Life Runs on Four DNA Letters. An Ordinary Bacterial Enzyme Just Read Eight Without Complaining.

Every organism that has ever been sequenced writes its genome in four letters: A, T, G and C. In 2019 chemists demonstrated an eight-letter version, adding four synthetic bases that pair with each other the way the natural ones do. They called it hachimoji DNA, from the Japanese for eight letters. The obvious question since then has been whether the machinery of a living cell can actually read it, or whether the extra letters are a chemistry demonstration that biology will refuse to touch.

Researchers at UC San Diego have now answered the reading half. In a study published September 2 in Nature Communications, a team led by Dong Wang, a professor at the Skaggs School of Pharmacy and Pharmaceutical Sciences, showed that E. coli RNA polymerase — the enzyme that transcribes DNA into RNA, and one of the most heavily studied proteins in biology — transcribes hachimoji templates accurately, handling the synthetic base pairs with the same fidelity it applies to natural ones.

The evidence is structural. Using high-resolution cryo-electron microscopy, the group captured the enzyme mid-transcription and resolved how its active site grips the unnatural pairs. The geometry is what makes it work: the polymerase is checking the shape and dimensions of the base pair sitting in its active site rather than interrogating the specific chemistry of the letters, so a synthetic pair that presents the right shape gets waved through.

That principle turns out to be even more forgiving than expected. In a companion paper published August 12 in the Proceedings of the National Academy of Sciences, the same group showed the enzyme recognizing a different synthetic base pair that lacks the hydrogen bonds normally considered essential to holding two DNA bases together. If hydrogen bonding is not strictly required for the enzyme to accept a pair, the design space for artificial genetic letters is substantially wider than the classical picture allows.

The practical target is storing and using information that natural biology cannot express. An eight-letter alphabet raises the information density of a strand of DNA, which matters for DNA data storage, and it opens the possibility of engineered organisms whose genetic code contains instructions no natural sequence can encode — including a built-in dependency that stops them surviving outside a laboratory that supplies the synthetic bases.

Reading is not the whole problem. A functioning synthetic genome needs replication, repair and translation to cope with the extra letters as well, and each of those systems has its own quality-control machinery that has never encountered anything but the standard four. But transcription was the step most people expected to be a wall, and the cryo-EM structures show the enzyme going through it without noticing that anything unusual is in front of it.