On April 16, 2026, *Science* published a paper that will make many biology students re-examine their textbook diagrams. A Stanford team led by Alex Gao discovered that DRT3, a bacterial defense system, can directly "draw" specific DNA sequences using its own protein structure—without any nucleic acid template.
The Central Dogma, Challenged
The central dogma we all memorized—DNA → RNA → protein, later amended with reverse transcription (RNA → DNA)—holds that information flow is always "nucleic acid-templated." Every known sequence-specific polymerase (DNA polymerase, RNA polymerase, reverse transcriptase) requires a nucleic acid template. DRT3 breaks this rule: its reverse transcriptase Drt3b uses its own amino acid side chains as a "mold" to precisely synthesize a fixed repeated DNA sequence.
How DRT3 Works
DRT3 is a three-part "molecular factory":
- Drt3a – a conventional reverse transcriptase that still works from a nucleic acid template, the "traditional craftsman"
- Drt3b – the "new artist" that completely abandons nucleic acids, using the protein's own geometry as a template
- A non-coding RNA – the third component of the complex
Why It Matters
In 1970, David Baltimore and others discovered reverse transcriptase—work that earned a Nobel Prize and established that information can flow from RNA back to DNA. More than half a century later, this discovery extends the concept of "template" from nucleic acids to proteins: for the first time, genetic information has a way of being written that does not depend on a nucleic acid template. It answers a long-standing question: are there exceptions in the living world? Yes.
Three Caveats
1. Simple repeats only. The product is just poly(GT/AC)—not DNA encoding arbitrary information. "Proteins writing genomes" remains far off. 2. Bacteria only, so far. It has only been observed in a bacterial defense system; whether it exists or could be harnessed in eukaryotic cells is unknown. 3. Mechanism unresolved. Exactly how Drt3b "reads" the spatial arrangement of amino acid side chains to determine sequence is still under investigation. The paper opens the question but doesn't finish it.
What Changes
This opens a new dimension in the "rules of biological encoding." In the short term, controllable repeat sequences like poly(GT/AC) may prove useful in nanomaterials and molecular data storage. In the long term, it suggests the central dogma may be just the most common implementation used by Earth's life—not the only possibility. For synthetic biology, "using protein shape as a template to build DNA" is a design route that previously did not exist.
Bottom line: This is not a case of the textbook being overturned—rather, that classic textbook diagram can finally gain one more dashed arrow.