Dahlmann, H. A., Vaidyanathan, V. G., & Sturla, S. J. (2009). Investigating the biochemical impact of DNA damage with structure-based probes: Abasic sites, photodimers, alkylation adducts, and oxidative lesions. Biochemistry, 48(40), 9347–9359.

Dahlmann et al. (2009) — Biochemical Impact of DNA Damage

Bibliographic Reference

Dahlmann, H. A., Vaidyanathan, V. G., & Sturla, S. J. (2009). Investigating the biochemical impact of DNA damage with structure-based probes: Abasic sites, photodimers, alkylation adducts, and oxidative lesions. Biochemistry, 48(40), 9347–9359. DOI: 10.1021/bi901059k.

Core Argument

DNA lesions — abasic sites, photodimers, alkylation adducts, and oxidative modifications — are the molecular starting point for mutagenesis. Understanding their biochemical behavior is essential for interpreting how they produce sequence artifacts when encountered by DNA polymerases during PCR and sequencing. The authors review chemical probe strategies for four lesion classes and distill the mechanistic principles governing polymerase behavior at damaged templates: the “A-rule” (adenine preferentially inserted opposite abasic sites), base-pairing versus size/exclusion effects in translesion synthesis, and lesion-specific polymerase stalling. These principles are the biochemical foundation for understanding FFPE-induced sequencing artifacts — the same lesion types characterized here are the ones that accumulate in formalin-fixed tissue and generate false-positive variant calls.

Methods

This is a review of chemical and biochemical approaches for probing DNA lesion structure and function. It covers:

• Abasic sites: Aldehyde-reactive probes (ARPs), methoxyamine, non-natural nucleoside probes for translesion synthesis studies
• TT photodimers: Lesion-specific exo/endonuclease probes, stereoelectronic effects on polymerase bypass
• Alkylation adducts (O⁶-Bn-G): Non-natural nucleoside probes under development
• Oxidative lesions (8-oxo-G): Detection and polymerase bypass characterization

Key Findings

• Abasic site formation is the most frequent spontaneous DNA lesion. Background rate: ~10,000 abasic sites per human genome per day. Depurination rate: 3 × 10⁻¹¹ s⁻¹ at physiological pH, ionic strength, and temperature. Depyrimidination is ~20-fold slower. Alkylation and oxidation accelerate abasic site formation substantially. Single base loss destabilizes the DNA duplex by 3–11 kcal/mol.

• The “A-rule” governs polymerase behavior at abasic sites. When DNA polymerases encounter an abasic site, they preferentially insert adenine opposite the lesion. This is the mechanistic basis for why abasic sites in FFPE DNA produce characteristic A:T insertions during PCR — the same A-rule described here operates during library amplification of damaged FFPE templates. The stabilized tetrahydrofuran (THF) analog is the standard experimental model for studying abasic site biochemistry.

• Abasic site structural variants produce different polymerase outcomes. Beyond the hemiacetal form (AP), the oxidative variant 2’-deoxyribonolactone (L) preferentially templates thymine (not adenine) in E. coli, and the bleomycin-induced C4-AP variant blocks primer extension by Klenow fragment. These variant-specific outcomes mean that the mutation spectrum from abasic sites depends on the chemical context of lesion formation — relevant to FFPE where oxidation and alkylation co-occur with spontaneous hydrolysis.

• Chemically reactive probes enable lesion detection and BER inhibition. Methoxyamine binding to abasic site aldehydes blocks AP endonuclease and inhibits DNA cleavage. In colon cancer xenografts, methoxyamine combined with temozolomide blocked BER of alkylation-induced AP sites, sensitizing tumors. This principle — chemically trapping lesions to prevent repair — is the conceptual precursor to UDG pretreatment of FFPE DNA (do2015-sequence-artifacts-ffpe): rather than trapping lesions, UDG actively creates abasic sites at uracil lesions to prevent their amplification.

• Translesion synthesis is governed by both base-pairing and steric effects. Studies with non-natural nucleotide probes demonstrated that polymerase insertion opposite lesions depends on both hydrogen-bonding capacity (base-pairing) and the physical size/shape of the incoming nucleotide (steric exclusion). This dual mechanism explains why different polymerases produce different artifact spectra from the same lesion type — relevant to the Steiert et al. (2023) observation that high-fidelity (family B) polymerases with uracil read-ahead function produce different FFPE artifact profiles than standard Taq.

Concepts Introduced or Used

Abasic site (AP site), A-rule, depurination, depyrimidination, translesion synthesis (TLS), aldehyde-reactive probe (ARP), 2’-deoxyribonolactone, tetrahydrofuran (THF) analog, 8-oxo-G, O⁶-alkylguanine, TT photodimer, base excision repair (BER), methoxyamine, polymerase stalling

Entities Referenced

• AP endonuclease, pol β, Klenow fragment, E. coli DNA polymerase I — polymerases with distinct lesion-bypass behaviors
• Methoxyamine — BER inhibitor; blocks AP endonuclease
• Temozolomide — alkylating chemotherapeutic; combined with methoxyamine in xenograft studies
• Leinamycin, acylfulvene, azinomycin epoxide, NNK, estrogen metabolites — genotoxins that promote abasic site formation

Limitations (as stated by authors)

• The review focuses on chemical probe development; quantitative lesion frequencies in specific biological contexts (including FFPE) are not addressed
• Lesion-specific probes for O⁶-Bn-G and 8-oxo-G are described as “under development” — not yet validated for biological samples
• Polymerase studies reviewed use primarily model templates with single, defined lesions; mixed-lesion templates (the reality in FFPE DNA) are not addressed
• The biochemical principles are established in prokaryotic systems (E. coli); eukaryotic polymerase behavior at the same lesions may differ

Relevance to Clonal Evolution

This paper provides the biochemical substrate for understanding FFPE artifacts at the molecular level. Its relevance operates at the foundation of the measurement pipeline:

1. The A-rule explains why abasic sites cause specific mutation types. When FFPE DNA containing abasic sites is amplified by PCR, the polymerase inserts adenine opposite the lesion per the A-rule → the sequence read shows an A where the original base was (typically G or C, since purines depurinate faster). This produces apparent T→A or C→A transversions in the sequenced strand — precisely the non-C→T artifact types that Steiert et al. (2023) documented as equally prevalent to deamination artifacts but not addressed by UDG treatment. The A-rule is the biochemical mechanism underlying the oxidation-derived artifacts that Steiert et al. observed.

2. Lesion-specific polymerase behavior means artifact spectra are polymerase-dependent. Different DNA polymerases process the same lesion differently: some stall (producing dropout), some follow the A-rule (producing A insertions), some insert T opposite oxidative variants (producing different errors). This provides the mechanistic basis for the Steiert et al. recommendation to use high-fidelity family B polymerases — they have lower bypass efficiency at lesions, reducing artifact formation by stalling rather than misincorporating.

3. Spontaneous lesion rates set the biochemical baseline. The background rate of ~10,000 abasic sites per genome per day means that even in fresh-frozen tissue, some DNA damage is present. FFPE fixation and storage accelerate these rates (through acid-catalyzed hydrolysis, oxidation, and deamination), but the baseline is non-zero. This means that zero-artifact sequencing does not exist — even fresh-frozen DNA carries background damage. The question is always about the artifact rate relative to the biological signal of interest (e.g., subclonal mutations at 1–5% VAF).

4. The Do & Dobrovic UDG strategy has a biochemical cost. UDG creates abasic sites at uracil lesions to prevent their amplification — but the resulting abasic sites then block polymerase extension, reducing amplifiable template count. This is the trade-off: UDG removes C→T artifacts at the cost of reducing the already-limited template pool. For low-input FFPE samples, this trade-off may do more harm than good — a consideration not addressed in Do & Dobrovic but directly implied by the biochemistry described here.