What Is RNA Polymerase and What Does It Do?

RNA polymerase is the enzyme that reads a cell’s DNA and builds a complementary strand of RNA from it, a process called transcription. Every living cell depends on at least one version of this enzyme to convert the genetic instructions stored in DNA into the RNA molecules that carry out or guide protein production, structural tasks, and gene regulation. The enzyme is not a single simple molecule but a complex molecular machine with moving parts, built-in error correction, and sensitivity to a variety of drugs and toxins. Its variations across bacteria, animals, plants, and viruses reveal how central it is to life and how evolution has adapted it to wildly different needs.

The Basic Job

At its core, RNA polymerase does one thing: it reads one strand of a DNA double helix and stitches together a chain of RNA building blocks (nucleotides) that match the DNA template. The enzyme clamps onto the DNA, pries the two strands apart locally to create a small “bubble,” and then moves along, adding one nucleotide at a time to the growing RNA strand. When the enzyme reaches a signal to stop, the finished RNA molecule is released.

That description sounds mechanical, and in many ways it is. The active center of the enzyme, where nucleotides are added, has been described as a molecular engine composed of fixed and moving parts that both carry out the chemistry and serve as the point where regulatory signals and drug molecules converge.¹ High-resolution structures have shown that the active site coordinates metal ions (magnesium) that are essential for catalysis, including a recently identified third metal site that had not been seen in cellular RNA polymerases before.² A pore beneath the active center allows fresh nucleotide substrates to enter and may also provide a route for the RNA transcript to exit during proofreading.³

Three Eukaryotic RNA Polymerases

Bacteria get by with a single type of RNA polymerase, but eukaryotic cells (the kind found in animals, plants, and fungi) have at least three. Each one handles a different category of genes, and they are structurally distinct from one another. Early biochemical work showed that the three enzymes differ in the sizes of their large subunits and some of their smaller ones, making them genuinely separate molecular machines rather than minor variants of each other.⁴

• RNA Polymerase I: transcribes the genes for the large ribosomal RNAs, the structural backbone of ribosomes. These genes are present in many copies and need to be transcribed heavily, since cells require enormous quantities of ribosomes to make proteins.
• RNA Polymerase II: transcribes protein-coding genes into messenger RNA, plus many noncoding RNAs. It is the most studied of the three because it handles the vast majority of genes whose expression varies between tissues, developmental stages, and disease states. Its promoters are highly variable, giving the cell fine-grained regulatory control.⁵
• RNA Polymerase III: transcribes small structural RNAs such as transfer RNAs (tRNAs) and 5S ribosomal RNA. Its promoters are more uniform than those of Pol II, and its initiation complexes are very stable, allowing rapid recycling so the enzyme can crank out many transcripts quickly.⁵

Despite their structural differences, the three polymerases share a conserved core, and each uses a surrounding ring of polymerase-specific helper proteins that account for their different abilities to recognize promoters and respond to regulatory cues.⁶

Getting Started on a Gene

RNA polymerase cannot simply land on any stretch of DNA and start transcribing. It needs to be guided to the correct starting position by a collection of helper proteins. In bacteria, a detachable subunit called sigma (σ) is responsible for recognizing the promoter, the DNA sequence that marks the beginning of a gene. Different sigma factors recognize different promoter sequences, which lets the bacterium redirect transcription toward different gene sets depending on environmental conditions.⁷

In eukaryotes, the startup process is far more elaborate. RNA Polymerase II, for instance, assembles at a promoter together with a set of general transcription factors (TFIIB, TFIID, TFIIE, TFIIF, and TFIIH) to form a large multiprotein-DNA complex that supports accurate initiation.⁸ The full assembly, known as the preinitiation complex, involves roughly 80 different proteins.⁹ That complexity is not just structural overhead. It provides numerous points at which the cell can say “go” or “stop,” layering regulatory precision that bacteria do not need.

Moving Along the Gene

Once transcription initiation is complete and the enzyme clears the promoter, it enters the elongation phase, steadily adding nucleotides to the growing RNA chain. This is not a smooth ride. The enzyme is a ratchet machine, oscillating between forward (productive) and backward (backtracked) states at many positions along the DNA.¹⁰ Forward movement is influenced by the interplay of base-pairing at the upstream edge of the transcription bubble: the RNA-DNA pairing at the rear of the bubble acts as a drag on forward motion, while DNA-DNA re-pairing just behind the bubble helps lock the enzyme in its forward position.¹¹

A critical moving part inside the enzyme is the trigger loop, a flexible element near the active site that flips between “open” and “closed” states with each nucleotide addition. Structural studies suggest that these open-and-closed transitions may be coupled to the grip and release of the DNA strand that is not being copied (the non-template strand), potentially helping coordinate the enzyme’s step-by-step advance along the gene.¹²

Catching Mistakes

RNA polymerase does not copy DNA with the near-perfect accuracy of DNA polymerase, but it is far from sloppy. When the wrong nucleotide is incorporated, it tends to fray away from the DNA template, which stalls the enzyme. The polymerase then backtracks by one position, and a built-in nuclease activity clips off a small piece of RNA containing the error, allowing transcription to resume from the corrected position.¹³

How the enzyme recovers from deeper backtracks, where it slides further than one position, depends on which polymerase is involved. Single-molecule experiments have shown that both Pol I and Pol II recover from shallow backtracks by simply diffusing forward again. For medium-depth backtracks, they rely on RNA cleavage. For very deep backtracks, recovery can stall entirely without help. Pol I has a built-in subunit (A12.2) that slows backward sliding and can cleave RNA up to about 20 nucleotides deep. Pol II, by contrast, depends on an external helper factor called TFIIS, which stimulates cleavage and can rescue the enzyme from backtracks of virtually any depth.¹⁴ This difference means that in eukaryotic cells, having enough TFIIS around is genuinely important for keeping Pol II productive.

How Transcription Ends

Bacteria use two main strategies to stop transcription. The first, called intrinsic termination, relies on a specific sequence in the newly made RNA that folds back on itself to form a hairpin structure, which destabilizes the elongation complex and causes the enzyme to fall off. The second depends on a ring-shaped motor protein called Rho.¹⁵ In the widely accepted model, Rho latches onto a cytosine-rich stretch of the nascent RNA, uses energy from ATP to chase down the polymerase, and either yanks the transcript out or pushes the enzyme forward until it dissociates.¹⁶

Eukaryotic termination is handled differently. For Pol II, it involves signals in the RNA that trigger cleavage and polyadenylation of the transcript, after which the polymerase continues briefly before being dislodged. The details vary among the three eukaryotic polymerases, but the common thread is that termination is tightly linked to the processing of the RNA transcript itself, rather than being a purely mechanical event.

The Tail That Orchestrates Everything

One feature that sets eukaryotic RNA Polymerase II apart from all other polymerases is a long, flexible tail on its largest subunit called the C-terminal domain, or CTD. This tail is not involved in the chemistry of RNA synthesis. Instead, it acts as a landing pad for dozens of other proteins that process, modify, and export the newly made RNA. The CTD undergoes extensive chemical modifications, especially phosphorylation, as the polymerase moves through different stages of the transcription cycle, and these modifications control which processing factors are recruited at each stage.¹⁷ In effect, the CTD turns Pol II into a coordination hub: it links the act of copying DNA with the capping, splicing, and polyadenylation of the resulting messenger RNA, so that these steps happen in order and on time.

Navigating Packed DNA in Eukaryotic Cells

In eukaryotes, DNA is not floating freely in the nucleus. It is wrapped tightly around protein spools called histones, forming structures called nucleosomes. This packaging creates a real physical barrier for RNA polymerase. When a single Pol II complex encounters a nucleosome, it can push through while the nucleosome survives, forming a small internal DNA loop that allows the histone spool to stay largely intact. But when multiple Pol II molecules are transcribing the same stretch in rapid succession, the histone proteins can be displaced entirely.¹⁸

Structural studies have revealed that elongation factors called Elf1 and Spt4/5 cooperate to lower these nucleosomal barriers and increase the enzyme’s ability to keep going. They reshape the leading edge of the transcription complex and adjust the nucleosome’s orientation to favor forward progress, suppressing pauses at specific stalling points on the nucleosome.¹⁹ Other work has shown that the nucleosome actually rotates relative to the polymerase, with about half a turn of DNA lifting off the histone surface to feed into the enzyme’s active-site cleft.²⁰ In living cells, Pol II also forms clusters at active genes. These clusters colocalize with the Mediator complex at enhancer elements and behave like phase-separated droplets, concentrating the transcription machinery where it is needed most.²¹

Plant-Specific Polymerases

Plants took the three standard eukaryotic polymerases and added two more. RNA Polymerases IV and V are unique to the plant kingdom and evolved from Pol II, though they have diverged substantially. Both produce noncoding RNAs that participate in a process called RNA-directed DNA methylation, a gene-silencing mechanism that helps plants defend against viral DNA and jumping genes (transposons).²² Structural analysis of Pol V confirms that while it retains the overall architecture of a multisubunit RNA polymerase, it has acquired specialized features for its noncoding RNA role in methylation.²³ No animals or fungi have polymerases IV or V. Their existence in plants underscores how evolutionary pressure can repurpose an existing enzyme for entirely new biological tasks.

Single-Subunit Polymerases in Mitochondria

The RNA polymerases discussed so far are all multisubunit machines, made of many protein parts assembled together. But there is a completely different lineage of RNA polymerase that works as a single protein chain. This is the type found in mitochondria (the energy-producing compartments of eukaryotic cells) and in certain bacteriophages like T7.

Mitochondria descended from ancient bacteria, so you might expect their RNA polymerase to resemble a bacterial one. Instead, the mitochondrial RNA polymerase of yeast and other eukaryotes is encoded in the nucleus and resembles the single-subunit RNA polymerase of bacteriophages T3 and T7, not the multicomponent bacterial enzyme.²⁴ The evolutionary origins of this single-subunit polymerase may trace back to a duplication and divergence of a DNA polymerase or reverse transcriptase gene, coinciding roughly with the acquisition of mitochondria through endosymbiosis.²⁵

The T7 phage RNA polymerase itself is remarkably self-sufficient. It consists of just one protein and can carry out transcription without any additional factors, recognizing its own promoter, initiating, elongating, and terminating all on its own.²⁶ During initiation, the enzyme stays bound to its promoter while synthesizing the first seven to ten nucleotides, accomplishing this through a dramatic rotation of its promoter-binding domain.²⁷ This simplicity has made T7 RNA polymerase a workhorse of biotechnology, as we will see shortly.

Viral RNA Polymerases

Viruses have pushed the concept of RNA polymerase into territory that cellular life does not occupy. Depending on the type of genome a virus carries, it may need an RNA-dependent RNA polymerase (to copy an RNA genome into more RNA), an RNA-dependent DNA polymerase (better known as reverse transcriptase, used by retroviruses like HIV), or a more conventional DNA-dependent RNA polymerase. These viral polymerases typically function as a single protein that handles promoter recognition, elongation, termination, and sometimes additional enzymatic tasks all in one package.²⁸ Their compactness makes them attractive drug targets, and many antiviral medications work by jamming these enzymes.

Drugs and Toxins That Target RNA Polymerase

Because RNA polymerase is essential for life, it is a prime target for both natural toxins and therapeutic drugs. Two of the best-studied examples illustrate different strategies of attack.

Alpha-amanitin, the lethal toxin in death cap mushrooms (Amanita phalloides), is a potent and specific inhibitor of RNA Polymerase II. It has no effect on Pol I or Pol III, and it works by binding directly to Pol II and interfering with the trigger loop, the moving element that drives nucleotide addition.²⁹ When the trigger loop is jammed, the enzyme can still inch along, but very slowly and with reduced ability to select the correct nucleotide.³⁰ This is why death cap poisoning is so devastating: it shuts down mRNA production across the body, and the liver, which processes the toxin, takes the worst hit.

Rifampicin, a cornerstone antibiotic against tuberculosis, targets the bacterial RNA polymerase instead. It binds deep within the DNA/RNA channel of the beta subunit, more than 12 angstroms away from the catalytic site itself, but it physically blocks the path of the growing RNA chain when the transcript is only two to three nucleotides long.³¹ The drug essentially lets the enzyme get started but then slams the door before the RNA can grow to any useful length. Some rifampicin-family compounds may also interact with a loop of the sigma factor, adding another layer of disruption during the earliest moments of transcription.³² Because the bacterial polymerase differs structurally from human polymerases, rifampicin can kill bacteria without harming the patient’s cells, which is the pharmacological basis of its safety.

RNA Polymerase in Biotechnology and mRNA Vaccines

The simplicity of phage T7 RNA polymerase has turned it into one of the most commercially important enzymes in molecular biology. If you received an mRNA vaccine for COVID-19, the mRNA in that vial was almost certainly synthesized in a test tube using T7 RNA polymerase. The enzyme is given a DNA template, a supply of nucleotide building blocks, and it churns out RNA transcripts in a process called in vitro transcription.

This process is powerful but imperfect. T7 RNA polymerase can produce unwanted byproducts, including double-stranded RNA (which can trigger unwanted immune responses), fragmented transcripts, and uncapped molecules. Manufacturers have addressed these issues through engineered versions of the enzyme and optimized reaction conditions to improve purity.³³ More recently, researchers have demonstrated scalable cell-free production of T7 RNA polymerase itself, achieving high purity and activity in liter-scale reactions, which could reduce cost and complexity for vaccine manufacturing.³⁴

An Ancient Enzyme with a Deep Evolutionary Record

The multisubunit RNA polymerases of bacteria, archaea, and eukaryotes all share a common ancestor, but the family tree is not symmetrical. Archaeal RNA polymerase is structurally far closer to eukaryotic Pol II than to the bacterial enzyme. The archaeal version shares the full set of subunit counterparts found in Pol II, including elements like the clamp-head and jaw domains that grip DNA.³⁵ Even the general transcription factors that archaea use to start transcription are conserved in eukaryotes, making the simpler archaeal system a useful model for studying how eukaryotic transcription works.³⁶

Phylogenetic analyses of the polymerase gene sequences place archaea as a coherent group closely related to the eukaryotic Pol II and Pol III lineages, suggesting that the eukaryotic nuclear transcription machinery arose from an ancestor shared with archaea rather than from bacteria.³⁷ The single-subunit polymerases of mitochondria and phages represent a separate evolutionary lineage entirely, one that likely originated from a DNA polymerase or reverse transcriptase gene rather than from the ancestral multisubunit enzyme.²⁵ So when biologists talk about “RNA polymerase,” they are really referring to at least two independently evolved families of enzymes that converged on the same essential function: copying DNA into RNA.