SHORT COMMUNICATION Experience-dependent increase in cAMP-responsive element binding protein in synaptic and nonsynaptic mitochondria of the rat hippocampus Lia R. M. Bevilaqua 1,2 , Martõ Ân Cammarota 1 , Gustavo Paratcha 1, Miguelina Levi de Stein 1 , Iva Ân Izquierdo 2 and JorgeH.Medina 1 1 Instituto de Biologia Celular y Neurociencias `Prof. Dr Eduardo de Robertis', Facultad de Medicina, Universidad de Buenos Aires (UBA), Paraguay 2155, Buenos Aires 1121, Argentina 2 Centro de Memo Âria, Departamento de Bioquõ Âmica, Instituto de Biocie Ãncias, Universidade Federal do Rio Grande do Sul (UFRGS), Porto Alegre 90035±003, Brasil Keywords: inhibitory avoidance; memory; synaptic plasticity; intracellular signalling Abstract Cyclic AMP-responsive element binding protein (CREB) plays a pivotal role in the formation of long-term memory in Drosophila, Aplysia, mice and rats. Recently, we were able to demonstrate that CREB and its serine 133 phosphorylated form p-CREB are localized in synaptic and nonsynaptic mitochondria of the rat brain. Here we report on the effect of a one-trial inhibitory avoidance training procedure on mitochondrial CREB from the rat hippocampus. This aversively motivated training task is associated with a time-dependent increase (34±35%) in both p-CREB and CREB immunoreactivities detected in synaptic mitochondria of the hippocampus. In nonsynaptic mitochondria, p-CREB levels increased in both trained and shocked animals. In addition to CREB, two CRE-element binding repressors, CREB-2 and CREM-1, were also detected in puri®ed brain mitochondria. No changes were observed in CREB-2 and CREM-1 immunoreactivities in hippocampal synaptic mitochondria after an inhibitory avoidance training. Taken together the present ®ndings represent the ®rst evidence showing that brain mitochondrial CREB may participate in plasticity- dependent changes associated with a behavioural training procedure. Introduction Frommolluskstomammals,memorycanbedividedintoatleasttwo temporally and mechanistically, albeit linked, distinct forms (McGaugh, 1966; Davis & Squire, 1984; Izquierdo et al., 1998): a protein and mRNA-synthesis-independent form that lasts minutes to hours (short-term memory) and a protein and mRNA-synthesis- dependent form that lasts several hours to days or longer (long-term memory, LTM). In the last few years a great body of evidence emerged demonstrating the importance of CREB in the molecular basis of LTM formation (see for references Silva et al., 1998). For instance, CREB-regulated transcription participates in long-term facilitation in Aplysia, a model of nonassociative learning (Dash et al., 1990). In Drosophila, the overexpression of a CREB repressor isoform blocks LTMforanodouravoidancelearning(Yin et al.,1994),whereasthe expression of an activating isoform of CREB facilitates LTM formation (Yin et al., 1995). In CREB knockout mice, LTM for spatial and aversively motivated learning tasks is impaired, whereas acquisition and short-term memory are normal (Bourtchuladze et al., 1994). In rats, we recently found that p-CREB immunolabelling in hippocampal neurons increased immediately after, and 3±6h after, a one-trial inhibitory avoidance training procedure (Bernabeu et al., 1997), a hippocampus-dependent learning process (Izquierdo & Medina,1997;Impey et al.,1998;Izquierdo et al.,1998).Inaddition, localmicroinjectionofCREBantisenseoligonucleotidesimpairsLTM fortwodifferentlearningtasksinrats(Guzowski&McGaugh,1997; Lamprecht et al.,1997).Furthermore,CRE-mediatedgeneexpression increased in the hippocampus after contextual conditioning or inhibitory avoidance tasks (Impey et al., 1998). Quite recently, we were able to demonstrate, by using immuno- blotting, immunoelectronmicroscopy and electrophoretic mobility shift assays, the presence of CREB and p-CREB in synaptic and nonsynaptic mitochondria of the rat brain (Cammarota et al., 1999). Therefore, in the present study we examine the effect of one-trial inhibitory avoidance training on hippocampal mitochondrial CREB. Materials and methods Training procedure Weused235maleWistarrats(age3months;weight:170±200g)from our own breeding stock. The animals were housed four to a 50 3 40 3 15cm steel cage with a grid-like stainless steel lid. The animalroomwasona12hlight:12hdarkcyclewithwaterandfood freely available. There were three experimental groups; group 1 animals withdrawn from their home cages and killed immediately (control group); group 2 and group 3 animals exposed to a 50 3 25 3 25cm acrylic box, with a frontal glass panel and a ¯oor madeofparallel1mmcalibrebronzebarsspaced0.8cmapart.A5-cm high,8cmwidewoodplatformwasplacedontheleftextremeofthe Correspondence: Dr Jorge H. Medina, as above. Email: jhmedina99@yahoo.com Received 10 May 1999, revised 7 July 1999, accepted 2 July 1999 European Journal of Neuroscience, Vol. 11, pp. 3753±3756, 1999 ã European Neuroscience Association