Single-molecule states link transcription factor binding to gene expression

BRM014 is a small molecule catalytic inhibitor of the ATPase activity of BRG1/BRM.

“We used the small molecule BRM014, a catalytic inhibitor of the ATPase activity of BRG1/BRM24”

• cites [24] Yudkovsky, N., Logie, C., Hahn, S. & Peterson, C. L. Recruitment of the SWI/SNF chromatin remodeling complex by transcriptional activators. Genes Dev. 13, 2369–2374 (1999). sourceCould not verify

  The ‘Transcriptional activators can target SWI/SNF remodeling activity’ section reports that remodeling depends on ATP, but does not discuss BRM014 or inhibition of BRG1/BRM ATPase activity. Searches for ‘BRM014 catalytic inhibitor BRG1 BRM’ and ‘small molecule inhibitor ATPase activity’ found no passage addressing the claimed inhibitor. (LLM, conf=0.78)

  Experimental — accuracy not yet measured

Single-molecule footprinting methods that use methyltransferases to mark accessible sites can provide quantitative measurements of binding states across individual molecules.

“single-molecule footprinting methods that use methyltransferases to mark accessible sites11–15 can provide quantitative measurements of these binding states across individual molecules”

• cites [11] Fryer, C. J. & Archer, T. K. Chromatin remodelling by the glucocorticoid receptor requires the BRG1 complex. Nature 393, 88–91 (1998). sourceNot checked — no full text of the source
• cites [12] Herschlag, D. & Johnson, F. B. Synergism in transcriptional activation: a kinetic view. Genes Dev. 7, 173–179 (1993). sourceCould not verify

  The “Branches in the pathway for transcriptional initiation” section mentions footprinting experiments showing that most DNA templates were bound within complexes. It does not describe methyltransferase-based single-molecule footprinting or quantitative measurement of binding states across individual molecules. Searches for those methods and measurements found no supporting passage in our copy; page numbers were unavailable. (LLM, conf=0.76)

  Experimental — accuracy not yet measured
• cites [13] Martinez-Corral, R. et al. Transcriptional kinetic synergy: a complex landscape revealed by integrating modeling and synthetic biology. Cell Syst. 14, 324–339 (2023). sourceCould not verify

  The Results section on the synthetic platform discusses transcription-factor binding and unbinding kinetics, while STAR★METHODS describes flow cytometry and fluorescence measurements (pages unavailable). Searches for single-molecule footprinting and methyltransferases found no passage describing methyltransferase marking of accessible sites or quantitative binding-state measurements across individual molecules. (LLM, conf=0.76)

  Experimental — accuracy not yet measured
• cites [14] Shipony, Z. et al. Long-range single-molecule mapping of chromatin accessibility in eukaryotes. Nat. Methods 17, 319–327 (2020). sourceSupported
• cites [15] Stergachis, A. B., Debo, B. M., Haugen, E., Churchman, L. S. & Stamatoyannopoulos, J. A. Single-molecule regulatory architectures captured by chromatin fiber sequencing. Science 368, 1449–1454 (2020). sourceNot checked — no full text of the source

VP48 is known to directly interact with chromatin-remodeling cofactors.

“Because VP48 is known to directly interact with chromatin-remodeling cofactors21–23”

• cites [21] Ackers, G. K., Johnson, A. D. & Shea, M. A. Quantitative model for gene regulation by lambda phage repressor. Proc. Natl Acad. Sci. USA 79, 1129–1133 (1982). sourceCould not verify

  The supplied copy describes a model of lambda phage repressor binding at the right operator (Abstract, page not specified). Searches for VP48 and chromatin-remodeling cofactors found no passage addressing whether they directly interact. (LLM, conf=0.75)

  Experimental — accuracy not yet measured
• cites [22] Kim, H. D. & O’Shea, E. K. A quantitative model of transcription factor-activated gene expression. Nat. Struct. Mol. Biol. 15, 1192–1198 (2008). sourceCould not verify

  The Discussion describes Pho4 binding and chromatin-remodeling kinetics at the PHO5 promoter, but does not address VP48 or direct interactions with remodeling cofactors. Searches for “VP48” and direct physical interactions found no passage supporting the attributed claim. (LLM, conf=0.76)

  Experimental — accuracy not yet measured
• cites [23] Neely, K. E. et al. Activation domain-mediated targeting of the SWI/SNF complex to promoters stimulates transcription from nucleosome arrays. Mol. Cell 4, 649–655 (1999). sourceNot checked — no full text of the source

Cooperativity between activation domains in controlling the transcription cycle is not required to explain the synergy between motif copy number and transcription.

“and cooperativity between activation domains in controlling the transcription cycle is not required to explain the synergy between motif copy number and transcription9.”

• cites [9] Mirny, L. A. Nucleosome-mediated cooperativity between transcription factors. Proc. Natl Acad. Sci. USA 107, 22534–22539 (2010). sourceSupported

The observation that p300 catalytic activity does not contribute to destabilization is inconsistent with a model by which histone acetylation reduces nucleosome-DNA affinity to facilitate TF binding.

“appears inconsistent with a model by which histone acetylation reduces nucleosome-DNA affinity to facilitate TF binding29,49,50”

• cites [29] Alerasool, N., Leng, H., Lin, Z.-Y., Gingras, A.-C. & Taipale, M. Identification and functional characterization of transcriptional activators in human cells. Mol. Cell 82, 677–695 (2022). sourceNot checked — no full text of the source
• cites [49] Mostafavi, S. et al. Parsing the interferon transcriptional network and its disease associations. Cell 164, 564–578 (2016). sourceCould not verify

  The Results section, “Mechanistic aspects of ISG induction: chromatin analyses” (page unavailable), discusses STAT2 binding and chromatin accessibility, but does not address the cited model that histone acetylation reduces nucleosome–DNA affinity to facilitate transcription-factor binding. Searches for that model and for p300 catalytic activity found no passage establishing the attributed point. The p300 observation is the citing authors’ framing, not the part this citation is offered to support. (LLM, conf=0.72)

  Experimental — accuracy not yet measured
• cites [50] Dogan, N. et al. Occupancy by key transcription factors is a more accurate predictor of enhancer activity than histone modifications or chromatin accessibility. Epigenetics Chromatin 8, 16 (2015). sourceCould not verify

  The Background describes TF-bound enhancers as flanked by nucleosomes bearing H3K27ac, and the Results and Discussion compare TF occupancy and histone marks as predictors of enhancer activity. No passage found addresses the cited model that acetylation reduces nucleosome–DNA affinity to facilitate TF binding; searches for nucleosome affinity, acetylation-mediated binding, and p300 catalytic activity did not resolve it. (LLM, conf=0.73)

  Experimental — accuracy not yet measured

Each bound TF adds linearly to the rate of promoter activation.

“We observe that each bound TF adds linearly to the rate of promoter activation2, and cooperativity between activation domains in controlling the transcription cycle is not required to explain the synergy between motif copy number and transcription9.”

• cites [2] Kelly, T. K. et al. Genome-wide mapping of nucleosome positioning and DNA methylation within individual DNA molecules. Genome Res. 22, 2497–2506 (2012). sourceCould not verify

  The Results section on nucleosome occupancy at transcription start sites links promoter chromatin patterns with expression level, but does not address whether each bound transcription factor adds linearly to the rate of promoter activation. Searches for that claim and for additive TF binding or promoter activation rates found no relevant passage in our copy; page numbers were unavailable. (LLM, conf=0.73)

  Experimental — accuracy not yet measured

Reducing the dose of dox tunes the effective rTetR-VP48 concentration as measured in vitro by an Electrophoretic Mobility Shift Assay.

“By reducing the dose of dox, we tuned the effective rTetR-VP48 concentration36 as measured in vitro by an Electrophoretic Mobility Shift Assay (Fig. S4L–O)”

• cites [36] Xiao, J. Y., Hafner, A. & Boettiger, A. N. How subtle changes in 3D structure can create large changes in transcription. Elife 10, e64320 (2021). sourceCould not verify

  The Materials and methods section discusses transcription-factor concentration and enhancer occupancy only in general terms. Searches for rTetR-VP48, doxycycline or dox, and electrophoretic mobility shift assays found no passage in our copy addressing whether reducing dox tunes effective rTetR-VP48 concentration or how that was measured. The copy provides no page numbers for these passages. (LLM, conf=0.75)

  Experimental — accuracy not yet measured

The decrease in wide footprints is consistent with the known downregulation of type I IFN response after activation.

“consistent with the known downregulation of type I IFN response after activation46.”

• cites [46] Manry, J. et al. Evolutionary genetic dissection of human interferons. J. Exp. Med. 208, 2747–2759 (2011). sourceCould not verify

  The Discussion addresses variation in type I interferon production and proposes that different subtypes may regulate overall activity, but it does not describe downregulation of the type I IFN response after activation. Searches for that phrase and for post-activation attenuation or negative feedback found no passage addressing the claimed timing or direction of change. No page number is available in our copy. (LLM, conf=0.72)

  Experimental — accuracy not yet measured

The relationship between TF occupancy and gene expression can be well fit by modifying the telegraph model to include a promoter on-rate proportional to the average number of TFs bound (Additive Activation model).

“This relationship between TF occupancy and gene expression can be well fit by modifying the commonly-used “telegraph” model of transcription31–33, which assumes that promoters can be either “on” (transcribing) or “off”, to include a promoter on-rate that is proportional to the average number of TFs bound (Additive Activation model; Fig. 3F, Methods)”

• cites [31] Brower-Toland, B. et al. Specific contributions of histone tails and their acetylation to the mechanical stability of nucleosomes. J. Mol. Biol. 346, 135–146 (2005). sourceNot checked — no full text of the source
• cites [32] Lasko, L. M. et al. Discovery of a selective catalytic p300/CBP inhibitor that targets lineage-specific tumours. Nature 550, 128–132 (2017). sourceCould not verify

  The passages on gene-expression profiling discuss the effects of the p300/CBP inhibitor A-485, not a model relating transcription-factor occupancy to expression (section labelled Abstract; page unavailable). Searches for the telegraph model, Additive Activation, and a promoter on-rate proportional to bound transcription factors found no passage addressing that claim. (LLM, conf=0.76)

  Experimental — accuracy not yet measured
• cites [33] Raj, A., Peskin, C. S., Tranchina, D., Vargas, D. Y. & Tyagi, S. Stochastic mRNA synthesis in mammalian cells. PLoS Biol. 4, e309 (2006). sourceSupported

The strong effects of p300 inhibition on promoter activation support previous reports on the role of p300 catalytic activity in assembling the pre-initiation complex and recruiting RNA polymerase.

“also supports previous reports on the role of the catalytic activity of p300 in assembling the pre-initiation complex (PIC) and recruiting RNA polymerase51.”

• cites [51] Corces, M. R. et al. Lineage-specific and single-cell chromatin accessibility charts human hematopoiesis and leukemia evolution. Nat. Genet. 48, 1193–1203 (2016). sourceCould not verify

  The Results section on regulatory networks of normal hematopoiesis discusses transcription-factor motifs and chromatin accessibility, not p300 catalytic activity. Searches for p300, pre-initiation complex assembly, and RNA polymerase recruitment found no passage addressing the attributed mechanism in our copy. (LLM, conf=0.76)

  Experimental — accuracy not yet measured

Prior to induction of TF binding (-dox), the synthetic reporter constructs were relatively inaccessible on average, with methylation probabilities of around 20%.

“Prior to the induction of TF binding (-dox), the synthetic reporter constructs were relatively inaccessible on average, with methylation probabilities of around 20% (Fig. 1E, gray)”

There is a strong preference for TFs bound to adjacent sites on molecules with nucleosomes.

“We see a strong preference for TFs bound to adjacent sites on molecules with nucleosomes, which represent the majority of molecules in our dataset (Fig. 2D)”

TF binding on nucleosome-free DNA is consistent with independent binding between TFs.

“TF binding on nucleosome-free DNA is consistent with independent binding between TFs”

• Also stated in: § TF occupancy exhibits cooperative behavior as a function of the number of binding sites · hedged

  “once we condition on molecules with no nucleosomes over the enhancer, we observe equal representation of molecular configurations (Fig. 2D) that is consistent with independent binding on free DNA”

Individual molecules with the same synthetic reporter sequence display diverse methylation patterns corresponding to distinct molecular binding configurations.

“Individual molecules with the same synthetic reporter sequence display diverse patterns of methylation that correspond to distinct and interpretable molecular binding configurations”

Increasing numbers of TetO sites in the enhancer produce dox-dependent, non-linear increases in gene expression.

“We observed dox-dependent, non-linear increases in gene expression for increasing numbers of TetOs in the enhancer (Fig. 1B, Fig. S1A)”

The TF binding distribution of each enhancer construct, even 7xTetO, reveals that very few molecules have all TetO sites occupied.

“The TF binding distribution of each enhancer construct, even those with high numbers of sites such as 7xTetO, revealed that very few molecules have all TetO sites occupied (Fig. 2C)”

The reduction in TetO occupancy with rTetR without VP48 could not be explained by a reduced TF concentration.

“that could not be explained by a reduced TF concentration (Fig. S3F)”

Average TF occupancy linearly contributes to promoter activation rates.

“average TF occupancy linearly contributes to promoter activation rates”

The average number of bound TFs exhibits cooperative behavior as a function of the number of binding sites.

“the average number of bound TFs exhibited cooperative behavior as a function of the number of binding sites (Fig. 2B)”

After applying the probabilistic model, substantial heterogeneity in molecular configurations is observed even on identical enhancer constructs, with variable nucleosome positions and degrees of TF binding.

“we observed substantial heterogeneity in the molecular configurations even on identical enhancer constructs, with variable nucleosome positions and degrees of TF binding (Fig. 1F, Fig. S1I–J)”

Predicted distributions of binding microstates from the Nucleosome Destabilization model were qualitatively similar to those observed in the measurements.

“We used the Nucleosome Destabilization model to generate predicted distributions of binding microstates which were qualitatively similar to those observed in our measurements (Fig. 2I)”

• AssumptionThe authors state this as an assumption: TF binding is independent in the thermodynamic model.
• AssumptionThe authors state this as an assumption: TF binding is at (non-equilibrium) steady state in the thermodynamic model.

Increasing the number of TetO sites results in a non-linear increase in the fraction of accessible and TF-bound enhancers, with a large increase between four and six TetO sites.

“Increasing the number of TetO sites resulted in a non-linear increase in the fraction of accessible and TF-bound enhancers, with a large increase between four and six TetO sites (Fig. 2A)”

The Simple Competition Model cannot reproduce the bimodality observed in the binding data.

“This two-parameter model, which allows only for TF-DNA and nucleosome-DNA interaction energies (Fig. 2F), cannot reproduce the bimodality observed in our binding data (Fig. 2C)”

• AssumptionThe authors state this as an assumption: TF binding is independent in the thermodynamic model.
• AssumptionThe authors state this as an assumption: TF binding is at (non-equilibrium) steady state in the thermodynamic model.

In a cell line expressing rTetR without VP48, there is an average reduction in TetO occupancy of 35% across the library.

“we first carried out the same SMF measurements in a cell line expressing rTetR without VP48 (Methods), thus unable to recruit coactivators, and observed an average reduction in TetO occupancy of 35% across the library (Fig. 2J–K)”

Different TFs possess distinct capacities or potencies once bound.

“different TFs possess distinct capacities (or potencies) once bound”

The capacity of TFs to activate transcription is proportional to the average number of bound TFs.

“we observe that the capacity of TFs to activate transcription is proportional to the average number of bound TFs”

Alternative models encoding other VP48-dependent molecular interactions are unable to similarly fit the observed data.

“Alternative models encoding other VP48-dependent molecular interactions are unable to similarly fit the observed data (Fig. S3E)”

Average TF occupancy directly correlates with accessibility as measured by ATAC-seq peak strength.

“This average occupancy directly correlates with accessibility as measured by ATAC-seq peak strength (Fig. S3A)”

After conditioning on molecules with the same number of TFs bound, there is still a dependence on the number of TetO sites.

“However, after conditioning on molecules with the same number of TFs bound, there is still a dependence on the number of TetO sites (Fig. 3H)”

Stimulation of rTetR-VP48/TetO reporter lines with saturating dox produces a time-dependent increase in rTetR-VP48 binding that reaches steady state by 8 hours.

“We observed a time-dependent increase in the binding of rTetR-VP48 which reached steady state by 8 hours, although a significant portion of the TF binding occurs while cells are being processed in buffers containing dox during the assay (Fig. 5A, Fig. S8A).”

Combining models for TF occupancy from TetO number, promoter activation from TF occupancy, and gene expression from promoter activation can largely explain gene expression as a function of the number of TetO sites.

“In combining our models for TF occupancy from TetO number (Nucleosome Destabilization model, 3 parameters for each background), promoter activation from TF occupancy (Additive Activation model, 1 parameter across all backgrounds), and gene expression from promoter activation (linear fit, 1 parameter across all backgrounds), we can largely explain gene expression as a function of the number of TetO”

Inhibiting p300 has strong effects on promoter activation.

“The observation that inhibiting p300 does have strong effects on promoter activation also supports previous reports on the role of the catalytic activity of p300 in assembling the pre-initiation complex (PIC) and recruiting RNA polymerase51.”

Pre-stimulation TF footprints are observed at ISREs genome-wide.

“and we observe pre-stimulation TF footprints at ISREs genome-wide (Fig. 4I).”

Average rTetR-VP48 occupancy remains linearly related to the active promoter fraction at each timepoint during the activation process.

“Average rTetR-VP48 occupancy remains linearly related to the active promoter fraction at each timepoint during this activation process (Fig. 5B)”

Kinetic model fits to wide footprint dynamics reveal a shorter binding half-time for ISGF3 than for rTetR-VP48.

“Kinetic model fits (as described in Fig. 5D) to wide footprint dynamics reveal a shorter binding half-time than rTetR-VP48, consistent with pre-accessibility enabling faster TF binding (however, a direct comparison cannot be made as this term also depends on TF concentration) (Fig. 5N, Fig. S8J).”

• AssumptionThe authors state this as an assumption: Each bound TF increases the activation rate of the promoter.
• AssumptionThe authors state this as an assumption: The promoter on-rate scales with average TF occupancy.

Average single-molecule methylation is linearly correlated with ATAC-seq signal strength.

“However, average single-molecule methylation is linearly correlated with ATAC-seq signal strength, suggesting that genome-wide chromatin accessibility measurements accurately report information regarding the average single-molecule state present in cells.”

There is an hours-long delay between rTetR-VP48 binding and transcriptional activation, measured by active promoter fraction and RT-qPCR, which does not reach steady state until between 8 and 24 hours.

“However, we observed an hours-long delay between rTetR-VP48 binding and transcriptional activation (as measured by both the active promoter fraction and RT-qPCR), which does not reach steady state until between 8 and 24 hours (Fig. 5B, Fig. S8B).”

Promoters and promoter-proximal enhancers of ISRE-containing ISGs show more pre-stimulation accessibility in ATAC-seq than non-ISGs.

“Consistent with our hypothesis, the promoters and promoter-proximal enhancers of ISRE-containing ISGs demonstrate more pre-stimulation accessibility in ATAC-seq compared to non-ISGs (Fig. 4H)”

Increasing ISRE copy number produces a non-linear increase in reporter expression, measured by Citrine fluorescence and mRNA levels, only after addition of IFN-β.

“We observe a non-linear increase in reporter expression (Citrine fluorescence and mRNA levels) with increasing ISRE copy number only after addition of IFN-β (Fig. 4B)”

Average TF occupancy continues to linearly contribute to the promoter activation rate across dox concentrations with the same slope.

“However, average TF occupancy continues to linearly contribute to the promoter activation rate across dox concentrations with the same slope (Fig. 3J)”

Narrow footprints are not predictive of the fraction of promoters active.

“While narrow footprints are not predictive of the fraction of promoters active (Fig. S6K)”

The single-molecule binding distributions of both synthetic TFs with activation domains (rTetR-VP48) and native TFs (IRF9) are best fit by a model requiring destabilization of nucleosome-DNA interactions across the entire regulatory element upon binding of at least one TF.

“The single-molecule binding distributions of both synthetic TFs with activation domains (rTetR-VP48) and native TFs (IRF9) are best fit by a model requiring the destabilization of nucleosome-DNA interactions across the entire regulatory element upon the binding of at least one TF.”

The fit promoter on-rate for ISGF3 is more than 5-fold faster than for rTetR-VP48.

“and the fit promoter on-rate is >5-fold faster (Fig. 5L,N, Fig. S8M).”

Both BAF and p300 inhibition decrease TF potency as fit by the Additive Activation model.

“In contrast, both BAF and p300 inhibition decrease TF potency as fit by the Additive Activation model (Fig. 3M)”

The fraction of active promoters is always proportional to the average number of bound TFs with an invariable slope across all backgrounds.

“However, the fraction of active promoters is always proportional to the average number of bound TFs with an invariable slope across all backgrounds (Fig. 3D)”

The IFN-β-stimulated fraction of active promoters increases non-linearly with increasing ISRE copy number.

“the IFN-β-stimulated fraction of active promoters increases non-linearly with increasing ISRE copy number (Fig. S6I)”

A significant portion of rTetR-VP48 binding occurs while cells are processed in buffers containing dox during the assay.

“although a significant portion of the TF binding occurs while cells are being processed in buffers containing dox during the assay (Fig. 5A, Fig. S8A).”

The core promoter and promoter-proximal ISRE-containing enhancers of ISG15, IFI6, and USP18 exhibit significant pre-stimulation accessibility with variable increases post-stimulation.

“These regulatory elements exhibit significant pre-stimulation accessibility with variable increases post-stimulation (Fig. 4L, Fig. S6P), consistent with ATAC-seq data at the endogenous genes (Fig. S6Q)”

Promoter dynamics closely follow those of the wide footprints.

“and the promoter dynamics closely follow those of the wide footprints, further suggesting that footprint width can discriminate between inactive IRF9 and active ISGF3 (Fig. 5L–M).”

p300 catalytic activity does not contribute to destabilization of nucleosome-DNA interactions.

“Our observation that p300 catalytic activity does not contribute to destabilization of nucleosome-DNA interactions appears inconsistent with a model by which histone acetylation reduces nucleosome-DNA affinity to facilitate TF binding29,49,50 and is instead consistent with recent observations that inhibiting p300 catalytic activity actually stabilizes interactions with chromatin51,52.”

The kinetic model describing TF binding, promoter activation, transcription, and translation accurately captures the temporal delays between successive steps of the central dogma across all enhancers in the rTetR-VP48/TetO reporter library.

“These models accurately capture the temporal delays between successive steps of the central dogma across all enhancers in our rTetR-VP48/TetO reporter library (Fig. 5E–F).”

• AssumptionThe authors state this as an assumption: Each bound TF increases the activation rate of the promoter.
• AssumptionThe authors state this as an assumption: The promoter on-rate scales with average TF occupancy.

Reducing dox causes a dox-dependent decrease in the average TF occupancy across TetO copy number.

“observed a dox-dependent decrease in the average TF occupancy across TetO copy number (Fig. 3I, Fig. S4P)”

The study identified a decoupling of chromatin accessibility and transcriptional activity in type I IFN response.

“We also identified a decoupling of chromatin accessibility and transcriptional activity in type I IFN response, challenging the common heuristic that accessibility implies transcriptional activity.”

The probability that a promoter is in the active state increases with the number of TFs bound on the same molecule when combining data across all enhancers or restricting to individual enhancers.

“The probability that we find a promoter in the active state increases with the number of TFs bound on the same molecule when we combine data across all enhancers (Fig. S4I) or restrict to individual enhancers (Fig. S4J)”

The fraction of active promoters remains linearly related to gene expression.

“Furthermore, the fraction of active promoters remains linearly related to gene expression (Fig. S6L)”

The apparent potency of rTetR-VP48 increases over time.

“and the apparent potency increases over time, confirming that promoters integrate information from multiple binding events (Fig. 5C).”

Average wide footprint occupancy remains predictive of the fraction of active promoters during activation.

“Average wide footprint occupancy remains predictive of the fraction of active promoters during activation, and the promoter dynamics closely follow those of the wide footprints, further suggesting that footprint width can discriminate between inactive IRF9 and active ISGF3 (Fig. 5L–M).”

An ISGF3 reporter demonstrates faster activation than an rTetR-VP48 reporter with the same maximum active promoter fraction.

“Furthermore, an ISGF3 reporter demonstrates faster activation when compared to an rTetR-VP48 reporter with the same maximum active promoter fraction, and the fit promoter on-rate is >5-fold faster (Fig. 5L,N, Fig. S8M).”

Regulatory elements exhibit a broad diversity of microscopic binding states, including states that are primarily nucleosome occupied.

“Our single-molecule occupancy data reveal that regulatory elements exhibit a broad diversity of microscopic binding states, including states that are primarily nucleosome occupied, suggesting that even highly active elements are likely accessible to regulatory proteins only in a subset of cells.”

BAF inhibition slows down rTetR-VP48 TF binding.

“As predicted, BAF inhibition slows down TF binding (Fig. 5G,J, Fig. S8H) while both inhibitors decrease the rate of promoter activation (Fig. 5H,J, Fig. S8I) and lead to a slower increase in potency (Fig. 5I).”

The core promoter and promoter-proximal ISRE-containing enhancers of ISG15, IFI6, and USP18 only express Citrine post-stimulation.

“and only express Citrine post-stimulation (Fig. 4M).”

The combined TF binding and promoter activation models correctly predict that BAF inhibition has a larger effect on expression than p300 inhibition.

“our combined TF binding and promoter activation models correctly predict that BAF inhibition has a larger effect on expression than p300 inhibition (Fig. 3N)”

For rTetR-VP48, a significant amount of binding occurs on ice during sample processing.

“For rTetR-VP48, since there is a significant amount of binding that happens on ice during sample processing (see Fig. S5A), the sample of 0 hours induction with dox present in buffers during handling was used as the zero-time sample.”

Thermodynamic and kinetic models quantitatively predict both the binding microstates observed at the enhancer and subsequent time-dependent gene expression.

“we develop thermodynamic and kinetic models that quantitatively predict both the binding microstates observed at the enhancer and subsequent time-dependent gene expression”

The binding of TFs is not simply determined by their DNA-binding domains; activation domains also play a large role in determining in vivo TF occupancy.

“This also implies that the binding of TFs is not simply determined by their DNA-binding domains, but rather that activation domains also play a large role in determining in vivo TF occupancy”

Activation domains enhance a transcription factor's capacity to compete with nucleosomes for binding to DNA in a BAF-dependent manner.

“We find that activation domains enhance a TF’s capacity to compete with nucleosomes for binding to DNA in a BAF-dependent manner”

This may help explain the poor prediction of in vivo TF occupancy from in vitro binding affinities, often using only DNA binding domains.

“This may help explain the poor prediction of in vivo TF occupancy from in vitro binding affinities (often using only DNA binding domains)48.”

• cites [48] Arimoto, K.-I., Miyauchi, S., Stoner, S. A., Fan, J.-B. & Zhang, D.-E. Negative regulation of type I IFN signaling. J. Leukoc. Biol. https://doi.org/10.1002/JLB.2MIR0817-342R (2018). sourceNot checked — no full text of the source

Comparisons between the synthetic activation system and ISREs suggest that cells have multiple ways to increase the rate by which promoters are activated.

“Comparisons between our synthetic activation system and ISREs suggests that cells have multiple ways to increase the rate by which promoters are activated.”

Footprint width can discriminate between inactive IRF9 and active ISGF3.

“further suggesting that footprint width can discriminate between inactive IRF9 and active ISGF3 (Fig. 5L–M).”

The synthetic approach allowed engineering GpC dinucleotides in strategic locations that enabled unambiguous assignment of molecular states across each molecule.

“Furthermore, the synthetic approach we describe here allowed us to engineer GpC dinucleotides in strategic locations that enabled unambiguous assignment of molecular states across each molecule.”

Each bound TF contributes additively to the rate of promoter activation.

“That each bound TF contributes additively to the rate of promoter activation suggests that, for this system, the observed non-linearity between gene expression and TetO copy number stems primarily from AD-dependent nucleosome destabilization via chromatin remodelers (allowing for increased TF binding) and not from additional synergy in promoter activation once TFs are bound (Fig. S4F–H)”

Each additional bound TF contributes the same amount to promoter activation.

“demonstrating that each additional bound TF contributes the same amount to promoter activation”

The observed decoupling challenges the common heuristic that accessibility implies transcriptional activity.

“challenging the common heuristic that accessibility implies transcriptional activity.”

TF binding dynamics are likely to differ between the interferon system and the rTetR-VP48 system.

“Since steady-state TF binding and DNA accessibility are higher pre-activation in the interferon system compared to the rTetR-VP48 system, TF binding dynamics are also likely to differ.”

The quantitative statistical mechanical model provides a clearly defined mechanism by which a TF might begin to establish accessibility at a regulatory element and facilitate binding of additional TFs, even without directly binding to nucleosomal DNA.

“Furthermore, our quantitative statistical mechanical model also provides a clearly defined mechanism by which a TF might begin to establish accessibility at a regulatory element and facilitate the binding of additional TFs, even without directly binding to nucleosomal DNA.”

The platform provides a facile mechanism to separately quantify the contributions of TF binding and TF potency on gene expression, features that are largely convolved in other assays.

“Overall, however, our platform for linking microscopic molecular states at a regulatory element to transcriptional states at a promoter provides a facile mechanism to separately quantify the contributions of TF binding and TF potency on gene expression, features that are largely convolved in other assays.”

The approach can only be applied to TFs with binding residence times longer than the timescale of the irreversible methylation reaction to avoid underestimating the true TF occupancy state at regulatory elements.

“Finally, this approach can only be applied to TFs with binding residence times longer than the timescale of the (irreversible) methylation reaction to avoid underestimating the true TF occupancy state at regulatory elements.”

The footprinting system was only studied with two distinct types of TFs that were readily inducible: rTetR-VP48 via dox and ISGF3 via IFN-β.

“First, we only studied two distinct types of TFs that were readily inducible (through dox addition for rTetR-VP48 and through stimulation with IFN-β for ISGF3).”

Genome-wide chromatin accessibility measurements accurately report information regarding the average single-molecule state present in cells.

“suggesting that genome-wide chromatin accessibility measurements accurately report information regarding the average single-molecule state present in cells.”

Promoters integrate information from multiple binding events.

“and the apparent potency increases over time, confirming that promoters integrate information from multiple binding events (Fig. 5C).”

The lower abundance and variable positioning of GpCs in the genome makes unambiguous assignment of molecular states much more challenging at non-engineered genomic loci.

“The lower abundance and variable positioning of GpCs in the genome makes this task much more challenging at non-engineered genomic loci.”

Potency is an intrinsic property of a TF that is independent of a TF's ability to bind DNA.

“This demonstrates that potency is an intrinsic property of a TF that is independent of a TF’s ability to bind DNA”

Pre-accessibility enables faster TF binding.

“consistent with pre-accessibility enabling faster TF binding (however, a direct comparison cannot be made as this term also depends on TF concentration) (Fig. 5N, Fig. S8J).”

TF potency depends on the activity of the cofactors that VP48 recruits.

“demonstrating that this potency depends on the activity of the cofactors that VP48 recruits”

Even highly active elements are likely accessible to regulatory proteins only in a subset of cells.

“suggesting that even highly active elements are likely accessible to regulatory proteins only in a subset of cells.”

ISGF3 has a shorter delay between TF binding and promoter activation than rTetR-VP48.

“consistent with a shorter delay between TF binding and promoter activation (Fig. 5B,M, Fig. S8L).”

Average TF binding occupancy and TF potency are separable parameters that multiplicatively predict transcriptional output.

“Overall, these results demonstrate that average TF binding occupancy and TF potency are separable parameters that multiplicatively predict transcriptional output (Fig. 3O)”

Activation domain:cofactor interactions play a substantial role in determining TF binding within chromatin.

“The necessity of this destabilization parameter, alongside empirical decreases in TF binding upon activation domain removal and cofactor inhibition, implies that activation domain:cofactor interactions play a substantial role in determining TF binding within chromatin.”

The fraction of active promoters can be interpreted as an accurate single-molecule metric for transcriptional output.

“demonstrating that the fraction of active promoters can be interpreted as an accurate single-molecule metric for transcriptional output”

ISREs are the critical sequence elements for driving and decoupling accessibility and expression at ISGs.

“demonstrating that ISREs are the critical sequence elements for driving and decoupling accessibility and expression at ISGs.”

Expression level remains directly proportional to the aggregate occupancy of activating TFs.

“demonstrating that expression level remains directly proportional to the aggregate occupancy of activating TFs (Fig. 4SM)”