Base Excision Repair

Base excision repair is the pathway that removes damaged or inappropriate bases that do not grossly distort the DNA double helix. This includes deaminated cytosine (uracil), deaminated adenine (hypoxanthine), oxidized guanine (8-oxo-7,8-dihydroguanine, or 8-oxoG), alkylated bases such as 3-methyladenine, and spontaneous abasic (apurinic/apyrimidinic, AP) sites generated by hydrolysis of the N-glycosidic bond. Because these lesions are chemically small, they do not kink the helix enough for the bulky, distortion-sensing machinery of nucleotide excision repair to notice them efficiently. Instead, cells deploy DNA glycosylases: small proteins, most in the 200–350 amino-acid range, that scan DNA by sliding along the minor groove and testing candidate bases by flipping them out of the duplex into an extrahelical position. A glycosylase does two things once it commits to a base: it catalyzes hydrolysis of the N-glycosidic bond linking the damaged base to the deoxyribose sugar, releasing the free base and leaving an abasic (AP) site, and in many cases it also nicks the phosphodiester backbone immediately adjacent to that site (bifunctional glycosylases with associated AP lyase activity). Monofunctional glycosylases (like uracil-DNA glycosylase) do only the first step, leaving the incision to a separate enzyme, AP endonuclease 1 (APE1 in humans, Nfo/Xth in E. coli). Either way, the outcome is a single-strand break flanked by a 3′-OH and, typically, a 5′-deoxyribose phosphate (5′-dRP) group that must be removed before ligation. The downstream steps — gap filling and ligation — are shared machinery, essentially a miniature, localized version of DNA Replication: An Overview confined to a one- or few-nucleotide patch, using the undamaged complementary strand as template. This is the crucial logical link to DNA Damage and Mutation: An Overview: BER is the pathway of choice precisely because base lesions preserve Watson-Crick geometry closely enough that template-directed resynthesis of one nucleotide is sufficient, unlike bulky lesions that block polymerases and require translesion strategies.

Step 1 — Damage recognition and base flipping. A glycosylase diffuses along DNA, transiently bending it and inserting an amino-acid wedge (often a conserved loop or a specific residue like a leucine or a wedge helix) into the minor groove at each base pair it samples. When it encounters a base that fits its recognition pocket poorly as a normal Watson-Crick partner but well as the target lesion, the enzyme rotates that nucleotide roughly 180° out of the helix and into an extrahelical active-site pocket, while the DNA backbone kinks to fill the vacated space (often with an inserted amino acid side chain occupying the gap left in the helix, "plugging the hole"). The pocket geometrically and chemically excludes normal bases: for example, uracil-DNA glycosylase's pocket is too small to accommodate the 5-methyl group of thymine, so it binds uracil but rejects thymine, giving near-absolute specificity.

Step 2 — Glycosidic bond cleavage. Once flipped and bound, the glycosylase catalyzes hydrolysis of the N-glycosidic bond between the base and C1′ of deoxyribose. Mechanistically this proceeds through an oxocarbenium-ion-like transition state at C1′, attacked by a water molecule (monofunctional glycosylases) or by an internal enzyme nucleophile — often the ε-amino group of a lysine, forming a covalent Schiff-base intermediate with the sugar (bifunctional glycosylases). The bifunctional mechanism then proceeds to AP lyase activity: β-elimination (or β,δ-elimination) cleaves the phosphodiester backbone 3′ to the abasic site, leaving a nick.

Step 3 — Incision by AP endonuclease. If the glycosylase was monofunctional, a free AP site remains, and APE1 hydrolyzes the phosphodiester bond 5′ to the AP site, generating a nick with a 3′-OH and a 5′-deoxyribose phosphate (5′-dRP) flap. APE1 is essential and abundant (roughly 10⁵–10⁶ molecules per human cell), reflecting how central this step is; APE1 knockout is embryonic lethal in mice.

Step 4 — Removal of the 5′-dRP blocking group. The 5′-dRP is not a normal nucleotide and blocks ligation, so it must be excised. In short-patch BER, DNA polymerase β (Pol β) uses its intrinsic dRP-lyase activity (in its 8-kDa N-terminal lyase domain) to remove the dRP group by β-elimination, while its polymerase domain simultaneously fills the single-nucleotide gap using the intact strand as template.

Step 5a — Short-patch BER (the dominant route, ~80–90% of BER events). Pol β inserts one correct nucleotide opposite the template base and removes the dRP moiety, leaving a nick with normal 3′-OH and 5′-phosphate termini. DNA ligase III, in complex with the scaffold protein XRCC1, seals this nick. XRCC1 has no catalytic activity itself but acts as a coordinating hub, physically bridging Pol β, Lig III, APE1, and PARP1 (poly-ADP-ribose polymerase, which detects the nick and helps recruit XRCC1).

Step 5b — Long-patch BER (used when the 5′-dRP is chemically resistant to Pol β's lyase, e.g., oxidized or reduced abasic sites). Here Pol δ or Pol ε, aided by the sliding clamp PCNA (the same clamp used in DNA Replication: An Overview), extends synthesis for 2–10 nucleotides, displacing the old strand containing the dRP as a short flap. The flap endonuclease FEN1 cleaves off this displaced flap, and DNA ligase I seals the remaining nick. Long-patch BER essentially borrows components of the normal replication fork toolkit. Both pathways converge on the same outcome: the original damaged base is gone, replaced through templated synthesis with the correct nucleotide, and the backbone is fully resealed with no net change in DNA length.

A closer look at recognition specificity. Different glycosylases patrol for different lesions, and their pockets are shaped to physically exclude the normal, undamaged base.

Note MUTYH's elegant logic: after 8-oxoG mispairs with adenine during replication, MUTYH removes the adenine, not the still-damaged guanine, giving the replication machinery another chance to insert correct C opposite 8-oxoG — while OGG1 handles the simpler case of 8-oxoG still correctly paired with C. Germline MUTYH mutations cause MUTYH-associated polyposis, a hereditary colorectal cancer syndrome, directly demonstrating this pathway's tumor-suppressive role in humans.

Discovery of the pathway: Lindahl's uracil-DNA glycosylase (1974). Tomas Lindahl, working on the chemical instability of DNA, reasoned that if cytosine deaminates to uracil at an appreciable rate under physiological conditions (he measured roughly 100–500 deamination events per mammalian genome per day at 37°C), cells must possess an enzyme to remove uracil from DNA, or mutation rates would be catastrophic. In 1974, Lindahl isolated an activity from E. coli extracts that released free uracil from DNA containing uracil, without releasing any nucleotide, ribonucleotide, or normal base — the first DNA glycosylase, uracil-DNA glycosylase (UNG). Critically, this enzyme did not act on RNA or on DNA containing thymine, showing precise substrate discrimination between uracil and the sterically similar but methylated thymine. This discovery essentially founded the base excision repair field and was central to Lindahl's share of the 2015 Nobel Prize in Chemistry (with Paul Modrich and Aziz Sancar, for mechanistic studies of DNA repair).

Proving base flipping structurally: crystal structures of glycosylase–DNA complexes. The mechanistic centerpiece of BER — that the enzyme physically rotates the damaged nucleotide out of the double helix — was not obvious from biochemistry alone and required structural proof. In 1995, the crystal structure of human uracil-DNA glycosylase bound to DNA (solved by John Tainer's group) showed the uracil base rotated 180° out of the double helix and inserted into a narrow, positively charged recognition pocket, with a "pinching" loop that intercalated into the DNA and a Leu residue that filled the gap left in the duplex. This structure directly visualized what had previously been inferred only indirectly, and it explained why UNG could discriminate uracil from thymine at the atomic level: the pocket has no room for a 5-methyl group. Subsequent structures of OGG1, MutY (bacterial MUTYH homolog), and AlkA generalized the base-flipping mechanism across essentially the entire glycosylase family, each with a differently shaped pocket matched to its own target lesion.

Proving base flipping kinetically: 2-aminopurine and fluorescence studies. Structures show a single frozen snapshot; to prove that base flipping is a real kinetic step on the reaction pathway (not a crystallization artifact), researchers used the fluorescent base analog 2-aminopurine (2-AP), whose fluorescence is quenched when stacked inside a normal duplex but increases sharply when flipped out into solution or an enzyme pocket. Studies (notably by James Stivers and colleagues in the late 1990s and 2000s) monitored 2-AP fluorescence in real time as UNG engaged DNA substrates, observing a rapid fluorescence increase consistent with extrahelical flipping that preceded catalysis, and stopped-flow kinetics resolved distinct steps of DNA binding, base flipping, and chemical cleavage. This time-resolved fluorescence evidence, combined with the static crystal structures, converted "base flipping" from a hypothesis into an experimentally verified mechanistic step.

Demonstrating the two sub-pathways biochemically: reconstitution with purified proteins. The short-patch versus long-patch distinction was established by reconstituting BER in vitro from purified components (work led by Samuel Wilson, Paul Modrich, and colleagues through the 1990s). Using defined AP-site-containing substrates and purified APE1, Pol β, FEN1, PCNA, and ligases, researchers showed that when the 5′-terminal sugar was a normal dRP, Pol β alone could complete repair via a single-nucleotide patch; but when the AP site was chemically modified (e.g., oxidized or reduced, making it resistant to Pol β's lyase activity), the reaction stalled unless PCNA, Pol δ/ε, and FEN1 were added, which then generated a longer, strand-displaced patch. This defined, in molecular terms, exactly which chemical property of the lesion determines pathway choice.

Genetic evidence — mouse knockouts. Complete loss of core BER genes is embryonic lethal or causes severe phenotypes in mice: Pol β knockout mice die perinatally with major developmental defects, and Ape1 (Apex1) knockout is embryonic lethal at a very early stage, underscoring that unrepaired AP sites and single-strand breaks, if left unresolved, are incompatible with life. Conditional and tissue-specific knockouts of Ogg1 in mice produce elevated 8-oxoG accumulation and increased mutation and tumor rates, directly linking the pathway to genome stability in vivo, consistent with human MUTYH-associated polyposis.

Example 1 — Quantifying the mutational cost of a missing glycosylase. Suppose a cell line is engineered to lack functional UNG (uracil-DNA glycosylase), and independent biochemical measurements show that spontaneous cytosine deamination occurs at a rate of 100 events per genome per day in wild-type cells (Lindahl's estimate, applicable to a genome of roughly 3 × 10⁹ bp), each converting a C:G pair to a U:G mismatch. If unrepaired, replication across a U:G site will pair the uracil with adenine, so half of daughter cells inherit a fixed C→T transition at that site (the other half retain a repaired-looking G on the old strand until the next round, but conservatively assume the mismatch is fixed at the next replication). Over 30 cell divisions (roughly the number separating a zygote from many adult tissue lineages), if none of these lesions are repaired: Total unrepaired deamination events ≈ 100/day × (assume ~1 day per division) × 30 divisions = 3,000 potential C→T transitions accumulating across the lineage, compared to a wild-type background where >99% are corrected by UNG-initiated BER before replication. This is why UNG-deficient cells show a dramatically elevated spontaneous C→T transition mutation rate in reporter assays — the deficit is not subtle, because deamination is constant and background, not something that can wait for an unrelated trigger. This calculation also explains why hypermutation of immunoglobulin genes deliberately co-opts this same chemistry: Somatic Hypermutation uses AID to deliberately deaminate cytosines in antibody genes, and cells partially suppress UNG-initiated repair at those loci to convert some of those U:G lesions into diversifying mutations rather than perfectly repairing them — a physiological example of BER output being tuned rather than simply maximized.

Example 2 — Diagnosing a hereditary colorectal cancer case using MUTYH logic. A 42-year-old patient presents with 30 colorectal adenomatous polyps (an intermediate number — too many for sporadic disease, too few for classic FAP with hundreds to thousands of polyps caused by APC mutations). Genetic testing finds the patient is biallelic (homozygous or compound heterozygous) for pathogenic MUTYH variants, with normal APC sequence. Reasoning through the mechanism: MUTYH normally removes adenine misincorporated opposite 8-oxoguanine during replication, giving APE1/Pol β/ligase a second chance to insert correct cytosine opposite the still-present 8-oxoG. Without functional MUTYH, A:8-oxoG mispairs are not corrected before the next round of replication. On the next replication cycle, the templating 8-oxoG (which base-pairs with A about as readily as with C in the polymerase active site) directs insertion of A opposite it again in one daughter strand, while the original A:8-oxoG mismatch, if it does replicate as an A:T pair in the other lineage, fixes a G:C → T:A transversion permanently. Because this happens stochastically at oxidized guanines genome-wide (oxidative damage is continuous, driven by cellular reactive oxygen species), the somatic mutation burden accumulates specifically as G:C→T:A transversions, enriched in tumor-suppressor genes like APC and KRAS, producing a distinctive polyposis and colorectal cancer phenotype: this is exactly the mutational signature (COSMIC Signature associated with MUTYH deficiency) found in MUTYH-associated polyposis tumors, confirming the mechanism from the patient's genome back to the enzyme. The clinical action: biallelic MUTYH carriers require intensive colonoscopic surveillance and prophylactic colectomy is considered given near-certain colorectal cancer risk if unmanaged.

Lindahl's insight reframed the genome not as an inert archive but as a molecule under continuous chemical assault, defended in real time by dedicated enzymatic machinery. He spent much of his career at Imperial Cancer Research Fund laboratories and later as director of Cancer Research UK's Clare Hall Laboratories (later part of the Francis Crick Institute), building much of the biochemical foundation for base excision repair using purified bacterial and mammalian enzymes. His 1974 discovery of uracil-DNA glycosylase was followed by systematic biochemical dissection of AP endonucleases, DNA ligases, and the coordination between them, establishing BER essentially as a complete pathway years before most of its human components were cloned.

Cancer therapy — PARP inhibitor synthetic lethality. PARP1 detects single-strand nicks generated during BER and helps recruit XRCC1/Lig III; when PARP1 is pharmacologically inhibited (e.g., olaparib, used clinically in BRCA1/2-mutant ovarian and breast cancers), unrepaired BER intermediates persist into replication, collapsing forks into double-strand breaks that require Homologous Recombination to fix. In BRCA-deficient tumors lacking that repair route, this becomes a synthetic-lethal vulnerability, illustrating direct translation from BER mechanism to an approved cancer drug class.

Hereditary cancer screening. Biallelic MUTYH mutation testing is now standard clinical practice for patients presenting with multiple colorectal adenomas and no APC mutation, directly informing surveillance intervals and prophylactic surgery decisions, as worked through above.

Antifolate and nucleoside analog chemotherapy. Drugs like 5-fluorouracil are partly metabolized to fraudulent nucleotides that get misincorporated into DNA and excised by BER-related glycosylase/repair cycling, and thymidylate synthase inhibition (methotrexate, 5-FU) increases dUTP misincorporation, driving "futile cycling" of BER that contributes to the cytotoxic mechanism of these chemotherapies in dividing cancer cells.

Neurodegeneration and aging research. Because post-mitotic neurons rely heavily on BER (rather than replication-coupled pathways) to manage oxidative damage accumulated over decades, BER gene variants and age-related decline in BER capacity are actively studied as contributors to neurodegenerative disease and general aging phenotypes.

Biotechnology and diagnostics. Purified UNG is a standard reagent in molecular diagnostics (the "UNG carryover-prevention" system) to destroy contaminating dU-containing PCR amplicons from previous reactions before a new PCR run, a direct practical repurposing of BER's founding enzyme.

Ask the tutor: How does base excision repair coordinate with translesion synthesis when an abasic site is encountered during active DNA replication, before BER has finished repairing it?