Abstract
Expanding the genetic code of living cells with noncanonical monomers (ncMs) relies on engineered aminoacyl-tRNA synthetases (aaRS) and their cognate tRNAs. Conventional aaRS engineering strategies rely on translation-dependent selection systems, limiting their utility for ncMs that are poorly accommodated by the native translational machinery. To address this limitation, we recently developed START, a translation-independent platform that selects Methanomethylophilus alvus pyrrolysyl-synthetase (MaPylRS) mutants based on their ability to acylate cognate tRNA(MaPyl). START uses barcoded tRNAs to encode the identity of distinct aaRS mutants in a library. Acylation by active aaRS mutants protects the corresponding tRNAs from periodate oxidation, and their identity is retrieved subsequently through sequencing. START was previously applied to genetically encode noncanonical α-amino acids. Here, we successfully applied START to engineer MaPylRS mutants capable of acylating tRNA(MaPyl) with diverse non-α-amino acid substrates with good efficiency and fidelity, including (R) and (S) enantiomers of a β(2)-hydroxy acid, a β(2)-amino acids, and a malonate. Several mutants exhibit notable polyspecificity across noncanonical backbones while maintaining selectivity against their α-amino acid counterparts. Using these novel enzymes, we demonstrate the ribosomal incorporation of both (R)- and (S)-β(2)-hydroxy acids into a luciferase reporter protein expressed in Escherichia coli with good efficiency and fidelity. These results imply that highly active engineered aaRS/tRNA pairs can overcome the recently established limitations of EF-Tu with respect to non-α-amino acid substrates. The engineered MaPylRS mutants also enabled the successful incorporation of both (R)- and (S)-β(2)-hydroxy acids into a protein expressed in mammalian cells, demonstrating for the first time that eukaryotic translation can accommodate non-α-backbones.