RNA Fidelity, DNA Proofreading, and Recombination

Three distinct strategies decide how much information a replicator can keep: the intrinsic fidelity of RNA replication, the exonuclease proofreading added by DNA replication, and recombination, which edits information at the level of templates and populations rather than of single base pairs. Together they set the effective mutation rate — and therefore where the error threshold falls and how stable a quasi-species can remain.

RNA replication: fast copying without proofreading

RNA replicases generally extend a growing strand without a 3′→5′ exonuclease domain. A misincorporated nucleotide is therefore neither recognised nor removed: replication is a single, non-editing synthesis step, and the error is carried into the daughter strand. The result is a high per-base error rate, on the order of one error per 10⁴ bases copied. Because the fraction of error-free progeny decays as genome length grows, long genomes cannot be maintained, so RNA genomes stay compact and sit close to the error threshold. The Qβ phage mutagenesis experiments illustrate this regime: instead of a single sharp wild-type sequence, the population appears as a broad mutant cloud distributed around a master sequence.

DNA replication: proofreading buys genome length

DNA polymerases usually carry a 3′→5′ exonuclease activity that inspects the newly added 3′ terminus. On detecting a mismatch, the enzyme reverses direction and hydrolyses the incorrect nucleotide before extension resumes, so the error is removed and the correct base is inserted. Proofreading layered on top of base-pairing selectivity lowers the error rate by two to three orders of magnitude. In error-threshold terms, a higher per-base accuracy q raises the largest genome that selection can maintain: with νmax ≈ ln σm ⁄ (1 − q), a smaller error rate (1 − q) relaxes the constraint and permits larger genomes with stable inheritance. DNA polymerase exonuclease mutants, which lose proofreading, show exactly the predicted increase in mutation rate and a reduced effective genome size.

Recombination: repair and reassortment at the template level

Recombination does not improve base pairing at the polymerase active site. It acts on templates and populations instead. Homologous recombination can reconstruct an intact sequence from two differently damaged templates; template switching can rescue a stalled or error-laden copy; and reassortment can combine beneficial variants into one background while purging deleterious ones. Its payoff depends on co-infection, available homology and the strength of selection, so from the quasi-species point of view recombination modifies the effective mutant load and the population-level mutation–selection balance rather than the per-base error rate itself.

Error threshold and quasi-species consequences

RNA's high error rate constrains νmax and keeps populations near the error threshold; DNA proofreading lowers the error rate and relaxes that constraint; recombination can reduce the effective mutant load or regenerate fit sequences, shifting the distribution of variants on which selection acts. The comparative principle follows directly: RNA fidelity favours rapid variation and compact information, DNA proofreading favours genome expansion and stable inheritance, and recombination favours population-level repair and modular innovation. These are complementary strategies, not mutually exclusive ones, and real replicators combine them to different degrees.

Comparative principle. RNA fidelity: rapid variation, compact genomes, high error load. DNA proofreading: order-of-magnitude lower error rate, νmax raised, large genomes inherited stably. Recombination: no change in base-pairing accuracy, but population-level repair, reassortment and purge of deleterious variants.

Three Fidelity Strategies

Left: RNA replication inserts a mismatched base and, lacking proofreading, leaves it in the daughter strand, so the per-base error rate stays high. Middle: a DNA polymerase's 3′→5′ exonuclease excises the mismatched 3′ nucleotide and inserts the correct one, lowering the error rate by orders of magnitude. Right: recombination joins two differently damaged templates to restore a functional sequence, repairing information at the population level rather than at a single base pair.

QUESTION 1

Why does RNA replication generally produce a higher per-base mutation rate than DNA replication?

RNA templates are single-stranded and therefore cannot be copied accurately at all.

RNA replicases proofread, but only after the genome has been fully synthesised.

RNA replicases typically lack a 3′→5′ exonuclease activity, so misincorporated nucleotides are not corrected during synthesis.

RNA bases pair less stably than DNA bases, so mismatches are thermodynamically favoured.

QUESTION 2

What does 3′→5′ exonuclease proofreading accomplish during DNA replication?

It repairs double-strand breaks by homologous recombination.

It unwinds the double helix ahead of the polymerase.

It removes an incorrectly paired nucleotide from the 3′ end before extension continues.

It chemically modifies the template so that only correct bases can be inserted.

QUESTION 3

If proofreading lowers the per-base error rate (1 − q), what is the predicted effect on the largest genome that selection can maintain, νmax?

νmax is unchanged, because it depends only on the selective advantage σm.

νmax becomes unlimited for any sequence, whatever its selective advantage.

νmax increases, because νmax grows as the error rate (1 − q) falls.

νmax decreases, because proofreading slows replication.

QUESTION 4

Which statement best captures what recombination contributes to the maintenance of information?

It lowers the per-base mutation rate by acting as a second proofreading exonuclease.

It edits mispaired bases directly at the polymerase active site.

It increases the selective superiority σm of every sequence in the population.

It restores functional sequences and combines variants at the template and population level without changing per-base copying fidelity.

QUESTION 5

In quasi-species terms, why do error-prone RNA genomes tend to stay short?

Shorter genomes mutate at a lower rate per base than longer ones.

The probability of producing an error-free copy decays with genome length, so longer genomes fall below the error threshold and are lost.

Shorter genomes are proofread more efficiently by the replicase.

Only short genomes can participate in recombination.

Case Study: Tuning Fidelity in an RNA Virus

Read the scenario below and answer the questions.

A 10-kb single-stranded RNA virus is copied by a replicase that lacks a 3′→5′ exonuclease. Direct sequencing of progeny shows roughly one error per 10⁴ bases copied. The virus persists as a broad mutant cloud around a master sequence, and an attempt to insert a large 15-kb foreign gene fails: the insert disappears within a few passages. Two experiments are proposed. Experiment (i): fuse an exonuclease domain onto the replicase. Experiment (ii): passage the virus at high multiplicity of infection, so that many cells are co-infected with different variants.

1. Using the error-threshold relation, explain why the 10-kb genome sits near the threshold and why the 15-kb insert is lost.

Answer:

The per-genome error load is the product of genome length and the per-base error rate: 10⁴ bases × 10⁻⁴ per base ≈ 1 error per replication. The error-free fraction is therefore about e⁻¹ ≈ 0.37, and since νmax ≈ ln σm ⁄ (1 − q) with ⁻⁴ and ln σm of order 1–2, the maintainable length is on the order of 10⁴ bases — exactly the size of the genome. The 10-kb genome therefore sits at, or just below, the error threshold, which is why it survives only as a mutant cloud around a master sequence. A 15-kb insert pushes the error load above 1 error per copy and the length above νmax, so most progeny carry the insert in a mutated form. Functional copies are progressively diluted and selection cannot accumulate them, so the insert is lost within a few passages.

2. Predict the consequences of experiment (i) adding exonuclease proofreading and experiment (ii) high-MOI co-infection with recombination, for the mutation rate, νmax, the mutant cloud and evolvability.

Answer:

Experiment (i) — engineered proofreading. Adding 3′→5′ exonuclease activity removes misincorporated nucleotides before extension continues, lowering the per-base error rate by roughly two to three orders of magnitude. Because νmax ≈ ln σm ⁄ (1 − q), a ten- to thousand-fold smaller error rate raises the maintainable genome length by a similar factor, so both the 10-kb genome and the 15-kb insert become stably inherited. The mutant cloud collapses toward the master sequence, giving sharper inheritance — but the standing supply of variation falls, so the population's capacity to explore sequence space (evolvability) decreases, and adaptation to changing conditions may slow.
Experiment (ii) — high-MOI co-infection and recombination. The per-base error rate is unchanged, because recombination is not proofreading. However, with many variants in the same cell, homologous recombination and template switching can reconstruct an intact sequence from two differently damaged templates, combine beneficial variants into single genomes, and purge deleterious variants. This lowers the effective mutant load and shifts the population-level mutation–selection balance, effectively buffering the genome against the error threshold without changing q. The benefit is conditional: it requires co-infection, sufficient sequence homology between the partners, and selection that favours the recombined products. Combining both strategies gives the largest νmax together with continued modular innovation.