Molecular Biology
RNA Polymerase
DNA-dependent RNA synthesis — promoters, sigma factor, and the eukaryotic Pol I/II/III trio
RNA polymerase is the enzyme that transcribes DNA into RNA — reading the template strand 3'→5' and stitching ribonucleotides into a growing chain 5'→3' at roughly 30 to 85 nucleotides per second. Unlike DNA polymerase it needs no primer: it starts a chain from scratch. It finds a gene by recognizing a promoter, runs the initiation–elongation–termination cycle, and proofreads not with a built-in exonuclease but by backtracking and cleaving its own transcript. Bacteria use one core enzyme (2α, β, β', ω) steered by a dissociable sigma factor, while eukaryotes split the job among three nuclear enzymes — Pol I for ribosomal RNA, Pol II for messenger RNA, and Pol III for tRNA and other small RNAs. Roger Kornberg solved the atomic structure of eukaryotic Pol II and won the 2006 Nobel Prize in Chemistry for it.
- Synthesis directionRNA built 5'→3'
- Elongation rate~30–85 nt/second
- PrimerNone — starts de novo
- Error rate~1 in 10⁴–10⁵
- Bacterial core2α, β, β', ω + σ factor
- Eukaryotic enzymesPol I, II, III
- Structure NobelKornberg, 2006
Interactive visualization
Press play, or step through manually. The visualization is yours to drive — try it before reading on.
Watch the 60-second explainer
A condensed visual walkthrough — narrated, captioned, under a minute.
Why RNA polymerase matters
- It is the gateway of gene expression. Nothing in the cell gets made without first being transcribed. RNA polymerase is the single enzyme that converts the static information in DNA into the working molecules — mRNA, rRNA, tRNA — that build and run the cell. Every decision about which genes are on or off ultimately funnels down to whether this enzyme is allowed to start.
- Regulation happens mostly at initiation. Cells rarely control gene output by degrading finished RNA; they control it by gating RNA polymerase recruitment and promoter escape. Transcription factors, enhancers, repressors, and chromatin state all converge on the question of whether polymerase engages a promoter — making this enzyme the central node of nearly all gene regulation.
- It is a premier antibiotic target. Rifampicin binds the bacterial β subunit and jams RNA extension; it is a frontline tuberculosis drug. Fidaxomicin targets bacterial RNA polymerase to treat Clostridioides difficile. Because bacterial and human polymerases differ structurally, these drugs kill the pathogen while sparing the host.
- It defines the pace of the central dogma. At 30 to 85 nucleotides per second, transcription sets the tempo for how fast a cell can respond to a signal. A 3,000-nucleotide human gene takes roughly a minute to transcribe, and ribosomes in bacteria latch onto the emerging mRNA and translate it before transcription even finishes.
- It underlies the deadliest natural toxins. α-Amanitin from the death-cap mushroom kills by shutting down RNA Pol II, halting mRNA synthesis and causing liver failure. The differential sensitivity of the three eukaryotic polymerases to this toxin is exactly how biochemists first proved there were three separate enzymes.
- It is the engine behind biotechnology. The bacteriophage T7 RNA polymerase, a single-subunit enzyme with an exquisitely specific promoter, drives most in-vitro transcription kits, mRNA vaccine manufacturing, and inducible expression systems. Billions of doses of mRNA vaccine were synthesized by T7 polymerase copying a DNA template in a tube.
- Its errors and stalls cause disease. Mutations in RNA Pol III subunits cause a form of hypomyelinating leukodystrophy; defective transcription-coupled repair, which relies on stalled Pol II as a damage sensor, underlies Cockayne syndrome. When the enzyme cannot start, run, or recover, whole tissues fail.
How RNA polymerase works, step by step
1. Promoter recognition. Transcription begins when the polymerase locates a promoter — a DNA sequence that marks the start of a gene. In bacteria the holoenzyme (core plus sigma) reads two conserved elements: the −35 box (consensus TTGACA) and the −10 Pribnow box (consensus TATAAT), separated by about 17 base pairs. In eukaryotes, Pol II does not read DNA directly; general transcription factors — TFIID (bearing the TATA-binding protein), TFIIA, TFIIB, TFIIF, TFIIE, and TFIIH — assemble a pre-initiation complex over the core promoter and hand the polymerase to the start site.
2. Open-complex formation. The closed complex, with polymerase perched on double-stranded DNA, isomerizes into the open complex by melting roughly 12 to 14 base pairs around the start site to expose the template strand. In eukaryotes the helicase subunit of TFIIH (XPB) uses ATP to pry the DNA apart, creating the transcription bubble.
3. Initiation and abortive cycling. The enzyme aligns two ribonucleoside triphosphates at the active site and forges the first phosphodiester bond — with no primer required. Early on it often releases short abortive transcripts of 2 to 10 nucleotides before committing, scrunching downstream DNA into itself to build the tension needed for promoter escape.
4. Promoter escape and elongation. Once the RNA reaches about 8 to 12 nucleotides, the polymerase breaks its promoter contacts, sigma factor (or the general transcription factors) is released, and the enzyme becomes a highly processive elongation complex. It maintains a moving transcription bubble of about 17 base pairs and an 8- to 9-base-pair RNA–DNA hybrid, adding nucleotides 5'→3' by nucleophilic attack of the RNA 3'-OH on the incoming NTP's α-phosphate, releasing pyrophosphate. The two-metal-ion mechanism, with catalytic magnesium ions coordinated by conserved aspartates, is shared with all nucleic-acid polymerases.
5. Proofreading on the fly. With no dedicated exonuclease, the enzyme corrects errors by backtracking: after a misincorporation it stalls, slides backward, extrudes the erroneous 3' RNA end into a secondary channel, and cleaves it — assisted by GreA/GreB in bacteria or TFIIS in eukaryotes — before resuming. Kinetic selection plus this backtrack-and-cleave editing hold the error rate near one in ten thousand to one hundred thousand.
6. Termination. Elongation ends when the enzyme releases the transcript and dissociates from DNA. Bacteria use intrinsic termination — a GC-rich RNA hairpin followed by a poly-U stretch that destabilizes the hybrid — or rho-dependent termination, where the ATP-driven Rho helicase chases and dislodges the polymerase. Eukaryotic Pol II terminates after the poly(A) signal is transcribed and the transcript is cleaved, with the Xrn2/Rat1 exonuclease degrading the downstream RNA and catching up to the polymerase in the "torpedo" model.
Common misconceptions
- "RNA polymerase reads DNA 5'→3'." It synthesizes RNA 5'→3', but it reads the template strand in the antiparallel 3'→5' direction. The new RNA is a complementary copy of the template and matches the coding (sense) strand's sequence, with uracil in place of thymine.
- "It needs a primer like DNA polymerase." The opposite is true — the ability to start a chain de novo is one of the defining features of RNA polymerase. In fact, DNA replication depends on this: primase, a specialized RNA polymerase, lays down the RNA primers that DNA polymerase then extends.
- "RNA polymerase can't proofread." It has no separate exonuclease, but it does proofread — by backtracking and hydrolyzing its own transcript. It is simply less accurate than a proofreading DNA polymerase, which is acceptable because RNA is transient and made in many copies.
- "Eukaryotes have one RNA polymerase like bacteria." Eukaryotes have three (I, II, III), plus mitochondrial and chloroplast enzymes, and plants add Pol IV and Pol V. Each nuclear enzyme is a large multi-subunit complex dedicated to a different class of RNA.
- "Sigma factor is part of the core enzyme." Sigma is a dissociable specificity subunit. The core (2α, β, β', ω) can polymerize but cannot find promoters; sigma joins to make the holoenzyme, guides promoter selection, then is released after promoter escape and recycles.
- "The promoter is transcribed into RNA." The promoter is a recognition and docking sequence; transcription generally starts just downstream of it. The +1 start site is the first transcribed base, and the −10 and −35 elements upstream are read as DNA, not copied into the transcript.
RNA polymerase vs DNA polymerase
| Property | RNA polymerase | DNA polymerase |
|---|---|---|
| Product | RNA (uracil, ribose) | DNA (thymine, deoxyribose) |
| Substrate | Ribonucleoside triphosphates (NTPs) | Deoxyribonucleoside triphosphates (dNTPs) |
| Synthesis direction | 5'→3' | 5'→3' |
| Template read | 3'→5' | 3'→5' |
| Primer needed | No — starts de novo | Yes — needs a free 3'-OH |
| Proofreading | Backtracking + transcript cleavage (no separate exonuclease) | Dedicated 3'→5' exonuclease |
| Error rate | ~1 in 10⁴–10⁵ | ~1 in 10⁷ (with proofreading) |
| Scope per cycle | Selected genes, many copies each | Whole genome, once |
| Speed | ~30–85 nt/s | ~500–1000 nt/s (bacterial replisome) |
Eukaryotic RNA Pol I vs Pol II vs Pol III
| Property | RNA Pol I | RNA Pol II | RNA Pol III |
|---|---|---|---|
| Location | Nucleolus | Nucleoplasm | Nucleoplasm |
| Main products | 45S pre-rRNA (28S, 18S, 5.8S) | mRNA, most snRNA/snoRNA, miRNA | tRNA, 5S rRNA, U6, 7SL |
| Share of cell transcription | ~50–60% by mass | Regulatory bulk of genes | Short, very abundant RNAs |
| CTD heptad repeats | No | Yes (52 in humans) | No |
| α-Amanitin sensitivity | Resistant | Very sensitive (nanomolar) | Moderately sensitive |
| Promoter type | Core + upstream (SL1/UBF) | TATA / initiator / core promoter | Often internal (type 1, 2, 3) |
| Termination | TTF-I factor | Poly(A) signal + torpedo | Run of T's on non-template strand |
Famous experiments and history
- Discovery of the enzyme (1959–1960). Independent groups led by Samuel Weiss, Jerard Hurwitz, Audrey Stevens, and others detected an enzyme in animal and bacterial extracts that incorporated ribonucleotides into RNA only when DNA was present — the first evidence for a DNA-dependent RNA polymerase. This directly demonstrated that DNA templates RNA, cementing the central dogma.
- The sigma factor (1969). Richard Burgess, Andrew Travers, John Dunn, and Ekkehard Bautz separated a dissociable subunit from purified E. coli RNA polymerase and showed that this factor — named sigma (σ) — was required for specific, promoter-directed initiation but not for RNA synthesis per se. Removing sigma left a core that transcribed randomly; adding it back restored promoter fidelity.
- Three eukaryotic polymerases resolved (1969). Robert Roeder and William Rutter chromatographically separated three distinct RNA polymerase activities from sea-urchin and rat nuclei, and their differential sensitivity to α-amanitin (Pol II ≫ Pol III ≫ Pol I) proved they were separate enzymes with separate jobs — the foundation of eukaryotic transcription.
- Atomic structure of Pol II (2001, Nobel 2006). Roger Kornberg's laboratory solved the crystal structure of yeast RNA Pol II — a 12-subunit, half-million-dalton machine — and later captured it mid-transcription with DNA and RNA in the active site. The structures revealed the bridge helix, trigger loop, and two-metal-ion catalysis, earning Kornberg the 2006 Nobel Prize in Chemistry (his father, Arthur Kornberg, had won for DNA polymerase in 1959).
- Rifampicin and the β subunit. Structural and genetic work showed that rifampicin binds a pocket in the β subunit (product of the rpoB gene) about 12 Å from the active site, sterically blocking the growing RNA at two to three nucleotides. Point mutations in rpoB confer resistance and are the molecular basis of rifampicin-resistant tuberculosis, now detectable by rapid molecular tests.
- Single-molecule transcription. Optical-trap experiments in the 2000s watched individual polymerases move along DNA one base at a time, directly measuring stepping, pausing, and backtracking. These experiments confirmed that the enzyme moves in single-nucleotide increments and that backtracking underlies both proofreading and regulated pausing.
Frequently asked questions
How is RNA polymerase different from DNA polymerase?
Both are template-directed nucleotidyl transferases that build a new strand 5'→3' while reading the template 3'→5', but they differ in four decisive ways. First, substrate and product: RNA polymerase uses ribonucleotide triphosphates and makes RNA (uracil, ribose sugar), while DNA polymerase uses deoxyribonucleotides and makes DNA (thymine, deoxyribose). Second, priming: RNA polymerase initiates de novo — it can join the first two nucleotides on a bare template with no 3'-OH to extend from — whereas DNA polymerase absolutely requires a pre-existing primer with a free 3'-OH, which is why cells make RNA primers for replication. Third, proofreading: most DNA polymerases carry a dedicated 3'→5' exonuclease domain and achieve error rates near one in ten million; RNA polymerase has no separate exonuclease and instead proofreads by backtracking and cleaving, reaching a more modest one error in ten thousand to one hundred thousand. Fourth, processivity and scope: DNA polymerase copies the whole genome once per cell cycle, while RNA polymerase transcribes selected genes repeatedly, producing many RNA copies from one template.
Why doesn't RNA polymerase need a primer?
RNA polymerase can catalyze the very first phosphodiester bond de novo because its active site stabilizes the initiating nucleotide directly, without needing a 3'-hydroxyl handed to it by another enzyme. In the initiation complex the enzyme holds two incoming ribonucleoside triphosphates in the i and i+1 sites and joins them; the initiating 5' nucleotide retains its triphosphate, which is why mature primary transcripts begin with a 5' triphosphate (later capped in eukaryotic mRNA). DNA polymerase, by contrast, cannot start a chain on its own — it can only extend an existing 3'-OH — so replication depends on primase, itself a specialized RNA polymerase, to lay down short RNA primers. This asymmetry is a fundamental division of labor: RNA polymerases start chains, DNA polymerases extend them, which is precisely why every DNA replication event begins with a stretch of RNA.
What is a sigma factor and what does it do?
In bacteria the catalytic core enzyme (subunit composition 2α, β, β', ω) can polymerize RNA but cannot find where to start. The sigma factor is a dissociable specificity subunit that binds the core to form the holoenzyme and recognizes promoter DNA — the −35 element (consensus TTGACA) and the −10 or Pribnow box (consensus TATAAT), spaced roughly 17 base pairs apart. Sigma positions the enzyme, helps melt the DNA to form the open complex, and then is typically released after the enzyme escapes the promoter, so the same sigma can recycle to another core. E. coli has a housekeeping sigma, σ70 (RpoD), plus alternative sigmas that redirect transcription to whole gene sets under stress: σ32 for heat shock, σ54 for nitrogen limitation, σS (RpoS) for stationary phase, and σ28 for flagellar genes. Swapping sigma factors is one of the simplest and fastest ways a bacterium reprograms its transcriptome.
What are RNA polymerase I, II, and III?
Eukaryotes split transcription among three distinct nuclear RNA polymerases, each dedicated to different products. RNA Pol I works in the nucleolus and makes the single large pre-rRNA (the 45S/47S precursor processed into 28S, 18S, and 5.8S rRNA) — the bulk of the cell's transcriptional output by mass. RNA Pol II makes all messenger RNA plus many small nuclear and small nucleolar RNAs and most microRNAs; its largest subunit carries a C-terminal domain (CTD) of tandem heptad repeats (consensus Tyr-Ser-Pro-Thr-Ser-Pro-Ser, 52 repeats in humans) whose phosphorylation couples transcription to capping, splicing, and polyadenylation. RNA Pol III makes short, abundant RNAs: all tRNAs, the 5S rRNA, U6 snRNA, and 7SL. A classic distinguishing test is sensitivity to the mushroom toxin α-amanitin: Pol II is exquisitely sensitive (inhibited at nanomolar levels), Pol III is moderately sensitive, and Pol I is essentially resistant. Plants add Pol IV and Pol V for silencing-related small RNAs.
How does RNA polymerase proofread if it lacks an exonuclease?
RNA polymerase has no separate 3'→5' exonuclease domain like proofreading DNA polymerases, yet it still corrects mistakes through a backtracking mechanism. When a wrong nucleotide is added, the mismatched 3' end is poorly positioned, elongation stalls, and the enzyme slides backward along the DNA and RNA. The frayed or extruded 3' RNA end then enters a secondary channel, and the enzyme's own active site cleaves the RNA endonucleolytically, removing the error along with a few correct nucleotides so synthesis can resume from a clean end. In bacteria this intrinsic cleavage is stimulated by the transcript-cleavage factors GreA and GreB; in eukaryotes the elongation factor TFIIS (SII) performs the analogous role for RNA Pol II. Combined with kinetic selection at the insertion step, this backtrack-and-cleave proofreading lowers the transcription error rate to roughly one in ten thousand to one hundred thousand — good enough because RNA is transient and made in many copies.
How does RNA polymerase know where to start and stop?
Starting is dictated by the promoter, a DNA sequence upstream of the gene. In bacteria the sigma factor reads the −35 and −10 elements; in eukaryotic Pol II genes, general transcription factors (TFIID with its TATA-binding protein, plus TFIIA, B, E, F, H) assemble the pre-initiation complex over the core promoter, defining the transcription start site. Stopping — termination — works differently by system. Bacteria use two modes: intrinsic (rho-independent) termination, where the RNA forms a GC-rich hairpin followed by a run of uracils that destabilizes the RNA–DNA hybrid and ejects the transcript; and rho-dependent termination, where the ATP-driven Rho helicase loads onto the RNA and chases the polymerase to dislodge it. Eukaryotic Pol II terminates after the poly(A) signal (AAUAAA) is transcribed and cleaved, often via the torpedo model in which the Xrn2/Rat1 exonuclease degrades the downstream RNA and catches up to the polymerase. Pol I uses the transcription-termination factor TTF-I, and Pol III stops at a simple run of thymines on the non-template strand (making a poly-U tail on the RNA).
Which drugs and toxins target RNA polymerase?
Because transcription is essential and its enzymes differ between organisms, RNA polymerase is a rich drug target. Rifampicin (rifampin) binds the β subunit of bacterial RNA polymerase, blocking extension of the nascent RNA past two or three nucleotides; it is a cornerstone of tuberculosis therapy, and resistance arises from point mutations in the rpoB gene. Fidaxomicin, used against Clostridioides difficile, inhibits bacterial RNA polymerase at the open-complex step. Because human polymerases are structurally distinct, these antibiotics spare the host. On the toxin side, α-amanitin from the death-cap mushroom Amanita phalloides binds the bridge helix of RNA Pol II and slows translocation, causing lethal liver failure at doses of a few milligrams; its polymerase selectivity (Pol II ≫ Pol III ≫ Pol I) is the basis of the classic assay that first defined the three eukaryotic enzymes. Actinomycin D intercalates DNA and blocks elongation, and is used both as chemotherapy and as a research tool.