Design and evolution of an enzyme with a non-canonical organocatalytic mechanism

Headnote

The combination of computational design and laboratory evolution is a powerful and potentially versatile strategy for the development of enzymes with new functions1-4. However, the limited functionality presented by the genetic code restricts the range of catalytic mechanisms that are accessible in designed active sites. Inspired by mechanistic strategies from small-molecule organocatalysis5, here we report the generation of a hydrolytic enzyme that uses N_δ-methylhistidine as a non-canonical catalytic nucleophile. Histidine methylation is essential for catalytic function because it prevents the formation of unreactive acyl-enzyme intermediates, which has been a long-standing challenge when using canonical nucleophiles in enzyme design6-10. Enzyme performance was optimized using directed evolution protocols adapted to an expanded genetic code, affording a biocatalyst capable of accelerating ester hydrolysis with greater than 9,000-fold increased efficiency over free N_δ-methylhistidine in solution. Crystallographic snapshots along the evolutionary trajectory highlight the catalytic devices that are responsible for this increase in efficiency. Ns-methylhistidine can be considered to be a genetically encodable surrogate of the widely employed nucleophilic catalyst dimethylaminopyridine11, and its use will create opportunities to design and engineer enzymes for a wealth of valuable chemical transformations.

Designing hydrolytic enzymes with sophisticated catalytic machineries reminiscent of those found in nature presents a formidable challenge6-9,12. Natural hydrolases often use a catalytic triad containing an activated nucleophile paired with hydrogen-bond donors to stabilize oxyanion intermediates13-15. Although computational approaches have enabled the accurate design of histidine, cysteine and serine nucleophiles, the formation of unreactive acyl-enzyme intermediates has compromised catalytic function6-9. Synthetic mimics of hydrolytic enzymes and self-assembled a-helical barrels derived from synthetic peptides are compromised by similar limitations10,16.

To create a functional hydrolase we selected a computationally designed enzyme for the Morita-Baylis-Hillman reaction, denoted BH32, as a template for catalytic remodelling. BH32 uses a histidine nucleophile (His23) built into the cap domain (HAD superfamily nomenclature) of haloacid dehalogenase from Pyrococcus horikoshii by introducing 12 active-site mutations predicted by the Rosetta software suite17. The catalytic activity of BH32 for its designed Morita-BaylisHillman reaction is modest (less than 1 turnover per day). Nevertheless, we were attracted by the potential to unlock new catalytic functions within the BH32 template by harnessing the reactivity of the designed histidine nucleophile. The hydrolytic activity of BH32 towards a series of monoacylated fluorescein...