Engineering life into new chemical transformations

Nature has had many generations to optimise its catalytic processes. As a result they are both extremely efficient and exquisitely selective. Biocatalysis is now very much a mainstay of industrial organic synthesis, particularly with the advent of protein engineering and directed evolution enabling the preparation of robust enzymes fine-tuned to a particular substrate. The importance of this technology was recognised in 2018 with the award of the chemistry Nobel prize to Frances Arnold for her pioneering work on directed evolution of enzymes. Despite significant advances, biocatalysis is limited to natural chemical reactions, precluding its wide application in general synthetic chemistry, where small molecule catalysts still have the upper hand. That said the repertoire of “unnatural” enzyme mediated transformations is increasing, particularly with engineered metalloenzyme systems. This article describes how development of new to Nature chemistry is evolving and what lies on the horizon.*

As the complexity of drug molecules increases, so the synthetic challenges become more pointed.In addition, increasing demand for green, atom-economic and sustainable methods of manufacture with low environmental impact introduce another level of complexity.1 This coupled with shortened development timelines and demand for lower cost medicines has prompted a reckoning with Nature, and as such the use of enzymes and biocatalytic processes for mainstream synthesis of complex organic substrates, particularly chiral molecules, has increased significantly over recent years. The introduction of protein engineering and- in a Darwinian parody- genetic mutation and the evolutionary selection of a particular favourable reaction trait (robustness, selectivity, solvent stability etc.) the fine tuning and optimization of an enzyme for a particular chemical transformation is now readily achievable. The historical approach of finding a wildtype enzyme that’s available to get you as close as possible to the required conversion and selectivity (but never quite close enough) is now a thing of the past. Notable examples of the success of such an approach is the directed evolution of a transaminase enzyme developed by Merck and Codexis for the manufacture of the diabetes drug Sitagliptin and the Baeyer−Villiger monooxygenase-mediated synthesis of the proton pump inhibitor Esomeprazoleby Codexis.2 The former ultimately replaced the rhodium based asymmetric hydrogenation of an enamine as the route of manufacture- the ketone being directly converted to the required amine enantiomer using the engineered transaminase (Scheme 1).3 Concerns over the fluctuating cost of rhodium and the compelling atom efficiency of the enzyme process meant the enzyme process ultimately won out.



Scheme 1. Chemocatalytic and biocatalytic approaches to Sitagliptin.

With the 2018 chemistry Nobel prize jointly awarded for biocatalytic engineering, and the recent recognition by the International Union of Pure and Applied Chemistry (IUPAC) that directed evolution of selected enzymes is one of the top emerging technologies that will change our world, the mandate for this approach is sealed.4 The future paradigm for biocatalysis moves beyond taming the innate activity of an enzyme for a particular transformation and expanding substrate scope. Engineering enzymes for new chemical reactions outside the scope of Nature’s toolbox- the so called ‘new to nature’ approach- offers an inherently greener and more sustainable alternative to existing catalytic processes that require harsh reaction conditions, toxic organometallic complexes and suffer from poor atom economy.5

Engineering an enzyme for non-natural chemistry relies partly on evolving natural substrate promiscuity. In addition genomic sequencing, chemomimetic biocatalysis, artificial metalloenzyme engineering and the introduction of unnatural amino acids into the protein have been employed, with or without the application of ‘unnatural’ co-factors. Combinations of these techniques with a smattering of in silico design and heavy dose of directed evolution have proved fruitful.

Various organocatalytic processes have been explored, including Aldol reactions, Michael-type additions and Knoevenagel condensations. Very recently a stereospecific biocatalytic Friedel-Crafts alkylation reaction was described by Balskus, demonstrating that mild and selective C-C bond forming reactions are possible.6

Arguably the most exciting area of research in this field is the development of non-natural chemistry catalysed by modified P450 enzymes. These proteins have evolved naturally with inherent promiscuity (and generally low selectivity) to oxidize and detoxify a broad range of organic molecules in a cellular environment. Engineering proteins in the laboratory has enabled improved enantio- and regio-selectivity for synthetically useful transformations, but more significantly, activity in non-hydroxylation chemistry can be switched on by directed evolution. Following the Seminal work by Frances Arnold, in which she showed that a heme protein could be engineered to generate and selectively transfer a carbene intermediate to an alkene and form a cyclopropane, a range of carbene and nitrene transfer chemistry has been reported, including C-H insertion of sulfonyl nitrenes to form C-N bonds.7 The use of Earth-abundant iron underpins the inherent sustainability of the processes. Later, Arnold also described the first enzyme-catalysed carbene insertion into a carbon-silicon bond using an engineered cytochrome c.8

Others have continued in a similar vein- in particular Hartwig has described substitution of the P450 heme iron with precious metals such as iridium in combination with protein mutations and co-factor modifications to further expand and open up new areas of reactivity. In particular carbene insertion into C-H, N-H and S-H bonds has been demonstrated.9 The P450 enzymes used for this chemistry are thermophilic and reconstituted with an Ir(Me)-PIX cofactor. They are currently the most active artificial metalloenzymes yet identified. Heme-centred enzymes evolved due to the abundance of iron in the biosphere. “Evolving” unnatural metalloenzymes with other metals has the potential to exceed the performance of the native enzymes as well as opening up new areas of reactivity.

A good example of the power of directed evolution in the context of P450 enzymes is the recent paper in the journal Scienceby Cho et al describing remote, tuneable C-H functionalisation to generate pharmaceutically relevant cyclic amides. In this approach acyl-protected hydroxylamines were used as carbonyl nitrene precursors and evolved P450 mutants, expressed in escherichia coli cells, directed enantioselective and regioselective C-H amidation via intramolecular cyclisation to the corresponding lactams.10 Mechanistically the heme iron forms an intermediate iron niterenoid species and the protein environment (P450 mutant proteins) dictates which of the prochiral C-H bonds undergoes insertion, generating the corresponding chiral lactam. From an initial screen of a panel of P450 enzymes and variants (dubbed P411 variants that had historically been shown to be effective in generating sulfonylnitrene intermediates) a weakly active, non-selective enzyme was identified and subjected to four rounds of directed evolution (mutating amino acid residues close to the heme iron in the active site). A variant was identified that gave almost complete selectivity for the beta-lactam (regioisomeric ratio up to 25:1) in high yield (>95%) and high enantiomeric excess (>90%) with exceptionally high efficiency (up to 1×10total turnovers). Further rounds of directed evolution yielded variants that gave selectively the beta, gamma or delta-lactams. The power of this technology was further demonstrated by diversifying a single substrate with regiodivergent C-H bonds (Scheme 2). Amidation of aliphatic, homobenzylic, and benzylic C(sp3)‒H bonds using these different mutant P450’s gave the corresponding beta, gamma or delta-lactams selectively in high (94%) enantiomeric excess. Cyclic amides and amines have been utilized extensively in the pharmaceutical industry and this approach will undoubtedly be utilized in drug discovery.

Scheme 2. ‘Lactam synthase’s’-mutated P450 enzymes- promote regioselective, enantioselective C-H insertion.

What about modalities not frequently encountered in Nature such as fluorination? Enzymatic fluorination has not evolved in the biosphere due in part to the limiting inherent physical properties of the fluoride ion (and to a lesser extent its relatively low natural abundance and hence bioavailability). A high heat of hydration renders desolvation of fluoride ion difficult, and its high electronegativity disfavours oxidation- factors commonly encountered in biochemical processing. That said fluorinase enzymes have been identified in bacteria and marine microorganisms and a small number of fluorine-containing metabolites have been characterised.11 Attempts to expand the substrate scope of natural fluorinase enzymes via protein engineering has resulted in limited success. However evolution of other enzymes that operate mechanistically on a similar level and exhibit promiscuous activities for abiological reactions have been utilised to introduce fluorinated functional groups. Insertion of fluoroalkyl-substituted carbenes (derived from the corresponding diazo compounds) into a-amino C(sp3)‒H bonds has very recently been realised using genetically encoded P450 enzymes that can be mutated to give the required absolute configuration in the products. As with the lactam example above, P411 variants were evolved to enable the insertion chemistry, which at present is unique to biocatalysis (Scheme 3). Remarkably there is no equivalent small molecule catalyst able to replicate this chemistry.12

Scheme 3. α‐Amino C−H fluoroalkylation catalysed by engineered cytochrome P450s.

Another example, again using carbenes derived from fluoroalkylated diazo precursors in combination with “carbene transferase” enzymes, is the synthesis of chiral a-trifluoromethylated organoboron intermediates. Insertion into the B-H bond of an N-heterocyclic carbene borane substrate using an enzyme derived from Rhodothermus marinuscytochrome c(Rma cytc) gave selectively the chiral trifluoromethylated boranes in high enantioselectivity (Scheme 4). With the aid of computer modelling, selective mutation enabled “evolution” of an active site that could accommodate a range of structurally diverse trifluorodiazo alkanes enabling access to these versatile boron-derived synthetic building blocks.13

Scheme 4. Biocatalytic synthesis of alpha-trifluoromethylated boranes.

One of the most significant technological advancements in synthetic chemistry in recent years is the use of photoredox catalysis to generate synthetically useful radical intermediates and promote novel reactivity that is difficult to achieve in other ways, giving access to new structural motifs relevant to drug discovery. Photoredox catalysis relies on electronic photoexcitation of transition metal complexes or organic dyes with visible light to enable single electron transfer (SET) and generation of synthetically useful organic radical intermediates. Historically, photoredox in biocatalysis has centred on controlling the redox state of co-factors. Todd Hyster and his team at Princeton have recently shown that photoredox catalysts can be used to instil unnatural catalytic activity in nicotinamide-dependent reductase enzymes.15 Under visible light irradiation a Xanthene-based photocatalyst (Rose Bengal) enabled a double bond-reductase to catalyse an enantioselective deacetoxylation (Scheme 5). Hyster suggests that protein binding attenuates the reduction potential of the substrate, localising the a-acyl radical to the enzyme active site where stereochemistry is controlled. This coupling of photoredox catalysis- to access new reaction space via generation of radical intermedites – and enzyme synergy- to increase selectivity- is an exciting future direction for “unnatural” biocatalysts.

Scheme 5. Photoredox and biocatalysis in synergistic asymmetric deacetoxylation – NtDBR: Nicotiana tabaccum(double bond reductase enzyme) RB: Rose Bengal (photocatalyst).

Biocatalysis has moved on significantly in the past few decades from simple hydrolytic process using wildtype enzymes to engineered systems designed for specific industrial manufacture. With ever increasing demands for sustainable green chemistry, biocatalysis offers rich potential.15 Efforts to operate outside of Natures biochemistry has driven some exciting developments as we continue to engineer life into new chemical transformations.


Written by John Studley, August 2019

* This article was previously published in the Monographic special issue: Catalysis & Biocatalysis – Chimica Oggi – Chemistry Today – vol. 37 (4) July/August 2019


  1. Broadening the scope of biocatalysis in sustainable organic synthesis, R. Sheldon and D. Bradly ChemSusChem 2019, 12, 2859.
  2. Sitagliptin: Y. Hsiao et al, J. Am. Chem. Soc.,2004, 126, 9918; C.K. Saville et al Science, 2010, 329, 305; Esomeprazole: D. Entwistle et al, J. Org. Chem. 2018, 83, 7453.
  3. Merck and Codexis won a green chemistry award for their work on the synthesis of Sitagliptin, see C&EN July 10th2006, 24.
  4. Innovation by evolution: bringing new chemistry to life, Frances Arnold Nobel lecture 8thDecember 2018,; April 1, 2019 – The International Union of Pure and Applied Chemistry (IUPAC) top ten emerging technologies in chemistry, Fernando Gomollen-Bel:
  5. The enzymology of organic transformations: A survey of named reactions in biological systems, -w. Liuet alAngew. Chem. Int. Ed. 201756, 3446; Design and evolution of enzymes for non-natural chemistry, F. Arnold et al, Current Opinion in Green and Sustainable Chemistry 2017, 7, 23.
  6. Fast Knoevenagel condensations catalysed by an artificial Schiff-base forming enzyme, X. Garrabou et al, Am. Chem. Soc., 2016, 138, 6972; Emergence of a catalytic tetrad during evolution of a highly active artificial aldolase, D.Hilvert et al, Nat. Chem. 2017, 9, 50; Stereochemical control of enzymatic carbon-carbon bond-forming Michael-type additions by “substrate engineering”, G. Poelarends et al, Eur. J. Org. Chem., 2016, 32,5350; Biocatalytic Friedel–Crafts alkylation using a promiscuous biosynthetic enzyme, E. Balskus et al, Angew. Chem. Int. Ed. 201958, 3151.
  7. Olefin cyclopropanation viacarbene transferase catalysed by engineered cytochrome P450 enzymes, F. Arnold et al, Science 2013, 339, 307; Enantioselective, intermolecular benzylic C-H amination catalysed by an engineered iron haem enzyme, F. Arnold et al, Chem. 2017, 9, 629.
  8. Directed evolution of cytochrome c for carbon-silicon bond formation: bringing silicon to life, F. Arnold et al, Science 2016, 354, 1048.
  9. Noble−metal substitution in hemoproteins: An emerging strategy for abiological catalysis, J. Hartwig and S. Natoli, Accounts Chem. Res. 2019, 52, 326; Biocatalytic oxidation reactions: A chemist’s perspective, F. Hollmann et al, Angew. Chem. Int. Ed. 2018, 57, 9238.
  10. Site-selective enzymatic C‒H amidation for synthesis of diverse lactams, I. Cho et al, Science 2019, 364, 575.
  11. Enzymatic fluorination and biotechnological developments of the fluorinase, O’Hagan and H. Deng, Chem. Rev. 2015, 2, 634; Halogenase engineering and its utility in medicinal chemistry, D. Sherman et al, Bioorg. Med. Chem. Let. 2018, 28, 1992.
  12. Enantiodivergent α‐amino C−H fluoroalkylation catalysed by engineered cytochrome P450s, J. Zhang et al, J. Am. Chem. Soc. 2019, 141, 9798.
  13. A biocatalytic platform for synthesis of chiral α-trifluoromethylated organoborons, K. Houk et al, ACS Cent. Sci. 2019, 5, 270.
  14. Catalytic promiscuity enabled by photoredox catalysis in nicotinamide-dependent oxidoreductases, T. Hyster et al, Chem. 2018, 10, 770.
  15. Extending the application of biocatalysis to meet the challenges of drug development (review article), N. Turner et al, Chem. 2018, 2, 409; The impact of recent developments in technologies which enable the increased use of biocatalysts, A. Maguire et al, Eur. J. Org. Chem. 2019, 23, 3713.