Bacteria enzyme synthesizes DNA

DNA polymerases are supposed to copy a nucleic-acid template. That is the basic rule — DNA from DNA, or DNA from RNA in the case of reverse transcriptases. But a Stanford team just described a bacterial defense system that partly breaks that expectation. In a paper published in *Science* in April 2026, they showed that one enzyme in the system writes a specific DNA strand without reading any DNA or RNA template at all. ### What actually got discovered? The system is called DRT3, a bacterial anti-phage defense module made of two reverse transcriptases, Drt3a and Drt3b, plus a noncoding RNA. Together they produce double-stranded DNA with an alternating repeat pattern — one strand is poly(GT), the other is poly(AC). Drt3a does the familiar thing and copies an ACACAC sequence embedded in the RNA. Drt3b is the weird one. It makes the complementary strand without a nucleic-acid guide. (science.org) ### Why is that such a big deal? Because “template-free” usually means sloppy. Biology already knows enzymes that add random tails, homopolymers, or very short motifs without a template. What it generally does not do is make a defined sequence by choosing one base, then another, in a controlled repeating pattern. The Science paper argues that Drt3b is doing exactly that — sequence-specific synthesis, but with the instructions coming from the protein itself. (science.org) ### So what is the protein using as a template? Basically, its own active site. The authors say conserved amino-acid residues inside Drt3b enforce the alternating base pattern during synthesis. In other words, the enzyme is not reading letters from DNA or RNA. It is using the shape and chemistry of its own protein pocket to favor the right nucleotide at the right step. That is why people are calling this protein-templated DNA synthesis, not just template-independent synthesis. (science.org) ### How do they know there was no hidden RNA or DNA there? The load-bearing evidence is structural. Cryo-EM at 2.6 angstrom resolution showed a 6:6:6 complex of Drt3a, Drt3b, and the noncoding RNA. In that structure, Drt3a sits with its RNA template in the expected role, but Drt3b does not have a nucleic-acid template positioned to instruct synthesis. That separation matters — it makes the claim much stronger than a weird biochemical assay alone would. (science.org) ### What is this doing for the bacteria? It looks like an antiviral weapon. DRT systems are part of bacterial anti-phage immunity, and DRT3 helped engineered *E. coli* resist phage infection in the experiments described in coverage of the paper. The idea is that bacteria are not making this odd DNA for housekeeping. They are making it as part of a defense response after detecting a viral trigger. (science.org) ### Does this overturn the central dogma? Not really. That headline is catchy, but the catch is that the central dogma is about information flow — DNA to RNA to protein — not a ban on exotic chemistry. What this result really challenges is a narrower textbook assumption: that sequence-specific DNA synthesis always needs a nucleic-acid template. That is still a big deal, just a more precise one. (phys.org) ### Could this matter outside bacterial immunity? Maybe, but slowly. The immediate importance is conceptual — polymerases can apparently do more than the standard two buckets of template-directed and loose template-free synthesis. Longer term, people will wonder whether this can inspire new enzymatic DNA-writing tools. But DRT3 is making a very repetitive product, not arbitrary custom genes, so this is not instant desktop DNA printing. (science.org) ### What is the bottom line? A bacterial enzyme did not start making unlimited DNA from nothing. It did something subtler and, in a way, stranger: it used protein structure itself as the instruction set for a defined DNA pattern. That widens the map of what biology can do — and it suggests there may be more “impossible” polymerases hiding in bacterial defense systems. (science.org)