Protein reported to synthesize DNA from scratch

Stanford researchers reported in a Science paper published on April 16 that a bacterial antiviral system can make one strand of DNA without copying a DNA or RNA template. The study described a defense complex called DRT3, found in bacteria, that produces a repeating double-stranded DNA product while blocking infection by phages, or bacterial viruses. The work was led by Pujuan Deng, Hyunbin Lee, Carlo Armijo, Haoqing Wang and Alex Gao at Stanford University. Social-media posts on May 13 recirculated images and commentary from the paper, presenting the finding as a challenge to a long-standing rule of molecular biology. ### Which protein is at the center of the claim? The Science paper identified DRT3 as a three-part bacterial defense system made of two reverse transcriptases, Drt3a and Drt3b, plus a noncoding RNA. The authors reported that the complex synthesizes alternating poly(GT/AC) double-stranded DNA. Drt3a followed a more familiar route. The paper said Drt3a uses a conserved ACACAC sequence in the noncoding RNA as a template to make the poly(GT) strand. (science.org) Drt3b was the unusual component: the authors wrote that it synthesizes the complementary poly(AC) strand without a nucleic acid template. ### Did the researchers really say DNA was made “from scratch”? The paper did not describe unrestricted, general-purpose DNA writing. (science.org) The authors reported a sequence-specific, repeating product — alternating poly(GT/AC) DNA — rather than arbitrary DNA sequences. They also described Drt3b as “protein-primed,” meaning synthesis still depends on the enzyme complex itself and does not amount to free-form DNA assembly. A separate commentary on the finding stressed the same limit. The CRISPeR FRENZY post argued that no biological “dogma” was broken because the enzyme appears specialized for a simple repetitive output rather than broad encoding of genetic information. That interpretation was the blog author’s, not the paper’s, but it matched the narrow product described in the study. (science.org) ### What evidence did the paper present for a protein template? The authors reported cryo-electron microscopy structures at 2.6 angstrom resolution showing a D3-symmetric 6:6:6 complex of Drt3a, Drt3b and the noncoding RNA. In the abstract, they said conserved active-site residues in Drt3b enforce the alternating base pattern needed to make the poly(AC) strand. Phys.org’s summary of the study said Drt3b uses its own protein structure as a mold and that the Stanford team tested the system by expressing DRT3 in E. coli and exposing the bacteria to phages. (mycrispr.blog) The report said the system was activated by detection of a specific viral protein, ST61, after which sequencing showed long, repetitive DNA strands with the expected alternating pattern. (science.org) ### Why are people calling it a challenge to a “4 billion year” rule? The paper itself said nucleic acid polymerases are generally grouped into template-directed enzymes, which copy existing nucleic acids, and template-independent enzymes, which usually make simpler products such as homopolymers or short motifs. The authors presented DRT3 as an expansion of that landscape because Drt3b appears to generate a defined complementary DNA sequence using protein features rather than a nucleic-acid guide. (phys.org) ScienceAlert, summarizing the study after publication, said the finding pointed to “an entirely new way that DNA can be synthesized.” That phrasing reflected how outside writers framed the result, while the paper itself described a “protein-templated mechanism for sequence-specific DNA synthesis.” ### How broad is the discovery so far? (science.org) The authors reported that DRT3 homologs are distributed across at least 20 bacterial phyla, suggesting the system is not a one-off oddity. At the same time, the published work focused on one bacterial defense pathway and one highly repetitive DNA product, not on ordinary genome replication in cells. Alex Gao’s lab lists the paper as a 2026 publication and notes that it was highlighted in Science News. (sciencealert.com) The next concrete step for readers is the primary paper itself — “Protein-templated synthesis of dinucleotide repeat DNA by an antiphage reverse transcriptase” — published in Science on April 16, 2026, with Gao listed as corresponding author. (gaolab.bio) (science.org)