The CREB Pathway

8. The CREB transcription factor

The identification and characterization of the CREB transcription factor resulted from studies

on the cAMP regulation as a response to hormonal stimulus (reviewed in (Shaywitz, 1999).

KID

Figure 7 : Evolutionary conservation of the CREB transcription factor family. Amino-acid sequence similarity between putative hydra CREB proteins and their vertebrate CREB and CREM counterparts. The predicted 248 amino-acid sequence of CREBα from Chlorohydra viridissima (cv.) is aligned with CREBβ (199 aa long) and CREBα (partial) from Hydra vulgaris (hv.), CREB∆ from human (Hoeffler, 1988 38), CREBα from rat (Gonzalez, 1989 148), CREMβ and CREMτ (Foulkes, 1992 149) from mouse. Grey shaded regions show identities or conservative changes between hydra and vertebrate sequences; points represent gaps introduced in the sequences to optimize the alignment. Boxes contain highly conserved regions that contain the target phosphorylation sites (KID), the DNA-binding region (basic region) and the dimerization domain (leucine zipper). The inverted triangle under the basic region indicates the position of the conserved splice site (according to Galliot et.al., 1995).

Glucagon or Epinephrine exposure lead to an increase in the intracellular cAMP levels, that in turn can enhance the transcription of the somatostatin and other neuropeptides gene (Montminy, 1986; Goodman, 1990). Systematic deletion analysis of the promoter region of the somatostatin gene revealed the functional role of the cAMP response element (CRE) in the

elicited response. This is an 8 bp palindromic sequence 5`-TGACGTCA-3` that is critical for cAMP inducibility.

Using CRE affinity chromatography, the CREB protein was isolated as the main CRE-binding protein among PC12 whole cell extracts, migrating as a 43-kDa protein (Montminy, 1987). The human CREB gene encoding a 327-residue protein was isolated (Hoeffler, 1988) and soon after the rat gene, coding for a 341-residue protein (Gonzalez, 1989). Since then, the CREB gene was cloned from evolutionarily distant species as Aplysia, Drosophila and Hydra (Figure 7). DNA-binding is mediated by a basic domain, a lysine and arginine rich stretch of aminoacids. In addition, CREB binds to its target sequence as a dimer (Habener, 1990).

Dimerization occurs through an evolutionarily- conserved structural motif located at the C-terminus of the protein, the heptad repeat of leucine residues, named the leucine zipper, just carboxy-terminal to the basic domain. The presence of these two domains places CREB within the family of bZIP transcription factors. The mouse and human genes are composed of 11 exons that encode several isoforms, resulting mainly by the alternative splicing of the 5 first exons of the gene (Hoeffler, 1990; Ruppert, 1992). Functional studies of these isoforms have revealed that according to the domains they encode, they act as either activators or repressors of transcription (Foulkes, 1992).

9. The CREB pathway

The CREB transcriptional activity is regulated by several phosphorylation events, among which the phosphorylation of the Ser 133 residue that is located within the kinase inducible domain (KID) (Gonzalez, 1989). KID is loaded with multiple phosphorylation target sites

for different kinases, such as

modulatory) and ATF1 (activating transcription factor) are CREB-related bZIP factors (see CREM sequence in Figure 7) that bind to the very same motif and whose transcriptional activities also depend on serine phosphorylation induced by various extracellular signals.

Figure 8: Post-transcriptional regulation of the mouse CREB gene.

The mouse CREB gene undergoes several splicing events (Ruppert, 1992 37) that produce isoforms with distinct transactivation potentials. The position of the intron located between exons 10 and 11 is identical in the Hydra CREB gene.

At the biological level, the regulation brought by the CREB transcription factor is critical for a variety of cellular processes, including proliferation, differentiation and adaptive responses.

CREB family members are believed to be important for learning and memory (Silva, 1998) and contribute to neuronal adaptation to drugs of abuse (Blendy, 1998). CREB activity is also important for hormonal control of metabolic processes, including regulation of neoglucogenesis by glucagon and insulin. During development, the Drosophila CREB gene is involved in the establishment of the dorsal ventral polarity by regulating genes like Ultrabithorax (Rose, 1997; Eresh, 1997; Andrew, 1997). In mice CREB is required during development for the establishment of a normal repertoire of T-cell lineages (Barton, 1996) and the absence of CREB leads to dwarfism and cardiac myopathy in the adult (Fentzke, 1998 ; Struthers, 1991). In Xenopus, the inhibition of CREB at blastula and early gastrula stages leads to severe posterior defects of the embryos reflected by strong spina bifida, whereas the inhibition of CREB at the beginning of neurulation resulted in stunted embryos with microcephaly (Lutz, 1999).

10. The CREB pathway in Hydra

Since many years, it was known that treating Hydra with cAMP could affect regeneration (Wolpert, 1974). In addition, it was shown that cAMP treatment could mimic the affect of HA during nerve cell differentiation (Holstein, 1986;

Fenger, 1994). By testing the DNA-binding activity of Hv nuclear extracts (NE) prepared at various times during regeneration onto the CRE, a significant modulation of the DNA-binding activity was observed early during regeneration (Galliot, 1995). This band-shift analysis was also performed using a species that is deficient for foot regeneration, Hydra oligactis, and no modulation of DNA-binding activity was observed in the absence of regeneration (Figure 9).

Therefore these modulations were likely not linked to the wound healing response but rather specifically induced at the time regeneration is initiated. In addition, slight differences in the CRE-binding complexes were observed according to the species. These results suggested that the putative hydra CREB transcription factor was implicated in the regulation of regeneration in Hydra. Thus Gallliot et al. (1995) screened Hv and Cv cDNA libraries with a guessmer oligonucleotide corresponding to the conserved basic region and identified in these two species a single CREB gene where both the bZIP and the KID domains were highly conserved (Figure 7).

Figure 9 : Regeneration-specific modulation of CRE DNA-binding activity. a) Hydra vulgaris nuclear extracts (NE) prepared 4 and 28 hours after cutting. This species regenerate both, apex and basal disk. b) Hydra oligactisNE prepared 2, 4 and 24 hours after cutting. This species is slower for apical regeneration and is deficient for basal regeneration (less than 5% of the animals will regenerate a basal disk). C1, C2: uncut controls.