Regioselective Control of SAM Halogenation through Structure-Guided Directed Evolution of a Hydroxide Adenosyltransferase

Enzymatic halogenation reactions are gaining attention because of their mild reaction conditions. <i>S</i>-Adenosylmethionine (SAM), a critical cofactor in biological systems, is utilized by SAM-dependent halogenases to produce 5'-halogenated adenosines for antibiotic biosynthesis.^1-3 Both SAM-dependent halogenases and hydroxide adenosyltransferase (HATase) belong to the domain of unknown function 62 (DUF-62) family, but HATase uses water as a substrate to produce adenosine.^3-5 Recently, we transformed a HATase into a halogenase by protein engineering. However, whether this protein could be further tuned for regioselective control of SAM halogenation remains unclear. To our knowledge, regioselective control of SAM halogenation has never been achieved with any SAM-dependent halogenases so far. Here, we identified key residues influencing regioselective cleavage of SAM with a HATase from <i>Thermotoga maritima</i> MSB8. Multiple rounds of directed evolution, aided by cell surface display in <i>E. coli</i>, generated HATase mutants that convert SAM to CH₃I and <i>S</i>-adenosylhomocysteine (SAH). Computational studies revealed that mutations altered the halide orientation, leading to changes in regioselectivity for SAM cleavage. Notably, these identified mutation sites conferred regioselective halogenation capability to other HATases. Our study unveils the potential of the DUF-62 superfamily for regioselective cleavage of SAM and expands the halide methyltransferase repertoire through engineered HATases.