Metabolomics,Unknown,Transcriptomics,Genomics,Proteomics

Dataset Information

Unraveling determinants of transcription factor binding outside the core binding site

ABSTRACT: Binding of transcription factors (TFs) to regulatory sequences is a pivotal step in the control of gene expression. Despite many advances in the characterization of sequence motifs recognized by TFs, our ability to quantitatively predict TF binding to different regulatory sequences is still limited. Here, we present a novel experimental assay termed BunDLE-seq that provides quantitative measurements of TF binding to thousands of fully designed sequences of 200 bp in length within a single experiment. Applying this binding assay to two yeast TFs we demonstrate that sequences outside the core TF binding site profoundly affect TF binding. We show that TF-specific models based on the sequence or DNA shape of the regions flanking the core binding site are highly predictive of the measured differential TF binding. We further characterize the dependence of TF binding, accounting for measurements of single and co-occurring binding events, on the number and location of binding sites and on the TF concentration. Finally, by coupling our in vitro TF binding measurements, and another application of our method probing nucleosome formation, to in vivo expression measurements carried out with the same template sequences now serving and promoters, we offer insights into mechanisms that may determine the different expression outcomes observed. Our assay thus paves the way to a more comprehensive understanding of TF binding to regulatory sequences, and allows the characterization of TF binding determinants within and outside of core binding sites. Binding experiments were carried out with one of two input libraries; Lib1 (also referred to as “main lib”) includes designed sequence variants of length 150bp, and Lib2 (also referred to as “constant flanks lib”) includes designed sequence variants of length 200bp. The sequences of the variants of interest can be found in the second column of the processed data file (BunDLE-seq_data, see 3 excel sheets). EXP1, EXP2 and EXP3 were carried out using Lib1, and EXP4 was carried out using Lib2. Binding experiments were carried out with different proteins (in different concentrations) – DNA was extracted from bound bands that were formed on the gel and amplified with a primer with a unique 5bp upstream tail specifying the identity of the originating band (also referred to as “band barcode”). Data from the following bands was used in the analysis: EXP1_band1, no-protein band – with band barcode “AACGA”, EXP1_band2, single Gcn4 (conc 1:8) bound band* – with band barcode “TCGTA” , EXP1_band3, histones bound band – with band barcode “AGATA”, EXP2_band1, no-protein band – with band barcode “AACGA”, EXP2_band2, single Gcn4 (conc 1:8) bound band* – with band barcode “AGATA”, EXP2_band2, single Gcn4 (conc 1:7) bound band – with band barcode “CACTT”, EXP2_band2, single Gcn4 (conc 1:6) bound band – with band barcode “TCGTA”, EXP2_band2, single Gcn4 (conc 1:5) bound band – with band barcode “CATCA”, EXP2_band2, single Gcn4 (conc 1:4) bound band – with band barcode “TATTA”, EXP2_band2, single Gcn4 (conc 1:3) bound band – with band barcode “AGTCA”, EXP2_band2, single Gcn4 (conc 1:2) bound band – with band barcode “ACGAA”, EXP2_band2, single Gcn4 (conc 1:1) bound band – with band barcode “ACAGA”, EXP2_band2, two Gcn4 (conc 1:1) bound band – barcoded with “GTAGA”, EXP3_band1, no-protein band – with band barcode “GCATA”, EXP3_band2, single Gal4 (conc 1:1) bound band – with band barcode “GATGT”, EXP3_band3, single Gal4 (conc 3:1) bound band – with band barcode “GTTCA”, EXP3_band4, two Gal4 (conc 3:1) bound band – with band barcode “CTCAA”, EXP4_band1, no-protein band – with band barcode “GAGTA”, EXP4_band2, single Gcn4 (conc 1:3) bound band – “TACCA”, EXP4_band3, no-protein band – with band barcode “TCGTT”, EXP4_band3, single Gal4 (conc 3:1) bound band – with band barcode “ACATC”, *biological replicates

ORGANISM(S): synthetic construct

SUBMITTER: Michal Levo

PROVIDER: E-GEOD-66143 | biostudies-arrayexpress |

REPOSITORIES: biostudies-arrayexpress

ACCESS DATA

Similar Datasets

Project description:Transcription factors direct gene expression, and so there is much interest in mapping their genome-wide binding locations. M-BM- Current methods do not allow for the multiplexed analysis of TF binding, and this limits their throughput. We describe a novel method for determining the genomic target genes of multiple transcription factors simultaneously. DNA-binding proteins are endowed with the ability to direct transposon insertions into the genome near to where they bind. The transposon becomes a M-bM-^@M-^\Calling CardM-bM-^@M-^] marking the visit of the DNA-binding protein to that location. A unique sequence M-bM-^@M-^\barcodeM-bM-^@M-^] in the transposon matches it to the DNA-binding protein that directed its insertion. The sequences of the DNA flanking the transposon (which reveal where in the genome the transposon landed) and the barcode within the transposon (which identifies the TF that put it there) are determined by massively-parallel DNA sequencing. To demonstrate the methodM-bM-^@M-^Ys feasibility, we determined the genomic targets of eight transcription factors in a single experiment. The Calling Card method promises to significantly reduce the cost and labor needed to determine the genomic targets of many transcription factors in different environmental conditions and genetic backgrounds. These data contain Ty5 insertion sites mapped by an Illumina GAII analyzer in the S. cerevisiae genome for the background strain without any Sir4 present (1 run), in strains expressing Sir4-tagged copies of three well-characterized TFs: Gal4, Leu3, and Gcn4 (1 run each), and a multiplex of eight Sir4-tagged TFs pooled in a single experiment (2 biological replicates), and insertions from the Thi2-Sir4 fusion expressed from its native locus in two conditions (1 run each). The format of each insertions file is [chromosome number] [position of genomic base] [direction of insertion] [number of reads at that position]. Raw sequencing data comes in two varieties. Paired-end data contains a 5 bp barcode at the beginning of read #2. Single-end data contains a 2 bp barcode on the beggining of read #1.

Project description:Transcription factors (TFs) bind to thousands of DNA sequences in mammalian genomes, but most of these binding events appear to have no direct effect on gene expression. It is unclear why only a subset of transcription factor bound sites are actively involved in transcriptional regulation, nor are the key genomic features known that discriminate between active and inactive TF binding events. Recent studies have identified promoter-distal RNA polymerase II (RNAP2) binding at enhancer elements, suggesting that these interactions may serve as a potential marker for active regulatory sequences. Despite these correlative analyses, a thorough functional validation of these genomic co-occupancies is still lacking. To characterize systematically the gene regulatory activity of DNA sequences underlying promoter-distal TF binding events that co-occur with RNAP2 and TF sites devoid of RNAP2 occupancy using a functional assay, we performed cis-regulatory element sequencing (CRE-seq) reporter assays. We tested more than 1,000 promoter-distal CCAAT/enhancer-binding protein beta (CEBPB) bound sites in two human cell types, HepG2 and K562. We found that CEBPB bound sites that co-occur with RNAP2 were more likely to exhibit enhancer activity in our reporter assay and these active sequences display significantly stronger RNAP2 read enrichment and local enhancer RNA (eRNA) activity compared to inactive sites. CEBPB-bound sites further maintained substantial cell type specificity, indicating that DNA sequence information at such sites can accurately convey cell type-specific regulatory information, supporting a strong role for local DNA sequence in distinguishing active from inactive TF-bound loci. By comparing our CRE-seq results to a comprehensive set of over 1,000 genome annotations, we identified a variety of genomic features that are strong predictors of regulatory element activity as well as cell type-specific activity. Collectively, our functional assay results indicate that RNAP2 occupancy can be used as a key genomic marker that can discriminate active versus inactive transcription factor bound sites.

Project description:To evaluate the effect of CG methylation on DNA binding of sequence-specific B-ZIP transcription factors (TFs) in a high-throughput manner, we enzymatically methylated the cytosine in the CG dinucleotide on protein binding microarrays. Two Agilent DNA array designs were used. One contained 40,000 features using de Bruijn sequences where each 8-mer occurs 32 times in various positions in the DNA sequence. The second contained 180,000 features with each CG containing 8-mer present three times. The first design was better for identification of binding motifs, while the second was better for quantification. Using this novel technology, we show that CG methylation enhanced binding for CEBPA and CEBPB and inhibited binding for CREB, ATF4, JUN, JUND, CEBPD and CEBPG. The CEBPB|ATF4 heterodimer bound a novel motif CGAT|GCAA 10-fold better when methylated. EMSA confirmed these results. CEBPB ChIP-seq data using primary female mouse dermal fibroblasts with 50X methylome coverage for each strand indicate that the methylated sequences well-bound on the arrays are also bound in vivo. CEBPB bound 39% of the methylated canonical 10-mers ATTGC|GCAAT in the mouse genome. After ATF4 protein induction by thapsigargin which results in ER stress, CEBPB binds methylated CGAT|GCAA in vivo, recapitulating what was observed on the arrays. This methodology can be used to identify new methylated DNA sequences preferentially bound by TF, which may be functional in vivo. To evaluate the effect of CG methylation on DNA binding of sequence-specific B-ZIP transcription factors (TFs) in a high-throughput manner, we enzymatically methylated the cytosine in the CG dinucleotide on protein binding microarrays. Two Agilent DNA array designs were used. One contained 40,000 features using de Bruijn sequences where each 8-mer occurs 32 times in various positions in the DNA sequence. The second contained 180,000 features with each CG containing 8-mer present three times. The first design was better for identification of binding motifs, while the second was better for quantification. Using this novel technology, we show that CG methylation enhanced binding for CEBPA and CEBPB and inhibited binding for CREB, ATF4, JUN, JUND, CEBPD and CEBPG. The CEBPB|ATF4 heterodimer bound a novel motif CGAT|GCAA 10-fold better when methylated. EMSA confirmed these results. CEBPB ChIP-seq data using primary female mouse dermal fibroblasts with 50X methylome coverage for each strand indicate that the methylated sequences well-bound on the arrays are also bound in vivo. CEBPB bound 39% of the methylated canonical 10-mers ATTGC|GCAAT in the mouse genome. After ATF4 protein induction by thapsigargin which results in ER stress, CEBPB binds methylated CGAT|GCAA in vivo, recapitulating what was observed on the arrays. This methodology can be used to identify new methylated DNA sequences preferentially bound by TF, which may be functional in vivo. Protein binding microarray (PBM) experiments were performed for a set of 8 mouse B-ZIP homodimers and one hetrodimer transcription factors. Briefly, the PBMs involved binding GST-tagged DNA-binding proteins to double-stranded and methylated or unmethylated 44K Agilent microarrays, containing a DeBruijn sequence design, in order to determine their sequence preferences. Details of the PBM protocol are described in Berger et al., Nature Biotechnology 2006.

Dataset Information

Unraveling determinants of transcription factor binding outside the core binding site

Similar Datasets

OmicsDI is part of the ELIXIR infrastructure