Expanding the enzyme universe: accessing non-natural reactions by mechanism-guided directed evolution

Abstract

High selectivity and exquisite control over the outcome of reactions entice chemists to use biocatalysts in organic synthesis. However, many useful reactions are not accessible because they are not in nature's known repertoire. In this Review, we outline an evolutionary approach to engineering enzymes to catalyze reactions not found in nature. We begin with examples of how nature has discovered new catalytic functions and how such evolutionary progression has been recapitulated in the laboratory starting from extant enzymes. We then examine non-native enzyme activities that have been exploited for chemical synthesis, with an emphasis on reactions that do not have natural counterparts. Non-natural activities can be improved by directed evolution, thus mimicking the process used by nature to create new catalysts. Finally, we describe the discovery of non-native catalytic functions that may provide future opportunities for the expansion of the enzyme universe.

Keywords: biocatalysis; enzymes; non-natural activity; promiscuity; protein engineering.

Figure 1

(A) Divergence of an ancestral enzyme having broad catalytic capabilities (denoted a, b, c, d, e) to more specialized enzymes (denoted A, B, C, D, E) that catalyze primarily one reaction; (B) Relationship between catalytic promiscuity and evolution of new function. A given protein sequence might catalyze multiple reactions. In the right circumstances, a catalyst with a low level of a promiscuous activity can be improved by mutation and natural (or artificial) selection so that it becomes specialized for a new function. For more discussion, see ref. .

Figure 2

(A) Chlorohydrolase activity of AtzA and aminohydrolase activity of TriA; (B) TriA and AtzA (98% AA identity) are believed to be related through a common ancestor similar to TriA. These catalytic functions can be interconverted with a few amino acid mutations.

Figure 3

(A) Hydrolysis reactions catalyzed by PTE and lactonase; (B) Putative evolutionary relationship between lactonase and PTE and their interconversion in the laboratory.^,

Figure 4

Members of MBL superfamily are functionally highly interconnected, as illustrated by Tokuriki and co-workers. The different reactions catalyzed by members of the superfamily are connected to one another via promiscuous enzymes (gray circles) that catalyze two or more reactions. Figure is reproduced from reference .

Figure 5

(A) Desaturation and hydroxylation reactions of oleic acid catalyzed by FAD2 (a desaturase) and LFAH12 (a hydroxylase); (B) Four mutations significantly increase hydroxylase activity of A. thaliana FAD2 desaturase, and a single mutation significantly increases desaturase activity of L. fendleri oleate hydroxylas\\e LFAH12.^,

Figure 6

(A) Promiscuous epoxide ring-opening activity of wild-type HHDH; (B) Application of HHDH in the synthesis of the atorvastatin side-chain and improvement of volumetric productivity using directed evolution.

Figure 7

(A) Mechanism of wild-type retaining β-glycosidase featuring catalytic acid/base and catalytic nucleophile residues where R′ = H or other sugar in low water concentration; (B) Engineered thioglycoligase via removal of catalytic acid/base residue, DNP = dinitrophenyl, Nuc = deoxythio sugar nucleophile as acceptor.

Figure 8

Use of engineered thioglycoligase for the synthesis of thiodisaccharides. Reported yields were after peracetylation of the thiodisaccharides. Wild-type enzymes do not catalyze this reaction.

Figure 9

(A) Precedents of carbene and nitrene reactivity with iron porphyrins.^, (B) Comparison of monooxygenation activity and non-natural carbene/nitrene transfer reactivity of P450-BM3. Top box, reaction of Fe(III) of P450-BM3 with O₂ and NAD(P)H generates compound I, the active species in monooxygenation. Bottom box, reaction of reduced Fe(II) with an activated species, followed by extrusion of N₂ generates a putative carbenoid/nitrenoid species.

Figure 10

P450-catalyzed non-natural carbene and nitrene transfer reactions: (A) styrene cyclopropanation;^, (B) cyclopropanation of N,N-diethyl-2-phenylacrylamide en route to levomilnacipran; (C) N-H insertion reaction; (D) intramolecular C-H amination;^, (E) intermolecular sulfimidation; (F) regioselective C-H amination by different P450 variants.

Figure 11

Cyclization of squalene catalyzed by SHC and promiscuous activity of SHC on homofarnesol.

Figure 12

(A) SHC-catalyzed cyclization of terpene-like substrates, red color indicates bond(s) formed during the reaction;^– (B) SHC-catalyzed Prins cyclization of citronellal for production of isopulegol, a precursor to menthol, and activity improvement via mutation.^,

Figure 13

Cyclization reactions of various substrates utilizing functional group initiators such as epoxide and aldehyde with AacSHC and identification of enzyme variants with improved cyclization activities.

Figure 14

(A) Mechanisms of Fe/αKG hydroxylase (path a, in red) and Fe/αKG halogenase (path b, in blue) where a common reactive intermediate, A, is diverted into two reaction pathways depending on the ligand environment around the Fe center; (B) Divergent outcome of SyrB2-catalyzed reactions of threonine-SyrB1 and norvaline-SyrB1, indicating the complex interplay between the protein fold and substrate positioning in determining the outcome of SyrB2-catalyzed reaction.

Figure 15

Use of O-acetylserine sulfhydrylase in the synthesis of unnatural β-substituted alanine derivatives and fermentation yields with E. coli strain W3110/pACcysE^fbr.

Figure 16

Mechanism of tryptophan synthase and some representative UAAs prepared with wild-type tryptophan synthase.^,