ABSTRACT: Purpose: Exploring the alteration in histone acetylation colocalized with CREB/CRTC binding, which are important for long-term memory maintenance Methods: The nuclei of Drosophila mushroom bodies were purified and the histone acetylation and CREB/CRTC binding were examined by ChIP-seq using specific antibodies. The sequencing was performed using SOLiD, and aligned to dm3 using Lifescope..The sequence reads that passed quality filters were analyzed by peak calling using PICS and ERD equipped on Strand NGS software. This method using two algorisms excludes false-positive calling of peaks. The PICS and ERD were run using a default setting except for the following parameters; for PICS, 120 bp as an average fragment length, 10 bp as a minimum distance between forward and reverse reads, 200 bp as a minimum distance between forward and reverse reads, 100 bp as a window width, with 5% false discovery rate; for ERD, 2 as an enrichment factor, 100 bp as a window size, 10 bp as a window slide size, 100 bp as a minimum region size. The peaks obtained by ERD analysis were filtered by an enrichment factor of 1.5, and a density of reads at 0.3 for H3K9Ac, 0.15 for H4K16Ac, 0.18 for CREB and 0.15 for CRTC. To determine the histone acetylation-increased sites, the peaks determined from 3 replicates of the trained samples were summed to analyze all peaks. Then the centers of the H3K9Ac and H4K16Ac peaks were extracted, and then used to create neighboring 400 bp windows. The mapped reads were counted in these 400 bp windows for each biological replicate of ChIP-seq analysis. The counted read numbers in individual regions were normalized by elav for H3K9Ac ChIP-seq, and gapdh2 for H4K16Ac ChIP. The normalized read numbers in individual windows obtained from the spaced trained flies and the naïve flies were analyzed by Student’s t test (p < 0.05). Standard deviations of the normalized read numbers in individual windows from 3 replicates of naïve flies were 0.07 in H3K9Ac-ChIP-seq and 0.10 in H4K16Ac-ChIP-seq. Significantly increased peaks showing a more than 1.2-fold increase (>2SD) were determined as sites with an increase in H3K9Ac or H4K16Ac. Results: Using an optimized data analysis workflow, the filtered reads amounted to 12-15 million reads for H3K9Ac in each of 3 replicates, 5-7 million reads for H4K16 in each of 3 replicates, 2.4 million reads for CREB, 2 million reads for CRTC, and 2.9 million reads for input. Using anti-H3K9Ac antibody, we found significant increases in H3K9Ac 1 day after training, and 75.0% of these mapped within 500bps of the transcriptional start sites (TSSs) of 1766 genes. Using anti-H4K16Ac antibody, we found significant increases in H4K16Ac 1 day after spaced training, and 61.6% of these mapped within 500bps of the TSSs of 1320 genes. Epigenetic changes in the vicinity of TSSs suggests that expression of these genes may be increased in LTM maintenance. We also examined CRTC and CREB binding 1 day after training. We identified 2390 CRTC binding sites, of which 79.6% of these were close to CREB binding sites. These CREB/CRTC binding sites mapped to 1394 genes, and were predominately located within 500bp from TSSs (55.4%). Of the 1394 CREB/CRTC target genes, 346 genes showed increases in H3K9Ac, 319 showed increases in H4K16Ac, and 135 showed increases in both. Conclusions: Our study represents the first detailed analysis of chromatin state related to memory in Drosophila mushroom bodied. The optimized data analysis workflows reported here should provide a framework for comparative investigations of tissue-specific epigenetic alterations.
