The dangerous art of completing the task
21 January 2026
For decades, molecular biologists have investigated in great detail how DNA replication starts and keeps going. Entire careers have been built on understanding the molecular choreography of initiation and elongation. However, for many years the final step, the conclusion of the duplication process, was assumed to be simple. Two replication machineries simply meet. The proteins involved disassemble. Everything is ligated into complete DNA chromosomes. Easy.
But what if it is not?
The latest study from our lab reveals that in bacteria, the moment when two replication forks meet – an inevitable, routine event in every cell cycle – is inherently destabilizing to the genome. Even when everything goes "right," termination can trigger local genomic instability. When key processing systems fail, the consequences can be rather impressive.
The final handshake hardly anybody studied
Think of DNA replication in bacteria like two high-speed trains (and I do mean high-speed; the replication machinery in the bacterial model Escherichia coli is incredibly fast!) traveling toward each other, each copying one half of the circular chromosome. They must meet with perfect precision. Too early or too late, and you have problems, which can be intuitively understood quite well. But here's what surprised us: even when they meet exactly on time, the handshake itself is dangerous.
We engineered E. coli strains where we could switch fork fusion events "on" and "off" at specific chromosomal locations. By integrating a recombination reporter cassette – essentially a genetic sensor that detects DNA rearrangements – we measured what happens at these meeting points.
The results were unambiguous. Wherever replication forks fused, recombination frequencies increased significantly. This was not in mutant cells with broken DNA repair systems. This was in wild-type bacteria with fully functional quality control machinery. Fork fusion itself triggers genomic instability in bacteria.
When safety nets fail
Cells are not helpless against this threat. Bacteria have evolved an arsenal of proteins to process the problematic DNA structures that form when forks meet. Chief among these are RecG helicase and a suite of 3' exonucleases, molecular machines that normally defuse dangerous intermediates before they can cause harm.
But what happens when these safety systems fail?
We systematically deleted genes encoding RecG and various exonucleases, then used whole-genome sequencing to visualize DNA synthesis across the chromosome. The effects were striking. Cells lacking just RecG showed moderate over-replication in the terminus region. Cells missing exonucleases showed even more modest effects. But cells lacking both RecG and even a single exonuclease showed massive over-replication, synthesis that peaks in the termination area and exceeded the synthesis started at the native replication origin.
This synergistic effect reveals something fundamental: fork fusion generates intermediates so problematic that cells need multiple redundant systems to manage them. Remove those systems, and the termination area becomes a zone of chaos. And we showed before that this chaos can even be so toxic under certain circumstances that cells are unable to survive.
The plot twist we did not see coming
Here is where the story takes an unexpected turn; one that surprised us as much as anyone.
We initially included analysis of R-loops (three-stranded RNA-DNA hybrids increasingly recognized as sources of genome instability) as a control experiment. Cells lacking RNase HI accumulate R-loops; cells lacking RecG helicase might show modest increases. But cells lacking Tus terminator protein? Tus is purely a DNA-binding protein with no known role in R-loop metabolism. We expected nothing.
We found elevated R-loop levels.
Initially, we were not sure what to make of this. To us the effect seemed somewhat tangential to our main story about fork fusion. The findings were interesting, sure, but given the relatively modest effects we treated them as a curiosity, as an outlook at what we would be looking at in future experiments. It was in part Prof. Gary Sharples, the external examiner of Daniel's viva, who gave a significant push to put the entire story in one manuscript. And when all peer reviewers responded with particular excitement to this finding, we realized that we had indeed stumbled onto something that apparently was as significant to at least the expert reviewers as what we thought of as the main message: replication termination also influences R-loop homeostasis.
The fact that both we and the reviewers found different aspects most compelling actually validates something important: the paper reveals the multiple layers of termination-associated genomic instability, each potentially significant in its own right. Fork fusion events will trigger recombination wherever forks meet, but on top of that increased R-loop levels can result in a more global destabilisation of genomic stability.
Why this matters
We believe that these findings help answer a longstanding evolutionary puzzle: why do bacteria exclusively use single replication origins?
Eukaryotes use hundreds or thousands of origins to replicate their genomes. Even archaea with bacterial-sized chromosomes often employ multiple origins. Bacteria could easily do the same. We know they tolerate additional origins without major problems. Yet no bacterial species naturally replicates its chromosome from more than one origin under normal growth.
Our results show that every termination event is a liability. Fork fusions trigger recombination, generate toxic intermediates requiring multiple processing pathways, and disrupt R-loop equilibrium. More origins mean more termination events – more opportunities for instability. There is no doubt that more than one factor will have shaped the evolution of bacterial chromosomes. However, if multiple termination events destabilise the genome, then there will be a selection pressure for fewer of those events.
A new lens on an old process
We began this project thinking about replication termination as a mechanical endpoint – forks meet, synthesis stops, done. We are ending it convinced that termination is an active, error-prone process that cells must carefully manage to maintain genome stability.
For bacteria, this means living with an inherent tension: they must complete replication termination every cell cycle, but doing so threatens the very genome integrity they're trying to preserve. The proteins and systems they've evolved – RecG, exonucleases, R-loop processing machinery are not just insurance against rare accidents. They're essential countermeasures against the routine dangers of finishing the job.
The implications extend beyond bacteria. If termination is problematic with just two fork fusions per cell cycle, what does this mean for eukaryotic cells managing hundreds to thousands of termination events? The mechanisms may differ, but the fundamental challenge remains: in DNA replication, stopping is as difficult as all other parts.
This research was published in Nucleic Acids Research: Goodall et al., "Termination of DNA replication drives genomic instability via multiple mechanisms" (2025)