Drt3b and DRT7 write DNA

DNA polymerases are supposed to copy a nucleic-acid template. That is the rule most of molecular biology quietly leuses everywhere — replication, repair, reverse transcription, all of it. But two bacterial defense systems now look like they can do something stranger. They can make sequence-specific DNA by reading features of a protein instead of reading an existing DNA or RNA strand. ### What actually got shown? The clearest published result is DRT3, a bacterial anti-phage system from work at Stanford that appeared in *Science* in April 2026. DRT3 has two reverse-transcriptase-like proteins, Drt3a and Drt3b, plus a noncoding RNA. Drt3a does something recognizable — it reads an ACACAC sequence in the RNA and makes the matching poly(GT) strand. Drt3b then makes the complementary poly(AC) strand, but the striking part is that it does so without any nucleic-acid template sitting in the active site. (science.org) ### So is this really “template-free”? Not exactly. The better phrase is protein-templated. Drt3b is not spraying out random DNA. It is making a very specific alternating sequence, and the evidence points to amino-acid residues near the active site acting like gates that enforce the A-C-A-C pattern. So the old rule is not “dead.” The rule is narrower than many people thought: sequence information may come from a protein architecture, not only from base-pairing to DNA or RNA. (science.org) ### Why did people connect this to DRT7? Because DRT7 seems to pull a related trick. A bioRxiv preprint posted in April 2026 describes DRT7 — another bacterial anti-phage reverse-transcriptase system — making protein-primed, sequence-specific poly(T) DNA without a complementary nucleic-acid template. In that model, the RT-like domain first writes poly(T), then a primase-polymerase domain extends poly(A), producing long palindromic A/T-rich DNA structures. (biorxiv.org) That paper is still a preprint, but it points in the same direction as DRT3: bacteria seem to have evolved multiple ways to encode DNA-synthesis rules in protein structure. ### Why were bacteria doing this in the first place? This is phage warfare, basically. DRT systems are part of bacterial antiviral defense. When a phage infects the cell, these enzymes switch on and generate unusual DNA products that help stop the infection, sometimes by triggering abortive infection — the cell sacrifices itself so the virus cannot spread. That matters because it means the weird chemistry is not a laboratory curiosity. (biorxiv.org) It is tied to a real survival function that evolution kept around. ### Does this break the central dogma? No — but it does dent a textbook simplification. The central dogma is about the flow of sequence information between DNA, RNA, and protein. These findings do not show proteins being copied back into genes in the usual hereditary sense. What they show is narrower and still important: enzymes can use protein features to control the order of nucleotide addition during DNA synthesis. That expands the known chemistry of information transfer without overturning the whole framework. (science.org) ### What was the decisive evidence? Structure did a lot of the heavy lifting. In the DRT3 paper, cryo-EM at 2.6 Å resolved a 6:6:6 complex of Drt3a, Drt3b, and the ncRNA. The key claim was not just that unusual DNA appeared in a tube. It was that the enzyme architecture and active-site arrangement matched a mechanism where Drt3b could enforce base alternation without a nucleic-acid guide sitting there. That is why this landed as a serious mechanistic claim, not just an odd biochemical artifact. (science.org) ### What does this open up? The obvious angle is synthetic biology. If proteins can be engineered to “write” defined DNA patterns without a standard template, that could become a new design space for molecular recording, programmable polymers, or antiviral tools. But the catch is that these systems are still specialized, repetitive, and tied to weird bacterial defense contexts. Nobody has shown a general-purpose protein that can write arbitrary DNA sequences this way. (science.org) ### Bottom line? The big shift is not that biology abandoned templates. It is that templates may come in more than one physical form. Drt3b already has a peer-reviewed case behind it, and DRT7 looks like a second example waiting for broader validation. If both hold up, “DNA writes DNA” stops being the whole story. Proteins can sometimes write it too. (science.org)