Transcription factors relay information from the external environment to gene regulatory networks that control cell physiology. To confer signalling specificity, robustness and coordination, these signalling networks use temporal communication codes, such as the amplitude, duration or frequency of signals. Although much is known about how temporal information is encoded, a mechanistic understanding of how gene regulatory networks decode signalling dynamics is lacking. Recent advances in our understanding of phase separation of transcriptional condensates provide new biophysical frameworks for both temporal encoding and decoding mechanisms. In this Perspective, we summarize the mechanisms by which transcriptional condensates could enable temporal decoding through signal adaptation, memory and persistence. We further outline methods to probe and manipulate dynamic communication codes of transcription factors and condensates to rationally control gene activation.

Transcription factors can form nuclear condensates at genomic sites, and condensates are thought to enhance transcriptional activity. In this issue of Cell, Saad et al. suggest that DNA binding prevents rather than facilitates condensate formation of particularly aggregation-prone transcription factors.

Developmental gene expression is a remarkably conserved process, yet most cis-regulatory elements (CREs) lack sequence conservation, especially at larger evolutionary distances. Some evidence suggests that CREs at the same genomic position remain functionally conserved independent of sequence conservation. However, the extent of such positional conservation remains unclear. Here, we profiled the regulatory genome in mouse and chicken embryonic hearts at equivalent developmental stages and found that most CREs lack sequence conservation. To identify positionally conserved CREs, we introduced the synteny-based algorithm interspecies point projection, which identifies up to fivefold more orthologs than alignment-based approaches. We termed positionally conserved orthologs 'indirectly conserved' and showed that they exhibited chromatin signatures and sequence composition similar to sequence-conserved CREs but greater shuffling of transcription factor binding sites between orthologs. Finally, we validated indirectly conserved chicken enhancers using in vivo reporter assays in mouse. By overcoming alignment-based limitations, we revealed widespread functional conservation of sequence-divergent CREs.

The mammalian cortex is comprised of cells classified into types according to shared properties. Defining the contribution of each cell type to the processes guided by the cortex is essential for understanding its function in health and disease. We use transcriptomic and epigenomic cortical cell-type taxonomies from mouse and human to define marker genes and putative enhancers and create a large toolkit of transgenic lines and enhancer adeno-associated viruses (AAVs) for selective targeting of cortical cell populations. We report creation and evaluation of fifteen transgenic driver lines, two reporter lines, and >1,000 different enhancer AAV vectors covering most subclasses of cortical cells. The tools reported here have been made publicly available, and along with the scaled process of tool creation, evaluation, and modification, they will enable diverse experimental strategies toward understanding mammalian cortex and brain function.

Humans have evolved an extraordinarily expanded and complex cerebral cortex associated with developmental and gene regulatory modifications. Human accelerated regions (HARs) are highly conserved DNA sequences with human-specific nucleotide substitutions. Although there are thousands of annotated HARs, their functional contribution to species-specific cortical development remains largely unknown. HARE5 is a HAR transcriptional enhancer of the WNT signalling receptor Frizzled8 that is active during brain development. Here, using genome-edited mouse (Mus musculus, Mm) and primate models, we demonstrated that human (Homo sapiens, Hs) HARE5 fine-tunes cortical development and connectivity by controlling the proliferative and neurogenic capacities of neural progenitor cells. Hs-HARE5 knock-in mice have significantly enlarged neocortices, containing more excitatory neurons. By measuring neural dynamics in vivo, we showed that these anatomical features result in increased functional independence between cortical regions. We assessed underlying developmental mechanisms using fixed and live imaging, lineage analysis and single-cell RNA sequencing. We discovered that Hs-HARE5 modifies radial glial cell behaviour, with increased self-renewal at early developmental stages, followed by expanded neurogenic potential. Using genome-edited human and chimpanzee (Pan troglodytes, Pt) neural progenitor cells and cortical organoids, we showed that four human-specific variants of Hs-HARE5 drive increased enhancer activity that promotes progenitor proliferation. Finally, we showed that Hs-HARE5 increased progenitor proliferation by amplifying canonical WNT signalling. These findings illustrate how small changes in regulatory DNA can directly affect critical signalling pathways to modulate brain development. Our study uncovered new functions of HARs as key regulatory elements crucial for the expansion and complexity of the human cerebral cortex.

Ultra-long-range genomic contacts, which are key components of neuronal genome architecture, constitute a biochemical enigma. This is because regulatory DNA elements make selective and stable contacts with DNA sequences located hundreds of kilobases away, instead of interacting with proximal sequences occupied by the exact same transcription factors. This is exemplified in olfactory sensory neurons (OSNs), in which only a fraction of LHX2-, EBF1- and LDB1-bound sites interact with each other, converging into highly selective multi-chromosomal enhancer hubs. To obtain biochemical insight into this process, here we assembled olfactory receptor (OR) enhancer hubs in vitro with recombinant proteins and enhancer DNA. Cell-free reconstitution of enhancer hubs revealed that OR enhancers form nucleoprotein condensates with unusual, solid-like characteristics. Assembly of these solid condensates is orchestrated by specific DNA motifs enriched in OR enhancers, which are likely to confer distinct homotypic properties on their resident LHX2-EBF1-LDB1 complexes. Single-molecule tracking and pulse-chase experiments in vivo confirmed that LHX2 and EBF1 assemble OR-transcription-competent condensates with solid properties in OSN nuclei, under physiological concentrations of protein. Thus, homophilic nucleoprotein interactions that are influenced by DNA sequence generate new types of biomolecular condensate, which might provide a generalizable explanation for the stability and specificity of long-range genomic contacts across cell types.

Transcription of genes is regulated by DNA elements such as promoters and enhancers, the activity of which are in turn controlled by many transcription factors. Owing to the highly complex combinatorial logic involved, it has been difficult to construct computational models that predict gene activity from DNA sequence. Recent advances in deep learning techniques applied to data from epigenome mapping and high-throughput reporter assays have made substantial progress towards addressing this complexity. Such models can capture the regulatory grammar with remarkable accuracy and show great promise in predicting the effects of non-coding variants, uncovering detailed molecular mechanisms of gene regulation and designing synthetic regulatory elements for biotechnology. Here, we discuss the principles of these approaches, the types of training data sets that are available and the strengths and limitations of different approaches.

Genome conformation underlies transcriptional regulation by distal enhancers, and genomic rearrangements in cancer can alter critical regulatory interactions. Here we profiled the three-dimensional genome architecture and enhancer connectome of 69 tumor samples spanning 15 primary human cancer types from The Cancer Genome Atlas. We discovered the following three archetypes of enhancer usage for over 100 oncogenes across human cancers: static, selective gain or dynamic rewiring. Integrative analyses revealed the enhancer landscape of noncancer cells in the tumor microenvironment for genes related to immune escape. Deep whole-genome sequencing and enhancer connectome mapping provided accurate detection and validation of diverse structural variants across cancer genomes and revealed distinct enhancer rewiring consequences from noncoding point mutations, genomic inversions, translocations and focal amplifications. Extrachromosomal DNA promoted more extensive enhancer rewiring among several types of focal amplification mechanisms. These results suggest a systematic approach to understanding genome topology in cancer etiology and therapy.

During cellular differentiation, enhancers transform overlapping gradients of transcription factors (TFs) to highly specific gene expression patterns. However, the vast complexity of regulatory DNA impedes the identification of the underlying cis-regulatory rules. Here, we characterized 64,400 fully synthetic DNA sequences to bottom-up dissect design principles of cell-state-specific enhancers in the context of the differentiation of blood stem cells to seven myeloid lineages. Focusing on binding sites for 38 TFs and their pairwise interactions, we found that identical sites displayed both repressive and activating function as a consequence of cell state, site combinatorics, or simply predicted occupancy of a TF on an enhancer. Surprisingly, combinations of activating sites frequently neutralized one another or gained repressive function. These negative synergies convert quantitative imbalances in TF expression into binary activity patterns. We exploit this principle to automatically create enhancers with specificity to user-defined combinations of hematopoietic progenitor cell states from scratch.

Measurements of gene expression or signal transduction activity are conventionally performed using methods that require either the destruction or live imaging of a biological sample within the timeframe of interest. Here we demonstrate an alternative paradigm in which such biological activities are stably recorded to the genome. Enhancer-driven genomic recording of transcriptional activity in multiplex (ENGRAM) is based on the signal-dependent production of prime editing guide RNAs that mediate the insertion of signal-specific barcodes (symbols) into a genomically encoded recording unit. We show how this strategy can be used for multiplex recording of the cell-type-specific activities of dozens to hundreds of cis-regulatory elements with high fidelity, sensitivity and reproducibility. Leveraging signal transduction pathway-responsive cis-regulatory elements, we also demonstrate time- and concentration-dependent genomic recording of WNT, NF-κB and Tet-On activities. By coupling ENGRAM to sequential genome editing via DNA Typewriter, we stably record information about the temporal dynamics of two orthogonal signalling pathways to genomic DNA. Finally we apply ENGRAM to integratively record the transient activity of nearly 100 transcription factor consensus motifs across daily windows spanning the differentiation of mouse embryonic stem cells into gastruloids, an in vitro model of early mammalian development. Although these are proof-of-concept experiments and much work remains to fully realize the possibilities, the symbolic recording of biological signals or states within cells, to the genome and over time, has broad potential to complement contemporary paradigms for how we make measurements in biological systems.

Patterns of transcriptional activity are encoded in our genome through regulatory elements such as promoters or enhancers that, paradoxically, contain similar assortments of sequence-specific transcription factor (TF) binding sites. Knowledge of how these sequence motifs encode multiple, often overlapping, gene expression programs is central to understanding gene regulation and how mutations in non-coding DNA manifest in disease. Here, by studying gene regulation from the perspective of individual transcription start sites (TSSs), using natural genetic variation, perturbation of endogenous TF protein levels and massively parallel analysis of natural and synthetic regulatory elements, we show that the effect of TF binding on transcription initiation is position dependent. Analysing TF-binding-site occurrences relative to the TSS, we identified several motifs with highly preferential positioning. We show that these patterns are a combination of a TF's distinct functional profiles-many TFs, including canonical activators such as NRF1, NFY and Sp1, activate or repress transcription initiation depending on their precise position relative to the TSS. As such, TFs and their spacing collectively guide the site and frequency of transcription initiation. More broadly, these findings reveal how similar assortments of TF binding sites can generate distinct gene regulatory outcomes depending on their spatial configuration and how DNA sequence polymorphisms may contribute to transcription variation and disease and underscore a critical role for TSS data in decoding the regulatory information of our genome.

The binding of multiple transcription factors (TFs) to genomic enhancers drives gene expression in mammalian cells. However, the molecular details that link enhancer sequence to TF binding, promoter state and transcription levels remain unclear. Here we applied single-molecule footprinting to measure the simultaneous occupancy of TFs, nucleosomes and other regulatory proteins on engineered enhancer-promoter constructs with variable numbers of TF binding sites for both a synthetic TF and an endogenous TF involved in the type I interferon response. Although TF binding events on nucleosome-free DNA are independent, activation domains recruit cofactors that destabilize nucleosomes, driving observed TF binding cooperativity. Average TF occupancy linearly determines promoter activity, and we decompose TF strength into separable binding and activation terms. Finally, we develop thermodynamic and kinetic models that quantitatively predict both the enhancer binding microstates and gene expression dynamics. This work provides a template for the quantitative dissection of distinct contributors to gene expression, including TF activation domains, concentration, binding affinity, binding site configuration and recruitment of chromatin regulators.

Cellular identity requires the concerted action of multiple transcription factors (TFs) bound together to enhancers of cell-type-specific genes. Despite TFs recognizing specific DNA motifs within accessible chromatin, this information is insufficient to explain how TFs select enhancers. Here we compared four different TF combinations that induce different cell states, analysing TF genome occupancy, chromatin accessibility, nucleosome positioning and 3D genome organization at the nucleosome resolution. We show that motif recognition on mononucleosomes can decipher only the individual binding of TFs. When bound together, TFs act cooperatively or competitively to target nucleosome arrays with defined 3D organization, displaying motifs in particular patterns. In one combination, motif directionality funnels TF combinatorial binding along chromatin loops, before infiltrating laterally to adjacent enhancers. In other combinations, TFs assemble on motif-dense and highly interconnected loop junctions, and subsequently translocate to nearby lineage-specific sites. We propose a guided-search model in which motif grammar on nucleosome fibres acts as signpost elements, directing TF combinatorial binding to enhancers.

DNA sequence-specific transcription factors (TFs) modulate transcription and chromatin architecture, acting from regulatory sites in enhancers and promoters of eukaryotic genes. How multiple TFs cooperate to regulate individual genes is still unclear. In yeast, most TFs are thought to regulate transcription via binding to upstream activating sequences, which are situated within a few hundred base pairs upstream of the regulated gene. Although this model has been validated for individual TFs and specific genes, it has not been tested in a systematic way. Here we integrated information on the binding and expression targets for the near-complete set of yeast TFs and show that, contrary to expectations, there are few TFs with dedicated activator or repressor roles, and that most TFs have a dual function. Although nearly all protein-coding genes are regulated by one or more TFs, our analysis revealed limited overlap between TF binding and gene regulation. Rapid depletion of many TFs also revealed many regulatory targets that were distant from detectable TF binding sites, suggesting unexpected regulatory mechanisms. Our study provides a comprehensive survey of TF functions and offers insights into interactions between the set of TFs expressed in a single cell type and how they contribute to the complex programme of gene regulation.

In the same way that the mRNA-binding specificities of transfer RNAs define the genetic code, the DNA-binding specificities of transcription factors (TFs) form the molecular basis of the gene regulatory code. The human gene regulatory code is much more complex than the genetic code, in particular because there are more than 1,600 TFs that commonly interact with each other. TF-TF interactions are required for specifying cell fate and executing cell-type-specific transcriptional programs. Despite this, the landscape of interactions between DNA-bound TFs is poorly defined. Here we map the biochemical interactions between DNA-bound TFs using CAP-SELEX, a method that can simultaneously identify individual TF binding preferences, TF-TF interactions and the DNA sequences that are bound by the interacting complexes. A screen of more than 58,000 TF-TF pairs identified 2,198 interacting TF pairs, 1,329 of which preferentially bound to their motifs arranged in a distinct spacing and/or orientation. We also discovered 1,131 TF-TF composite motifs that were markedly different from the motifs of the individual TFs. In total, we estimate that the screen identified between 18% and 47% of all human TF-TF motifs. The novel composite motifs we found were enriched in cell-type-specific elements, active in vivo and more likely to be formed between developmentally co-expressed TFs. Furthermore, TFs that define embryonic axes commonly interacted with different TFs and bound to distinct motifs, explaining how TFs with a similar specificity can define distinct cell types along developmental axes.

Cis-regulatory elements (CREs) control gene expression and are dynamic in their structure and function, reflecting changes in the composition of diverse effector proteins over time. However, methods for measuring the organization of effector proteins at CREs across the genome are limited, hampering efforts to connect CRE structure to their function in cell fate and disease. Here we developed PRINT, a computational method that identifies footprints of DNA-protein interactions from bulk and single-cell chromatin accessibility data across multiple scales of protein size. Using these multiscale footprints, we created the seq2PRINT framework, which uses deep learning to allow precise inference of transcription factor and nucleosome binding and interprets regulatory logic at CREs. Applying seq2PRINT to single-cell chromatin accessibility data from human bone marrow, we observe sequential establishment and widening of CREs centred on pioneer factors across haematopoiesis. We further discover age-associated alterations in the structure of CREs in murine haematopoietic stem cells, including widespread reduction of nucleosome footprints and gain of de novo identified Ets composite motifs. Collectively, we establish a method for obtaining rich insights into DNA-binding protein dynamics from chromatin accessibility data, and reveal the architecture of regulatory elements across differentiation and ageing.

The evidence that signal- and ligand-dependent pathways function by activating regulatory enhancer programs suggests that a 'checkpoint' strategy may underline activation of many diversely regulated enhancers. Here we report a molecular mechanism common to several acute signal- and ligand-dependent enhancer activation programs based on release of a shared enhancer RNA (eRNA) transcription checkpoint. It requires recruitment of a DNA-dependent protein kinase catalytic subunit (DNA-PKcs)-phosphorylated RING finger repressor (Krüppel-associated box)-associated protein 1 (KAP1) as a modulator, inhibiting its association with 7SK and E3 small ubiquitin-like modifier (SUMO) ligase activity on the CDK9 subunit of positive transcription elongation factor b (P-TEFb). This facilitates formation of an activated P-TEFb complex, licensing eRNA elongation. Overcoming this checkpoint for signal-dependent enhancer activation occurs in diverse pathways, including estrogen receptor-α, NF-κB-regulated proinflammatory stimulation, androgen receptor and neuronal depolarization. Therefore, a specific strategy required to convert a basal state enhancer P-TEFb complex to an active state to release a conserved checkpoint is apparently employed by several functionally important signal-regulated regulatory enhancers to implement the instructions of the endocrine and paracrine system.

The human genome contains millions of candidate cis-regulatory elements (cCREs) with cell-type-specific activities that shape both health and many disease states. However, we lack a functional understanding of the sequence features that control the activity and cell-type-specific features of these cCREs. Here we used lentivirus-based massively parallel reporter assays (lentiMPRAs) to test the regulatory activity of more than 680,000 sequences, representing an extensive set of annotated cCREs among three cell types (HepG2, K562 and WTC11), and found that 41.7% of these sequences were active. By testing sequences in both orientations, we find promoters to have strand-orientation biases and their 200-nucleotide cores to function as non-cell-type-specific 'on switches' that provide similar expression levels to their associated gene. By contrast, enhancers have weaker orientation biases, but increased tissue-specific characteristics. Utilizing our lentiMPRA data, we develop sequence-based models to predict cCRE function and variant effects with high accuracy, delineate regulatory motifs and model their combinatorial effects. Testing a lentiMPRA library encompassing 60,000 cCREs in all three cell types further identified factors that determine cell-type specificity. Collectively, our work provides an extensive catalogue of functional CREs in three widely used cell lines and showcases how large-scale functional measurements can be used to dissect regulatory grammar.

Nucleotide changes in gene regulatory elements are important determinants of neuronal development and diseases. Using massively parallel reporter assays in primary human cells from mid-gestation cortex and cerebral organoids, we interrogated the cis-regulatory activity of 102,767 open chromatin regions, including thousands of sequences with cell type-specific accessibility and variants associated with brain gene regulation. In primary cells, we identified 46,802 active enhancer sequences and 164 variants that alter enhancer activity. Activity was comparable in organoids and primary cells, suggesting that organoids provide an adequate model for the developing cortex. Using deep learning we decoded the sequence basis and upstream regulators of enhancer activity. This work establishes a comprehensive catalog of functional gene regulatory elements and variants in human neuronal development.

Most phenotype-associated genetic variants map to noncoding regulatory regions of the human genome, but their mechanisms remain elusive in most cases. We developed a highly efficient strategy, Perturb-multiome, to simultaneously profile chromatin accessibility and gene expression in single cells with CRISPR-mediated perturbation of master transcription factors (TFs). We examined the connection between TFs, accessible regions, and gene expression across the genome throughout hematopoietic differentiation. We discovered that variants within TF-sensitive accessible chromatin regions in erythroid differentiation, although representing <0.3% of the genome, show a ~100-fold enrichment for blood cell phenotype heritability, which is substantially higher than that for other accessible chromatin regions. Our approach facilitates large-scale mechanistic understanding of phenotype-associated genetic variants by connecting key cis-regulatory elements and their target genes within gene regulatory networks.

Sequence-based machine-learning models trained on genomics data improve genetic variant interpretation by providing functional predictions describing their impact on the cis-regulatory code. However, current tools do not predict RNA-seq expression profiles because of modeling challenges. Here, we introduce Borzoi, a model that learns to predict cell-type-specific and tissue-specific RNA-seq coverage from DNA sequence. Using statistics derived from Borzoi's predicted coverage, we isolate and accurately score DNA variant effects across multiple layers of regulation, including transcription, splicing and polyadenylation. Evaluated on quantitative trait loci, Borzoi is competitive with and often outperforms state-of-the-art models trained on individual regulatory functions. By applying attribution methods to the derived statistics, we extract cis-regulatory motifs driving RNA expression and post-transcriptional regulation in normal tissues. The wide availability of RNA-seq data across species, conditions and assays profiling specific aspects of regulation emphasizes the potential of this approach to decipher the mapping from DNA sequence to regulatory function.

We present a gene-level regulatory model, single-cell ATAC + RNA linking (SCARlink), which predicts single-cell gene expression and links enhancers to target genes using multi-ome (scRNA-seq and scATAC-seq co-assay) sequencing data. The approach uses regularized Poisson regression on tile-level accessibility data to jointly model all regulatory effects at a gene locus, avoiding the limitations of pairwise gene-peak correlations and dependence on peak calling. SCARlink outperformed existing gene scoring methods for imputing gene expression from chromatin accessibility across high-coverage multi-ome datasets while giving comparable to improved performance on low-coverage datasets. Shapley value analysis on trained models identified cell-type-specific gene enhancers that are validated by promoter capture Hi-C and are 11× to 15× and 5× to 12× enriched in fine-mapped eQTLs and fine-mapped genome-wide association study (GWAS) variants, respectively. We further show that SCARlink-predicted and observed gene expression vectors provide a robust way to compute a chromatin potential vector field to enable developmental trajectory analysis.

Cell-type-specific regulatory elements such as enhancers can direct expression of recombinant adeno-associated viruses (AAVs) to specific cell types, but this approach is limited by the relatively small packaging capacity of AAVs. In this study, we used spatial genomics to show that transcriptional crosstalk between individual AAV genomes provides a general method for cell-type-specific expression of large cargo by separating distally acting regulatory elements into a second AAV genome. We identified and profiled transcriptional crosstalk in AAV genomes carrying 11 different enhancers active in mouse brain. We developed spatial genomics methods to identify and localize AAV genomes and their concatemeric forms in cultured cells and in tissue, and we demonstrate here that transcriptional crosstalk is dependent upon concatemer formation. Finally, we leveraged transcriptional crosstalk to drive expression of a 3.2-kb Cas9 cargo in a cell-type-specific manner with systemically administered engineered AAVs, and we demonstrate AAV-delivered, minimally invasive, cell-type-specific gene editing in wild-type mice that recapitulates known disease phenotypes.

Thyrotropin (TSH) is the master regulator of thyroid gland growth and function. Resistance to TSH (RTSH) describes conditions with reduced sensitivity to TSH. Dominantly inherited RTSH has been linked to a locus on chromosome 15q, but its genetic basis has remained elusive. Here we show that non-coding mutations in a (TTTG) short tandem repeat (STR) underlie dominantly inherited RTSH in all 82 affected participants from 12 unrelated families. The STR is contained in a primate-specific Alu retrotransposon with thyroid-specific cis-regulatory chromatin features. Fiber-seq and RNA-seq studies revealed that the mutant STR activates a thyroid-specific enhancer cluster, leading to haplotype-specific upregulation of the bicistronic MIR7-2/MIR1179 locus 35 kb downstream and overexpression of its microRNA products in the participants' thyrocytes. An imbalance in signaling pathways targeted by these micro-RNAs provides a working model for this cause of RTSH. This finding broadens our current knowledge of genetic defects altering pituitary-thyroid feedback regulation.

Gene expression is controlled by dynamic localization of thousands of regulatory proteins to precise genomic regions. Understanding this cell type-specific process has been a longstanding goal yet remains challenging because DNA-protein mapping methods generally study one protein at a time. Here, to address this, we developed chromatin immunoprecipitation done in parallel (ChIP-DIP) to generate genome-wide maps of hundreds of diverse regulatory proteins in a single experiment. ChIP-DIP produces highly accurate maps within large pools (>160 proteins) for all classes of DNA-associated proteins, including modified histones, chromatin regulators and transcription factors and across multiple conditions simultaneously. First, we used ChIP-DIP to measure temporal chromatin dynamics in primary dendritic cells following LPS stimulation. Next, we explored quantitative combinations of histone modifications that define distinct classes of regulatory elements and characterized their functional activity in human and mouse cell lines. Overall, ChIP-DIP generates context-specific protein localization maps at consortium scale within any molecular biology laboratory and experimental system.

Transcription factors (TFs) can define distinct cellular identities despite nearly identical DNA-binding specificities. One mechanism for achieving regulatory specificity is DNA-guided TF cooperativity. Although in vitro studies suggest that it may be common, examples of such cooperativity remain scarce in cellular contexts. Here, we demonstrate how "Coordinator," a long DNA motif composed of common motifs bound by many basic helix-loop-helix (bHLH) and homeodomain (HD) TFs, uniquely defines the regulatory regions of embryonic face and limb mesenchyme. Coordinator guides cooperative and selective binding between the bHLH family mesenchymal regulator TWIST1 and a collective of HD factors associated with regional identities in the face and limb. TWIST1 is required for HD binding and open chromatin at Coordinator sites, whereas HD factors stabilize TWIST1 occupancy at Coordinator and titrate it away from HD-independent sites. This cooperativity results in the shared regulation of genes involved in cell-type and positional identities and ultimately shapes facial morphology and evolution.

Convergent transcription, that is, the collision of sense and antisense transcription, is ubiquitous in mammalian genomes and believed to diminish RNA expression. Recently, antisense transcription downstream of promoters was found to be surprisingly prevalent. However, functional characteristics of affected promoters are poorly investigated. Here we show that convergent transcription marks an unexpected positively co-regulated promoter constellation. By assessing transcriptional dynamic systems, we identified co-regulated constituent promoters connected through a distinct chromatin structure. Within these cis-regulatory domains, transcription factors can regulate both constituting promoters by binding to only one of them. Convergent promoters comprise about a quarter of all active transcript start sites and initiate 5'-overlapping antisense RNAs-an RNA class believed previously to be rare. Visualization of nascent RNA molecules reveals convergent cotranscription at these loci. Together, our results demonstrate that co-regulated convergent promoters substantially expand the cis-regulatory repertoire, reveal limitations of the transcription interference model and call for adjusting the promoter concept.

How enhancers control target gene expression over long genomic distances remains an important unsolved problem. Here we investigated enhancer-promoter communication by integrating data from nucleosome-resolution genomic contact maps, nascent transcription and perturbations affecting either RNA polymerase II (Pol II) dynamics or the activity of thousands of candidate enhancers. Integration of new Micro-C experiments with published CRISPRi data demonstrated that enhancers spend more time in close proximity to their target promoters in functional enhancer-promoter pairs compared to nonfunctional pairs, which can be attributed in part to factors unrelated to genomic position. Manipulation of the transcription cycle demonstrated a key role for Pol II in enhancer-promoter interactions. Notably, promoter-proximal paused Pol II itself partially stabilized interactions. We propose an updated model in which elements of transcriptional dynamics shape the duration or frequency of interactions to facilitate enhancer-promoter communication.

Human accelerated regions (HARs) are conserved genomic loci that have experienced rapid nucleotide substitutions following the divergence from chimpanzees. HARs are enriched in candidate regulatory regions near neurodevelopmental genes, suggesting their roles in gene regulation. However, their target genes and functional contributions to human brain development remain largely uncharacterized. Here we elucidate the cis-regulatory functions of HARs in human and chimpanzee induced pluripotent stem (iPS) cell-induced excitatory neurons. Using genomic and chromatin looping information, we prioritized 20 HARs and their chimpanzee orthologues for functional characterization via single-cell CRISPR interference, and demonstrated their species-specific gene regulatory functions. Our findings reveal diverse functional outcomes of HAR-mediated cis-regulation in human neurons, including attenuated NPAS3 expression by altering the binding affinities of multiple transcription factors in HAR202 and maintaining iPS cell pluripotency and neuronal differentiation capacities through the upregulation of PUM2 by 2xHAR.319. Finally, we used prime editing to demonstrate differential enhancer activity caused by several HAR26;2xHAR.178 variants. In particular, we link one variant in HAR26;2xHAR.178 to elevated SOCS2 expression and increased neurite outgrowth in human neurons. Thus, our study sheds new light on the endogenous gene regulatory functions of HARs and their potential contribution to human brain evolution.

A central question in biology is how macromolecular machines function cooperatively. In bacteria, transcription and translation occur in the same cellular compartment, and can be physically and functionally coupled. Although high-resolution structures of the ribosome-RNA polymerase (RNAP) complex have provided initial mechanistic insights into the coupling process, we lack knowledge of how these structural snapshots are placed along a dynamic reaction trajectory. Here we reconstitute a complete and active transcription-translation system and develop multi-colour single-molecule fluorescence microscopy experiments to directly and simultaneously track transcription elongation, translation elongation and the physical and functional coupling between the ribosome and the RNAP in real time. Our data show that physical coupling between ribosome and RNAP can occur over hundreds of nucleotides of intervening mRNA by mRNA looping, a process facilitated by NusG. We detect active transcription elongation during mRNA looping and show that NusA-paused RNAPs can be activated by the ribosome by long-range physical coupling. Conversely, the ribosome slows down while colliding with the RNAP. We hereby provide an alternative explanation for how the ribosome can efficiently rescue RNAP from frequent pausing without requiring collisions by a closely trailing ribosome. Overall, our dynamic data mechanistically highlight an example of how two central macromolecular machineries, the ribosome and RNAP, can physically and functionally cooperate to optimize gene expression.

Enhancers control gene expression and have crucial roles in development and homeostasis. However, the targeted de novo design of enhancers with tissue-specific activities has remained challenging. Here we combine deep learning and transfer learning to design tissue-specific enhancers for five tissues in the Drosophila melanogaster embryo: the central nervous system, epidermis, gut, muscle and brain. We first train convolutional neural networks using genome-wide single-cell assay for transposase-accessible chromatin with sequencing (ATAC-seq) datasets and then fine-tune the convolutional neural networks with smaller-scale data from in vivo enhancer activity assays, yielding models with 13% to 76% positive predictive value according to cross-validation. We designed and experimentally assessed 40 synthetic enhancers (8 per tissue) in vivo, of which 31 (78%) were active and 27 (68%) functioned in the target tissue (100% for central nervous system and muscle). The strategy of combining genome-wide and small-scale functional datasets by transfer learning is generally applicable and should enable the design of tissue-, cell type- and cell state-specific enhancers in any system.

Sickle cell disease (SCD) is a prevalent, life-threatening condition attributable to a heritable mutation in β-hemoglobin. Therapeutic induction of fetal hemoglobin (HbF) can ameliorate disease complications and has been intently pursued. However, safe and effective small-molecule inducers of HbF remain elusive. We report the discovery of dWIZ-1 and dWIZ-2, molecular glue degraders of the WIZ transcription factor that robustly induce HbF in erythroblasts. Phenotypic screening of a cereblon (CRBN)-biased chemical library revealed WIZ as a previously unknown repressor of HbF. WIZ degradation is mediated by recruitment of WIZ(ZF7) to CRBN by dWIZ-1, as resolved by crystallography of the ternary complex. Pharmacological degradation of WIZ was well tolerated and induced HbF in humanized mice and cynomolgus monkeys. These findings establish WIZ degradation as a globally accessible therapeutic strategy for SCD.

The POU2F3-POU2AF2/3 transcription factor complex is the master regulator of the tuft cell lineage and tuft cell-like small cell lung cancer (SCLC). Here, we identify a specific dependence of the POU2F3 molecular subtype of SCLC (SCLC-P) on the activity of the mammalian switch/sucrose non-fermentable (mSWI/SNF) chromatin remodeling complex. Treatment of SCLC-P cells with a proteolysis targeting chimera (PROTAC) degrader of mSWI/SNF ATPases evicts POU2F3 and its coactivators from chromatin and attenuates downstream signaling. B cell malignancies which are dependent on the POU2F1/2 cofactor, POU2AF1, are also sensitive to mSWI/SNF ATPase degraders, with treatment leading to chromatin eviction of POU2AF1 and IRF4 and decreased IRF4 signaling in multiple myeloma cells. An orally bioavailable mSWI/SNF ATPase degrader significantly inhibits tumor growth in preclinical models of SCLC-P and multiple myeloma without signs of toxicity. This study suggests that POU2F-POU2AF-driven malignancies have an intrinsic dependence on the mSWI/SNF complex, representing a therapeutic vulnerability.

The contribution of rare noncoding genetic variation to common phenotypes is largely unknown, as a result of a historical lack of population-scale whole-genome sequencing data and the difficulty of categorizing noncoding variants into functionally similar groups. To begin addressing these challenges, we performed a cis association analysis using whole-genome sequencing data, consisting of 1.1 billion variants, 123 million noncoding aggregate-based tests and 2,907 circulating protein levels in ~50,000 UK Biobank participants. We identified 604 independent rare noncoding single-variant associations with circulating protein levels. Unlike protein-coding variation, rare noncoding genetic variation was almost as likely to increase or decrease protein levels. Rare noncoding aggregate testing identified 357 conditionally independent associated regions. Of these, 74 (21%) were not detectable by single-variant testing alone. Our findings have important implications for the identification, and role, of rare noncoding genetic variation associated with common human phenotypes, including the importance of testing aggregates of noncoding variants.

Age-associated B cells (ABCs) accumulate during infection, aging, and autoimmunity, contributing to lupus pathogenesis. In this study, we screened for transcription factors driving ABC formation and found that zinc finger E-box binding homeobox 2 (ZEB2) is required for human and mouse ABC differentiation in vitro. ABCs are reduced in haploinsufficient individuals and in mice lacking in B cells. In mice with toll-like receptor 7 (TLR7)-driven lupus, ZEB2 is essential for ABC formation and autoimmune pathology. ZEB2 binds to +20-kb myocyte enhancer factor 2b ()'s intronic enhancer, repressing MEF2B-mediated germinal center B cell differentiation and promoting ABC formation. ZEB2 also targets genes important for ABC specification and function, including . ZEB2-driven ABC differentiation requires JAK-STAT (Janus kinase-signal transducer and activator of transcription), and treatment with JAK1/3 inhibitor reduces ABC accumulation in autoimmune mice and patients. Thus, ZEB2 emerges as a driver of B cell autoimmunity.

Transcriptional regulation, which involves a complex interplay between regulatory sequences and proteins, directs all biological processes. Computational models of transcription lack generalizability to accurately extrapolate to unseen cell types and conditions. Here we introduce GET (general expression transformer), an interpretable foundation model designed to uncover regulatory grammars across 213 human fetal and adult cell types. Relying exclusively on chromatin accessibility data and sequence information, GET achieves experimental-level accuracy in predicting gene expression even in previously unseen cell types. GET also shows remarkable adaptability across new sequencing platforms and assays, enabling regulatory inference across a broad range of cell types and conditions, and uncovers universal and cell-type-specific transcription factor interaction networks. We evaluated its performance in prediction of regulatory activity, inference of regulatory elements and regulators, and identification of physical interactions between transcription factors and found that it outperforms current models in predicting lentivirus-based massively parallel reporter assay readout. In fetal erythroblasts, we identified distal (greater than 1 Mbp) regulatory regions that were missed by previous models, and, in B cells, we identified a lymphocyte-specific transcription factor-transcription factor interaction that explains the functional significance of a leukaemia risk predisposing germline mutation. In sum, we provide a generalizable and accurate model for transcription together with catalogues of gene regulation and transcription factor interactions, all with cell type specificity.

We set out to exhaustively characterize the impact of the cis-chromatin environment on prime editing, a precise genome engineering tool. Using a highly sensitive method for mapping the genomic locations of randomly integrated reporters, we discover massive position effects, exemplified by editing efficiencies ranging from ∼0% to 94% for an identical target site and edit. Position effects on prime editing efficiency are well predicted by chromatin marks, e.g., positively by H3K79me2 and negatively by H3K9me3. Next, we developed a multiplex perturbational framework to assess the interaction of trans-acting factors with the cis-chromatin environment on editing outcomes. Applying this framework to DNA repair factors, we identify HLTF as a context-dependent repressor of prime editing. Finally, several lines of evidence suggest that active transcriptional elongation enhances prime editing. Consistent with this, we show we can robustly decrease or increase the efficiency of prime editing by preceding it with CRISPR-mediated silencing or activation, respectively.

Transcription factor (TF) DNA-binding dynamics govern cell fate and identity. However, our ability to pharmacologically control TF localization is limited. Here we leverage chemically driven binding site restriction leading to robust and DNA-sequence-specific redistribution of PU.1, a pioneer TF pertinent to many hematopoietic malignancies. Through an innovative technique, 'CLICK-on-CUT&Tag', we characterize the hierarchy of de novo PU.1 motifs, predicting occupancy in the PU.1 cistrome under binding site restriction. Temporal and single-molecule studies of binding site restriction uncover the pioneering dynamics of native PU.1 and identify the paradoxical activation of an alternate target gene set driven by PU.1 localization to second-tier binding sites. These transcriptional changes were corroborated by genetic blockade and site-specific reporter assays. Binding site restriction and subsequent PU.1 network rewiring causes primary human leukemia cells to differentiate. In summary, pharmacologically induced TF redistribution can be harnessed to govern TF localization, actuate alternate gene networks and direct cell fate.

Transposon-encoded tnpB and iscB genes encode RNA-guided DNA nucleases that promote their own selfish spread through targeted DNA cleavage and homologous recombination. These widespread gene families were repeatedly domesticated over evolutionary timescales, leading to the emergence of diverse CRISPR-associated nucleases including Cas9 and Cas12 (refs. ). We set out to test the hypothesis that TnpB nucleases may have also been repurposed for novel, unexpected functions other than CRISPR-Cas adaptive immunity. Here, using phylogenetics, structural predictions, comparative genomics and functional assays, we uncover multiple independent genesis events of programmable transcription factors, which we name TnpB-like nuclease-dead repressors (TldRs). These proteins use naturally occurring guide RNAs to specifically target conserved promoter regions of the genome, leading to potent gene repression in a mechanism akin to CRISPR interference technologies invented by humans. Focusing on a TldR clade found broadly in Enterobacteriaceae, we discover that bacteriophages exploit the combined action of TldR and an adjacently encoded phage gene to alter the expression and composition of the host flagellar assembly, a transformation with the potential to impact motility, phage susceptibility, and host immunity. Collectively, this work showcases the diverse molecular innovations that were enabled through repeated exaptation of transposon-encoded genes, and reveals the evolutionary trajectory of diverse RNA-guided transcription factors.

Methods to characterize the functional effects of genetic variants of uncertain significance (VUSs) have been limited by incomplete coverage of the mutational space. In clinical oncology, drug resistance arising from VUSs can prevent optimal treatment. Here we introduce PEER-seq, a high-throughput method based on prime editing that can evaluate the functional effects of single-nucleotide variants (SNVs). PEER-seq introduces both intended SNVs and synonymous marker mutations using prime editing and deep sequences the endogenous target regions to identify the introduced SNVs. We generate and functionally evaluate 2,476 SNVs in the epidermal growth factor receptor gene (EGFR), including 99% of all possible variants in the canonical tyrosine kinase domain. We determined resistance profiles of 95% of all possible EGFR protein variants encoded in the whole tyrosine kinase domain against the common tyrosine kinase inhibitors afatinib, osimertinib and osimertinib in the presence of the co-occurring substitution T790M, in PC-9 cells. Our study has the potential to substantially improve the precision of therapeutic choices in clinical settings.

Many enhancers control gene expression by assembling regulatory factor clusters, also referred to as condensates. This process is vital for facilitating enhancer communication and establishing cellular identity. However, how DNA sequence and transcription factor (TF) binding instruct the formation of high regulatory factor environments remains poorly understood. Here we developed a new approach leveraging enhancer-centric chromatin accessibility quantitative trait loci (caQTLs) to nominate regulatory factor clusters genome-wide. By analyzing TF-binding signatures within the context of caQTLs and comparing episomal versus endogenous enhancer activities, we discovered a class of regulators, 'context-only' TFs, that amplify the activity of cell type-specific caQTL-binding TFs, that is, 'context-initiator' TFs. Similar to super-enhancers, enhancers enriched for context-only TF-binding sites display high coactivator binding and sensitivity to bromodomain-inhibiting molecules. We further show that binding sites for context-only and context-initiator TFs underlie enhancer coordination, providing a mechanistic rationale for how a loose TF syntax confers regulatory specificity.