biostudies-arrayexpressAlexander GlahsDrosophila melanogasterhttps://www.ebi.ac.uk/biostudies/studies/E-MTAB-15767In this study, I identify novel regulatory nodes and potential signal transduction pathways that govern neurogenic tissue specification and differentiation. The Drosophila melanogaster central nervous system develops from three columns of neuroectodermal cells that give rise to neural stem cells called neuroblasts. I investigated the genomic regulatory state of cells derived from the ventral column (VC) and the intermediate column (IC) of the neurectoderm (NE) as they specify, diversify and differentiate. In this thesis I investigated how cell populations become distinct from one another, and how the genomic regulatory state within these cell populations changes over developmental time. As post-translational histone modification occupancy and chromatin accessibility has been linked to distinct aspects of gene regulation, I assessed their tissue specific genome wide distributions and used them as a proxy to investigate the genomic regulatory state in a developing embryo. Differential peak calling, differential enrichment analysis and machine learning was employed to identify spatiotemporally regulated cis- regulatory modules (CRMs) and differentially regulated genes (DRGs). I could show that I can accurately predict tissue specific CRMs by validating a subset of de novo predicted differentially regulated CRMs with reporter lines. The spatiotemporal regulatory trajectory of differentially regulated elements was used to refine the existing models for the gene regulatory networks that govern neuronal differentiation, to associate distal regulatory elements with core promoters, and to identify enriched sequence motifs within active CRMs. Taken together the datasets and algorithms in this study provide a comprehensive atlas of novel CRMs of the developing nervous system in Drosophila, novel insights into the regulatory nodes and edges during neurogenesis and a valuable resource to refine existing models of neurogenesis specifically and tissue differentiation generally.biostudies-arrayexpressSequencing - wo biological replicates of every mark were pair- end sequenced with 100-bp reads using an Illumina High-seq 2500 platform.Sample Collection - This section describes the batch isolation of tissue-specific chromatin for immunoprecipitation (BiTS-ChIP) for analysis of histone modifications as described by Stefan Bonn, Robert P Zinzen et al. 2012. This method utilizes dissociated formaldehyde cross-linked embryos expressing a cell type–specific nuclear marker. “Fixed nuclei are isolated and sorted using FACS on the basis of the cell type–specific nuclear marker. Tissue-specific chromatin is extracted, sheared by sonication and used for ChIP-seq or other analyses.” Covalent cross-linking before embryo dissociation preserves the chromatin context and increases stability during FACS of nuclei. This allows sorting high purity material that can be used for ChIP seq. The collection, isolation, sorting, ChIP and library prep protocols were optimized at several positions to isolate rare cell populations with high purity and low input material. The optimized protocol described below allows to generate sequencing libraries within 4 days. To investigate the histone modifications (H3K4me3, H3K4me1, H3K79me3, H3K27ac and H3K27me3) in columnar cells of the developing Drosophila nervous system, chromatin immunoprecipitation (ChIP) followed by high-throughput DNA sequencing (ChIP-seq) experiments were performed. The experiments were conducted on an Illumina HiSeq device in collaboration with Madlen Sohn and Mirjam Feldkamp, from the Next Generation Sequencing core facility of the Max Delbrück Center for Molecular Medicine in Berlin, Germany. The dissociated nuclei were stained with antibody to the GFP and RFP tags (1:100) for 1 h, washed with 10 ml of PBT for 10 min, resuspended in 3 ml of PBT and incubated for 1 h in the dark with Alexa488-conjugated donkey anti-mouse secondary antibody (1:100; A-21202, Invitrogen). Nuclei were centrifuged for 1 min at 1,000g and were resuspended in 6 ml of PBTB (5% BSA in PBT). The suspension was then divided into 1ml aliquots in 5-ml tubes (352063, Falcon) and brought to 3 ml by adding PBTB. Proteinase inhibitors was added to PBTB shortly before use (Roche Complete, 1 tablet per 50 ml) and the solution was filtered through a 0.22-μm-pore filter to remove undissolved albumin particles. Immediately before sorting, the nuclei were dissociated by passing the samples through a 22-gauge needle 7 times and then filtering them through a cell strainer (322350, Falcon). Nuclear samples were run on a BD FACS Aria II Flow Cytometry Cell Sorter from Becton Dickinson using the D FACSDiva software. A 488-nm solid state, 24-mW laser, was used as primary laser for excitation of the Alexa-488 fluorophore. A 405-nm solid state diode, 25-mW laser was used for excitation of the DAPI fluorophore that was used to stain chromatin in nuclei. A 70μm nozzle was used for sorting. An accurate drop- delay value was determined by BD Accudrop and the drop delay is calculated automatically by the system and adjusts to maintain a constant droplet breakoff. The differential pressure during the sorts was kept low to limit illumination variations and prevent clogging. Sample acquisition rates was adjusted to be 10.000 -15.000 events per second. Temperature on both sample and collection tubes was kept at 4°C during sorting. The sort decision gate was based on a combination of scatter, pulse width and fluorescence parameters. A detailed gating procedure used for sorting is described in the Results section. Sorting purity was assessed by repeated acquisition of a small aliquot of sorted material in the BD FACS Aria II Flow Cytometry Cell Sorter and assayed under an epifluorescent microscope. Samples with >10% DAPI-positive, Alexa488-negative events were discarded. Post-acquisition analysis was performed using FlowJo for Windows. I optimized the ChIP protocol for 1.000.000 input cells but were successful to ChIP from only 10.000 sorted cells. Immediately after sorting, 10^6 nuclei were centrifuged at 3,500g for 5 min at 4°C, and the pellet was resuspended in 75 μl of complete Lysis Buffer tL1 (Lysis Buffer tL1 + Protease Inhibitor Cocktail - PIC) (Diagenode Cat. No. C01010130) after the supernatant was transferred to a low-binding tube (710176, Biozym Scientific). Samples were incubated on ice for 5 minutes. 225 μl of complete HBSS (HBSS + PIC) was added to the samples and mixed by manual agitation. he chromatin was sheared into about 200-bp fragments using a Diagenode BioRuptor. Samples were sonicated for 8 cycles of 30 seconds interrupted by 30 seconds of cooling in the integrated water bath at 4°C. The chromatin was then centrifuged for 2 min at 14,000g, and the supernatant was transferred to a low-binding tube (710176, Biozym Scientific). Chromatin supernatant was either snap frozen in liquid nitrogen or directly used for immunoprecipitation. A small aliquot was used to measure DNA concentration and fragment size. This sub-sample (5%) was reverse crosslinked. For reverse-crosslinking RNaseA was added to a final concentration of 50μg/ml. The sample was incubated for 30min at 37°C. The sample was then adjusted to 0.5%SDS.ProteinaseK was added to 0.5mg/ml and the sample was incubated for at 1h at 37°C. Cross-linking was reversed over night at 65°C. The DNA was then extracted using UltraPureTM Phenol:Chloroform:Isoamyl Alcohol (25:24:1, v/v) and Phase Lock GelTMtubes. DNA was precipitated over night at -20°C and quantified using the Qubit® dsDNA HS Assay Kit (Life Technologies) according to manufacturer’s instructions. To the remaining chromatin sample 300 μl of complete ChIP Buffer tC1 (ChIP Buffer tC1 + PIC) was added and mixed by inverting the sample. 2 μg of antibody against PTMs was added to the samples. Antibodies detecting H3 (ab1791), H3K4me3 (ab71998), H3K4me1 (ab8895), H3K27ac (ab4729), H3K36me3 (ab9050) and H3K79me3 (ab2621) were purchased from Abcam and antibody to H3K27me3 was obtained from Diagenode (C15410069). Samples were rotated at 40 rpm for 14 hours at 4°C. Magnetic immunoprecipitation and washes were performed according to manufacturer’s instructions provided by the Diagenode MicroChIP kit. In short, 10 μl of washed Protein G coated magnetic beads were added to each IP samples. Samples were rotated at 40 rpm for 2 hours at 4°C and washed four times for 4 minutes on a rotator at 4°C. Supernatants were removed by using a magnetic rack that immobilized the magnetic Protein-G beads. The IP conditions for the individual antibodies were optimized for recovery and enrichment using chromatin from 10^6 nuclei to yield enough material for subsequent library generation (>2ng). After washing the beads 200 μl of Elution Buffer tE1 was added to the tubes. The tubes were rotated at 40 rpm for 30 minutes at room temperature. 20 μl of previously set aside input is purified in parallel. The beads were removed with the magnetic rack and the supernatant was transferred into new low- binding tubes.8 μl of Elution Buffer tE2 was added and the samples were incubated at 65°C overnight.Nucleic Acid Extraction - The DNA was then extracted using UltraPureTM Phenol:Chloroform:Isoamyl Alcohol (25:24:1, v/v) and Phase Lock GelTMtubes. DNA was precipitated over night at -20°C and quantified using the Qubit® dsDNAHSAssay Kit (Life Technologies) according to manufacturer’s instructions. 235 The quality of each IP was assessed by real-time PCR (see Figure 17). Real-time PCR primer combinations described in Bonn, Zinzen, Girardot, et al., 2012; Bonn, Zinzen, Perez-Gonzalez, et al., 2012 were used. The combinations used were: H3-ChIP – osk/twi-promoter; H3K4me1-ChIP – Rpl32-5’/osk; input, H3K27ac-ChIP, H3K79me3- ChIP and H3K4me3-ChIP – Rpl32-promoter/Rpl32-5’. For H3K27me3-ChIP I used a primer set targeting Mef2 which is a critical regulator in heart development and cardiac gene expression that is repressed in neuronal cells and compared it to the signal at the constitutively active Rpl32 promoter.Growth Protocol - Embryos were collected from ten large cylindrical population cages (29 cm × 48 cm), each seeded with 30 g of flies. Embryos were collected for 6 days after seeding. After performing at least three 1-h pre-lays, 2-h embryo collections (using 15-cm molasses plates streaked with yeast paste) typically yield 0,5-3 g dry weight of tightly staged embryos, which are aged as for another 4, 6, or 8 hours for the 4-6h, 6-8h, 8-10h timepoint respectively. Collections of IC-eGFP and VC-RFP embryos at the 4-6h, 6–8 h and 8-10h stage of development were collected, dechorionated, and fixed as described in Sandmann, T.et al 2006 (Sandmann, Jakobsen and Furlong, 2007).Library Construction - Libraries were prepared with the NEB NEBNext® Ultra™ II DNA Library Prep Kit for Illumina® according to manufacturer’s instructions. Size selection was performed via agarose gel electrophoresis in a 2% high purity DNA-free agarose gel or with AMPure XP beads according to manufacturer’s instructions. Library quality was assessed on a 2100 Bioanalyzer system (Agilent).Sample Treatment - In brief, embryos were dechorionated for 2 Minutes in PBS with 0.1%TritonX-100 and 6%(v/v) Na-hypochlorite(bleach). Embryos were rinsed under running deionized water for at least 2 Minutes. Rinsed embryos were transferred into 40ml fixation and cross- linking solution (aqueous phase: 50mM HEPES (pH8.0), 1mM EDTA, 0.5mM EGTA, 100mM NaCl, 1%(v/v) formaldehyde; organic phase: 100%(v/v) n-heptane) and shaken vigorously for 15 Minutes. To quench the fixation, the embryos were vigorously shaken for 1 Minutes in PBS with 0.1%TritonX-100 and 125mM glycine. Afterwards embryos were washed 3 times in 40 ml ice cold PBS and then drained on sieves. The embryos were weighed, shock frozen in liquid nitrogen and stored at -80°C. 50 μl of embryos was prepared to survey and validate an appropriate stage distribution of the collection. To achieve this the vitelline membrane was removed by shaking in methanol:heptane(1:1) for 90 seconds. The embryos were washed at twice with absolute methanol and store at-20°C. The aliquot of embryos was later rehydrated and mounted for microspcopic assessment of the stage distribution. Fixed embryos can be stored at −80 °C or used immediately.OrganizationMINSEQE ScoreAssays and DataProcessed DataMAGE-TAB FilesData Transformation - The Illumina bcl2fastq Conversion Software was used to demultiplex sequencing data and converts base call (BCL) files into FASTQ files. The following parameters were used “--barcode-mismatches 1 -r 1 -d 1 -p 1 -w 1”. To examining the sequencing errors and trimming low quality reads, the quality assessment of the raw data was performed by using FastQC. Quality statistics like per base sequence quality and duplication levels were assessed (Andrews, S. (2010) “FastQC: A Quality Control Tool for High Throughput Sequence Data”, Available at: http://www.bioinformatics.babraham.ac.uk/projects/fastqc/). The pre-processing workflow was implemented in Galaxy (Giardine et al., 2005; Afgan et al., 2018) which was set up and made available to me by the MDC bioinformatics core facility. Reads that passed the quality control were aligned to the reference Drosophila melanogaster genome (version dm6) using Bowtie2 (version 0.12.8) with the following parameters: “-1 -2 -fr -X 600 -no-mixed -N 1” (Langmead and Salzberg, 2012). The output options were chosen to obtain sorted bam files from, Bowtie2 global alignment. The Dedup tool from SAM Tools was used to deduplicate the bam files (Li et al., 2009). 236 Biological replicates were compared to evaluate their reproducibility. The data from every experiment was corrected for library size as previously described by Bonn, Zinzen, Girardot et al. 2012 (Bonn, Zinzen, Girardot, et al., 2012). In brief, the genome size of dm6 was used as the reference to apply the correction, which has the advantage of direct genomic coverage readout. Bonn and Zinzen et al. 2012 defined RPGC normalization as follows: “The coverage enrichment E of a library is E = R * l / M, where M is the genome size, R the number of reads in the library and l the length of the mapped fragments. Thus, E represents the expected coverage of the library if the reads were uniformly distributed. For a given base pair in the genome overlapped by r reads, its corrected score s is: s= r/E. We call this corrected value “Read Per Genomic Coverage” (RPGC).” The median value of the RPGC was calculated for 20bp windows across the genome. The obtained values were compared between biological replicates using Pearson correlation (Results section 8.5). The background correction for chromatin modifications was performed using stage and tissue matched H3 ChIP-seq as the background model. The RPGC normalized H3 signal was subtracted from the RPGC normalized histone modification signals for visualization in IGV. In cases where only one track per condition is shown the average of the RPGC and background subtracted tracks is shown. bamCoverage from the Deeptools2 package was used to generate bigwig files for visualization in IGV. The following parameters were used: “-e –ignoreDuplicates -bs 20”UnknownTranscriptomicsGenomicsProteomicsIllumina HiSeq 2500ChIP-seqDrosophila melanogasterAlexander GlahsfalseAtlas of the chromatin landscape in the developing nervous system of Drosophila melanogaster - ChIPseq and ATACseq - time series of tissue specific neuroblast histone PTMsIn this study, I identify novel regulatory nodes and potential signal transduction pathways that govern neurogenic tissue specification and differentiation. The Drosophila melanogaster central nervous system develops from three columns of neuroectodermal cells that give rise to neural stem cells called neuroblasts. I investigated the genomic regulatory state of cells derived from the ventral column (VC) and the intermediate column (IC) of the neurectoderm (NE) as they specify, diversify and differentiate. In this thesis I investigated how cell populations become distinct from one another, and how the genomic regulatory state within these cell populations changes over developmental time. As post-translational histone modification occupancy and chromatin accessibility has been linked to distinct aspects of gene regulation, I assessed their tissue specific genome wide distributions and used them as a proxy to investigate the genomic regulatory state in a developing embryo. Differential peak calling, differential enrichment analysis and machine learning was employed to identify spatiotemporally regulated cis- regulatory modules (CRMs) and differentially regulated genes (DRGs). I could show that I can accurately predict tissue specific CRMs by validating a subset of de novo predicted differentially regulated CRMs with reporter lines. The spatiotemporal regulatory trajectory of differentially regulated elements was used to refine the existing models for the gene regulatory networks that govern neuronal differentiation, to associate distal regulatory elements with core promoters, and to identify enriched sequence motifs within active CRMs. Taken together the datasets and algorithms in this study provide a comprehensive atlas of novel CRMs of the developing nervous system in Drosophila, novel insights into the regulatory nodes and edges during neurogenesis and a valuable resource to refine existing models of neurogenesis specifically and tissue differentiation generally.2025-10-30T00:00:00Z2026-05-27T12:54:30.884Z2025-11-14T09:18:54.694ZE-MTAB-15767ERP182426EFO_0002944EFO_0004170EFO_0003789EFO_0002692EFO_0005518EFO_0003816EFO_0004184EFO_0003969