Novel adenine base editor shows superior precision

Researchers from YolTech Therapeutics and academic institutions in Shanghai used a novel evolutionary approach to create a more precise gene editing tool for converting A•T to G•C base pairs, combining structural clustering with sequential amino acid saturation mutagenesis.

They first identified candidate proteins through AI-based structural prediction and identified a naturally active adenine deaminase from Hafnia paralvei (hpTadA). This enzyme was then evolved using multiple rounds of directed evolution, with saturation mutagenesis libraries targeting specific segments of consecutive amino acids.

The resulting hpABE5.20 editor demonstrated efficiency comparable to the current gold standard ABE8e but with a more refined editing window. Testing across 16 genomic sites showed that hpABE5.20 achieved editing efficiencies close to or higher than ABE8e at 15 out of 16 sites whilst significantly reducing bystander editing.

Notably, the new editor showed approximately four-fold lower guide RNA-dependent DNA off-target activity and significantly reduced RNA off-target editing compared to ABE8e. When used to correct disease-causing mutations in cellular models of phenylketonuria, hpABE5.20 maintained high on-target efficiency while delivering up to 2.59 times greater editing precision.

The team further demonstrated the editor's therapeutic potential in human hematopoietic stem cells, where it efficiently induced gamma-globin expression at levels superior to current clinical protocols for treating β-hemoglobinopathies.

In non-human primates, a single intravenous dose of hpABE5.20 delivered via lipid nanoparticles achieved sustained PCSK9 editing, with protein levels reduced by approximately 80% and stable therapeutic outcomes maintained for at least 15 months.

The research was led by Drs Zijun Wang, Yuxuan Wu, and Yuming Lu from YolTech Therapeutics. The findings were published as a preprint on bioRxiv on 16 May 2025.