medizinische genetik 2021; 33(2): 147–155 Janna Leiz, Maria Rutkiewicz, Carmen Birchmeier, Udo Heinemann*, and Kai M. Schmidt-Ott* Technologies for profling the impact of genomic variants on transcription factor binding https://doi.org/10.1515/medgen-2021-2073 Received February 3, 2021; accepted June 24, 2021 Introduction Abstract: Transcription factors (TFs) bind DNA in a Transcription factors (TFs) are regulatory proteins that sequence-specifc manner and thereby regulate target bind DNA in a sequence-specifc manner. Along with chro- gene expression. TF binding and its regulatory activity is matin accessibility and histone modifcation, TFs regulate highly context dependent, and is not only determined by the expression of target genes depending on cell type, de- specifc cell types or diferentiation stages but also relies velopmental stage, or external signals; for a detailed re- on other regulatory mechanisms, such as DNA and chro- view on the regulatory epigenome and histone modifca- matin modifcations. Interactions between TFs and their tions see [1, 2]. DNA binding sites are critical mediators of phenotypic Over 1,000 potential TFs have been identifed in hu- variation and play important roles in the onset of disease. mans. In general, TFs are highly conserved among species A continuously growing number of studies therefore at- and recognize specifc nucleotide sequences or motifs in tempts to elucidate TF:DNA interactions to gain knowl- non-coding regulatory regions of the genome. Binding is edge about regulatory mechanisms and disease-causing based on the complementarity of the DNA sequence and variants. Here we summarize how TF-binding characteris- protein structure. As recognition sites are rather short, tics and the impact of variants can be investigated, how usually 6–12 bases, and every TF has numerous bind- bioinformatic tools can be used to analyze and predict ing sites throughout the genome, many TFs bind coop- TF:DNA binding, and what additional information can be eratively as multimers to ensure highly specifc and sta- obtained from the TF protein structure. ble interactions. To add an additional layer of control to Keywords: transcriptional regulation, genomic variants, the rigorously regulated process of gene expression, many TF:DNA binding, binding prediction TFs recruit cofactors or depend on binding of specifc lig- ands [3]. Though the majority of the genomic DNA is usually densely packed in nucleosomes and higher-order struc- tures, making it inaccessible to TFs, a special class of TFs, *Corresponding authors: Udo Heinemann, Max-Delbrück-Center for called “pioneer factors,” are able to bind to their recogni- Molecular Medicine in the Helmholtz Association (MDC), tion sites even in condensed chromatin. They thereby in- Macromolecular Structure and Interaction, Berlin, Germany, e-mail: duce changes to the chromatin structure that enable bind- [email protected]; and Kai M. Schmidt-Ott, ing of other factors needed to initiate transcription [4]. To Charité–Universitätsmedizin Berlin, Corporate Member of Freie amplify the expression of a protein-coding target gene, Universität Berlin and Humboldt-Universität zu Berlin, Department TFs guide RNA polymerase II to gene promoter regions of Nephrology and Intensive Care Medicine, Hindenburgdamm 30, 12203 Berlin, Germany; and Max-Delbrück-Center for Molecular to start the transcription process (Figure 1). TFs can also Medicine in the Helmholtz Association (MDC), Molecular and block binding sites for other proteins and thereby, depend- Translational Kidney Research, Robert-Rössle-Str. 10, 13125 Berlin, ing on the specifc context, act as repressors. How DNA Germany, e-mail: [email protected] sequence variants in TF binding sites impact site recog- Janna Leiz, Charité–Universitätsmedizin Berlin, Corporate Member nition and transcriptional regulation is difcult to pre- of Freie Universität Berlin and Humboldt-Universität zu Berlin, Department of Nephrology and Intensive Care Medicine, dict. Hindenburgdamm 30, 12203 Berlin, Germany; and Though we tend to think of TFs as either occupying Max-Delbrück-Center for Molecular Medicine in the Helmholtz particular sequences in the chromatin or not, many TFs Association (MDC), Molecular and Translational Kidney Research, Robert-Rössle-Str. 10, 13125 Berlin, Germany, e-mail: [email protected] Maria Rutkiewicz, Max-Delbrück-Center for Molecular Medicine in the Helmholtz Association (MDC), Macromolecular Structure and Carmen Birchmeier, Max-Delbrück-Center for Molecular Medicine in Interaction, Berlin, Germany, e-mail: the Helmholtz Association (MDC), Developmental Biology and Signal [email protected] Transduction, Berlin, Germany, e-mail: [email protected] Open Access. © 2021 Leiz et al., published by De Gruyter. This work is licensed under the Creative Commons Attribution 4.0 International License. 148 | J. Leiz et al., Technologies for profling the impact of genomic variants on transcription factor binding Figure 1: Transcriptional activation is mediated by chromatin state and transcription factor binding. Densely condensed chromatin (closed) prevents transcription factors and other proteins needed to initiate transcription from binding, thereby inhibiting gene expression. Open chromatin on the other hand is accessible for proteins to bind to promoter and enhancer regions. Transcription factors can either directly guide RNA polymerase II to promoter regions of target genes or work in cooperation with other factors and mediators to assemble a tran- scription initiation complex. TF, transcription factor; Pol II, RNA polymerase II; CoF, cofactor; M, mediator. are controlled in a more complex manner. Some TFs are sociated with altered target gene expression and disease produced in response to external or internal stimuli in a [9, 10]. This emphasizes the necessity of examining ge- pulsed, oscillating, or sustained manner. Consequently, netic variants in the context of patient clinical phenotypes. the genes that are switched on by such TFs depend on For example, single-nucleotide variants (SNVs) in the pro- their expression dynamics [5, 6]. The TF MyoD, a mas- moter region of coagulation factor F9 have been associ- ter regulatory factor of skeletal muscle known for its abil- ated with a specifc subtype of the blood-clotting disor- ity to initiate the muscle-specifc diferentiation program, der hemophilia B. More than 20 diferent SNVs in three is one example. In proliferating muscle precursor cells, distinct clusters afecting highly conserved base pair po- MyoD oscillates with a periodicity of 2–3 hours, thereby sitions in the binding sites of the TFs HNF4α, ONECUT1/2, keeping the cells in an undiferentiated state. This oscil- and C/EBPα have been identifed. All three binding sites latory MyoD pattern changes before the cells diferentiate. are in close proximity, and disruption of either one of When MyoD expression becomes sustained, cells start to them signifcantly decreases promoter activity, indicating diferentiate and undergo fusion into myotubes [7]. This a possible cooperation of HNF4α, ONECUT1/2, and C/EBPα implies that the biological function of TF binding should to control F9 expression. This example shows how alter- not only be investigated statically, but in a time resolved ations of single bases in TF-binding motifs can disrupt fashion. TF:DNA interactions and critically alter gene expression The impact of mutations in genes coding for TFs and leading to disease [11]. resulting structural changes of the proteins have been in- Here, we review diferent in vitro and in vivo tech- vestigated and linked to diseases in many studies. Be- niques that can be used to investigate TF:DNA binding cause the protein-coding exome covers less than 2 % of characteristics and describe how the impact of variants on the human genome, the focus of variant analysis has been the binding capacity can be measured. We outline the re- expanded to non-coding regulatory regions (Krude et al. cently made advances of bioinformatic tools and machine this edition, Guo et al. this edition). Approximately 80 % learning approaches to predict and analyze TF-binding of all genome-wide association study (GWAS) hits afect sites (TFBSs) in the human genome. We further review the non-coding DNA and many of them are thought to act techniques for deriving quantitative TF:DNA afnity data through diferential TF binding [8]. Although it is often and discuss how they can be related to high-resolution challenging to identify disease-associated variants and to structures of TFs bound to their DNA targets. We fnish by prove causation, several studies show that variants in the discussing future challenges relating structures of target regulatory genome alter TF:DNA interactions and are as- DNA-bound TFs with quantitative afnity data. J. Leiz et al., Technologies for profling the impact of genomic variants on transcription factor binding | 149 How can transcription factors and How can we measure transcription their binding motifs be factor:DNA interactions and the investigated? impact of DNA variants on transcription factor binding? There are several techniques to identify potential reg- ulatory domains with TFBSs and to assay chromatin The ChIP-seq techniques map the TF-binding DNA sites accessibility on a genomic scale. Widely used in an in a genome-wide fashion, but result in low spatial res- ever-increasing number of publications are chromatin olution and permit only limited assessment of binding immuno-precipitation followed by sequencing