Bacterial enzyme makes DNA without template

DNA polymerases are supposed to copy a template. That is the rule you learn early, and it is one of the load-bearing ideas in molecular biology. But a bacterial virus-defense system just bent that rule in a very strange way. In a paper published in *Science* on April 16, 2026, a Stanford team showed that one enzyme in a bacterial complex can build a defined DNA strand without reading DNA or RNA at all. ### What is the thing they found? The system is called DRT3. It shows up in bacteria as part of their anti-phage toolkit — basically, a way to fight viruses that infect bacteria. DRT3 is not one enzyme doing one job. It is a three-part machine: two reverse transcriptases, named Drt3a and Drt3b, plus a noncoding RNA. Together they make a double-stranded DNA product with a repeating pattern: alternating GT on one strand and AC on the other. (science.org) ### Why is that weird? Because sequence-specific DNA synthesis normally needs a nucleic-acid guide. A polymerase reads DNA or RNA and adds the matching bases. There are enzymes that add bases without a template, but those products are usually messy, random-ish tails, homopolymers, or very short simple motifs. The weird part here is not merely “no template.” It is “no nucleic-acid template, yet still a defined repeating sequence.” (science.org) ### So what does each enzyme do? Drt3a is the familiar half of the story. It uses a conserved ACACAC sequence embedded in the noncoding RNA as a template and writes the complementary poly(GT) strand. That is odd but still legible in standard biology terms — RNA guides DNA synthesis, like reverse transcriptase is supposed to do. Drt3b is the shocker. It builds the complementary poly(AC) strand in the complete absence of a DNA or RNA template. (science.org) ### Wait — how can Drt3b know what to add? The team’s answer is: the protein itself acts like the template. Cryo-EM structures at 2.6 Å resolution showed a 6:6:6 complex of Drt3a, Drt3b, and the RNA, and the structural work pointed to conserved amino-acid residues in Drt3b’s active site that enforce the alternating pattern. Basically, instead of base-pairing against a nucleic-acid strand, the enzyme’s own shape and chemistry steer which nucleotide comes next. (science.org) That is why the paper calls it protein-templated DNA synthesis. ### Does this actually help bacteria? Yes — that is not just a test-tube curiosity. When the researchers put DRT3 into *E. coli* and challenged the cells with phages, the system protected the bacteria. The defense appears to switch on in response to a specific phage protein called ST61, and then the machinery produces these long repeating DNA molecules. So the strange chemistry seems tied to a real antiviral function, not an accidental side reaction. (science.org) ### Is this the first weird DRT story? No, and that is part of why this area is suddenly hot. Another DRT system, DRT2, was shown in separate work to make entirely new defensive genes by rolling-circle reverse transcription from a noncoding RNA, then turn those products into toxic “never-ending” proteins that halt growth during phage attack. DRT3 is different, but it fits the same bigger pattern: bacterial reverse transcriptases are doing much weirder information processing than anyone expected a few years ago. (phys.org) ### Does this rewrite the central dogma? Not really in the dramatic social-media sense. DNA is still DNA, RNA is still RNA, and proteins are not being reverse-copied into genes the way a ribosome reads mRNA. The narrower but still big change is this: sequence-defined DNA synthesis does not always require a nucleic-acid template. Protein structure can, at least in this special system, carry enough information to guide the sequence outcome. That expands what biologists think polymerases can do. (biorxiv.org) ### Why do people care beyond bacterial immunity? Because if this mechanism proves generalizable, it could matter for both basic science and tools. On the basic side, it opens a new category of biochemical information transfer — protein-to-DNA patterning. On the applied side, any enzyme that can write defined DNA without a conventional template gets synthetic biologists interested fast. The catch is that this is a very specialized bacterial defense machine making a simple repeating product, not a universal programmable DNA printer. (science.org) ### Bottom line? The real news is not that biology’s rules collapsed. It is that one rule turned out to have a hidden exception. A bacterial anti-virus enzyme can write a defined DNA strand by using its own protein architecture as the guide — and that is a much bigger deal than it sounds at first glance. (science.org)