Chem Soc Rev

View Article Online

TUTORIAL REVIEW View Journal

Biocatalysis engineering: the big picture

Published on 13 March 2017. Downloaded by University of Newcastle on 13/03/2017 16:02:53.

Cite this: DOI: 10.1039/c6cs00854b

Roger A. Sheldon*ab and Pedro C. Pereirab

Received 28th November 2016
DOI: 10.1039/c6cs00854b
rsc.li/chem-soc-rev
In this tutorial review we describe a holistic approach to the invention, development and optimisation of
biotransformations utilising isolated enzymes. Increasing attention to applied biocatalysis is motivated by
its numerous economic and environmental benefits. Biocatalysis engineering concerns the development
of enzymatic systems as a whole, which entails engineering its dierent components: substrate
engineering, medium engineering, protein (enzyme) engineering, biocatalyst (formulation) engineering,
biocatalytic cascade engineering and reactor engineering.

Key learning points

(1) A historical and state of the art overview of applied biocatalysis.
(2) The advantages and limitations of biocatalysis.
(3) The dierent components of biocatalysis engineering: substrate, medium, protein, biocatalyst and bioreactor engineering.
(4) Examples of the application of biocatalysis engineering in the improvement and invention of biotransformations.
(5) The importance of following a holistic approach where combinations of the dierent components give the optimum results.

1. Introduction
The use of enzymes, before their existence was even known,
dates back thousands of years to bread- and cheese-making,
beer brewing and wine-making. Food and beverage processing
together with animal feed and detergents still account for more
than 60% of the total enzyme market, the remainder consisting
mainly of starch processing, leather, and pulp and paper, and
this is reflected in the three highest volume industrial enzymes
being proteases, amylases, and glucose isomerase.1 The first
commercial enzyme preparations were produced in the late
19th and early 20th centuries and include the use of dried
calves stomachs in cheese manufacture and pancreatic extracts
in laundry cleaning. It was the development of recombinant
DNA technology in the 1970s, however, which provided the
basis for the shift from the traditional animal and plant sources
to much more ecient production with recombinant micro-
organisms, thus aording cheaper and higher purity enzymes.
a
Molecular Sciences Institute, School of Chemistry, University of the Witwatersrand,
Johannesburg 2050, South Africa. E-mail: roger.sheldon@wits.ac.za;
Tel: +27 11-7176727
b
Department of Biotechnology, Delft University of Technology, Section BOC, van der
Maasweg 9, 2629 HZ, Delft, The Netherlands. E-mail: r.a.sheldon@tudelft.nl;
Fax: +31 (0)15 2781415; Tel: +31 (0)15 2782675
Electronic supplementary information (ESI) available. See DOI: 10.1039/c6cs00854b
Today almost all industrial enzymes are recombinant, with the
exception of those used in food processing.
Biocatalysis generally refers to the use of enzymes, or
enzyme containing cells, in chemical transformations. There
are basically two types of biotransformations: those involving
growing (microbial) cells and resting cells, respectively. The
former are fermentations, the main goal of which, in vivo, is
to produce more microbial cells but in vitro the objective
is generally the production of chemical products. Buchner
showed in 1897 that the degradation of glucose into ethanol
by living yeast, found by Pasteur in 1860, didnt actually require
living yeast but only the catalytic components he extracted from
the cells. Nonetheless, the term fermentation continued
to be used for transformations with growing microorganisms.
The catalytically active substances present in cell extracts,
later shown to be proteins, were given the name en-zymes,
meaning in-yeast. While acknowledging that advances in
metabolic pathway engineering and the underpinning synthetic
biology2 have revolutionised the microbial production of chemical
and pharmaceutical products, in this review we will focus on
the use of isolated enzymes as biocatalysts.
Organic synthesis with the aid of enzymes has been known
since the early 20th century but until fairly recently, with
some notable exceptions, such as the synthesis of beta lactam
antibiotics,3 was not widely used in industrial processes.

This journal is The Royal Society of Chemistry 2017 Chem. Soc. Rev.

Tutorial Review
This situation changed in the mid-1980s following two
coincidental developments. First, the landmark publication
by Zaks and Klibanov4 in 1984, describing enzymatic catalysis
at 100 1C in organic media, demonstrated that many enzymes
were actually more thermally stable in organic solvents, such as
toluene, than in water. This was a revelation for most synthetic
organic chemists since conventional wisdom held that enzymes
work well only in water. It led to the realisation among organic
chemists that biocatalysis had a much broader scope in organic
synthesis than was previously imagined, totally and lastingly
Published on 13 March 2017. Downloaded by University of Newcastle on 13/03/2017 16:02:53.
changing the biocatalysis landscape. Second, at about the same
time, recognition of the importance of stereoisomerism in drug
action signaled the advent of regulations by, for example,
the FDA which required that producers separate and test both
enantiomers of chiral drugs. In commercial practice this
generally led to the decision to produce and market the drug
as a single, active enantiomer. It created a need for cost-
eective methods for enantioselective synthesis and biocatalytic
methods were obvious candidates since enzymes are usually
highly enantioselective.
Hence, biocatalysis evolved from an academic curiosity to an
industrially attractive technology for the enantioselective synth-
esis of APIs.5 However, in the early 1990s a major obstacle to
their widespread use was the number of commercially available
enzymes. This was largely limited to those already in use in the
food and beverages and detergent industries: mainly proteases,
lipases, esterases and glycosidases, often from animal origin.
Over the last two decades this situation has changed dramati-
cally thanks to modern biotechnology. As a result of high
throughput DNA sequencing, more than 20 000 whole genome
sequences have become publicly available. Thanks to advances
in recombinant DNA technology it is now possible to identify a
target gene, by in silico analysis of a genome sequence data
base, have the gene synthesised chemically within weeks, ready
for cloning into a host production organism and at relatively
Roger Sheldon (www.sheldon.nl) is
a recognised authority on Green
Chemistry. He developed the E
factor for assessing the environ-
mental impact of chemical
processes. He is currently Distin-
guished Professor of Biocatalysis
Engineering at the University of
the Witwatersrand (SA). He
received the RSC 2010 Green
Chemistry Award and the Biocat
2010 lifetime achievement. He was
Roger A. Sheldon elected a Fellow of the Royal
Society in 2015. He has a PhD
and DSc from Leicester University (UK) and was Professor at Delft
University (NL) (19912007), CEO of CLEA Technologies (2006
2015), VP R&D at DSM-Andeno (19801990) and with Shell Research
Amsterdam (19691980).
Chem. Soc. Rev.
View Article Online
Chem Soc Rev
low cost. Consequently, more enzymes are available and they
can be produced for commercially acceptable prices, making
the notion that enzymes are expensive obsolete.
Furthermore, protein engineering techniques, such as directed
(in vitro) evolution,6 have enabled the engineering of enzymes
such that they exhibit pre-defined properties with regard to,
inter alia, substrate specificity, activity, selectivity, stability and
pH optimum. Lastly, the development of eective immobilisation
techniques has improved their operational stability and cost-
eectiveness by facilitating recovery and recycling.7
Consequently, in the last decade biocatalysis has been
integrated into mainstream organic synthesis, particularly in the
pharmaceutical industry.8 Indeed, Turner and OReilly9 proposed the
introduction of guidelines and rules for biocatalytic retrosynthesis
to aid synthetic chemists in identifying where biocatalysts could be
applied to the synthesis of target molecules.
The broad application of biocatalysis can also be attributed to its
numerous environmental and economic benefits. Enzymes are
natures sustainable catalysts. They are derived from renewable
resources and are biocompatible, biodegradable, essentially non-
hazardous and non-toxic. Biocatalysis avoids the use of scarce
precious metals and the associated, often prohibitive, costs of
removing traces of noble metals, to an acceptable ppm level, from
end products. Enzymatic reactions are performed under mild
conditions (physiological pH and ambient temperature and
pressure) in water, often without the need for functional-group
activation and protection and deprotection steps required
in conventional organic syntheses. This translates to shorter
syntheses, higher selectivities and purer products in processes
that are more resource and energy ecient and generate less
waste than conventional routes. Moreover, enzymatic processes
can be conducted in standard multi-purpose batch reactors and,
hence, do not require any extra investment, e.g. for high-pressure
equipment. Lastly, most enzymatic reactions are conducted
under roughly the same conditions of temperature and pressure
Pedro Cherando Pereira graduated
in chemical engineering at the
Instituto Superior Tecnico (Lisbon,
Portugal), specializing in bio-
technology and doing his research
graduation project with Prof.
Sheldons group, at the Technical
University of Delft (The
Netherlands). After graduating, he
conducted organic chemistry
research at the same University
in Lisbon with Prof. M. M.
Pedro C. Pereira Marques, but later returned to
Prof. Sheldons group, where he
obtained his PhD (Biocatalysis Engineering: Developing Enzymatic
Systems) and a post-doc fellowship. Outside academia, he worked
several years with ViaZym, a biotech company, as an enzyme
technology scientist.
This journal is The Royal Society of Chemistry 2017

Chem Soc Rev
and, hence, it is relatively easy to integrate multiple reactions
into eco-ecient catalytic cascade processes (see later).
Wild-type enzymes have evolved over millions of years to be
able to convert their natural substrates in high rates in vivo. It
is not surprising therefore, that they have a problem with
sustaining a high activity and productivity with non-natural
substrates under more challenging conditions in vitro, such
as high substrate concentrations and non-aqueous media.
Nonetheless, spectacular results have been obtained in some
cases with wild-type enzymes. An early example is the enzymatic
Published on 13 March 2017. Downloaded by University of Newcastle on 13/03/2017 16:02:53.
synthesis, developed and commercialised by DSM-Andeno in the
1980s, of the key chiral intermediate in the manufacture of the
anti-hypertensive drug, diltiazem. It involved a highly enantio-
selective hydrolysis of a chiral glycidate ester catalysed by
Thermomyces lanuginosus lipase (E.C. 3.1.1.3), otherwise known
as lipolase (Fig. 1a),10 an inexpensive, readily available enzyme
used in a wide variety of industrial applications.
More recently, Pfizer developed an extremely eective
chemoenzymatic process (Fig. 1b), using the same lipase, for
the manufacture of pregabalin,11 the active ingredient of the
CNS drug Lyricas. The key enzymatic step was conducted at an
impressive 3 M (765 g L 1) substrate concentration in a largely
aqueous process with dramatically reduced organic solvent
usage. Compared to the first generation manufacturing process,
the new one aorded a higher yield and a five-fold reduction in
the E factor from 86 to 17.
Notwithstanding these and other impressive examples
of the industrial application of wild-type enzymes, with most
non-natural substrates optimisation of the biocatalytic step is
necessary for commercial viability. Enter biocatalysis engineering
which is concerned with optimisation of the use of enzymes as
catalysts, either by modifying the enzyme or the reaction system.
A biocatalytic process comprises many variables, of which the
Fig. 1 Diltiazem and pregabalin processes.
This journal is The Royal Society of Chemistry 2017
View Article Online
Tutorial Review
biocatalyst is merely one part. In this tutorial we shall describe
dierent strategies for biocatalysis engineering aimed at
improving existing biotransformations or creating new ones.
These include engineering of the substrate, the reaction
medium, the protein, and post-translational modification of
the enzyme. Finally, reactor engineering in combination with
downstream processing facilitates product recovery and re-use
of the biocatalyst. Hence, the holistic approach of biocatalysis
engineering can aord an environmentally and economically
viable biotransformation by optimising all the variables.
2. Substrate engineering
Substrate engineering involves varying the substrate structure,
leading to both the optimisation of existing biotransformations
and the invention of new ones. It is often referred to as the use
of non-natural substrates, which gave rise to the term enzyme
promiscuity.12 A powerful consequence of this property is that it
can lead to the discovery of totally dierent reaction types.
Knowledge of the reaction mechanism is almost a conditio sine
qua non for designing a strategy for substrate engineering. For
example, the in vivo hydrolysis of triglycerides, to a mixture of
glycerol and fatty acids, is catalysed by lipases. These enzymes
belong to the group of serine proteases, a common feature of
which is the involvement of a serinehistidineaspartate triad
in the so-called electron relay mechanism (Fig. 2). The acyl
enzyme intermediate is formed by reaction of the substrate
with the OH group of a serine residue in the active site which is
assisted by the His and Asp groups. Subsequent reaction of the
acylenzyme intermediate with water affords the free fatty acid.
Close inspection of this mechanism easily leads to the idea that
other, non-natural nucleophiles (acyl acceptors) could possibly
be used instead of water.
Chem. Soc. Rev.
 View Article Online

Tutorial Review Chem Soc Rev

amides from esters. However, the use of the simplest nitrogen

nucleophile, ammonia, had not been described. Hence, a new
enzymatic reaction, ester ammoniolysis, was invented when
ethyl octanoate was allowed to react with a saturated solution of
ammonia in tert-butanol in the presence of lipases (Fig. 3a).13
Most lipases showed some activity but Candida antarctica
lipase B (CaLB) and Thermomyces lanuginosus lipase (lipolase)
were the most active. The method was used,14 for example, to
produce white crystals of oleic acid amide in 94% yield by
reaction of olive oil (triolein) with a saturated solution of
Published on 13 March 2017. Downloaded by University of Newcastle on 13/03/2017 16:02:53.

ammonia in tert-butanol at 60 1C (Fig. 3b). Amides could

Fig. 2 The electron relay mechanism.
Indeed, it was already known in the early 1990s that amines,
could be used as the nucleophile, leading to the formation of
also be prepared from the corresponding acids in a one-pot
procedure involving CaLB catalysed esterification followed by
CaLB catalysed ammoniolysis of the resulting ester.
Enzymatic ammoniolysis is also an eective method for
the kinetic resolution of chiral carboxylic acid esters and
ammoniolysis of amino acid esters, catalysed by both lipases
and serine proteases, aord the corresponding amino amides.
For example racemic phenylglycine methyl ester was converted
to D-phenylglycine amide, an intermediate in the manufacture
of semi-synthetic penicillins and cephalosporins (Fig. 3c).14
Another striking example of non-natural nucleophiles in an
enzymatic reaction involves halohydrin dehalogenases (HHDHs).
In vivo HHDHs catalyse the ring closure of chlorohydrins to
epoxides and the reverse reaction: ring opening of an epoxide
with chloride ion (Fig. 4). The mechanism involves a halide ion
binding site and one might expect that halide ion could be
exchanged with other nucleophiles which could subsequently

Fig. 3 Enzymatic ester ammoniolysis.

Chem. Soc. Rev. This journal is The Royal Society of Chemistry 2017

Chem Soc Rev
Published on 13 March 2017. Downloaded by University of Newcastle on 13/03/2017 16:02:53.
Fig. 4 Epoxide ring opening reactions catalysed by HHDH.
react with the epoxide. Indeed, HHDHs are highly promiscuous
enzymes that can accept at least nine other anions as nucleo-
philes in the enantioselective ring opening of epoxides,15
forming the basis for the highly enantioselective synthesis of
a broad range of b-substituted alcohols. The use of cyanide ion
as the nucleophile, for example, is particularly interesting as it
generates a new CC bond. It has been applied in a process,
commercialised by Codexis, for the synthesis of a key inter-
mediate in the manufacture of atorvastatin,16 the active ingre-
dient of Pfizers cholesterol lowering agent Lipitor (Fig. 4).
A more recent, remarkable example of enzyme promiscuity
designed by intuitive, mechanism-based substrate engineering
is cyclopropanation via carbene transfer catalysed by an engi-
neered cytochrome P450-dependent monooxygenase.17 In vivo
these enzymes catalyse the aerobic oxidation of a wide variety of
organic substrates via a high-valent oxo-iron species as the
active oxidant. Reaction of the latter with an olefin leads to
transfer of the oxygen atom (oxene transfer) to the double bond,
aording an epoxide (Fig. 5). By analogy, the formation of a high-
valent iron-carbenoid species from a diazo ester as a co-substrate
was envisaged. Reaction of this putative intermediate with an
olefin would result in carbene transfer to the double bond,
Fig. 5 Enzymatic cyclopropanation by carbene transfer.
This journal is The Royal Society of Chemistry 2017
View Article Online
Tutorial Review
aording a cyclopropane. This proved to be the case: reaction of
styrene with ethyl diazo acetate, in the presence of 0.2 mole%
of P450BM3 aorded the expected cyclopropane carboxylic
ester (Fig. 5).
Finally, it is worth emphasizing that substrate engineering
can be approached in a systematic way based on an eective
combination of a knowledge of the three dimensional structure
of the active site of the enzyme and the mechanism of the
enzymatic transformation with a good understanding of
organic reaction mechanisms in general.
3. Medium engineering
As we already noted (see above), the 1984 paper of Zaks
and Klibanov served as a wake-up call to organic chemists.
Enzymes generally function optimally in water but they are also
active in organic solvents. Indeed, biocatalysis in organic media
has several benefits: many organic substrates are at best
only sparingly soluble in water, and some reactions e.g. (trans)-
esterifications and amidations, cannot be conducted in water
owing to equilibrium limitations and/or competing product
hydrolysis. Additional benefits of non-aqueous biocatalysis are
easier product recovery from volatile organic solvents and
elimination of microbial contamination. In short, engineering
of the reaction medium can be used to optimise the synthetic
potential of biocatalytic transformations.
Although enzymes are able to perform as suspensions in
organic solvents, their catalytic eciencies are generally two
orders of magnitude lower than those observed in water. It
should be noted, however, that it is dicult to meaningfully
compare the rates of reactions conducted in a homogeneous
enzyme solution in water with reactions catalysed by the same
enzyme as a heterogeneous suspension of a solid in an organic
solvent. Moreover, the environmental issues associated with
the use of volatile organic solvents (VOCs) represent another
drawback of reactions in organic media. In this context it
should also be mentioned that many biocatalytic reactions
are performed, both on laboratory and industrial scale, in an
Chem. Soc. Rev.

Tutorial Review
aqueous biphasic system, consisting of a, preferably environ-
mentally attractive, organic solvent and water. The reaction
takes place in the aqueous phase and the substrate and product
are dissolved primarily in the organic phase. A good example is
the three enzyme process for atorvastatin intermediate referred
to earlier which is conducted in aqueous ethyl acetate.
The best solvent is no solvent but if a solvent (diluent) is
needed supercritical carbon dioxide, scCO2, is a potentially
attractive alternative to VOCs. It is non-flammable, non-toxic
and is the only readily available, inexpensive solvent that is
Published on 13 March 2017. Downloaded by University of Newcastle on 13/03/2017 16:02:53.
supercritical under conditions (31 1C and 7.4 MPa) which are
conducive to biocatalysis. In short, a natural catalyst in a
natural solvent. Furthermore, supercritical fluids combine the
solubilising capacity of a liquid with the low viscosity of a gas,
resulting in high rates of mass transfer, and the product can be
recovered by simply reducing the pressure. Enzymes generally
exhibit good activity and stability in scCO2 but there are two
properties which can have a detrimental eect: (i) amino groups
in lysine residues can react with CO2 to aord carbamates, and
(ii) CO2 can react with water to give carbonic acid resulting in a
drop in pH.
Immobilisation of the enzyme using the CLEA methodology
(see later) suppresses inactivation by converting free amino
groups on the surface. In the kinetic resolution of 1-phenyl-
ethanol by CaLB-CLEA catalysed transesterification in scCO2 in
continuous mode, reaction rates were higher than those
observed in n-hexane.18 A benefit of using liquid CO2 is that
it can be maintained under relatively modest pressure (4.5 MPa
at 10 1C). In another recent development transesterification of
triglycerides in scCO2 catalysed by lipases immobilised in
monoliths in packed bed reactors was used for continuous
biodiesel production.19
The activity of enzymes in organic solvents can be increased
by lyophilisation in the presence of relatively large amounts of
salts such as potassium chloride. This led to the idea that
suspension of an enzyme in a room temperature ionic liquid
(IL), with its salt- and water-like character, could aord signifi-
cant rate enhancements compared to organic solvents. ILs are
composed entirely of ions and are liquid at or close to ambient
temperature. They have been widely advocated as potentially
attractive alternatives to volatile organic solvents (VOCs).
The first examples of biocatalysis in a water-free IL are
the transesterifications and amidations (Fig. 6) catalysed by a
Fig. 6 Biocatalysis in water-free ILs.
Chem. Soc. Rev.
View Article Online
Chem Soc Rev
suspension of CaLB in anhydrous [bmim][PF6] and [bmim][BF4].20
To ensure that they were anhydrous, both the IL and the enzyme
were dried over phosphorus pentoxide prior to use. Although the
rates were only marginally higher than those observed in the best
organic solvents these results demonstrated the compatibility
of enzymes with ILs and marked the advent of prolific studies
of biocatalysis in ILs in the next 15 years.
The use of ILs as reaction media can also result in increased
enantioselectivity or stability resulting from conformational
changes of enzymes in IL media. Based on their ability to dissolve
large amounts of highly polar substrates they are potentially
interesting media for biotransformations of carbohydrates,
nucleosides and polysaccharides which are, at best, only
sparingly soluble in common organic solvents.
An important driver for using ILs is the possibility of
replacing VOCs with non-volatile ILs, thereby circumventing
the risk of air pollution. However, ILs have significant solubility
in water and first generation ILs, such as dialkylimidazolium
and tetraalkylammonium salts exhibit aquatic ecotoxicity and
poor biodegradability. Moreover, their preparation often involves
circuitous processes, making them prohibitively expensive.
Hence, the current trend is towards the rational design of task
specific, biocompatible ILs that combine cost-eectiveness with
a low environmental footprint. Cholinium ILs, for example, are
prepared by reaction of inexpensive choline hydroxide with a
wide variety of carboxylic acids or amino acids.21
Another class of IL comprises the protic ionic liquids (PILs).
The latter are easily prepared, by simply mixing a tertiary amine
with a (carboxylic) acid, and exhibit better biodegradability and
lower toxicity than the corresponding quaternary ammonium
salts. Moreover, they have suitable H-bond donating properties
for interaction with and stabilisation of enzymes and, when
combined with carboxylate anions, they are self-buering. PILs
derived from a variety of tertiary amines and carboxylic acids
were successfully used, for example, as the reaction medium for
enantioselective transesterification of a-methyl benzyl alcohol
(Fig. 7) catalysed by CaLB-CLEA (see Section 5).22
An important issue regarding reactions in ILs is: how can the
product be separated from the ionic liquid? An attractive option
is to use supercritical carbon dioxide, scCO2, to extract the
product. The IL has no measurable solubility in scCO2 but
the latter is highly soluble in the IL phase and can extract
hydrophobic molecules. This provides the basis for biphasic
This journal is The Royal Society of Chemistry 2017

Chem Soc Rev
Published on 13 March 2017. Downloaded by University of Newcastle on 13/03/2017 16:02:53.
Fig. 7 CaLB-CLEA catalysed transesterification in PILs.
biocatalysis in which the scCO2 acts as a mobile phase for the
continuous removal of products from the IL.23 The product is
obtained by decompression of the scCO2 which can be recycled
by recompression. The concept was applied, for example, to
the CaLB catalysed kinetic resolution of 1-phenylethanol in
high enantioselectivities (ee 4 99.9%) and with good opera-
tional stability.
In addition to second generation ILs and PILs another class
of neoteric solvents has emerged in recent years: so-called deep
eutectic solvents (DES).24 They are easy to synthesise, by mixing
certain solid salts, in dierent proportions, with a hydrogen
bond donor (HBD) such as urea and glycerol, and gently heating
the mixture. For example combining choline chloride (m.p.
302 1C) with urea (m.p. 132 1C) aords a DES with a melting
point of 12 1C. They have similar properties to ILs and, because
they are generally made from naturally occurring, biocompatible
substrates, they are non-toxic and biodegradable. Choline
chloride (ChCl), for example, is a readily available, inexpensive
feed additive produced in 41 mio tonnes per annum, urea is a
common fertiliser and glycerol is a byproduct of biodiesel
manufacture. Hence, ChCl/urea (ChCl/U; 1 : 2 molar ratio) and
ChCl/glycerol (Ch/Gly; 1 : 2) are readily available, inexpensive,
biocompatible and biodegradable. DESs have also been prepared
from ChCl and a variety of renewable HBDs such as lactic acid,
amino acids and carbohydrates (Fig. 8).
Kazlauskas and coworkers25 first reported that various
hydrolases show excellent activity retention in DESs. For
example, immobilised CaLB (Novozym 435) catalysed the trans-
esterification and amidation of ethyl valerate with 1-butanol
and 1-butylamine, respectively (Fig. 9) in ChCl/U (1 : 2) and
ChCl/Gly (1 : 2), exhibiting conversions comparable with those
observed in toluene. These results were surprising since urea is
known to be a potent protein denaturant. A plausible explana-
tion is that the urea is hydrogen bonded to the ChCl thus
This journal is The Royal Society of Chemistry 2017
View Article Online
Tutorial Review
Fig. 8 Examples of cholinium ILs and DESs.
preventing its diusion into the core of the protein. A further
surprising observation was that transesterifications in ChCl/Gly
(8 M glycerol) gave more than 90% conversion with o0.5%
glyceryl ester formation, which can similarly be explained by
assuming that the glycerol is hydrogen-bonded to the ChCl,
preventing its access to the active site.
4. Protein engineering
As noted above, wild-type enzymes are often not eective with
non-natural substrates under the harsh conditions of industrial
processes which leads to low selectivities, activities and volumetric
productivities. Hence, they need to be re-designed to give high
space time yields and high selectivities at high substrate concen-
trations and low enzyme loadings. Moreover, this process should
not take millions of years. This can be achieved using directed
evolution, also known as in vitro evolution or evolution in the test
tube, to generate libraries of mutant enzymes to be screened
for improved properties. Alternatively, space-time yields can be
increased by simply using higher enzyme concentrations but,
in addition to being less economical, this generally leads to
problems in down-stream processing owing to the formation of
dicult to separate emulsions.
Rational design by site-directed mutagenesis (SDM), a genetic
engineering tool introduced by Smith and coworkers26 in the
late 1970s, involves the creation of so-called point mutations
whereby a given amino acid at a predetermined site in the
protein is replaced by one of the other 19 canonical amino acids.
An important shortcoming of SDM is that detailed information
is required regarding the three dimensional structure and
mechanism of the enzyme. In contrast, random mutagenesis
requires no structural information and in the early 1990s error-
prone polymerase chain reaction, ep-PCR, was used to generate
libraries of mutants. A seminal paper, published by the group
of Arnold27 in 1993, involved the use of ep-PCR to generate
Chem. Soc. Rev.
 View Article Online

Tutorial Review Chem Soc Rev

Fig. 9 Biocatalysis in deep eutectic solvents.

Published on 13 March 2017. Downloaded by University of Newcastle on 13/03/2017 16:02:53.

mutants of subtilisin E, an industrially important serine protease,

with increased stability in organic solvents such as DMF.
Although ep-PCR is an eective means to introduce changes
into a protein, in natural evolution the eect is amplified by an
additional mechanism: recombination. Hence, techniques
were developed to mimic the latter in vitro. A landmark in the
development of directed evolution was the report by Stemmer
in 1994, of DNA shuing.28 In this technique DNase is used to
fragment a set of parent genes and a primer-free PCR step
recombines fragments from diverse parent genes into new
chimeras which can then be cloned into an expression vector
and screened. This sequence can then be repeated until mutant
enzymes with the sought after improved properties are obtained
Fig. 10 Three enzyme process for atorvastatin intermediate.
in a Darwinian type evolution in vitro. Although early studies
of directed evolution were mainly focused on increasing the
stabilities of enzymes the involvement of organic chemists Table 1 Evolution of a KRED/GDH biocatalyst
turned attention to improving another important property of
Parameter Wild-type Best variant
enzymes, namely stereoselectivity. A pioneering, proof-of-concept
for improving enantioselectivity by directed evolution was TTN catalyst 3000 4100 000
TTN NADP 4000 420 000
reported by Reetz and Jaeger and coworkers29 in 1997. STY (g L 1 day 1) 80 600
Over the subsequent decades DNA shuing, and other Yield (%) 80 495
directed evolution strategies have been widely applied in ee (%) 99.8 499.9
[Enzyme] (g L 1) 100
the improvement of existing activities30 and evolving new [Substrate] (g L 1) 80 200
activities of enzymes.31 Indeed, it is now possible to predefine Reaction time (h) 24 10
the parameters of the dream process and then use directed Phase separation 41 h ca. 1 min
Work-up Complex Very simple
evolution to modify the biocatalyst to fit it. This contrasts
dramatically with the traditional approach of modifying the
process to fit the commercially available biocatalyst. Indeed, exhibited by the wild-type KRED. The parameters of the wild-type
key advances in DNA sequencing and gene synthesis, in combi- KRED/GDH combination are compared with the best variant,
nation with directed evolution techniques, have paved the way obtained after several rounds of DNA shuing, in Table 1. Because
for the widespread industrial application of biocatalysis, parti- of the much lower enzyme loading, no emulsion problems were
cularly in the synthesis of active pharmaceutical ingredients encountered and phase separation required less than one minute.
(APIs), in the last decade.32 A pertinent example is the two-step, Similarly, the activity of the wild-type HHDH in the hydro-
three enzyme process for the synthesis of a key intermediate for cyanation step was extremely low and the enzyme showed poor
atorvastatin (Fig. 10) in which a key step is a HHDH catalysed stability in the presence of substrate and product and strong
conversion of a chlorohydrin to a cyanohydrin (see Section 2).16 product inhibition. After many rounds of DNA shuing and
Proof-of-concept on a lab scale was obtained with the wild-type screening, in the presence of increasing product concentrations,
KRED, GDH and HHDH. In the first step the chlorohydrin the inhibition was largely overcome and the HHDH activity
product was obtained in 85% yield and 499.5% ee. increased 42500 fold. The wild-type HHDH is compared with
However, the activities were too low for commercial viability the best variant in Table 2. The substrate loading was increased
and the use of high enzyme loadings led to emulsion formation from 20 to 140 g L 1, the enzyme loading reduced from 30 to
and problematical product recovery. DNA shuing was used16 1.2 g L 1 and the reaction time to completion from 72 to 5 h.
to improve the activity and stability of the KRED and the This process was substantially greener than previous pro-
GDH, while maintaining the nearly perfect enantioselectivity cesses when assessed according to the twelve principles of green

Chem. Soc. Rev. This journal is The Royal Society of Chemistry 2017
 View Article Online

Chem Soc Rev Tutorial Review

Table 2 Evolution of a HHDH biocatalyst by DNA shuing required and were subject to acetone inhibition. Subsequent
directed evolution using DNA shuffling was focused on improv-
Parameter Process design Wild-type Best variant
1
ing the activity and stability. After an initial three rounds
[Substrate] (g L ) 120 20 140 of evolution the product/enzyme mass ratio was improved 400
[Enzyme] (g L 1) 1.5 30 1.2
Catalyst productivity (g g 1) 80 0.7 117 fold from 1 : 50 to 8 : 1. Further improvements were obtained
STY (g L 1 day 1) 4360 7 672 by optimising the reaction medium to a toluene/water mixture
Isolated yield (%) 490 67 92 and raising the reaction temperature to 40 1C. The final variant
Chemical purity (%) 498 498 498
ee (%) 499.5 499.5 499.5 exhibited a 3000-fold improvement and the process could be
Reaction time (h) 8 72 5 performed in a slurry-to-slurry mode at 100 g L 1 substrate
Phase separation (min) o10 460 o1
loading and 3 g L 1 enzyme concentration in 24 h at 45 1C,
Published on 13 March 2017. Downloaded by University of Newcastle on 13/03/2017 16:02:53.

affording the (S)-hydroxy ester in 495% yield and 499.9% ee.

chemistry. The E factor (kg waste per kg product) for the overall
process is 5.8 if water is excluded and 18 if it is included. The
main contributors to the E factor are solvent (ethyl and butyl
acetates) losses (51%), sodium gluconate (25%) and NaCl and
Na2SO4 (combined 22%). The three enzymes and the NADP
cofactor accounted for o1% of the waste. The dramatic improve-
ment in the key properties of the three enzymes is the epitome of
the power of directed evolution in enabling green-by-design,
economically viable biocatalytic processes that would otherwise
have remained laboratory curiosities.
In the last decade KRED catalysed ketone reductions have
become economically and environmentally attractive alternatives
to stoichiometric reductions with chiral borane reagents and
rhodium catalysed asymmetric hydrogenations.33 The success
of KRED catalysed reductions is largely a result of the develop-
ment of eective cofactor regeneration systems and the use of
protein engineering to improve their catalytic properties and,
hence, commercial viability. A striking example of the latter is
provided by the synthesis of the key intermediate in the manu-
facture of montelukast, the active ingredient of Singulair, an
anti-asthmatic agent. No commercially available KRED showed
any activity towards the ketone substrate (Fig. 11).34 The very low
solubility of the ketone substrate, even in waterorganic solvent
mixtures, was a limiting factor. An initial screening of KREDs
afforded a few with 499.9% (S)-selectivity, using isopropanol for
cofactor regeneration. However, activities were three orders of
magnitude lower than that required for commercial viability and
the enzymes were unstable at the high level of organic solvents
Yet another protein engineering tour de force is provided by
the synthesis of the key intermediate for sitagliptin, the active
ingredient of the antidiabetic agent, Januvia. Teams at Codexis
and Merck used a palette of protein engineering techniques,
to develop a transaminase for the conversion of the ketone
precursor to sitagliptin (Fig. 12).35 The starting point was
an (R)-selective transaminase (TA), possessing the necessary
catalytic machinery to perform the desired transformation but
which lacked any activity towards the prositagliptin substrate.
A combination of computer-aided catalyst design of the active
site and site saturation mutagenesis was used to produce a
transaminase that, at an enzyme loading of 10 g L 1 and a
substrate loading of 2 g L 1 gave 0.7% conversion in 24 h.
As with the montelukast example the prositagliptin is only
sparingly soluble (o1 g L 1) in water, necessitating substantial
amounts of DMSO as a cosolvent.
In order to achieve the parameters required for commercial
viability the enzyme characteristics had to be improved to
withstand the harsh conditions of 100 g L 1 substrate, 1 M
isopropylamine, 425% DMSO, a temperature of 440 1C for
24 h to give a product with 499.9% ee. This was achieved using
directed evolution with, inter alia, DNA shuing. The final
variant contained 27 mutations and, of the 17 noncatalytically
essential amino acids predicted to interact with the substrate,
10 were mutated. Using 6 g L 1 of this variant in 50% DMSO
at 45 1C, 100 g L 1 substrate was converted to sitagliptin of
499.95% ee (the other isomer was below detection level) in
92% yield. Compared with the rhodium catalysed asymmetric
hydrogenation of an eneamine which it replaced, the biocatalytic
process provided a 13% increase in overall yield, a 53% increase
in productivity (kg L 1 day 1), a 19% reduction in E factor, the
elimination of all heavy metals and a reduction in total manu-
facturing costs. Moreover, the biocatalytic process is conducted in
standard multipurpose reactors obviating the need for specialised
high pressure hydrogenation equipment. Furthermore, the
enzymes developed for sitagliptin synthesis were shown to have
broad scope in the synthesis of chiral amines, which is of general
interest in the pharmaceutical industry.

5. Biocatalyst engineering
Having identified an enzyme for the targeted transformation,
optimised its properties using substrate, medium and protein
Fig. 11 Montelukast intermediate by KRED reduction. engineering techniques, the enzyme is expressed in a microbial

This journal is The Royal Society of Chemistry 2017 Chem. Soc. Rev.

Tutorial Review
Published on 13 March 2017. Downloaded by University of Newcastle on 13/03/2017 16:02:53.
Fig. 12 Synthesis of sitagliptin by biocatalytic transamination.
production host with a GRAS (Generally Regarded as Safe) status to
enable its overproduction in large amounts at relatively low cost.
The next step is to identify an eective formulation of the enzyme.
Enzymes are soluble in water and it is dicult and costly to recover
them from aqueous euents, although this can be achieved using
ultrafiltration. Hence, many enzymes are used on a single use,
throw-away basis. The enzyme costs per kg of product can be
substantially reduced by immobilisation of the enzyme to form a
free-flowing powder, i.e. by creating a heterogeneous catalyst to
enable catalyst recovery and reuse,7 thereby resulting in process
simplification combined with a higher product quality and a
smaller environmental footprint. It constitutes post-translational
biocatalyst engineering, as opposed to (pre-translational) protein
engineering. Additionally, enzyme immobilisation generally
results in increased stability, by suppressing the unfolding
and, hence, denaturation of the enzyme. This enables, in turn,
the use of a much broader choice of solvents.
A case in point is the aforementioned (R)-selective trans-
aminase (TA) developed, using a combination of computer-aided
modelling and directed evolution, for the synthesis of sitagliptin.
The optimum protocol involved the use of aqueous DMSO as the
reaction medium. Subsequently, Merck scientists36 investigated
the immobilisation of the TA on a variety of polymer-based
resins and compared the immobilisates with the lyophilised free
TA. The best results were obtained using the TA adsorbed on a
highly hydrophobic octadecyl functionalised polymethacrylate
resin with 4% loading of the TA, resulting in 45% activity
recovery. The immobilised TA was shown to be active in a variety
of organic solvents, the best results being obtained in the
environmentally attractive, isopropyl acetate which combined a
favourable solubility of the ketone substrate with increased
stability of the TA compared to other solvents. The TA was found
Chem. Soc. Rev.
View Article Online
Chem Soc Rev
to be remarkably robust in dry isopropyl acetate at 50 1C, with a
deactivation rate of 0.5% h 1 over 6 days. When water-saturated
isopropyl acetate was used no deactivation was observed over the
same time period and 10 consecutive cycles were obtained, with
no detectable loss of activity over a 200 h period. In contrast, the
soluble TA was completely denatured and showed no activity in
the organic solvent.
The combination of increased stability and reuse of the
immobilised TA aorded a 490% reduction in the enzyme
loading compared to the soluble enzyme. Interestingly, the
immobilised TA was used successfully in the synthesis of chiral
amines, in high enantiomeric purity, from the corresponding
ketone precursors. The use of an organic solvent represents a
significant advantage compared to the aqueous protocol which
requires the use of buer and continuous pH control during
the reaction. Organic solvent is then used to extract the product
and the mixture filtered to remove the denatured enzyme. The
remaining aqueous euent constitutes a solvent contaminated
waste stream. In contrast, the use of an immobilised enzyme in
an environmentally attractive organic solvent obviates the need
for buer, continuous pH control and laborious removal of the
denatured enzyme. This substantially simplifies the work-up
and reduces the processing time and the amount of waste
generated. Moreover, the enzyme can be reused over and over
again and the bottom line is a more commercially attractive
process than the rhodium catalysed asymmetric hydrogenation
of an enamine which it replaced.
Enzyme immobilisation usually results in some loss of activity,
compared with the soluble enzyme, but this is more than
compensated by the increase in stability coupled with reusability
which can aord dramatic cost reductions.1 Two important
landmarks were the immobilisation of glucose isomerase and
This journal is The Royal Society of Chemistry 2017

Chem Soc Rev
penicillin G amidase. Glucose isomerase catalyses the isomer-
isation of D-glucose to D-fructose and is one of the most impor-
tant industrial enzymes in use today, driven by the extensive use
of high fructose corn syrup (HFCS) as a sweetener in food and
beverages. The immobilisation of glucose isomerase dates back
more than 50 years and all current commercial applications
involve the use of an immobilised form. It is the largest process
involving an immobilised enzyme; more than 500 tons are used
to produce ca. 106 tons of HFCS, corresponding to productivities
of 410.000 kg kg 1, on an annual basis. Penicillin G amidase
Published on 13 March 2017. Downloaded by University of Newcastle on 13/03/2017 16:02:53.
(PGA) catalyses the hydrolysis of penicillin G to 6-amino peni-
cillanic acid (6-APA), a key intermediate in the manufacture of
penicillins and cephalosporins with an annual production of
12 000 tons.3 Productivities of 600 kg kg 1 are obtained and are
necessary for commercial viability. It is worth emphasizing that
the two products, HFCS and 6-APA, have very dierent values but
so do the respective enzymes. Penicillin G amidase is a more
expensive enzyme than glucose isomerase. Hence, in both cases
high productivities are needed to achieve economic viability.
Methods for enzyme immobilisation can be divided into three
categories: binding to a solid support (carrier), entrapment and
cross-linking. Binding to a prefabricated support can be physical
(e.g. hydrophobic), ionic, or covalent. Physical binding is often
too weak to keep the enzyme fixed to the carrier under rigorous
industrial conditions of high reactant and product concentra-
tions and high ionic strength (for an exception see the example of
the immobilised TA discussed above). Ionic binding is generally
stronger and covalent binding precludes leaching of the enzyme
from the surface but has the disadvantage that if the enzyme is
irreversibly deactivated, both the enzyme and the (often costly)
support are rendered unusable. Typical supports are synthetic
resins, biopolymers, such as polysaccharides, or inorganic solids
such as (mesoporous) silicas.7
Entrapment involves inclusion of an enzyme in an organic
or inorganic polymer matrix, such as polyacrylamide and silica
solgel, respectively, or a membrane device such as a hollow
fiber or a microcapsule. Physical constraints generally are
too weak, however, to prevent enzyme leakage entirely and
additional covalent attachment is often required. Entrapment
generally requires the synthesis of the polymeric matrix in the
presence of the enzyme.
The use of a carrier inevitably leads to dilution of activity,
owing to the introduction of a large portion of non-catalytic
ballast, ranging from 90% to 499%, resulting in lower space-
time yields and productivities. Increasing the enzyme loading
doesnt help as loss of activity occurs because some of the
enzyme molecules become inaccessible when they consist of
multiple layers on the carrier surface or are situated deeply
within the carrier pores. Consequently, there is an increasing
interest in carrier-free immobilised enzymes, such as cross-
linked enzyme aggregates (CLEAs)37 which oer clear advan-
tages: high catalyst and volume productivities, high stability
and low production costs owing to the exclusion of an addi-
tional (expensive) carrier.
CLEAs are prepared by simply precipitating the enzyme from
a solution in aqueous buer, at the optimum pH, by adding a
This journal is The Royal Society of Chemistry 2017
View Article Online
Tutorial Review
Fig. 13 Cross-linked enzyme aggregates.
salt, such as ammonium sulfate, or a water miscible organic
solvent or polymer, followed by cross-linking with a bifunctional
reagent. Precipitation of the enzyme aords physical aggregates
of enzyme molecules, held together by non-covalent bonding
without perturbation of their tertiary structure, that is without
denaturation. Subsequent cross-linking of these aggregates
renders them permanently insoluble while maintaining their
pre-organised superstructure, and, hence their catalytic activity
(Fig. 13). Selective precipitation from an aqueous medium is
often used to purify enzymes and, hence, the CLEA methodology
essentially combines purification and immobilisation into a
single unit operation that does not require a highly pure enzyme.
It can be used, for example, for the direct immobilisation of an
enzyme from crude cell lysate. Glutaraldehyde is generally the
cross-linker of choice as it is inexpensive and readily available in
commercial quantities.
The methodology has been commercialised by CLEA Tech-
nologies (www.cleatechnologies.com) and has been applied to a
broad range of hydrolases, oxidoreductases, lyases and trans-
ferases. In a further elaboration, smart magnetic CLEAs were
prepared by conducting the cross-linking in the presence of
functionalised magnetic particles.38 These mCLEAs can be separated
magnetically on an industrial scale, using standard commercial
equipment, to aord novel combinations of bioconversions and
downstream processing. It is particularly eective for applications
involving suspended solids where separation of the immobilised
enzyme from other solids is problematical. Magnetic CLEAs,
can be used, for example, in the conversion of lignocellulose to
2nd generation biofuels.39
Interestingly, two or more enzymes can be co-immobilised
in a so-called combi-CLEA by co-precipitating the enzymes from
aqueous buer and cross linking the aggregates. For example, a
combi-CLEA of a KRED and GDH was developed40 and used
repeatedly in the reduction of the keto ester in the synthesis of the
key atorvastatin intermediate discussed in Section 4. It proved
to be a robust regeneration system for the pyridine nucleotide
co-factor, exhibiting increased thermal and pH stability, high
substrate tolerance and long-term operational stability. The
authors predicted that the technology would be widely applied
in the cost-eective production of chiral alcohols.
Combi-CLEAs are also finding applications in food and
beverages processing. For example, Wilson and coworkers
used a combi-CLEA of two glycosidases, a-L-arabinosidase and
Chem. Soc. Rev.

Tutorial Review
b-glucosidase, for aroma enhancement in wine.41 Similarly, a
trienzyme magnetic combi-CLEA of an a-amylase, pectinase
and cellulase was used for clarification of fruit juices.42
6. Biocatalytic cascade processes:
cell-free synthetic biology
Brevity is the soul of synthesis and the ultimate in green
catalytic methodologies is to telescope multi-step syntheses
Published on 13 March 2017. Downloaded by University of Newcastle on 13/03/2017 16:02:53.
into one-pot catalytic cascades obviating the need for time-
consuming isolation and purification of intermediates. This is
truly emulating the elegant orchestration of enzymatic steps in
metabolic pathways in the living cell. Such cascading syntheses
have several advantages: fewer unit operations, less solvent,
and reactor volume, shorter cycle times, higher volumetric and
space time yields and less waste, which translates to substantial
economic and environmental benefits. Furthermore, coupling
of reactions can be used to drive equilibria towards product
thus avoiding the need for excess reagents. On the other hand,
there are problems to be overcome: catalysts are often incom-
patible with each other and optimum conditions can be very
dierent and catalyst recovery and recycle complicated. Nature
solves the incompatibility problem by compartmentalising
enzymes in dierent parts of the cell. Hence, compartmentalisa-
tion via immobilisation can be the solution in enzymatic cascades.
Moreover, biocatalytic processes generally proceed under roughly
the same conditions in water at ambient temperature and
pressure facilitating their integration in cascade processes.
Biocatalytic cascade processes have become a focus of atten-
tion in recent years, largely motivated by the envisaged environ-
mental and economic benefits.43 The rates of sequential
biocatalytic cascades can be substantially increased by simulating
the close proximity of the enzymes extant in microbial cells by
co-immobilisation of the respective enzymes in, for example,
combi-CLEAs such as the KREDGDH combi-CLEA discussed
in Section 5. An example of a cascade involving a combi-CLEA is
the trienzyme CLEA consisting of a (S)-hydroxynitrile lyase from
Fig. 14 Conversion of benzaldehyde to (S)-mandelic acid with a trienzyme-
CLEA.
Chem. Soc. Rev.
View Article Online
Chem Soc Rev
Manihot esculenta, a non-selective nitrilase from Pseudomonas
fluotrescens EBC 191 and an amidase from Rhodococcus
erythropolis, which catalysed the enantioselective conversion of
benzaldehyde to (S)-mandelic acid in 90% yield and 499% ee
(Fig. 14).44 It was originally developed as a hydroxynitrile lyase/
nitrilase bienzyme combi-CLEA but large amounts of the corres-
ponding amide were formed as a byproduct of the nitrilase-
catalysed hydrolysis and an amidase was added to convert the
amide side product to the acid. Rates were higher than with
mixtures of the respective CLEAs.
7. Reactor engineering and down-
stream processing
In the context of down-stream processing we note that sub-
stantial cost reductions of biocatalytic processes can often be
achieved by employing in situ product removal (ISPR), which
can shift unfavourable equilibria and circumvent product
inhibition and degradation of labile products. However, in this
article we have concentrated on the integration of enzyme
formulation, enzyme reactor configuration and downstream
processing as a means to optimising biocatalytic processes.
The choice of reactor configuration is intimately connected
with downstream processing and is largely influenced by the
industry segment and characteristic product volumes. In the
pharmaceutical and fine chemical industries product volumes
are relatively small and processes are generally conducted in
multi-purpose stirred tank reactors (STRs), equipped with a
propeller stirrer. When the process involves the cost-eective
application of an immobilised enzyme, filtration or centrifuga-
tion is generally used to recover and reuse the biocatalyst.
A problem that is often encountered in the use of STRs is
mechanical attrition of the immobilised enzyme, resulting
from shear forces caused by the propeller stirrer, leading to
the formation of pulverised particles which are dicult to
separate. Several concepts have been used to cope with this
problem. For example, in a membrane slurry reactor (MSR),
which can easily be implemented by making minor modifica-
tions to an existing STR, the immobilised enzyme is retained
inside the reactor because it is too large to pass through the
pores of a membrane patch in the reactor wall. In contrast, the
substrate and product can be pumped in and out of the reactor.
The MSR concept enables the use of a broad range of catalyst
particle sizes including the relatively small particles of CLEAs.
High catalyst loadings, longer catalyst life-times owing to
less mechanical stress, and high volumetric and catalyst pro-
ductivities are some of the many advantages of this system. The
biotransformation and catalyst separation are combined into
a single operation. Its practical utility was demonstrated in
the industrially important hydrolysis of penicillin G to 6-APA
catalysed by penicillin G amidase CLEA.45
The tea-bag concept for containing solid catalysts was
originally used by Houghten46 in the combinatorial synthesis
of peptides. A polystyrene-based resin catalyst was contained in
solvent-permeable polypropylene bags which were dangled as
This journal is The Royal Society of Chemistry 2017

Chem Soc Rev
tea bags in the reaction solution. The so-called perfusion basket
reactor (BR) constitutes a refinement of the tea bag concept.
It consists of a metallic filtration membrane-like module that
retains the immobilised biocatalyst. The BR was used success-
fully, for example, in the degradation of endocrine-disrupting
chemicals in aqueous euents catalysed by laccase-CLEAs.47
Another variation on this theme is the spinning basket
reactor which was used in the enzymatic degumming of rice
bran oil catalysed by an immobilised lecitase.48 In yet another
elaboration, a rotating flow cell reactor, consisting of a catalyst-
Published on 13 March 2017. Downloaded by University of Newcastle on 13/03/2017 16:02:53.
containing compartment attached to a propeller stirrer, was
developed by the company SpinChem (www.spinchem.com). It
was used, for example, by Bornscheuer and coworkers in a
transamination catalysed by an immobilised transaminase.49
An important benefit of this novel variation on the basket
reactor is that, in addition to protecting the biocatalyst from
shear forces caused by the stirrer, it greatly accelerates mass
transfer, thus aording substantially higher reaction rates and
creating the possibility to use much smaller reactors.
Alternatively, the attrition of immobilised enzymes, as a result
of mechanical stirring, can be circumvented by using a bubble
column reactor (BCR). Liese and coworkers, for example, used a
bubble column reactor to conduct the Novozym 435 catalysed
esterification of polyols under solvent-free conditions.50 The pro-
ducts are so-called emollient esters used as cosmetic ingredients.
Commercially viable production of these highly viscous materials
is not feasible in conventional STRs or fixed bed reactors (FBRs).
The water formed in the reaction is removed by entrainment with
pressurised air. The latter also serves to mix the reactants without
causing any significant attrition of the catalyst.
In the oils and fats processing industry continuous pro-
cesses with immobilised enzymes in fixed bed reactors (FBRs)
are the norm. This generally requires relatively large particles in
order to avoid large pressure drops over the column but a
compromise generally has to be found because a large particle
size can lead to mass transfer limitations. A fluidised bed can
be used as an alternative, whereby the reactants are introduced
at the bottom and flow upwards through the column. However,
this requires that the particles are of sucient density that
precludes them being blown out of the column at operational
flow rates. Alternatively, smart magnetic particles, for example
magnetic CLEAs, with relatively small particle size, could be
held in place in a magnetically stabilised fluidised bed.
A particularly challenging type of process with regard to
conducting the reaction and downstream processing is one
which involves suspended solids. In the synthesis of beta lactam
antibiotics, by enzymatic coupling of side chains to the penicillin
or cephalosporin nucleus, the product is precipitated from the
reaction mixture which presents a challenge for recovering the
immobilised enzyme. One possibility, which is practiced on an
industrial scale, involves the use of a sieve bottom STR whereby
the product can diuse through the sieve bottom while the
larger catalyst particles are retained in the reactor and can be
reused with the following batch.
The presence of suspended solids is often the case in the
enzymatic hydrolysis of polysaccharides, such as starch and
This journal is The Royal Society of Chemistry 2017
View Article Online
Tutorial Review
lignocellulose, in the production of first and second generation
biofuels, respectively. In the production of bioethanol, for
example, the hydrolysis step and the subsequent fermentation
can be integrated in a one-pot simultaneous saccharification
and fermentation (SSF) process. As already discussed in
Section 5, in this case smart magnetic CLEAs of the hydrolytic
enzymes could be used to enable their facile recovery and reuse,
with obvious cost benefits.
8. Conclusions and prospects
Buchners studies on cell-free biotransformations, for which
he received the Nobel Prize, was a sign of times to come.
Biocatalysis has benefitted enormously from advances in mole-
cular biology and biotechnology and has become a mainstream
technology for the sustainable production of chemicals, from
pharmaceutical ingredients to other specialty chemicals and
even bulk chemicals, and applications in food and beverage
processing. Moreover, applications of biocatalysis are expected
to increase even further in the future as a result of the transi-
tion from an unsustainable economy based on fossil fuels to a
sustainable bio-based economy. Hence, there is a continuing
need for optimisation of biotransformations from both an
economic and environmental viewpoint. We expect that the holistic
approach of biocatalysis engineering will play an important
enabling role in this development.
References
1 R. DiCosimo, J. McAulie, A. J. Poulose and G. Bohlmann,
Chem. Soc. Rev., 2013, 42, 64376474.
2 G. Stephanopoulos, ACS Synth. Biol., 2012, 1, 514525.
3 K. A. Buchholz, Appl. Microbiol. Biotechnol., 2016, 100,
38253839 and references therein.
4 A. Zaks and A. M. Klibanov, Science, 1984, 224, 12491251.
5 R. Holt, Spec. Chem., 2013, 2123.
6 M. T. Reetz, J. Am. Chem. Soc., 2013, 135, 1248012496.
7 R. A. Sheldon and S. van Pelt, Chem. Soc. Rev., 2013, 42,
62236235.
8 G. W. Huisman and S. J. Collier, Curr. Opin. Chem. Biol.,
2013, 17, 284292 and references therein.
9 N. Turner and E. OReilly, Nat. Chem. Biol., 2013, 9, 285288.
10 L. A. Hulshof and J. H. Roskam, Eur. Pat. Appl., EP343714 A1
19891129, 1989.
11 C. A. Martinez, S. Hu, Y. Dumond, J. Tao, P. Kellcher and
L. Tully, Org. Process Res. Dev., 2008, 12, 392398.
12 R. J. Kazlauskas and U. T. Bornscheuer, in Comprehensive
Chirality, ed. E. M. Carreira and H. Yamamoto, 2012, vol. 7,
pp. 465480.
13 M. C. de Zoete, A. C. Kock-van Dalen, F. van Rantwijk and
R. A. Sheldon, Chem. Commun., 1993, 1831.
14 F. van Rantwijk, M. A. P. J. Hacking and R. A. Sheldon,
Monatsh. Chem., 2000, 131, 549569 and references therein.
15 G. Hasnaoui-Dijoux, M. Majeric Elenkov, J. H. Lutje Spelberg,
B. Hauer and D. B. Janssen, ChemBioChem, 2008, 9, 10481051.
Chem. Soc. Rev.

Tutorial Review
16 S. K. Ma, J. Gruber, C. Davis, L. Newman, D. Gray, A. Wang,
J. Grate, G. W. Huisman and R. A. Sheldon, Green Chem.,
2010, 12, 8186 and references therein.
17 P. S. Coelho, E. M. Brustad, A. Kannan and F. H. Arnold,
Science, 2013, 339, 307310.
18 H. R. Hobbs, B. Kondor, P. Stephenson, R. A. Sheldon, N. R.
Thomas and M. Poliako, Green Chem., 2006, 6, 816821.
19 R. E. Gumba, S. Saallah, M. Misson, C. M. Ongkudon and
A. Anton, Biofuel Res. J., 2016, 11, 431447.
20 R. Madeira Lau, F. van Rantwijk, K. R. Seddon and
Published on 13 March 2017. Downloaded by University of Newcastle on 13/03/2017 16:02:53.
R. A. Sheldon, Org. Lett., 2000, 2, 41894191.
21 R. A. Sheldon, Chem. Eur. J., 2016, 22, 1298412999. and
references therein.
22 A. P. de los Ros, F. van Rantwijk and R. A. Sheldon, Green
Chem., 2012, 14, 15841588.
23 P. Lozano, T. De Diego, D. Carrie, M. Vaultier and
J. L. Iborra, Chem. Commun., 2002, 692693.
24 E. L. Smith, A. P. Abbott and K. S. Ryder, Chem. Rev., 2014,
114, 1106011082 and references therein.
25 J. T. Gorke, F. Srienc and R. Kazlauskas, Chem. Commun.,
2008, 12351237.
26 C. A. Hutchison, S. Phillips, M. H. Edgell, S. Gillam, P. Jahnke
and M. Smith, J. Biol. Chem., 1978, 253, 65516560.
27 C. Economou, K. Chen and F. H. Arnold, Biotechnol. Bioeng.,
1992, 39, 658662.
28 W. P. Stemmer, Nature, 1994, 370, 389391.
29 M. T. Reetz, A. Zonta, K. Schimossekl, K. Liebeton and
K.-E. Jaeger, Angew. Chem., Int. Ed., 1997, 36, 28302832.
30 Z. Sun, Y. Witmark, J.-E. Backvall and M. T. Reetz,
Chem. Eur. J., 2016, 22, 50465054 and references therein.
31 H. Renata, J. Wang and F. H. Arnold, Angew. Chem., Int. Ed.,
2015, 54, 33513367.
32 U. T. Bornscheuer, G. W. Huisman, R. J. Kazlauskas, S. Lutz,
J. C. Moore and K. Robins, Nature, 2012, 485, 185194.
33 G. W. Huisman, J. Liang and A. Krebber, Curr. Opin.
Chem. Biol., 2010, 14, 122129.
Chem. Soc. Rev.
View Article Online
Chem Soc Rev
34 J. Liang, J. Lalonde, B. Borup, V. Mitchell, E. Mundor,
N. Trinh, D. A. Kochrekar, R. N. Cherat and G. G. Pai, Org.
Process Res. Dev., 2010, 14, 193198.
35 C. K. Saville, J. M. Maney, E. C. Mundor, J. C. Moore,
S. Tam, W. R. Jarvis, J. C. Colbeck, A. Krebber, F. J. Fleitz,
J. Brands, P. N. Devine, G. W. Huisman and G. J. Hughes,
Science, 2010, 329, 305308.
36 M. D. Truppo, H. Strotman and G. Hughes, ChemCatChem,
2012, 4, 10711074.
37 R. A. Sheldon, S. Van Pelt, S. Kanbak-Aksu, J. Rasmussen
and M. H. A. Janssen, Aldrichimica Acta, 2013, 46(3), 8193.
38 R. A. Sheldon, M. J. Sorgedrager and B. Kondor, WO2012/
023847 A2, CLEA Technologies B.V., 2012.
39 A. Bhattacharya and B. L. Pleschke, J. Mol. Catal. B: Enzym.,
2015, 115, 140150.
40 C. Ning, E. Su, Y. Tian and D. Wei, J. Biotechnol., 2014, 184, 710.
41 K. Ahumada, P. Urrutia, A. Illanes and L. Wilson, Food
Bioprod. Process., 2015, 94, 555560.
42 U. V. Sojitra, S. S. Nadar and V. K. Rathod, Food Chem., 2016,
213, 296305.
43 R. Xue and J. M. Woodley, Bioresour. Technol., 2012, 115,
183195.
44 A. Chmura, S. Rustler, M. Paravidino, F. van Rantwijk, A. Stolz
and R. A. Sheldon, Tetrahedron: Asymmetry, 2013, 24, 12251232.
45 M. J. Sorgedrager, D. Verdoes, H. van der Meer and
R. A. Sheldon, Chim. Oggi, 2008, 26, 2325.
46 R. A. Houghten, Proc. Natl. Acad. Sci. U. S. A., 1985, 82,
51315135.
47 H. Cabana, J. P. Jones and S. N. Agathos, Biotechnol. Bioeng.,
2009, 102, 15821592.
48 G. Sheelu, G. Kavitha and N. W. Fadnavis, J. Am.
Oil Chem. Soc., 2008, 85, 739748.
49 H. Mallin, J. Muschiol, E. Bystrom and U. T. Bornscheuer,
ChemCatChem, 2013, 5, 35293532.
50 L. Hilterhaus, O. Thum. and A. Liese, Org. Process Res. Dev.,
2008, 12, 618625.
This journal is The Royal Society of Chemistry 2017