= Solution
\b[Replication fidelity uses successive discrimination and repair systems, because no single <enzyme> detects every error.] The important distinctions are between an incorrect <base pair>, a slipped sequence register, the wrong sugar incorporated into <DNA>, a damaged template and an incompletely or incorrectly restored <chromosome>. Repair must also identify which information should be retained: recognizing a mismatch without strand discrimination could fix the wrong sequence.
A polymerase can select an incorrect <deoxynucleoside triphosphate> and insert the wrong base through imperfect discrimination, transient base configurations or an altered <active site> environment. If the mismatch is extended and survives into another replication round, it can become a stable <base substitution>. Many replicative polymerases possess a separate 3′→5′ <exonuclease> proofreading site. A mismatched primer end tends to slow extension and transfer toward this site; removal of the terminal <nucleotide> permits another attempt at correct synthesis. This is distinct from the polymerase's forward synthetic reaction and from proofreading in <biological translation>. In mice, selective disruption of <DNA polymerase> δ proofreading increased <mutation> and <tumour> susceptibility, demonstrating an organismal consequence of unrepaired copying errors.<High incidence of epithelial cancers in mice deficient for DNA polymerase δ proofreading by Robert E. Goldsby and others>[Proofreading-domain mutation, mutator phenotype and tumour development]{ref}
Post-replicative <DNA mismatch repair> corrects many errors that escape the polymerase. In <eukaryotes>, <MutSα>, formed by MSH2 and MSH6, recognizes base mismatches and many short insertion–deletion loops; <MutSβ>, containing MSH2 and MSH3, has an important role in larger loops. Recruitment of MutL complexes, strand incision, excision, resynthesis and ligation replaces a stretch of the newly synthesized strand. Nascent-strand discontinuities and replication-associated factors including <PCNA> help direct the process. Purified human activities reconstituted nick-directed repair from defined substrates, showing how a mismatch and a strand-break signal can support directional replacement.<Human mismatch repair: reconstitution of a nick-directed bidirectional reaction by Nicoleta Constantin and others>[Reconstitution using defined nicked substrates and human factors]{ref} The classic <Escherichia coli> <MutH> system instead uses transiently undermethylated daughter <DNA> at suitable sites; that methyl-directed mechanism is not a universal description of eukaryotic <DNA mismatch repair> or even of every bacterium.
<replication slippage> creates a different substrate. In repetitive sequence, transient misalignment can leave an extra loop in the nascent or template strand. A retained nascent-strand loop can produce an insertion; copying past a looped-out template segment can produce a deletion. Short repeats are therefore vulnerable to length changes, and coding-region insertions or deletions can cause a frameshift if their length is not a multiple of three. <DNA mismatch repair> recognizes many such loops, although substrate size and structure affect efficiency. Yeast mismatch-repair mutants destabilized simple repetitive tracts, linking repair defects with <microsatellite instability>.<Destabilization of tracts of simple repetitive DNA in yeast by mutations affecting DNA mismatch repair by Micheline Strand and others>[Repeat-length instability in repair mutants]{ref} This is a DNA-sequence frameshift mechanism, distinct from a ribosome changing <reading frame> during <biological translation>.
The layers can be represented by conditional escape probabilities,
$$
\epsilon_{\mathrm{final}}=\epsilon_{\mathrm{insert}}q_{\mathrm{proof}}q_{\mathrm{MMR}},
$$
where the second and third factors refer to escape among errors reaching the corresponding stage. As a purely illustrative calculation, $10^{-5}\times10^{-2}\times10^{-2}=10^{-9}$ per copied base. These are not asserted universal measured rates: mismatch identity, sequence, polymerase and repair state all matter. The multiplicative logic explains why disabling a later layer can reveal errors that had previously escaped detection without eliminating <DNA> synthesis itself.
Polymerases can also incorporate a <ribonucleotide> instead of a deoxyribonucleotide. <base pairing> may be correct, so ordinary mismatch recognition is insufficient. <RNase H2> can initiate <ribonucleotide excision repair> by incising near an embedded <ribonucleotide>, followed by <DNA> synthesis, removal of the displaced segment and ligation. Reconstitution established this dedicated pathway.<RNase H2-initiated ribonucleotide excision repair by Justin L. Sparks and others>[RNase H2-initiated incision and downstream repair reactions]{ref} <RNase H1> processing of an RNA–DNA hybrid should not be equated with removal of a single isolated <ribonucleotide>. Separately, normal <RNA> primers on <Okazaki fragments> must be processed and the remaining nicks sealed; persistent primers, flaps or unligated junctions threaten <chromosome> integrity even without an incorrect base.
A chemically damaged template can miscode or stall a fork. Small base lesions can be addressed by <base excision repair>, whereas <nucleotide excision repair> removes a damaged oligonucleotide around many bulky, helix-distorting lesions. These pathways repair chemical damage, not every mismatch made on an undamaged template. Replication across an oxidized base can require coordinated recognition of the lesion and the incorrectly paired partner so that repair does not convert damage into a fixed <mutation>. <translesion synthesis> permits passage across some lesions with specialized polymerases; it is damage tolerance, not necessarily lesion removal or error-free copying. <template switching> can provide another route around a block.
Stalled or collapsed forks can leave gaps, broken <DNA> and structural rearrangements. <homologous recombination> can use a sister chromatid to restore information and restart appropriate substrates; inappropriate pairing or <template switching> can instead change sequence organization. <non-homologous end joining> can restore continuity after a break but need not preserve every base, particularly when end processing is required. Replication checkpoints help stabilize forks and delay progression, while under-replication or rereplication requires control of origin use as well as lesion processing. Numerical chromosome-segregation errors are a related genome-stability problem but are not corrected by simply excising one mismatched base. \b[Base selectivity and proofreading act during synthesis, mismatch and <ribonucleotide> repair address distinct residual substrates, and damage/fork-repair pathways restore templates and continuity with different fidelity limits.]
Back to article page
