TOR, which is critical for the phosphorylation of the S-6 kinase. Through inactivation of the S-6 kinase, rapamycin reduces the translation of a set of mRNAs containing an oligopyrimidine tract immediately after the N7 methyl guanosine cap. The family of mRNAs that contain the polypyrimidine tract includes proteins that are themselves important components of the translational machinery, such as ribosomal proteins (S3, S6, S14, and S24) and translational elongation factors eF-1 and eF-1. Rapamycin also blocks the phosphorylation of eIF4EBP, which affects the translation of another class of mRNAs containing a high degree of secondary structure in their 5' ends, such as cyclin D. Rapamycin has recently been shown to cause degradation of eIF4G, a component of mRNA cap recognition complex eIF4F. Unfortunately, rapamycin is not completely specific and can have effects on cell functions other than translation. When emetine is continuously perfused onto the synapses that receive five pulses of 5-HT, the block of long-term facilitation is already evident at 24 hr, indicating that unlike synaptic capture, local protein synthesis for the induction of long-term facilitation is required early and may therefore be necessary to initiate the retrograde signal from the synapse to the nucleus. This presynaptic retrograde signal is likely to be important for activating CREB-dependent transcription in the nucleus and for the translation of the new proteins required for the expression of the long-term plasticity. These newly expressed proteins are then shipped back to the originally stimulated synapse (where they are captured and incorporated into functional release sites), and to all other synapses as well. This component of protein synthesis can be blocked by emetine but not by rapamycin. Thus, local protein synthesis has two components that serve two distinct functions at two different sites. At the site of initiation for synapse-specific facilitation, a non-rapamycin-sensitive component of local protein synthesis is required for the retrograde signal. At both this site and at the site of capture, rapamycin-sensitive local protein synthesis is required to mark the synapse for persistence of the functional and structural change (Casadio, 1999).

These findings suggests that activation of transcription by one set of synapses in a synapse-specific way activates downstream genes that affect all nonactivated synapses, any one of which presumably can now be recruited by a single pulse of 5-HT. Moreover, this period of capture persists in time over a period of between 1 and 4 hr and extends in space over distances between synapses of hundreds of microns. Retrograde signals, nuclear loops, marking signals, and time windows of 1 to 4 hr for capture introduce a new temporal and spatial dimension in the integrative actions of the neuron. To give but one example, the classical temporal and spatial summative processes of neurons typically involve the associative interaction of signals separated by seconds. This is evident, for example, in the associative activation of the NMDA receptor by glutamate and voltage, or the associative enhancement of facilitation by 5-HT and activity. These forms of temporal summation allow short-term associative influences of the sort required for behavioral associative events, such as Pavlovian conditioning. As these arguments make clear, capture, which can persist for 1 to 4 hr, adds a new dimension to the temporal summative capabilities of a single neuron. What might be the function of such long-range and long-term associations? One possibility is that they are important for arousal or for flashbulb memories, for charged events that produce a memory for relevant events that occurred during a period of hours surrounding an important event. Similarly, there are forms of learning such as bait shyness with a long conditioned stimulus-unconditioned stimulus interval that could be accommodated by this mechanism. This new level of integration raises a further question: Will this integrative action also coordinate the action of inhibitory or facilitatory input? The sensory neurons undergo presynaptic inhibition, induced by the peptide FMRFamide. Is inhibition also branch specific, and if so, what are the interactive mechanisms that coordinate long-term inhibition and facilitation (Casadio, 1999).

At least two temporally and mechanistically distinct forms of memory are conserved across many species: short-term memory that persists minutes to hours after training and long-term memory (LTM) that persists days or longer. In general, repeated training trials presented with intervening rest intervals (spaced training) is more effective than massed training (the same number of training trials presented with no or short intervening rest intervals) in producing LTM. LTM requires de novo protein synthesis, and cAMP response element-binding protein (CREB) may be one of the transcription factors regulating the synthesis of new proteins necessary for the formation of LTM. Rats given massed fear conditioning training show no or weak LTM, as measured by fear-potentiated startle, compared with rats given the same amount of training but presented in a spaced manner. Increasing CREB levels specifically in the basolateral amygdala via viral vector-mediated gene transfer significantly increases LTM after massed fear training. These results suggest that CREB activity in the amygdala serves as a molecular switch for the formation of LTM in fear conditioning (Josselyn, 2001).

The effects of CREB overexpression on LTM formation are specific, in terms of (1) biochemistry, (2) anatomy, (3) stimulus processing, (4) time course, and (5) training protocol used: (1) the facilitatory effects of CREB in this task depend on phosphorylation at Ser133, because similar overexpression of mCREB, with a single point mutation at this phosphoacceptor site, does not facilitate LTM; (2) the enhancing effects of HSV-CREB on LTM formation show anatomical specificity to the basolateral amygdala because infusions of HSV-CREB into brain regions surrounding the amygdala or into a control region (the caudate nucleus) failed to enhance LTM formation after massed training; (3) the LTM facilitation produced by overexpression of CREB cannot be attributed to nonspecific effects on conditioned stimulus or unconditioned stimulus processing; (4) the temporal specificity of the effects of CREB overexpression suggests the critical involvement of LTM encoding rather than retrieval processes and (5) overexpression of CREB does not alter short term memory produced by massed training or LTM produced by spaced training, indicating that overexpression of CREB enhances the formation of LTM only under specific training conditions, such as massed training (Josselyn, 2001).

Restricted and regulated expression in mice of VP16-CREB, a constitutively active form of CREB, in hippocampal CA1 neurons lowers the threshold for eliciting a persistent late phase of long-term potentiation (L-LTP) in the Schaffer collateral pathway. This L-LTP has unusual properties in that its induction is not dependent on transcription. Pharmacological and two-pathway experiments suggest a model in which VP16-CREB activates the transcription of CRE-driven genes and leads to a cell-wide distribution of proteins that prime the synapses for subsequent synapse-specific capture of L-LTP by a weak stimulus. This analysis indicates that synaptic capture of CRE-driven gene products may be sufficient for consolidation of LTP and provides insight into the molecular mechanisms of synaptic tagging and synapse-specific potentiation (Barco, 2002).

The encoding of new memories in the nervous system is thought to require long-lasting changes in the strength of specific synaptic connections between neurons mediated by specific alterations in gene expression. One important synaptic model for encoding memories is long-term potentiation (LTP). In LTP, as in memory storage, it is possible to distinguish between stages of storage. There is an early, short-term stage (E-LTP), which lasts minutes, and a later, long-term stage (L-LTP), which lasts much longer. The long-lasting synaptic process (L-LTP) shares with long-term memory (LTM) the requirement for the synthesis of new mRNA and protein. The finding of a transcriptional requirement for long-lasting forms of synaptic plasticity has raised a fundamental question in the study of learning-related plasticity: does the activation of transcription in the nucleus mean that the critical unit of long-term synaptic plasticity is the cell nucleus, or can long-term synaptic plasticity somehow be restricted to the single synapse? If the unit of long-term storage is the synapse, what mechanisms restrict the action of the newly expressed gene products to some synapses but not to others? Studies of these questions in Aplysia and rats using two-pathway experiments indicate that both short-term and long-term facilitation can be synapse specific (Barco, 2002).

Frey and Morris (1997, 1998) first delineated synaptic capture in the mammalian brain. They found that once transcription-dependent LTP has been induced at one pathway, the long-term process can be 'captured' at a second pathway receiving a single train, a stimulation that would normally produce only E-LTP. Thus, the stimulus for the short-term process serves not only to produce transient facilitation at that synapse, but can also mark and stabilize facilitation at any synapse of the neuron by capturing, for that synapse, the newly expressed gene products. In the years following its initial description, synaptic capture in mammalian hippocampus has remained uncharacterized, and many aspects remain unclear. Does capture share the same molecular machinery that is used by L-LTP? What gene products are distributed cell-wide when the long-term process is turned on at one synapse? What is the molecular nature of the tag that marks active synapses (Barco, 2002)?

Studies of synaptic capture at the synapses between the sensory and motor neurons of the gill-withdrawal reflex in Aplysia have shed light on some of these questions. In Aplysia, synapse-specific facilitation requires the activity of the transcriptional activator CREB-1 (the cAMP responsive element binding protein) in the nucleus, as well as a PKA-mediated covalent signal to mark the stimulated synapses and local protein synthesis to stabilize that mark. Furthermore, injection of CRE sequence oligonucleotides in Aplysia neurons selectively inhibits long-term facilitation with no effect in short-term synaptic plasticity, suggesting that the cAMP responsive element (CRE)-driven gene products are good candidates for the priming molecules that need to be captured at the marked synapse. Indeed, injection into the cell body of phosphorylated CREB-1 gives rise to long-term facilitation in all the synapses of the sensory neuron by seeding these synapses with the protein products of CRE-driven genes. However, this facilitation is not maintained unless the synapse is marked by the short-term process (Barco, 2002 and references therein).

Substantial evidence in experimental systems ranging from mollusks to humans suggests that CREB acts as one of the core components in the molecular switch that converts short- to long-term synaptic plasticity and short- to long-term memory. In Aplysia, opposing forms of CREB (CREB1a activator and CREB1b and CREB2 repressor) produce opposite effects on long-term facilitation. Similarly, opposing forms of CREB also produce opposite effects on long-term memory in transgenic flies (Barco, 2002 and references therein).

Mammals also seem to require activation of CRE-dependent transcription for long-term memory. Intrahippocampal infusion of CREB antisense oligonucleotides disrupts long-term spatial memory in rats, but does not affect short-term memory. In mice, there is increased expression of a CRE-driven lacZ reporter construct following stimuli that produce L-LTP and after training on a hippocampus-dependent task. Moreover, CREB is phosphorylated in the CA1 pyramidal cells by electrical stimuli that induce LTP and after training in hippocampus-dependent tasks. CREB is a target of PKA, CaMKIV, and the MAPK cascade, all of which have been implicated in L-LTP. Therefore, CREB is a strong candidate for the activation of CRE-driven gene expression observed during memory formation. Indeed, LTP and long-term, but not short-term, memory was defective in mice homozygous for a genetic deletion of alphadelta isoforms of CREB. However, in contrast to the evidence for the direct role of CREB in long-term synaptic plasticity in invertebrates, the situation in mammals is less clear. Both memory and the deficits in LTP in mice with genetic deletion of the alphadelta isoform of CREB have been found to be sensitive to gene dosage and genetic background, indicating that the activity of other genes can compensate for loss of CREB. In fact, the CREB partial knockout mice show strong upregulation of other CRE binding transcription factors. In addition, transgenic mice overexpressing a dominant-negative mutant of CREB in amygdala do not show any deficit in LTP or memory, although overexpression of CREB in the same region, using viral expression vectors, enhances memory (Barco, 2002 and references therein).

To explore the role played in hippocampal synaptic plasticity by CRE-driven genes, transgenic mice were generated in which it is possible to induce, in a regulated manner, the expression of a constitutively active CREB protein. This chimeric protein, VP16-CREB, was obtained by replacing the first transactivation domain of CREB with the acidic transactivation domain of Herpes simplex virus (HSV) VP16. Equivalent chimeric proteins bind to CRE sequences in tissue-specific promoters and behave like CREB activated by phosphorylation in both adipocytes and in cultured neurons (Barco, 2002 and references therein).

When expressed in the postsynaptic neurons of the Schaffer collateral pathway, VP16-CREB binds to CREs and regulates transcription of several downstream genes thought to play an important role in LTP and memory formation. Expression of VP16-CREB is sufficient to facilitate establishment of hippocampal L-LTP in an input-specific manner by enhancing synaptic capture, much as is the case with phospho CREB-1 in Aplysia. These findings provide an opportunity to examine some of the molecular mechanisms and phenotypic characteristics of capture in both transgenic and wild-type synapses (Barco, 2002).

Expression of VP16-CREB in postsynaptic neurons of the Schaffer collateral pathway facilitates the establishment of long-lasting LTP in hippocampal slices by allowing a single tetanic train, which normally produces E-LTP, to produce L-LTP. Although these data do not specify what role CREB-1 by itself plays in hippocampal LTP, it is clear from these results that the regulation of gene expression by the CREB family of transcription factors plays an essential role in the consolidation of LTP. These results also indicate that the input specificity of persistent changes in synaptic strength is determined not only by nuclear events but also by synaptic events, such as the interactions between plasticity proteins and synaptic tags. Thus, the activation of the CREB family of transcription factors that takes place under physiological conditions in hippocampal neurons might only reflect the potential to induce a lasting change, rather that the commitment to do so (Barco, 2002).

The data indicate that CRE-mediated transcription is one of the prerequisites for the consolidation of long-term synaptic changes. Specifically, VP16-CREB activity can lead to a cell-wide priming for LTP by seeding the synaptic terminals with proteins and mRNAs required for the stabilization and capture of L-LTP. These gene products can then be used productively for L-LTP when a given synapse is tagged by brief synaptic stimulation of the sort normally needed for E-LTP. Thus, in addition to CRE-mediated gene expression, synaptic tagging is a second prerequisite for LTP consolidation. This result is similar to that in Aplysia where the injection of phospho-CREB into sensory neurons paired to a single pulse of serotonin initiates a synapse-specific long-term facilitation that persists for several days (Barco, 2002).

In turn, the finding in VP16-CREB mice that a single tetanus to any branch can capture the long-term process has allowed an investigation of properties of 'synaptic capture' in transgenic animals and then to confirm its features in wild-type animals using two-pathway experiments. This analysis has allowed the delineation of seven key features of the capture process in mammalian hippocampus:

  1. L-LTP in one pathway appears not to affect basal transmission in the second pathway, but reduces in that pathway the threshold for establishing L-LTP (Barco, 2002).
  2. L-LTP mediated by synaptic capture is dependent on the activity of NMDA receptors, as is L-LTP mediated by four trains stimulation (Barco, 2002).
  3. L-LTP elicited by a single 100 Hz train in the second pathway is sustained for several hours and resistant to depotentiation, again resembling L-LTP obtained by four trains (Barco, 2002).
  4. Once CRE-driven gene products have been efficiently captured, further stimulation will not lead to additional sustained reinforcement of synaptic connections. This last feature is suggested by the ceiling effect observed after repeated tetanization and the occlusion of facilitated L-LTP by forskolin or after previous potentiation of the same pathway (Barco, 2002).
  5. The tag appears to involve PKA. The result with the PKA inhibitor KT5720 suggests that PKA is necessary for tagging the synapse, but is not necessary for maintenance of the captured potentiation process, a view consistent with the resistance of synaptic capture to depotentiation. Once the gene products have been captured, PKA activity is no longer needed and can be obliterated by activation of phosphatases during depotentiation with no effect on potentiated synapses. PKA seems to play a double role in the consolidation of LTP -- (1) it acts locally to mark a synapse that occurs following one tetanus; (2) it leads to nuclear activation with stronger activation following four tetani (Barco, 2002).
  6. New protein synthesis is necessary for maximal sustained potentiation in the second pathway. Translation is required whereas transcription is not for the facilitated LTP observed in VP16-CREB mice, suggesting that stabilization of the maximal synaptic change might require new protein synthesis. Among the mRNAs induced by VP16-CREB, some might be translated only after tagging of the synapse -- likely this translation will take place locally in stimulated synapses. These results are consistent with studies in Aplysia suggesting that tagging has two components: a mark for capture mediated by PKA and a mechanism for long-term stabilization mediated by local protein synthesis (Barco, 2002).
  7. The proteins necessary for priming the captured L-LTP represent CRE-driven gene products. As a consequence of VP16-CREB expression, certain CRE-driven genes, such as BDNF, dynorphin, and c-fos, are upregulated. Some of these genes, such as BDNF, play a role in synaptic plasticity. Other genes remain to be identified (Barco, 2002).

The gene expression necessary for the consolidation of LTP, for synaptic growth and for memory storage is regulated by two types of balancing mechanisms. In a given synaptic terminal, a balance between both phosphatase and kinase activities gates the synaptic signals that reaches the nucleus. In the nucleus, a second balance exists between transcriptional activators and repressors. This balance gates transcription activation. The use of a constitutively active form of CREB has allowed both gates to be bypassed and has allowed action directly on the nuclear output by activating expression of a specific set of genes required for L-LTP. The threshold for L-LTP in these mice is therefore determined locally by the threshold for synaptic tagging instead of that for nuclear activation (Barco, 2002).

What is the consequence of this threshold shift in memory and learning capability? In VP16-CREB mice, the frequency-response function reflecting the induction of persistent changes in synaptic efficacy shows an upward shift for a range of frequencies. For example, 1 Hz stimulation produces depression in wild-type mice but not in the mutants, and 10 Hz induces a much larger LTP in mutants than in wild-type mice. In addition, potentiated synapses in these animals cannot depotentiate, making these synaptic changes irreversible. Training these mice in a spatial learning task might cause too many synapses within the hippocampal network to become strongly and irreversibly potentiated, preventing the storage of new information. However, regulated expression of this protein shortly before or during the task might allow one trial learning and flashbulb memory. To clarify this question, VP16-CREB mice are currently being characterized in a number of behavioral tasks (Barco, 2002).

In conclusion, the line of transgenic mice described here represents a useful tool for deciphering the genetic program required for LTP consolidation. This tool has been used to provide some initial molecular insights into synaptic tagging for synapse-specific potentiation in the hippocampus, and results suggest that synaptic capture of CRE-driven gene products may be sufficient for the establishment of the late phase of LTP (Barco, 2002).

To examine the role of C/EBP-related transcription factors in long-term synaptic plasticity and memory storage, the tetracycline-regulated system was used and a broad dominant-negative inhibitor of C/EBP (EGFP-AZIP) was expressed in the forebrain of mice. This inhibitor preferentially interacts with several inhibiting isoforms of C/EBP. EGFP-AZIP also reduces the expression of ATF4, a distant member of the C/EBP family of transcription factors that is homologous to the Aplysia memory suppressor gene ApCREB-2. Consistent with the removal of inhibitory constraints on transcription, an increase was found in the pattern of gene transcripts in the hippocampus of EGFP-AZIP transgenic mice and both a reversibly enhanced hippocampal-based spatial memory and LTP. These results suggest that several proteins within the C/EBP family including ATF4 (CREB-2) act to constrain long-term synaptic changes and memory formation. Relief of this inhibition lowers the threshold for hippocampal-dependent long-term synaptic potentiation and memory storage in mice (Chen, 2003; full text of article).

Expression of VP16-CREB, a constitutively active form of CREB, in hippocampal neurons of the CA1 region lowers the threshold for eliciting the late, persistent phase of long-term potentiation (L-LTP) in the Schaffer collateral pathway. This VP16-CREB-mediated L-LTP differs from the conventional late phase of LTP in not being dependent on new transcription. This finding suggests that in the transgenic mice the mRNA transcript(s) encoding the protein(s) necessary for this form of L-LTP might already be present in CA1 neurons in the basal condition. High-density oligonucleotide arrays were used to identify the mRNAs differentially expressed in the hippocampus of transgenic and wild-type mice. Then the contributions were explored of the most prominent candidate genes revealed by this screening, namely prodynorphin, BDNF, and MHC class I molecules, to the facilitated LTP of VP16-CREB mice. It was found that the overexpression of brain-derived neurotrophic factor accounts for an important component of this phenotype (Barco, 2005).

This screen for overexpressed genes in VP16-CREB mice revealed more than a dozen ESTs encoding MHC I antigens. A role for MHC I molecules in synaptic plasticity was first suggested by an unbiased screen for genes involved in synaptic plasticity during development of the visual system. It has been proposed that MHC I expression has a postsynaptic role acting as a 'synaptic glue' in stabilizing appropriate synaptic connections, a view consistent with the results in VP16-CREB mice. The synaptic plasticity and LTP phenotype of mice deficient for MHC I signaling has been studied and it was found that in these mutants the normal segregation of retinal afferent axons in central targets during development is disrupted and synaptic plasticity in hippocampal neurons of adult animals is altered. In particular, LTP in the Schaffer collateral pathway is enhanced and LTD was absent. Based on these findings it has been concluded that MHC I molecules may also promote the elimination, by pruning, of inappropriate synapses during development and in the adult. This hypothesis, therefore, suggests that the enhanced LTP observed in MHC I- deficient mice reflects an absence of the normal pruning of synaptic connections in the hippocampus during development (Huh, 2000; Barco, 2005).

To identify the MHC I variants expressed in the hippocampus of VP16-CREB mice, RT-PCR of hippocampal mRNA was performed by using primers targeted to a segment that varies considerably among different class members of the MHC I antigen family. Both wild-type and transgenic mice expressed multiple MHC I antigens in hippocampal neurons, although the level of expression for the different MHC I subfamilies, both classical and nonclassical, was higher in mutant mice. CD3? is a component of the class I MHC receptor that is expressed in many cell types, including neurons, and its depletion disrupts MHC I signaling. An enhanced response to 1 × 100 Hz stimulation in CD36-/- mice is consistent with the enhanced response to 4 × 100 Hz stimulation reported by Huh (2000) and attributed by them to the absence of pruning of synaptic connections during development. The exploration of the consequences of MHC I disruption on enhanced LTP in VP16-CREB mice will require the generation of mutant mice in which MHC I deficiency occurs in the adult brain, thus bypassing the developmental consequences of its early disruption. MHC I antigens are expressed at low levels in the prenatal hippocampus, but their expression increases as hippocampal neurons mature, and these molecules may come to play a role in modulating adult synaptic plasticity (Barco, 2005).

The identification of BDNF in the unbiased screen for genes that are important in late-phase LTP reinforces the already strong evidence supporting a critical role for BDNF in some forms of L-LTP. Inhibition of endogenous BDNF or of the signaling through its TrkB receptor impairs the late phase of some forms of LTP, indicating that BDNF is required for its full expression. Furthermore, BDNF expression increases after induction of LTP in the CA1 region of the hippocampus and after learning-related events. The genomic structure of the BDNF gene is unusually complex. This gene contains multiple promoters that drive the expression of transcripts bearing different noncoding exons spliced upstream of a common 3′ exon that encompasses the entire encoding sequence. The detailed characterization of the rat BDNF gene has revealed the existence of at least four promoters located upstream of these alternatively spliced exons that enable a precise regulation of BDNF expression in different cell types or in response to different stimuli. The structure of the mouse gene is very similar to that of rat (93% homology), but the mouse gene appears to contain an additional promoter and alternative exon located in a sequence homologous to rat intron II (Barco, 2005 and references therein).

To identify the BDNF transcripts upregulated by VP16-CREB, a quantitative RT-PCR analysis with primer pairs specific for each one of the five upstream exons was performed. The expression of mRNAs bearing exons 1, 2, and 3 were found to be significantly increased in VP16-CREB mice whereas the expression of exons 4 and 5 were not affected. Interestingly, exons 1, 2, and 3 are clustered within 1.5 kb at the 5′ end of the gene and 15 kb apart from exons 4 and 5. The analysis of the genomic sequence upstream of exon 1 has revealed the existence of regulatory elements responsible for the Ca2+-mediated activation of PI, including a CRE site that overlaps with a USF binding element. CRE elements have also been reported upstream of PII, although their functionality has not been confirmed. Previous studies on the rat BDNF gene suggested that the promoters PI and PII shared common regulatory elements. Indeed, a neural-restrictive silencer element (NRSE) located in intron 1 contributes to the regulation of both promoters. Because the molecular interactions that underlie the activation of transcription by VP16-CREB and phospho-CREB are different, the transactivation driven by the VP16 domain might have a broader range than that of CREB. It is therefore possible that the binding of VP16-CREB to the CRE element in PI also activates the expression driven by promoter PII and even PIII, which are located relatively close downstream of a functional CRE site but have not been previously reported as being regulated by CREB. Strikingly, no effect of VP16-CREB expression on the expression of BDNF driven by promoter PIV was detected, in spite of the presence of a conserved CRE site upstream of this promoter. Experiments in rat neuronal cultures have found that membrane depolarization induces the transcription from both P1 and PIII (PIV in the mouse). Some of these experiments suggest that PIII responds to depolarization but not to an increase of internal cAMP, suggesting that the activation of CREB is not sufficient to drive rat BDNF PIII expression, a view that might explain the negative results for mouse promoter PIV (Barco, 2005 and references therein).

To evaluate the component of the facilitated L-LTP in VP16-CREB mice that is due to an enhanced expression of BDNF, two experiments were carried out. (1) Hippocampal slices from VP16-CREB mice and from wild-type littermates were incubated with the BDNF scavenger TrkB receptor body (TrkB-Fc). This protein scavenges unbound BDNF and blocks its interaction with cellular receptors. Although normal E-LTP was detected in wild-type mice in the presence of this protein, the late expression of the enhanced LTP of VP16-CREB mice was significantly reduced after incubation with TrkB-Fc. (2) BDNF heterozygous mice (BDNF+/−), in which expression of BDNF mRNA is reduced to about half the level of wild-type mice, were crossed with VP16-CREB transgenics. It was found that genetic reduction of BDNF expression largely blocks the enhanced LTP phenotype observed in VP16-CREB mice but does not affect E-LTP in wild-type mice. These two experiments taken together strongly indicate that a significant component contributing to the L-LTP phenotype of VP16-CREB mice results from overexpression of BDNF (Barco, 2005).

In an earlier study of the VP16-CREB mice, it was proposed that the activity of constitutively active CREB leads to a cell-wide distribution of CRE-driven gene products in the postsynaptic CA1 neuron that primes the synapses for subsequent synapse-specific capture of L-LTP by a single tetanus, so giving rise to the facilitated L-LTP phenotype. This study shows that BDNF is one of these CRE-driven gene products and can, by itself, be responsible for an important component of the facilitated L-LTP. Because facilitated L-LTP in VP16-CREB mice can be equivalent in molecular terms to synaptic capture mediated L-LTP in wild-type animals, BDNF may also contribute to the processes of synaptic tagging and capture in normal mice. To test this prediction, two-pathway experiments were carried out in slices of BDNF heterozygous mice. Two independent inputs (S1 and S2) to the same CA1 neuronal population were stimulated and it was found that synaptic capture was dramatically impaired in BDNF+/− mice (Barco, 2005).

To further explore the role of BDNF in synaptic capture, mice were used in which the genetic deletion was restricted either to the entire forebrain, including both the CA3 and CA1 pyramidal neurons of the hippocampus, or only to the postsynaptic CA1 neurons. Thus, the individual contributions of pre- and post-synaptic sources of BDNF could be assessed. It was found that both types of BDNF mutants had normal 4 × 100 Hz-induced L-LTP in S1 but exhibited a defect in synaptic capture in S2. This defect was more pronounced in the case of the forebrain BDNF−/−(CA3-CA1) mutants, suggesting that BDNF may play a dual role in synaptic capture. (1) The decay of the late phase observed in the more restricted BDNF−/−(CA1) mutants suggests a late, postsynaptic role in the maintenance of captured L-LTP, a view consistent with the results in VP16-CREB and BDNF+/− mice, with the facilitation of the late phase of LTP by exogenous BDNF and with the late enhancement of BDNF expression observed in postsynaptic CA1 neurons after L-LTP induction. (2) The rapid decay of captured LTP in slices lacking presynaptic BDNF suggests that the presynaptic release of BDNF into the synaptic cleft after tetanic stimulation may participate in the postsynaptic tagging of the synapse (Barco, 2005).

Although the role of BDNF in hippocampal LTP has been extensively studied, some aspects of its function are just beginning to be understood. For example, BDNF seems to participate only in some forms of LTP. Also, the presynaptic or postsynaptic source of BDNF and its cellular targets in different forms of LTP are only now being delineated. This study further clarifies these issues in two ways. (1) Similar to experiments using exogenously applied BDNF, the experiments in VP16-CREB mice suggest that, regardless of the site of BDNF release, increased levels of BDNF in the synaptic cleft lead to a facilitation of LTP in CA3-CA1 synapses, probably by acting on both pre- and post-synaptic targets. An enhanced release of BDNF accumulated in postsynaptic spines of CA1 neurons of VP16-CREB mice after tetanic stimulation may contribute to sustaining an otherwise transient potentiation by stimulating local protein synthesis or enhancing the neurotransmitter release from presynaptic CA3 neurons. (2) These experiments on synaptic capture suggest distinct roles for pre- or post-synaptically released BDNF. Interestingly, a recent study in neuronal primary cultures has demonstrated that BDNF-induced plasticity exhibits a bimodal profile and has an early presynaptic component and a later postsynaptic component. These results indicate that a similar bimodal action can be also observed in hippocampal slices. Thus, presynaptically released BDNF contributes to the formation of those forms of LTP that recruit a presynaptic component and might participate in tagging the synapse for subsequent capture of late-phase LTP, while postsynaptically released BDNF might contribute to the maintenance of different forms of L-LTP at late times, including facilitated L-LTP in VP16-CREB mice and synaptically captured L-LTP (Barco, 2005).

LTP is not a unitary phenomenon, but a family of different processes. It is most likely for this reason that hundreds of molecules and more than a dozen molecular pathways have been implicated in the induction, expression, and maintenance of the family of LTP processes. The list of candidate genes revealed by this study suggests the existence of complex interactions between various gene networks. Nevertheless, this analysis provides insights into the molecular mechanism underlying a specific aspect of the phenotype of VP16-CREB mice: their enhanced L-LTP in the Schaffer collateral pathway. It was found that the overexpression of BDNF by itself importantly contributes to the L-LTP phenotype of VP16-CREB mice and that reducing BDNF compromises the phenotype. Further physiological studies might reveal the contribution of other candidate genes to other aspects of the complex phenotype of VP16-CREB mice and explain the causes and consequences of their altered expression (Barco, 2005).

The transcription factor CREB is critical for several forms of experience-dependent plasticity in a range of species and is commonly activated in neurons by calcium/calmodulin-dependent protein kinase IV (CaMKIV). Surprisingly, little is known about the neural circuit adaptations caused by activation of CaMKIV and CREB. Viral-mediated gene transfer in vivo was used to examine the consequences of acute expression of constitutively active forms of CaMKIV and CREB on synaptic function in the rodent hippocampus. Acute expression of active CaMKIV or CREB causes an enhancement of both NMDA receptor-mediated synaptic responses and long-term potentiation (LTP). This is accompanied by electrophysiological and morphological changes consistent with the generation of 'silent synapses', which provide an ideal substrate for further experience-dependent modifications of neural circuitry and which may also be important for the consolidation of long-term synaptic plasticity and memories (Marie, 2005).

A simple but compelling hypothesis to explain the increases in NMDAR-mediated transmission and LTP (without effects on LTD) due to expression of CaMKIVCA and CREBCA is that activation of CREB leads to the generation of so-called silent synapses. These synapses are termed silent because they contain only NMDARs and no (or undetectable levels of) AMPARs and thus are functionally silent at resting membrane potentials. They are 'unsilenced' during LTP by the insertion of AMPARs into the synaptic plasma membrane. More silent synapses would, by definition, increase NMDAR-mediated synaptic responses without affecting AMPAR EPSCs and also would enhance the magnitude of LTP by providing synapses that had not previously been potentiated. Furthermore, since they are silent, such synapses cannot express LTD (Marie, 2005 and references therein).

Two different electrophysiological assays were carried out to determine whether CREBCA-expressing neurons expressed more silent synapses compared to neighboring uninfected control cells. The first approach, which provided important initial support for the existence of silent synapses and their unsilencing during LTP, involved comparing the coefficient of variation (CV) of the AMPAR EPSCs and NMDAR EPSCs. The CV measures the trial-to-trial variability of synaptic responses and varies inversely with quantal content. In general, the lower the CV of EPSCs the larger the number of synapses that contribute to the measured synaptic response. Thus, when a mixture of silent and functional synapses contribute to an evoked synaptic current, the CV of NMDAR EPSCs is less than that of AMPAR EPSCs. It is predicted that if the increase in NMDAR-mediated synaptic transmission in CREBCA-expressing neurons is due to an increase in silent synapses, a decrease should be observed in the within-cell ratio of the CV of NMDAR EPSCs to that of AMPAR EPSCs. Consistent with this prediction, this CV ratio was significantly smaller in CREBCA-expressing cells. A comparison of the raw CV measurements between the two cell populations suggests that this change is due to a decrease in the CV of NMDAR EPSCs in the CREBCA-expressing neurons, not an increase in the CV of AMPAR EPSCs. A similar decrease in the ratio of the CV of NMDAR EPSCs to that of AMPAR EPSCs was observed in neurons expressing CaMKIVCA compared to uninfected controls and CaMKIVDN-expressing neurons (Marie, 2005).

The second approach used minimal stimulation techniques to produce a mixture of synaptic responses and failures. When small numbers of silent and functional synapses are activated, the proportion of failures decreases (i.e., the proportion of successes increases) when EPSCs are sampled at +40 mV versus -65 mV, because silent synapses do not contribute to EPSCs at hyperpolarized membrane potentials but do contribute at depolarized membrane potentials at which NMDARs can pass current. This difference in success rates at depolarized versus hyperpolarized membrane potentials is another important piece of evidence that supports the existence of silent synapses. In uninfected cells, there was a modest increase in the percent of successes at +40 mV when compared to -65 mV. This difference in success rate at the two membrane potentials is significantly larger in cells expressing CREBCA. By comparing the success rates at the two membrane potentials and assuming that the average probability of release at silent and functional synapses is similar, it is possible to directly estimate the proportion of silent synapses per cell. Such a calculation resulted in an estimate of the proportion of silent synapses to be 19% ± 7.2% in uninfected neurons and 41% ± 4.8% in CREBCA-infected neurons. This difference in the proportion of silent synapses can be seen most dramatically when CREBCA-expressing cells and neighboring uninfected cells in the same slice are directly compared. Together, the changes in the CV ratios and the relative change in success rates at depolarized versus hyperpolarized membrane potentials in CREBCA-expressing neurons provide strong evidence supporting the hypothesis that activation of CREB caused generation of silent synapses (Marie, 2005).

Silent synapses could be generated by one of two general mechanisms: the removal of AMPARs from synapses that originally contained both AMPARs and NMDARs or the production of new synapses that contain only NMDARs. Since there was no change in AMPAR-mediated transmission in CREBCA neurons, the generation of silent synapses is probably due to the creation of new silent synapses rather than the silencing of preexisting AMPAR-containing synapses. To further test this hypothesis, an examination was made to determine whether the density of dendritic spines, the postsynaptic site of excitatory synapses, was increased in CaMKIVCA- and CREBCA-expressing neurons. Neurons were filled with the fluorophore Alexa 568 via the whole-cell patch-clamp recording pipette, and three-dimensional images of apical secondary dendrites were reconstructed from confocal Z stacks. Images were acquired and analyzed blind without knowledge of the protein being expressed. All types of spines were counted, including stubby, mushroom, and thin spines. A visual count of the number of spines per unit length of dendrite revealed that spine density was significantly increased in both CaMKIVCA- and CREBCA-expressing neurons compared to neurons expressing GFP. This morphological change is consistent with the hypothesis that these molecular manipulations caused the generation of new silent synapses (Marie, 2005).

If CREB activation leads to the generation of silent synapses, cells expressing CREBCA, when compared to control cells, should have a relative increase in the number of synapses containing NMDARs but not the number of synapses containing AMPARs. In a final set of experiments, this prediction was tested using dissociated hippocampal neuronal cultures, a preparation in which immunocytochemical visualization of endogenous NMDARs and AMPAR at individual synapses is possible. Specifically, the density of synaptic NMDAR and AMPAR puncta was compared; these were defined by colocalization with the presynaptic active zone proteins Piccolo or Bassoon, between cells infected with the CREBCA virus versus cells infected with the virus that expresses GFP alone. Consistent with the hypothesis, neurons expressing CREBCA have significantly more synaptic NMDAR puncta per unit length of dendrite, while there was no significant change in the density of synaptic AMPAR puncta in the CREBCA-expressing neurons (Marie, 2005).

Since silent synapses provide an ideal substrate for LTP, the increase in LTP magnitude due to CaMKIV and CREB can be explained by the increase in the proportion of silent synapses, although additional mechanisms cannot by ruled out. For example, LTP is also enhanced in transgenic mice expressing active CREB for more prolonged durations (3-7 days), an effect that was attributed to a CREB-mediated enhancement of the production of proteins that need to be 'captured' by synapses to consolidate or maintain LTP. While the two data sets and hypotheses are not mutually exclusive, and both mechanisms may importantly contribute to the synaptic adaptations caused by CREB activation, many of the results in this previous work can be explained by the CREB-dependent generation of silent synapses (Marie, 2005).

CaMKIVCA caused many of the same synaptic changes as CREBCA but also increased AMPAR-mediated synaptic transmission, an effect that could be due to actions of CaMKIV on other substrates. A hypothesis that might explain all of these results is that expression of CaMKIVCA, in addition to increasing the proportion of silent synapses via activation of CREB, also stimulates the growth of the dendritic arbor, an effect that in cultured neurons has been shown to require factors in addition to CREB. This would account for the increase in AMPAR EPSCs and mEPSC frequency in CaMKIVCA-expressing neurons, assuming that more functional synapses occurred on the extended dendritic tree. It is noted, however, that no direct evidence was presented that the synaptic effects of CaMKIVCA, which mimic those caused by CREBCA, are in fact due to phosphorylation and activation of CREB. It is conceivable that, despite their very similar effects, CREB does not participate in mediating the actions of CAMKIVCA in CA1 pyramidal cells in vivo. Other work, specifically on the role of CaMKIV in synaptic plasticity, used mutant mice either lacking CaMKIV or expressing a dominant-negative form of CaMKIV and presented different conclusions about its role in the late versus the early phases of LTP. No significant effects of CAMIVDN on the first hour of LTP were observed, and the effects of CaMKIVCA were most apparent 50-60 min after the induction protocol, results that are most consistent with a role for CaMKIV signaling in the later phases or so-called maintenance of LTP (Marie, 2005).

The experiments using CaMKIVCA do not directly address the issue of which signaling pathways normally lead to CREB activation during various forms of synaptic and experience-dependent plasticity. While CaMKIV is certainly a prime candidate, other signaling cascades can activate CREB and thus can presumably lead to the synaptic modifications reported in this study. Another limitation of these experiments is that, by necessity, the synaptic consequences of overexpression of active CaMKIV and CREB were studied, not the consequences of the activation of endogenous CaMKIV or CREB. While overexpression of recombinant proteins has been a valuable approach that is routinely used to examine the functions of individual proteins, it is possible that the effects caused by overexpression of the active CaMKIV and CREB do not exactly mimic those caused by the activation of endogenous CaMKIV and CREB. To address that issue, however, will require the ability to identify, in living tissue, the individual neurons in which CaMKIV and CREB have been activated by some in vivo experience. This is a particularly challenging task in a structure like the hippocampus, in which the neural representation of experiences, such as exploring a novel environment, appear to be encoded sparsely such that only a modest proportion (30%-40%) of CA1 pyramidal cells are activated and presumably modified by signaling to the nucleus. Thus, at the present time, it is not possible to directly demonstrate that activation of endogenous CaMKIV and CREB in CA1 pyramidal cells by some in vivo experience causes the same effects as those reported in this study (Marie, 2005).

What advantages for experience-dependent plasticity might the CREB-mediated generation of new silent synapses provide? One possibility is that silent synapses provide a new substrate that will be devoted entirely to the neural circuit adaptations that store new memories or experiences. According to this hypothesis, the activation of CREB would not only contribute to the maintenance of the synaptic weight changes that mediate long-lasting memories but also, simultaneously, would provide neurons with naive synapses ideal for participation in future neural circuit adaptations. Indeed, the computational advantages of such structural changes in circuit wiring have recently been highlighted. The morphological changes in spine number and shape that have been reported to accompany LTP can be viewed as consistent with this hypothesis. When single synapses/spines are activated using caged glutamate, a stimulus that rapidly elicits LTP but is unlikely to potently activate nuclear signaling, preexisting spines enlarge, but new spines or filopodia are not generated. In contrast, strong tetanic activation of multiple synapses, an ideal stimulus for signaling to the nucleus, results in the generation of new spines or filopodia. This process requires at least 20 min and therefore cannot be part of the initial increase in synaptic strength during LTP. Instead, one function of CREB activation may be to stabilize and maintain these new postsynaptic structures such that they become silent synapses. Consistent with this hypothesis, it has been reported that activation of CREB is required for the increase in spine density in cultured neurons caused by prolonged exposure to estradiol (Marie, 2005 and references therein).

A related but distinct hypothesis is that the generation of new silent synapses may be required for the consolidation of long-term memories, a process in which CREB has been implicated. This might explain the intriguing finding that consolidation of long-term memories requires the reactivation of NMDARs after initial learning. In this case, the silent synapses generated by CREB activation provide a critical substrate for the further NMDAR-dependent synaptic modifications that are required for the long-term maintenance/consolidation of recently formed memories (Marie, 2005).

It is worthwhile noting that the significance of these results does not depend on activation of CREB playing a critical role in the late phase and maintenance of LTP, a hypothesis that remains controversial. CREB-dependent transcription has been implicated in many forms of adaptive experience-dependent plasticity as well as in pathological states such as addiction and depression. Thus, the present findings have wide-ranging implications, especially if other cell types in the mammalian brain respond to CREB activation in the same manner. The approach taken here, which allows for temporal and spatial control of transgene expression in a cell-restricted manner, should prove useful for further investigation of the critical question of which CRE-driven gene products mediate the synaptic consequences of CREB activation (Marie, 2005).

Late-phase long-term potentiation (L-LTP) and long-term memory depend on the transcription of mRNA of CRE-driven genes and synthesis of proteins. However, how synaptic signals propagate to the nucleus is unclear. This study reports that the CREB coactivator TORC1 (transducer of regulated CREB activity 1) undergoes neuronal activity-induced translocation from the cytoplasm to the nucleus, a process required for CRE-dependent gene expression and L-LTP. Overexpressing a dominant-negative form of TORC1 or down-regulating TORC1 expression prevented activity-dependent transcription of CREB target genes in cultured hippocampal neurons, while overexpressing a wild-type form of TORC1 facilitated basal and activity-induced transcription of CREB target genes. Furthermore, overexpressing the dominant-negative form of TORC1 suppressed the maintenance of L-LTP without affecting early-phase LTP, while overexpressing the wild-type form of TORC1 facilitated the induction of L-LTP in hippocampal slices. These results indicate that TORC1 is essential for CRE-driven gene expression and maintenance of long-term synaptic potentiation (Zhou, 2006).

A key feature of memory processes is to link different input signals by association and to preserve this coupling at the level of synaptic connections. Late-phase long-term potentiation (L-LTP), a form of synaptic plasticity thought to encode long-term memory, requires gene transcription and protein synthesis. This study reports that a recently cloned coactivator of cAMP-response element-binding protein (CREB), called transducer of regulated CREB activity 1 (TORC1), contributes to this process by sensing the coincidence of calcium and cAMP signals in neurons and by converting it into a transcriptional response that leads to the synthesis of factors required for enhanced synaptic transmission. Evidence is provided that TORC1 is involved in L-LTP maintenance at the Schaffer collateral-CA1 synapses in the hippocampus (Kovacs, 2007).

The requirement for enhanced CREB1 expression in consolidation of long-term synaptic facilitation and long-term excitability in sensory neurons of Aplysia

Accumulating evidence suggests that the transcriptional activator cAMP response element-binding protein 1 (CREB1) is important for serotonin (5-HT)-induced long-term facilitation (LTF) of the sensorimotor synapse in Aplysia. Moreover, creb1 is among the genes activated by CREB1, suggesting a role for this protein beyond the induction phase of LTF. The time course of the requirement for CREB1 synthesis in the consolidation of long-term facilitation was examined using RNA interference techniques in sensorimotor cocultures. Injection of CREB1 small-interfering RNA (siRNA) immediately or 10 h after 5-HT treatment blocked LTF when measured at 24 and 48 h after treatment. In contrast, CREB1 siRNA did not block LTF when injected 16 h after 5-HT treatment. These results demonstrate that creb1 expression must be sustained for a relatively long time to support the consolidation of LTF. In addition, LTF is also accompanied by a long-term increase in the excitability (LTE) of sensory neurons (SNs). Because LTE was observed in the isolated SN after 5-HT treatment, this long-term change was intrinsic to that element of the circuit. LTE was blocked when CREB1 siRNA was injected into isolated SNs immediately after 5-HT treatment. These data suggest that 5-HT-induced CREB1 synthesis is required for consolidation of both LTF and LTE (Liu, 2011).

Upregulation of CREB-mediated transcription enhances both short- and long-term memory

Unraveling the mechanisms by which the molecular manipulation of genes of interest enhances cognitive function is important to establish genetic therapies for cognitive disorders. Although CREB is thought to positively regulate formation of long-term memory (LTM), gain-of-function effects of CREB remain poorly understood, especially at the behavioral level. To address this, four lines of transgenic mice were generated expressing dominant active CREB mutants (CREB-Y134F or CREB-DIEDML) in the forebrain; these lines exhibited moderate upregulation of CREB activity. The transgenic lines improved not only LTM but also long-lasting long-term potentiation in the CA1 area in the hippocampus. However, enhanced short-term memory (STM) was also observed in contextual fear-conditioning and social recognition tasks. Enhanced LTM and STM could be dissociated behaviorally in these four lines of transgenic mice, suggesting that the underlying mechanism for enhanced STM and LTM are distinct. LTM enhancement seems to be attributable to the improvement of memory consolidation by the upregulation of CREB transcriptional activity, whereas higher basal levels of BDNF, a CREB target gene, predicted enhanced shorter-term memory. The importance of BDNF in STM was verified by microinfusing BDNF or BDNF inhibitors into the hippocampus of wild-type or transgenic mice. Additionally, increasing BDNF further enhanced LTM in one of the lines of transgenic mice that displayed a normal BDNF level but enhanced LTM, suggesting that upregulation of BDNF and CREB activity cooperatively enhances LTM formation. These findings suggest that CREB positively regulates memory consolidation and affects memory performance by regulating BDNF expression (Suzuki, 2011).

Creb and addiction

Dopamine is a classical neurotransmitter that can influence neurologic functions ranging from movement to emotion, memory, and reinforcement. The behavioral aspects of reward and motivation are known to be modulated by the nigrostriatal and mesolimbic pathways. Both nigrostriatal and mesolimbic pathways target the basal ganglia, extensively interconnected subcortical nuclei found in the telencephalon. The primary striatal targets of these dopamine-containing systems are different. The nigrostriatal system (originating in one area of the substantia nigra) primarily innervates the caudoputamen (found in the dorsal striatum). The mesolimbic system (originating from a different area of the substantia nigra) innervates a number of forebrain structures, including the ventral striatum, which is formed by the nucleus accumbens and olfactory tubercle. Although both the dorsal and ventral striatum (both part of the basal ganglia) share a striatal architecture, and are targets of dopamine-containing afferents from the midbrain, they differ from one another in neurochemical organization and in their connectivity with other brain regions. The dorsal striatum is engaged primarily in processing information from the cerebral cortex and thalamus and is involved in motor control, aspects of cognitive control, and some forms of learning and memory. The ventral striatum receives inputs from limbic structures, including the amygdala, the hippocampus, and the limbic prefrontal cortex; it projects back to limbic structures via the ventral pallidum. These limbic-associated circuits are thought to underlie motivational and viscero-affective aspects of neurologic function served by the ventral striatum. The midbrain dopamine pathways (originating in the substantia nigra) innervating the dorsal and ventral striatum are critically involved in controlling these functions, including, for the ventral striatum, the reinforcing properties of psychostimulant and other drugs (Liu, 1998 and references).

A distinct feature of much reinforcement-based learning is that the modified behaviors develop with time and are highly sensitive to the temporal organization of events. The time-dependent nature of the learning suggests that the underlying neural plasticity may share similar time sensitivity at the molecular level. The mesolimbic dopamine-ventral striatum system undergoes a series of cellular and molecular changes in response to psychoactive and addictive drugs. The cAMP response element-binding protein (CREB) is an activity-dependent transcription factor that is involved in neural plasticity. The kinetics of CREB phosphorylation are thought to be important for gene activation, with sustained phosphorylation being associated with downstream gene expression. If so, the duration of CREB phosphorylation might serve as an indicator for time-sensitive plastic changes in neurons. To screen for regions potentially involved in dopamine-mediated plasticity in the basal ganglia, organotypic slice cultures were used to study the patterns of dopamine- and calcium-mediated CREB phosphorylation in the major subdivisions of the striatum. Different durations of CREB phosphorylation are evoked in the dorsal and ventral striatum by activation of dopamine D1-class receptors, known to act through adenylate cyclase. The same D1 stimulus elicits (1) transient phosphorylation in the matrix of the dorsal striatum; (2) sustained phosphorylation in limbic-related structures (ventral striatum) including striosomes, the nucleus accumbens (ventral striatum), the fundus striati, and the bed nucleus of the stria terminalis; and (3) prolonged phosphorylation (up to 4 hr or more) in cellular islands in the olfactory tubercle. Elevation of Ca2+ influx by stimulation of L-type Ca2+ channels, NMDA, or KCl induces strong CREB phosphorylation in the dorsal striatum but not in the olfactory tubercle. These findings differentiate the response of CREB to dopamine and calcium signals in different striatal regions and suggest that dopamine-mediated CREB phosphorylation is persistent in limbic-related regions of the neonatal basal ganglia. The downstream effects activated by persistent CREB phosphorylation may include time-sensitive neuroplasticity modulated by dopamine (Liu, 1998).

Chronic morphine administration increases levels of adenylyl cyclase and cAMP-dependent protein kinase (PKA) activity in the locus coeruleus (LC), which contributes to the severalfold activation of LC neurons that occurs during opiate withdrawal. A role for the transcription factor cAMP response element-binding protein (CREB) in mediating the opiate-induced upregulation of the cAMP pathway has been suggested, but direct evidence is lacking. The morphine-induced increases in adenylyl cyclase and PKA activity in the LC are associated with selective increases in levels of immunoreactivity of types I and VIII adenylyl cyclase and of the catalytic and type II regulatory subunits of PKA. Antisense oligonucleotides directed against CREB were used to study the role of this transcription factor in mediating these effects. Infusion (5 d) of CREB antisense oligonucleotide directly into the LC significantly reduces levels of CREB immunoreactivity. This effect is sequence-specific and not associated with detectable toxicity. CREB antisense oligonucleotide infusions completely block the morphine-induced upregulation of type VIII adenylyl cyclase but not of PKA. The infusions also blocked the morphine-induced upregulation of tyrosine hydroxylase but not of Gialpha, two other proteins induced in the LC by chronic morphine treatment. Electrophysiological studies reveal that intra-LC antisense oligonucleotide infusions completely prevent the morphine-induced increase in spontaneous firing rates of LC neurons in brain slices. This blockade is completely reversed by addition of 8-bromo-cAMP (which activates PKA) but not by addition of forskolin (which activates adenylyl cyclase). Intra-LC infusions of CREB antisense oligonucleotide also reduces the development of physical dependence to opiates, based on attenuation of opiate withdrawal. Together, these findings provide the first direct evidence that CREB mediates the morphine-induced upregulation of specific components of the cAMP pathway in the LC that contribute to physical opiate dependence (Lane-Ladd, 1997).

Creb and tumorogenesis

Members of the ATF/CREB family of eukaryotic transcription factors contain the bZIP structural motif and mediate their transcriptional activities via heterodimerization with ATF and AP-1 family members. Quenching of CREB-associated proteins by a dominant-negative CREB (KCREB) that is mutated within its DNA-binding domain decreases the radiation resistance of human melanoma cells. The purpose of this study was to determine the role of CREB in tumor growth and metastasis in human melanoma using KCREB. Highly metastatic MeWo human melanoma cells were transfected with the KCREB expression vector and subsequently analysed for changes in their tumorigenic and metastatic potential. Expression of KCREB in MeWo human cells decreases their tumorigenic and metastatic potential in nude mice, as compared with parental and control transfected cells. The KCREB-transfected cells display downregulation of 72 kDa collagenase type IV (MMP-2) mRNA expression and activity and decreased invasiveness through Matrigel-coated filters. Transcriptional activities mediated by the CAT gene driven by the MMP-2 promoter are decreased 14- to 45-fold in KCREB-transfected cells. The cell-surface adhesion molecule MCAM/MUC18, involved in the metastasis of human melanoma, is downregulated in the KCREB-transfected cells. These data indicate that CREB and its associated proteins play an important role through their transcriptional activities in the acquisition of the metastatic phenotype of human melanoma cells (Xie, 1997).

Table of contents

CrebB-17A: Biological Overview | Regulation | Developmental Biology | Effects of Mutation | References

Home page: The Interactive Fly © 1995, 1996 Thomas B. Brody, Ph.D.

The Interactive Fly resides on the
Society for Developmental Biology's Web server.