Replication Fork Stalling and Restart at DNA Damage

What fork stalling is and where it happens

A replication fork is the Y-shaped junction where the double helix is unwound and each strand is copied. In E. coli the fork advances at about 1,000 nt/s; in human cells each of tens of thousands of forks moves at roughly 1–3 kb/min (~20–50 nt/s). Because bacteria replicate their entire ~4.6 Mb genome from a single origin, a permanently stalled fork is lethal—yet stalling happens at least once per cell cycle.

Stalling means synthesis stops. The physical triggers include:

• Bulky lesions (UV-induced cyclobutane pyrimidine dimers, cisplatin crosslinks) that block the polymerase template.
• Single-strand nicks in the template, which convert a fork into a broken end.
• Tightly bound proteins and transcription–replication collisions (R-loops).
• Nucleotide starvation—e.g., hydroxyurea depleting dNTPs by inhibiting ribonucleotide reductase.
• Difficult-to-replicate sequences like G-quadruplexes, trinucleotide repeats, and common fragile sites.

A hallmark of a stalled fork is uncoupling: the replicative helicase keeps unwinding while the polymerase stops, exposing long stretches of single-stranded DNA coated by SSB (bacteria) or RPA (eukaryotes).

The restart mechanism, step by step

Restart is a decision tree, not a single reaction. The nature of the block dictates which route is taken:

• Step 1 — Sensing. Exposed ssDNA is bound by SSB/RPA. In eukaryotes RPA-ssDNA recruits ATR via ATRIP; ATR phosphorylates CHK1, which stabilizes forks, slows origin firing, and buys time.
• Step 2 — Fork reversal. Specialized translocases run the fork backward. The two newly synthesized strands anneal to each other and the parental strands re-pair, extruding a four-way Holliday-junction-like "chicken-foot" structure. This pulls the nascent 3' ends away from the lesion.
• Step 3 — Lesion handling. The regressed arm gives repair enzymes access to the blocking lesion, or the nascent strand can template past it (template switching), or a translesion polymerase copies across it.
• Step 4 — Restoration. The reversed fork is restored (branch migration/nuclease processing), and in bacteria PriA recognizes the fork, recruits PriB/PriC and DnaC, and reloads the replicative helicase DnaB—an origin-independent restart. A new replisome assembles and synthesis resumes.

Key molecules and characteristic numbers

Bacterial machinery:

• RecG — a ~76 kDa superfamily-2 helicase that drives fork reversal, simultaneously translocating on leading- and lagging-strand templates to form a Holliday junction; it works against large opposing forces to strip proteins off the fork.
• RuvABC — branch-migrates (RuvAB) and resolves (RuvC endonuclease) Holliday junctions.
• PriA — the master restart helicase; loads DnaB origin-independently at ~85% of restart events.
• RecA — forms nucleoprotein filaments for strand invasion; the eukaryotic homolog is RAD51.

Eukaryotic machinery: Three SNF2-family DNA translocases catalyze reversal—SMARCAL1 (recruited via RPA32), ZRANB3 (recruited via K63-polyubiquitinated PCNA), and HLTF. Each cooperates with RAD51, which both promotes reversal and coats the regressed arm to shield it from MRE11, DNA2, and EXO1 nucleases. BRCA1/BRCA2 and FANCD2 are essential fork protectors—their loss lets nucleases chew back nascent DNA at rates of hundreds of nucleotides. A single human S phase fires tens of thousands of origins; under replication stress, thousands of forks stall and are processed by these pathways.

How stalled forks are studied and regulated

The field advanced through several techniques with characteristic readouts:

• DNA fiber / DNA combing labels tracts with sequential CldU/IdU pulses, letting researchers measure fork speed, stalling, and nascent-strand degradation (a shortened green tract after hydroxyurea flags an unprotected reversed fork).
• Electron microscopy of psoralen-crosslinked DNA directly visualizes reversed forks; landmark EM work (Neelsen & Lopes) showed reversal is a physiological, genome-wide response, not an artifact.
• Single-molecule assays and 2-D gel electrophoresis resolve X-shaped reversed intermediates.

Regulation centers on the ATR–CHK1 (eukaryotes) and SOS / RecA-LexA (bacteria) responses. ATR limits origin firing and stabilizes stalled replisomes so forks don't collapse; loss of ATR causes forks to break and generates DNA damage. Post-translational control—PCNA ubiquitination (by RAD18/RAD6 for TLS; by HLTF/SHPRH for template switching), and SMARCAL1 phosphorylation/ubiquitylation—chooses between reversal, TLS, and repriming. Timing matters: too much reversal without protection is toxic, so the pathways are tightly balanced.

How it differs from related processes

Fork stalling/restart is often confused with its neighbors:

• vs. DNA repair (excision repair). Nucleotide excision repair removes a lesion from duplex DNA anywhere in the genome; fork restart specifically rescues a replication machine that has hit a block, and may bypass rather than remove the lesion.
• vs. fork collapse. A stalled fork is intact and restartable; a collapsed fork has lost its replisome or broken into a double-strand end, requiring recombination (RecBCD/BRCA-RAD51) to rebuild.
• vs. translesion synthesis. TLS is one option within restart—a low-fidelity polymerase (Pol V, Pol η) copies over the lesion—whereas reversal and template switching are error-free bypass routes.
• vs. origin re-firing. Dormant-origin firing brings a fresh fork from the other direction; PriA-mediated restart re-launches synthesis at the same stalled fork, independent of any origin.

The unifying theme is damage tolerance: the cell prioritizes finishing replication and preserving fork integrity, deferring definitive repair when necessary.

Why it matters: disease, drugs, and open questions

Replication fork biology sits at the heart of cancer and genome-instability disorders. Mutations in SMARCAL1 cause Schimke immuno-osseous dysplasia; defects in the Fanconi anemia pathway (FANCD2, FANCM) and in BRCA1/BRCA2 destabilize forks and drive tumorigenesis. Because BRCA-deficient cells cannot protect reversed forks, their nascent DNA is degraded and their genomes fragment—the basis for PARP-inhibitor (olaparib) synthetic lethality and for chemosensitivity to cisplatin.

Therapeutically, ATR and CHK1 inhibitors (e.g., ceralasertib, prexasertib) push cancer cells with high replication stress into fork collapse and mitotic catastrophe—an actively pursued oncology strategy.

Open questions remain:

• What determines the choice among reversal, TLS, and repriming at a given fork?
• How is the timing and extent of fork reversal limited to prevent toxic over-remodeling?
• How do PrimPol-dependent gaps get filled, and when do they trigger the checkpoint?
• Can fork-protection factors be drugged to overcome PARP-inhibitor resistance?

Major stalled-fork processing pathways and their molecular machinery

Pathway	Core enzymes	What happens	Cost / risk
Fork reversal	RecG (bacteria); SMARCAL1, ZRANB3, HLTF + RAD51 (eukaryotes)	Parental strands re-anneal, nascent strands pair, forming a 4-way "chicken-foot" junction	Regressed arm vulnerable to MRE11/DNA2/EXO1 nucleases if unprotected
Recombination restart	RecA/RAD51, RuvABC, PriA	Strand invasion / D-loop restores a fork; junction resolved and replisome reloaded	Can cause crossovers, deletions if misregulated
Translesion synthesis (TLS)	Pol IV (DinB), Pol V (UmuD'2C); Pol η, ι, κ, REV1 (eukaryotes)	Low-fidelity polymerase copies directly across the lesion	Mutagenic (error rate up to ~10⁻²–10⁻³)
Repriming / gap tolerance	PrimPol (eukaryotes); DnaG-dependent (bacteria)	Synthesis restarts downstream, leaving a ssDNA gap to fill later	Gaps must be sealed post-replicatively
Direct helicase reload	PriA/PriB/PriC + DnaC → DnaB (bacteria)	Replicative helicase reloaded at an intact fork; origin-independent restart	Requires undamaged fork structure