Experiments were conducted with adult mice. Eight mice were not injected and served as a baseline

resent in all lobes of the MB and widespread throughout the brain. Notably, the expression NP1617 and NP3558 pattern overlapped in the MB. The lack of NP1617-driven expression in the ‘/’ neurons suggests that the enhancer elements for the / and neurons may be distinct from those that drive expression in the ‘/’ neurons. In a separate approach to examine endogenous HDAC4 expression, a fly strain harbouring an HDAC4-trapped EYFP generated by the Cambridge Protein Trap Project was obtained. This consists of an artificial exon containing the EYFP gene flanked by splice acceptor and donor sequences, which is inserted into the 2nd intron of the HDAC4 gene, 1310 bp downstream from the 3′ end of the 2nd exon. This results in an internal incorporation of EYFP into the HDAC4 protein. The normal viability of the flies suggests that insertion of EYFP does not appear to inactivate or severely reduce endogenous HDAC4, in contrast to some hypomorphs of HDAC4 that are hemizygous lethal due to a role in embryo development. EYFP expression was extremely low and we could not detect it by confocal microscopy on whole mount brains without amplifying the signal via immunohistochemistry with an anti-GFP antibody. Although the level of expression in whole-mount brains was low and difficult to distinguish from background, we did observe a weak signal in the MB that was never observed in control brains, which is in agreement with the MB expression driven by the two enhancer trap lines. Characterisation of HDAC4 overexpression in the mushroom body As long-term courtship memory is dependent on an intact MB, we chose to examine importance of HDAC4 in LTM by modulating HDAC4 levels in the MB and assessing the effect on LTM using the courtship suppression assay. We generated transgenic flies containing an N-terminus FLAG-tagged HDAC4 gene driven by a UASG-hsp70min enhancer-promoter. Expression was induced in the adult fly brain using the GAL4 driver OK107. We selected this driver for its ability to facilitate high expression in Kenyon cells, the intrinsic neurons of the MB. Drosophila HDAC4 plays a role in embryo segmentation, therefore in order to obviate potential effects of HDAC4 overexpression on development, we restricted HDAC4 expression to adult brains with the temporal and regional gene expression targeting system. In 2569287 this system, flies are raised at 19C, at which GAL4-mediated gene expression is inhibited by a temperature sensitive mutant of GAL80. When the temperature is raised to 30C, GAL80ts is inactivated and transgene expression ensues. To aid visualisation of the general anatomy of the brain, wholemount brains were subjected to 7473193 immunohistochemistry with an antibody to the Class I HDAC Rpd3, which is expressed in all neuronal nuclei, and co-labelled with anti-FLAG to detect 3 HDAC4 Regulates Long-Term Memory doi: 10.1371/journal.pone.0083903.g002 4 HDAC4 Regulates Long-Term Memory HDAC4. In contrast to the nuclear Rpd3, the majority of HDAC4 was 870281-82-6 localized to the MB lobes. HDAC4 was not detected in the brains of flies maintained at 19C. HDAC4 was also localized to the calyx, the dendritic field of the Kenyon cells, and in a subset of Kenyon cell nuclei. Within the nuclei, HDAC4 was localized to discrete regions, appearing as punctate nuclear bodies, which have been previously observed when HDAC4 is expressed in cultured cells. Overexpression of HDAC4 impairs LTM The repeat training courtship assay was used to assess 24 hour LTM. This assay measures the a