When DNA is just DNA
A new collaborative paper, led by Professor Ed Bolt's group at the University of Nottingham, has just appeared in Nucleic Acids Research – and it has been quite a journey to get here.
I should be upfront: I am not a CRISPR specialist. The intellectual engine behind this project is Ed Bolt, whose lab has been working on CRISPR adaptation for many years. Our group's contribution was to provide the microscopy data that allowed Cas1-Cas2 complexes to be visualised directly inside living bacterial cells. Ed's group, together with collaborators at the University of Zagreb led by Professor Ivana Ivančić-Baće, provided the molecular and genetic framework, and together we made sense of what we were seeing. This is what collaborative science looks like when it works well.
The paper was never supposed to be this large. The story began with observations left over from an earlier study published in 2023 – data that in the end did not fit into the context of this earlier publication. The plan was relatively modest: Ed had been invited to provide a paper for a special issue, and we had planned to use this as an opportunity for a focused follow-up, a few clarifications, some loose ends tied up. That sort of thing.
Science, as usual, had other ideas. The deadline did not quite work with the time we needed to develop the story. Other things got in the way. In the end we had to abandon the idea to contribute to the special issue.
However, a loss is often also an opportunity. As we looked more carefully at the data and combined it with new experiments, a much larger picture began to emerge. More than a year later, the result is a paper that addresses one fundamental questions in CRISPR biology.
The problem with DNA
The question at the heart of the paper sounds deceptively simple: how does a bacterial immune system tell the difference between its own DNA and the DNA of an invading virus?
The obvious answer – that the cell must have some way of recognising foreign DNA directly – turns out to be rather more complicated than it looks. To a protein, DNA is DNA. The chemical structure of viral DNA is essentially identical to the bacterium's own chromosome. There are no colours, no labels, no shapes to distinguish one from the other at the molecular level. Distinguishing self from foreign is, it turns out, a genuinely difficult problem.
Part of the answer may lie in PAM sequences – short DNA motifs that flank the sequences Cas1-Cas2 captures, and which are required for the later interference step that destroys returning invaders. Because PAMs are not carried into the CRISPR locus when a new spacer is stored, the cell's own immune memory is protected from self-attack. But there is an important problem: PAM sequences are common throughout the bacterial chromosome, so they alone cannot fully explain how Cas1-Cas2 selects what to capture in the first place. That is the question this paper addresses.
Our findings suggest that perhaps bacteria do not solve it by directly identifying foreign DNA at all.
A proxy for invasion
For a virus to establish a successful infection, it needs to replicate its DNA rapidly inside the host cell. This process generates specific DNA intermediates – single-stranded gaps, partial duplexes, tailed intermediates – that are either very similar or in many cases identical to the structures produced during the cell's own DNA replication and repair. The Cas1-Cas2 complex, which is responsible for capturing new DNA sequences to build CRISPR immune memory, is strongly attracted to precisely these structures.
The implication is interesting. Rather than asking "is this DNA foreign?", the cell may effectively be asking "is there unusually active DNA replication going on here?" – and using the answer as a proxy for infection. It is a subtle distinction, but an important one. Instead of recognising invaders directly, the cell may be monitoring the consequences of invasion.
This was shown by directly visualising Cas1-Cas2 complexes inside living bacterial cells – something that had not been done before, and where our microscopy contribution came in. The complexes accumulated at DNA structures generated behind advancing replication forks, the regions where gap repair normally occurs. When replication was active, the complexes were present. When replication was halted, they largely disappeared. When gap repair was disrupted – making the relevant DNA structures more abundant and more accessible – Cas1-Cas2 activity increased substantially.
Viewed from this angle, viral replication becomes its own undoing. The very process that allows a virus to multiply inside a cell simultaneously generates the DNA structures that allow the bacterial immune system to capture fragments of viral DNA and build memory against future infections. The host is using the invader's own replication machinery against it.
Where the paper came from
There are a few additional aspects about this paper that I want to mention, because they matter.
The first author, Amin Hashemloo, originally joined my laboratory as an undergraduate student on work experience. He went on to complete his Final Year Project with us, and continued working in the lab through the following summer. The microscopy work in this paper is, in very large part, his. This publication would not exist in its current form without him, and seeing that contribution recognised in a major international journal is one of the satisfying aspects of this work.
In addition, Ed Bolt, Ivana Ivančić-Baće and I all worked in the laboratory of Professor Bob Lloyd at the University of Nottingham. Incidentally, we never directly collaborated while we were there – our paths diverged as we each established independent research groups – but we found our way back to each other years later, and this paper is one result of that. Bob's influence on all three of our scientific careers has been considerable, and though indirectly, he still has influenced the paper by the way his mentoring has developed us into scientists. It feels right to acknowledge that here.
Research papers present polished final stories. What they rarely show is the years of discussion, dead ends, unexpected observations and gradual shifts in thinking that shape the result. This paper is a good example. What began as a small attempt to explain a few lingering observations eventually became a study that may change how we think about one of the central questions in CRISPR biology.
Sometimes very interesting discoveries are hiding in data meant to be used to simply tie up some loose ends.
The full version of the paper can be accessed here.
Similar and related posts: