Bacterial immune system works in reverse, baffling scientists

They weren't. Experiments showed the bacteria could still defend themselves against phages. The system was functional; the researchers simply couldn't see it working with their usual tools. "Whenever we tried the experiments we normally use to study retrons, they would fail," Shipman said, describing what he called "a long molecular odyssey" of inventing new methods.

What is the reversed mechanism?

The team eventually mapped the sequence of events:

• Healthy cells carry little detectable DNA and no noticeable toxin.
• A phage's arrival triggers a sharp increase in DNA production.
• The accumulating DNA interacts with an RNA molecule that had been blocking toxin production.
• That interaction lifts the restriction, and the bacterium rapidly makes the toxic protein.

The toxin is an exceptionally small protein that inserts into the bacterium's cell membrane and destabilizes it, killing the infected cell. In other words, the DNA acts as an activation switch rather than a restraint. "It's almost the mirror-image version of how we traditionally view retrons, because the DNA switches on toxin production instead of inhibiting an already-produced toxin," Zhang said.

A key breakthrough came from Spacer-seq, a technique previously developed in Shipman's lab that repurposes CRISPR-Cas machinery — bacteria's own antiviral system — to record molecular activity inside cells. "We had just about given up hope that any DNA was being made when this approach finally revealed a very, very low level of production in uninfected cells," Zhang said.

What does this change for medicine?

The honest answer: nothing yet, but the starting points are real. Shipman's lab is already exploring two applications. One involves synthetic molecular sensors and switches that regulate gene activity — systems that could control when particular genes turn on or off, useful for studying and eventually treating disease. "This system represents a unique way to potentially build a molecular sensor and switch encoded in DNA and RNA, that no one has thought of before," Zhang said.

The other involves the toxin itself. Because these tiny proteins disrupt bacterial membranes and kill cells, researchers are investigating whether they could be adapted into new antimicrobial treatments against harmful bacteria. Both lines of research remain preliminary; neither has produced a therapy.

The study also raises unresolved evolutionary questions. Zhang wonders whether type VI retrons are older or newer than other retrons, or a stepping stone between types. "This study highlights why it's so important for a molecular technology development lab like ours to keep looking to nature for inspiration," Shipman said.

The paper, "Reverse transcribed ssDNA derepresses translation of a retron antiviral protein," includes co-authors from the University of Copenhagen and UC San Francisco, with funding from the National Science Foundation, the Gordon and Betty Moore Foundation, the European Research Council, and others.