A Struggle Against Forgetting

One of the most remarkable characteristics of living organisms is their ability to alter their behavior by learning. Human beings excel at processing new information, retaining it in their memory, and using it to develop new ideas. Studies in the mollusk Aplysia showed that learning involves the strengthening and increased effectiveness of preexisting synapses. Over time, it has become clear that synaptic plasticity is the fundamental mechanism for learning and memory. Short-term memory, lasting minutes to hours, involves to a large extent a serotonin-mediated release of cAMP, which activates protein kinase A (PKA), leading to the covalent modifications of S-type K channels and delayed-rectifier K channels. These modifications, which cause spike broadening, an increase in Ca influx, and enhanced transmitter release, do not require protein synthesis. Long-term memory, lasting days to weeks, differs from short-term memory and requires the synthesis of new proteins. Studies in Aplysia, the fruit fly Drosophila, and later in rodents have revealed that this process involves activation of the cAMP response element binding protein (CREB)mediated transcriptional cascade. Inactivation of CREB or inhibition of its phosphorylation interferes with learning and memory (see reviews1,2). The CREB family of transcription factors includes CREB, CREM (cAMP response element modulator), and ATF-1 (activating transcription factor 1). Each member of the family contains a basic DNA-binding domain at the C-terminus and a leucine zipper domain that mediates homodimerization or heterodimerization of these factors (bZIP). All three factors bind to a cAMP response cis-regulatory element (CRE) that consists of the sequence TGACGTCA. The core sequence CGTCA is conserved in almost all CREs and is essential for their function. CREB contains several additional domains including a kinase-inducible domain (KID) that separates the two glutamine-rich transcription activator domains, Q1 and Q2/CAD. Q2/CAD interacts with components of the basal transcriptional machinery and is necessary and sufficient for basal CRE-mediated gene expression. Phosphorylation of Ser133 within the KID facilitates the binding of CREBbinding protein (CBP) and p300, both of which serve as coactivators of the transcriptional machinery. Through its endogenous histone acetyl transferase activity, CBP acetylates histones, which leads to the unraveling of the chromatin and facilitates transcription. CREB has two main isoforms that function as transcription activators (CREB and CREB ). Alternative splicing of the CREM transcript determines whether CREM isoforms function as transcriptional activators (CREM and CREM ) or repressors (S-CREM or ICER). Thus, isoforms that lack the activation domains will bind to the CRE and repress transcription. This model of CREB/CREM activation is simple but far from conclusive, since not all Ser133 phosphorylation events lead to transcriptional activation. Of note, in all these models, the steady-state levels of CREB polypeptides in the cells remain unchanged (see reviews2–4). By contrast, the CREM isoform ICER (inducible cAMP early repressor) is a potent inhibitor of CREBand CREM-mediated transcription and is rapidly induced by cAMP. On induction, ICER polypeptides are rapidly degraded by the ubiquitin-proteasome pathway.5 In addition to cAMP/PKA-inducible activation of CREB, Greenberg and colleagues have shown that CREB can function as a Ca -inducible transcription factor. Membrane depolarization in neurons leads to calcium entry through L-type Ca channels (ICa,L). These calcium cations can bind to calmodulin (CaM) and activate CaMKI, CaMKII, and CaMKIV, each of which can phosphorylate Ser133. Calcium can also activate the Ras/ERK pathway with similar consequences. Neuronal growth factors can activate the CREB pathway through receptor tyrosine kinases, which after ligand binding and dimerization can also activate the Ras/ERK pathway. In addition, the phosphatidylinositol 3-kinase/Akt pathway can activate CREB-mediated transcription after stimulation with IGF-1. Finally, the stress-activated kinase, SAPK2/p38MAPK can also activate the CREB-mediated pathway through several downstream kinases (see reviews2–4). There are more than 500 genes that contain a cis-CRE regulatory element in their promoter region. The list of channels that contain CRE includes Kv1.5, Kv3.1, Cav1.2, aquaporin 2, and CFTR.2,6 Na /K -ATPase also contains CRE. Of note, some of the genes that contain CRE are critical to cell survival. Indeed, in neuronal cells, CREB phosphorylation correlates with neuronal survival. In PC12 cells, hypoxia induces the expression of CREB-dependent Bcl-2 and protects the cells from apoptosis. Similarly, dentate gyrus neurons are protected by CREB-mediated induction of Bcl2.7,8 In the heart, the overexpression of a dominant-negative CREB mutant where Ser133 was mutated to alanine resulted in transgenic mice with dilated cardiomyopathy with extenThe opinions expressed in this editorial are not necessarily those of the editors or of the American Heart Association. From the Bioelectricity Laboratory (E.J.F., K.R., G.K.), Cardiovascular Division, Brigham and Women’s Hospital, Harvard Medical School, Boston, Mass; Cardiovascular Engineering Inc (G.F.M.), Holliston, Mass. Correspondence to Gideon Koren, MD, Bioelectricity Laboratory, Brigham and Women’s Hospital, Cardiovascular Division, 75 Francis St, Boston, MA 02115. E-mail [email protected] (Circ Res. 2003;93:384-386.) © 2003 American Heart Association, Inc.