Primase and the RNA Primer

DNA polymerases are chemically constrained by their catalytic mechanism. Each new nucleotide is added by a nucleophilic attack of a 3′-OH group on the α-phosphate of an incoming dNTP, releasing pyrophosphate. This means the enzyme's active site must already be clamped onto a base-paired 3′ end before it can catalyze anything. There is no equivalent chemistry, in any known DNA polymerase, for joining two free nucleotides together from scratch on a template with no primer. This "primer-dependence" is not a quirk of one polymerase family — it is true of Pol I, Pol III, and every eukaryotic replicative polymerase (Pol α, δ, ε) that has been characterized.

Primase is the enzyme that breaks this deadlock. It is a specialized, low-fidelity RNA polymerase that can initiate a chain de novo, pairing two ribonucleotides together on a single-stranded DNA template without requiring a pre-existing end. Because RNA polymerases (including primase and cellular transcription machinery) are mechanistically capable of de novo initiation while DNA polymerases are not, cells exploit this asymmetry: primase writes a short RNA stretch, and DNA polymerase then takes over, extending from the RNA's terminal 3′-OH as though it were the beginning of a new DNA chain. The resulting RNA segment is deliberately temporary. It is typically 5–10 nucleotides long, complementary to the template, and chemically distinguishable from the surrounding DNA (ribose sugars, occasional mismatches, no proofreading). Every new strand of DNA ever synthesized by a replicative polymerase — on both the leading and lagging strand, in every domain of life — begins its existence as a short piece of RNA. This single mechanistic fact explains why lagging-strand synthesis must be discontinuous, foreshadowing the story of Leading and Lagging Strand Synthesis and the necessary primer-removal machinery of Okazaki Fragment Maturation and DNA Ligase.

Step 1 — Fork opening exposes single-stranded template. As the replicative helicase unwinds the double helix (see The Replication Fork and Strand Separation), two single-stranded templates are exposed, coated by single-strand binding proteins (SSB in bacteria, RPA in eukaryotes) that keep them from reannealing or forming hairpins.

Step 2 — Primase is recruited to the fork. In E. coli, primase (the DnaG protein) does not travel continuously with the fork. Instead it transiently docks onto the hexameric helicase DnaB, forming a loose, dynamic complex sometimes called the primosome. DnaG binds DnaB, is activated, synthesizes a primer, and then dissociates — it is not a stable, always-present passenger. In eukaryotes, primase is permanently tethered as a subunit of a four-protein complex called Pol α-primase: two primase subunits (PRI1/PRI2, catalytic and regulatory) plus the noncatalytic and catalytic subunits of DNA polymerase α (POLA1/POLA2). This physical fusion of "the enzyme that starts RNA" and "the enzyme that starts DNA" is itself a mechanistic solution to the priming problem.

Step 3 — De novo RNA synthesis. Primase reads the single-stranded DNA template 3′→5′ and synthesizes a complementary RNA chain 5′→3′, exactly as any RNA polymerase would, but without needing a promoter or a primer of its own. It begins by binding two free ribonucleoside triphosphates in a specific recognition sequence context (often a pyrimidine-rich initiation site) and catalyzes formation of the first phosphodiester bond between them — the one truly de novo bond-forming step in the entire replication process. It then extends this short RNA by a few more nucleotides.

Step 4 — Length control and handoff. Primase products are short and imprecise: typically 5–10 nt in bacteria (DnaG), about 7–12 nt in eukaryotes' initial RNA primer (further extended in DNA by Pol α to about 20–30 nt total before handoff to Pol δ or ε). Primase itself has no proofreading exonuclease, so RNA primers have a relatively high error rate — tolerable because the entire primer will be destroyed and replaced with DNA later. Once the primer reaches sufficient length, a polymerase switch occurs: the replicative DNA polymerase displaces or is loaded onto the RNA–DNA junction (in eukaryotes, this involves the sliding clamp PCNA and clamp loader RFC handing off from Pol α to Pol δ/ε), and DNA synthesis proceeds from the primer's 3′-OH.

Step 5 — Different fates on the two strands. On the leading strand, priming happens essentially once per origin, since continuous synthesis in the same direction as fork movement means one primer suffices for very long stretches of DNA. On the lagging strand, because synthesis must always proceed 5′→3′ but the fork exposes template in the opposite orientation, DNA polymerase must synthesize in short bursts, each requiring its own new primer, laid down repeatedly by primase as the fork advances. This is why lagging-strand synthesis is fragmented into Okazaki fragments, each beginning with its own RNA primer.

Okazaki's pulse-labeling experiments (Reiji and Tsuneko Okazaki, late 1960s). The first clue that replication was not a smooth, continuous process on both strands came from brief radioactive pulse-labeling of replicating E. coli DNA with [³H]-thymidine. When cells were labeled for very short times (seconds), the newly made DNA sedimented on alkaline sucrose gradients as short fragments (roughly 1,000–2,000 nucleotides in bacteria), not as long continuous chains. With longer chase times, these short fragments were progressively joined into high-molecular-weight DNA. This directly demonstrated that at least part of replication proceeds through small, transient intermediates — now called Okazaki fragments — that are later ligated together, exactly as predicted for lagging-strand synthesis.

Demonstrating that the fragments begin with RNA, not DNA. A separate line of evidence pinned down the chemical nature of the fragment's 5′ end. Several complementary approaches converged:

• Sensitivity to ribonuclease: short pulse-labeled fragments, when treated with alkali or RNase H, lost a small 5′ segment, showing that segment was RNA (alkali specifically hydrolyzes RNA's 2′-OH-containing phosphodiester backbone, leaving DNA intact).
• Genetic and biochemical isolation of primase (DnaG): temperature-sensitive dnaG mutants in E. coli stalled DNA replication at the nonpermissive temperature specifically in lagging-strand synthesis, mapping a distinct gene product to primer synthesis and confirming its enzymatic identity as an RNA polymerase-like activity, distinct from DNA Pol III.
• In vitro reconstitution: purified DnaG, DnaB helicase, SSB, and Pol III could reconstitute coordinated fork movement and fragment synthesis only when primase was included; omitting primase (or blocking it) abolished lagging-strand DNA synthesis even though single-stranded template and DNA polymerase were both present and fully active — directly proving DNA polymerase's absolute dependence on a primer supplied by a separate enzyme.
• Sequencing the 5′ termini of Okazaki fragments: mapping showed a short ribonucleotide stretch (~5–10 nt in bacteria) at the extreme 5′ end of each fragment, exactly at the position predicted for a primase start site, with the rest of the fragment being ordinary deoxyribonucleotide DNA.

Together, these results established the now-standard model: replication forks generate many short RNA-primed DNA fragments on the lagging strand, primase is the enzyme responsible for the RNA starts, and DNA polymerase's inability to initiate de novo is a real, testable biochemical property, not merely inferred from structure.

Eukaryotic confirmation. Later work (from the 1980s onward) purified Pol α as an obligate primase-polymerase complex from yeast and human cells, and showed biochemically that Pol α-primase synthesizes a defined RNA primer (about 7–10 nt) followed immediately, in the same complex, by a short DNA extension (~20 nt) — after which a polymerase-switching mechanism hands the primer terminus to Pol δ or Pol ε for processive synthesis. Structural studies (X-ray crystallography and cryo-EM of primase and of the CMG helicase–Pol α-primase complex) later visualized the active site directly, confirming a two-metal-ion nucleotidyl transfer mechanism shared with other primase/polymerase superfamilies, and showing exactly how the enzyme measures out a primer of characteristic length before releasing it.

Example 1 — Counting primers on a bacterial chromosome. The E. coli chromosome is about 4.6 × 10⁶ base pairs, replicated bidirectionally from a single origin (oriC), so each fork travels roughly 2.3 × 10⁶ bp before forks meet at the terminus. Suppose Okazaki fragments average 1,000 nucleotides (a realistic bacterial figure).

• Leading strand per fork: continuous synthesis from a single primer laid at the origin → 1 primer per fork per leading strand (2 forks × 1 = 2 primers total for both leading strands).
• Lagging strand per fork: number of fragments ≈ 2.3 × 10⁶ / 1,000 ≈ 2,300 fragments, each needing its own primer.
• Two forks moving in opposite directions from oriC → total lagging-strand primers ≈ 2 × 2,300 = 4,600 primers.
• Grand total RNA primers made to replicate the entire chromosome once ≈ 4,600 (lagging) + 2 (leading) ≈ 4,602 primase events.

This calculation shows something important quantitatively: the overwhelming majority of primase activity in a replicating bacterium is devoted to the lagging strand, not the leading strand — consistent with primase acting as a repeatedly-recruited, transient enzyme rather than a stably bound one.

Example 2 — A clinical/molecular reasoning scenario: a hypothetical DNA ligase deficiency vs. a primase defect. Suppose you are given two cell lines with replication defects and asked to distinguish, from a single experiment, whether the defect lies in primase or in DNA ligase I (which normally seals the nick left after RNA primer removal, foreshadowing Okazaki Fragment Maturation and DNA Ligase). You pulse-label both cell lines briefly with [³H]-thymidine and run alkaline sucrose gradients:

• If the defect is in DNA ligase, Okazaki fragments would still be initiated normally (primase still works) but would fail to be joined. You would see an accumulation of short, unligated fragments even after a long chase — the fragments form, but never coalesce into high-molecular-weight DNA.
• If the defect is in primase, fragments could not even be initiated properly. You would instead see a drastic reduction in the overall rate of lagging-strand DNA synthesis from the earliest labeling times, because DNA polymerase has no primer to extend at all — not merely unjoined fragments, but a shortage of fragments in the first place.

The key discriminating observation: ligase mutants show normal fragment formation with a chase defect (short fragments persist), while primase mutants show a primary synthesis defect (fewer/shorter labeled products from the start, because the priming step itself — the true rate-limiting event for each fragment — is broken). This logic mirrors how E. coli dnaG(ts) mutants were originally distinguished from ligase mutants in the classic genetic dissection of the replication pathway.

Reiji Okazaki, working with his wife and close collaborator Tsuneko Okazaki at Nagoya University in the mid-to-late 1960s, designed the pulse-chase experiments that first revealed short, discontinuous DNA fragments during replication. Using extremely brief exposures of replicating E. coli to radioactive thymidine, followed by alkaline sucrose gradient sedimentation, they showed that nascent DNA first appears as short pieces (later named Okazaki fragments in his honor) before being joined into long strands. This discovery, alongside the subsequent identification of primase (DnaG) and the characterization of RNA at fragment 5′ ends by Okazaki's own group and others, established the discontinuous, primer-dependent model of lagging-strand synthesis that is now textbook fact. Reiji Okazaki died in 1975 from leukemia attributed to radiation exposure from the atomic bombing of Hiroshima in 1945, which he had survived as a child; Tsuneko Okazaki continued the research program and remained a leading figure in Japanese molecular biology for decades afterward.

Understanding primase and RNA priming has direct consequences well beyond textbook mechanism. Cancer biology and chemotherapy: because rapidly dividing tumor cells depend heavily on efficient replication initiation, components of the priming and polymerase-switching machinery (Pol α-primase, and downstream primer-processing enzymes) are studied as potential targets for selectively slowing cancer cell proliferation, and defects in primer maturation pathways are linked to genome instability phenotypes exploited by some existing chemotherapeutics. Antibiotic development: bacterial primase (DnaG) is structurally distinct enough from human Pol α-primase that it has been pursued as a target for novel antibacterials, aiming to halt bacterial DNA replication without affecting human replication machinery. Human genetic disease: mutations affecting primer removal and processing (downstream of priming, in genes such as those encoding FEN1 or DNA ligase I) cause genome instability syndromes and have been linked to cancer predisposition, underscoring that the entire prime-synthesize-remove-seal pathway must work correctly, not just the initial priming step. Biotechnology: the biochemical logic of primase-dependent initiation directly underlies laboratory techniques such as primer-based PCR, Sanger sequencing, and isothermal amplification methods (e.g., helicase-dependent amplification), all of which exploit the same fundamental biochemical requirement — that a polymerase needs a free 3′-OH end before it can act — that cells solve naturally with primase.

Ask the tutor: Why does using a fused Pol α-primase complex, rather than a separate free-floating primase, make sense for eukaryotic lagging-strand priming specifically?