Translesion Synthesis
A replication fork travels at roughly 1,000–2,000 nucleotides per minute in bacteria. It hits chemical damage on the template strand thousands of times per cell cycle. A single unrepaired thymine dimer or abasic site can stop the high-fidelity replicase completely. This happens because its tight active site cannot accommodate a distorted or missing base. If the cell simply let every stalled fork collapse, damage would be lethal. It would kill the cell long before it could ever cause a mutation. Instead, cells carry a second, deliberately sloppy set of DNA polymerases. Their entire job is to jam a nucleotide opposite the lesion, however imperfectly. This lets replication continue. This is translesion synthesis (TLS): a damage-tolerance mechanism, not a repair mechanism. It does not remove the lesion — that is the business of Nucleotide Excision Repair and Base Excision Repair. It merely lets the replication machinery get past the lesion, at the cost of accuracy. TLS polymerases have open, solvent-exposed active sites. They have little or no proofreading activity. This is the opposite design philosophy from the replicative polymerases discussed under Proofreading and Replication Fidelity. That tradeoff is the whole story: survival now, mutation later. Why does this matter for a cell, and for a student of molecular biology? Because it reveals that fidelity is not a fixed property of "DNA replication" as a monolithic process. It is a dial that cells turn deliberately. They swap a high-fidelity enzyme for a low-fidelity one at exactly the moment and place where fidelity would otherwise cost the cell its life. TLS is also medically important. It is a major source of mutagenesis after UV or chemotherapeutic DNA damage. It underlies the classic phenomenon of UV-induced ("SOS") mutagenesis in bacteria. Its dysregulation in cancer cells (loss of Pol η in xeroderma pigmentosum variant, XPV) directly causes a human cancer-predisposition syndrome. This article works through the molecular architecture of TLS polymerases. It covers the polymerase-switching mechanism that hands the primer terminus from replicase to bypass polymerase and back. It also covers the classic genetic and biochemical experiments that established the pathway. Finally, it presents worked quantitative examples of mutation rates and switching kinetics.
- Translesion synthesis (TLS) uses specialized, low-fidelity Y-family (and some B-family) DNA polymerases to replicate past DNA lesions that block replicative polymerases.
- TLS polymerases have open active sites lacking a rigid steric checking pocket, and most lack 3′→5′ proofreading exonuclease activity — they trade accuracy for the ability to fit distorted templates.
- Access to the primer terminus is controlled by a "polymerase switch," triggered in eukaryotes by mono-ubiquitination of PCNA (by Rad6/Rad18) and in bacteria by RecA*-mediated cleavage of LexA (the SOS response).
- Some lesions are bypassed accurately ("error-free" TLS, e.g., Pol η across a cyclobutane pyrimidine dimer); others are bypassed with high error rates ("error-prone" TLS, e.g., Pol ι, Pol κ, Rev1/Pol ζ).
- Loss of Pol η causes xeroderma pigmentosum variant (XPV), proving TLS is a genuine, non-redundant human pathway rather than a bacterial curiosity.
- TLS is the principal source of UV- and chemical-carcinogen-induced point mutations, and its overactivity in tumors contributes to chemotherapy resistance and to mutagenesis during cancer evolution.
Translesion synthesis rescues a stalled replication fork by substituting an open-active-site, non-proofreading polymerase for the replicase at the exact moment the fork meets a lesion, trading fidelity for survival.
The core idea
Translesion synthesis is the process by which specialized DNA polymerases replicate directly across DNA lesions — cyclobutane pyrimidine dimers, abasic sites, bulky adducts, cisplatin crosslinks — that stereochemically block the replicative polymerases (Pol III in bacteria, Pol δ/ε in eukaryotes). These replicative enzymes have narrow, rigid active sites that check Watson-Crick geometry before catalysis and possess proofreading exonucleases; a damaged or missing base simply does not fit, and the polymerase stalls, leaving single-stranded template exposed behind the fork. TLS polymerases, mostly of the Y family (Pol η, Pol ι, Pol κ in humans; Pol IV/DinB and Pol V/UmuD′₂C in E. coli), plus the B-family Pol ζ, have solvent-accessible active sites with little or no room-checking and essentially no proofreading, so they can accommodate a lesion and still add a nucleotide, sacrificing fidelity for continuity. The key conceptual distinction from other DNA repair systems is that TLS does not repair anything: the lesion remains in the DNA after TLS, to be removed later (if at all) by excision repair. TLS is a tolerance mechanism operating in cis, at the fork itself, distinguishing it from post-replication daughter-strand-gap repair mechanisms that use the sister chromatid as a template. Two outcomes are possible depending on which TLS polymerase handles which lesion: error-free bypass, where the correct base is inserted opposite the lesion because the polymerase's active site geometry happens to complement it (the paradigm case is Pol η opposite a thymine-thymine cyclobutane dimer, inserting two A's just as Pol III would), and error-prone bypass, where an incorrect or non-complementary nucleotide is inserted, seeding a point mutation at the next round of replication. A useful vocabulary distinction: inserter polymerases (Pol η, Pol ι, Pol κ, Rev1) place a nucleotide directly opposite the lesion, while the extender polymerase Pol ζ (a Pol B-family enzyme built from the Rev3/Rev7 subunits) is comparatively poor at inserting opposite damage but is specialized for extending a primer terminus that is mismatched or sits just past a lesion — a step the inserter polymerases themselves often cannot do efficiently. Many bypass events therefore require a two-polymerase mechanism: one enzyme inserts, a second extends, and then the replicative polymerase resumes.
Mechanism
Step 1 — Fork stalling. The replicative polymerase (Pol III core in bacteria; Pol δ or Pol ε in eukaryotes) encounters a lesion on the template strand that distorts the helix (a UV-induced cyclobutane pyrimidine dimer or 6-4 photoproduct, an abasic site left by spontaneous depurination or by a glycosylase of Base Excision Repair, or a bulky adduct such as a benzo[a]pyrene-guanine lesion). The polymerase's tight active site and its proofreading exonuclease cannot process the abnormal template, so nucleotide addition halts. The replicative helicase, however, can continue unwinding, generating single-stranded DNA (ssDNA) ahead of the stalled polymerase.
Step 2 — ssDNA signal generation. The exposed ssDNA is rapidly coated by single-strand binding protein — SSB in bacteria, RPA (replication protein A) in eukaryotes. This nucleoprotein filament is the universal "damage alarm" signal in both kingdoms, though it triggers different downstream switches.
Step 3a — The bacterial switch (SOS response). RecA polymerizes on the RPA/SSB-coated ssDNA to form an activated nucleoprotein filament, RecA. RecA is a co-protease: it stimulates autocleavage of the LexA repressor, de-repressing roughly 40 SOS genes in E. coli, including umuD, umuC (encoding Pol V), dinB (Pol IV), recA itself, and excision-repair genes. RecA also cleaves UmuD to UmuD′, and the active mutasome is UmuD′₂C — Pol V — which is recruited to the stalled fork by RecA itself and by the β-clamp.
Step 3b — The eukaryotic switch (PCNA ubiquitination). RPA-coated ssDNA recruits the Rad6–Rad18 ubiquitin ligase complex, which mono-ubiquitinates a specific lysine (K164) on PCNA, the sliding clamp encircling the primer-template junction. Mono-ubiquitinated PCNA has increased affinity for the ubiquitin-binding domains (UBZ or UBM motifs) found in Y-family polymerases, and this is the molecular basis of the polymerase switch: Pol δ/ε is exchanged for a TLS polymerase (Pol η, Pol ι, or Pol κ) at the primer terminus, without releasing the clamp from DNA. Polyubiquitination of PCNA (via Mms2-Ubc13 building K63-linked chains) instead channels the lesion into an error-free, template-switching pathway that uses the newly synthesized sister strand — a distinct fork-rescue route, not TLS itself.
Step 4 — Lesion bypass (insertion). The TLS polymerase, now holding the primer terminus at the clamp, inserts one or more nucleotides opposite the lesion. Its open, more solvent-exposed active site — lacking the tight steric "fidelity checkpoint" of a B-family replicase — tolerates the distorted geometry. Whether the inserted base is correct depends on the specific lesion–polymerase pairing:
- Pol η is specialized for CPDs: it inserts two A's opposite a thymine-thymine dimer with an efficiency and fidelity approaching an undamaged template — genuinely accurate ("error-free") bypass.
- Pol ι and Pol κ handle other lesions (minor-groove adducts, 8-oxoguanine, abasic sites) with variable and generally much lower fidelity.
- Rev1, formally a deoxycytidyl transferase rather than a true templated polymerase, inserts a C opposite abasic sites and many bulky N2-dG adducts using a protein-template mechanism (it flips the templating base out and uses an arginine side chain, not Watson-Crick pairing, to specify C).
Step 5 — Extension. Because a nucleotide inserted opposite a lesion (or a Rev1-generated C:lesion pair) often leaves a distorted, poorly paired primer terminus, a second TLS polymerase — most often Pol ζ (Rev3L–Rev7 catalytic core, in complex with Rev1 and PolD2/PolD3 accessory subunits) — takes over to extend a few nucleotides past the lesion until the primer terminus is normal duplex DNA again.
Step 6 — Switch-back. Once the primer terminus has cleared the lesion, PCNA is deubiquitinated (by USP1 in humans) and the replicative polymerase is handed the primer terminus again, restoring high-fidelity synthesis. The lesion itself is still present in the template strand; it will be excised later by NER or BER, or simply diluted out in the daughter duplex that inherited the newly synthesized (formerly damaged) strand's complement. The net outcome: the fork survives, but the sequence opposite the lesion may now be mutated. If the mutation escapes mismatch repair, it becomes fixed in the genome at the next replication cycle.
The evidence
1. Radman and Witkin: discovery of "SOS mutagenesis" (1970s). Miroslav Radman and Evelyn Witkin, working independently and then jointly, found that UV irradiation of E. coli induces a coordinated set of responses — filamentous growth, prophage induction, and increased mutagenesis — that required an intact recA and lexA system. Critically, mutagenesis induced by UV was not simply a passive consequence of damage: it required active gene induction, since recA or lexA(Ind⁻) mutants (which cannot de-repress the SOS regulon) showed almost no UV-induced mutagenesis even though the DNA damage itself was identical. This was the first solid evidence that mutagenesis is an inducible, genetically programmed cellular response — "SOS mutagenesis" — rather than an unavoidable chemical byproduct of damage. It directly implied the existence of damage-inducible, mutagenic polymerase activity, later shown to be Pol V (UmuD′₂C).
2. Echols, Goodman, Walker: biochemical reconstitution of UmuD′₂C (Pol V) activity (1990s). Genetic epistasis had shown that umuD, umuC, and recA were required for UV mutagenesis, but direct biochemical proof came when Roger Woodgate, Myron Goodman, and colleagues purified UmuD′₂C and RecA and reconstituted lesion bypass in vitro on a primer-template containing a defined abasic site or thymine dimer. Purified UmuD′₂C, together with RecA*, the β-clamp, and single-stranded DNA-binding protein, synthesized DNA directly across the lesion — something Pol III core alone could not do on the same template — establishing Pol V as a genuine, biochemically sufficient translesion polymerase, not merely a genetic requirement inferred indirectly.
3. Masutani, Hanaoka, and colleagues: identification of human Pol η and its link to xeroderma pigmentosum variant (1999). Patients with the clinical syndrome xeroderma pigmentosum have severe UV sensitivity and skin cancer predisposition; most have mutations in NER genes (XPA–XPG). But a subset, historically called XP variant (XPV), have completely normal nucleotide excision repair yet are still extremely UV-sensitive and cancer-prone. Chikahide Masutani, Fumio Hanaoka, and coworkers showed that XPV cells lack a functional copy of POLH, the gene encoding DNA polymerase η, and that transfection of wild-type POLH into XPV cells restored accurate translesion synthesis across cyclobutane pyrimidine dimers. This was the decisive piece of evidence that TLS is a distinct, essential human pathway: NER can be entirely intact and a person can still develop severe UV pathology purely from losing one TLS polymerase, proving TLS operates independently of, and cannot be substituted by, excision repair.
4. Crystal structures of Y-family polymerases (Aggarwal, Yang, and colleagues, early 2000s). X-ray crystal structures of Sulfolobus Dpo4 (an archaeal Y-family enzyme used as a structural model) and later of human Pol η bound to a CPD-containing template directly visualized the mechanistic basis for bypass: the active site is markedly more open and solvent-exposed than that of replicative B-family polymerases, lacking the tight "fingers"-domain closure that normally checks Watson-Crick geometry before catalysis, and the CPD can be accommodated within the active site cleft with both damaged thymines still base-pairing (imperfectly) with the two inserted adenines. These structures converted the biochemical/genetic inference of "open active site, low fidelity" into a directly observed physical structure, and explained quantitatively why Pol η has a per-base error rate several orders of magnitude higher than replicative polymerases on undamaged DNA, yet is more accurate than replicases specifically at a CPD (since replicases cannot use the CPD template at all).
5. Stewart, Chen, and colleagues: PCNA ubiquitination as the polymerase switch (2002, yeast and human systems). Using yeast genetics, Stefan Jentsch's and later other groups showed that Rad18 (an E3 ubiquitin ligase) and Rad6 (E2) mono-ubiquitinate PCNA at lysine 164 specifically after DNA damage, and that mutating this lysine (pol30-K164R in yeast) abolishes damage-induced recruitment of Y-family polymerases to chromatin without affecting normal replication. This established the molecular switch mechanism directly, linking a specific post-translational modification to polymerase exchange at the fork — the eukaryotic analog of the RecA*/LexA circuit in bacteria.
Worked examples
Example 1 — Quantifying the fidelity cost of TLS. Suppose a human fibroblast accumulates roughly 100,000 UV-induced cyclobutane pyrimidine dimers (CPDs) after a moderate UVB dose, and NER removes 90% of them before the next S phase, leaving 10,000 CPDs still present in the genome when replication forks arrive. Assume:
- 70% of these residual CPDs are bypassed by Pol η, with an error rate of roughly 1 mispaired base per 300 CPDs bypassed (since Pol η is largely accurate on T-T dimers, but not perfectly so, and its fidelity is worse on the less common C-containing dimers).
- 30% are bypassed by Pol ι or Pol κ instead (e.g., at sites where the dimer geometry or sequence context disfavors Pol η, or where Pol η is unavailable), with a much higher error rate of roughly 1 mispair per 5 lesions bypassed.
Expected mutation-causing events:
This arithmetic (illustrative, using representative literature-consistent error-rate orders of magnitude rather than a single measured dataset) makes a key point vivid: even though Pol η handles the majority of lesions, the minority handled by lower-fidelity backup polymerases contributes the overwhelming share of mutations — roughly 96% of the mutagenic burden in this hypothetical, despite processing only 30% of the lesions. This is exactly the logic behind why XPV patients (who lack Pol η and must route essentially all CPD bypass through the error-prone backups) show dramatically elevated UV mutagenesis and skin cancer incidence, even with fully intact excision repair.
Example 2 — Genetic epistasis in E. coli: distinguishing TLS from other damage tolerance pathways. A classic experiment design: measure UV survival and UV-induced mutation frequency (reversion of a his⁻ auxotroph to his⁻+) in four E. coli strains.
| Strain genotype | UV survival | UV-induced his⁺ revertants |
|---|---|---|
| wild-type | moderate | high |
| ΔumuDC (no Pol V) | moderate (similar to WT) | strongly reduced |
| ΔrecA | very low | ~none |
| ΔpolB (no Pol II) | moderate (similar to WT) | modestly altered |
Reasoning: ΔrecA cells die en masse after UV because RecA is required both for homologous recombination-based fork rescue and for SOS induction — it is upstream of everything, so its loss is catastrophic for survival. ΔumuDC cells, in contrast, survive UV about as well as wild-type (because other tolerance pathways — recombination, other TLS polymerases like Pol II or Pol IV — can still rescue forks) but show a large, specific drop in UV-induced mutagenesis. This dissociation — normal survival but abolished mutagenesis — is the epistatic signature that identifies Pol V (UmuD′₂C) as specifically the mutagenic TLS polymerase, distinguishing its role from RecA's more general, mutagenesis-independent role in fork rescue and recombination. This kind of separation-of-function genetics (survival vs. mutagenesis as two independently measurable phenotypes) is exactly how the "UV mutagenesis requires an active gene product, not just damage" conclusion was rigorously established.
Common misconceptions & traps
| Misconception | Correction |
|---|---|
| "TLS repairs the DNA damage." | TLS bypasses the lesion so replication can continue; the lesion itself remains in the template and must still be removed by NER/BER or diluted by later replication. |
| "TLS polymerases are just defective versions of the replicase." | They are actively selected, structurally distinct enzymes (Y-family fold) specialized for open, damage-tolerant active sites — a functional adaptation, not a broken one. |
| "All translesion synthesis is mutagenic." | Some TLS is accurate ("error-free"), notably Pol η opposite T-T CPDs; mutagenesis depends on which polymerase handles which lesion. |
| "Loss of a TLS polymerase always increases mutation rate." | Loss of an accurate TLS polymerase (like Pol η) increases mutation rate because lesions get shunted to less accurate backups; loss of an inherently error-prone polymerase can sometimes decrease UV mutagenesis. |
| "The SOS response and translesion synthesis are the same thing." | SOS is the broader bacterial transcriptional damage response (~40 genes, including recombination and cell-division genes); TLS via Pol IV/Pol V is only one output of it. |
| "TLS only happens in bacteria." | Y-family polymerases and PCNA ubiquitination-based switching are conserved from bacteria to humans; loss of human Pol η causes a real cancer-predisposition disease (XPV). |
The scientist behind it
Evelyn Witkin began studying X-ray and UV mutagenesis in bacteria in the 1940s, at a time when radiation-induced mutation was assumed to be a straightforward physical-chemical consequence of DNA damage. Working first at Cold Spring Harbor and later at Rutgers, she isolated UV-resistant and UV-sensitive E. coli mutants and noticed something that did not fit the simple picture: some mutants were resistant to UV killing yet showed almost no UV-induced mutagenesis, meaning survival and mutagenesis could be genetically uncoupled. This observation — decades before "SOS" had a name — was the conceptual seed of the entire translesion-synthesis field: mutagenesis after damage is an active, inducible cellular choice, encoded by specific genes, not an inevitable physical byproduct of the lesion itself. Working alongside Miroslav Radman, who proposed the unifying "SOS" model of coordinately induced damage responses, Witkin's genetics and Radman's model together predicted the existence of damage-inducible mutagenic polymerases years before UmuD′₂C (Pol V) was purified and shown biochemically to synthesize DNA across lesions.
Origins · 1 minEvelyn M. Witkin and how this was discovered
Where this matters today
Cancer predisposition and diagnosis. Xeroderma pigmentosum variant (XPV), caused by loss-of-function mutations in POLH (Pol η), is a direct clinical demonstration of TLS's importance: patients have completely normal excision repair yet develop severe photosensitivity and markedly elevated skin cancer risk purely because CPD bypass is rerouted onto error-prone backup polymerases. Genetic testing for POLH mutations is used in the differential diagnosis of XP-like photosensitivity syndromes.
Chemotherapy resistance. Many chemotherapeutic agents (cisplatin, alkylating agents) work by generating bulky DNA adducts and crosslinks intended to block replication and kill tumor cells. Tumor cells that upregulate TLS polymerases — particularly Pol η and the Pol ζ/Rev1 extension complex — can bypass these adducts and continue proliferating, contributing to acquired chemoresistance. This has made TLS polymerases, especially Rev1 and Pol ζ, active targets for small-molecule inhibitors intended to re-sensitize tumors to platinum-based chemotherapy.
Mutational signatures in cancer genomics. Genome sequencing of tumors reveals characteristic "mutational signatures" — patterns of base substitution linked to specific mutagenic processes. UV-associated tumors (melanoma) show a signature dominated by C→T transitions at dipyrimidine sites, the direct fingerprint of TLS polymerase misincorporation opposite unrepaired CPDs; this signature is now used clinically and forensically to infer UV exposure history from tumor genomes.
Antibiotic development and bacterial mutagenesis. Because SOS-induced TLS (Pol IV, Pol V) contributes to bacterial adaptive mutagenesis and antibiotic-resistance evolution under stress (including sub-lethal antibiotic exposure that itself induces the SOS response), inhibitors of RecA* or of the Umu proteins are being explored as "anti-evolution" adjuvants intended to slow the emergence of resistance during antibiotic treatment.
Basic understanding of mutagenesis and evolution. TLS is a conceptual cornerstone for understanding that mutation rate is not a fixed genomic constant but a regulated, damage-responsive variable — directly relevant to models of stress-induced mutagenesis and to how populations under environmental pressure can generate genetic variation faster than baseline replication error alone would predict, connecting molecular mechanism to population-level evolutionary dynamics.
Ask the tutor: How does the choice between translesion synthesis and recombination-based template switching get decided at a given stalled fork?
Is translesion synthesis a form of DNA repair?
No, and this is the single most important distinction to keep straight. DNA repair pathways like Nucleotide Excision Repair and Base Excision Repair physically remove the damaged base or nucleotide and resynthesize correct DNA using the intact complementary strand as a template. Translesion synthesis does the opposite: it leaves the lesion in place and simply lets a specialized polymerase synthesize a new strand across it, tolerating the damage rather than removing it. The lesion persists in the template strand and must eventually be excised by NER/BER, or it is effectively "diluted" because only one of the two daughter duplexes inherits it. TLS is best classified as a damage-tolerance mechanism, alongside recombination-based template switching, not as a repair mechanism.
Why don't replicative polymerases like Pol δ or Pol III just bypass lesions themselves?
Replicative (B-family) polymerases evolved for maximum fidelity on undamaged templates, achieving roughly one error per 10⁷ bases after proofreading (as discussed under Proofreading and Replication Fidelity). This fidelity comes from a tight, closed active site that geometrically checks Watson-Crick base pairing before allowing catalysis, plus a 3′→5′ exonuclease that excises mismatches. A damaged or missing template base does not fit this checking geometry at all, so the polymerase simply cannot catalyze the reaction — it stalls rather than making an error. TLS polymerases sacrifice this precision, using open active sites that can physically accommodate a distorted base, at the cost of much higher error rates even on undamaged DNA.
What is the difference between "error-free" and "error-prone" translesion synthesis?
Both terms describe TLS across the same lesion but with different fidelity outcomes, depending on which polymerase performs the bypass. Error-free TLS inserts the correct, Watson-Crick-complementary base opposite the lesion — the classic example is Pol η inserting two adenines opposite a thymine-thymine cyclobutane pyrimidine dimer, essentially replicating the pre-damage sequence. Error-prone TLS inserts an incorrect or non-complementary base, seeding a point mutation once that strand is replicated again. The same lesion bypassed by a different polymerase (say, Pol ι instead of Pol η) can shift the outcome from largely error-free to substantially error-prone, which is exactly what happens pathologically in Pol η-deficient (XPV) cells.
How does the bacterial SOS response relate to translesion synthesis?
The SOS response is the broader transcriptional program in bacteria, induced when RecA polymerizes on single-stranded DNA exposed at damaged or stalled forks, forming RecA and triggering autocleavage of the LexA repressor. This de-represses roughly 40 genes involved in recombination, cell-division arrest, excision repair, and — critically for TLS — the umuDC genes encoding Pol V, and dinB encoding Pol IV. Translesion synthesis via Pol IV and Pol V is thus one specific, mutagenic output of the much larger SOS regulon, not synonymous with it. This is why SOS induction and TLS-mediated mutagenesis can be experimentally separated using strains carrying only umuDC* deletions.
What happens if TLS itself fails at a lesion — is the cell doomed?
Not necessarily; TLS is one of several fork-rescue mechanisms available to the cell. If TLS polymerases cannot bypass a lesion, cells can instead use recombination-based template switching, where the nascent strand invades the newly synthesized sister duplex and uses it (rather than the damaged template) to extend past the gap, an essentially error-free alternative regulated by polyubiquitination of PCNA via the Mms2-Ubc13 pathway. Persistent, unresolved fork stalling can also trigger fork collapse into a double-strand break, engaging homologous recombination repair machinery, or in extreme cases apoptosis. TLS is best thought of as the fastest, most direct rescue option, with recombination as a slower but more accurate backup.
Why does Rev1 count as a translesion polymerase if it barely uses templated base-pairing?
Rev1 is biochemically unusual: rather than reading the templating base by Watson-Crick pairing, it flips the damaged or abasic templating base out of the helix and uses a conserved arginine side chain within its own active site as a structural "protein template," directing insertion of a cytosine almost regardless of what the true templating base was. Functionally this still counts as translesion synthesis because it adds a nucleotide directly opposite an otherwise blocking lesion. Rev1's more essential role, though, is scaffolding: it has protein-interaction domains that recruit Pol ζ and other Y-family polymerases to the stalled fork, making it as much an organizer of the TLS machinery as an inserter polymerase itself.
Does translesion synthesis occur during normal, undamaged replication?
Occasionally, yes — TLS polymerases can be recruited even at non-canonical DNA structures or naturally occurring damage such as spontaneous abasic sites, and some TLS polymerases (notably Pol η) also participate in replicating difficult sequences like common fragile sites and G-quadruplex-forming regions, plus a specialized role in somatic hypermutation at immunoglobulin loci, where Pol η's intrinsically low fidelity is co-opted deliberately to diversify antibody genes. But under normal, damage-free bulk genome replication, TLS polymerases are largely excluded from the fork because mono-ubiquitinated PCNA (the signal that recruits them) is a damage-induced modification, kept at low levels otherwise.