Recent Patents on Biotechnology 2007, 1, 1-9 1
1872-2083/07 $100.00+.00 © 2007 Bentham Science Publishers Ltd.
Engineering Enzymes for Biocatalysis
Paul A. Dalby*
The Advanced Centre for Biochemical Engineering, Department of Biochemical Engineering, University College
London, Torrington Place, London WC1E 7JE, UK
Received: October 1, 2006; Accepted: October 7, 2006; Revised: October 13, 2006
Abstract: Protein engineering techniques have been available for over two decades beginning with the development of
methods for genetic engineering. Since that time, the engineering of enzymes has advanced rapidly along with a
revolution in the range and efficiency of new techniques and strategies for designing and evolving proteins in the
laboratory. While recent advances in high-throughput screening techniques are permitting larger libraries of enzymes to be
screened more rapidly, a combination of new genetic tools and computational methods are enabling the efficient
application of random mutagenesis targeted to areas of enzyme structures that are more likely to elicit the desired
enhancement of their biocatalytic properties. Meanwhile the rational design of enzyme properties, in particular, by
computational design is showing great potential. This review article summarises recent and important patents relating to
the engineering of enzymes for biocatalysis.
Keywords: Rational design, protein engineering, directed evolution, enzyme engineering, biocatalysis, computational design.
INTRODUCTION
Natural enzymes are capable of catalysing reactions with
up to 1017
-fold rate accelerations [1], and with exquisite
control of regio- and stereochemistry. Such control makes
the use of enzymes very attractive as an alternative to
traditional catalysts in the synthesis of complex and highvalue
molecules, especially where chemical routes are difficult
to implement [2,3]. Enzymes have evolved over billions
of years to operate most effectively under physiological
conditions, on a narrow range of natural substrates, and
usually at concentrations in the low mM range. By contrast,
an efficient industrial synthetic process may require a
biocatalytic enzyme to operate on non-natural substrates, and
under conditions in which the enzyme becomes unstable or
inactive, such as extremes of temperature, pH and pressure,
after repeated or prolonged use, or in the presence of organic
solvents which facilitate substrate solubility or product
extraction [4,5]. As a result of these requirements, by far the
majority of naturally occurring enzymes are not suitable for
industrial-scale biocatalytic processes without further
modification of the enzyme itself. Many examples of the
engineering of enzymes for their improved use as biocatalysts
have been published including: increased activity on
novel substrates [6,7]; altered enantioselectivity [8-10];
enhanced stability at extreme temperatures [11,12], pH [13]
or in organic solvents [14,15]; and the preference for alternative
enzyme cofactors [16]. Many more examples have
been reviewed in detail [5,17,18] and cannot all be discussed
here.
Two broad approaches have been taken to engineer the
amino-acid sequences of enzymes, namely rational design
*Address correspondence to this author at The Advanced Centre for
Biochemical Engineering, Department of Biochemical Engineering,
University College London, Torrington Place, London WC1E 7JE, UK; Tel:
+44 020 7679 2962; Fax: +44 020 7383 2348; E-mail: p.dalby@ucl.ac.uk
and directed evolution. Rational design applies the growing
knowledge of enzyme structure and function to make one or
more amino-acid changes that are predicted to elicit the
desired improvements to enzyme function. Such knowledge
is often based upon the bioinformatical analysis of protein
sequences or amino-acid propensities, generalised rules
derived by characterising the effect of mutations upon
enzyme properties, and by implementation of molecular
potential functions that enable the effect of mutations upon
structure to be predicted [19-21].
On the other hand, directed evolution techniques mimic
natural evolution processes such as random mutagenesis [22]
and sexual recombination [23-26]. Just as Nature has
produced its enzymes by the process of evolution without
using any 'knowledge' of enzyme structure and function,
directed evolution similarly permits us to engineer enzymes
without understanding them in great detail. Computational
methods also provide a valuable tool for the evolutionary
design of proteins. Such tools can potentially be used to
obtain useful information from protein sequences and
structures that can guide the random mutagenesis of proteins
making directed evolution more efficient [27]. Alternatively,
the use of molecular potential functions can be used to
predict the effects of mutations on protein structure and
stability for libraries of enzyme variants generated in silico.
The latest developments and applications of these tools will
also be discussed.
Recent advances in both the rational and evolutionary
design of enzymes will be outlined below. However, these
two general approaches are not mutually exclusive and
techniques for improving directed evolution strategies may
include the knowledge-based design of enzyme variant
libraries, as will also be discussed. Overall, this review is
intended to discuss the protein engineering of enzymes for
biocatalysis and associated new patents. While many of the
methods are of general applicability to protein engineering,
2 Recent Patents on Biotechnology 2007, Vol. 1, No. 1 Paul A. Dalby
only those of interest to modifying enzymes for biocatalysis
are included. The final section brings together examples of
the applications of protein engineering to various biocatalytic
enzymes, particularly those appearing as published patents.
RATIONAL DESIGN STRATEGIES
The degree of success of rational design studies depends
largely on the enzyme property to be altered. For example,
the expression level of a protein is readily improved by
ensuring the use of the DNA codons for which the host cell
has the greatest preference [28]. Many successful studies
have applied rational design to the improved conformational
stability of enzymes, through the introduction of disulphide
bonds [29,30], proline residues [31,32], or the mutation of
protein sequences towards the consensus for a given enzyme
family [33]. An alternative approach that uses combinatorial
mutation towards the consensus sequence is discussed later
as a directed evolution strategy [34]. Since these successful
applications, progressively more challenging goals have been
achieved using rational design.
An ability to improve the aqueous solubility of membrane-bound
enzymes is of potential use in the application of
these enzymes as biocatalysts, and also for determining their
three-dimensional structures which would critically enable
further enzyme design work. An automated computational
design strategy has recently been demonstrated to effectively
solubilise two membrane-associated proteins, phospholamban
and KcsA [35]. In the example of phospholamban
(PLB), the authors first predicted the sites for mutagenesis,
in the absence of a PLB structure [36]. They used the experimentally
observed effects of mutations upon pentameric
protein structure formation, to define a perturbation index as
a sinusoidal function of residue position in helices. This
enabled eleven solvent exposed residues to be predicted as
shown in Fig. (1). While one residue, F32, was mutated to a
more chromogenic tyrosine residue to assist experimental
characterisation, the remaining ten were subjected to an
energy minimisation designed to optimise the net free
energy, α-helical propensity, and water solubility of the
protein. Sequence diversity was introduced using a sequence
entropy term in the potential function. In the PLB example,
the new water-soluble variant was found to retain the
pentameric form and degree of helical structure, and also to
be stabilised upon phosphorylation, in the same manner as
the wild-type PLB. More recently, the formation of this
water-soluble variant of PLB has subsequently enabled the
crystal structure of its tetrameric form to be obtained (Fig.
(1)), providing information about the PLB structure that was
previously difficult to obtain [37]. While the techniques
described are limited to the solubilisation of helical
membrane-associated proteins, there is a clear potential for
the application of this approach to the solubilisation of
membrane associated enzymes with highly α-helical content.
Some striking examples of the de novo design of protein
structures have been achieved based on existing knowledge
[38-41], and using computational design algorithms [42-44].
Computational design has been successfully used to
introduce binding functionality into natural protein
structures, such as the design of TNT, L-lactate, D-lactate,
and serotonin binding-sites into Escherichia coli periplasmic
Fig. (1). Water-soluble phospholamban structure.
One half of the tetrameric water-solubilised phospholamban structure
(1YOD.pdb), showing antiparallel helices. Residues identified
as lipid-exposed by mutagenesis studies and a perturbation index
analysis are indicated (green) for one helix only.
binding proteins (PBP) [45,46], and also to redesign the
DNA-sequence specificity of an endonuclease [47].
Furthermore, computational design has been used to add
binding functionality, such as calcineurin-binding [48], to
protein scaffolds that were themselves obtained previously
by rational design.
Enzyme functionality has also been introduced into
protein scaffolds using computational design methods,
demonstrating the potential application of such methods to
biocatalysis. For example, the ability to hydrolyse paranitrophenyl
acetate ester has been engineered into the
catalytically inert thioredoxin protein scaffold [49]. In a
spectacular example, Hellinga and coworkers introduced
triose phosphate isomerase (TIM) activity into the bacterial
ribose-binding protein (RBP), a periplasmic receptor that has
no known catalytic activity, using the computational design
Engineering Enzymes Recent Patents on Biotechnology 2007, Vol. 1, No. 1 3
algorithm described above [45,50]. The nascent enzyme
functionality was then further improved by experimental
directed evolution to achieve a rate enhancement of 3.4x105
over the uncatalysed reaction.
The de novo design of a protein structure that also has
partial enzyme functionality has been demonstrated for a
four-helix bundle with stoichiometric ferridoxase activity,
obtained by introducing a diiron-binding site [51]. However,
the creation of an enzyme functionality that has an efficiency
comparable with natural enzymes, remains elusive using
rational design techniques alone, most likely due to the
greater complexity of the problem, and the reliance on static
protein structure models and transition-state theory for
predicting enzyme structure-function relationships. These
models omit the potentially significant contributions of
enzyme dynamics and quantum tunnelling towards catalysis
[52,53]. Recent efforts to study the function of enzymes
using molecular dynamics simulations and experimental
probes of structural dynamics are expected to lead to
improved rational design methods for enzyme functionality
[54]. While the rational design of a highly efficient catalytic
activity into an inert protein scaffold remains challenging, it
is worth noting that the improvement of an existing enzyme
activity using computational design methods has been
recently demonstrated [55]. In this example, the activity of
chorismate mutase from E. coli was improved by 50%
despite modelling an ab initio calculated transition-state
structure bound into only a static active-site configuration of
the enzyme.
DIRECTED EVOLUTION STRATEGIES
While rational design methods seek to design beneficial
mutations or protein sequences by applying empirically
derived rules or theoretical models, directed evolution uses a
combinatorial approach to create libraries of enzymes from
which enhanced variants can be identified experimentally.
The two most important aspects to consider when devising
directed evolution programmes are: a) the availability of
suitable screening or selection-based methods for identifying
'hits' with improvements of the desired enzyme property; and
b) an appropriate mutagenesis strategy given the extent of
change required in the target property, and the existing
availability of information relating to the target enzyme
structure and function. These two aspects are discussed
below along with recent related innovations that impact on
the application of directed evolution to enzyme biocatalysis.
SCREENING AND SELECTION METHODS
For directed evolution to be successful and efficient, it is
essential that good screening or selection tools are employed
for accurately identifying enzymes with improvements in the
desired properties. Furthermore, the screening or selectionbased
method available for the target activity or enzyme
property critically dictates the enzyme variant library size
that can be practically tested and, therefore, the most
appropriate strategy for library design. In general, the ability
to isolate variants from larger libraries will improve the
chances of success in finding an enzyme with the desired
enhanced properties. Screening methods analyse individual
colonies cultured on agar plates, in microtitre plates, or in
microfluidic chips [56], and are typically used to screen in
the range of 103
-106
individual enzyme variants [57].
However, selection-based methods can use phage display,
cell surface display, ribosomal display, or plasmid display, to
isolate high-performing enzymes from libraries ranging up to
1014
variants. While selection-based methods enable much
greater sequence diversity to be explored, they rely upon
indirect selection for catalytic activity such as binding to
transition-state analogues, whereas screening methods bring
the advantage of direct evaluation of catalytic proficiency
[58]. Screening methods also permit enzyme variants to be
assessed in accurately controlled, and non-physiological
environments [4] which may otherwise prohibit the use of
selection-based techniques.
The gap between library sizes that can be practicably
analysed by screening and by selection-based techniques is
gradually being diminished, for example by using
fluorescence-activated cell-sorting (FACS) of cell surface
displayed enzymes [59], capable of sorting up to 107
enzyme
variants per hour. A recent high-throughput screening
technique has been developed in which water-in-oil-in-water
emulsion droplets are sorted by FACS as outlined in Fig. (2),
to identify the best enzyme variants [58,60]. This technique
takes advantage of in vitro compartmentalization, in which
library encoding DNA and an in vitro transcription/
translation reaction mixture create enzyme variants within an
emulsion droplet [61]. The droplets also contain a fluorogenic
enzyme substrate, such that enzyme variants can then
be assessed and sorted by FACS. The DNA that encodes the
highest performing variants can then also be isolated from
the same droplet and used for sequencing. In an alternative
version of this technique, single bacterial cells expressing
enzyme variants cytoplasmically or displayed on the cell
surface, can be encapsulated in water-in-oil-in-water emulsions
and screened by FACS [62].
LIBRARY CONSTRUCTION
The first methods of directed evolution employed mutagenesis
that was targeted across entire genes by error-prone
PCR [63] or recombination of homologous genes by DNA
shuffling [23,64-66]. While these techniques provided the
groundbreaking benefits of being able to enhance the
properties of enzymes without the need for any knowledge of
the enzyme structure or its relationship to enzyme function,
they still suffer from a number of deficiencies that present a
barrier to their wider application. For example, the gains in
enzyme activity or other desired properties obtained by these
techniques, may not always sufficiently improve an
industrial biocatalytic process such that an investment in
them is seen as worthwhile. On the other hand, performing
directed evolution experiments require a degree of
experience sufficient to overcome the difficulties in creating
large libraries with techniques that depended on inefficient
DNA ligation reactions [67].
Since the early directed evolution experiments a number
of advances have been made that help to overcome these
issues, in particular, the development of mutagenesis techniques
that avoid the need for DNA ligations. One example
was the introduction of mutator strains which enable
mutations to accumulate in plasmids harboured within a
bacterial strain deficient in DNA repair mechanisms [68].
However, such strains were of limited use as the chromo-
4 Recent Patents on Biotechnology 2007, Vol. 1, No. 1 Paul A. Dalby
somal DNA of the host strain was also susceptible to
mutagenesis, resulting in long-term instability in cell
viability. A recent technique significantly reduces this
problem by using a highly error-prone DNA polymerase that
preferentially mutates ColE1 plasmids [69,70]. The errorprone
DNA polymerase was created by introducing three
point mutations based upon homology with Taq polymerase
I. It has been shown to introduce an 80,000-fold greater level
of mutagenesis on average, than wild-type DNA polymerase
I, and targets plasmid genes with a 400-fold preference over
chromosomal gene mutagenesis.
Many developments have also been made that improve
the potential gains obtained by directed evolution strategies.
While the ability to isolate variants from larger libraries will
lead to a greater chance of success in finding an enzyme with
the desired enhanced properties, there may be limitations
imposed by the library size that can be practicably assessed
by the available screening or selection-based methods. It is,
therefore, recognised that an ability to improve the quality of
enzyme libraries would be beneficial, for example, by
promotion of the diversity or uniqueness of clones [71], or
by reducing the number of enzyme residues subjected to
random mutagenesis, without incurring any loss in
evolutionary potential.
Most DNA shuffling methods utilise the homology
between parent sequences to reassemble fragments that have
been generated using either restriction-enzyme digests, or
short bursts of PCR. Consequently it is difficult to recombine
sequences with less than approximately 70% DNA sequence
similarity, thus potentially limiting the range of genetic
diversity, and therefore, enzyme functionality that can be
explored. Recent methods have enabled the non-homologous
random recombination (NRR) of DNA sequences [26,72,73].
While most NRR methods typically achieve only a single
crossover [26,72], a few enable multiple crossovers to occur,
including a technique in which DNA hairpins are added to a
mixture of DNA fragments before reassembly by ligation
with T4 DNA ligase [74,75]. The use of T4 DNA ligase
enables the recombination of DNA fragments independently
of their homology, whereas the addition of the DNA
hairpins, which can only ligate at one end, permits the range
of DNA sequence lengths obtained to be carefully controlled.
Fig. (2). Selection of active enzymes in double emulsion microdroplets by fluorescence-activated cell sorting (FACS).
(1) Water-in-oil-in-water emulsion droplets are passed through a fluorescence-activated cell sorter. (2) Each of these double emulsion
droplets contains a variant gene which is transcribed and translated to a mutant enzyme. Active enzymes convert fluorogenic substrate into
fluorescent product. (3) Laser excitation produces fluorescence for double emulsion droplets containing active enzyme. (4) Droplets are
sorted by FACS into those containing active and inactive enzyme variants.
Engineering Enzymes Recent Patents on Biotechnology 2007, Vol. 1, No. 1 5
In an alternative approach to NRR, a method that mimics
exon shuffling [76] has been described in an example which
recombined two distantly related polymerase genes, to
identify novel DNA polymerases [77,78]. The approach,
called structure-based combinatorial protein engineering
(SCOPE), requires structural knowledge or sequence analysis
of the parent enzymes to design PCR primers that enable
recombination to occur at defined structural boundaries.
Amino-acid mutations introduced by traditional errorprone
PCR tend to be biased towards those that can be
obtained by single-base substitutions, typically accessing
only 5.7 alternative residues on average [79]. Greater
sequence diversity can be obtained by applying saturation
mutagenesis to one or more specific sites using degenerate
oligonucleotides that encode for potentially all twenty
amino-acids. In one approach, every single site in a protein
can be systematically mutagenised independently to identify
the optimal residues at each position [80]. The best
mutations can then be recombined with the aim of combining
the enhancements endowed by each individual mutation.
Another recent method enables sequence saturation
mutagenesis (SeSaM) to occur with less bias than errorprone
PCR, at random positions within the targeted
sequence, yet with only a single 1-3bp mutation appearing in
each variant [81]. The method, outlined schematically in Fig.
(3), creates random-length DNA gene fragments by PCR
doped with α-phosphothioate nucleotides (dNTPαS), and
selective cleavage of the sites of dNTPαS incorporation
using iodine. The 3' ends of each DNA fragment are then
'tailed' with typically 1-3 units of the universal base deoxyinosine
(dITP), before elongation to the full length genes by
PCR with a single-stranded template. A subsequent PCR to
amplify the product allows promiscuous incorporation of
standard bases at only the sites of dITP in the template DNA.
The incorporation of random saturation mutations at
specific multiple sites provides the potential of identifying
pairs of mutations that result in synergistic effects on enzyme
properties. This type of library can be achieved using
oligonucleotide PCR primers that encode for potentially all
twenty amino-acids at each position, in a modification [82]
of the original Quikchange® protocol as supplied by
Stratagene (La Jolla, CA). The original protocol employs Pfu
DNA polymerase to extend mutagenic primers annealed to a
template plasmid [83]. This then creates a non-methylated
nicked plasmid progeny containing mutations defined by the
oligonucleotide primers. Subsequent digestion with the
methylated-DNA specific DpnI restriction enzyme then
removes the methylated parental plasmid DNA that is
obtained from cell lysates, leaving only the mutated plasmid
DNA progeny. The improved method significantly enhances
the ability to introduce mutations at multiple sites by
including a blend of Pfu DNA polymerase with Taq
polymerase, Pfu flap endonuclease (PfuFEN-1), and Taq
DNA ligase in the Quikchange® PCR reaction. Adaptation
of the Quikchange methods are also available for performing
site-directed mutagenesis of genes in plasmids, with
improved transformation directly into Bacillus strains, and
without the need for an intermediate transformation of E.
coli [84].
Fig. (3). Sequence saturation mutagenesis (SeSaM) method.
The SeSaM method is schematically outlined. Step 1: creation of
random length DNA gene fragments by PCR doped with αphosphothioate
nucleotides (dNTPαS), and selective cleavage of
the sites of dNTPαS incorporation using iodine. Step 2: Tailing of
3' ends with the universal base deoxyinosine (dITP) (black squares).
Step 3: Elongation to the full length genes by PCR with a singlestranded
template. Step 4: PCR to amplify the product, allowing
promiscuous incorporation of standard bases (white & gray
squares) at sites that base-pair only with the dITP in the template
DNA.
Mutagenesis at specific sites of a plasmid DNA generally
makes use of PCR for amplification of mutagenised DNA, as
discussed in the examples above. However, the use of
mismatch endonucleases, mixed with a proofreading polymerase,
dNTPs and a ligase, potentially obviates the PCR
step by simply excising and repairing mismatched DNA in
the heteroduplex formed between parental template plasmid
6 Recent Patents on Biotechnology 2007, Vol. 1, No. 1 Paul A. Dalby
and a mutagenic oligonucleotide [85]. This new method
presents an efficient alternative for site-directed mutagenesis,
and also for the recombination of point mutations from
several gene variants.
The functionality represented by the standard twenty
amino-acids may be limiting in terms of engineering novel
enzyme functions. For example, the incorporation of nonnatural
amino-acids into proteins may potentially permit
enzyme chemistries that are not observed in Nature. It is now
possible, through groundbreaking work, to incorporate nonnatural
amino-acids into proteins at specific sites during their
expression from bacterial cells [86,87]. This was achieved by
the directed evolution of a tyrosyl t-RNA synthetase to
enable it to charge the t-RNA which is complementary to the
little used amber codon TAG, with a non-natural amino-acid.
IDENTIFYING TARGET SITES FOR MUTAGENESIS
Alongside the development of techniques for creating
different types of library as discussed above, the methods
available for identifying key residues within an enzyme to
target with mutagenesis have also advanced considerably.
Decreasing the redundancy of enzyme libraries by targeting
mutagenesis has the potential to allow more useful sequencespace
to be searched, and makes more efficient use of the
practical limits often imposed by the throughput of analytical
tools.
Two general approaches are taken. In the first, aligned
sequence information from homologous enzymes can be
used to identify residues that vary in Nature, which may be
linked to various enzyme properties. It has recently been
observed that the consensus protein sequence for a particular
enzyme has greater stability in many cases than the wild-type
enzymes [88]. A recent method for improving enzyme
stability takes advantage of this observation by creating
libraries of enzymes that contain variations only at sites that
deviate from the consensus sequence [89]. Interestingly, a
similar method has also recently been used to invert the
cofactor specificity of a lactate dehydrogenase enzyme from
NAD+
to NADP+
[34].
The second general approach for identifying key residues
uses computational techniques, often though not necessarily,
the same as those applied to rational design. In these
examples, the conformational stabilities of enzyme variants
generated in silico, are estimated. The degree of loss in
conformational stability estimated by such a method has
been shown to correlate well with mutagenic hot-spots that
were determined experimentally, and could therefore
potentially be used for identifying amino acids in a protein
that are tolerant to mutagenesis [90-92].
A similar concept has been used for pre-screening
libraries in silico in a technique dubbed Protein Design
Automation (PDA), which eliminates sequences that are
incompatible with the protein fold of the target enzyme
[93,94]. In this example, the 7x1023
potential variants created
by in silico mutagenesis of nineteen active-site residues of βlactamase,
were reduced down to just 172,800. The
remaining variants were designed and screened experimentally
in a single round of directed evolution to produce
entirely novel enzyme variants with a 1280-fold increase in
cefotaxime resistance.
EXAMPLES OF ENZYME ENGINEERING FOR
BIOCATALYSIS
Many enzymes have been modified by protein engineering
methods and some recently patented examples are
summarised below.
The targeted random mutagenesis of seven amino-acid
residues of human butyrylcholinesterase expressed in
mammalian cells, was recently used to evolve an increased
activity towards cocaine hydrolysis [95]. Such an enzyme
has therapeutic potential for treating severe cases of cocaine
toxicity. Seven residues that line the active-site gorge when
aligned to the available structure of the homologous
acetylcholinesterase, were chosen for mutagenesis. Previous
biochemical data that identified sites important for cocaine
hydrolysis were also used to guide the choice of the seven
target sites. Several enzyme variants were found that had up
to 100-fold greater activity towards cocaine hydrolysis than
the wild-type butyrylcholinesterase.
Using a range of protein engineering techniques,
including error-prone PCR, saturation mutagenesis and DNA
shuffling, mutants of toluene 4-monooxygenase, toluene-oxylene
monooxygenase and toluene-4-monooxygenase from
various Pseudomonads have been obtained with improved
activities towards benzene, toluene, nitrobenzene, nitrophenols,
catechols, o-methoxyphenol, and o-Cresol [96]. Such
biocatalytic reactions can be used in the synthesis of phenol,
catechol, nitrophenols, nitrocatechols, 3-methoxy-catechol,
1,2,3-trihydroxybenzene, methoxyhydroquinone, methylhydroquinone,
4-methylresorcinol, methylhydroquinone, and
pyrogallol.
B12-dependent dehydratases with improved reaction
kinetics have been obtained using error-prone PCR and
oligonucleotide-directed mutagenesis to target the DhaB1
gene, encoding the α-subunit of glycerol dehydratase [97].
Such enzymes are useful in the production of glycerol and
1,3-propanediol.
Nitrilases can be used in biocatalytic processes to convert
nitriles to carboxylic acids. Recently, error-prone PCR has
been used to generate mutants of an Acidovorax facilis
nitrilase having improved activity towards 3-hydroxynitriles
[98]. Isolated enzyme variants were capable of up to 1.9-fold
improved activities when used to convert 3-hydroxynitrile to
3-hydroxycarboxylic acid, 3-hydroxyvaleronitrile to 3-hydroxyvaleric
acid, or 3-hydroxybutyronitrile to 3-hydroxybutyric
acid.
In another example highlighting the general applicability
of error-prone PCR, two carotenoid biosynthesis genes were
simultaneously mutated randomly to improve the production
of astaxanthin from beta-carotene [99]. Targeting of the two
enzymes, CrtO ketolase and CrtZ hydroxylase, simultaneously
resulted in mutations within both genes and a net
increase in the astaxanthin biosynthesis yield of up to 20%
above that obtained with the wild-type enzyme pathway.
Error-prone PCR has also been used to improve the
activity of 5'-xanthylic acid (XMP) aminase, which is
potentially useful for the biocatalytic synthesis of 5'-guanylic
acid (GMP) from XMP in microbial fermentations [100]. Six
improved mutants were found, with between two and six
Engineering Enzymes Recent Patents on Biotechnology 2007, Vol. 1, No. 1 7
accumulated mutations, and the best mutant having 3.8-fold
greater activity resulting from three mutations. Xylanases are
useful enzymes added to animal feed to aid the digestion of
feed grains containing xylan. The growing need to sterilise
the feed grain requires a xylanase enzyme that can retain
activity after heat treatment, as well as being active at
physiological digestive conditions (pH 3.5-5.0, 40°C). A
recent patent [101] has combined an engineered disulphide
bond, previously introduced into a homologous xylanase
enzyme, and the Q162H mutation previously reported for
xylanases [102]. Previously, the disulphide bond was shown
to only marginally increase the xylanase thermostability,
whereas the Q162H mutation had no effect on thermostability.
However, the two mutations combined together
result in a xylanase enzyme with a thermal denaturation midpoint
increase of 14°C, and which also retains 40% of the
original activity after a 30 minute treatment at 70°C, and at
least 30% activity at pH3.5-6, and 40-60°C.
A similar product, 5'-inosinic acid (IMP), is synthesised
in an industrial process that uses mutant acid phosphatases
from E. blattae, for which the phosphomonoesterase activity
has been reduced to less than 40% of the wild-type enzyme
due to the mutations G74D or I153T [103]. Decreased esterase
activity allows the mutant enzymes to synthesise IMP
from inosine.
Another enzyme that can be used as a supplement to
animal feeds is phytase. This enzyme is useful during
digestion for releasing inorganic phosphate from phytate,
which contains over 70% of the phosphate in plant material.
A variant of the E. coli pH 2.5 acid phosphatase has been
shown, rather unexpectedly, to exhibit increased phytase
activity upon digestion of the enzyme with a mixture of
trypsin and pepsin [104].
Alpha-amylase is an enzyme that is commonly used for
the degradation of starch during textile and paper desizing,
and also as a component of household detergents. Several
examples of engineered alpha-amylases have been recently
patented, including variants with improved activity and/or
thermostability [105,106]; solvent stability [107], and
resistance to multimerisation [108].
Polysaccharide lyases have potential uses in
polysaccharide sequencing and degradation. A recent patent
describes rationally designed variants of the polysaccharide
lyase, chondroitinase B, and also potential uses in the
inhibition of anticoagulation, angiogenesis and maternal
malarial infection [109]. Active-site residues involved in
catalysis and substrate binding were targeted by site-directed
mutagenesis. While mutagenesis of the catalytic residues
were damaging to enzyme activity, the mutation of a
substrate binding arginine (R364) to alanine resulted in an
altered product profile to that of wild-type enzyme after the
digestion of dermatan sulfate [110].
CURRENT & FUTURE DEVELOPMENTS
In recent years, significant advances have been made in
the understanding of protein structure and function that have
also enabled rational design techniques to become more
successful and further reaching than previously. In particular,
the use of computational design algorithms with advanced
features has made it possible to introduce nascent enzyme
functionality into proteins, and in at least one example, to
improve the efficiency of an existing enzyme activity.
Advances in rapid screening techniques, new methods for
creating genetic diversity, and computational tools for
predicting hot-spots in protein sequences, have also pushed
forward the boundaries that can be reached using directed
evolution. The challenge now resides in refining computational
methods for more accurate prediction of the effect of
mutations on enzyme activity, potentially including the
effects of protein dynamics. Meanwhile, improved understanding
of natural enzyme evolution and better prediction of
important residues within enzymes will enable directed
evolution techniques to become much more powerful.
ABBREVIATIONS
CrtO = Carotene ketolase gene
CrtZ = Carotene hydroxylase gene
dITP = Deoxyinosine triphosphate
dNTPαS = α-Phosphothioate nucleotides
FACS = Fluorescence-activated cell-sorting
GMP = 5'-Guanylic acid
IMP = 5'-Inosinic acid
KcsA = Potassium ion-channel protein gene
NAD+
= Nicotinamide adenine dinucleotide
NADP = Nicotinamide adenine dinucleotide
phosphate
NRR = Non-homologous random recombination
PCR = Polymerase chain reaction
PDA = Protein design automation
Pfu = Pyrococcus furiosus
PLB = Phospholamban
RBP = Ribose binding protein
SCOPE = Structure-based combinatorial protein
engineering
SeSaM = Sequence saturation mutagenesis
TIM = Triose phosphate isomerase
TNT = Trinitrotoluene
XMP = 5'-Xanthylic acid
ACKNOWLEDGEMENTS
The author has no conflicts of interest that are directly
relevant to the contents of this manuscript.
REFERENCES
[1] Radzicka A, Wolfenden R. A proficient enzyme. Science 1995;
267: 90-93.
[2] Schoemaker HE, Mink D, Wubbolts MG. Dispelling the mythsbiocatalysis
in industrial synthesis. Science 2003; 299: 1694-
1697.
[3] Straathof AJ, Panke S, Schmid A. The production of fine
chemicals by biotransformations. Curr Opin Biotech 2002; 13:
548-556.
[4] Hibbert EG, Baganz F, Hailes HC, et al. Directed evolution of
biocatalytic processes. Biomol Eng 2005; 22: 11-19.
8 Recent Patents on Biotechnology 2007, Vol. 1, No. 1 Paul A. Dalby
[5] Sylvestre J, Chautard H, Cedrone F, Delcourt M. Directed
evolution of biocatalysts. Org Process Res Dev 2006; 10: 562-
571.
[6] Glieder A, Farinas ET, Arnold FH. Laboratory evolution of a
soluble, self-sufficient, highly active alkane hydroxylase. Nat
Biotechnol 2002; 20: 1135-1139.
[7] Yoshikuni Y, Ferrin TE, Keasling JD. Designed divergent
evolution of enzyme function. Nature 2006; 440: 1078-1082.
[8] Reetz MT, Zonta A, Schimossek K, Liebeton K, Jaeger KE.
Creation of enantioselective biocatalysts for organic chemistry by
in vitro evolution. Angew Chem Int Edit 1997; 36: 2830-2832.
[9] May O, Nguyen PT, Arnold FH. Inverting enantioselectivity by
directed evolution of hydantoinase for improved production of Lmethionine.
Nat Biotechnol 2000; 18: 317-320.
[10] Carr R, Alexeeva M, Enright A, Eve TSC, Dawson MJ, Turner
NJ. Directed evolution of an amine oxidase possessing both broad
substrate specificity and high enantioselectivity. Angew Chem Int
Edit 2003; 42: 4807-4810.
[11] Merz A, Yee MC, Szadkowski H, et al. Improving the catalytic
activity of a thermophilic enzyme at low temperatures.
Biochemistry 2000; 39: 880-889.
[12] Miyazaki K, Wintrode PL, Grayling RA, Rubingh DN, Arnold
FH. Directed evolution study of temperature adaptation in a
psychrophilic enzyme. J Mol Biol 2000; 297: 1015-1026.
[13] Cherry JR. Directed evolution of microbial oxidative enzymes.
Current Opin Biotech 2000; 11: 250-254.
[14] Moore JC, Arnold FH. Directed evolution of a para-nitrobenzyl
esterase for aqueous-organic solvents. Nat Biotechnol 1996; 14:
458-467.
[15] Hao JJ, Berry A. A thermostable variant of fructose bisphosphate
aldolase constructed by directed evolution also shows increased
stability in organic solvents. Protein Eng Des Sel 2004; 17: 689-
697.
[16] Joo H, Lin Z, Arnold FH. Laboratory evolution of peroxidemediated
cytochrome P450 hydroxylation. Nature 1999; 399:
670-673.
[17] Dalby PA. Optimising enzyme function by directed evolution.
Curr Opin Struct Biol 2003; 13: 500-505.
[18] Turner NJ. Directed evolution of enzymes for applied
biocatalysis. Trends Biotech 2003; 21: 474-478.
[19] Schymkowitz J, Borg J, Stricher F, Nys R, Rousseau F, Serrano
L. The FoldX web server: an online force field. Nucleic Acids
Res 2005; 33: W382-W388.
[20] Monsellier E, Bedouelle H. Improving the stability of an antibody
variable fragment by a combination of knowledge-based
approaches: validation and mechanisms. J Mol Biol 2006; 362:
580-593.
[21] Poole AM, Ranganathan R. Knowledge-based potentials in
protein design. Curr Opin Struct Biol 2006; 16: 508-513.
[22] Chen K, Arnold FH. Tuning the activity of an enzyme for unusual
environments: sequential random mutagenesis of subtilisin E for
catalysis in dimethylformamide. Proc Natl Acad Sci USA 1993;
90: 5618-5622.
[23] Stemmer WPC. Rapid evolution of a protein in vitro by DNA
shuffling. Nature 1994; 370: 389-391.
[24] Zhao H, Giver L, Shao Z, Affholter JA, Arnold FH. Molecular
evolution by staggered extension process (StEP) in vitro
recombination. Nat Biotechnol 1998; 16: 258-261.
[25] Coco WM, Levinson WE, Crist MJ, et al. DNA shuffling method
for generating highly recombined genes and evolved enzymes.
Nat Biotechnol 2001; 19: 354-359.
[26] Ostermeier M, Shim JH, Benkovic SJ. A combinatorial approach
to hybrid enzymes independent of DNA homology. Nat
Biotechnol 1999; 17: 1205-1209.
[27] Bloom JD, Meyer MM, Meinhold P, Otey CR, MacMillan D,
Arnold FH. Evolving strategies for enzyme engineering. Curr
Opin Struct Biol 2005; 15: 447-452.
[28] Hale RS, Thompson G. Codon optimization of the gene encoding
a domain from human type 1 neurofibromin protein results in a
threefold improvement in expression level in Escherichia coli.
Protein Expres Purif 1998; 12: 185-188.
[29] Shimaoka M, Lu CF, Salas A, Xiao T, Takagi J, Springer TA.
Stabilizing the integrin alpha M inserted domain in alternative
conformations with a range of engineered disulfide bonds. Proc
Natl Acad Sci USA 2002; 99: 16737-16741.
[30] Clarke J, Fersht AR. Engineered Disulfide Bonds As Probes of
the Folding Pathway of Barnase - Increasing the Stability of
Proteins Against the Rate of Denaturation. Biochemistry 1993;
32: 4322-4329.
[31] Suzuki Y. The proline rule - A strategy for protein thermal
stabilization. P Jpn Acad B-Phys 1999; 75: 133-137.
[32] Choi EJ, Mayo SL. Generation and analysis of proline mutants in
protein G. Protein Eng Des Sel 2006; 19: 285-289.
[33] Lehmann M, Wyss M. Engineering proteins for thermostability:
the use of sequence alignments versus rational design and
directed evolution. Current Opin Biotech 2001; 12: 371-375.
[34] Flores H, Ellington AD. A modified consensus approach to
mutagenesis inverts the cofactor specificity of Bacillus
stearothermophilus lactate dehydrogenase. Protein Eng Des Sel
2005; 18: 369-377.
[35] Slovic, A.M., Summa, C.M., Saven, J.G., DeGrado, W.F.:
US2004215400 (2004).
[36] Slovic AM, Summa CM, Lear JD, DeGrado WF. Computational
design of a water-soluble analog of phospholamban. Protein Sci
2003; 12: 337-348.
[37] Slovic AM, Stayrook SE, North B, DeGrado WF. X-ray structure
of a water-soluble analog of the membrane protein
phospholamban: Sequence determinants defining the topology of
tetrameric and pentameric coiled coils. J Mol Biol 2005; 348:
777-787.
[38] DeGrado WF, Wasserman ZR, Lear JD. Protein design, a
minimalist approach. Science 1989; 243: 622-628.
[39] Hecht MH, Richardson JS, Richardson DC, Ogden RC. Denovo
Design, Expression, and Characterization of Felix - A 4-Helix
Bundle Protein of Native-Like Sequence. Science 1990; 249: 884-
891.
[40] Kamtekar S, Schiffer JM, Xiong HY, Babik JM, Hecht MH.
Protein Design by Binary Patterning of Polar and Nonpolar
Amino-Acids. Science 1993; 262: 1680-1685.
[41] Quinn TP, Tweedy NB, Williams RW, Richardson JS,
Richardson DC. Betadoublet - De-Novo Design, Synthesis, and
Characterization of A Beta-Sandwich Protein. Proc Natl Acad Sci
USA 1994; 91: 8747-8751.
[42] Kuhlman B, Dantas G, Ireton GC, Varani G, Stoddard BL, Baker
D. Design of a novel globular protein fold with atomic-level
accuracy. Science 2003; 302: 1364-1368.
[43] Kortemme T, Ramirez-Alvarado M, Serrano L. Design of a 20amino
acid, three-stranded beta-sheet protein. Science 1998; 281:
253-256.
[44] Summa CM, Rosenblatt MM, Hong JK, Lear JD, DeGrado WF.
Computational de novo design, and characterization of an
A(2)B(2) diiron protein. J Mol Biol 2002; 321: 923-938.
*[45] Hellinga, H.W., Looger, L.L., Dwyer, M.A.: US2004229290
(2004).
[46] Looger LL, Dwyer MA, Smith JJ, Hellinga HW. Computational
design of receptor and sensor proteins with novel functions.
Nature 2003; 423: 185-190.
[47] Ashworth J, Havranek JJ, Duarte CM, et al. Computational
redesign of endonuclease DNA binding and cleavage specificity.
Nature 2006; 441: 656-659.
[48] Ghirlanda G, Lear JD, Lombardi A, DeGrado WF. From synthetic
coiled coils to functional proteins: Automated design of a receptor
for the calmodulin-binding of calcineurin. J Mol Biol 1998; 281:
379-391.
[49] Bolon DN, Mayo SL. Enzyme-like proteins by computational
design. P Natl Acad Sci USA 2001; 98: 14274-14279.
[50] Dwyer MA, Looger LL, Hellinga HW. Computational design of a
biologically active enzyme. Science 2004; 304: 1967-1971.
[51] Pasternak A, Kaplan S, Lear JD, DeGrado WF. Proton and metal
ion-dependent assembly of a model diiron protein. Protein Sci
2001; 10: 958-969.
[52] Sutcliffe MJ, Scrutton NS. Enzyme catalysis: over-the-barrier or
through-the-barrier? Trends Biochem Sci 2000; 25: 405-408.
[53] Sutcliffe MJ, Scrutton NS. A new conceptual framework for
enzyme catalysis. Hydrogen tunnelling coupled to enzyme
dynamics in flavoprotein and quinoprotein enzymes. Eur J
Biochem 2002; 269: 3096-3102.
[54] Agarwal PK. Enzymes: An integrated view of structure, dynamics
and function. Microb Cell Fact 2006; 5: 2.
[55] Lassila JK, Keeffe JR, Oelschlaeger P, Mayo SL. Computationally
designed variants of Escherichia coli chorismate mutase
Engineering Enzymes Recent Patents on Biotechnology 2007, Vol. 1, No. 1 9
show altered catalytic activity. Protein Eng Des Sel 2005; 18:
161-163.
[56] Fu AY, Spence C, Scherer A, Arnold FH, Quake SR. A
microfabricated fluorescence-activated cell sorter. Nat Biotechnol
1999; 17: 1109-1111.
[57] Aharoni A, Griffiths AD, Tawfik DS. High-throughput screens
and selections of enzyme-encoding genes. Curr Opin Chem Biol
2005; 9: 210-216.
[58] Mastrobattista E, Taly V, Chanudet E, Treacy P, Kelly BT,
Griffiths AD. High-throughput screening of enzyme libraries: In
vitro evolution of a beta-galactosidase by fluorescence-activated
sorting of double emulsions. Chem Biol 2005; 12: 1291-1300.
[59] Olsen MJ, Stephens D, Griffiths D, Daugherty P, Georgiou G,
Iverson BL. Function-based isolation of novel enzymes from a
large library. Nat Biotechnol 2000; 18: 1071-1074.
*[60] Griffiths, A.D., Weitz, D., Link, D., Ahn, K., Bibette, J.:
US2006078888 (2006).
[61] Tawfik, D.S., Bernath, K., Magdassi, S., Peisajovich, S.G.:
WO06051552 (2006).
[62] Aharoni A, Amitai G, Bernath K, Magdassi S, Tawfik DS. Highthroughput
screening of enzyme libraries: Thiolactonases evolved
by fluorescence-activated sorting of single cells in emulsion
compartments. Chem Biol 2005; 12: 1281-1289.
[63] Zhou YH, Zhang XP, Ebright RH. Random mutagenesis of genesized
DNA-molecules by use of PCR with Taq DNA-polymerase.
Nucleic Acids Res 1991; 19: 6052-6052.
[64] Zhao H, Giver L, Shao Z, Affholter JA, Arnold FH. Molecular
evolution by staggered extension process (StEP) in vitro
recombination [see comments]. Nat Biotechnol 1998; 16: 258-
261.
[65] Stemmer, W.P.: US6995017 (2006).
[66] Short, J.M.: US2004248143 (2004).
[67] Hibbert EG, Dalby PA. Directed evolution strategies for
improved enzymatic performance. Microb Cell Fact 2005; 4: 29.
[68] Greener A, Callahan M, Jerpseth B. An efficient random
mutagenesis technique using an Escherichia coli mutator strain.
Mol Biotechnol 1997; 7: 189-195.
*[69] Camps M, Naukkarinen J, Johnson BP, Loeb LA. Targeted gene
evolution in Escherichia coli using a highly error-prone DNA
polymerase I. P Natl Acad Sci USA 2003; 100: 9727-9732.
[70] Camps, M., Loeb, L.A.: WO03102213 (2003).
[71] Drummond DA, Iverson BL, Georgiou G, Arnold FH. Why higherror-rate
random mutagenesis libraries are enriched in functional
and improved proteins. J Mol Biol 2005; 350: 806-816.
[72] Sieber V, Martinez CA, Arnold FH. Libraries of hybrid proteins
from distantly related sequences. Nat Biotechnol 2001; 19: 456-
460.
[73] Lutz S, Ostermeier M, Moore GL, Maranas CD, Benkovic SJ.
Creating multiple-crossover DNA libraries independent of
sequence identity. P Natl Acad Sci USA 2001; 98: 11248-11253.
[74] Bittker JA, Le BV, Liu DR. Nucleic acid evolution and
minimization by nonhomologous random recombination. Nat
Biotechnol 2002; 20: 1024-1029.
[75] Liu, D.R., Bittker, J.A., Liu, J.M.: US2005260655 (2005).
[76] Kolkman JA, Stemmer WP. Directed evolution of proteins by
exon shuffling. Nat Biotechnol 2001; 19: 423-428.
[77] O'Maille PE, Bakhtina M, Tsai MD. Structure-based
combinatorial protein engineering (SCOPE). J Mol Biol 2002;
321: 677-691.
[78] O'Maille, P.E., Noel, J.P.: WO05118861 (2005).
[79] Miyazaki K, Arnold FH. Exploring nonnatural evolutionary
pathways by saturation mutagenesis: rapid improvement of
protein function. J Mol Evol 1999; 49: 716-720.
[80] Delcourt, M.: WO06027471 (2006).
[81] Schwaneberg, U.: WO05035757 (2005).
[82] Cline, J.M., Hogrefe, H.H.: WO03025118 (2003).
[83] Bauer, J.C., Wright, D.A., Braman, J.C., Geha, R.S.: US5932419
(1999).
[84] Leeflang, C., Van der Kleij, W.A.H.: WO04064744 (2004).
[85] Padgett, H.S., Vaewhongs, A.A., Vojdani, F.S., Smith, M.L.,
Lindbo, J.A., Fitzmaurice, W.P.: WO03066809 (2003).
[86] Wang L, Brock A, Herberich B, Schultz PG. Expanding the
Genetic Code of Escherichia coli. Science 2001; 292: 498-500.
*[87] Schultz, P.G., Wang, L., Andersen, J.C., Chin, J.W., Liu, D.R.,
Magliery, T.J., Meggers, E.L., Mehl, R.A., Pastrnak, M., Santoro,
S.W., Zhang, Z.: US2005250183 (2005).
[88] Lehmann M, Pasamontes L, Lassen SF, Wyss M. The consensus
concept for thermostability engineering of proteins. Biochim
Biophys Acta Protein Struct Mol Enzymol 2000; 1543: 408-415.
[89] Aehle, W., Ramer, S.W., Schellenberger, V.: WO05040344
(2006).
[90] Voigt CA, Mayo SL, Arnold FH, Wang ZG. Computational
method to reduce the search space for directed protein evolution.
Proc Natl Acad Sci USA 2001; 98: 3778-3783.
[91] Voigt CA, Mayo SL, Arnold FH, Wang ZG. Computationally
focusing the directed evolution of proteins. J Cell Biochem Suppl
2001; Suppl 37: 58-63.
[92] Voigt, C.A., Wang, Z.G., Arnold, F.H., Mayo, S.L.:
WO03091835 (2003).
[93] Hayes RJ, Bentzien J, Ary ML, Hwang MY, Jacinto JM,
Vielmetter J, Kundu A, Dahiyat BI. Combining computational
and experimental screening for rapid optimization of protein
properties. P Natl Acad Sci USA 2002; 99: 15926-15931.
*[94] Mayo, S.L., Dahiyat, B.I., Gordon, D.B., Street, A., Su, Y.:
US2006019316 (2006).
[95] Watkins, J.D., Pancook, J.D.: US2003153062 (2003).
[96] Wood, T.K., Vardar, G.: US2006051782 (2006).
[97] Der-ing, L.: WO04056963 (2004).
[98] Payne, M.S., Dicosimo, R., O'Keefe, D.P.: WO06023520 (2006).
[99] Tang, X.-S., Cheng, Q., Shyr, J.Y., Tao, L.: WO06072078 (2006).
[100] Pan, J.G., Jung, H.C., Kim, E.J., Lee, H.S., Park, Y.H., Kim, H.S.,
Han, J.K., Lee, J.N., Oh, K.H., Kim, J.H., Oh, Y.S., Sim, J.I.,
Hong, K.K., Choi, K.O., Kim, H.S., Baek, M.J., Kang, T.S.:
WO06065076 (2006).
[101] Sung, W.L., Tolan, J.S.: US20067060482 (2006).
[102] Sung, W.L., Yaguchi, M., Ishikawa, K.: US5759840 (1998).
[103] Ajinomoto, K.K., Mihara, Y., Utagawa, T., Hideaki, Y., Asano,
Y.: WO9637603 (1996).
[104] Lei, X.G.: US2003072844 (2003).
[105] van der Laan, J.M., Aehle, W.: WO9535382 (1995).
[106] Svendsen, A., Kjaerulff, S., Bisgaard-Frantzen, H., Andersen, C.:
EP1676913 (2006).
[107] Bessler, C., Wieland, S., Maurer, K.-H.: WO06037484 (2006).
[108] Bessler, C., Wieland, S., Maurer, K.-H.: WO06037483 (2006).
[109] Pojasek, K., Raman, R., Sasisekharan, R.: WO03102160 (2003).
[110] Pojasek K, Raman R, Kiley P, Venkataraman G, Sasisekharan R.
Biochemical characterization of the chondroitinase B active site. J
Biol Chem 2002; 277: 31179-31186.