
==== Front
Mol Cell
Mol Cell
Molecular Cell
1097-2765
1097-4164
Cell Press

S1097-2765(24)00663-4
10.1016/j.molcel.2024.08.009
Article
Genome access is transcription factor-specific and defined by nucleosome position
Grand Ralph Stefan 125
Pregnolato Marco 135
Baumgartner Lisa 1
Hoerner Leslie 1
Burger Lukas 14
Schübeler Dirk dirk@fmi.ch
136∗
1 Friedrich Miescher Institute for Biomedical Research, 4058 Basel, Switzerland
2 Zentrum für Molekulare Biologie der Universität Heidelberg (ZMBH), DKFZ-ZMBH Alliance, 69120 Heidelberg, Germany
3 Faculty of Sciences, University of Basel, 4003 Basel, Switzerland
4 Swiss Institute of Bioinformatics, 4058 Basel, Switzerland
∗ Corresponding author dirk@fmi.ch
5 These authors contributed equally

6 Lead contact

19 9 2024
19 9 2024
84 18 34553468.e6
28 7 2023
14 6 2024
6 8 2024
© 2024 The Authors
2024
https://creativecommons.org/licenses/by-nc-nd/4.0/ This is an open access article under the CC BY-NC-ND license (http://creativecommons.org/licenses/by-nc-nd/4.0/).
Summary

Mammalian gene expression is controlled by transcription factors (TFs) that engage sequence motifs in a chromatinized genome, where nucleosomes can restrict DNA access. Yet, how nucleosomes affect individual TFs remains unclear. Here, we measure the ability of over one hundred TF motifs to recruit TFs in a defined chromosomal locus in mouse embryonic stem cells. This identifies a set sufficient to enable the binding of TFs with diverse tissue specificities, functions, and DNA-binding domains. These chromatin-competent factors are further classified when challenged to engage motifs within a highly phased nucleosome. The pluripotency factors OCT4-SOX2 preferentially engage non-nucleosomal and entry-exit motifs, but not nucleosome-internal sites, a preference that also guides binding genome wide. By contrast, factors such as BANP, REST, or CTCF engage throughout, causing nucleosomal displacement. This supports that TFs vary widely in their sensitivity to nucleosomes and that genome access is TF specific and influenced by nucleosome position in the cell.

Graphical abstract

Highlights

• Over 100 motifs tested for their ability to recruit transcription factors in the cell

• Single-molecule footprinting reveals that only few motifs enable TF binding

• Proficient TFs contain diverse DNA-binding domains, tissue specificity, and functions

• Binding is sensitive to the exact nucleosome position

Packaging of DNA into nucleosomes is an important part of restricting and specifying the genomic binding patterns of transcription factors (TFs). Grand et al. quantify, at single-molecule resolution, how sensitive TF-recruiting DNA motifs are to chromatin-based occlusion and how exact nucleosome position impacts TF binding in the genome.

Keywords

chromatin
nucleosome
transcription factors
pioneer factors
genome
accessibility
Published: August 28, 2024
==== Body
pmcIntroduction

Development of multicellular organisms relies on the establishment of specific gene expression patterns. Key players in this process are transcription factors (TFs), proteins that bind short DNA sequence motifs in regulatory regions.1,2 In eukaryotes, this binding occurs in a chromatinized genome, where DNA methylation, nucleosome position, and modification can affect motif recognition.3,4,5,6 Fundamentally, the wrapping of 147 bp of DNA around histone proteins to form nucleosomes reduces access to >95% of the DNA, presenting a major barrier for the binding of most TFs.6,7,8 High resolution studies into how TFs bind to a nucleosomal substrate in vitro have identified diverse strategies for motif recognition and engagement,6,9,10,11,12,13,14 with some TFs suggested to favor nucleosomal embedded motifs.11,15 These TFs, also referred to as pioneer factors, are thought to mediate the first encounter with closed chromatin to initiate chromatin opening and enable binding of chromatin-sensitive factors, eventually promoting gene activation. It remains difficult, however, to test whether in vitro binding to reconstituted nucleosomes reflects motif recognition by individual TFs in the context of the cell. Here, regulatory regions are composed of complex assemblies of multiple TF motifs and dinucleotide frequencies, and chromatin remodeling enzymes move and evict nucleosomes. Importantly, these enzymes have been shown to not only be required for proper binding of many TFs16 but, in the case of SWI/SNF, also to continuously restore open chromatin.17,18 This suggests that motif binding and nucleosome rearrangement by TFs happens repeatedly, even at sites that reside in open chromatin. It is further unclear whether chromatin engagement and opening is limited to pioneer factors that drive developmental changes or whether it is a feature also found in broadly expressed TFs.11,12,19,20,21

Despite its fundamental relevance for our understanding of genomic regulation, our knowledge of TF sensitivity to chromatin in vivo remains incomplete, and different mechanisms of engagement with nucleosomal DNA have been proposed. For example, the reprogramming factors OCT4, SOX2, and KLF4 are thought to bind partial motifs via a subset of their DNA-binding domains (DBDs), which can adapt to the surface of the nucleosome and enable stable binding.11 Moreover, to exert its pioneer activity and generate an open chromatin region, OCT4 recruits the chromatin remodeler BRG1, which stabilizes binding of SOX2 at OCT4-occupied sites.22 Notably, these factors have been suggested to bind preferentially to inaccessible chromatin,11,15 a feature that is in contrast with the general preference of most TFs to bind open chromatin. In vitro, OCT4-SOX2 can engage with a singular consensus motif at the entry-exit site of a highly positioned nucleosome,10 while engagement can occur more internally in the context of a less-positioned nucleosome with multiple OCT4 motifs.23 This suggests that motif position, number, and nucleosome breathing govern OCT4-SOX2 binding. The paradigm pioneer factor FOXA1, instead, is thought to bind nucleosomal DNA because of the similarity in the structure of its DBD to the linker histone H1 and is able to displace nucleosomes without recruiting remodelers in vitro.13,24

In the context of the cell, however, these TFs occupy just a minor fraction of their consensus motif occurrences across the genome,10,25 arguing for a more complex regulation of motif access, even for pioneer factors. For example, at inaccessible chromatin, FOXA1 has been shown to depend on the binding of the non-pioneer factor HNF4a to access a subset of its motifs in vivo as a function of DNA affinity.26 Despite the large number of in vivo studies that use genomics to investigate TF binding, nucleosome positioning and epigenetic marks, the results remain largely correlative and do not reveal the effect of local chromatin state on the ability of individual TFs to bind their motifs.15,22,27,28,29 In particular, inferring chromatin sensitivity from genomic binding is difficult because different and multiple motifs often co-occur within the same regulatory region,30,31,32,33 making it hard to determine the interplay between chromatin and individual TFs. To circumvent this limitation, we engineered a reductionist approach that enables us to simultaneously evaluate in vivo occupancy of individual TF motifs and nucleosomes in different chromatin contexts using single-molecule footprinting (SMF), also known as NOMe-seq.34,35 With this approach, we comprehensively surveyed most-known motifs of TFs robustly expressed in mouse embryonic stem cells (mESCs). This survey of over 100 single motifs reveals a set of TFs, including the pioneer factors OCT4 and SOX2, which are able to engage with their consensus motif in a sequence with unphased nucleosome positioning. We further tested the ability of these chromatin-competent TFs to access their motif on nucleosomal DNA by specifically positioning a nucleosome in the cell. Strikingly, we found that the pioneer factors OCT4-SOX2 showed a drastic reduction in binding when the motif was nucleosome occupied, similar to non-pioneer factors, a result that we also observe genome wide. On the other hand, TFs such as BANP, REST, and CTCF were able to fully engage with their motif and drive nucleosome disruption independent of nucleosome position. These results demonstrate that, in the cellular context, the sensitivity to nucleosomes is strongly TF dependent and not necessarily linked to a potential developmental function of the respective TF, with nucleosome-internal sites being generally less accessible. These findings provide further evidence that chromatin sensitivity can contribute to shaping specific TF-binding patterns in the genome.

Results

A reductionist approach to systematically survey individual TF motifs for factor recruitment and nucleosome rearrangement

Throughout mammalian genomes, TFs predominantly bind in promoter and enhancer regions, which are complex assemblies of sequence motifs.30,31,32,33 To profile and compare the ability of individual TFs to bind and rearrange nucleosomes in vivo, we reasoned that their binding needs to be quantified in the same chromatinized template outside the context of regulatory regions. We set out to define the sensitivity of TFs to nucleosomes by introducing single copies of their consensus motifs into a defined sequence and genomic context in mESCs, and quantifying TF binding and local chromatin accessibility. More specifically, we inserted single occurrences of consensus TF motifs into a previously described, in vitro-derived sequence that has high CG density and low nucleosome affinity,36 resembling endogenous promoter regions, and integrated this sequence into a defined locus in mESCs using recombinase-mediated cassette exchange (RMCE),37 hereafter referred to as the RMCE-locus (Figure 1A). To quantify and compare the occupancy of a large set of TF motifs in parallel, we required an approach that is factor agnostic and able to directly read TF and nucleosome occupancy without enrichment steps. This can be achieved with SMF, which uses a DNA methyltransferase as a sensor and bisulfite sequencing as a readout.34 Here, we used it to simultaneously quantify TF and nucleosome occupancy on individual DNA molecules and, thus, to identify subpopulations of bound and unbound states for each tested sequence. This approach has been shown to work genome wide as well as at engineered genomic loci.34,35,36,38,39 First, we monitored the chromatin state of our background sequence in the absence of a motif using SMF following genomic integration. This revealed an intermediate accessibility profile, indicative of unphased nucleosomes and in line with the low nucleosome affinity of this sequence (Figure S1A). The sequence also displayed low levels of endogenous DNA methylation (Figure S1A), reflecting that it is relatively CpG rich,40 and, accordingly, absence of the heterochromatic histone modification histone H3 lysine 9 trimethylation (H3K9me3) (Figure S1B). We further did not detect enrichment of the active histone modification histone H3 lysine 27 acetylation (H3K27ac), indicating that the sequence does not intrinsically resemble an active regulatory region (Figure S1B). In the context of this sequence, we sought to test all TFs that are expressed in mESCs and for which a consensus motif was available via the JASPAR database (STAR Methods). This definition yielded a set of 114 TF motif sequences, to which we added 14 scrambled motifs that served as controls. To increase the throughput, these were inserted in pools of 10 constructs after we verified that the signal obtained for single insertions and pools was highly reproducible (Figure S1C). Combined, we obtained high-quality SMF data for 121 out of a total of 128 different constructs (Table S1), enabling detailed and quantitative analysis of motif occupancy and nucleosome position.Figure 1 A comprehensive survey of motifs for their individual ability to recruit TFs

(A) Schematic overview of single-molecule footprinting (SMF): following targeted genomic integration of single-TF-motif occurrences, nuclei are isolated and treated with the GpC methyltransferase M. CviPI. Footprinted DNA is bisulfite converted and sequenced, enabling the quantification of TF and nucleosome occupancy on individual DNA molecules.

(B) Top: footprint detected over the REST motif compared with the scrambled control. SMF signal is displayed as a percentage of reads being unmethylated at each individual GCH position. Mean of biological replicates is plotted (dots) and interpolated (lines) together with SEM (shaded area). Bottom: mean footprint of combined biological replicates displayed in a heatmap format (n = 2).

(C) Left: footprints detected at single-motif occurrences for TFs expressed in mESCs (see STAR Methods, n = 107). Motifs’ names indicate the annotated cognate TF on JASPAR. Data are sorted by footprint amplitude (SMF score) and interpolated mean of biological replicates is plotted (n = 2). Right: line plots show average SMF signal of respective clusters. Red bar indicates location of TF motif.

(D) Tissue specificity of TFs in each SMF cluster (see STAR Methods).

(E) Expression level (log2 RPKM) of TFs in different clusters. Black lines correspond to median, boxes to first and third quartiles, and whiskers to the maximum and minimum values of each distribution (n = 2).

(F) Length of motifs in the different clusters. Black lines correspond to median, boxes to first and third quartiles, and whiskers to the maximum and minimum values of the distribution.

Number of motifs per cluster: cluster 1 = 10; cluster 2 = 14; cluster 3 = 83. n.s., not significant, ∗∗∗p < 0.001 (two-tailed, unpaired Mann-Whitney U test).

See also Figure S1.

Systematic survey of TF motifs reveals marked variability in factor recruitment and chromatin opening

Analysis of GpC methylation over each motif in the library revealed variable footprint strength, allowing us to cluster motifs into three main classes of TF footprints (Figures S1D and S1E; STAR Methods). Clusters 1 and 2, representing ∼25% of tested motifs, showed strong and weak footprints, respectively, while there was no discernible footprint or prominent nucleosome rearrangement for the third cluster (Figures 1B, 1C, S1F, and S1G). The latter agrees with the notion that most TFs only engage with a small subgroup of motif occurrences in the genome25,26,41 and is in line with a model where the recruitment of most TFs requires additional features such as motif multimers or motif combinations.26,35,42,43 An example is illustrated by the combined OCT4-SOX2 motif, which shows a clear footprint, while the individual OCT4 and SOX2 motifs show no reproducible signal (Figures 1C, S1E, and S1G), resembling the genome-wide binding of OCT4 and SOX2.10 Notably, the few additional composite motifs included in our screen did not show a difference in TF binding or nucleosome rearrangement compared with the individual motifs (Figures 1C, S1E, and S1G).

We also noted the lack of a footprint for NRF1 (Figure S1H), even though this TF has been shown to give a signal in the genome.35 We had previously shown by chromatin immunoprecipitation (ChIP) that NRF1 can bind in the presence of a neighboring bound REST motif (which is a member of cluster 1), likely due to the ability of this TF to cause chromatin opening and hypomethylation of DNA.44 Indeed, testing a similar construct with SMF shows an increase in footprint signal at the NRF1 motif as well as a shift in the nucleosome position in presence of the REST motif (Figure S1H). This indicates increased binding of NRF1, which was independently validated by ChIP-qPCR (Figure S1I). This argues that the absence of signal at the individual NRF1 motif is a result of chromatin-mediated inhibition of binding that can be weakened by adding additional motifs. We speculate that this is similarly the case for other non-footprintable motifs tested, yet it remains formally possible that additional features, such as TF-binding dynamics and residence time, limit our ability to detect binding events for some TFs using SMF in our experimental setup.

To ask whether there was a common feature that defined footprintable TFs, we explored several parameters. The ability to create a TF footprint and, in addition, to rearrange nucleosomes did not relate to the level or tissue specificity of TF expression, presence of intrinsically disordered regions (IDRs), or utilized class of DBD (Figures 1D, 1E, and S1J; Table S1). Motif length, however, is on average significantly longer for TFs capable of footprinting (Figure 1F), in line with the notion that binding strength, in principle, increases with motif length due to extended contacts between the DBD and the DNA.45

Interestingly, motifs from the first two classes displayed a spectrum of different impacts on nucleosomes, ranging from no rearrangement to nearly complete displacement and generation of open chromatin. In particular, a group of targets, representing 3% of tested motifs, showed strong phased nucleosomes adjacent to the motif and, thus, open chromatin in the majority of cells. These motifs included those for CTCF, REST, and BANP, which also showed a stronger TF footprint compared with other motifs in the same class (Figures 1B and 1C), likely reflecting high occupancy. This is in line with previous genome-wide analyses, where these TFs have been described to bind their strongest motifs at a high frequency, coinciding with phased nucleosomes.36,46,47

These results indicate that a subset of single-motif occurrences are sufficient to recruit the cognate TF within a chromatinized sequence occupied by unphased nucleosomes.

A sequence that positions nucleosomes in vitro fails to do so in chromatin

Having identified motifs that are bound by their cognate TFs, we next wanted to ask whether binding is affected by motif position relative to a nucleosome. This is motivated by the fact that TF binding varies greatly along nucleosomal DNA when tested in vitro, with internally positioned motifs being generally less accessible than those placed at non-nucleosomal and entry-exit sites.6,10 Our initial survey does not inform on this question, as nucleosomes are not phased within the tested sequence context but, instead, reside at different positions between cells (Figure S1A).48 With the aim of generating a phased nucleosome in the chromosomal context, we inserted the Widom 601 (W601) sequence that has high affinity for nucleosomes in vitro49—and for which there is some evidence that it can position nucleosomes in vivo.50,51,52 Following genomic insertion, however, the footprinting profile of the W601 sequence displayed an intermediate accessibility profile similar to the construct used for the screen, indicative of unphased nucleosomes (Figure 2A). Indeed, single-molecule analysis of the SMF reads revealed the presence of a nucleosome at different positions spanning the entire length of the sequence, leading to a defined position only in ∼10% to 30% of cells at any given time (Figure 2A). This result shows that even a sequence that strongly positions nucleosomes in vitro does not position them in vivo. This is in line with previous results in yeast,53,54 and suggests that chromatin remodeler activity likely overrides sequence-intrinsic features that favor specific nucleosome positioning in vitro.Figure 2 TF-assisted positioning of a nucleosome in the context of the cell

(A) Top left: nucleosome positioning at the W601 sequence measured by SMF following targeted genomic integration (brown bar: minimal 601) (n = 3, mean ± SEM is plotted as in Figure 1B). Bottom left: protected and accessible regions observed on individual DNA molecules from SMF data. Right: average profile of nucleosome occupancy and position in the individual single-molecule clusters and relative quantification (% of all single molecules, n = 1000).

(B) Left: nucleosome positioning and REST binding at a REST-bound site in the genome, as measured by MNase-seq and ChIP-seq, respectively. Right: average profile of REST footprint and adjacent phased nucleosome of the highlighted region, as measured by SMF (n = 2, mean ± SEM is plotted as in Figure 1B).

(C) Top left: average profile of the REST footprint and adjacent phased nucleosome when the REST motif (blue bar) was inserted upstream of the unphased nucleosome positioning sequence used in the motif screen (n = 3, mean ± SEM is plotted as in Figure 1B). Bottom left: protected and accessible regions observed on individual DNA molecules from SMF data. Right: average profiles of REST and nucleosome occupancy and position in the individual clusters and relative quantification (% of all single molecules, n = 1000).

See also Figure S2.

TF-mediated nucleosome phasing enables precise nucleosome positioning in vivo

Throughout the genome, phased nucleosomes tend to be located adjacent to transcriptional start sites (TSSs) or in the vicinity of certain TF-binding sites.48 For example, well-phased nucleosomes are observed upstream and downstream of bound CTCF and REST motifs, as measured by MNase-seq,16 as well as SMF.34,35 We similarly detected such well-phased nucleosomes at regions bound by either REST or CTCF using SMF, as indicated by the presence of highly accessible DNA between the TF and nucleosome footprints, which suggested the presence of a well-positioned nucleosome in the majority of cells (Figures 2B and S2A). To determine whether we could employ TF-mediated nucleosome phasing to generate a well-positioned nucleosome at the RMCE-locus, we placed a REST motif at one end of the unphased nucleosome sequence used for the screen. Following integration, SMF indeed detected high REST occupancy and a phased nucleosome next to the TF footprint in the majority of cells (Figure 2C). Here, the non-nucleosomal region between the positioning TF and the nucleosome showed little variability, resulting in the presence of a nucleosome at a defined position in ∼70% of cells (hereafter referred to as a REST-phased nucleosome). Importantly, the slight variability observed suggests that this is a labile nucleosome that displays physiological dynamics, a feature that is thought to be required for TFs to engage with nucleosomal substrates.23,55 Thus, our strategy enables precise phasing of a natural nucleosome in vivo and, in turn, should enable us to assess the ability of TFs to bind as a function of nucleosome position in the cellular context.

Nucleosomes impair binding in a TF-specific fashion

Having generated a well-positioned nucleosome in chromatin, we next tested the ability of several TFs to bind a nucleosomal versus a non-nucleosomal motif. We focused on motifs that generated detectable footprints in our initial screen and included some motifs that did not show footprints as well as scrambled motifs as controls. For this, we inserted 17 motifs and six scrambled motif controls in the center of the REST-phased nucleosome construct and engineered it so that the motifs would be in a non-nucleosomal or more-nucleosome-occupied region by placing the REST motif either 67 or 137 bp upstream of the test motifs, respectively (Figures S3A and S3B). These constructs were inserted into the RMCE-locus followed by SMF, which displayed high replicate reproducibility (Figures S3C and S3D). Here, all tested motifs that showed a footprint signal in the unphased nucleosome sequence also footprinted when placed in the non-nucleosomal region next to REST (Figures 3A and S3E–S3H). This indicates that binding of REST, which functions as a transcriptional repressor, does not interfere with TF recruitment in this setting. By contrast, the majority of motifs, including the heterodimeric motif of the pioneer TFs OCT4-SOX2, showed a decreased footprint compared with the non-nucleosomal region when placed at a nucleosome-internal position (Figures 3B, 3C, and S3L). Furthermore, they did not cause significant nucleosome displacement in the majority of cells (Figures 3B and S3H–S3J). Notable exceptions were CTCF, REST, and BANP, which were able to bind and cause major nucleosome rearrangement, even when the motif was placed more internally on the nucleosome (Figure 3B). This suggests that nucleosome presence can impair genomic binding in a TF-specific manner and is in line with in vitro data that show that the majority of TFs have a reduced affinity for nucleosome-embedded motifs.6,10Figure 3 Differential occupancy of tested TF motifs at a non-nucleosomal versus a nucleosomal site

(A) Top: average profile of the REST footprint and adjacent phased nucleosome when the REST motif (blue bar) was inserted 67 bp upstream of the tested motif (red bar) in the unphased nucleosome positioning sequence used in the motif screen (n = 3, mean is plotted as in Figure 1B). Bottom: heatmap of the SMF detected at single-motif occurrences for selected TFs (n = 17) in the non-nucleosomal region of the REST-phased nucleosome context (non-nucleosomal motifs). SMF is plotted relative to the maximum SMF signal over the REST motif (see STAR Methods). Data are sorted by footprint amplitude (SMF score) and interpolated mean of biological replicates is plotted (n ≥ 2).

(B) Top: average profile of the REST footprint and adjacent phased nucleosome when the REST motif (blue rectangle) was inserted 137 bp upstream of the tested motif (red rectangle) in the unphased nucleosome positioning sequence used in the motif screen (n = 3, mean is plotted as in Figure 1B). Bottom: heatmap of the SMF detected at single-motif occurrences for selected TFs (n = 17) in the nucleosomal region of the REST-phased nucleosome context (nucleosomal motifs). Data are plotted as in (B) (n ≥ 2).

(C) Quantification of the footprint for each tested TF motif when placed in the non-nucleosomal or nucleosomal region (the upstream REST motif was not used to compute the SMF score) (n ≥ 2, mean ± SEM).

See also Figure S3.

Motif tiling reveals preference for binding non-nucleosomal DNA and nucleosome entry-exit sites

The position of the motif relative to the nucleosome has the potential to influence binding, with the dyad presumably being less accessible than the entry-exit sites.6,10,56 As our previous setting focused on two single, distinct motif positions, we sought to test whether TF binding is increasingly impeded as the motif approaches the dyad in vivo. For this, we selected five TFs from different functional classes (repressor, structural, and activator) and harboring different DBDs (zinc fingers, BEN, POU-HMG, and basic helix-loop-helix) and carried out a more detailed dissection. We tiled their motifs at seven positions across the nucleosome and performed SMF following genomic integration at the RMCE-locus (Figures 4A, 4B, S4A, and S4B). This revealed again that all tested TFs were able to bind in the non-nucleosomal region (tiles 1 and 2) and the first linker (tile 7); however, only CTCF and REST were able to fully engage with their motif and move the nucleosome in the majority of the cells at nucleosome-internal motif locations (tiles 3, 4, 5, and 6; Figure 4C). This was also true for BANP (although to a lesser extent), which appeared slightly sensitive to a motif positioned in the nucleosome (Figures 4B and 4C). For BHLHE40 and OCT4-SOX2, the magnitude of the TF footprint drastically decreased as the motif was placed more internally on the nucleosome, reaching a minimum around the dyad (tile 5) and increasing again toward the first linker (Figures 4C–4E). Notably, at tile 5, the average profile of SMF reads corresponding to nucleosomal occupied DNA differed slightly between the OCT4-SOX2 motif and the scrambled control (Figure 4E). It is possible that OCT4-SOX2 recognizes a partial motif at this position,10,11 which leads to distortion of nucleosomal DNA. However, no changes in the overall nucleosome footprint were observed, suggesting that this effect was not sufficient to cause nucleosome rearrangement. Finally, we wanted to test the possibility that these results depend on using the TF REST to phase the nucleosome. To do so, we repeated the entire experimental set, but now employing a CTCF motif to mediate nucleosome phasing. Even using this strong insulator, all tested TF motifs displayed footprints almost identical to the ones obtained with the REST-phased nucleosome, arguing that the sensitivity of the tested TFs is to the nucleosome position itself rather than to the TF employed to phase the nucleosome (Figures S4C–S4E). This is further supported by the fact that binding of all tested TFs is similarly strong in the first linker and the non-nucleosomal region next to REST or CTCF.Figure 4 Motif position along the nucleosome influences occupancy in a TF-specific fashion

(A) Scheme illustrating motif tiling throughout the REST-phased nucleosome construct.

(B) Average profiles of SMF data over the BANP motif across the seven motif positions illustrating that BANP can bind to all positions (n = 3, mean ± SEM is plotted as in Figure 1B).

(C) Quantification of the TF-bound fraction at each motif position in the REST-phased nucleosome construct for five different motifs (n ≥ 2, mean ± SEM).

(D and E) Average profiles of SMF data over the OCT4-SOX2 motif in a non-nucleosomal (D, tile 2) or nucleosomal (E, tile 5) tile compared with scrambled controls illustrating reduced binding at the dyad (n ≥ 2, mean ± SEM is plotted as in Figure 1B).

(F) Chromatin immunoprecipitation followed by quantitative PCR of OCT4 at a non-nucleosomal (tile 2) or nucleosomal (tile 5) motif position. ChIP enrichment is relative to an OCT4-bound endogenous region (n = 2, mean ± SEM). scr, scrambled motif.

See also Figure S4.

To test whether the observed footprints detected at nucleosomal positions indeed require presence of the motif-binding TF, we monitored the SMF traces at the tested locus upon depletion of individual factors. First, for BANP, this was accomplished by utilizing an inducible degron system that enables targeted depletion of this TF.36 This revealed that the footprint over the BANP motif is lost upon factor degradation and the displaced nucleosome returns to its native position (Figure S4F). In addition, we employed an induced neuronal differentiation system to similarly test the dependence for the REST motif, as REST is not expressed in neurons.36,57,58 This differentiation-specific protein depletion resulted in loss of footprint at the REST motif in the tested constructs (Figure S4G). This shows, for both BANP and REST, that the observed footprint and nucleosome displacement are strictly dependent on the presence of the respective protein and excludes the possibility that the observed effects on nucleosome position are due to a potential impact of the motif sequence itself, even for long motifs like that of REST.

We next wondered whether we could extend this observation to more TFs by utilizing the single-molecule data from the larger screen of nucleosome sensitivity (Figure 3B). In particular, we wanted to assess whether the decrease in SMF signal of the TF was due to the general presence of a nucleosome or whether the nucleosome position relative to the motif also affected TF engagement. We reasoned that if a TF favored a specific nucleosome location to access its motif, this would result in a lower representation of that nucleosome position on individual DNA molecules. To test this, we defined three nucleosome populations, based on the position of the nucleosome relative to the REST motif in the REST-phased nucleosome in the absence of a TF motif, and quantified those populations in the presence of each motif (Figure S4H). This revealed that the majority of TFs preferentially engage motifs that are only partially masked by the nucleosome, while CTCF and REST and, to a lesser extent, BANP and p53, are also able to bind and drive nucleosome rearrangement when their motif is fully masked.

Together, these results suggest that, in cells, TFs preferentially engage with motifs that reside in non-nucleosomal regions and at nucleosomal entry-exit sites, in line with previous in vitro data.6,10,56

Nucleosome obstruction of internal sites strongly reduces binding of OCT4

Because the SMF approach is factor agnostic, it does not provide information on potential simultaneous binding of different proteins to the same position and DNA molecule if accessibility is not altered. In our particular setting, we would not be able to distinguish whether a TF is not engaging with a motif on nucleosomal DNA or whether it is binding without causing nucleosome rearrangement. This seems particularly relevant because motifs embedded in nucleosomes have been proposed to be a favored binding substrate of OCT4-SOX2 in vivo,11,15 even though recent biochemical and structural data have challenged this view.10

To test whether this might explain the observed decrease in a TF footprint at internal nucleosome sites, we focused on the OCT4-SOX2 motif and tested actual binding of OCT4 to a motif located in the non-nucleosomal region and one close to the nucleosome dyad using a factor-specific approach. ChIP-qPCR revealed strong OCT4 binding to the non-nucleosomal region next to REST or CTCF, as observed by SMF, at levels comparable with a highly enriched endogenous locus (Figures 4F and S4G). By contrast, the OCT4 ChIP signal at the motif positioned over the nucleosomal dyad is close to background levels (Figures 4F and S4I). This result argues that OCT4-SOX2 struggle to bind nucleosomal embedded motifs and, further, that the slight distortion observed in the average SMF profile of the nucleosome when the OCT4-SOX2 motif is close to the dyad (Figure 4E) likely reflects weak and unstable engagement of the TFs with the motif. Overall, these findings provide additional evidence that even the binding of two established pioneer factors is impeded by the presence of a nucleosome on their motif in cells.

Nucleosome presence impairs OCT4 binding at endogenous OCT4-SOX2 motifs

The results from the RMCE-locus predict that OCT4 binding should occur preferentially in non-nucleosomal regions throughout the genome. To test this model, we performed ChIP-seq for OCT4 and measured binding at OCT4-SOX2 motifs in a 250-bp window around bound CTCF sites, where non-nucleosomal and nucleosomal regions can be identified by MNase-seq. Indeed, we observed an enrichment of OCT4 binding in non-nucleosomal regions compared with that over the first phased nucleosome (Figure 5A), strongly resembling OCT4 behavior at the RMCE-locus. We then wondered whether global changes in nucleosome phasing would affect OCT4 binding at these sites. We took advantage of the fact that CTCF binding and, in turn, CTCF-dependent phasing of nucleosomes are drastically reduced upon loss of the ISWI chromatin remodeler, which can be achieved in mESCs by genetic deletion of its catalytic subunit SNF2H.16,59 More specifically, we measured OCT4 binding at OCT4-SOX2 motifs genome wide in mESCs, in the presence and absence of SNF2H, and focused our analysis on those sites adjacent to bound CTCF motifs whose binding is lost upon SNF2H depletion. Strikingly, in the absence of SNF2H, we detected a significant loss of OCT4 binding at sites residing in the non-nucleosomal region adjacent to CTCF, coinciding with increased nucleosomal presence, as measured by MNase-seq (Figures 5B and 5C). Notably, we did not detect significant changes more distally, likely due to the smaller changes in nucleosome occupancy, which might not be sufficient to alter OCT4 binding. This shows that the sensitivity of OCT4 to nucleosomes observed at the RMCE-locus is also observed when perturbing nucleosome phasing at endogenous sites covering a range of motif scores, and this is also compatible with the fact that only 20% of strong OCT4-SOX2 motifs are actually bound throughout the genome.10Figure 5 Nucleosome presence restricts OCT4 binding in the genome

(A) Running average (k = 15) of OCT4 binding at OCT4-SOX2 motifs in the ±250-bp window around CTCF-bound sites in wild-type cells.

(B) Single-locus example of OCT4 binding and nucleosome positioning (MNase-seq) at an OCT4-SOX2 motif (red bar) next to a CTCF-bound motif (blue bar), before and after genetic deletion of SNF2H.

(C) Loss of OCT4 binding at OCT4-SOX2 motifs adjacent to bound CTCF sites that lose binding in SNF2H knockout cells. Each bar represents the average fold change of OCT4 binding over 15 OCT4-SOX2 motifs measured in four biological replicates. Error bars indicate SEM. Significant fold changes are highlighted in red (p < 0.05, permutation test, see STAR Methods). Average MNase fold change between SNF2H KO and WT at these sites is displayed on the top.

See also Figure S5.

Discussion

Here, we comprehensively surveyed the ability of TFs to engage with over a hundred singular motif occurrences at a defined genomic position, revealing that sensitivity to nucleosomes is not only factor specific but also a function of motif position. This study of most-known motifs with expressed TFs in mESCs has been enabled by a reductionist approach that quantifies TF recruitment as a function of nucleosome presence and phasing on individual DNA molecules at a coverage of thousands of molecules per tested sequence. It illustrates the value of varying sequences to decode gene regulation60 and provide direct in vivo evidence of the TF-specific inhibitory effect of nucleosomes on TF binding, in support of a scenario more complex than a binary distinction of TFs into pioneer versus non-pioneer factors.61 Instead, the ability to bind in a chromatinized genome can be framed in a broader context of chromatin sensitivity, which we show to be a continuous spectrum and highly TF specific.

Within the context of unphased nucleosomes, ∼75% of tested motifs do not lead to detectable footprints or nucleosome rearrangement, instead likely requiring multiple motifs and TF-binding combinations to generate stable binding. This is experimentally supported by our results combining the OCT4 with the SOX2 motif and the dependence of NRF1 on the REST motif. It suggests that most TFs cannot robustly engage with isolated chromatinized motif occurrences, which is reflected in the fact that most TFs only bind a small subsets of their motif occurrences in the genome.25,41,62 However, we cannot rule out that our experimental setup, like other methods for detecting binding, is limited in its ability to distinguish weak binding, complete absence of engagement, or general inability to cause footprints.35 Regardless, it seems noteworthy that our test sequence can be considered euchromatic as it has low nucleosome affinity, is CpG rich, predominantly DNA unmethylated, and harbors no feature of heterochromatin besides the presence of nucleosomes. It will be interesting to test whether engagement with these motifs changes upon alteration of the neighboring sequences, motif multimerization or combinations, and motif syntax, to explore rules of additive or combinatorial effects that likely account for the general openness of regulatory regions.35,63,64,65,66,67,68,69 While such experiments go beyond the scope of the current study, these questions can, in principle, be addressed in a highly controlled way with the presented experimental framework.

It seems notable that ∼25% of tested singular motifs are indeed able to recruit TFs, as detected by footprinting, while only a few TFs, including REST, CTCF, and BANP, create very strong footprints and open chromatin. This reflects the genuine ability of these TFs to engage with singular motif occurrences, as they also tend to occupy the majority of their genomic sites, as measured by ChIP-seq.36,46,47 Interestingly, these three factors are broadly expressed with no known function in driving cell fate in development, a characteristic previously suggested to be a feature of pioneer factors.70,71

Among all studied motifs, the amplitude of the footprint seems to be independent of several features, such as expression level, being broadly expressed versus tissue specific, DBD class, or presence of IDRs, with only motif length showing moderate predictive power. This indicates that although more potential contacts between the DBD and DNA can favor motif engagement, establishing stable binding in a chromatinized genome might require a combination of features. These seem to go beyond the contribution of the motif sequence and are likely TF specific, such as the dependence on cofactors (e.g., remodelers) that these TFs utilize61 or the presence of specific chromatin environments. A key next step will be to decode combinatorial activity, which we observe here for the OCT4 and SOX2 heterodimer TFs and that had already been suggested from genome-wide analyses based on SMF data.35 OCT4 and SOX2 appear not to engage on their own, but establish detectable binding when combined on the OCT4-SOX2 motif. Even then, the magnitude of their combined footprint is moderate, only causing nucleosome rearrangement in ∼15% of the cell population, which is in line with their moderate occupancy at even strong motif occurrences genome wide.10

The proposed model—that some TFs engage a nucleosomal substrate, potentially even preferred to naked DNA alone11,13,15,20,27,72—remained difficult to test in vivo, where nucleosomes are not often found in precise positions. In our experimental setup, the W601 sequence, which strongly positions nucleosomes in vitro, fails to do so upon insertion into mESCs. It is tempting to speculate that remodeler activity overrides any potential phasing, arguing that sequence-intrinsic positioning of nucleosomes plays only a minor role in the dynamic environment of the cell.73 Instead, we show that TF-mediated nucleosome positioning can generate a highly phased nucleosome. This enabled us to show experimentally that nucleosome-occupied motifs are less favorably bound by almost all TFs we tested, including OCT4-SOX2, with only REST, CTCF, and BANP displaying the ability to displace a nucleosome, even when the motif is placed close to the dyad. We further confirmed the preference of OCT4-SOX2 for binding to non-nucleosomal DNA or entry-exit sites with an orthologous approach that entailed reducing genome-wide CTCF-mediated nucleosomal phasing. This resulted in reduced OCT4 binding at OCT4-SOX2 motifs with increased nucleosome presence, mirroring OCT4 behavior at the RMCE-locus. We cannot formally exclude the possibility that a TF-phased nucleosome might not fully recapitulate other nucleosomal substrates. However, the similarity in binding proximal to the TF and in the first linker, using either REST or CTCF (two very different TFs) to phase, does not indicate a TF-specific effect. Moreover, this binding preference and response to nucleosome position at the RMCE-locus and genome wide is fully in line with previous in vitro binding and structural data.10,23 It suggests an important role for nucleosome position and mobility in governing OCT4-SOX2 binding in vivo, as is further evidenced by its dependence on SWI/SNF remodeling.16,17,22 Thus, it is tempting to speculate that further cooperativity with other TFs might contribute to the previously observed binding of these factors at sites of closed chromatin.11,15,74,75,76 Overall, our results confirm that nucleosomes have a general inhibitory role, with TFs engaging preferentially at non-nucleosomal and entry-exit sites. This is in line with previous in vitro data6,10 and also exposes factor-specific nucleosome sensitivity depending on motif position. Further understanding of the contribution of different genomic and epigenetic features on TF binding will require us to determine their individual and combinatorial effects in a controlled setting, ideally linking biochemical and structural information with reductionist approaches in the cellular context. This can be achieved, for example, by testing the effect of motif multiplicity, combinations, and mutations, as well as varying the surrounding sequence. In addition, targeted epigenetic editing,77,78 or removal of cofactors such as chromatin remodelers,17,18 can be implemented to delineate the contribution of epigenetic modifications and binding partners in defining chromatin sensitivity. This should lead to an even more integrated understanding of how specific sequence contexts and chromatin states guide TF-binding patterns throughout the genome.

Limitations of the study

Although the SMF approach utilized in this study allows the simultaneous survey of a number of TFs, it does not provide direct information about the identity of the factors the observed footprints are derived from. Unless confirmed by complementary approaches such as ChIP, or through the loss of a footprint upon factor-specific knockdown experiments, we cannot ascertain that the observed footprints are caused by the TF linked to the underlying motif through the literature. Moreover, the possibility cannot be excluded that the ability of a TF to create a footprint upon binding depends on the factor’s residence time or other unknown sequence, TF, or chromatin properties, leaving the possibility that we may not be able to detect all binding events.

STAR★Methods

Key resources table

REAGENT or RESOURCE	SOURCE	IDENTIFIER	
Antibodies	
	
Anti-Histone H3 (tri methyl K9) antibody - ChIP Grade	Abcam	Cat#ab8898; RRID:AB_306848	
Anti-Histone H3 (acetyl K27) antibody - ChIP Grade	Abcam	Cat#ab4729; RRID:AB_2118291	
Anti-NRF1 antibody [EPR5554(N)] - ChIP Grade	Abcam	Cat#ab175932; RRID:AB_2629496	
Oct-4A (C30A3C1) Rabbit mAb (ChIP Formulated)	Cell Signaling Technology	Cat #5677; RRID:AB_10547892	
	
Chemicals, peptides, and recombinant proteins	
	
Dulbecco’s Modified Eagle Medium (DMEM)	Gibco	Cat #11965092	
Gelatin from porcine skin	Sigma	Cat #G2625	
Fetal Bovine Serum	Gibco	Cat #A5256701	
Glutamax (100X)	Gibco	Cat #35050061	
MEM Non-essential aminoacids solution (100X)	Gibco	Cat #11140050	
2-mercaptoethanol	Sigma	Cat #M3148	
Leukemia inhibitory factor (LIF)	Produced in-house	N/A	
Lipofectamine 3000	Thermo Fisher Scientific	Cat #L3000001	
Mouse Embryonic Stem Cell Nucleofector Kit	Lonza	Cat #VPH-1001	
Ganciclovir	Sigma	Cat #G2536	
DMEM/F12 with Glutamax	Gibco	Cat #10565018	
50x B27 supplement without vitamin A	Gibco	Cat#12587010	
100X N2 supplement	Gibco	Cat#17502048	
Doxycycline	Takara	Cat#631311	
dTAG-13	Sigma	Cat#SML2601	
DPBS, no calcium, no magnesium	Gibco	Cat#14190144	
Trypsin-EDTA (0.25%)	Gibco	Cat#25200056	
GpC Methyltransferase (M.CviPI)	New England Biolabs	Cat#M0227	
S-adenosylmethionine (SAM)	New England Biolabs	Cat#86867-01-8	
Glycogen	Roche	Cat#10901393001	
AMPure XP Beads	Beckman Coulter	Cat#A63881	
Dynabeads Protein A for Immunoprecipitation	Thermo Fisher Scientific	Cat# 10001D	
KAPA HiFi Uracil+	Roche	Cat#KK2801	
PowerUp™ SYBR™ Green Master Mix for Qpcr	Applied Biosystem	Cat#A25742	
	
Critical commercial assays	
	
EZ DNA methylation-gold kit	Zymo	Cat#D5006	
NEBNext ChIP-seq Library Prep Master Mix Set	New England Biolabs	Cat#E6240	
NEBNext Ultra II DNA Library Prep Kit	New England Biolabs	Cat#E7645	
	
Deposited data	
	
ChIP-seq CTCF in mESCs WT	Barisic et al.16	GEO: GSM3058327; GSM3058329	
ChIP-seq CTCF in mESCs Snf2h KO	Barisic et al.16	GEO: GSM3058328; GSM3058330	
MNase-seq in mESCs WT	Barisic et al.16	GEO: GSM3058339; GSM3058340	
MNase-seq in mESCs Snf2h KO	Barisic et al.16	GEO: GSM3058341; GSM3058342	
ChIP-seq CTCF in mESCs WT	Chronis et al.79	GEO: GSM747535	
ChIP-seq OCT4-SOX2 in mESCs WT	Stadler et al.44	GEO: GSM2417142	
SMF of TF motifs in different chromatin contexts in mESCs WT	This study	GEO: GSE242067	
	
Experimental models: Cell lines	
	
Mouse embryonic stem cells - TC-1 line	Lienert et al.37	N/A	
Mouse embryonic stem cells - TC-1 line - BANP-dTAG	Grand et al.36	N/A	
Mouse embryonic stem cells - TC-1 line - NGN2 cassette	Grand et al.36	N/A	
	
Oligonucleotides	
	
RMCE locus primer forward (for bisulfite-converted DNA): TGATGATGAGAAGAAGAAATGATG	This study	N/A	
RMCE locus primer reverse (for bisulfite-converted DNA): CCACCACAAACAATACCAATAATAA	This study	N/A	
OCT4 RMCE ChIP-qPCR primer forward: TTACCGCCTTTGAGTGAGCT	This study	N/A	
OCT4 RMCE ChIP-qPCR primer reverse: GGTTACGCGGCTGGTCTAT	This study	N/A	
OCT4 endogenous ChIP-qPCR primer forward: TGATTGACTACAGAATCGGAGC	This study	N/A	
OCT4 endogenous ChIP-qPCR primer reverse: GCTATTCTTATGCGCCTCTGC	This study	N/A	
NRF1 RMCE ChIP-qPCR primer forward: TCACACCGCATAGACCAGC	This study	N/A	
NRF1 RMCE ChIP-qPCR primer reverse: CAACCGTAACCGATTTTGCCA	This study	N/A	
NRF1 endogenous ChIP-qPCR primer forward: GTTAACACTCTCTCCCCGCC	This study	N/A	
NRF1 endogenous ChIP-qPCR primer reverse: CTCGGAGAGGGGTTGCTAGA	This study	N/A	
	
Recombinant DNA	
	
Plasmids for RMCE integration carrying TF motifs inserted in a low nucleosome affinity sequence	This study	N/A	
Plasmid for RMCE integration carrying the Widom 601 sequence	This study	N/A	
	
Software and algorithms	
	
R studio 4.4.0 or higher	N/A	https://www.r-project.org	
Bioconductor QuasR	Gaidatzis et al.80	https://doi.org/10.18129/B9.bioc.QuasR	
Bioconductor ComplexHeatmap	Gu et al.81	https://doi.org/10.18129/B9.bioc.ComplexHeatmap	
Trimmomatic	Bolger et al.82	http://www.usadellab.org/cms/?page=trimmomatic	
Bowtie	Langmead et al.83	https://bowtie-bio.sourceforge.net/index.shtml	
Bioconductor Gviz	Hahne et al.84	https://doi.org/10.18129/B9.bioc.Gviz	

Resource availability

Lead contact

Further information and requests for resources and reagents should be directed to and will be fulfilled by the lead contact, Dirk Schübeler (dirk@fmi.ch).

Materials availability

Plasmids and cell lines generated by this study are available from the FMI under a material transfer agreement.

Data and code availability

• Next-generation sequencing data generated in this study are deposited at Gene Expression Omnibus (GEO) with accession number GSE242067. CTCF ChIP-seq data in mESCs from Barisic et al.16 are available at GEO with accession numbers: GSM3058327 and GSM3058329 (WT replicates) - GSM3058328 and GSM3058330 (Snf2h KO replicates). Mnase-seq data in mESCs from Barisic et al.16 are available at GEO with accession numbers: GSM3058339 and GSM3058340 (WT replicates) - GSM3058341 and GSM3058342 (Snf2h KO replicates). CTCF ChIP-seq data in mESCs from Chronis et al.79 are available at GEO with accession number: GSM747535 (WT replicate). OCT4-SOX2 ChIP-seq data in mESCs from Stadler et al.44 are available at GEO with accession number: GSM2417142 (WT replicate).

• This paper does not report original code. The analyses were performed using base R functions and publicly available packages. All analysis steps and packages are detailed in the STAR Methods section.

• Any additional information required to reanalyze the data reported in this paper is available from the lead contact upon request.

Experimental model and study participant details

Cell culture

Mouse ES cells, TC-1 line, carrying an RMCE cassette at a previously defined genomic locus37 were used for all experiments and cultured as described in Mohn et al.85 Briefly, plates were coated with 0.2% gelatin and cells were maintained in Dulbecco’s Modified Eagle Medium (DMEM), supplemented with 15% Fetal Bovine Serum, 1X Glutamax and 1X Non-essential amino acids, 2-mercaptoethanol (1:100,000) and leukemia inhibitory factor (LIF). No synchronization was performed before any of the experiments.

Method details

RMCE insertion

RMCE insertion was performed as previously described.37,86 In brief, for individual construct insertion, 250,000 cells were transfected with 1 μg of L1-insert-1L vector and 500 ng of pIC-Cre using Lipofectamine 3000 Reagent following manufacturer protocol. For pooled insertions, 4 million cells were electroporated (Amaxa Nucleofection, Lonza) with 25 μg of L1-insert-1L vector pool (up to 10 different constructs) and 15 μg of pIC-Cre. In both cases, negative selection with Ganciclovir at a final concentration of 3 μM was carried out starting 2 days after transfection for a total of 10 days.

Neuronal differentiation

Neuronal differentiation of mESCs was induced as described previously.36 In brief, mESCs harboring a Dox-inducible NGN2 expression cassette were plated on gelatine-coated dishes in proliferation medium (DMEM/F12 with Glutamax, 1× B27 supplement without vitamin A, 1× N2 supplement) supplemented with 1 μg ml−1 doxycycline and cultured for 3 days.

Targeted depletion of BANP

Targeted depletion of BANP was achieved as previously described.36 Briefly, BANP degron-tagged mESCs were seeded on gelatin-coated dishes and treated with dTAG13 (500 nM) for 24 h to degrade BANP.

Selection of transcription factors motifs and generation of constructs

Transcription factors expressed in mESCs were ranked based on their mean RPKM value according to Domcke et al.87 and a threshold of log2 RPKM ≥ 2 was applied to ensure robust expression. Motifs were taken from JASPAR database88 by selecting the maximum PWM, followed by the addition of GCT and TGCA sequences at their 5′ and 3′, respectively, to maximize the possibility to detect a footprint even for factors that do not have a GpC in their motif. For BANP, the motif was not available in JASPAR, so the motif used in36 was selected. For the constructs where REST and CTCF are used to position nucleosomes, two slightly different top motifs were chosen to avoid repeats in the constructs with two REST or CTCF motifs. Finally, all motifs were expanded to 28bp in length by adding random bases on either end. Scrambled motifs were obtained by randomly shuffling the motif bases while maintaining any CpG dinucleotides. The final constructs were generated by placing the 28mers at the desired position inside a previously described, in-vitro derived sequence, with low nucleosome affinity (LNA), high CpG content (O/E 1.34), which is not part of any known regulatory region.49,89 For RMCE, the constructs were finally cloned into a plasmid using a multiple cloning site flanked by two inverted L1 Lox sites.

Single-molecule footprinting

Single-molecule footprinting was performed as previously described34,36 with some modifications. After Ganciclovir selection, 250,000 cells were collected, washed once with PBS and incubated 10 min on ice in 1ml of ice-cold lysis buffer (10 mM Tris pH 7.5, 10 mM NaCl, 3 mM MgCl2, 0.1 mM EDTA and 0.5% NP-40) to extract nuclei. After centrifugation (800 g, 5 min, 4ºC), nuclei were washed once with 250ul of ice-cold wash buffer (10 mM Tris pH 7.5, 10 mM NaCl, 3 mM MgCl2 and 0.1 mM EDTA) and centrifuged again (800g, 5min, 4ºC). Samples were then resuspended in 100 ul of 1x M.CviPI buffer, mixed with 200ul of GpC methyltransferase reaction mix on ice (1x M.CviPI buffer, 1mM SAM, 300nM sucrose, 200U M.CviPI: M0227L, NEB) and then incubated at 37ºC for 15 min. Reaction was stopped by adding 300ul of Stop Solution (20 mM Tris-HCl pH 8.0, 600 mM NaCl, 1% SDS, 10 mM EDTA) pre-warmed at 37ºC. To remove proteins and RNA, samples were treated first with proteinase K (200 ug/ml) at 65ºC for 16h and then with 10ul of RNAse A (10 mg/ml) for 30 min at 37ºC. Finally, genomic DNA was extracted with phenol-chloroform purification and isopropanol precipitation as follows. Samples were mixed with 600 ul of a 1:1 phenol-chloroform solution, vortexed for 1 min at maximum speed and centrifuged at 10,000g for 10 min. The supernatant was then transferred to new tubes and mixed with 600ul of chloroform, vortexed and centrifuged again as before. The supernatant was transferred to new tubes and mixed with 600ul of isopropanol and 1ul glycogen and incubated 10 min at RT before centrifugation (max speed, 1h, 4ºC). The pellet was then washed with 500 ul of ice-cold 70% ethanol and centrifuged (max speed, 15 min, 4ºC). The supernatant was discarded and the pellet was dried for 15 min at 37ºC and resuspended in 22 ul ddH2O overnight at RT. NanoDrop was used for quantification and 2 μg of DNA were converted using the EZ DNA methylation-gold kit (Zymo).

DNA amplification and bisulfite sequencing

Target amplicons were amplified from bis-converted DNA using KAPA HiFi Uracil+ with the following program: 95ºC, 4min; [98ºC, 20s; 60ºC, 15s, 72ºC, 20s]x35; 75ºC, 5min; 4ºC, hold. DNA was then purified with AMPure XP Beads (0.6X ratio) and used as input for library preparation. Sequence libraries were prepared with the NEBNext ChIP-seq Library Prep Master Mix Set, doing 8 cycles of amplification, and loaded on an Illumina MiSeq (500 bp paired-end).

ChIP-qPCR and ChIP-seq

ChIP samples were prepared as previously described16 with following modifications: (1) Diagenode Bioruptor Pico was used to sonicate chromatin with 20 cycles, 20sec ON and 40sec OFF, (2) protein A magnetic Dynabeads Magnetic beads were used. For immunoprecipitation of H3K27ac, 25 μg of chromatin were combined with 2 μg of H3K27ac antibody. For immunoprecipitation of H3K9me3, 25 μg of chromatin were combined with 2 μg of H3K9me3 antibody. For immunoprecipitation of OCT4, 70 μg of chromatin was combined with 5 μg of Oct-4a antibody. For immunoprecipitation of NRF1, 70 μg of chromatin was combined with 5 μL of NRF1 antibody.

For ChIP-qPCR, input and immunoprecipitated materials were resuspended in 40ul and 1ul was used as input for qPCR. Amplification was carried out in 20ul reaction with the PowerUp SYBR Green Master Mix using the following program: 95ºC, 10min; [98ºC, 15s; 60ºC, 1min]x40, ramping 1.6 ºC/s. To measure H3K27ac and H3K9me3 enrichment at the unphased nucleosome construct and at a negative region, a primer pair targeting the center of the construct and a primer pair targeting an intergenic region (Mm10 genome build: chr12:47,899,674-47,899,815) were used. Enrichments were then compared against their matched positive region, namely a primer pair targeting a H3K27ac-enriched region (Mm10 genome build: chr6:125,165,470-125,165,602) or a primer pair targeting a H3K9me3-enriched IAP repeat element. To measure OCT4 enrichment at tile 2 and 5 in the REST- and CTCF-phased nucleosome construct, a primer pair encompassing the region from tile 2 to tile 5 was used and compared against a primer pair targeting a single genomic OCT4-bound OCT4-SOX2 motif (Mm10 genome build: chr8:92,740,685-92,746,111). To measure NRF1 enrichment at the unphased or REST-phased nucleosome construct, a primer pair encompassing the region from tile 2 to tile 5 was used and compared against a primer pair targeting a single genomic NRF1-bound site (Mm10 genome build: chr8:70,865,657-70,866,052).

For ChIP-seq, input and immunoprecipitated materials underwent library preparation (NEBNext Ultra II DNA Library Prep Kit) using 12 cycles of PCR amplification. Libraries were sequenced on the NovaSeq6000 paired-end 2x50bp.

Quantification and statistical analysis

TF Features

Tissue specificity

Gene expression data for the 107 surveyed TFs in 17 different mouse tissues were obtained from90 and for each tissue a threshold of log2 RPKM ≥ 1 was used to define expressed TFs. Factors were then grouped based on their presence in a certain proportion of the tissues: (0%;25%], (25%;50%], (50%;75%], (75%;100%], with the first and last group representing tissue-specific and ubiquitous distribution respectively. IDR fraction: To predict IDRs for the 107 surveyed TFs, the amino acid sequence of each TF was obtained from https://www.uniprot.org/ and an IDR score was computed using the idpr Bioconductor package version 1.8.091 with standard parameters.

Generation of average profiles and single-molecule analysis

Read mapping to the relative constructs was carried out using the Bioconductor package QuasR with default settings for bisulfite-converted templates80 and after adaptor trimming with Trimmomatic version 0.32.82 Methylation of Cs in the CpG and GpC contexts was then quantified using qMeth, excluding GCG and GCCG contexts, where endogenous and exogenous methylation cannot be distinguished, and Cs with coverage <10. Such GpCs and CpGs were defined as GCH and HCG, respectively. In all figures, the SMF data are shown as 1 minus methylation on the Y axis and bp position on the X axis. HCG raw methylation is shown on the right Y axis in Figures S1A and S1H. TF motifs are shown in blue (nucleosome-positioning TF) or red (test TF). Heatmaps were generated using the function Heatmap from the ComplexHeatmap package version 2.14.0.81 For heatmaps in Figures 3A, 3B, S3G, and S3K, SMF signal is normalized as follows: SMF value of each GCH is divided by the max SMF value reached within the upstream REST motif and multiplied by 100. This assumes that the upstream REST motif has stable occupancy across constructs and uses this value to internally normalize each sample and facilitate comparison. Single-molecule analysis was performed using QuasR in combination with custom functions to cluster reads according to their GCH methylation pattern, with 0 and 1 reflecting unmethylated and methylated status respectively. For each construct, reads from all replicates were pooled and filtered by GCH coverage. Then, 1000 reads were randomly sampled and used for downstream analyses. If less than 1000 reads were available after filtering, all remaining reads were used. To define single-molecule populations on the Widom 601 and REST-phased nucleosome constructs, reads were clustered using kmeans with k = 5, after assessing the optimal k using the Elbow Method (Figure S2B and S2C). To match single-molecule reads to the three nucleosomal clusters of the REST-phased nucleosome construct (Figure S4D), the euclidean distance between each read and the cluster centroids was calculated. Then, reads were assigned to the closest centroid.

For all the other single-molecule analyses, three clusters were defined as follows. Unbound: >80% of GCH must be methylated; TF bound: average of GCH methylation at the motif must be equal to zero after rounding (<0.5). This ensures that the motif is occupied. Then, at least one of the first two GCHs upstream of the motif and at least one of the first two GCHs downstream of the motif must be methylated. This ensures presence of non-nucleosomal DNA around the bound TF and accounts for different footprint sizes. When resolution or coverage was limited (tiles 1, 6 and 7), a subset of non-nucleosomal GCHs was selected for each condition and maintained across constructs; Nucleosome: all reads that do not fall in the previous two clusters.

SMF score and motif clustering

To identify footprints over each motif in the unphased nucleosome context, we selected different GCH positions to calculate a SMF score (Figures S1D and S3E–S3H). The GCH coordinates range from -2 to +2, with position 0 (•) being the centermost GCH of the motif. Positions -1 and +1 (▲) represent the GCH adjacent to the motif (upstream and downstream respectively) and positions -2 and +2 (◼) represent the GCHs upstream and downstream of those respectively. When no GCH was present inside the motif, position 0 coincided with position +1.

To account for differences in TF footprint size, we calculated two scores for each motif and then took the maximum. Score 1 was obtained by calculating the difference in SMF signal between position 0 and positions -1 and +1, respectively, and summing them up. This score makes the assumption of a narrow TF footprint where occupancy is high over the central part of the motif and drops already at either end. Score 2 was obtained with the same procedure, but using positions -2 and +2 instead to calculate the difference. This score makes the assumption of a wider TF footprint where occupancy is high over the entire motif and drops at the neighboring GCHs outside of the motif.

The maximum score was calculated for each replicate and the replicates mean score of the “no motif” construct was used as baseline. A strong SMF score was defined as at least two-fold higher compared to the SMF score of the “no motif” construct. Scores greater than the “no motif” baseline, but less than the two-fold threshold, were defined as weak. Scores below the “no motif” baseline were defined as negative.

To identify footprints over each motif in the two REST-phased nucleosome contexts, the same procedure was applied, but taking into account only positions 0, +1 and +2, as, in these contexts, all constructs have similar SMF signal over the upstream GCHs due to the presence of REST. Score 2 was not calculated for the “no motif” construct due to insufficient GCH content.

Reference genome and annotation

Sequencing reads were aligned to the mouse GRCm38/mm10 genome assembly (GCA_000001635.2, Dec 2011), obtained from ftp://hgdownload.soe.ucsc.edu/.goldenPath/mm10/. Gene annotation was obtained from the Bioconductor package TxDb.Mmusculus.UCSC.mm10.knownGene (version 3.10.0). CTCF and OCT4-SOX2 predicted motifs were obtained as previously described92 using publicly available datasets (GSM74753579 and GSM241714244 respectively).

ChIP-seq and MNase-seq data analysis

To explore binding of OCT4 in response to CTCF and nucleosome occupancy, publicly available CTCF ChIP-seq (GSM3058327, GSM3058328, GSM3058329, GSM305833016) and MNase-seq (GSM3058339, GSM3058340, GSM3058341, GSM305834216) datasets in Snf2h-WT and Snf2h-KO mESCs were used. To align ChIP-seq and MNase-seq reads to the genome, the qAlign function from the QuasR Bioconductor package version 1.32.0,80 which in internally uses bowtie,83 was used with default parameters, except for paired=“fr” which was used for NovaSeq paired-end reads. Only reads with a unique match to the genome were reported as alignments. First, quality control was performed for each individual replicate and, then, replicates were merged for the final analyses, except in Figure 5C. Samples were normalized using a scaling factor based on library size, which was obtained as the ratio of the smallest library size and the library size of each sample. Signal was calculated by counting reads in the regions of interest for each sample and log2 transformation after addition of a pseudocount of 8.

For all analyses involving counts in or relative to regions (QuasR function qCount and qProfile), paired-end reads were shifted towards the 3′-end using the “halfInsert” parameter, while single-end reads were shifted towards the 3′-end by 70 bp, which is the shift that generates the largest overlap between read density profiles relative to predicted motifs. Metaplots around TF binding sites were normalized by multiplying each profile by the respective scaling factor and then dividing it by the total number of regions under consideration. Smoothing was then performed by summing counts in a 51-bp running window. For single locus plots, profiles were obtained with qProfile, scaled by the respective scaling factor and plotted using the Gviz package version 1.42.184

For analyses of OCT4 binding near CTCF sites, OCT4-SOX2 and CTCF motifs were selected as follows: CTCF ChIP-seq was used to identify all CTCF motifs that are bound in SNF2H WT (log2 enr > 1.5) and those that are bound in SNF2H WT and lose binding in SNF2H KO (log2 WT enr > 1.5 and log2 KO-WT fold change ≤ -1). OCT4-SOX2 motifs were selected to be within 250bp distance from distal CTCF sites and with a motif score ≥ 16. Distal CTCF sites were selected to be outside of promoter regions, defined as a 2 kb window around known TSSs. To take into account potential differences in OCT4 ChIP efficiency between SNF2H WT and SNF2H KO samples, ChIP signal was further normalized using read counts at OCT4-SOX2 motifs (score ≥ 16) residing in promoters and not overlapping with SNF2H-dependent CTCF motifs (≥ 2 kb distance). The counts were divided by the smallest and used as additional scaling factor. This normalization is based on the observation that chromatin accessibility and nucleosome positioning is not affected around TSSs upon SNF2H deletion.16

Permutation test for OCT4 ChIP enrichment near CTCF sites

To test for significance of changes in OCT4 binding as a function of distance from CTCF, and thus of nucleosome positioning, a permutation test was conducted under the null hypothesis that any effect observed is not dependent on motif position relative to CTCF and the nucleosome. For each replicate, OCT4 log2 read counts at OCT4-SOX2 motifs nearby CTCF sites (see previous section for motif selection) were ordered based on position relative to CTCF from nearest to farthest. Positions were then randomly permuted and the rolling mean (n = 15) was computed for the new permuted dataset. Finally, replicates were averaged and used to calculate the change in OCT4 binding between SNF2H KO and SNF2H WT conditions at each mean position. This process, which was repeated 10,000 times, is exactly the same as for the true unpermuted data shown in Figure 5C except for the initial random permutation of positions. Empirical (one-sided) p values for each mean position were calculated as the fraction of permuted datasets with a fold-changed at least as small as the observed fold-change, i.e. P(permΔ ≤ obsΔ), where permΔ is the binding change derived from the permuted datasets and obsΔ is the binding change derived from the real data (mean of replicates).

Supplemental information

Document S1. Figures S1–S5 and Table S1

Document S2. Article plus supplemental information

Acknowledgments

We thank members of the D.S. laboratory for their critical feedback on the study and manuscript, Arnaud Krebs and Nicolas Thomä for project discussions and feedback on the manuscript, and Sebastien Smallwood and the functional genomics platform of the FMI for next-generation sequencing support. D.S. acknowledges support from the Novartis Research Foundation, the Swiss National Science Foundation (310030B_176394 to D.S.), and the European Research Council under the European Union’s (EU) Horizon 2020 research and innovation programme grant agreements (ReadMe-667951 and DNAccess-884664 ). R.S.G. acknowledges Long-Term Fellowships from EMBO and from the EU Horizon 2020 Research and Innovation Program under the Marie Skłodowska-Curie grant (705354 ). The funders had no role in study design, data collection and analysis, decision to publish, or preparation of the manuscript.

Author contributions

R.S.G., M.P., and D.S. conceived and planned the experiments. R.S.G. and M.P. performed the majority of the experiments and data analysis. L. Baumgartner performed and analyzed the SMF and ChIP-qPCR for NRF1 and the SMF for the BANP and REST depletion. L. Burger assisted with computational data analysis. L.H. assisted with ChIP-qPCR and cell-line maintenance. D.S. supervised the project. R.S.G., M.P., and D.S. interpreted the results and wrote the manuscript.

Declaration of interests

The authors declare no competing interests.

Supplemental information can be found online at https://doi.org/10.1016/j.molcel.2024.08.009.
==== Refs
References

1 D’haeseleer P. What are DNA sequence motifs? Nat. Biotechnol. 24 2006 423 425 10.1038/NBT0406-423 16601727
2 Inukai S. Kock K.H. Bulyk M.L. Transcription factor–DNA binding: beyond binding site motifs Curr. Opin. Genet. Dev. 43 2017 110 119 10.1016/J.GDE.2017.02.007 28359978
3 Li E. Beard C. Jaenisch R. Role for DNA methylation in genomic imprinting Nature 366 1993 362 365 10.1038/366362a0 8247133
4 Padeken J. Methot S.P. Gasser S.M. Establishment of H3K9-methylated heterochromatin and its functions in tissue differentiation and maintenance Nat. Rev. Mol. Cell Biol. 23 2022 623 640 10.1038/s41580-022-00483-w 35562425
5 Yin Y. Morgunova E. Jolma A. Kaasinen E. Sahu B. Khund-Sayeed S. Das P.K. Kivioja T. Dave K. Zhong F. Impact of cytosine methylation on DNA binding specificities of human transcription factors Science 356 2017 eaaj2239 10.1126/SCIENCE.AAJ2239
6 Zhu F. Farnung L. Kaasinen E. Sahu B. Yin Y. Wei B. Dodonova S.O. Nitta K.R. Morgunova E. Taipale M. The interaction landscape between transcription factors and the nucleosome Nature 562 2018 76 81 10.1038/S41586-018-0549-5 30250250
7 Kornberg R.D. Chromatin structure: A repeating unit of histones and DNA Science 184 1974 868 871 10.1126/SCIENCE.184.4139.868 4825889
8 Michael A.K. Thomä N.H. Reading the chromatinized genome Cell 184 2021 3599 3611 10.1016/J.CELL.2021.05.029 34146479
9 Dodonova S.O. Zhu F. Dienemann C. Taipale J. Cramer P. Nucleosome-bound SOX2 and SOX11 structures elucidate pioneer factor function Nature 580 2020 669 672 10.1038/s41586-020-2195-y 32350470
10 Michael A.K. Grand R.S. Isbel L. Cavadini S. Kozicka Z. Kempf G. Bunker R.D. Schenk A.D. Graff-Meyer A. Pathare G.R. Mechanisms of OCT4-SOX2 motif readout on nucleosomes Science 368 2020 1460 1465 10.1126/SCIENCE.ABB0074 32327602
11 Soufi A. Garcia M.F. Jaroszewicz A. Osman N. Pellegrini M. Zaret K.S. Pioneer Transcription Factors Target Partial DNA Motifs on Nucleosomes to Initiate Reprogramming Cell 161 2015 555 568 10.1016/J.CELL.2015.03.017 25892221
12 Cirillo L.A. McPherson C.E. Bossard P. Stevens K. Cherian S. Shim E.Y. Clark K.L. Burley S.K. Zaret K.S. Binding of the winged-helix transcription factor HNF3 to a linker histone site on the nucleosome EMBO J. 17 1998 244 254 10.1093/EMBOJ/17.1.244 9427758
13 Cirillo L.A. Lin F.R. Cuesta I. Friedman D. Jarnik M. Zaret K.S. Opening of compacted chromatin by early developmental transcription factors HNF3 (FoxA) and GATA-4 Mol. Cell 9 2002 279 289 10.1016/S1097-2765(02)00459-8 11864602
14 Fernandez Garcia M. Moore C.D. Schulz K.N. Alberto O. Donague G. Harrison M.M. Zhu H. Zaret K.S. Structural Features of Transcription Factors Associating with Nucleosome Binding Mol. Cell 75 2019 921 932.e6 10.1016/J.MOLCEL.2019.06.009 31303471
15 Soufi A. Donahue G. Zaret K.S. Facilitators and Impediments of the Pluripotency Reprogramming Factors’ Initial Engagement with the Genome Cell 151 2012 994 1004 10.1016/J.CELL.2012.09.045 23159369
16 Barisic D. Stadler M.B. Iurlaro M. Schübeler D. Mammalian ISWI and SWI/SNF selectively mediate binding of distinct transcription factors Nature 569 2019 136 140 10.1038/S41586-019-1115-5 30996347
17 Iurlaro M. Stadler M.B. Masoni F. Jagani Z. Galli G.G. Schübeler D. Mammalian SWI/SNF continuously restores local accessibility to chromatin Nat. Genet. 53 2021 279 287 10.1038/s41588-020-00768-w 33558757
18 Schick S. Grosche S. Kohl K.E. Drpic D. Jaeger M.G. Marella N.C. Imrichova H. Lin J.G. Hofstätter G. Schuster M. Acute BAF perturbation causes immediate changes in chromatin accessibility Nat. Genet. 53 2021 269 278 10.1038/s41588-021-00777-3 33558760
19 Hsu H.T. Chen H.M. Yang Z. Wang J. Lee N.K. Burger A. Zaret K. Liu T. Levine E. Mango S.E. TRANSCRIPTION. Recruitment of RNA polymerase II by the pioneer transcription factor PHA-4 Science 348 2015 1372 1376 10.1126/SCIENCE.AAB1223 26089518
20 McDaniel S.L. Gibson T.J. Schulz K.N. Fernandez Garcia M. Nevil M. Jain S.U. Lewis P.W. Zaret K.S. Harrison M.M. Continued Activity of the Pioneer Factor Zelda Is Required to Drive Zygotic Genome Activation Mol. Cell 74 2019 185 195.e4 10.1016/J.MOLCEL.2019.01.014 30797686
21 Perlmann T. Wrange O. Specific glucocorticoid receptor binding to DNA reconstituted in a nucleosome EMBO J. 7 1988 3073 3079 10.1002/J.1460-2075.1988.TB03172.X 2846275
22 King H.W. Klose R.J. The pioneer factor OCT4 requires the chromatin remodeller BRG1 to support gene regulatory element function in mouse embryonic stem cells eLife 6 2017 10.7554/ELIFE.22631
23 Sinha K.K. Bilokapic S. Du Y. Malik D. Halic M. Histone modifications regulate pioneer transcription factor cooperativity Nature. 619 2023 378 384 10.1038/s41586-023-06112-6 37225990
24 Clark K.L. Halay E.D. Lai E. Burley S.K. Co-crystal structure of the HNF-3/fork head DNA-recognition motif resembles histone H5 Nature 364 1993 412 420 10.1038/364412A0 8332212
25 Lupien M. Eeckhoute J. Meyer C.A. Wang Q. Zhang Y. Li W. Carroll J.S. Liu X.S. Brown M. FoxA1 translates epigenetic signatures into enhancer-driven lineage-specific transcription Cell 132 2008 958 970 10.1016/J.CELL.2008.01.018 18358809
26 Hansen J.L. Loell K.J. Cohen B.A. A test of the pioneer factor hypothesis using ectopic liver gene activation eLife 11 2022 10.7554/ELIFE.73358
27 Van Oevelen C. Collombet S. Vicent G. Hoogenkamp M. Lepoivre C. Badeaux A. Bussmann L. Sardina J.L. Thieffry D. Beato M. C/EBPα Activates Pre-existing and De Novo Macrophage Enhancers during Induced Pre-B Cell Transdifferentiation and Myelopoiesis Stem Cell Rep. 5 2015 232 247 10.1016/J.STEMCR.2015.06.007
28 Sherwood R.I. Hashimoto T. O’Donnell C.W. Lewis S. Barkal A.A. Van Hoff J.P. Karun V. Jaakkola T. Gifford D.K. Discovery of directional and nondirectional pioneer transcription factors by modeling DNase profile magnitude and shape Nat. Biotechnol. 32 2014 171 178 10.1038/nbt.2798 24441470
29 Wapinski O.L. Vierbuchen T. Qu K. Lee Q.Y. Chanda S. Fuentes D.R. Giresi P.G. Ng Y.H. Marro S. Neff N.F. Hierarchical Mechanisms for Direct Reprogramming of Fibroblasts to Neurons Cell 155 2013 621 635 10.1016/J.CELL.2013.09.028 24243019
30 Moorman C. Sun L.V. Wang J. De Wit E. Talhout W. Ward L.D. Greil F. Lu X.J. White K.P. Bussemaker H.J. van Steensel B. Hotspots of transcription factor colocalization in the genome of Drosophila melanogaster Proc. Natl. Acad. Sci. USA 103 2006 12027 12032 10.1073/PNAS.0605003103 16880385
31 Li X.Y. Thomas S. Sabo P.J. Eisen M.B. Stamatoyannopoulos J.A. Biggin M.D. The role of chromatin accessibility in directing the widespread, overlapping patterns of Drosophila transcription factor binding Genome Biol. 12 2011 R34 10.1186/GB-2011-12-4-R34
32 ENCODE Project Consortium An integrated encyclopedia of DNA elements in the human genome Nature.. 489 2012 57 74 10.1038/nature11247 22955616
33 Adams D. Altucci L. Antonarakis S.E. Ballesteros J. Beck S. Bird A. Bock C. Boehm B. Campo E. Caricasole A. BLUEPRINT to decode the epigenetic signature written in blood Nat. Biotechnol. 30 2012 224 226 10.1038/NBT.2153 22398613
34 Kelly T.K. Liu Y. Lay F.D. Liang G. Berman B.P. Jones P.A. Genome-wide mapping of nucleosome positioning and DNA methylation within individual DNA molecules Genome Res. 22 2012 2497 2506 10.1101/GR.143008.112 22960375
35 Sönmezer C. Kleinendorst R. Imanci D. Barzaghi G. Villacorta L. Schübeler D. Benes V. Molina N. Krebs A.R. Molecular Co-occupancy Identifies Transcription Factor Binding Cooperativity In Vivo Mol. Cell 81 2021 255 267.e6 10.1016/J.MOLCEL.2020.11.015 33290745
36 Grand R.S. Burger L. Gräwe C. Michael A.K. Isbel L. Hess D. Hoerner L. Iesmantavicius V. Durdu S. Pregnolato M. BANP opens chromatin and activates CpG-island-regulated genes Nature 596 2021 133 137 10.1038/S41586-021-03689-8 34234345
37 Lienert F. Wirbelauer C. Som I. Dean A. Mohn F. Schübeler D. Identification of genetic elements that autonomously determine DNA methylation states Nat. Genet. 43 2011 1091 1097 10.1038/NG.946 21964573
38 Krebs A.R. Imanci D. Hoerner L. Gaidatzis D. Burger L. Schübeler D. Genome-wide Single-Molecule Footprinting Reveals High RNA Polymerase II Turnover at Paused Promoters Mol. Cell 67 2017 411 422.e4 10.1016/J.MOLCEL.2017.06.027 28735898
39 Pardo C.E. Darst R.P. Nabilsi N.H. Delmas A.L. Kladde M.P. Simultaneous Single-Molecule Mapping of Protein-DNA Interactions and DNA Methylation by MAPit CP Molecular Biology 95 2011 21 22 10.1002/0471142727.MB2122S95
40 Krebs A.R. Dessus-Babus S. Burger L. Schübeler D. High-throughput engineering of a mammalian genome reveals building principles of methylation states at CG rich regions eLife 3 2014 e04094 10.7554/ELIFE.04094
41 Slattery M. Zhou T. Yang L. Dantas Machado A.C. Gordân R. Rohs R. Absence of a simple code: how transcription factors read the genome Trends Biochem. Sci. 39 2014 381 399 10.1016/J.TIBS.2014.07.002 25129887
42 Swinstead E.E. Miranda T.B. Paakinaho V. Baek S. Goldstein I. Hawkins M. Karpova T.S. Ball D. Mazza D. Lavis L.D. Steroid Receptors Reprogram FoxA1 Occupancy through Dynamic Chromatin Transitions Cell 165 2016 593 605 10.1016/J.CELL.2016.02.067 27062924
43 Levo M. Avnit-Sagi T. Lotan-Pompan M. Kalma Y. Weinberger A. Yakhini Z. Segal E. Systematic Investigation of Transcription Factor Activity in the Context of Chromatin Using Massively Parallel Binding and Expression Assays Mol. Cell 65 2017 604 617.e6 10.1016/J.MOLCEL.2017.01.007 28212748
44 Stadler M.B. Murr R. Burger L. Ivanek R. Lienert F. Schöler A. van Nimwegen E. Wirbelauer C. Oakeley E.J. Gaidatzis D. DNA-binding factors shape the mouse methylome at distal regulatory regions Nature 480 2011 490 495 10.1038/NATURE10716 22170606
45 Kribelbauer J.F. Rastogi C. Bussemaker H.J. Mann R.S. Low-Affinity Binding Sites and the Transcription Factor Specificity Paradox in Eukaryotes Annu. Rev. Cell Dev. Biol. 35 2019 357 379 10.1146/ANNUREV-CELLBIO-100617-062719 31283382
46 Johnson D.S. Mortazavi A. Myers R.M. Wold B. Genome-wide mapping of in vivo protein-DNA interactions Science 316 2007 1497 1502 10.1126/SCIENCE.1141319 17540862
47 Plasschaert R.N. Vigneau S. Tempera I. Gupta R. Maksimoska J. Everett L. Davuluri R. Mamorstein R. Lieberman P.M. Schultz D. CTCF binding site sequence differences are associated with unique regulatory and functional trends during embryonic stem cell differentiation Nucleic Acids Res. 42 2014 774 789 10.1093/NAR/GKT910 24121688
48 Lai B. Gao W. Cui K. Xie W. Tang Q. Jin W. Hu G. Ni B. Zhao K. Principles of nucleosome organization revealed by single-cell MNase-seq Nature 562 2018 281 285 10.1038/S41586-018-0567-3 30258225
49 Lowary P.T. Widom J. New DNA sequence rules for high affinity binding to histone octamer and sequence-directed nucleosome positioning J. Mol. Biol. 276 1998 19 42 10.1006/JMBI.1997.1494 9514715
50 Gracey L.E. Chen Z.Y. Maniar J.M. Valouev A. Sidow A. Kay M.A. Fire A.Z. An in vitro-identified high-affinity nucleosome-positioning signal is capable of transiently positioning a nucleosome in vivo Epigenetics Chromatin 3 2010 13 10.1186/1756-8935-3-13 20594331
51 Perales R. Zhang L. Bentley D. Histone occupancy in vivo at the 601 nucleosome binding element is determined by transcriptional history Mol. Cell. Biol. 31 2011 3485 3496 10.1128/MCB.05599-11 21690290
52 Subtil-Rodríguez A. Reyes J.C. BRG1 helps RNA polymerase II to overcome a nucleosomal barrier during elongation, in vivo EMBO Rep. 11 2010 751 757 10.1038/EMBOR.2010.131 20829883
53 Zhang Y. Moqtaderi Z. Rattner B.P. Euskirchen G. Snyder M. Kadonaga J.T. Liu X.S. Struhl K. Intrinsic histone-DNA interactions are not the major determinant of nucleosome positions in vivo Nat. Struct. Mol. Biol. 16 2009 847 852 10.1038/NSMB.1636 19620965
54 Lancrey A. Joubert A. Duvernois-Berthet E. Routhier E. Raj S. Thierry A. Sigarteu M. Ponger L. Croquette V. Mozziconacci J. Boulé J.B. Nucleosome Positioning on Large Tandem DNA Repeats of the ‘601’ Sequence Engineered in Saccharomyces cerevisiae J. Mol. Biol. 434 2022 167497 10.1016/J.JMB.2022.167497
55 Zaret K.S. Pioneer Transcription Factors Initiating Gene Network Changes Annu. Rev. Genet. 54 2020 367 385 10.1146/ANNUREV-GENET-030220-015007 32886547
56 Yan C. Chen H. Bai L. Systematic Study of Nucleosome-Displacing Factors in Budding Yeast Mol. Cell 71 2018 294 305.e4 10.1016/J.MOLCEL.2018.06.017 30017582
57 Thoma E.C. Wischmeyer E. Offen N. Maurus K. Sirén A.L. Schartl M. Wagner T.U. Ectopic expression of neurogenin 2 alone is sufficient to induce differentiation of embryonic stem cells into mature neurons PLoS One 7 2012 e38651 10.1371/JOURNAL.PONE.0038651
58 Zhang Y. Pak C.H. Han Y. Ahlenius H. Zhang Z. Chanda S. Marro S. Patzke C. Acuna C. Covy J. Rapid single-step induction of functional neurons from human pluripotent stem cells Neuron 78 2013 785 798 10.1016/J.NEURON.2013.05.029 23764284
59 Wiechens N. Singh V. Gkikopoulos T. Schofield P. Rocha S. Owen-Hughes T. The Chromatin Remodelling Enzymes SNF2H and SNF2L Position Nucleosomes adjacent to CTCF and Other Transcription Factors PLoS Genet. 12 2016 e1005940 10.1371/JOURNAL.PGEN.1005940
60 de Boer C.G. Taipale J. Hold out the genome: a roadmap to solving the cis-regulatory code Nature. 625 2024 41 50 10.1038/s41586-023-06661-w 38093018
61 Isbel L. Grand R.S. Schübeler D. Generating specificity in genome regulation through transcription factor sensitivity to chromatin Nat. Rev. Genet. 23 2022 728 740 10.1038/s41576-022-00512-6 35831531
62 Hansen L. Mariño-Ramírez L. Landsman D. Differences in local genomic context of bound and unbound motifs Gene 506 2012 125 134 10.1016/J.GENE.2012.06.005 22692006
63 Madsen M.S. Siersbæk R. Boergesen M. Nielsen R. Mandrup S. Peroxisome proliferator-activated receptor γ and C/EBPα synergistically activate key metabolic adipocyte genes by assisted loading Mol. Cell. Biol. 34 2014 939 954 10.1128/MCB.01344-13 24379442
64 Siersbæk R. Scabia V. Nagarajan S. Chernukhin I. Papachristou E.K. Broome R. Johnston S.J. Joosten S.E.P. Green A.R. Kumar S. IL6/STAT3 Signaling Hijacks Estrogen Receptor α Enhancers to Drive Breast Cancer Metastasis Cancer Cell 38 2020 412 423.e9 10.1016/J.CCELL.2020.06.007 32679107
65 Wunderlich Z. Mirny L.A. Different gene regulation strategies revealed by analysis of binding motifs Trends Genet. 25 2009 434 440 10.1016/J.TIG.2009.08.003 19815308
66 Spitz F. Furlong E.E.M. Transcription factors: from enhancer binding to developmental control Nat. Rev. Genet. 13 2012 613 626 10.1038/NRG3207 22868264
67 Berman B.P. Nibu Y. Pfeiffer B.D. Tomancak P. Celniker S.E. Levine M. Rubin G.M. Eisen M.B. Exploiting transcription factor binding site clustering to identify cis-regulatory modules involved in pattern formation in the Drosophila genome Proc. Natl. Acad. Sci. USA 99 2002 757 762 10.1073/PNAS.231608898 11805330
68 Avsec Ž. Weilert M. Shrikumar A. Krueger S. Alexandari A. Dalal K. Fropf R. McAnany C. Gagneur J. Kundaje A. Zeitlinger J. Base-resolution models of transcription-factor binding reveal soft motif syntax Nat. Genet. 53 2021 354 366 10.1038/S41588-021-00782-6 33603233
69 Grossman S.R. Engreitz J. Ray J.P. Nguyen T.H. Hacohen N. Lander E.S. Positional specificity of different transcription factor classes within enhancers Proc. Natl. Acad. Sci. USA 115 2018 E7222 E7230 10.1073/PNAS.1804663115 29987030
70 Mayran A. Drouin J. Pioneer transcription factors shape the epigenetic landscape J. Biol. Chem. 293 2018 13795 13804 10.1074/JBC.R117.001232 29507097
71 Iwafuchi-Doi M. Zaret K.S. Cell fate control by pioneer transcription factors Development 143 2016 1833 1837 10.1242/DEV.133900 27246709
72 Zaret K.S. Carroll J.S. Pioneer transcription factors: establishing competence for gene expression Genes Dev. 25 2011 2227 2241 10.1101/GAD.176826.111 22056668
73 Struhl K. Segal E. Determinants of nucleosome positioning Nat. Struct. Mol. Biol. 20 2013 267 273 10.1038/NSMB.2506 23463311
74 Roberts G.A. Ozkan B. Gachulincová I. O’Dwyer M.R. Hall-Ponsele E. Saxena M. Robinson P.J. Soufi A. Dissecting OCT4 defines the role of nucleosome binding in pluripotency Nat. Cell Biol. 23 2021 834 845 10.1038/S41556-021-00727-5 34354236
75 Friman E.T. Deluz C. Meireles-Filho A.C.A. Govindan S. Gardeux V. Deplancke B. Suter D.M. Dynamic regulation of chromatin accessibility by pluripotency transcription factors across the cell cycle eLife 8 2019 10.7554/ELIFE.50087
76 Cernilogar F.M. Hasenöder S. Wang Z. Scheibner K. Burtscher I. Sterr M. Smialowski P. Groh S. Evenroed I.M. Gilfillan G.D. Pre-marked chromatin and transcription factor co-binding shape the pioneering activity of Foxa2 Nucleic Acids Res. 47 2019 9069 9086 10.1093/NAR/GKZ627 31350899
77 de Mendoza A. Nguyen T.V. Ford E. Poppe D. Buckberry S. Pflueger J. Grimmer M.R. Stolzenburg S. Bogdanovic O. Oshlack A. Large-scale manipulation of promoter DNA methylation reveals context-specific transcriptional responses and stability Genome Biol. 23 2022 163 10.1186/S13059-022-02728-5 35883107
78 Policarpi C. Dabin J. Hackett J.A. Epigenetic editing: Dissecting chromatin function in context BioEssays 43 2021 e2000316 10.1002/BIES.202000316
79 Chronis C. Fiziev P. Papp B. Butz S. Bonora G. Sabri S. Ernst J. Plath K. Cooperative Binding of Transcription Factors Orchestrates Reprogramming Cell 168 2017 442 459.e20 10.1016/J.CELL.2016.12.016 28111071
80 Gaidatzis D. Lerch A. Hahne F. Stadler M.B. QuasR: quantification and annotation of short reads in R Bioinformatics 31 2015 1130 1132 10.1093/BIOINFORMATICS/BTU781 25417205
81 Gu Z. Eils R. Schlesner M. Complex heatmaps reveal patterns and correlations in multidimensional genomic data Bioinformatics 32 2016 2847 2849 10.1093/BIOINFORMATICS/BTW313 27207943
82 Bolger A.M. Lohse M. Usadel B. Trimmomatic: a flexible trimmer for Illumina sequence data Bioinformatics 30 2014 2114 2120 10.1093/BIOINFORMATICS/BTU170 24695404
83 Langmead B. Trapnell C. Pop M. Salzberg S.L. Ultrafast and memory-efficient alignment of short DNA sequences to the human genome Genome Biol. 10 2009 R25 10.1186/GB-2009-10-3-R25
84 Hahne F. Ivanek R. Visualizing genomic data using Gviz and bioconductor Methods Mol. Biol. 1418 2016 335 351 10.1007/978-1-4939-3578-9_16 27008022
85 Mohn F. Weber M. Rebhan M. Roloff T.C. Richter J. Stadler M.B. Bibel M. Schübeler D. Lineage-specific polycomb targets and de novo DNA methylation define restriction and potential of neuronal progenitors Mol. Cell 30 2008 755 766 10.1016/J.MOLCEL.2008.05.007 18514006
86 Feng Y.Q. Seibler J. Alami R. Eisen A. Westerman K.A. Leboulch P. Fiering S. Bouhassira E.E. Site-specific chromosomal integration in mammalian cells: highly efficient CRE recombinase-mediated cassette exchange J. Mol. Biol. 292 1999 779 785 10.1006/JMBI.1999.3113 10525404
87 Domcke S. Bardet A.F. Adrian Ginno P. Hartl D. Burger L. Schübeler D. Competition between DNA methylation and transcription factors determines binding of NRF1 Nature 528 2015 575 579 10.1038/NATURE16462 26675734
88 Mathelier A. Fornes O. Arenillas D.J. Chen C.Y. Denay G. Lee J. Shi W. Shyr C. Tan G. Worsley-Hunt R. JASPAR 2016: a major expansion and update of the open-access database of transcription factor binding profiles Nucleic Acids Res. 44 2016 D110 D115 10.1093/NAR/GKV1176 26531826
89 Yang J.G. Madrid T.S. Sevastopoulos E. Narlikar G.J. The chromatin-remodeling enzyme ACF is an ATP-dependent DNA length sensor that regulates nucleosome spacing Nat. Struct. Mol. Biol. 13 2006 1078 1083 10.1038/NSMB1170 17099699
90 Li B. Qing T. Zhu J. Wen Z. Yu Y. Fukumura R. Zheng Y. Gondo Y. Shi L. A Comprehensive Mouse Transcriptomic BodyMap across 17 Tissues by RNA-seq Sci. Rep. 7 2017 4200 10.1038/S41598-017-04520-Z 28646208
91 McFadden W.M. Yanowitz J.L. idpr: A package for profiling and analyzing Intrinsically Disordered Proteins in R PLoS One 17 2022 e0266929 10.1371/JOURNAL.PONE.0266929
92 Ginno P.A. Gaidatzis D. Feldmann A. Hoerner L. Imanci D. Burger L. Zilbermann F. Peters A.H.F.M. Edenhofer F. Smallwood S.A. A genome-scale map of DNA methylation turnover identifies site-specific dependencies of DNMT and TET activity Nfiat. Commun. 11 2020 2680 10.1038/s41467-020-16354-x
