Nuclear RNA-seq in Drosophila mushroom body to determine Kdm4B target genes

Purpose: In Drosophila, we tried to determine the neural activity-dependent transcription controlled by Kdm4B. We conducted nuclear RNA-seq analysis in mushroom body in Drosophila, either in naive state or after optogenetic activation. Methods: The nuclei of mushroom body were collected by immunoprecipitation-based method (Hirano Y., et al, 2016, Nature Communications, also see extraction protocol below). mRNA was purified from the mushroom body nuclei with oligo dT magnetic beads, and used to prepare library. After deep sequencing in triplicate using illumine Hiseq X, the adaptors were trimmed via Trimommatic, followed by mapping to the Drosophila reference genome, dm6 from UCSC using STAR. The reads with low mapping quality below 8 and the non-primary mapped reads were eliminated. The filtered reads were analyzed with HTSeq-count to obtain numbers of the reads mapped on exons, which was further analyzed on R using DESeq2 Results: Using an optimized data analysis workflow, we mapped about 9-18 million reads for ChIP-seq, and 30-40 million reads for nuclear RNA-seq to Drosophila reference genome, dm6 In ChIP-seq, binding of all proteins was enriched nearby the transcriptional start site (TSS). The binding sites determined were 1,237 for CoRest-C, 2,717 for Rpd3, and 6,684 for CBP. Importantly, among CoRest-C binding sites, overlapping sites of CoRest-C and CBP were 1,038/1,237 (83.9%), those of CoRest-C and Rpd3 were 789/1,237 (63.8%), and those of CoRest-C, Rpd3, and CBP were 704/1,237 (56.9%), suggesting that these three proteins colocalize on the specific genomic regions In nuclear RNA-seq, we found that upregulation of gene expression is robust at 10-min after 5-min optogenetic activation. The genes showing significant difference were 2,702 at this time point, in which 1,582 genes showed increase, and the rest showed decrease in the expression. Among the 1,582 increased genes, 1,055 genes were bound by CBP, supporting the idea that CBP is an important factor in activity-dependent transcription. Furthermore, 338 was bound by CoRest-C, 254 of which showed colocalization with CBP and Rpd3. Conclusions: Our study represents activity-dependent transcription in mushroom body in Drosophila, which is mediated by Kdm4B. Overall design: nuclear RNA-seq in mushroom body nuclei from WT Drosophila

研究目的：在果蝇（Drosophila）中，我们旨在探究由Kdm4B调控的依赖神经活动的转录过程。我们分别在未刺激状态，以及光遗传激活（optogenetic activation）后的果蝇蕈形体（mushroom body）中开展了核RNA测序（nuclear RNA-seq）分析。
研究方法：我们采用基于免疫沉淀（immunoprecipitation）的方法收集果蝇蕈形体的细胞核（该方法源自Hirano Y.等人2016年发表于《Nature Communications》的研究，具体提取流程见下文）。随后从蕈形体细胞核中利用寡聚dT磁珠纯化mRNA，并用于构建测序文库。使用Illumina Hiseq X进行三次重复深度测序后，通过Trimmomatic去除接头序列，再利用STAR软件将测序reads比对至UCSC数据库发布的果蝇参考基因组dm6。过滤掉比对质量值低于8的reads以及非唯一比对reads后，使用HTSeq-count统计比对至外显子区域的reads数目，随后在R环境中通过DESeq2进行后续分析。
研究结果：通过优化后的数据分析流程，我们分别将约900万至1800万的ChIP测序（ChIP-seq）reads、3000万至4000万的核RNA测序reads比对至果蝇参考基因组dm6。在ChIP-seq结果中，所有蛋白的结合位点均富集于转录起始位点（transcription start site, TSS）附近。本研究鉴定得到的结合位点数目分别为：CoRest-C 1237个、Rpd3 2717个、CBP 6684个。值得注意的是，在CoRest-C的结合位点中，与CBP存在重叠的位点占比83.9%（1038/1237），与Rpd3存在重叠的位点占比63.8%（789/1237），同时与CoRest-C、Rpd3及CBP三者均存在重叠的位点占比56.9%（704/1237），提示这三种蛋白可在特定基因组区域共定位。在核RNA测序分析中，我们发现经过5分钟光遗传激活后10分钟，基因表达上调现象尤为显著。该时间点下存在显著表达差异的基因共2702个，其中1582个基因表达上调，其余基因表达下调。在这1582个上调基因中，有1055个存在CBP结合位点，这支持了CBP是依赖神经活动的转录过程中的关键调控因子这一观点。此外，有338个上调基因存在CoRest-C结合位点，其中254个同时与CBP和Rpd3存在共定位。
研究结论：本研究揭示了果蝇蕈形体中由Kdm4B介导的依赖神经活动的转录调控过程。
整体实验设计：对野生型（wild type, WT）果蝇的蕈形体细胞核进行核RNA测序。