We have reported previously that cAMP-response-element-binding protein (CRE
B) was phosphorylated in a cell-cycle-dependent manner, showing that it was
phosphorylated at early S-phase at casein kinase II target sites. To asses
s the possible involvement of CREB in cell cycle progression, CREB expressi
on vector was transiently transfected into various cells. Unexpectedly we f
ound that transfection with CREB expression vector resulted in an abundance
of dead cells. Morphological examination revealed that these cells had und
ergone apoptosis. The coincidence of CREB overexpression and apoptosis indu
ction at the individual cell level was confirmed by a immunohistochemical s
tudy. To confirm that overexpression of CREB was the cause of apoptosis, a
dominant-negative mutant of CREB, KCREB, was coexpressed with the wild type
. The co-existence of KCREB effectively rescued CREB-mediated apoptosis in
a dose-dependent manner, verifying that apoptosis was truly a specific effe
ct of overexpressed CREB and not an artifact of the transfection procedure.
Deletion analysis indicates that neither the Q1 transactivation domain, wh
ich functions in transcription, nor the kinase-inducible domain, in which a
cluster of various kinase targets exists, is necessary; however, the Q2 tr
ansactivation domain is required for the induction of apoptosis. A more pre
cise study indicates that the four-residue stretch Glu-Glu-Ala-Ala at the m
ost C-terminal region of the Q2 domain is especially important for the indu
ction of apoptosis. Thus overexpressed CREB induces apoptosis by transmitti
ng certain signals from the C-terminal portion of the Q2 domain. Possible r
oles of cell-cycle-regulated phosphorylation and also an elevation of the i
ntracellular cAMP level in CREB-induced apoptosis are suggested.