Transcription is the process by which cells copy the information stored in DNA into RNA, and it follows a surprisingly consistent core mechanism across virtually all life on Earth: an enzyme called RNA polymerase pries open the DNA double helix, reads one strand as a template, and stitches together a complementary RNA molecule one building block at a time. The details, though, differ considerably between bacteria, archaea, and the cells of plants and animals, with eukaryotic cells layering on dozens of additional protein factors, elaborate quality-control steps, and even RNA processing that happens while the transcript is still being made. Understanding how all of this works has been one of the most productive areas of modern molecular biology, with structural snapshots now available for nearly every stage of the process.
How Transcription Begins in Bacteria
Bacterial transcription starts when RNA polymerase, together with a detachable subunit called sigma, locates a promoter sequence on the DNA. The sigma factor is what steers the enzyme to the right stretch of DNA and helps it recognize specific signals embedded in the sequence. But finding the promoter is not enough. The enzyme initially sits on the DNA in what researchers call a “closed complex,” unable to access the individual strands. It then undergoes a conformational shift to an “open complex,” in which roughly a dozen base pairs of DNA are melted apart so the template strand is exposed and ready to be read.
Studies of the two major bacterial sigma factors, σ70 and σ54, have shown that both forms of the enzyme start in a default closed state that cannot engage in single-strand binding, and must be actively converted to an open form before transcription can begin.1PubMed Central. Promoter opening by sigma(54) and sigma(70) RNA polymerases: sigma factor-directed alterations in the mechanism and tightness of control Cryo-electron microscopy has captured an intermediate state in which promoter DNA is caught at the entrance of the RNA polymerase cleft before being fully loaded, revealing that the transition from closed to open involves large-scale conformational changes in both the protein and the DNA.2PubMed Central. Structures of Bacterial RNA Polymerase Complexes Reveal the Mechanism of DNA Loading and Transcription Initiation Once the bubble is open, the enzyme begins synthesizing a short RNA. After producing a transcript about ten nucleotides long, the sigma factor typically releases, and the enzyme transitions into a stable elongation mode.
Eukaryotic Initiation Is a Much Bigger Production
In the cells of animals, plants, and fungi, transcription of protein-coding genes is handled by RNA Polymerase II (Pol II), one of three distinct RNA polymerases. Pol I makes ribosomal RNA and Pol III makes transfer RNA and other small RNAs. All three share a similar catalytic core, but they differ substantially on their surfaces, which accounts for their ability to recognize different gene classes.3PubMed. Structure of eukaryotic RNA polymerases
For Pol II, just showing up at a promoter is not sufficient. A large assembly called the pre-initiation complex (PIC) must first be built from Pol II plus a suite of general transcription factors: TFIID, TFIIA, TFIIB, TFIIF, TFIIE, and TFIIH.4PubMed Central. Structural Insights into the Eukaryotic Transcription Initiation Machinery TFIID typically makes first contact with the promoter by recognizing a short AT-rich sequence. The other factors then pile on in an ordered sequence, culminating in the recruitment of Pol II itself. TFIIH is particularly important because it contains a helicase motor (the XPB subunit) that uses the energy of ATP to pry open the DNA around the start site, generating the transcription bubble. Research has shown that TFIIH’s role extends beyond just melting DNA: it also helps the polymerase escape the promoter and avoid stalling during the very first nucleotides of RNA synthesis, and these two functions depend on different stretches of DNA downstream of the start site.5PubMed Central. TFIIH action in transcription initiation and promoter escape requires distinct regions of downstream promoter DNA
Promoter escape is a genuine bottleneck. The polymerase often makes several short RNAs, releases them, and starts over before it finally clears the promoter and commits to making a full-length transcript. Once it does escape, many of the general transcription factors fall away, and the enzyme settles into elongation.
The Elongation Cycle
Whether in bacteria or eukaryotes, the catalytic heart of transcription is remarkably similar. The polymerase maintains a short hybrid of about eight to nine base pairs where the newly made RNA is paired with the DNA template strand. At the front (downstream) edge, it continuously unwinds the DNA; at the back (upstream) edge, it re-anneals the DNA and peels the RNA away.
Each cycle of nucleotide addition follows what biochemists call the two-metal-ion mechanism: two magnesium ions in the active site coordinate the incoming RNA building block (a nucleoside triphosphate) and help catalyze the chemical bond that adds it to the growing chain.6PubMed Central. Nucleotide addition and cleavage by RNA polymerase II: Coordination of two catalytic reactions using a single active site After the bond forms, the polymerase shifts forward by one nucleotide position (translocates) and the cycle repeats. Structural work on the simpler T7 bacteriophage RNA polymerase has shown that two distinct conformations of the enzyme alternate during this cycle: one stabilized by the incoming nucleotide and one that drives the translocation step forward.7PubMed. The structural mechanism of translocation and helicase activity in T7 RNA polymerase
Single-molecule experiments using optical traps, in which a single bacterial RNA polymerase molecule is tethered and its movement along DNA is tracked in real time, have shown that the enzyme generates substantial force as it moves. When researchers applied opposing force to individual polymerase molecules, they found that the motor slows in a way that is fundamentally different from cytoskeletal motors. At high loads, the polymerase appears to stall by sliding backward several base pairs rather than simply stopping at a single step.8PubMed. Force and velocity measured for single molecules of RNA polymerase
Backtracking and Error Correction
That backward sliding is not just a response to physical force. It is a built-in feature that the cell exploits for quality control. During normal transcription, the polymerase oscillates among three states: a pre-translocated state right after adding a nucleotide, a post-translocated state ready to accept the next one, and a backtracked state in which the enzyme has reversed along the DNA and extruded the 3ʹ end of the RNA out through a secondary channel. Crystal structures of Pol II in this backtracked state revealed a dedicated binding site, dubbed the “P site,” that captures the first backtracked nucleotide. When the enzyme has incorporated a wrong base or encountered damaged DNA, the RNA end threads into this pocket, which triggers a proofreading response.9PubMed Central. Structural basis of transcription: backtracked RNA polymerase II at 3.4 angstrom resolution
Proofreading predominantly works by cutting out the mismatched RNA rather than simply reversing the addition. In eukaryotes, the cleavage factor TFIIS inserts into the polymerase’s secondary channel and stimulates the active site to chop off a short fragment (usually a dinucleotide) from the RNA 3ʹ end. Bacteria use analogous factors called GreA and GreB. Cryo-EM reconstructions of E. coli RNA polymerase have captured the entire reaction pathway from a backtracked complex through GreB binding, RNA cleavage, and reactivation with a new substrate nucleotide poised for extension.10PubMed Central. Structural Basis of Transcription: RNA Polymerase Backtracking and Its Reactivation This cleavage-and-restart mechanism does not catch every mistake, but it substantially improves fidelity over what base-pairing alone would achieve.
How Transcription Stops
Termination, the process of releasing the completed RNA and disassembling the transcription complex, works quite differently depending on the organism.
Bacteria use three recognized termination pathways. Intrinsic termination relies entirely on sequences encoded in the DNA and nascent RNA: a stable RNA hairpin forms in the exit channel of the polymerase, followed by a run of uridine residues that form a weak hybrid with the template, and the combination destabilizes the complex enough to cause release. Rho-dependent termination involves a ring-shaped motor protein called Rho that latches onto the RNA and uses ATP to translocate along it, eventually catching up to the polymerase and prying the complex apart. A third, less common pathway involves the DNA translocase Mfd, which pushes stalled polymerases off the template.11PubMed. Mechanisms of Bacterial Transcription Termination
Eukaryotic termination for protein-coding genes is tied to RNA processing. After Pol II transcribes past the polyadenylation signal (a sequence that marks where the mature RNA should end), the endonuclease CPSF73 cuts the nascent RNA. This cleavage has two consequences: the upstream fragment gets a poly-A tail and becomes the mature mRNA, while the downstream fragment is left attached to the still-moving polymerase with a free, unprotected 5ʹ end. An exonuclease called XRN2 then degrades this leftover RNA, chasing the polymerase like a torpedo. At the same time, the enzyme protein phosphatase 1 (PP1) slows the polymerase down over the termination zone, making it easier for XRN2 to catch up. Depleting either CPSF73 or XRN2 causes extensive read-through past normal termination sites, confirming that both the cleavage event and the torpedo are essential parts of a unified termination mechanism.12PubMed Central. A unified allosteric/torpedo mechanism for transcriptional termination on human protein-coding genes
Transcribing Through Chromatin
Bacterial DNA floats relatively freely in the cell, but eukaryotic DNA is wound around histone proteins into structures called nucleosomes. Each nucleosome wraps about 147 base pairs of DNA and presents a serious physical obstacle to a polymerase trying to read through it. The enzyme can force its way past, but it does so slowly and often stalls.
Cells solve this problem with histone chaperones, proteins that temporarily loosen or disassemble nucleosomes ahead of the polymerase and put them back together behind it. The best-studied is FACT (facilitates chromatin transcription). Single-molecule experiments using optical tweezers have shown that FACT dramatically reduces the strength of the nucleosomal barrier by promoting the unwrapping of DNA from one side of the nucleosome, which in turn weakens contacts near the central axis and cuts the time needed for the polymerase to get through.13PubMed Central. FACT weakens the nucleosomal barrier to transcription and preserves its integrity by forming a hexasome-like intermediate Structural studies have captured snapshots of the entire process: FACT adapts its shape to each successive intermediate as the nucleosome is disassembled ahead of the polymerase, then helps reassemble it behind, preserving both chromatin structure and the chemical marks on histones that carry epigenetic information.14PubMed. Structural basis of nucleosome disassembly and reassembly by RNAPII elongation complex with FACT
This reassembly step matters because losing nucleosomes during transcription would erase the histone modifications that help the cell remember which genes should be active or silent. Chromatin is not just packaging; it is a regulatory layer, and the transcription machinery has evolved to read through it without scrambling the message.
RNA Processing Happens While Transcription Is Still Going
One of the more striking features of eukaryotic transcription is that RNA processing does not wait until the transcript is finished. Capping, splicing, and 3ʹ end formation all occur while the RNA is still tethered to the elongating polymerase. The link between transcription and processing is physical: Pol II has a long, unstructured tail called the C-terminal domain (CTD) that extends from its largest subunit. In vertebrates this tail consists of 52 repeating units of a seven-amino-acid motif, and in yeast it has 26.15PubMed Central. The RNA polymerase II CTD coordinates transcription and RNA processing
As the polymerase progresses through a gene, different positions within each repeat get tagged with phosphate groups. Early in transcription, serine-5 phosphorylation recruits the capping enzyme to modify the 5ʹ end of the RNA. Later, serine-2 phosphorylation brings in splicing and polyadenylation factors. Experiments have demonstrated that the CTD independently stimulates capping, splicing, and 3ʹ cleavage, and that these processing events are also interdependent with one another.16Genes & Development. Capping, splicing, and 3′ processing are independently stimulated by RNA polymerase II: different functions for different segments of the CTD The result is an assembly-line arrangement where the polymerase acts as a moving platform that coordinates every step of mRNA maturation.
Enhancers, Mediator, and Long-Range Activation
Eukaryotic genes are not regulated solely by sequences right next to them. Enhancers, regulatory DNA elements that can sit tens or even hundreds of thousands of base pairs away from the gene they control, play a major role in determining when and where a gene is turned on. The prevailing model holds that enhancers physically loop through 3D space to contact the promoter of their target gene, and a large multi-protein complex called Mediator helps stabilize this contact and relay activation signals to Pol II.
Depletion experiments have confirmed that when Mediator is removed from cells, the frequency of enhancer-promoter interactions drops, and gene expression falls sharply along with it.17PubMed Central. The Mediator complex regulates enhancer-promoter interactions Mediator subunits are dynamically recruited to enhancers during gene activation, and knocking them down impairs specific biological processes that depend on enhancer-driven transcription.18PubMed Central. Mediator facilitates transcriptional activation and dynamic long-range contacts at the IgH locus during class switch recombination The picture that emerges is one in which Mediator acts as a bridge: it reads the activation signals delivered by transcription factors bound at enhancers and translates them into the assembly of the general transcription machinery at the promoter.
Phase Separation and Transcriptional Condensates
A newer and more contentious idea in the field is that some of the concentration of transcription factors at active genes is driven by liquid-liquid phase separation, the same physics that causes oil droplets to form in water. Many components of the transcription machinery, including sequence-specific transcription factors, Mediator, and Pol II itself, contain disordered, low-complexity protein regions that can coalesce into droplet-like condensates in the test tube.19PubMed Central. Merging Established Mechanisms with New Insights: Condensates, Hubs, and the Regulation of RNA Polymerase II Transcription These condensates could, in principle, locally concentrate the factors needed for transcription and help partition them away from the rest of the cell.20PubMed Central. Phase separation in transcription factor dynamics and chromatin organization
The debate centers on whether phase separation is truly a driving organizational principle inside living cells or whether the droplets seen in purified systems are an artifact of unusually high protein concentrations. Evidence on both sides is accumulating, and the field has not reached consensus. What is clear is that many transcriptional regulators have the biophysical properties needed to phase-separate, and at least some super-enhancer clusters show behavior consistent with condensate formation.
Transcription Beyond the Nucleus
Eukaryotic cells run a parallel transcription system inside mitochondria, the organelles responsible for energy production. Mitochondria carry their own small circular genome and transcribe it using a dedicated single-subunit RNA polymerase called POLRMT, which is structurally related to bacteriophage polymerases rather than to the multi-subunit enzymes in the nucleus. POLRMT requires two initiation factors, TFAM and TFB2M, to begin transcription. Recent cryo-EM structures of human mitochondrial initiation complexes have captured three intermediate states showing how the promoter is melted and how the start site is selected, including a slippage mechanism in which the initial RNA can begin from a position one nucleotide upstream of the canonical start.21PubMed. Human mitochondrial RNA polymerase structures reveal transcription start site and slippage mechanism
Archaea add another twist. Their cells look like bacteria under a microscope, but their transcription machinery is strikingly eukaryotic: they use a TATA-binding protein and a factor related to TFIIB to initiate transcription. Yet archaeal gene regulation often resembles bacterial systems, with small repressor proteins binding near promoters. A recently characterized archaeal transcription factor was found to use a eukaryotic-style DNA-binding motif while functioning like a bacterial repressor, suggesting that archaea may represent an evolutionary midpoint between the two regulatory strategies.22Cell. An archaeal metabolite-sensing transcription factor bridges prokaryotic and eukaryotic regulatory features
Drugs That Target Transcription
Because transcription is essential for every living cell, it is an attractive target for drugs. Rifampicin, one of the cornerstone antibiotics for treating tuberculosis, works by plugging the RNA exit channel of bacterial RNA polymerase, blocking transcript elongation. Crucially, it does not bind eukaryotic polymerases, which is why it can kill bacteria without poisoning human cells. On the eukaryotic side, the mushroom toxin α-amanitin binds tightly to Pol II and locks the enzyme’s trigger loop in an inactive conformation, halting transcription. This toxin is the main reason that ingesting certain species of Amanita mushrooms is fatal: it shuts down mRNA production in liver cells. Both rifampicin and α-amanitin have been indispensable research tools, and resistant mutations in the polymerase have helped map exactly which parts of the enzyme are critical for each step of catalysis.
Transcriptional Bursting
If you could watch a single gene in a living cell, you would not see it producing RNA at a smooth, constant rate. Instead, genes tend to fire in bursts: a period of active transcription producing several RNA molecules in quick succession, followed by a quiet interval. This pattern, called transcriptional bursting, means that two genetically identical cells sitting side by side can have very different amounts of a given mRNA at any given moment. Computational models fitted to single-cell data have been used to infer the kinetics of these bursts, and more refined models suggest that burst frequency and burst size are often underestimated or overestimated depending on which mathematical framework is used.23PubMed Central. Inferring transcriptional bursting kinetics from single-cell snapshot data using a generalized telegraph model The biological causes of bursting are still being worked out, but chromatin remodeling, transcription factor binding dynamics, and enhancer looping all seem to play roles. The practical upshot is that gene expression is inherently noisy, and cells have evolved downstream buffering mechanisms to cope with that noise.