Quantification of Cooperativity in Heterodimer-DNA Binding Improves the Accuracy of Binding Specificity Models

Learning Useful Representations of DNA Sequences From ChIP-Seq Datasets for Exploring Transcription Factor Binding Specificities

Deep learning has been successfully applied to surprisingly different domains. Researchers and practitioners are employing trained deep learning models to enrich our knowledge. Transcription factors (TFs)are essential for regulating gene expression in all organisms by binding to specific DNA sequences. Here, we designed a deep learning model named SemanticCS (Semantic ChIP-seq)to predict TF binding specificities. We trained our learning model on an ensemble of ChIP-seq datasets (Multi-TF-cell)to learn useful intermediate features across multiple TFs and cells. To interpret these feature vectors, visualization analysis was used. Our results indicate that these learned representations can be used to train shallow machines for other tasks. Using diverse experimental data and evaluation metrics, we show that SemanticCS outperforms other popular methods. In addition, from experimental data, SemanticCS can help to identify the substitutions that cause regulatory abnormalities and to evaluate the effect of substitutions on the binding affinity for the RXR transcription factor. The online server for SemanticCS is freely available at http://qianglab.scst.suda.edu.cn/semanticCS/.

A Comprehensive Drosophila melanogaster Transcription Factor Interactome

Combinatorial interactions among transcription factors (TFs) play essential roles in generating gene expression specificity and diversity in metazoans. Using yeast 2-hybrid (Y2H) assays on nearly all sequence-specific Drosophila TFs, we identified 1,983 protein-protein interactions (PPIs), more than doubling the number of currently known PPIs among Drosophila TFs. For quality assessment, we validated a subset of our interactions using MITOMI and bimolecular fluorescence complementation assays. We combined our interactome with prior PPI data to generate an integrated Drosophila TF-TF binary interaction network. Our analysis of ChIP-seq data, integrating PPI and gene expression information, uncovered different modes by which interacting TFs are recruited to DNA. We further demonstrate the utility of our Drosophila interactome in shedding light on human TF-TF interactions. This study reveals how TFs interact to bind regulatory elements in vivo and serves as a resource of Drosophila TF-TF binary PPIs for understanding tissue-specific gene regulation.

Combinatorial regulation by transcription factors (TFs) is one mechanism for achieving condition and tissue-specific gene regulation. Shokri et al. mapped TF-TF interactions between most Drosophila TFs, reporting a comprehensive TF-TF network integrated with previously known interactions. They used this network to discern distinct TF-DNA binding modes.

MODER2: first-order Markov modeling and discovery of monomeric and dimeric binding motifs

MOTIVATION

Position-specific probability matrices (PPMs, also called position-specific weight matrices) have been the dominating model for transcription factor (TF)-binding motifs in DNA. There is, however, increasing recent evidence of better performance of higher order models such as Markov models of order one, also called adjacent dinucleotide matrices (ADMs). ADMs can model dependencies between adjacent nucleotides, unlike PPMs. A modeling technique and software tool that would estimate such models simultaneously both for monomers and their dimers have been missing.

RESULTS

We present an ADM-based mixture model for monomeric and dimeric TF-binding motifs and an expectation maximization algorithm MODER2 for learning such models from training data and seeds. The model is a mixture that includes monomers and dimers, built from the monomers, with a description of the dimeric structure (spacing, orientation). The technique is modular, meaning that the co-operative effect of dimerization is made explicit by evaluating the difference between expected and observed models. The model is validated using HT-SELEX and generated datasets, and by comparing to some earlier PPM and ADM techniques. The ADM models explain data slightly better than PPM models for 314 tested TFs (or their DNA-binding domains) from four families (bHLH, bZIP, ETS and Homeodomain), the ADM mixture models by MODER2 being the best on average.

AVAILABILITY AND IMPLEMENTATION

Software implementation is available from https://github.com/jttoivon/moder2.

SUPPLEMENTARY INFORMATION

Supplementary data are available at Bioinformatics online.

Unraveling the Hierarchy of cis and trans Factors That Determine the DNA Binding by Peroxisome Proliferator-Activated Receptor γ

Peroxisome proliferator-activated receptor γ (PPARγ) is a nuclear receptor essential for adipocyte development and the maintenance of the alternatively polarized macrophage phenotype. Biochemical studies have established that as an obligate heterodimer with retinoid X receptor (RXR), PPARγ binds directly repeated nuclear receptor half sites spaced by one nucleotide (direct repeat 1 [DR1]). However, it has not been analyzed systematically and genome-wide how cis factors such as the sequences of DR1s and adjacent sequences and trans factors such as cobinding lineage-determining transcription factors (LDTFs) contribute to the direct binding of PPARγ in different cellular contexts. We developed a novel motif optimization approach using sequence composition and chromatin immunoprecipitation with high-throughput sequencing (ChIP-seq) densities from macrophages and adipocytes to complement de novo motif enrichment analysis and to define and classify high-affinity binding sites. We found that approximately half of the PPARγ cistrome represents direct DNA binding; both half sites can be extended upstream, and these are typically not of equal strength within a DR1. Strategically positioned LDTFs have greater impact on PPARγ binding than the quality of DR1, and the presence of the extension of DR1 provides a remarkable synergy with LDTFs. This approach of considering not only nucleotide frequencies but also their contribution to protein binding in a cellular context is applicable to other transcription factors.

Modular discovery of monomeric and dimeric transcription factor binding motifs for large data sets

In some dimeric cases of transcription factor (TF) binding, the specificity of dimeric motifs has been observed to differ notably from what would be expected were the two factors to bind to DNA independently of each other. Current motif discovery methods are unable to learn monomeric and dimeric motifs in modular fashion such that deviations from the expected motif would become explicit and the noise from dimeric occurrences would not corrupt monomeric models. We propose a novel modeling technique and an expectation maximization algorithm, implemented as software tool MODER, for discovering monomeric TF binding motifs and their dimeric combinations. Given training data and seeds for monomeric motifs, the algorithm learns in the same probabilistic framework a mixture model which represents monomeric motifs as standard position-specific probability matrices (PPMs), and dimeric motifs as pairs of monomeric PPMs, with associated orientation and spacing preferences. For dimers the model represents deviations from pure modular model of two independent monomers, thus making co-operative binding effects explicit. MODER can analyze in reasonable time tens of Mbps of training data. We validated the tool on HT-SELEX and ChIP-seq data. Our findings include some TFs whose expected model has palindromic symmetry but the observed model is directional.

Modulation of RXR-DNA complex assembly by DNA context

Retinoid X Receptors (RXRs) act as dimer partners for several nuclear receptors including itself, binding to genomic DNA response elements and regulating gene transcription with cell and gene specificity. As homodimers, RXRs bind direct repeats of the half-site (A/G)G(G/T)TCA separated by 1 nucleotide (DR1) and little variability of this consensus site is observed for natural DR1s. However, these variations are responsible of the modulation of RXR receptors function through differential binding affinity and conformational changes. To further our understanding of the molecular mechanisms underlying RXR-DNA interactions, we examined how RXR DBDs bind to different DR1s using thermodynamics, X-ray crystallography and nuclear magnetic resonance (NMR) spectroscopy. We show that the half-site sequences modulate the binding cooperativity that results from the protein-protein contacts between the two DBDs. Chemical shifts perturbation NMR experiments revealed that sequence variations in half-sites induce changes that propagate from the protein-DNA interface to the dimerization interface throughout the DBD fold.

Eicosapentaenoic acid induces DNA demethylation in carcinoma cells through a TET1-dependent mechanism

In cancer cells, global genomic hypomethylation is found together with localized hypermethylation of CpG islands within the promoters and regulatory regions of silenced tumor suppressor genes. Demethylating agents may reverse hypermethylation, thus promoting gene re-expression. Unfortunately, demethylating strategies are not efficient in solid tumor cells. DNA demethylation is mediated by ten-eleven translocation enzymes (TETs). They sequentially convert 5-methylcytosine (5mC) to 5-hydroxymethylcytosine (5hmC), which is associated with active transcription; 5-formylcytosine; and finally, 5-carboxylcytosine. Although α-linolenic acid, eicosapentaenoic acid (EPA), and docosahexaenoic acid, the major n-3 polyunsaturated fatty acids, have anti-cancer effects, their action, as DNA-demethylating agents, has never been investigated in solid tumor cells. Here, we report that EPA demethylates DNA in hepatocarcinoma cells. EPA rapidly increases 5hmC on DNA, inducing p21Waf1/Cip1 gene expression, which slows cancer cell-cycle progression. We show that the underlying molecular mechanism involves TET1. EPA simultaneously binds peroxisome proliferator-activated receptor γ (PPARγ) and retinoid X receptor α (RXRα), thus promoting their heterodimer and inducing a PPARγ-TET1 interaction. They generate a TET1-PPARγ-RXRα protein complex, which binds to a hypermethylated CpG island on the p21 gene, where TET1 converts 5mC to 5hmC. In an apparent shuttling motion, PPARγ and RXRα leave the DNA, whereas TET1 associates stably. Overall, EPA directly regulates DNA methylation levels, permitting TET1 to exert its anti-tumoral function.-Ceccarelli, V., Valentini, V., Ronchetti, S., Cannarile, L., Billi, M., Riccardi, C., Ottini, L., Talesa, V. N., Grignani, F., Vecchini, A., Eicosapentaenoic acid induces DNA demethylation in carcinoma cells through a TET1-dependent mechanism.

Assessing sufficiency and necessity of enhancer activities for gene expression and the mechanisms of transcription activation

Enhancers are important genomic regulatory elements directing cell type-specific transcription. They assume a key role during development and disease, and their identification and functional characterization have long been the focus of scientific interest. The advent of next-generation sequencing and clustered regularly interspaced short palindromic repeat (CRISPR)/Cas9-based genome editing has revolutionized the means by which we study enhancer biology. In this review, we cover recent developments in the prediction of enhancers based on chromatin characteristics and their identification by functional reporter assays and endogenous DNA perturbations. We discuss that the two latter approaches provide different and complementary insights, especially in assessing enhancer sufficiency and necessity for transcription activation. Furthermore, we discuss recent insights into mechanistic aspects of enhancer function, including findings about cofactor requirements and the role of post-translational histone modifications such as monomethylation of histone H3 Lys4 (H3K4me1). Finally, we survey how these approaches advance our understanding of transcription regulation with respect to promoter specificity and transcriptional bursting and provide an outlook covering open questions and promising developments.

Recruitment and delivery of the fission yeast Rst2 transcription factor via a local genome structure counteracts repression by Tup1-family corepressors

Transcription factors (TFs) determine the transcription activity of target genes and play a central role in controlling the transcription in response to various environmental stresses. Three dimensional genome structures such as local loops play a fundamental role in the regulation of transcription, although the link between such structures and the regulation of TF binding to cis-regulatory elements remains to be elucidated. Here, we show that during transcriptional activation of the fission yeast fbp1 gene, binding of Rst2 (a critical C2H2 zinc-finger TF) is mediated by a local loop structure. During fbp1 activation, Rst2 is first recruited to upstream-activating sequence 1 (UAS1), then it subsequently binds to UAS2 (a critical cis-regulatory site located approximately 600 base pairs downstream of UAS1) through a loop structure that brings UAS1 and UAS2 into spatially close proximity. Tup11/12 (the Tup-family corepressors) suppress direct binding of Rst2 to UAS2, but this suppression is counteracted by the recruitment of Rst2 at UAS1 and following delivery to UAS2 through a loop structure. These data demonstrate a previously unappreciated mechanism for the recruitment and expansion of TF-DNA interactions within a promoter mediated by local three-dimensional genome structures and for timely TF-binding via counteractive regulation by the Tup-family corepressors.

Coop-Seq Analysis Demonstrates that Sox2 Evokes Latent Specificities in the DNA Recognition by Pax6

Sox2 and Pax6 co-regulate genes in neural lineages and the lens by forming a ternary complex likely facilitated allosterically through DNA. We used the quantitative and scalable cooperativity-by-sequencing (Coop-seq) approach to interrogate Sox2/Pax6 dimerization on a DNA library where five positions of the Pax6 half-site were randomized yielding 1024 cooperativity factors. Consensus positions normally required for the high-affinity DNA binding by Pax6 need to be mutated for effective dimerization with Sox2. Out of the five randomized bases, a 5' thymidine is present in most of the top ranking elements. However, this thymidine maps to a region outside of the Pax half site and is not expected to directly interact with Pax6 in known binding modes suggesting structural reconfigurations. Re-analysis of ChIP-seq data identified several genomic regions where the cooperativity promoting sequence pattern is co-bound by Sox2 and Pax6. A highly conserved Sox2/Pax6 bound site near the Sprouty2 locus was verified to promote cooperative dimerization designating Sprouty2 as a potential target reliant on Sox2/Pax6 cooperativity in several neural cell types. Collectively, the functional interplay of Sox2 and Pax6 demands the relaxation of high-affinity binding sites and is enabled by alternative DNA sequences. We conclude that this binding mode evolved to warrant that a subset of target genes is only regulated in the presence of suitable partner factors.

Transcription factor-DNA binding: beyond binding site motifs

Sequence-specific transcription factors (TFs) regulate gene expression by binding to cis-regulatory elements in promoter and enhancer DNA. While studies of TF-DNA binding have focused on TFs' intrinsic preferences for primary nucleotide sequence motifs, recent studies have elucidated additional layers of complexity that modulate TF-DNA binding. In this review, we discuss technological developments for identifying TF binding preferences and highlight recent discoveries that elaborate how TF interactions, local DNA structure, and genomic features influence TF-DNA binding. We highlight novel approaches for characterizing functional binding site motifs that promise to inform our understanding of how TF binding controls gene expression and ultimately contributes to phenotype.

Systematic dissection of genomic features determining transcription factor binding and enhancer function

Enhancers regulate gene expression through the binding of sequence-specific transcription factors (TFs) to cognate motifs. Various features influence TF binding and enhancer function-including the chromatin state of the genomic locus, the affinities of the binding site, the activity of the bound TFs, and interactions among TFs. However, the precise nature and relative contributions of these features remain unclear. Here, we used massively parallel reporter assays (MPRAs) involving 32,115 natural and synthetic enhancers, together with high-throughput in vivo binding assays, to systematically dissect the contribution of each of these features to the binding and activity of genomic regulatory elements that contain motifs for PPARγ, a TF that serves as a key regulator of adipogenesis. We show that distinct sets of features govern PPARγ binding vs. enhancer activity. PPARγ binding is largely governed by the affinity of the specific motif site and higher-order features of the larger genomic locus, such as chromatin accessibility. In contrast, the enhancer activity of PPARγ binding sites depends on varying contributions from dozens of TFs in the immediate vicinity, including interactions between combinations of these TFs. Different pairs of motifs follow different interaction rules, including subadditive, additive, and superadditive interactions among specific classes of TFs, with both spatially constrained and flexible grammars. Our results provide a paradigm for the systematic characterization of the genomic features underlying regulatory elements, applicable to the design of synthetic regulatory elements or the interpretation of human genetic variation.

A systems perspective on brown adipogenesis and metabolic activation

Brown adipocytes regulate energy expenditure via mitochondrial uncoupling. This makes these fat cells attractive therapeutic targets to tackle the burgeoning issue of obesity, which itself is coupled to insulin resistance, type 2 diabetes, cardiovascular and fatty liver disease. Recent research has revealed a complex network underlying brown fat cell differentiation and thermogenic activation, involving secreted factors, signal transduction, metabolic pathways and gene regulatory components. Given that brown fat is now reported to be present in adult humans, it is desirable to harness the knowledge from each network module to design effective therapeutic strategies. In this review, we will present a systems perspective on brown adipogenesis and the subsequent metabolic activation of brown adipocytes by integrating signaling, metabolic and gene regulatory modules with a specific focus on known 'druggable' targets within each module.

The Role of Peroxisome Proliferator-Activated Receptor Gamma (PPARG) in Adipogenesis: Applying Knowledge from the Fish Aquaculture Industry to Biomedical Research

The tropical freshwater zebrafish has recently emerged as a valuable model organism for the study of adipose tissue biology and obesity-related disease. The strengths of the zebrafish model system are its wealth of genetic mutants, transgenic tools, and amenability to high-resolution imaging of cell dynamics within live animals. However, zebrafish adipose research is at a nascent stage and many gaps exist in our understanding of zebrafish adipose physiology and metabolism. By contrast, adipose research within other, closely related, teleost species has a rich and extensive history, owing to the economic importance of these fish as a food source. Here, we compare and contrast knowledge on peroxisome proliferator-activated receptor gamma (PPARG)-mediated adipogenesis derived from both biomedical and aquaculture literatures. We first concentrate on the biomedical literature to (i) briefly review PPARG-mediated adipogenesis in mammals, before (ii) reviewing Pparg-mediated adipogenesis in zebrafish. Finally, we (iii) mine the aquaculture literature to compare and contrast Pparg-mediated adipogenesis in aquaculturally relevant teleosts. Our goal is to highlight evolutionary similarities and differences in adipose biology that will inform our understanding of the role of adipose tissue in obesity and related disease.

The Genetics of Transcription Factor DNA Binding Variation

Most complex trait-associated variants are located in non-coding regulatory regions of the genome, where they have been shown to disrupt transcription factor (TF)-DNA binding motifs. Variable TF-DNA interactions are therefore increasingly considered as key drivers of phenotypic variation. However, recent genome-wide studies revealed that the majority of variable TF-DNA binding events are not driven by sequence alterations in the motif of the studied TF. This observation implies that the molecular mechanisms underlying TF-DNA binding variation and, by extrapolation, inter-individual phenotypic variation are more complex than originally anticipated. Here, we summarize the findings that led to this important paradigm shift and review proposed mechanisms for local, proximal, or distal genetic variation-driven variable TF-DNA binding. In addition, we discuss the biomedical implications of these findings for our ability to dissect the molecular role(s) of non-coding genetic variants in complex traits, including disease susceptibility.

Quantitative profiling of selective Sox/POU pairing on hundreds of sequences in parallel by Coop-seq

Cooperative binding of transcription factors is known to be important in the regulation of gene expression programs conferring cellular identities. However, current methods to measure cooperativity parameters have been laborious and therefore limited to studying only a few sequence variants at a time. We developed Coop-seq (cooperativity by sequencing) that is capable of efficiently and accurately determining the cooperativity parameters for hundreds of different DNA sequences in a single experiment. We apply Coop-seq to 12 dimer pairs from the Sox and POU families of transcription factors using 324 unique sequences with changed half-site orientation, altered spacing and discrete randomization within the binding elements. The study reveals specific dimerization profiles of different Sox factors with Oct4. By contrast, Oct4 and the three neural class III POU factors Brn2, Brn4 and Oct6 assemble with Sox2 in a surprisingly indistinguishable manner. Two novel half-site configurations can support functional Sox/Oct dimerization in addition to known composite motifs. Moreover, Coop-seq uncovers a nucleotide switch within the POU half-site when spacing is altered, which is mirrored in genomic loci bound by Sox2/Oct4 complexes.