Researchers at ISTA report that bacteria near starvation use bacteriophages and their own CRISPR-Cas defences to move DNA between cells, at more than 100 times the rate of spontaneous mutation.

Bacteria are not supposed to have sex. They reproduce by splitting in two, and the copy each daughter cell receives is, barring errors, identical to the parent’s. Genetic novelty has to come from somewhere else — from mutation, or from the various mechanisms by which bacteria acquire DNA from their surroundings. A study from the Institute of Science and Technology Austria argues that one of those mechanisms amounts to something closer to sex than has been appreciated, and that it switches on when bacteria are starving.

The work, published in Molecular Biology and Evolution, describes E. coli cells under starvation stress using bacteriophages — viruses that infect bacteria — together with their own CRISPR-Cas immune systems to transfer DNA between cells, which the recipients then incorporate into their genomes and pass to their descendants.

The defence system doing something other than defending

The surprise lies in which machinery is involved. CRISPR-Cas is a bacterial immune system: it stores fragments of viral genetic material and uses them to recognise and destroy returning invaders. Its repurposing as a laboratory gene-editing tool has made it one of the most famous acronyms in biology. Here it appears to be participating in the opposite of defence.

“The twist is that the bacteria make use of CRISPR-Cas, a defense system against bacterial viruses that later became the basis for gene-editing technologies,” said Călin Guet, a professor at ISTA and head of the Guet group. “It turns out that bacteria use these viruses to shuttle DNA between themselves. Furthermore, recipient cells can incorporate this DNA into their own genomes and transmit it vertically to their offspring.”

The image the researchers reach for is pollination: bacteriophages carrying bacterial DNA from cell to cell somewhat as insects carry pollen between flowers. The virus is the vector; the genetic reshuffling is the consequence.

From herd immunity to genetic exchange

The path to the finding ran through an apparently unrelated question. Pavel Payne, now a postdoctoral researcher at ISTA and the study’s first author, had shown during his doctoral work that bacterial immunity can produce something resembling herd immunity — the immune cells in a population shield their non-immune neighbours as well as themselves.

“This is where things got more interesting,” Payne said. “Herd immunity not only protects the bacteria that directly carry the immune system.” That observation raised a further question. “The remaining question was whether the bacteria could use this mechanism to reshuffle their DNA, and if so, how often.”

The answer, according to the study, is yes — and the rate is the part worth noting. The process remains uncommon in bacterial populations in absolute terms, but it occurs at more than 100 times the rate of spontaneous mutation. Since spontaneous mutation is the baseline source of new genetic variation in an asexual organism, a mechanism running two orders of magnitude faster is not a rounding error. The authors conclude that phage immunity can make a substantial contribution to genetic variation.

Why starvation would be the trigger

The process appears specifically when bacteria are close to starvation, and the researchers report a gene linking it to the starvation response. The release does not name that gene, which is unfortunate — it is the mechanistic hinge of the paper and the detail a reader would most want to follow up.

The timing makes evolutionary sense, and Payne frames it that way. “Sex is widespread in nature because it can speed up adaptation,” he said. “From an evolutionary perspective, sexual reproduction can be especially advantageous under stressful conditions. Our findings suggest that E. coli can use a comparable DNA reshuffling strategy under starvation stress.”

The logic is standard population genetics. Recombination breaks up existing combinations of genes and generates new ones, which is wasteful when the current combination is working and valuable when it is not. A population that is comfortable has little to gain from reshuffling; a population that is starving has little to lose. That stress should induce genetic exchange is a prediction the field has made before, in various forms. What is new here is the specific machinery.

Payne also raises a more speculative possibility: that the same mechanism might help purge harmful mutations from a population. In sexually reproducing organisms, recombination allows deleterious mutations to be separated from beneficial ones and eliminated, a function that asexual lineages lack and that leaves them vulnerable to the slow accumulation of damage. Whether bacterial DNA shuffling does the same job is, for now, an idea rather than a finding.

A century of mechanisms, and a gap in the middle

Bacterial gene transfer is not a new subject. Natural competence — the ability of some bacteria to take up naked DNA from their environment — was discovered almost a century ago. Conjugation, in which cells connect directly and pass genetic material, was found in the 1940s, with transduction, the transfer of DNA by viruses, following a few years later. Evidence that bacteria might undergo something like a primitive sexual process emerged in the 1990s.

What has been missing is a clear account of the conditions under which such a process runs and at what rate. A mechanism that exists in principle but operates vanishingly rarely is a curiosity; one that runs at a hundred times the mutation rate under conditions bacteria routinely encounter is a contributor to their evolution.

Two caveats are worth carrying. This is laboratory work in E. coli, the most thoroughly domesticated bacterium in science, and starvation in a flask is not starvation in a gut or a soil. The release reports no information on whether the process operates in other species or in natural environments. And the word “sex” is doing analogical work here: there are no gametes, no mating types and no obligatory exchange — only DNA moving between cells and being inherited. The analogy is useful precisely because it names a function, not a resemblance.

---

Notes

1. “Herd immunity underlies homologous recombination in stationary phase bacteria.” Molecular Biology and Evolution, 2026. DOI: 10.1093/molbev/msag238.
2. Institute of Science and Technology Austria. “Bacterial immunity—or a form of sex?” EurekAlert!, 1 October 2026. https://www.eurekalert.org/news-releases/1145956.