Promoter-proximal pausing is a widespread checkpoint in gene expression where RNA polymerase II (the enzyme that copies DNA into RNA) starts transcribing a gene but then stalls about 30 to 100 base pairs downstream of the start site. Rather than racing straight through a gene, the polymerase halts, held in place by specific protein complexes, and waits for a signal to continue. This seemingly wasteful idling turns out to be one of the most important regulatory steps in animal cells, affecting everything from embryonic development to cancer progression to how viruses commandeer our cellular machinery.
How the Pause Gets Established
Once RNA polymerase II (Pol II) begins copying a gene, it does not get very far before two protein complexes clamp down and freeze it in place. One is called DSIF, and the other is NELF (negative elongation factor). Together, they convert the polymerase from an active copying machine into something more like a loaded spring under tension. Structural studies using cryo-electron microscopy have revealed how NELF accomplishes this at the atomic level: it tilts the DNA-RNA hybrid inside the polymerase into an unfavorable angle, blocks the entry site where new building blocks would normally arrive, and physically restrains moving parts of the enzyme that would otherwise let it advance.1PubMed Central. Structure of paused transcription complex Pol II-DSIF-NELF NELF also prevents a helper factor called TFIIS from binding, which would otherwise help the polymerase get unstuck.
The association of DSIF and NELF with the polymerase is not instantaneous. Work in fruit flies showed that these factors need the polymerase to have already produced a short stretch of RNA, at least 18 nucleotides long, before they can latch on. The larger subunit of DSIF physically contacts this emerging RNA strand, and once it has a grip, it recruits NELF to complete the pause.2PubMed Central. Interactions between DSIF, NELF, and the Drosophila RNA polymerase II transcription elongation complex So the pause is not a random accident. It is a precisely staged event that only engages once the polymerase has begun transcribing but before it gets very far into the gene body.
What Releases the Pause
A paused polymerase sitting near a promoter is poised for action but going nowhere until it receives a green light. That signal comes from a kinase called P-TEFb, which adds phosphate groups to both the polymerase and DSIF, effectively flipping them from a braking configuration into a go configuration. NELF falls away, DSIF switches from a repressor into an activator of elongation, and the polymerase accelerates into the gene body to produce a full-length RNA.3PubMed Central. P-TEFb: Finding its ways to release promoter-proximally paused RNA polymerase II
P-TEFb itself is tightly controlled. Much of the P-TEFb in a cell is sequestered in an inactive form, locked up in a complex with a small RNA called 7SK. Cells can recruit either the active or the inactive form to promoter regions, depending on the context. This means the decision of whether and when to release the pause is layered with multiple controls, not a simple on/off switch.
DNA Features That Encourage Pausing
Not every gene pauses to the same degree. The DNA sequence around the start of a gene helps determine how pronounced the pause will be. Genes that show strong pausing tend to be unusually rich in the nucleotide G (guanine) in the region flanking the transcription start site. This G-richness is partly explained by binding sites for common transcription factors upstream of the start, but downstream the explanation is different: G-rich sequences can fold into unusual four-stranded structures called G-quadruplexes. A study of human genes found that pausing correlates with the presence of G-quadruplex-forming motifs at the beginning of the first intron, on the strand not being used as a template.4PubMed Central. G4 motifs correlate with promoter-proximal transcriptional pausing in human genes
Beyond G-quadruplexes, the broader tendency of promoter-proximal DNA to form alternative secondary structures also matters. Genes with paused polymerases have promoter regions that are energetically more favorable for forming these non-standard DNA shapes than non-paused genes. Structure-probing experiments confirmed that such structures actually exist in living human cells, and manipulating them in a test tube at a well-studied heat shock gene directly altered how the polymerase paused.5PubMed Central. Alternative DNA secondary structure formation affects RNA polymerase II promoter-proximal pausing in human DNA shape, in other words, is not just a backdrop; it actively participates in setting the pause.
The Nucleosome Barrier
There is also a physical obstacle. Just downstream of most promoters sits a well-positioned protein-DNA structure called the +1 nucleosome, where DNA wraps around a core of histone proteins. The polymerase, already slowed by NELF and DSIF, collides with this nucleosome. Cryo-electron microscopy structures of elongation complexes stalled on nucleosomal DNA show that without help from TFIIS, the polymerase backtracks and arrests because NELF and the nucleosome physically conflict with one another. When TFIIS is present, the polymerase can advance far enough to partially unwrap the nucleosome, establishing a more stable paused state, but it still cannot easily push past.6PubMed Central. Structural insights into promoter-proximal pausing of RNA polymerase II at +1 nucleosome The +1 nucleosome thus acts as a second fence reinforcing the pause, working alongside the protein factors to keep the polymerase in check.
A Built-In Quality Control Step
Pausing is not just about holding the polymerase until the right activation signal arrives. It also serves as a checkpoint to make sure the transcription machinery has been properly assembled. The newly made RNA needs a chemical cap added to its front end, and the elongation complex needs to be fully loaded with the right cofactors before it commits to copying an entire gene, which can be hundreds of thousands of base pairs long. Recent work has identified two molecular enforcers of this quality control: one called Integrator, which can dissociate and recycle improperly assembled complexes, and an enzyme called CRL3-ARMC5, which tags defective polymerases for destruction.7Molecular Cell. Promoter Proximal Pausing: A Nexus of Gene Regulation
Backtracking is a related phenomenon. Sometimes a paused polymerase slides backward along the DNA, losing its grip on the growing RNA strand. This is surprisingly common within the first 2,000 base pairs after the start site and at gene ends where transcription is shutting down. The factor TFIIS rescues these backtracked complexes by triggering the polymerase’s built-in ability to cut away the misaligned RNA and try again. When researchers blocked this rescue mechanism using a dominant-negative version of TFIIS, paused polymerases lost their ability to resume elongation, and stress-responsive genes became harder to activate.8PubMed Central. Widespread Backtracking by RNA Pol II Is a Major Effector of Gene Activation, 5′ Pause Release, Termination, and Transcription Elongation Rate The pause, then, is not a single clean stop but a dynamic process of forward and backward movements, with quality-control machinery constantly testing whether the polymerase is ready to proceed.
Why Bother Pausing at All
From the cell’s perspective, pausing solves two practical problems. The first is speed. A gene whose promoter already has a loaded, paused polymerase sitting on it can be switched on almost instantly, because the slow initial steps of assembling the transcription machinery have already been completed. All that is needed is the release signal. The second is synchrony: when a developmental signal sweeps across a tissue, cells need to turn on the same gene at roughly the same time. If every cell had to start from scratch, assembling the transcription machinery on an empty promoter, each cell would activate the gene at a slightly different moment, and collective cellular behavior would fall out of coordination.
Computational modeling has shown that regulating transcription at the elongation step (through pause release) inherently produces faster, more synchronous gene activation than regulating it at the initiation step (loading a polymerase onto the promoter in the first place), regardless of the specific rate parameters involved.9PLoS Computational Biology. Transcriptional Regulation: Effects of Promoter Proximal Pausing on Speed, Synchrony and Reliability The biological reality matches. In fruit fly embryos, replacing the promoter of a gene critical for mesoderm formation (snail) with a promoter from a non-paused gene resulted in less synchronous gene activation across cells and a progressive failure of the coordinated tissue folding that forms the mesoderm during gastrulation.10Cell. Promoter Proximal Pausing: A Nexus of Gene Regulation Pausing, in short, is what lets embryonic development proceed with precision rather than chaos.
Rapid Stress Responses
The heat shock response is the textbook example of pausing in action. When mammalian cells experience a sudden temperature spike or other proteotoxic stress, they need to ramp up production of protective proteins within minutes. The master regulator of this response, HSF1, turns out to work primarily by releasing paused polymerases into productive elongation rather than by loading new polymerases onto promoters. Genome-wide analyses showed that HSF1 controls only a fraction of the genes that change during heat shock, but it does so specifically by triggering pause release, likely through recruiting P-TEFb.11PubMed Central. Mammalian Heat Shock Response and Mechanisms Underlying Its Genome-wide Transcriptional Regulation
The broader transcriptional response to heat shock, including genes not controlled by HSF1, is also largely modulated at the level of pause release. And this system has a memory: a single heat exposure in embryonic fibroblasts primed their daughter cells to respond faster to a subsequent heat shock. The mechanism involved an increase in promoter-proximal pausing at heat-responsive genes between stress episodes, so that more polymerases were loaded and ready to fire when the next stress hit, combined with faster pause release once it did.12PubMed Central. Stress-induced transcriptional memory accelerates promoter-proximal pause release and decelerates termination over mitotic divisions Cells were essentially cocking more springs and releasing them more quickly.
Enhancer RNAs as Molecular Decoys
Enhancers are stretches of DNA, sometimes far from the genes they regulate, that boost transcription. Many enhancers are themselves transcribed into short RNAs called enhancer RNAs, or eRNAs, whose function was mysterious for years. One answer turns out to involve pause release. In neurons, when immediate early genes (the kind that respond within minutes to a stimulus like a burst of neural activity) are activated, the eRNAs produced from nearby enhancers act as decoys that lure NELF away from the paused polymerase. By sequestering the brake, the eRNAs help the polymerase transition into productive elongation.13PubMed Central. Enhancer RNA facilitates NELF release from immediate early genes This is a satisfying connection: the enhancers are not just signaling abstractly but physically removing the factor that holds transcription back.
Genome Stability and R-Loops
Paused polymerases come with a potential hazard. When RNA polymerase stalls, the newly made RNA strand can thread back into the DNA double helix, displacing one DNA strand and forming a three-stranded structure called an R-loop. R-loops expose single-stranded DNA, which is fragile and vulnerable to breakage. There is a strong positive correlation between genes with high pausing and genes that accumulate R-loops at their promoters.14Nature Communications. Attenuation of RNA polymerase II pausing mitigates BRCA1-associated R-loop accumulation and tumorigenesis
Cells have surveillance machinery to manage this risk. One such system involves a complex called SOSS-INTAC, which recognizes the single-stranded DNA exposed by R-loops and stimulates the termination of paused transcription, clearing the polymerase before R-loop-induced damage accumulates. When the key component of this system, SSB1, is deleted or carries cancer-associated mutations, R-loops build up genome-wide.15PubMed Central. R-loop-dependent promoter-proximal termination ensures genome stability Separately, the tumor suppressor BRCA1, well-known for its role in breast cancer susceptibility, partners with an RNA-DNA helicase called SETX to resolve R-loops at transcription termination sites. Loss of BRCA1 leads to accumulation of R-loops and DNA damage at these sites.16Molecular Cell. BRCA1/SETX Complex Resolves R-Loops and Suppresses Transcription-Associated DNA Damage at Human Termination Regions The paused polymerase, then, is a double-edged sword: essential for regulation, but a source of genomic vulnerability if its R-loop byproducts are not kept in check.
How Cancer Hijacks the Pause
The oncogene MYC (often written c-Myc) is the most frequently amplified gene in human tumors, and its connection to pausing is direct. MYC protein promotes the expression of growth and proliferation genes by recruiting P-TEFb to their promoters, driving the release of paused polymerases into productive elongation.17PubMed Central. c-Myc regulates transcriptional pause release When MYC is overexpressed, as it frequently is in cancers, the pause-release machinery at its target genes goes into overdrive.
In prostate cancer, MYC overexpression creates an additional twist: it increases pausing at genes controlled by the androgen receptor, effectively suppressing the normal androgen-driven transcriptional program. This disruption of pause release at androgen receptor targets appears to contribute to tumor initiation and progression.18PubMed Central. MYC drives aggressive prostate cancer by disrupting transcriptional pause release at androgen receptor targets In liver cancer, CDK9, the catalytic subunit of P-TEFb, is required for MYC to exert its transcriptional effects. Inhibiting CDK9 pharmacologically or with genetic tools produced strong anti-tumor effects that scaled with how much MYC a tumor expressed.19Genes & Development. CDK9-mediated transcription elongation is required for MYC addiction in hepatocellular carcinoma
A fusion protein found in a childhood muscle cancer called alveolar rhabdomyosarcoma, PAX3-FOXO1, also operates through pause release. When this oncoprotein is degraded experimentally, its target genes show a clear signature: polymerase density in the gene body drops while paused polymerase accumulates at the promoter, meaning the polymerases are being loaded but not released.20Molecular Cell. PAX3-FOXO1 regulates RNA polymerase II pause release in alveolar rhabdomyosarcoma These examples collectively paint a picture of pause release as a control point that multiple oncogenic pathways converge on.
HIV and the Tat Connection
The human immunodeficiency virus offers a particularly clear illustration of how pause release can be exploited. After HIV integrates into the host genome, its own transcription faces the same pausing checkpoint as any cellular gene. Without help, Pol II copies only short transcripts from the viral promoter and then stalls. The viral protein Tat solves this problem by recruiting P-TEFb directly to the stalled polymerase via a structured RNA element called TAR that forms at the beginning of the viral transcript.21PubMed Central. HIV Tat/P-TEFb Interaction: A Potential Target for Novel Anti-HIV Therapies Before Tat accumulates, host cell activation pathways (including NF-κB) provide some basal P-TEFb recruitment to the viral promoter, but Tat massively amplifies the process.22Molecular Cell. Promoter Proximal Pausing: A Nexus of Gene Regulation This makes the Tat-P-TEFb interaction an appealing drug target: disrupting it could prevent viral transcription without directly attacking the reverse transcriptase or protease steps that current antiretrovirals focus on.
The Evolutionary Origins of Pausing
Promoter-proximal pausing as we see it in flies and mammals is not universal across life. Comparative studies spanning species across the tree of life found that unicellular eukaryotes, such as yeasts, do show some accumulation of polymerase near their transcription start sites, but it looks more like a gentle slowdown than a true pause. This “proto-paused” state appears to have transitioned into the longer-lived, sharply focused pause seen in animals. The transition coincided with the evolution of new subunits in NELF and the 7SK complex, the very molecules that stabilize the pause and regulate P-TEFb availability.23PubMed Central. Evolution of promoter-proximal pausing enabled a new layer of transcription control In other words, as multicellular animals became more complex and needed tighter, faster, and more coordinated gene control, evolution built out the molecular toolkit for pausing. It was not invented from scratch but elaborated from something simpler that already existed.
This evolutionary perspective helps explain why pausing is so pervasive in animal genomes. Genome-wide mapping using precision run-on sequencing (PRO-seq) consistently shows polymerase piling up 30 to 100 base pairs downstream of the start site across metazoan species.24PubMed Central. Evolution of promoter-proximal pausing enabled a new layer of transcription control Earlier work developing PRO-seq had already established that Pol II accumulates not only at promoters but also at intron-exon junctions used for splicing and at the ends of genes where the RNA gets processed, suggesting that polymerase pausing and slowdown serve regulatory roles at multiple points along a gene, not just the promoter.25PubMed Central. Precise maps of RNA polymerase reveal how promoters direct initiation and pausing
Measuring Pausing With New Precision
Much of what we know about pausing comes from sequencing-based methods that capture a snapshot of where polymerases sit at a given moment. But a snapshot does not tell you how long a polymerase has been sitting there. A method called TV-PRO-seq (time-variant PRO-seq) addressed this by estimating genome-wide pausing durations at single-base resolution. Applied to human cells, it produced a counterintuitive finding: polymerases near promoters pause more frequently than those in other regions of the genome, but each individual pause event is actually shorter than pauses elsewhere.26PubMed Central. Timing RNA polymerase pausing with TV-PRO-seq The large pileup of polymerase at promoters, then, reflects a high frequency of short pauses rather than a single long stall, a distinction that matters for understanding how quickly the system can respond to signals.
Drug Development Targeting CDK9
Because pause release through P-TEFb is a bottleneck that multiple cancer pathways depend on, the catalytic component of P-TEFb, the kinase CDK9, has become a drug target. One compound, AZD4573, was designed as a potent and selective CDK9 inhibitor with a deliberately short duration of action. Brief pulses of CDK9 inhibition rapidly reduced the phosphorylation marks on Pol II that signal active elongation and simultaneously collapsed levels of Mcl-1, a survival protein that many blood cancers depend on. This led to cancer cell death across multiple types of blood cancer cell lines.27PubMed. Discovery of AZD4573, a Potent and Selective Inhibitor of CDK9 That Enables Short Duration of Target Engagement for the Treatment of Hematological Malignancies The rationale for keeping the drug exposure brief is practical: CDK9 is needed for normal transcription everywhere in the body, so a drug that blocked it continuously would be toxic to healthy cells. By engineering a compound that hits CDK9 hard but clears quickly, the therapeutic window opens enough to kill cancer cells dependent on high CDK9 activity while sparing normal tissue that can tolerate a transient dip.
The connection between MYC levels and sensitivity to CDK9 inhibition makes these drugs especially interesting for tumors driven by MYC amplification, where conventional targeted therapies have struggled. MYC itself has long been considered “undruggable” because it lacks the kind of deep binding pocket that small molecules typically need to grab onto. Going after the pause-release machinery that MYC depends on is an indirect but potentially effective workaround.