Recovery of Information Stored in Modified DNA with an Evolved Polymerase

DNA
is increasingly being explored as an alternative medium for
digital information storage, but the potential information loss from
degradation and associated issues with error during reading challenge
its wide-scale implementation. To address this, we propose an atomic-scale
encoding standard for DNA, where information is encoded in degradation-resistant
analogues of natural nucleic acids (xNAs). To better enable this approach,
we used directed evolution to create a polymerase capable of transforming
2′-O-methyl templates into double-stranded
DNA. Starting from a thermophilic, error-correcting reverse transcriptase,
RTX, we evolved an enzyme (RTX-Ome v6) that relies on a fully functional
proofreading domain to correct mismatches on DNA, RNA, and 2′-O-methyl templates. In addition, we implemented a downstream
analysis strategy that accommodates deletions that arise during phosphoramidite
synthesis, the most common type of synthesis error. By coupling and
integrating new chemistries, enzymes, and algorithms, we further enable
the large-scale use of nucleic acids for information storage.