Files

Abstract

This dissertation describes the development and characterization of engineered Fe(II)- and α-ketoglutarate-dependent oxygenases (FeDOs) as site-selective biocatalysts for native and non-native C–H functionalization reactions. These enzymes provide an alternative to conventional organic synthesis by acting on unactivated C–H bonds with high selectivities under aqueous and mild conditions. Functionalization occurs via a C–H abstraction step followed by radical rebound of the new functional group. Previous attempts at non-native catalysis using FeDOs relied on the presence of an exogenous ligand that participates in the rebound step. Controlling the selectivity of rebound between native and non-native groups has been challenging, with most cases of non-native FeDO biocatalysis resulting in large amounts of the native product. The study presented herein addresses this challenge by engineering FeDOs with novel levels of selectivity towards non-native reactions.Chapter 1 provides a general background of site-selective enzymatic C–H functionalization, protein engineering, and directed evolution. Additionally, the family of FeDOs is introduced and described, with a focus on hydroxylating FeDOs and Fe(II)- α-ketoglutarate-dependent halogenases (FeDHs). Finally, several examples are presented in which these enzymes have been used as site-selective biocatalysts, with the goal of depicting the current state of the field and its challenges. Chapter 2 introduces the hydroxylase SadA and its variant SadA D157G. SadA D157G had been previously reported to carry out non-native chlorination of N-succinyl-L-leucine, albeit with poor chemoselectivity. This variant was engineered for improved expression via MBP-fusion to give SadX. Additionally, the activity of SadX and SadX variants in the presence of different exogenous anions is characterized. Finally, the reaction of SadX in the presence of NaOCN is shown to enable carbamate formation. Chapter 3 discusses the directed evolution of SadX for non-native reactivities, including chlorination and azidation. Conditions for screening SadX libraries in the presence of exogenous anions are established. This effort resulted in SadX variants with an increased chemoselectivity towards azidation of N-succinylated compounds. The resulting evolution campaign also offered insight into some of the underlying features of the SadX scaffold that control activity, chemoselectivity, and site selectivity. Chapter 4 focuses on exploring the native hydroxylase activity of SadA. An improved hydroxylase with higher activity on the model substrate N-succinyl-L-leucine is described. Additionally, the effect of mutations accumulated during the directed evolution described in Chapter 3 is evaluated when these mutations are introduced in the wild-type SadA scaffold. A series of engineered hydroxylases is obtained, with notable levels of activity and selectivity on a variety of N-succinylated compounds. Chapter 5 describes a novel methodology to incorporate next-generation sequencing and machine learning into directed evolution, with a focus on cost-efficient techniques with low experimental burden. This workflow is validated by applying it to the directed evolution libraries discussed in Chapter 3. A small set of predicted sequences is experimentally evaluated, resulting in a variant with improved azidation activity.

Details

Actions

PDF

from
to
Export
Download Full History