# Chapter 24. Transcription Elongation, Pausing, Termination, and Promoter-Proximal Decisions

## Scope Note

Transcription does not become biologically uninteresting after initiation. Once a polymerase leaves a promoter, the enzyme must hold a nucleic-acid scaffold, add ribonucleotides, move along a template, respond to obstacles, recruit factors, coordinate nascent RNA processing, and eventually stop. This chapter treats the elongation phase of transcription as a regulated sequence of molecular decisions. The main objects are the transcription elongation complex, the paused or backtracked states that interrupt nucleotide addition, the promoter-proximal checkpoint used by many metazoan RNA polymerase II genes, and the termination pathways that release or process transcripts. Polymerase architecture and catalytic chemistry are introduced in Chapters [20](chapter1019.md), [21](chapter1020.md), [22](chapter1021.md), and [23](chapter1022.md). Co-transcriptional RNA folding, capping, splicing, modification, and ribonucleoprotein assembly are developed in [Chapter 25](chapter1024.md).

## Executive Summary

An elongation complex is a polymerase-template-product assembly that remains bound while the polymerase adds nucleotides to a growing RNA. The complex contains an active site, a transcription bubble or analogous template-product scaffold, a short hybrid between the product RNA and template, channels for incoming nucleotides and exiting RNA, and surfaces for regulatory factors. Elongation complexes are processive, meaning that one polymerase can add many nucleotides before dissociating. Processivity is not the same as constant speed. A processive polymerase can pause, backtrack, arrest, recover, terminate, or be released into faster productive elongation. Structural and biochemical work now makes clear that elongation is a state landscape rather than a simple conveyor belt.

Promoter-proximal pausing is one of the best-studied post-initiation decisions in gene regulation. In many metazoan protein-coding genes and noncoding transcription units, RNA polymerase II initiates, synthesizes a short RNA, and pauses tens of nucleotides downstream of the transcription start region. DSIF and NELF stabilize the paused complex, while release into productive elongation commonly requires phosphorylation by P-TEFb and remodeling of the paused polymerase-factor assembly. This checkpoint allows a cell to regulate transcript production after polymerase recruitment and initiation, which is useful for genes that must respond rapidly to developmental, stress, immune, or signaling cues.

Termination is the transition in which transcription stops and the elongation complex is resolved. Bacteria use intrinsic termination, in which an RNA structure and weak RNA-DNA hybrid destabilize the complex, and Rho-dependent termination, in which an ATP-dependent RNA translocase catches a susceptible polymerase. Archaea possess factor-dependent termination mechanisms that should not be reduced to either bacterial or eukaryotic stereotypes. Eukaryotic RNA polymerase II termination is coupled to RNA processing, cleavage and polyadenylation, torpedo-like exonuclease action, allosteric changes in the elongation complex, chromatin context, Integrator action at selected loci, and promoter-proximal termination pathways.

Elongation also shapes genome maintenance. A moving or stalled polymerase can signal transcription-coupled repair, collide with replication machinery, generate topological stress, expose single-stranded DNA, create or stabilize R-loops, and require chromatin restoration after passage. These outcomes are not uniformly harmful or protective. Transcription can expose DNA lesions to repair systems, but unresolved polymerases, RNA-DNA hybrids, and collisions with replication forks can create genome instability. R-loops illustrate the context dependence: an RNA-DNA hybrid can participate in regulated promoter-proximal termination in some settings, but persistent or mislocalized R-loops can threaten genome integrity.

Nascent RNA methods let researchers observe transcription in progress, but these methods do not directly report every kinetic parameter. GRO-seq and PRO-seq capture engaged polymerases that can extend labeled nucleotides under run-on conditions. NET-seq-like approaches recover polymerase-associated RNA ends. Transient transcriptome sequencing and related metabolic-labeling methods capture recently made RNA over a pulse. Live-cell imaging follows reporter transcripts or transcription sites over time. Each method measures a different operational definition of "nascent RNA." Inferring initiation frequency, pause entry, pause duration, elongation rate, premature termination, RNA processing delay, or RNA decay requires kinetic models, perturbations, and orthogonal validation.

## Concept Inventory

- **Elongation complex:** a polymerase-template-product assembly that remains bound while adding nucleotides to a nascent RNA or related nucleic-acid product. In DNA-dependent transcription, the complex includes RNA polymerase, the DNA template, the nascent RNA, a transcription bubble, and a short DNA-RNA hybrid. The same term can also be used more broadly for RNA-dependent synthesis systems when the template and product are RNA, but the DNA-dependent transcription complex is the main subject of this chapter.
- **Processivity:** the number of nucleotides an enzyme adds before dissociating from its template. Processivity is a property of template retention, not a promise of uniform velocity. A highly processive polymerase may still pause for milliseconds, seconds, or longer when it encounters particular sequences, structures, chromatin barriers, regulatory factors, DNA damage, or another molecular machine.
- **Promoter-proximal pausing:** the accumulation of RNA polymerase II in a paused state near a promoter shortly after initiation. It is a post-initiation checkpoint, not simply a failure to initiate. A high signal near the promoter can also arise from high initiation, premature termination, RNA stability, mapping bias, or chromatin accessibility, so promoter-proximal pausing should be inferred with appropriate assays and controls.
- **Pause release:** the transition from a promoter-proximal paused state into productive elongation. In metazoan Pol II systems, pause release often involves phosphorylation of the Pol II C-terminal domain, DSIF, NELF, or associated factors by P-TEFb and related kinase activities. Pause release should not be inferred from total polymerase occupancy alone, because occupancy is shaped by entry into the pause, exit from the pause, initiation, premature termination, and decay of short RNAs.
- **Negative elongation factor:** a metazoan factor complex that stabilizes promoter-proximal paused Pol II and helps shape the paused regulatory state. The name can be misleading if it is read as "generic inhibitor." NELF can act as a structured regulatory platform that helps create, maintain, patrol, or remodel paused Pol II complexes.
- **DRB sensitivity-inducing factor:** a Spt4/5-family elongation factor complex that binds Pol II and participates in both pausing and productive elongation. Its functional relatives include archaeal Spt4/5 and bacterial NusG-family proteins. DSIF illustrates a recurring principle: an elongation factor can restrain one state of transcription and support another state after phosphorylation or partner-factor exchange.
- **Integrator:** a metazoan RNA-processing and termination complex. It was first associated with small nuclear RNA biogenesis, but it is now understood to act more broadly at selected Pol II transcription units, including promoter-proximal transcripts. Integrator can cleave nascent RNA and help terminate transcription, although its outcome is locus-specific and should not be reduced to one universal function.
- **R-loop:** a three-stranded nucleic-acid structure containing an RNA-DNA hybrid and a displaced single-stranded DNA strand. R-loops can form behind, within, or near transcription complexes when nascent RNA anneals to template DNA. They can have regulatory roles, but they are also associated with genome-instability risks when persistent, mislocalized, or poorly resolved.
- **Transcription-replication conflict:** occurs when transcription machinery and DNA replication machinery obstruct one another on the same DNA template. Head-on encounters, codirectional encounters, highly transcribed regions, stable RNA-DNA hybrids, and DNA-bound protein complexes can produce different outcomes. These conflicts are treated in more detail in [Chapter 98](chapter1093.md).
- **Nascent RNA method:** an assay that captures newly synthesized RNA, polymerase-associated RNA, or recently labeled RNA to infer transcription dynamics. The method family includes run-on approaches, polymerase-associated RNA end mapping, metabolic labeling, and live-cell imaging. Different methods report different molecules and time windows, so the phrase "nascent RNA" must be interpreted operationally.

## What to Know Before Reading This Chapter

Transcription uses a nucleoside triphosphate substrate, a template, and a polymerase active site to synthesize RNA in the 5′ to 3′ direction. The template is read in the opposite direction. In DNA-dependent transcription, local DNA strands separate to form a transcription bubble, and the growing RNA remains paired to the template over a short RNA-DNA hybrid before exiting the polymerase. Chapters [20](chapter1019.md) and [21](chapter1020.md) describe the catalytic center, bridge helix, trigger loop, clamp, and other structural elements that support nucleotide addition. This chapter assumes those details but repeats the functional logic where needed.

The reader should also distinguish recruitment, initiation, promoter escape, elongation, pausing, and termination. Recruitment places polymerase and initiation factors near a promoter. Initiation begins RNA synthesis. Promoter escape commits a polymerase to leave the initiation region. Elongation is the repeated cycle of nucleotide addition and translocation along the template. Pausing is a slowed or stopped state in which the polymerase usually remains template-bound. Termination resolves the elongation complex so RNA and polymerase no longer remain in a productive template-bound state.

A recurring example in this chapter is a metazoan Pol II gene with a promoter-proximal paused polymerase. At such a gene, Pol II may be present near the promoter even before a full-length mRNA is made. The key regulatory decision is not simply "is Pol II recruited?" but "does the initiated polymerase enter productive elongation, terminate early, or remain paused?" A second recurring example is a bacterial operon, where elongation and termination are influenced by RNA structures, Rho access, Nus factors, ribosome position, and coupling between transcription and translation.

The evidence base is diverse. Cryogenic electron microscopy and crystallography reveal molecular states of elongation complexes. In vitro biochemistry tests how defined nucleic-acid scaffolds, factors, and sequences change pausing or termination. Genome-wide assays locate polymerases or nascent RNA fragments across genes. Imaging follows transcription sites in living cells. Genetic and degron perturbations test factor requirements. No single evidence type is sufficient for all claims. A promoter-proximal sequencing peak, a structural snapshot, and a live-cell reporter trace answer related but non-identical questions.

## 24.1. Elongation complexes, nucleic acid scaffolds, and processivity

An elongation complex is the physical machine that converts a template into RNA after initiation. In bacterial and eukaryotic DNA-dependent transcription, the polymerase maintains a transcription bubble in which a short stretch of DNA is unwound. The template strand passes through the active-site region, the non-template strand is displaced and then rewound, and the nascent RNA forms a short hybrid with the template before separating from DNA and exiting through an RNA channel. The polymerase must therefore solve several problems at once: nucleotide selection, phosphodiester-bond formation, forward movement, template retention, product handling, and prevention of premature dissociation.

The basic nucleotide-addition cycle has repeated steps. First, the active site receives an incoming ribonucleoside triphosphate that can pair with the next template base. Second, catalytic metal ions and active-site residues position the 3′ hydroxyl of the growing RNA to attack the alpha phosphate of the incoming nucleotide. Third, pyrophosphate is released. Fourth, the polymerase translocates so the next template base enters the active site. This simplified cycle is useful, but real elongation complexes do not move as metronomes. Sequence context, folding of the nascent RNA, DNA supercoiling, protein factors, chromatin barriers, DNA lesions, nucleotide availability, and collisions with other machines can bias the polymerase toward forward elongation, pausing, backtracking, or termination.

Processivity is often misunderstood. A polymerase with high processivity can synthesize a long RNA without falling off the template, but processivity does not mean high speed or uninterrupted motion. A paused polymerase can remain highly stable on DNA. The distinction matters because regulatory systems often exploit stable paused states. A paused elongation complex can wait for a signal, allow RNA folding, recruit processing factors, provide time for ribosomes or spliceosomes to engage, or create a kinetic window for termination factors. Conversely, a long-lived stalled polymerase can become an obstacle that must be rescued or removed.

![Figure 24.1. Elongation State Landscape](../assets/figures/chapter1023_figure1.png)

**Figure 24.1. Elongation State Landscape.** Productive elongation, elemental pausing, hairpin-stabilized pausing, backtracking, arrest, rescue by transcript cleavage, pause release, and termination are competing states of a template-bound RNA polymerase. This schematic replaces the intuitive picture of smooth polymerase motion with a state-transition view in which kinetic competition among forward synthesis, backward movement, factor binding, RNA structure formation, and complex stability determines the outcome at any point along the template. Elongation factors do not simply speed or slow the polymerase; they shift the relative probabilities of entering, remaining in, or exiting each state.

Pausing has several mechanistic forms. In an elemental pause, the polymerase slows because the active-site and nucleic-acid scaffold are temporarily misaligned for rapid nucleotide addition. In a hairpin-stabilized bacterial pause, an RNA hairpin in the exit channel stabilizes a paused state. In backtracking, the polymerase reverse-translocates, and the 3′ end of the nascent RNA moves away from the active site. A backtracked complex cannot immediately add the next nucleotide because the RNA 3′ hydroxyl is no longer positioned for catalysis. Recovery can occur by forward diffusion or by cleavage of the extruded RNA segment, which creates a new aligned 3′ end. Bacterial Gre factors provide a well-studied rescue mechanism for backtracked RNAP, and backtracking also links pausing to genome-instability risk in broader transcription systems. Arrest refers to a particularly long-lived inactive state, often requiring specialized rescue factors.

Elongation factors control these state transitions rather than merely increasing polymerase speed. Bacterial NusG, archaeal Spt4/5, and eukaryotic Spt4/5 or DSIF-related factors bind near the nucleic-acid channels of elongation complexes. These factors can increase processivity, affect pausing, recruit additional proteins, and influence termination or antitermination. The direction of the effect depends on the polymerase, factor modification state, partner factors, and sequence context. A useful rule is that Spt4/5-family factors organize the elongation interface. They do not encode a single universal outcome.

**Table 24.1. Elongation Factors and Their Functional Effects.** Key transcription elongation factors from bacteria, archaea, and eukaryotes, with the organism, polymerase, primary elongation state each factor affects, what it can promote or restrain, and an important caveat for interpretation.

| Factor | Organism/Domain | Polymerase | Primary State Affected | Can Promote | Can Restrain | Key Caveat |
| --- | --- | --- | --- | --- | --- | --- |
| **NusG** | Bacteria | RNAP | Processivity and pause suppression | Forward elongation; Rho loading on nascent RNA | Pausing at some sequence contexts | Effect depends on operon context and ribosome occupancy of the nascent RNA |
| **NusA** | Bacteria | RNAP | Hairpin-stabilized pausing | RNA hairpin formation; intrinsic termination | Rapid forward elongation | Can enhance termination signals but also participates in antitermination complexes |
| **Gre factors (GreA/GreB)** | Bacteria | RNAP | Backtracked state | Transcript cleavage; rescue from arrested complexes | None directly | Stimulate intrinsic RNA cleavage to realign the 3′ end after backtracking |
| **Spt4/5 (archaeal)** | Archaea | RNAP | Promoter-proximal elongation | Processivity; factor recruitment | Early pausing | Eukaryotic-like subunit composition operating in archaeal genome and factor context |
| **DSIF (Spt4/Spt5)** | Metazoa, yeast | Pol II | Pausing and productive elongation | Pausing when unphosphorylated; productive elongation when phosphorylated by CDK9 | Each state depending on phosphorylation status | Same factor can restrain or promote elongation after P-TEFb-dependent modification |
| **NELF** | Metazoa | Pol II | Promoter-proximal pausing | Paused complex stability; pause entry | Productive elongation | Occupies multiple conformations on paused Pol II; not a simple generic elongation brake |
| **P-TEFb (CDK9/cyclin T)** | Metazoa | Pol II | Pause release | Productive elongation; DSIF conversion; NELF dissociation or remodeling | Prolonged pausing | Kinase complex; phosphorylates Pol II CTD, DSIF, and NELF; activity inhibited by DRB |
| **7SK RNP (7SK, LARP7, MePCE, HEXIM)** | Metazoa | P-TEFb availability | Pause-release competence | Reversible sequestration and mobilization of P-TEFb | Uncontrolled or mistimed kinase availability | RNP composition and conformation are dynamic; abundance alone does not report active P-TEFb flux |
| **Integrator** | Metazoa | Pol II | Promoter-proximal and snRNA transcription | RNA cleavage and termination at selected loci | Productive elongation at some loci | Outcomes are locus-specific; not a universal off switch for all early Pol II transcripts |
| **FACT** | Metazoa, yeast | Pol II | Nucleosome passage and chromatin restoration | Histone exchange and chromatin restoration behind Pol II | Can increase pausing at nucleosome barriers | Histone chaperone linking nucleosome management to Pol II elongation state in vivo |

Chromatin adds another layer to eukaryotic elongation. A nucleosome is not simply a passive bead on DNA. It is a histone-DNA complex that can impede polymerase movement, require partial DNA unwrapping, recruit or require histone chaperones, and influence the restoration of chromatin behind a polymerase. Structural work on chromatin transcription has shown how Pol II-associated factors, histone chaperones, and nucleosome geometry help determine whether a polymerase pauses, passes, or leaves altered chromatin in its wake. FACT, for example, can help maintain chromatin architecture and influence Pol II pausing in vivo, illustrating that pausing can depend on nucleosome management rather than only on DNA sequence or soluble pausing factors.

The nucleic-acid scaffold itself carries regulatory information. A GC-rich or AT-rich DNA segment, a weak RNA-DNA hybrid, an RNA hairpin, a pause-inducing sequence, or an RNA that can pair back to DNA can change elongation state. In bacteria, intrinsic terminators depend on a nascent RNA hairpin followed by a weak U-rich RNA-DNA hybrid. In eukaryotes, nascent RNA structures, splice sites, cleavage sites, R-loops, and RNA-binding proteins influence the kinetic environment that Pol II experiences. These examples emphasize a central point for the rest of the chapter: elongation is governed by kinetic competition among forward synthesis, pausing, factor binding, RNA folding, RNA processing, and termination.

Boundary cases are important. RNA-dependent RNA polymerases and reverse transcriptases use different templates and product-template hybrids, but they share the general problem of processive synthesis on a nucleic-acid scaffold. Organellar and viral polymerases may have streamlined factor sets or specialized termination signals. Some polymerases terminate by intrinsic template signals, while others require accessory factors or processing events. The conserved theme is not a single molecular pathway, but the need to stabilize synthesis long enough to make RNA and destabilize synthesis at the correct regulatory point.

## 24.2. Promoter-proximal pausing as a regulatory decision point

Promoter-proximal pausing is a regulatory state in which Pol II begins transcription but pauses near the promoter before entering productive elongation. The paused polymerase is usually found tens of nucleotides downstream of the transcription start site, although exact positions vary by gene, cell type, and assay. This arrangement separates initiation from full-length transcript production. A gene can have recruited Pol II and short nascent RNA without producing abundant mature mRNA. For many regulated metazoan genes, the question is therefore not only whether a promoter recruits Pol II, but whether the promoter-proximal elongation complex is released, held, or terminated.

The canonical metazoan pausing pathway involves DSIF and NELF. DSIF binds Pol II during early elongation. NELF associates with the early elongation complex and helps stabilize the paused state. P-TEFb, a kinase complex containing CDK9 and cyclin T, phosphorylates targets that include the Pol II C-terminal domain and pausing-associated factors. These phosphorylation events promote release from the paused state, NELF dissociation or remodeling, conversion of DSIF into a productive elongation factor, and recruitment of processing and chromatin factors. This description is a framework rather than a fixed script; exact dependencies differ among loci and biological conditions.

The supply of active P-TEFb is itself regulated by the 7SK ribonucleoprotein, or 7SK RNP. 7SK is an abundant structured nuclear RNA that assembles with proteins including LARP7 and methylphosphate capping enzyme (MePCE); HEXIM proteins can join this scaffold and bind P-TEFb in an inhibited state. Regulatory signals and competing factors redistribute P-TEFb between the sequestered 7SK particle and transcriptionally active complexes. The causal logic is therefore not that 7SK permanently switches elongation off. The 7SK RNP buffers and mobilizes a kinase pool, so its composition and RNA conformation can alter how rapidly particular loci gain pause-release activity.

Structural analysis supports a conformational-switching model in which alternative 7SK RNA and protein arrangements change partner accessibility. Such structures identify plausible molecular states but do not, by themselves, establish the order or rate of switching in living cells. Strong cellular evidence combines acute perturbation of 7SK, LARP7, MePCE, HEXIM, or P-TEFb with measurements that distinguish initiation, pause entry, pause duration, release, and early termination. Total RNA abundance or a static P-TEFb co-immunoprecipitation cannot resolve those alternatives.

![Figure 24.2. Promoter-Proximal Pausing and Release](../assets/figures/chapter1023_figure2.png)

**Figure 24.2. Promoter-Proximal Pausing and Release.** After initiating transcription and synthesizing a short RNA, RNA polymerase II pauses tens of nucleotides downstream of the transcription start site, where DSIF and NELF stabilize the paused elongation complex. A regulatory inset shows P-TEFb partitioning between a HEXIM-containing 7SK RNP pool and active transcription complexes. Phosphorylation by released P-TEFb targets the Pol II C-terminal domain, DSIF, and NELF, promoting NELF remodeling or dissociation, conversion of DSIF into a productive elongation-supporting factor, and entry of Pol II into productive elongation. The figure distinguishes recruitment, initiation, pausing, kinase-pool mobilization, and pause release as events that can be regulated independently and that do not always lead to full-length transcript production.

NELF is a useful example of why names can oversimplify mechanism. "Negative elongation factor" suggests a simple brake, but structural and mechanistic studies show that NELF can occupy distinct conformations on paused Pol II, interact with the emerging RNA and polymerase surface, and create a regulated platform for pausing and pause release. In some contexts, NELF stabilizes a pause long enough for regulated responses. In other contexts, loss of NELF can alter initiation, premature termination, chromatin, or RNA processing indirectly. A careful interpretation asks which transition changed: pause entry, pause lifetime, release into elongation, termination, or RNA stability.

Promoter architecture contributes to pausing. TFIID, promoter sequence, core-promoter elements, transcription start-site choice, and the early transcribed region can influence whether the polymerase enters a paused state. Evidence that TFIID enables promoter-proximal pausing links the initiation machinery in [Chapter 21](chapter1020.md) to the elongation checkpoint discussed here. Chromatin marks can also help. H3K4me3, a histone H3 lysine 4 trimethylation mark enriched near many active promoters, has been connected to promoter-proximal pause release in specific experimental contexts. The mechanistic lesson is not that one mark universally releases all paused Pol II, but that promoter chromatin, initiation factors, and elongation factors form one regulatory system.

The biological logic of promoter-proximal pausing is easiest to see at rapidly inducible genes. If Pol II is already initiated and paused near a promoter, a signal-dependent pathway can trigger productive elongation faster than a pathway that must first open chromatin, assemble initiation machinery, and recruit Pol II from scratch. Heat-shock genes and immediate-response genes have often served as conceptual examples. The same logic applies more broadly to developmental regulators, immune-response genes, and genes whose output must be tuned by signaling pathways. Pausing also provides a checkpoint at which low-quality, cryptic, or unproductive initiation events can be terminated rather than extended into full transcripts.

Promoter-proximal pausing is common in metazoans, but it is not a universal eukaryotic rule. Budding yeast has different Pol II pausing behavior and lacks the same NELF-centered system. Plants, animals, fungi, and protists differ in factor repertoires, chromatin environments, gene architecture, and transcriptional bursting. Archaea also show promoter-proximal elongation control in some systems, but archaeal mechanisms use archaeal transcription factors and genome contexts rather than a metazoan NELF pathway. A reader should therefore learn the general principle, post-initiation control, before generalizing the exact machinery.

Promoter-proximal pausing also intersects with early RNA processing. The first 20 to 60 nucleotides of a Pol II transcript include the interval in which capping enzymes can engage the polymerase C-terminal domain and the emerging RNA 5′ end. Pause duration can change the time available for capping, early RNA surveillance, and promoter-proximal termination. These links are developed in [Chapter 25](chapter1024.md) and [Chapter 26](chapter1025.md), but the key point here is that pausing is not an isolated transcription phenomenon. It is a kinetic window in which the identity of the nascent RNA and elongation complex can be inspected and modified.

**Table 24.2. Termination Pathways Across Domains.** Major transcription termination pathways, their required signals, factor dependencies, RNA-processing links, primary supporting evidence, and a common misconception to avoid for each.

| Pathway | Required Signal | Factor Requirement | RNA Processing Link | Main Evidence | Common Misconception to Avoid |
| --- | --- | --- | --- | --- | --- |
| **Bacterial intrinsic termination** | GC-rich RNA hairpin followed by a U-rich tract in the nascent RNA | None; intrinsic to the RNA and DNA-RNA hybrid | None required | Biochemical assays and mutational analysis of terminator sequence elements | The hairpin alone is sufficient; the weak rU-dA hybrid is equally necessary for complex destabilization |
| **Bacterial Rho-dependent termination** | C-rich, G-poor rut site in accessible nascent RNA | Rho ATP-dependent RNA translocase | None required | Genetics, in vitro reconstitution, and structural studies | Rho acts indiscriminately on all transcripts; ribosome occupancy and translation status gate Rho access |
| **Archaeal factor-dependent termination** | Archaeal termination signals in the DNA or RNA | Archaeal-specific termination factors | Not coupled to eukaryotic polyadenylation | Genetic and biochemical experiments in archaeal model systems | Archaeal termination is a simple intermediate between bacterial and eukaryotic; it is mechanistically distinct from both |
| **Eukaryotic cleavage-coupled Pol II termination** | Polyadenylation signal and downstream sequence elements | Cleavage and polyadenylation factors; XRN2 5′-to-3′ exonuclease | Coupled to pre-mRNA cleavage and polyadenylation | Torpedo and allosteric model experiments; mutational and depletion studies | The torpedo and allosteric models are mutually exclusive; cleavage, exonuclease action, and allosteric changes can cooperate |
| **Integrator-associated promoter-proximal termination** | Promoter-proximal Pol II transcript | Integrator complex | RNA cleavage without polyadenylation | Genomic occupancy, RNA-seq, and biochemical depletion experiments | Integrator terminates all early Pol II transcripts uniformly; its outcomes are locus- and context-specific |
| **Pol III oligo(dT)-linked termination** | Short U-rich RNA arising from oligo(dT) in the non-template strand | Pol III-intrinsic mechanism | Coupled to tRNA, 5S rRNA, and snRNA 3′ end formation | In vitro transcription and genetic studies with Pol III transcription units | Pol III termination is mechanistically identical to Pol II termination; each polymerase uses distinct signals |

Do not overgeneralize promoter-proximal signal. A peak of Pol II ChIP-seq, PRO-seq, GRO-seq, or NET-seq signal near a promoter can be consistent with pausing, but it is not by itself a rate constant for pausing. The same peak can be shaped by initiation frequency, polymerase density, pause entry, pause lifetime, release rate, premature termination, cleavage of short RNAs, mapping of short reads, and RNA degradation. Strong evidence for altered pausing usually combines location-specific nascent RNA data with perturbations, time-resolved measurements, factor occupancy, and models that separate initiation from pause release.

> **Box 24.1. High Promoter Signal Is Not Always Pausing**
>
> A peak of Pol II, PRO-seq, GRO-seq, or NET-seq signal near a promoter is consistent with promoter-proximal pausing but is not by itself evidence that pausing specifically changed. The same peak can arise from:
>
> - Higher initiation rate producing more polymerases entering the proximal region
> - Longer pause duration keeping each polymerase near the promoter for more time
> - Slower pause release reducing the fraction of polymerases that enter the gene body
> - More premature termination generating stable short capped RNAs that persist near the transcription start site
> - Read-mapping or RNA-recovery bias favoring short fragments at gene 5′ ends
> - RNA stability differences affecting how long short transcripts accumulate before degradation
>
> Strong evidence for altered pausing combines location-specific nascent RNA data with perturbations of specific factors, time-resolved measurements, factor occupancy profiles, and quantitative models that separately estimate initiation rate, pause entry, pause duration, and release rate.

## 24.3. Termination mechanisms across bacteria, archaea, eukaryotes, and viruses

Termination is the resolution of an elongation complex. A polymerase that keeps synthesizing beyond the biologically appropriate endpoint can make aberrant RNA, interfere with downstream genes, collide with replication machinery, or fail to recycle. A polymerase that terminates too early can truncate functional transcripts. Termination therefore combines chemistry, RNA structure, protein factors, kinetic timing, and genome organization. It is helpful to define termination broadly as cessation of productive synthesis with release, processing, or stable disengagement of the RNA and template. Different systems satisfy this definition in different ways.

Bacterial intrinsic termination depends on information in the nascent RNA and DNA template. The classic intrinsic terminator contains a GC-rich inverted repeat followed by a U-rich tract in the RNA. The inverted repeat can fold into a hairpin as the RNA exits the polymerase. The hairpin and the weak rU-dA hybrid destabilize the elongation complex, favoring release of the RNA and polymerase. This mechanism illustrates kinetic competition: the hairpin must form at the right time, the polymerase must be susceptible to termination, and elongation must not outrun the RNA-structure signal. Nus factors, RNA sequence, translation by ribosomes, and RNA-binding proteins can alter this competition.

Bacterial Rho-dependent termination uses a protein motor rather than an intrinsic RNA hairpin as the central driver. Rho is an ATP-dependent RNA translocase that loads onto accessible nascent RNA, often at C-rich and G-poor rut sites, and translocates along RNA toward a paused RNA polymerase. When Rho catches a susceptible elongation complex, it promotes termination. Translation strongly affects this pathway because ribosomes can cover nascent RNA and prevent Rho loading. This is why transcription-translation coupling is central to bacterial termination logic. A transcript that is being efficiently translated may be protected from Rho, whereas untranslated, poorly translated, antisense, damaged, or regulatory transcripts may become Rho substrates.

Antitermination is the mirror image of termination control. Some bacterial systems modify RNA polymerase so that it reads through terminators. Bacteriophage lambda provides historically important examples in which phage factors and host Nus proteins allow the polymerase to bypass termination signals and express downstream genes. Ribosomal RNA operons also use antitermination-like mechanisms to support long, efficient transcription units. These examples matter because they show that termination is not only a signal at the end of a gene. It is a regulatable decision distributed across the transcription unit.

![Figure 24.3. Cross-Domain Termination Mechanisms](../assets/figures/chapter1023_figure3.png)

**Figure 24.3. Cross-Domain Termination Mechanisms.** Bacterial intrinsic termination driven by an RNA hairpin and weak rU-dA hybrid, Rho-dependent termination by an ATP-dependent RNA translocase, archaeal factor-dependent termination, eukaryotic cleavage-coupled Pol II termination, Integrator-associated promoter-proximal termination, Pol III oligo(dT)-linked termination, and selected viral termination strategies are arranged on a shared comparison diagram. All mechanisms resolve an elongation complex at a regulated point, but they differ in required signals, factor dependencies, and processing links, illustrating that no single universal termination pathway exists.

Archaeal termination has often been underrepresented in textbooks. Archaea have transcription machinery with eukaryotic-like features, but archaeal genome organization and factor repertoires are distinct. Factor-dependent archaeal termination has been demonstrated experimentally, supporting the view that archaeal transcription termination is mechanistically specific and cannot be treated as a simple bacterial or eukaryotic variant. Archaeal promoter-proximal elongation regulation also shows that early elongation decisions are not restricted to metazoan Pol II systems. The boundary case is important: "archaeal" does not mean "primitive bacterial-like," and "eukaryotic-like polymerase" does not imply identical termination machinery.

Eukaryotic Pol II termination at many protein-coding genes is coupled to 3′ end processing. When Pol II transcribes a polyadenylation signal and downstream sequence elements, cleavage and polyadenylation factors assemble on the nascent RNA and Pol II-associated platform. The pre-mRNA is cleaved, the upstream RNA is prepared for poly(A) tail addition, and the downstream RNA remains associated with the elongating polymerase. In the torpedo model, a 5′ to 3′ exonuclease such as XRN2 degrades the downstream RNA and catches Pol II, promoting termination. In allosteric models, passage through the polyadenylation region changes the elongation complex so that Pol II becomes more termination-prone. Current views do not require these models to be mutually exclusive. Cleavage, exonuclease action, pausing, chromatin, factor exchange, and allosteric changes can cooperate.

Promoter-proximal termination is a different eukaryotic termination problem. Instead of ending a mature mRNA after a coding region, the cell may terminate short, promoter-associated transcripts, enhancer RNAs, divergent transcripts, cryptic transcripts, or low-quality early Pol II products. Integrator is a central player in several such contexts. Integrator can cleave nascent RNA and promote termination or processing at selected Pol II transcription units, including small nuclear RNA genes and promoter-proximal transcripts. Integrator should not be described as a universal off switch. Its effect depends on locus type, associated factors, RNA features, and the competing activities of productive elongation, processing, and surveillance.

> **Box 24.2. Termination Can Be Proximal**
>
> Transcription termination is not limited to the downstream end of protein-coding genes. Pol II can terminate or be processed early at:
>
> - Promoter-proximal transcripts, including divergent and upstream antisense RNAs, which may be cleaved by Integrator
> - Enhancer RNAs and other short noncoding transcription units, which are often rapidly terminated or degraded after brief elongation
> - Small nuclear RNA genes, where Integrator-coupled cleavage defines the mature 3′ end
> - Low-quality or unproductive initiation events at cryptic promoters, where early termination prevents aberrant RNA accumulation
>
> Promoter-proximal termination therefore functions as a quality checkpoint: it allows initiation to occur while preventing unintended elongation into downstream genes or noncoding territories.

Other nuclear RNA polymerases use additional termination strategies. Pol I termination is specialized for ribosomal RNA transcription and nucleolar organization, and Pol III termination often responds to short oligo(dT) tracts in the non-template DNA that produce U-rich RNA termini. These systems are treated in Chapters [21](chapter1020.md), [39](chapter1037.md), and [42](chapter1039.md), with Y and vault RNA examples in [Chapter 45](chapter1042.md). They are mentioned here because they reinforce the broader principle: termination is polymerase-specific, but every termination system must solve the same physical problem of resolving a stable elongation complex at the right location.

Viral transcription expands the range of solutions. Some viruses use host Pol II and therefore inherit host pausing, elongation, capping, and termination constraints. Others encode their own polymerases or transcription-replication complexes. Viral polymerases can terminate at defined signals, switch templates, produce nested transcripts, or synthesize discontinuous RNAs. The appropriate classification is mechanistic rather than taxonomic: ask what the template is, what the polymerase is, how the product is handled, what signal stops synthesis, and whether termination is coupled to processing, template switching, or polymerase recycling.

## 24.4. Transcription-repair, transcription-replication, and chromatin coupling

An elongating polymerase changes the physical state of DNA. It unwinds a short region, creates torsional stress ahead of and behind the transcription bubble, exposes template and non-template strands, leaves nascent RNA near DNA, and recruits many proteins. These properties make transcription a source of RNA and a force that reshapes genome maintenance. The same elongation complex can help a cell detect DNA damage, obstruct repair, collide with replication, generate R-loops, or alter chromatin inheritance.

Transcription-coupled repair is the clearest example of transcription as a genome-surveillance signal. If a lesion blocks elongating RNA polymerase on the transcribed strand, the stalled complex can recruit repair machinery that preferentially repairs the lesion. The logic is efficient: a gene being actively transcribed is likely to matter to the cell, and a polymerase stalled at damage provides positional information. The same logic creates a hazard. A polymerase that cannot be moved or processed can become a roadblock. Productive repair therefore requires recognition of the stalled complex, remodeling or removal of proteins, repair of DNA, and restoration of transcription competence.

Transcription-replication conflicts arise when replication machinery and transcription machinery attempt to use the same DNA. A codirectional encounter occurs when replication and transcription move in the same direction. A head-on encounter occurs when they move toward one another. Head-on conflicts are often more disruptive because the two machines directly oppose each other, but codirectional conflicts can also be harmful if a replication fork catches a slow or stalled polymerase. In bacteria, genome organization often biases highly transcribed genes, especially rRNA operons, toward codirectional orientation with replication. In eukaryotes, conflicts are shaped by replication timing, transcription level, chromatin state, R-loops, fragile sites, and long genes.

![Figure 24.4. Transcription, Replication, Repair, and R-Loops](../assets/figures/chapter1023_figure4.png)

**Figure 24.4. Transcription, Replication, Repair, and R-Loops.** An elongating RNA polymerase can recruit transcription-coupled repair factors when stalled at a DNA lesion on the transcribed strand, collide head-on or codirectionally with a replication fork, form an R-loop when the nascent RNA hybridizes back to the template DNA strand, or require histone chaperones and chromatin remodelers to restore nucleosome architecture after passage. Regulatory and genome-stability-threatening outcomes of these interactions are marked separately, emphasizing that the same event — such as R-loop formation — can be beneficial or harmful depending on position, lifetime, and resolution machinery.

R-loops provide a mechanistic bridge between transcription, termination, and genome stability. An R-loop forms when RNA hybridizes with the DNA template and displaces the non-template DNA strand. R-loops can form co-transcriptionally when the nascent RNA remains near complementary DNA, especially in regions with favorable sequence, negative supercoiling, RNA-processing delays, or impaired RNA-binding proteins. R-loops can influence promoter activity, termination, recombination, chromatin marks, and DNA damage responses. They are not inherently artifacts or inherently lesions. Their meaning depends on position, lifetime, protein context, and resolution mechanisms.

Promoter-proximal R-loop-associated termination is a concrete example of R-loop regulation. In some mammalian contexts, R-loop-dependent promoter-proximal termination can limit aberrant transcription and support genome stability. This finding does not mean all R-loops are protective. Persistent R-loops can expose single-stranded DNA, impede replication, and recruit processing or repair activities in ways that become harmful. R-loop biology is also assay-sensitive. Antibody-based detection, nuclease treatments, strand specificity, RNA processing, and extraction conditions can influence signals. Strong claims about R-loops should combine orthogonal detection methods with perturbations of hybrid-resolving enzymes or RNA-processing factors.

Chromatin coupling gives elongation a memory-like consequence. As Pol II moves through a nucleosome, DNA must become accessible and then be repackaged. Histone chaperones, chromatin remodelers, elongation factors, histone marks, and polymerase-associated factors help manage this transition. If nucleosomes are not restored properly behind Pol II, cryptic promoters may become accessible, histone modifications may be redistributed, and later transcription or replication events may change. Structural and single-molecule work increasingly shows that polymerases can reshape chromatin fibers and that chromatin architecture feeds back on pausing and elongation.

Co-transcriptional RNA modification and RNA-binding proteins can also feed into termination and genome stability. DDX21-mediated co-transcriptional m6A-linked regulation has been connected to transcription termination and genome stability in mammalian cells. The important lesson is not that one modification pathway explains all termination, but that nascent RNA state, RNA-binding proteins, chromatin, and polymerase behavior can form feedback loops. This chapter introduces those loops; [Chapter 25](chapter1024.md) treats co-transcriptional RNA folding and modification in more detail.

The bacterial case adds translation. In bacteria, transcription and translation can be physically coupled because ribosomes can begin translating an mRNA before transcription is complete. Ribosome position influences nascent RNA accessibility, Rho loading, RNA folding, and RNA decay. A stalled ribosome can expose RNA regions or alter transcription speed indirectly, while a paused polymerase can affect ribosome progression. Transcription-translation coupling therefore links elongation, termination, RNA quality control, and protein synthesis in one kinetic system.

## 24.5. Nascent RNA methods and kinetic interpretation

Nascent RNA methods are designed to observe RNA synthesis close to the time it occurs. This goal sounds straightforward, but "nascent" has multiple operational meanings. It may mean RNA still attached to a polymerase, RNA that can be extended during a nuclear run-on reaction, RNA labeled during a short metabolic pulse, RNA imaged at a transcription site, or RNA captured through a structure or proximity method while it is still being processed. Each definition is useful. None is a complete movie of transcription.

Run-on methods such as global run-on sequencing and precision run-on sequencing capture engaged polymerases that can extend nascent RNA in isolated nuclei or permeabilized cells in the presence of labeled nucleotides. Their strength is positional information about transcriptionally engaged polymerases and promoter-proximal pausing. Their limitations include run-on conditions, nucleotide incorporation biases, mapping constraints, and the fact that a polymerase must be capable of extension under assay conditions to be detected. A polymerase that is crosslinked, arrested in a nonextendable state, released before capture, or associated with a rapidly degraded RNA may be underrepresented.

NET-seq-like methods capture RNA ends associated with polymerase complexes. These methods can provide nucleotide-resolution information about polymerase position and pausing, especially when combined with immunoprecipitation of a polymerase or factor. The signal, however, depends on the stability of the polymerase-RNA complex, RNA cleavage, library construction, alignment of short reads, and the biochemical conditions used to isolate complexes. A sharp RNA-end peak can indicate pausing, but it can also be influenced by RNA cleavage, sequence-dependent recovery, or processing intermediates.

Metabolic-labeling approaches pulse cells with modified nucleosides so that recently synthesized RNA can be enriched and sequenced. Transient transcriptome sequencing and related approaches are powerful for estimating synthesis and decay, especially when pulse and chase designs are analyzed quantitatively. Their limitations are different from run-on methods. Label uptake, nucleotide metabolism, incorporation efficiency, RNA processing, RNA decay, and cell-state changes during labeling all affect interpretation. A short pulse improves temporal resolution but may reduce signal; a long pulse improves recovery but blurs kinetic events.

Live-cell imaging follows transcription in real time, often by inserting RNA stem-loop arrays into reporter transcripts and expressing fluorescent coat proteins that bind those stem loops. Imaging can reveal transcriptional bursting, changes in transcription-site intensity, stochastic splice-site choice, and the relationship between initiation, elongation, splicing, and release. Imaging is unusually direct about time, but it is still model-dependent. Fluorescent signal depends on reporter design, stem-loop folding, coat-protein binding, photophysics, RNA movement, maturation, and degradation. Reporters may not fully reproduce endogenous chromatin or RNA processing.

**Table 24.3. Nascent RNA Sequencing and Imaging Methods.** Comparison of major nascent RNA method families, the molecule each captures, the strongest and weakest inferences each supports, the primary technical bias, and how results can be validated.

| Method Family | Captured Molecule | Strongest Inference | Weak Inference | Main Bias | Validation Strategy |
| --- | --- | --- | --- | --- | --- |
| **GRO-seq / PRO-seq** | Engaged polymerases that extend labeled nucleotides under run-on conditions | Polymerase position and promoter-proximal pausing index | Absolute elongation rate or initiation frequency in isolation | Non-extendable, crosslinked, or already-released polymerases are underrepresented | Spike-in controls; inhibitor perturbations; comparison with orthogonal polymerase occupancy methods |
| **NET-seq and related end-mapping** | Pol II-associated RNA 3′ ends recovered by immunoprecipitation | Nucleotide-resolution pausing sites along the template | Distinction between pausing and RNA cleavage or processing intermediates | Complex stability and RNA cleavage during cell lysis and isolation | Factor-specific immunoprecipitation; orthogonal read-end mapping; cleavage-site mutation controls |
| **TT-seq / metabolic labeling** | Recently synthesized RNA pulse-labeled with modified nucleosides | RNA synthesis rate and RNA decay rate estimation | Real-time polymerase position or pause entry rate | Label uptake efficiency; RNA processing events occurring during the pulse window | Pulse-chase time courses; independent RNA half-life measurements; spike-in normalization |
| **Live-cell imaging** | Reporter RNA at active transcription sites via stem-loop and coat-protein systems | Transcriptional bursting dynamics and stochastic behavior over time | Endogenous elongation rate or chromatin-specific behavior of unmodified loci | Reporter stem-loop design, coat-protein binding efficiency, photophysics, and RNA movement | Endogenous genomic locus tagging; dual-color burst assays; comparison with PRO-seq data |
| **Nascent structure / contact mapping (e.g., CAR-SPLASH)** | Crosslinked nascent pre-mRNA structural contacts and proximity pairs | Co-transcriptional RNA folding and kinetic splicing coupling | Complete RNA folding pathway or thermodynamic stability of individual structures | Crosslink efficiency; proximity ligation biases; computational filtering of spurious contacts | Orthogonal structure-probing methods; perturbation of RNA-binding proteins or splicing factors |

Nascent RNA structure and proximity methods connect elongation to RNA folding. For example, CAR-SPLASH and related approaches can capture nascent pre-mRNA structures implicated in kinetic coupling and alternative splicing. These methods are conceptually important because a nascent RNA is not merely a line of sequence exiting a polymerase. It can fold, contact other RNA regions, bind proteins, and influence downstream processing before the transcript is complete. Yet structure-capture methods depend on crosslinking, proximity ligation, fragmentation, enrichment, and computational filtering. Their outputs identify supported contacts or structural signals, not complete folding pathways by themselves.

**Table 24.4. Interpreting Promoter-Proximal Sequencing Signal.** Common observations from promoter-proximal Pol II profiling experiments, the mechanistic explanations each is consistent with, an experimental approach that helps discriminate among explanations, and a conclusion the data alone do not support.

| Observed Pattern | Possible Causes | Discriminating Experiment | Unsupported Conclusion to Avoid |
| --- | --- | --- | --- |
| **Increased promoter-proximal Pol II or PRO-seq peak** | Higher initiation rate; longer pause duration; slower pause release; more premature termination; stable capped short RNAs | Acute factor depletion with time-resolved PRO-seq; measure initiation versus gene-body Pol II ratio separately | "Pause entry rate or pause duration specifically increased" |
| **Decreased gene-body Pol II or PRO-seq signal after factor loss** | Reduced pause release; increased early termination; reduced initiation; slower elongation with polymerase redistribution | Combine Pol II ChIP with RNA stability controls and independent initiation-rate measurements | "Elongation rate decreased" |
| **Sharp RNA 3′-end peak at +30 to +60 bp in NET-seq** | Paused Pol II; RNA cleavage intermediate; short-read mapping artifact; stable capped short RNA | Cross-validate with GRO-seq and RNA stability assay; examine read-length distribution | "Polymerase is stationary at this precise position in all cells" |
| **Unchanged total Pol II ChIP signal but reduced gene-body RNA output** | Altered pause release without change in polymerase density; altered RNA processing or decay downstream | Time-resolved nascent RNA labeling plus independent RNA stability measurement | "Pol II occupancy faithfully reports mRNA production rate" |

Kinetic interpretation requires explicit variables. A promoter-proximal signal can be increased by faster initiation into a pause, longer pause duration, slower pause release, higher premature termination with stable short RNAs, altered RNA cleavage, or improved recovery of short reads. Gene-body signal can increase because more polymerases enter productive elongation, because elongation slows, because RNA processing changes, or because decay changes. A drop in downstream signal after factor depletion might mean defective pause release, increased early termination, reduced initiation, slower elongation with redistribution of polymerases, or RNA instability. The assay does not choose among these explanations automatically.

> **Box 24.3. Nascent RNA Methods Need Kinetic Models**
>
> A pile of sequencing reads near a promoter or a fluorescent spot at a transcription site in live-cell imaging is a measurement, not a rate constant. To extract kinetic information from nascent RNA data, explicit models with defined variables are required.
>
> - A higher PRO-seq signal at a gene could mean faster initiation, a longer pause, or slower release from the pause — the reads alone do not distinguish these alternatives
> - A drop in gene-body signal after factor depletion could reflect reduced pause release, increased premature termination, or lower initiation
> - A burst of fluorescence at an imaging spot reports integrated reporter RNA accumulation, which depends on initiation, elongation, RNA cleavage, coat-protein binding, RNA movement, and export
>
> Good kinetic inference uses time resolution, spike-in controls, perturbations of individual pathway steps, independent RNA stability measurements, and mathematical models with explicit variables for each rate. Models that assume a single elongation rate for all genes, ignore premature termination, or treat all promoter-proximal reads as paused polymerase are likely to misinterpret the underlying biology.

Good kinetic studies therefore combine methods. A perturbation of NELF, DSIF, P-TEFb, Integrator, XRN2, Rho, or a chromatin factor should ideally be analyzed with a design that separates initiation from elongation and termination. Time-resolved perturbations, acute depletion, inhibitor washout, spike-ins, orthogonal polymerase occupancy measurements, RNA stability controls, and mathematical models can help. In vitro systems establish direct mechanisms but may omit chromatin or cellular factor competition. Genome-wide methods reveal scale but often compress heterogeneous molecules into averaged profiles. Live-cell imaging reveals dynamics but may rely on engineered reporters. The best evidence triangulates across these strengths and weaknesses.

## Biological Contexts Across Organisms and Transcript Classes

Bacteria couple elongation to translation, operon architecture, attenuation, and Rho surveillance. A bacterial mRNA can be translated while it is being made, so ribosome occupancy changes the accessibility of the nascent RNA to Rho and RNA structures. Regulatory leaders, riboswitches, attenuators, and small RNAs can influence whether transcription continues or terminates. These topics are developed in Chapters [78](chapter1073.md) through [81](chapter1076.md), but their physical basis is the elongation-state competition described here.

Archaea provide a comparative system in which eukaryotic-like polymerase architecture operates in archaeal genome and factor contexts. Promoter-proximal elongation regulation and factor-dependent termination show that archaeal transcription has its own regulatory logic. Archaeal examples are especially useful for separating deeply conserved polymerase properties from lineage-specific factor systems.

Metazoan Pol II genes use promoter-proximal pausing, chromatin-dependent elongation, enhancer-promoter regulation, early termination, and extensive co-transcriptional processing. Protein-coding genes must coordinate elongation with capping, splicing, cleavage, polyadenylation, export competence, and surveillance. Noncoding transcription units, including enhancer RNAs, divergent promoter transcripts, promoter upstream transcripts, and many unstable RNAs, often face early termination or rapid decay. A failure to distinguish productive mRNA synthesis from pervasive early transcription leads to overestimation of functional transcript output.

Stable RNA genes have specialized elongation and termination demands. Small nuclear RNA genes can use Integrator-linked processing and termination. Ribosomal RNA genes require high transcriptional output and specialized termination and processing in nucleolar contexts. tRNA and other Pol III transcripts often use compact transcription units and termination signals suited to Pol III. These transcript classes are covered in later chapters, but the common theme is that each RNA class uses elongation and termination choices appropriate to its maturation pathway.

Viral systems can exploit host elongation control or encode alternative polymerase behavior. Some viruses use host Pol II and manipulate pausing, elongation, capping, splicing, cleavage, and termination. Other viruses use viral RNA-dependent RNA polymerases, discontinuous transcription, template switching, or transcription-replication complexes. Viral examples caution against defining termination only by host Pol II conventions.

## Technology, Computational, Clinical, and Engineering Links

Transcription elongation is a druggable and engineerable process, but drug effects must be interpreted carefully. CDK9 and transcriptional kinase inhibitors can alter pause release and productive elongation, which is relevant to cancer biology, inflammation, viral transcription, and transcriptional addiction models. Nucleoside analogs and polymerase inhibitors can affect viral or cellular polymerases, although antiviral RNA polymerase drugs are treated mainly in Chapters [160](chapter1143.md) and [162](chapter1145.md). Because elongation factors often affect many genes, therapeutic windows depend on context, dose, target selectivity, and the distinction between global transcription suppression and selective pathway modulation.

Reporter engineering uses elongation logic to measure gene regulation. Stem-loop reporters, intronic reporters, MS2 or PP7 systems, and synthetic pause or terminator elements can test how sequence and factors affect transcriptional dynamics. These tools are powerful when the engineered transcript preserves the relevant chromatin, promoter, splicing, and RNA-processing context. They can mislead when a reporter is removed from its native locus or when stem-loop arrays change RNA processing.

Computational analysis of nascent RNA data often requires models that estimate initiation, pause release, elongation rate, termination, and decay from read density or time courses. A simple pileup plot is descriptive. A kinetic model is interpretive. Models become more reliable when the experiment contains time resolution, spike-ins, replicate perturbations, independent RNA stability estimates, and explicit uncertainty. Models become less reliable when they assume one elongation rate for all genes, ignore premature termination, or treat all promoter-proximal reads as paused polymerase.

Synthetic biology can use terminators, attenuators, antiterminators, riboswitches, and engineered pausing sequences to build regulatory circuits. In bacteria, terminator strength is a standard design variable. In eukaryotic systems, synthetic transcription units must account for promoter-proximal pausing, cryptic polyadenylation, transcriptional interference, chromatin effects, and RNA surveillance. A design that works in plasmids may fail at a genomic locus because elongation and termination are embedded in chromatin and nuclear RNA processing.

## Recent Consensus

The current consensus is that elongation is a regulated, state-rich phase of transcription. Polymerases switch among productive, paused, backtracked, arrested, termination-prone, and rescued states. These states are controlled by nucleic-acid sequence, scaffold geometry, RNA folding, elongation factors, chromatin, RNA processing, DNA topology, and molecular collisions.

Promoter-proximal pausing is a major Pol II checkpoint in many metazoan genes. DSIF, NELF, P-TEFb, promoter architecture, chromatin, and RNA-processing factors form a regulatory network that determines whether early Pol II becomes productive, remains paused, or terminates. The same machinery should not be generalized without qualification to all eukaryotes or all genes.

Termination is mechanistically diverse and integrated with RNA processing. Intrinsic and Rho-dependent bacterial termination, factor-dependent archaeal termination, cleavage-coupled Pol II termination, Integrator-associated termination, Pol I and Pol III termination, and viral termination strategies differ in signals and factors, but all resolve elongation complexes at regulated points.

Nascent RNA methods are indispensable but model-dependent. The field increasingly treats promoter-proximal peaks, gene-body read density, live-cell transcription spots, and nascent RNA contacts as measurements that require kinetic interpretation rather than direct visual readouts of single molecular events.

## Open Questions, Controversies, Deprecated Models, and Common Misconceptions

Open questions:

- How many mechanistically distinct promoter-proximal paused states exist in living cells? Structural work supports multiple NELF and paused Pol II conformations, but genome-wide assays often compress these states into one promoter-proximal signal.
- How does Integrator choose among processing, termination, and regulatory outcomes? Integrator has broad promoter-proximal roles, but locus-specific determinants remain incompletely resolved.
- When is an R-loop a regulated intermediate, and when is it a genome-instability lesion? R-loop-dependent promoter-proximal termination supports a regulatory role in some contexts, but persistent R-loops can threaten genome integrity.
- How accurately can genome-wide nascent RNA profiles be converted into absolute rate constants? Methods are improving, but many datasets still cannot fully separate initiation, pause entry, pause duration, elongation, premature termination, processing, and decay.

Common misconceptions:

- "A high promoter-proximal Pol II peak always means stronger pausing." A promoter peak can reflect initiation, pausing, release, premature termination, RNA stability, or technical recovery. The correct question is which kinetic step changed.
- "Elongation factors simply speed up polymerases." Many elongation factors stabilize pauses, recruit processing or termination factors, alter chromatin interactions, or change polymerase susceptibility to other regulators.
- "Termination happens only at the end of protein-coding genes." Termination can occur near promoters, at enhancer RNAs, within noncoding transcription units, downstream of polyadenylation sites, at small RNA genes, and at viral or organellar signals.
- "R-loops are always damage." R-loops can be regulatory intermediates, but their persistence, position, and resolution determine whether they become harmful.

Deprecated or weakened claims:

- Deprecated simplification: "The torpedo model and allosteric model of Pol II termination are mutually exclusive." Current evidence supports cooperation among cleavage, exonuclease action, allosteric changes, pausing, chromatin, and processing-factor exchange.