Computer-aided engineering of a transglycosylase for the glucosylation of an unnatural disaccharide of relevance for bacterial antigen synthesis

Article abstract of DOI:10.1021/cs501288r

The exploration of chemo-enzymatic routes to complex carbohydrates has been hampered by the lack of appropriate enzymatic tools having the substrate specificity for new reactions. Here, we used a computer-aided design framework to guide the construction of a small, diversity-controlled library of amino acid sequences of an α-transglucosylase, the sugar binding subsites of which were re-engineered to enable the challenging 1,2-cis-glucosylation of a partially protected β-linked disaccharide allyl (2-deoxy-2-trichloroacetamido-β-d-glucopyranosyl)-(1→2)-α-l-rhamnopyranoside, a potential intermediate in the synthesis of Shigella flexneri cell-surface oligosaccharides. The target disaccharide is not recognized by the parental wild-type enzyme and exhibits a molecular structure very distinct from that of the natural α-(1→4)-linked acceptor. A profound reshaping of the binding pocket had thus to be performed. Following the selection of 23 amino acid positions from the first shell, mutations were sampled using RosettaDesign leading to a subset of 1515 designed sequences, which were further analyzed by determining the amino acid variability among the designed sequences and their conservation in evolutionary-related enzymes. A combinatorial library of 2.7 × 10⁴ variants was finally designed, constructed, and screened. One mutant showing the desired and totally new specificity was successfully identified from this first round of screening. Impressively, this mutant contained seven substitutions in the first shell of the active site leading to a drastic reshaping of the catalytic pocket without significantly perturbing the original specificity for sucrose donor substrate. This work illustrates how computer-aided approaches can undoubtedly offer novel opportunities to design tailored carbohydrate-active enzymes of interest for glycochemistry or synthetic glycobiology.

Full text of DOI:10.1021/cs501288r

Research Article
pubs.acs.org/acscatalysis
Computer-Aided Engineering of a Transglycosylase for the
Glucosylation of an Unnatural Disaccharide of Relevance for Bacterial
Antigen Synthesis
Verges,†
,‡,∥,#
Δ,○,⊥
Emmanuelle Cambon,
,†C,‡,l∥aire Moulis,
†,‡,∥,#
Sophie Barbe,
†,‡,∥,#
Δ,○
Step
Magali Remaud-Simeo
hane Salamone,Δ,○,□
n,*,†,‡,∥
Alizee
Yann Le Guen,
and Isabelle Andre*
́
́
†,‡,∥
Laurence A. Mulard,
́
́
†
‡
∥
Universite
́
de Toulouse; INSA,UPS,INP; LISBP, 135 Avenue de Rangueil, F-31077 Toulouse, France
CNRS, UMR5504, F-31400 Toulouse, France
INRA, UMR792 Ingen
́
ierie des System
̀
́ ́
es Biologiques et des Procedes, F-31400 Toulouse, France
Δ
Institut Pasteur, Unite
́
de Chimie des Biomolec
́
ules, 28 rue du Dr. Roux, F-75724 Paris Cedex 15, France
○
CNRS UMR3523, Institut Pasteur, F-75015 Paris, France
⊥Universite
Paris Descartes Sorbonne Paris Cite, Institut Pasteur, F-75015 Paris, France
S Supporting Information
́
́
*
ABSTRACT: The exploration of chemo-enzymatic routes to complex
carbohydrates has been hampered by the lack of appropriate enzymatic
tools having the substrate specificity for new reactions. Here, we used a
computer-aided design framework to guide the construction of a small,
diversity-controlled library of amino acid sequences of an α-trans-
glucosylase, the sugar binding subsites of which were re-engineered to
enable the challenging 1,2-cis-glucosylation of a partially protected β-
linked disaccharide allyl (2-deoxy-2-trichloroacetamido-β-D-glucopy-
ranosyl)-(1→2)-α-L-rhamnopyranoside, a potential intermediate in the
synthesis of Shigella flexneri cell-surface oligosaccharides. The target
disaccharide is not recognized by the parental wild-type enzyme and
exhibits a molecular structure very distinct from that of the natural α-
(
1→4)-linked acceptor. A profound reshaping of the binding pocket had
thus to be performed. Following the selection of 23 amino acid positions from the first shell, mutations were sampled using
RosettaDesign leading to a subset of 1515 designed sequences, which were further analyzed by determining the amino acid
variability among the designed sequences and their conservation in evolutionary-related enzymes. A combinatorial library of 2.7
4
×
10 variants was finally designed, constructed, and screened. One mutant showing the desired and totally new specificity was
successfully identified from this first round of screening. Impressively, this mutant contained seven substitutions in the first shell
of the active site leading to a drastic reshaping of the catalytic pocket without significantly perturbing the original specificity for
sucrose donor substrate. This work illustrates how computer-aided approaches can undoubtedly offer novel opportunities to
design tailored carbohydrate-active enzymes of interest for glycochemistry or synthetic glycobiology.
KEYWORDS: oligosaccharide, amylosucrase, computer-aided library design, enzyme engineering, chemo-enzymatic synthesis,
Shigella flexneri
INTRODUCTION
including sequence alignments combined with or without
structural data, phylogenetic analyses, and identification of
mutation correlations within protein superfamilies occupy an
increasing place in the rationalization of enzyme evolution to
design focused libraries, guide amino acid substitutions, and
■
Enzyme engineering is a major instrument of modern
biocatalysis and novel enzyme-based process development.
Advances in bioinformatics, biophysics, biochemistry, and
screening instrumentation provided all their contribution to
rationalize or accelerate the optimization, adaptation, and even
1
,2
5
narrow down the size of the library. Also, computational protein
3
6−9
the conception of enzymes. Learning from the natural evolution
design (CPD) is developing quickly,
and despite existing
process has allowed decrypting some of the natural mechanisms
that shape protein fitness landscapes and has served to develop
laboratory methods of evolution and to tailor catalysts for
Received: August 29, 2014
Revised: January 7, 2015
Published: January 9, 2015
4
synthetic purposes. In particular, computer-aided approaches
©
2015 American Chemical Society
1186
DOI: 10.1021/cs501288r
ACS Catal. 2015, 5, 1186−1198

ACS Catalysis
Research Article
Figure 1. Reaction catalyzed by NpAS from sucrose in the presence of natural acceptor (maltose) and targeted reaction by redesigned amylosucrase (in
green) using the non-natural acceptor D′A′. Catalytic mechanism is decomposed in two sequential steps: the first step (enzyme glucosylation step) leads
to the formation of a β-glucosyl enzyme intermediate, whereas the second one (enzyme deglucosylation step) corresponds to the transglucosylation of
the acceptor molecule (natural or non-natural). NpAS: Neisseria polysaccharea amylosucrase; sucrose, α-D-glucopyranosyl-(1,2)-β-D-fructofuranoside;
Fru, Fructose; Maltose, Glucopyranosyl-(1−4)-α-D-glucopyranose; A′, allyl α-L-rhamnopyranoside; D′, 2-deoxy-2-trichloroacetamido-β-D-
glucopyranosyl; E, α-D-Glucopyranosyl
limitations in the treatment of flexibility, sequence-conformation
exploration, solvent consideration, and energy functions, it
provides extremely useful information to create libraries of
situation is nicely exemplified in Shigella flexneri, a family of Gram
negative bacteria. Most of the O-specific polysaccharides (O-SP)
of the different S. flexneri types and subtypes show repeating units
with the same linear tetrasaccharide backbone, and differing
mainly in the site-specific α-D-glucosylation and O-acetylation of
10−17
defined sequences.
An increasing number of examples
demonstrate their potential to draw evolutionary paths leading to
desired catalytic functions with a good efficacy and accuracy.
However, these computer-aided approaches have been scarcely
applied to carbohydrate-active enzymes (CAZymes), although
they could significantly accelerate the exploration of the
combinatorial mutations, especially when the engineering of
multiple carbohydrate binding sites is necessary to confer the
2
1,22
this common tetrasaccharide.
Recently, we demonstrated
that the active site of the α-retaining NpAS could be reshaped to
enable the site-selective α-D-glucosylation of two non-natural
acceptor monosaccharides, namely, methyl α-L-rhamnopyrano-
1
9,20
side and allyl 2-acetamido-2-deoxy-α-D-glucopyranoside.
A
more ambitious design relying on the re-engineering of several
binding subsites of NpAS was undertaken here to render the
enzyme able to glucosylate disaccharide D′A′, featuring a
partially protected D-glucosamine (D′) β-(1,2)-linked to a
partially protected L-rhamnoside (A′) (Figure 1, Figure S1 of
the Supporting Information). The unnatural disaccharide
acceptor (D′A′) was designed so as to open the way to a set of
serotype-specific S. flexneri O-SP motifs after appropriate
enzymatic α-D-glucosylation (Figure S1 of the Supporting
Information). In the following, emphasis was put on the
conversion of the D′A′ acceptor into the ED′A′ trisaccharide,
18
desired substrate specificity to the enzyme.
Although the number of natural CAZymes able to act in a
regio- or stereospecific manner on a vast range of glycosidic
molecules is continuously expanding in gene databases, access to
catalysts that will address a particular need is not a trivial issue. In
particular, the lack of appropriate enzymatic tools with requisite
substrate and bond specificities has hampered extensive
exploration of enzymatic or chemo-enzymatic routes to complex
carbohydrates and the design of biocatalysts, which in
combination with chemistry could help bypassing some
constraints encountered in glycochemistry, such as the need of
protection of highly reactive hydroxyl groups. Such constraints
can be integrated into the (re)design of enzymes, purposely
featuring an α-D-glucopyranosyl residue (E) at O-4 (Figure 1).
D′
This glucosylation pattern is present in the O-SP from S. flexneri
21,22
serotypes 1a, 1b, and 1d.
The CPD approach was first used
“
tailored” to enter the chemo-enzymatic pathway at a specific and
for the generation of a primary sequence library, which was then
rationally reduced in size by integration of additional information
regarding conserved amino acids or correlated pairs of amino-
acids found in homologous enzymes. Reasons for adopting such
a tactic were motivated by the objectives addressed in this work:
(1) conserving the original specificity toward donor substrate,
while (2) creating a novel specificity for a complex acceptor
substrate not recognized by the parental wild-type enzyme; (3)
requiring thus the redesign at multiple binding subsites of a very
constrained active site; and (4) needing of downsizing mutant
libraries due to the availability of only low- and medium-
throughput screening assays.
programmed stage and thus enable thorough exploration of
novel pathways toward defined carbohydrate-based struc-
tures.
1
9,20
This is the challenge addressed in this study where a computer-
aided approach was followed to redesign several sugar binding
subsites of a sucrose-utilizing transglucosylase, the amylosucrase
from Neisseria polysaccharea (NpAS). The objective was to adapt
the enzyme to the transglucosylation of a non-natural
disaccharide, structurally highly divergent from its natural
substrates, with the long-term goal of developing in vitro
chemo-enzymatic pathways to access complex oligosaccharides
featuring α-D-glucopyranosyl side chains. Indeed, α-D-glucosyla-
tion is frequently encountered in natural polysaccharides either
in the backbone or as side-chain substitutions. The latter
Using the framework disclosed herein, a library of defined
amino-acid sequences encoding only for the desired combina-
tions was constructed and screened. An α-transglycosylase,
1
187
DOI: 10.1021/cs501288r
ACS Catal. 2015, 5, 1186−1198

ACS Catalysis
Research Article
Figure 2. View of amylosucrase organization. Constitutive domains of NpAS are colored respectively in purple for the N-domain, in cyan for the A-
domain ((α/β) catalytic barrel), in red for the B-domain (loop 3 of catalytic barrel), in orange for the B′ domain (extension of the loop 7 from catalytic
barrel), and green for the C-domain. (A) Overall view of NpAS in complex with docked ED′A′ showing the 23 selected positions forming the binding
site; (B) cross-sectional view of the active site with ED′A′ occupying subsites −1, +1 and +2; (c) schematic representation of the amylosucrase topology
and location of the 23 selected position.
whose natural specificity is the synthesis of α-(1,4)-linked
compatible with further chemical transformation into building
blocks permitting chain elongation at both the reducing and
nonreducing ends. We reasoned that the use of lightly protected
acceptor substrates should in theory fulfill these criteria without
preventing enzyme remodeling. In the following, disaccharide
D′A′, namely, allyl (2-deoxy-2-trichloroacetamido-β-D-glucopyr-
anosyl)-(1→2)-α-L-rhamnopyranoside (Figure 1), correspond-
ing to a protected version of part of the backbone tetrasaccharide
repeating unit of several S. flexneri surface polysaccharides
23−25
glucopyranosyl polymers from sole sucrose,
was converted
into an enzyme displaying a novel substrate specificity for the α-
D-glucosylation at OH-4D′ of the β-(1,2)-linked disaccharide
D′A′ for which there is no equivalent reaction yet reported in the
literature.
RESULTS AND DISCUSSION
■
Candidate Enzyme and Selected Disaccharide. Adding
to the fact that a wealth of structural information is already
available on NpAS enabling the use of computer-aided
engineering strategies, this enzyme appeared to be a promising
and malleable candidate for the challenge undertaken herein,
which is the conception of a mutant able to regiospecifically
glucosylate, from sucrose donor, a lightly protected disaccharide
acceptor. As seen in Figure 1, the structure of disaccharide D′A′
is very different from that of the natural acceptor (maltose). It is
noteworthy that this molecule is not recognized at all by the wild-
type enzyme, and its accommodation necessitates major redesign
of both +1 and +2 subsites to allow its glucosylation during the
second step of the catalytic reaction (Figure 1). In contrast to
most chemo-enzymatic syntheses of complex oligosaccharides
2
1,22
(
Figure S1 of the Supporting Information),
was designed as
the selected acceptor. Besides being orthogonal to most
protecting groups, the allyl moiety does not bring in major
steric hindrance. Moreover, it was previously shown to be
compatible with NpAS remodeling and successful chain
elongation of the disaccharide resulting from the enzymatic
19
glucosylation step of a monosaccharide acceptor. The remote
effect of acetamides on chemical glycosylation reactions was
30
reported, and the synthesis of oligosaccharides bearing N-
acetylated aminosugars usually involves building blocks whereby
the acetamido function is masked. In this context, we have
successfully used the trichloroacetyl moiety as a N-protecting
group for the aminosugar building blocks involved in the
chemical synthesis of oligosaccharides representative of the O-SP
26−29
described to date,
the strategy under investigation is based
3
1,32
on an early stage enzymatic glucosylation, followed by chemical
chain elongation of the obtained product. Toward this aim, it is
critical that the products of the enzymatic glucosylation step are
from different S. flexneri types and subtypes.
In agreement
with our previous achievements, the 2 acetamide in disaccharide
D
DA is masked as a trichloroacetamide moiety in acceptor D′A′.
1
188
DOI: 10.1021/cs501288r
ACS Catal. 2015, 5, 1186−1198

ACS Catalysis
Research Article
Design and Construction of Directed Sequence
this stage, we decided to not directly generate this library of
mutants. Indeed, in some cases, the predicted structures differed
only by very small scoring differences that might have been
irrelevant given all approximations made in the calculations. On
the other hand, the combinatorial space of this library remains
Libraries. Crystallographic structure of NpAS in complex with
25
the maltoheptaose product was used as a template to manually
force docking of the desired glucosylated product (allyl α-D-
glucopyranosyl-(1→4)-(2-deoxy-2-trichloroacetamido-β-D-glu-
copyranosyl)-(1→2)-α-L-rhamnopyranoside, ED′A′) into the
enzyme active site. Examination of docked ED′A′ showed that all
binding interactions with the glucopyranosyl moiety (E) at the
11
huge, around 5.5 × 10 sequences and still too broad to be
handled experimentally. The reduction of library size was thus
pursued on the basis of a specific sequence analysis consisting in
determining the amino acid variability within the 1515 designed
sequences and their conservation in the evolutionary related
−
1 subsite, which is responsible for enzyme specificity toward
sucrose, were preserved in comparison with crystallographic data
on complexes. After excluding all essential amino acid residues
for either catalysis or specific recognition of sucrose, we selected
altogether 23 mutable positions from the first shell of amino acid
residues of subsites +1, +2, and +3, which were likely by mutation
to favor recognition of D′A′. These 23 positions are located in
structurally distinct regions of the enzyme, mainly on loops 3, 4,
and 7, which surround the catalytic center (Figure 2). If all 23
positions were considered to be mutable by one of the 20
possible amino acid residues, the theoretical space would be as
34
enzymes to subfamily GH13_4 of glycoside-hydrolases, which
share a common specificity toward sucrose substrate donor.
On the basis of the amino acid occurrence observed for each
position in the 1515 designs, mutations that had low occurrence
(≤10%) were discarded. Whenever the presence of wild-type
amino acids was found over or equal to 5%, it was maintained at
the given position in the final designed library. Although at some
positions the wild-type amino acid was not (or only scarcely)
observed in the design results (frequency of occurrence less than
5%), it was kept or reintroduced in the final library at this position
when it was found highly conserved in the GH13_4 subfamily
(Shannon entropy less than 0.2%). The decision to only
reintroduce wild-type amino acids at very specific positions is a
compromise considering that a high occurrence in the subfamily
GH13_4 may be critical for sucrose substrate specificity but that
the preservation of wild-type signature at all the positions would
have led to a too large library. Indeed, if the wild-type amino acid
had been conserved at each designable position, the size of the
2
9
large as 8.4 × 10 , clearly out of reach of available screening
methods. Prior methods focusing on the use of degenerate
oligonucleotides encoding for a limited set of amino acids (such
as NDT or NDK codons encoding for only 12 amino acids) have
33
been proposed to reduce the library size. However, these
methods randomly introduce a reduced set of amino acids
without taking into account the structural context and the
mechanistic constraints. To tackle the customized construction
of a small, semi-random, and diversity-controlled library, we
proposed to adopt an approach enabling to focus only on the
most promising mutations derived from computational
predictions and that were refined by additional integration of
7
5
library would have been as large as ca. 10 compared to 1.2 × 10
using a position-dependent reintroduction of wild-type amino
acids. To further reduce the size of the combinatorial library,
pairwise correlation analysis of predicted mutations was carried
out. On this basis, only significant interposition correlations
23−25,34
structural and biochemical knowledge.
Considering the
2
3 selected positions and using the enzyme design protocol
described in Figure 3, we sampled mutations at each position by
performing 20 000 independent runs of RosettaDesign. The
(
residue-pair frequency ≥ 5%) were kept in order to only
promote beneficial combinations of mutations. Such removal of
correlated pair combinations displaying low residue-pair
frequency allowed narrowing down further the combinatorial
2
0 000 generated sequences were subsequently scored and
filtered using various criteria including predicted favorable
enzyme-ligand binding interaction and complex stability. This
filtering process yielded a subset of 1515 designed sequences. At
4
space size to 2.7 × 10 sequences. The expected outcome of this
approach was to achieve a sharper exploration of the diversity in
the regions of higher interest in comparison to the exorbitant size
of theoretical combinatorial libraries generated if all 20 amino
acid mutations had been allowed at each of the 23 positions
23
(
corresponding in theory to 20 possible sequences).
To do so, diversity of the library was controlled in the library
construction by mixing synthetic oligonucleotides that encoded
for most frequent amino acid mutations and taking into account
favorable inter-residue correlations observed in sequences
predicted by computational design. Considering the 23 selected
positions from the first shell and using a computational approach
involving the different filters above-described (Figure 3), the
2
3
diversity was drastically narrowed down from 20 theoretical
4
possible sequences to 2.7 × 10 . To ensure the exploration of all
the designed sequences, three sublibraries were constructed
(
Table 1).
In the designed libraries, the amino acid sequences were
expected to contain from 9 up to 17 mutations, avoiding any
representation of the parental gene. To construct the libraries,
the gene was first amplified using six couples of nonmutated
primers positioned all along the sequence to generate six DNA
wild-type fragments surrounding the five regions containing
designed mutations (Step 1 in Figure S2 of the Supporting
Information). These fragments were then purified and used as
primers in subsequent PCR reactions in which both wild-type
Figure 3. Computer-aided approach for enzyme and sequence library
design. The size of the theoretical combinatorial sequence space is
indicated at each stage of the strategy.
1
189
DOI: 10.1021/cs501288r
ACS Catal. 2015, 5, 1186−1198

ACS Catalysis
Research Article
Table 1. Summary of the Predicted Mutations and Authorized Combinations of Mutations at the 23 Selected Positions in the
Construction of Sequence Libraries 1−3
and mutated primers were used to amplify longer fragments
incorporating the mutations (Step 2 in Figure S2 of the
Supporting Information) with an equimolar representation of
each desired mutation. The amplicons obtained from step 2 were
mixed two by two (Step 3 in Figure S2 of the Supporting
Information). At step 4, the five amplified long fragments
incorporating the mutations were mixed together in equimolar
amounts. The entire gene was finally amplified.
Table 2. Experimental Screening of Libraries 1, 2, and 3 on
Sucrose and Analysis of Active Clones
library
1
2
3
theoretical number of clones
number of screened clones
13 824
∼23 000
5
6912
∼23 000
24
6912
∼17 000
30
number of active clones on
sucrose using pH-based solid
medium assay
Primary Screening of the Library for Utilization of
Sucrose As Donor Substrate. Screening Results. The
libraries were first screened on sucrose using a colorimetric
number of active clones on
sucrose using the DNS assay
1
24
30
1
9,20,35
pH-based assay developed earlier on solid medium,
which
−3
1.0 × 10−3
1.7 × 10−3
%
of overall hit rate
4.3 × 10
enables the detection of the release of acids upon fructose uptake
and decreases the number of mutants to be subsequently
screened for D′A′ glucosylation. Out of the 63 000 screened
clones, the number of active clones represented only 0.09% of the
overall library, showing the drastic impact of some mutations on
proper catalytic recognition of sucrose or enzyme viability.
Altogether, 59 sucrose-active variants were identified. Of these,
a
average number of mutations in
active clones
13
7.8
9.5
minimum number of mutations in
active clones
-
-
5
5
maximum number of mutations
in active clones
13
17
a
One single mutant was isolated from library 1.
55 variants were confirmed by DiNitroSalicylic acid (DNS)
assays (Table 2).
mutant from library 1 turned out to be more stable than wild-type
NpAS (Tm of 48.6 °C), although overall, these mutants were less
active on sucrose. Conversely, mutants from library 3 had lower
Tm values than wild-type NpAS but were, however, more active
on sucrose. Altogether these results seem to indicate that gain in
mutant stability is often associated with a loss of activity on
sucrose as previously observed for F290K, A289P-F290L,
Analysis of these 55 variants revealed that about 25% and 70%
of active clones from library 2 and 3 are more active on sucrose
than the wild-type enzyme (Figure S3 of the Supporting
Information). The only mutant identified from library 1 is less
active than wild-type NpAS. Overall, 28 clones originating mostly
from library 3 were found with equal or higher activity on sucrose
cleavage (see Figure S3 of the Supporting Information).
To discriminate transglucosylases from sucrose-hydrolases,
the reaction products were analyzed by HPLC. Overall, 25 of the
1
9,20
A289P-F290C, and A289P-F290I mutants of NpAS.
Conformity of Constructed Libraries with Computational
Predictions. The 55 selected clones were sequenced (Table S2 of
the Supporting Information). On average, clones were found to
contain 8.8 mutations per clone, with each clone displaying a
different mutational rate. Among all libraries, library 3 yielded a
higher number of amino acid mutations per clone (Table 2). In
all, 20% and 16.7% of sequences from libraries 2 and 3,
respectively, are totally conformed to the design. The maximum
number of errors in the designed sequences is 5 and they were
introduced most of the time at the same positions (331, 398, 437,
439 or 445). About 73% and 76% of the encoded amino acid
sequences from libraries 2 and 3, respectively, contained at the
5
5 variants (45.5%) synthesized sucrose isomers and malto-
oligosaccharides. Thirteen mutants synthesized products differ-
ing from those produced by parental wild-type NpAS. Seventeen
mutants (30.9%) behaved as pure sucrose-hydrolases. The
thermal stability of active mutants purified in deep-well plates
was assessed by differential scanning fluorimetry (DSF) in the
absence of a bound ligand using the procedure previously
2
0
reported (Figure S4 of the Supporting Information). Overall,
Tm values of the 55 mutants ranged from 43.7 ± 0.2 °C to 51.5 ±
0.1 °C. Notably, most mutants resulting from library 2 and the
1
190
DOI: 10.1021/cs501288r
ACS Catal. 2015, 5, 1186−1198

ACS Catalysis
Research Article
Figure 4. Amino acid composition by position of sucrose-active hits resulting from screening of designed libraries: library 2 (A), library 3 (B).
most three errors compared to the design. The variation at each
position is reported in Figure 4. Amino acid mismatches
correspond to wild-type amino acid incorporations or sub-
stitutions not allowed by the design. Indeed, several sequences
from libraries 2 and 3 contained undesired wild-type amino acids
at positions 228, 290, 300, 331, 437, 439, and 445 with a
particularly higher frequency at positions 437, 439, and 445.
These discrepancies may be attributed to the length of the
oligonucleotide primers that could affect the proper hybridation
of the fragments (case of the region comprised between positions
Scheme 1. Chemical Synthesis of Disaccharide D′A′ (1) from
a
L-Rhamnose and D-Glucosamine Hydrochloride
434 and 448). A higher error rate is also observed when
mutations are located at primer extremities (case of mismatching
found at position 398, library 3) (Figure 4).
Site-Selective α-D-Glucosylation of Disaccharide D′A′
(
1). Disaccharide Synthesis. Multigram amounts of disaccharide
D′A′ (1) were necessary to allow for library screening and active
mutant analysis. Toward this aim, disaccharide 1 was chemically
synthesized in 9 steps starting from commercially available L-
rhamnose and D-glucosamine (Scheme 1). The use of ester and
acetal protecting groups was favored so as to prevent any
interference with the allyl aglycon and trichloroacetamide moiety
at the final deprotection step. Thus, the known allyl rhamno-
3
6
side 3 was protected regioselectively at position O-3 and O-4
by taking advantage of the selectivity of the 1,2-diacetal chemistry
3
7,38
for 1,2-diequatorial diol systems.
The preliminary con-
version of butane-2,3-dione into the more reactive species
a
3
9
Reagents and conditions: (a) see ref 36; (b) 2,2,3,3-Tetramethoxy-
2
,2,3,3-tetramethoxybutane was found advantageous, and
butane, CSA, MeCN, reflux, 30 min (88%); (c) see ref 40; (d)
TMSOTf, DCM, −15 °C → rt, 40 min (83%); (e) 90% aq TFA, rt, 50
min (86%); (f) MeONa, MeOH, rt, 2 h (92%) ; CSA:
Camphorsulfonic acid, TMSOTf: trimethylsilyltrifluoro-
methanesulfonate, DCM: dichloromethane, rt: room temperature;
TFA: trifluoroacetic acid.
alcohol 4 was isolated in a good 88% yield upon reaction of
triol 3 with the latter. TMSOTf-mediated glycosylation of
acceptor 4 with the known trichloroacetimidate glucosaminyl
4
0
donor 5 yielded the crystallized disaccharide 6 in a non-
optimized 83% yield, when run on a multigram scale. Acidic
hydrolysis of the acetal protecting group gave diol 7 (86%), and
subsequent deacetylation of the latter furnished the target allyl
glycoside 1 (92%).
−1
6
70 and P2: 832 g.mol , respectively), corresponding to the
transfer of one or two glucosyl units onto D′A′ (Figure 5).
Characterization of the Mutant Displaying a Novel D′A′
Specificity and Synthesized Products. Sequencing of the
mutant F3 that displays a totally novel substrate specificity
toward D′A′, not detected for the wild-type NpAS, revealed the
presence of 7 mutations (R226L-I228V-F290Y-E300V-V331T-
Secondary Screening of Mutants on D′A′ Glucosylation
and Identification of Hits. All 55 sucrose-active mutants and
wild-type NpAS were tested for their ability to glucosylate
disaccharide D′A′. One mutant (F3) glucosylated D′A′ with a
conversion of 2%. LC-MS analysis of the enzymatic mixtures
−1
indicated a molecular weight increase by 162 g.mol (P1) and
G396S-T398V). This mutant has a T of 45.5 °C instead of 48.6
°C for the wild-type enzyme. Reactions in the presence of sole
m
1
3
24 g.mol− (P2) for the novel UV-detectable compounds (P1:
1
191
DOI: 10.1021/cs501288r
ACS Catal. 2015, 5, 1186−1198

ACS Catalysis
Research Article
Figure 5. Comparison of product profiles obtained with wild-type NpAs (red) and mutant F3 (blue) after 24 h using 146 mM sucrose and 146 mM D′A′.
HPLC analysis was performed (C18 phase column (Venusil AQ 50 mm) with acetonitrile/water (17:83, v/v) as eluent, 30 °C; 1 mL/min) with UV
detection at 205 and 220 nm. The HPLC profile of D′A′ (t = 17.3 min) is shown in green for reference. P1 (t = 15.3 min) and P2 (t = 12.1 min):
r
r
r
−1
glucosylation reaction products of 670 and 832 g·mol , respectively. A zoom of the product P1 and P2 is embedded in the figure.
sucrose showed that the specific activity on sucrose of the mutant
was 1.8-fold higher than that of the wild-type enzyme. As seen in
Figure 6A, the mutant F3 produces a higher amount of turanose
(
6
34.2% versus 25.3% for wild-type), more maltose (14.9% vs
.1% for wild-type) and also shorter oligosaccharides from DP 3
to 7 (29.6% vs 3.1% for wild-type) compared to the wild-type
enzyme. Indeed, no insoluble α-glucans (DP > 25, 20%) were
produced. When D′A′ (146 mM) was added to the reaction
medium, the product profile was modified due to the
glycosylation of D′A′ leading to products P1 and P2,
representing 0.6% and 3.7% of the transferred glucopyranosyl
residues, respectively. Glucosylation of D′A′ was also accom-
panied by an increase of small size maltooligosaccharides and
trehalulose production and a pronounced decrease of turanose
synthesis (twice less than that of the wild-type) (Figure 6B).
Following mass spectrometry analysis, P1 and P2 molecular
structures were determined by NMR analysis of tetrasaccharide
P2. Data showed that first glucosylation of D′A′ occurred at OH-
Figure 6. Yields of reaction products analyzed by HPAEC-PAD (Dionex
Carbo-Pack PA100 column; 30 °C; 1 mL/min; gradient of sodium
acetate (from 6 to 300 mM in 28 min) in 150 mM NaOH) of wild-type
NpAS and mutant F3 after 24 h reaction. (A) In the presence of sole
sucrose (146 mM) or (B) in the presence of both sucrose (146 mM) and
D′A′ (146 mM). G: Glucose; ED′A′: allyl α-D-glucopyranosyl-(1→4)-
4
to provide allyl α-D- glucopyranosyl-(1→4)-2-deoxy-2-
D′
trichloroacetamido-β-D-glucopyranosyl-(1→2)-α-L-rhamnopy-
ranoside, also named ED′A′ (or P1 in HPLC analysis).
Subsequent chain elongation of the product of glucosylation
(
2-deoxy-2-trichloroacetamido-β-D-glucopyranosyl)-(1→2)-α-L-rham-
ED′A′ at OH-6 provided allyl α-D-glucopyranosyl-(1→6)-α-D-
E
nopyranoside; EED′A′: allyl α-D-glucopyranosyl-(1→6)-α-D-glucopyr-
anosyl-(1→4)-(2-deoxy-2-trichloroacetamido-β-D-glucopyranosyl)-
glucopyranosyl-(1→4)-(2-deoxy-2-trichloroacetamido-β-D-glu-
copyranosyl)-(1→2)-α-L-rhamnopyranoside, also named
EED′A′ (or P2 in HPLC analysis).
(1→2)-α-L-rhamnopyranoside.
The production of ED′A′ and EED′A′ using various sucrose/
D′A′ ratio was investigated (Table 3). Increasing amounts of
D′A′ allows to produce the highest amount of glycosylated
acceptors (2.2 mM of ED′A′, 8.9 mM of EED′A′ at sucrose/
acceptor ratio 1/8) in 24 h. The best compromise between the
1
192
DOI: 10.1021/cs501288r
ACS Catal. 2015, 5, 1186−1198

ACS Catalysis
Research Article
Table 3. Yields of D′A′ Conversion and Quantification of Mono-Glucosylated (P1) and Di-Glucosylated (P2) Products by the
Mutant F3 Using Different Donor/Acceptor Molar Ratios after 24 h of Reaction
sucrose/D′A′ ratio
sucrose/D′A′ ratio ([suc] = 73 mM)
sucrose/D′A′ ratio ([suc] = 146 mM)
([D′A′] = 73 mM)
1
/1
1/2
1/4
1/8
1/1
2/1
4/1
8/1
a
D′A′ conversion (%)
2
2.17
0.5
1.7
1.1
3.7
1.9
2.2
8.9
3.0
0.6
3.0
0.3
1.9
4.3
0.3
2.8
5
b
monoglucosylated D′A′ (ED′A′ or P1) , mM
diglucosylated D′A′(EED′A′ or P2) , mM
0.3
1.2
0.3
3.3
c
2.5
3.7
a
b
Conversion of D′A′ was calculated as follows: [P1] mM + [P2] mM/[D′A′] mM *100 Monoglucosylated D′A′ (ED′A′, P1) determined from
HPLC analysis (C18 phase column (Venusil AQ 50 mm) with acetonitrile/water (17:83, v/v)) using calibration curves established with D′A′
c
Diglucosylated D′A′ (EED′A′, P2) determined from HPLC analysis (C18 phase column (Venusil AQ 50 mm) with acetonitrile/water (17:83, v/
v)) using calibration curves established with D′A′
conversion rate and yield of acceptor reaction products is
obtained for the sucrose/acceptor ratio of 1/1, starting with 146
mM sucrose. Regardless of the sucrose/acceptor ratio used, the
production level of the monoglucosylated acceptor (up to 2.2
mM) remained always lower than that of EED′A′, suggesting
that ED′A′ is a better acceptor than disaccharide D′A′.
to be mainly due to the increase of K , suggesting that the
presence of disaccharide D′A′ affects sucrose binding.
m
Kinetic parameters were also determined in the presence of
sucrose (200 mM) and D′A′ acceptor (concentrations varying
from 5 to 200 mM). Due to the limited solubility of D′A′, the
enzyme could never be saturated, indicating a very low affinity of
D′A′ for mutant F3 leading to a poor catalytic efficiency value.
Although the catalytic efficiency of the mutant is low, we can
expect that additional rounds of directed evolution could help to
further improve it. Nevertheless, the design, based on molecular
modeling predictions, successfully conferred the novel and
desired acceptor specificity without substantially affecting the
original activity toward sucrose.
The steady-state kinetic parameters (k , K , k /K of
cat
m
cat
m)
transglucosylation reactions of both the wild-type NpAS and the
mutant F3 were first determined upon varying sucrose
concentration from 5 to 600 mM (Table 4). The mutant F3
Table 4. Apparent Kinetic Parameters Determination of
Mutant F3 for Sucrose Donor only (from 5 to 600 mM) and in
a
the Presence of Both Sucrose and D′A′ (See Foonote )
Structural Insight on Novel Substrate Specificity. Our
efforts to crystallize the mutant F3 were unsuccessful. Therefore,
sucrose only (from 5 to 600 mM)
NpAS
mutant F3
a three-dimensional model of the mutant was first built using the
b
Km apparent (mM)
50.2
1.3
21.58
0.82
23,25,41
X-ray structure of the parental NpAS as template.
Both
b
−1
kcat (s
)
sucrose and trisaccharide ED′A′, the reaction product of interest,
were subsequently docked into the active site. Inspection of the
models showed that nearly all seven mutations were carried by
k /K (s−1 mM−1
b
)
0.026
0.040
cat
m
sucrose (from 5 to 600 mM) + D′A′ (200 mM)
c
e
Km, suc (mM)
_
71
loops 3 and 4 which have been shown to be highly flexible by
c
−1
e
kcat, suc (s
)
_
1.75
0.024
19,20
molecular dynamics simulations
and assumed to play a
c
(s−1 mM−1
e
k /K
cat
)
_
m, suc
determinant role in the catalytic mechanism of the enzyme. As a
result of the introduction of these mutations, the active site of the
enzyme is widened up at subsites +1 and +2 enabling the
accommodation of the ED′A′, and in particular its N-
trichloroacetyl and allyl substituents, while the productive
binding of sucrose is not disturbed, as confirmed by experiment
D′A′ (from 5 to 200 mM) + sucrose (200 mM)
d
−1
−1
e
3.5 × 10−4
k /K
(s mM
)
_
cat
m, D′A′
a
Sucrose concentration at 200 mM and various D′A′ concentrations
from 5 to 200 mM or 200 mM D′A′ and sucrose from 5 to 600 mM.
b
Catalytic parameters (K , k , k /K ) were determined in the
m
cat
cat
m
c
presence of sucrose only from 5 to 600 mM. Catalytic parameters
(
Figure 7). Indeed, detailed analysis of interactions established
(
K , k , k /Km, suc) were determined by measuring the initial rate of
m
cat cat
fructose production in the presence of constant acceptor concentration
between the mutant F3 and ED′A′ revealed that mutation of
Glu300 and Arg226, which are involved in a salt bridge
interaction in the parental wild-type enzyme, are mutated by
smaller aliphatic residues, valine and leucine, respectively, likely
enhancing the plasticity of the active site and favoring the
productive binding of D′A′. Mutation of Thr398 by a valine
residue enables the nesting of the allyl aglycon in a small
depression, while mutation of Gly396 by a serine residue brings
an additional hydrogen bonding interaction which stabilizes
further the allyl group. The mutation V331T also enables to
introduce a stabilizing hydrogen bonding interaction with the 2D′
N-trichloroacetyl group. Introduction of the Tyr residue at
position 290 requires the creation of enough space, provided by
the I228V mutation, to accommodate the extra hydroxyl group
compared to the Phe residue present in wild-type enzyme.
Altogether these seven mutations enabled a subtle interplay
between the structure, the flexibility, the stabilization and the
molecular recognition, which cannot yet be predicted altogether
by computational methods. As a result, the combination of these
mutations enabled the reshaping of the enzyme active site that
d
(
200 mM) and sucrose from 5 to 600 mM. Catalytic efficiency k /
cat
Km, D′A′ was determined by measuring the initial rate of ED′A′
formation from 200 mM sucrose (fixed concentration) and 5 to 200
mM D′A′. As D′A′ was not soluble at concentrations higher than 200
mM, no saturation could be observed and only k /K
determined from the slope of the linear regression. Not determined
was
cat
m, D′A′
e
due to D′A′ non recognition.
showed a standard saturation kinetic behavior like the wild-type
enzyme. The Km value was found diminished by 2.3 fold in
comparison to the NpAS value, indicating a higher affinity of the
mutant for sucrose. Conversely, the kcat value of the mutant F3 is
decreased by 1.6 fold compared to parental NpAS. The catalytic
efficiency of the enzymes (k /K values) was determined by
cat
m
linear regression of the transglucosylation initial rates of
measured as variable substrate concentration (sucrose donor or
acceptor) (Table 3). In the presence of 200 mM D′A′ acceptor
(
considered as a saturating concentration), the catalytic
efficiency for sucrose consumption decreased and this appeared
1
193
DOI: 10.1021/cs501288r
ACS Catal. 2015, 5, 1186−1198

ACS Catalysis
Research Article
improvements are still necessary, these new findings open the
way to alternatives to ongoing developments toward chemically
synthesized oligosaccharides representing fragments from the O-
5
1,52
SP of various S. flexneri type 1 strains.
However, the low hit
rate emphasizes the extreme complexity of redesigning
amylosucrase, given its double-step mechanism that involves
the sequential binding of a donor (sucrose) and an acceptor
substrate along the catalytic reaction, and the constrained active
site topology which needs to be adapted to D′A′ disaccharide
which presents considerable structural differences with natural
acceptors.
Indeed, the novel substrate specificity of the mutant was
obtained upon introduction of 7 combined amino acid mutations
in the first shell of the active site, enabling a drastic reshaping of
the catalytic binding site to accommodate non-natural acceptor
D′A′ in a productive manner without perturbing too much its
original specificity for donor sucrose. Not only these mutations,
mainly introduced in flexible loops, enabled a change in substrate
specificity but they also enhanced the stability of the enzyme in
comparison to the parental wild-type NpAS. The catalytic
efficiency of the mutant for trisaccharide production was still low
compared to that observed for the natural donor sucrose. This is
due to the difficulty of this design and can also be attributed to
the limits of CPD approaches and experimental difficulties
encountered in the construction of the library. Nonetheless, the
combined approach used herein was efficient to identify in only
one run of screening the evolutionary path toward a totally new
and unnatural catalytic activity. Obviously, this opens now the
route to further enzyme improvement via either computational
methods, directed evolution or mixed strategies.
Figure 7. Molecular docking of ED′A′ in the active site of a built 3D
model of the mutant F3 displaying novel acceptor specificity toward
D′A′. The 7 mutations (R226L-I228V-F290Y-E300V-V331T-G396S-
T398V) are highlighted in cyan color (hydrogen atoms have been
omitted for clarity purpose). In gray lines are shown the remaining
positions targeted by the computational design but not mutated in the
mutant F3.
was beneficial for the glucosylation of D′A′ while preserving
sucrose recognition. Interestingly, the diglucosylated molecule,
EED′A′, was the main glucosylation product. NMR analysis of
this molecule revealed the presence of an unusual α-(1,6) linkage
that might confer higher conformational flexibility to enable
productive binding of ED′A′ in subsites +1, +2, and+3. This
could also explain to some extent why ED′A′ appears to be a
better acceptor than D′A′. Of note, similar observations were
42
THEORETICAL AND EXPERIMENTAL PROCEDURES
found in our recent work on the glucosylation of flavonoids.
■
These results also suggest a high plasticity of subsite +3 enabling
Computational Modeling and Design. Detailed proce-
dures are provided in Supporting Information and Methods.
Three-dimensional model of NpAS in complex with the desired
glucosylated product (ED′A′) bound into the enzyme active site
was constructed starting from the crystallographic complex with
accommodation of A′. Recent work on the engineering of the
4
3
active site also provides new directions to explore in order to
better control the elongation of gluco-oligosaccharides and in
particular, to avoid the formation of unwanted EED′A′.
25
maltoheptaose, a reaction product (PDB ID: 1MW0). ED′A′
was manually docked into the active site of the amylosucrase by
superimposing the glucopyranosyl unit (E) of ED′A′ onto the
glucopyranosyl unit of the crystallographic maltoheptaose
CONCLUSIONS
■
Computer-aided design was for the first time applied to the
engineering of several subsites of an α-transglycosylase. The
difficulty of this work was to target a limited number of mutations
allowing to keep a donor specificity for sucrose and tailor the
acceptor specificity to render NpAS able to recognize a β-linked
and partially protected disaccharide. Interestingly, NpAS has only
been described up to now as being able to recognize α-linked
disaccharides. The engineering approach included the use of
RosettaDesign software, the selection of amino-acid sequences
favoring inter-residue correlations predicted by computational
design, and the experimental control of the desired combinations
of mutations. The medium size library was easily screened to
yield, after only a single screening run, one mutant displaying the
expected substrate specificity, thus showing the efficacy of the
approach combining both computer-aided design and con-
4
4
located at subsite −1. The molecular all-atom ff 99SB,
45
46
Glycam06 and gaff force fields were used for the proteins,
carbohydrates, and other nonstandard groups, respectively. The
molecular system was then subjected to several cycles of
minimizations applying different levels of restraints using the
Sander module of the Amber 9 software package. Visual
inspection of the modeled complex, as well as the crystallo-
graphic complex of the amylosucrase with the donor substrate,
the sucrose (PDB: 1JGI), was carried out using PyMOL software
47
̈
(Schrodinger, Portland, OR, USA). Based on the analysis of
these minimized complexes and our knowledge on amino acid
residues involved in catalysis and/or specific recognition of
sucrose, 23 positions (226−229, 289−300, 331, 332, 396−398,
434−439, 444−448) of the active site were selected to be
sampled for mutations. Version 3.2 of the Rosetta software suite
(Rosetta Commons, UW techTransfer, Seattle, WA, U.S.A.) was
used to redesign the NpAS active site in order to enhance the
interaction with ED′A′. A protocol was used that tests mutations
and side-chain conformations at the 23 selected positions and
samples variations in the translation and rotation of the docked
ED′A′. All amino acid types, except Cysteine, were authorized at
the 23 designable positions. Sixty residues surrounding the 23
struction of sequence libraries. Indeed, out of a library of 2.7 ×
4
1
0 variants, the mutant F3 regioselectively glucosylates D′A′,
what was not detected for the parental wild-type enzyme. A novel
transglucosylase, whose substrate specificity was purposely
tailored toward a non-natural acceptor, allyl 2-deoxy-2-
trichloroacetamido-β-D-glucopyranosyl-(1→2)-α-L-rhamnopyr-
anoside (D′A′), to yield a potential intermediate in the chemo-
enzymatic synthesis of oligosaccharides representative of S.
flexneri serotypes 1a, 1b, or 1d O-SPs. Obviously, although
1
194
DOI: 10.1021/cs501288r
ACS Catal. 2015, 5, 1186−1198

ACS Catalysis
Research Article
2
4
designable residues were allowed to relax toward alternate side-
chain rotamers. Enzyme backbone atoms, nondesigned or
nonrepacked side-chains as well as the internal conformation
of ED′A′ were held fixed. The initial ED′A′ conformation from
which perturbations were made was taken from the minimized
model of the ED′A′-enzyme complex. The design protocol based
on the use of RosettaDesign and described in details in the
Supporting Information and Methods was run 20 000
independent times. The output sequences were filtered based
on different Rosetta scoring and filtering schemes. The 1515
unique sequences resulting from this filtering were used to guide
library design. The library design approach developed was based
on the following: (1) the determination from a multiple
sequence alignment of 1515 designed sequences of the
position-dependent amino acid residue frequency, a Correlated
Mutation score at position pairs and the residue-pair frequency;
had R = 0.49 (cHex/EtOAc 3:2); [α]
= −229 ° (c 1.0,
f
D
1
CHCl ); H NMR (CDCl ): δ 5.93−5.83 (m, 1H, C HCH ),
3
3
2
5.29−5.23 (m, J
= 17.2 Hz, 1H, CHCH H ), 5.19−5.16
trans
a
b
(m, J = 10.4 Hz, 1H, CHCH H ), 4.81 (d, J = 1.2 Hz, 1H,
cis
a b 1,2
H-1), 4.19−4.13 (m, 1H, -OCH H ), 4.00−3.95 (m, 2H,
a
bAll
-OCH H , H-3), 3.92 (dd, J = 3.2 Hz, 1H, H-2), 3.82 (dq, J
a
bAll
2,3
4,5
= 9.8 Hz, 1H, H-5), 3.70 (pt, J3,4 = 9.9 Hz, 1H, H-4), 3.26 (s, 3H,
-OCH ), 3.23 (s, 3H, -OCH ), 2.42 (bs, 1H, OH), 1.31 (s, 3H,
CH ), 1.28 (s, 3H, CH ), 1.26 (d, J = 6.2 Hz, 3H, H-6).
3
3
1
3
C
3
3
5,6
NMR (CDCl ): δ 134.0 (CHCH ), 117.6 (CHCH ), 100.3
3
2
2
(C ), 100.0 (C ), 99.0 (C-1), 70.0 (C-2), 68.5 (C-4), 68.4
quat.
quat.
(C-3), 68.1 (-OCH2All), 66.7 (C-5), 48.2 (-OCH ), 47.8
(-OCH ), 18.0 (CH ), 17.8 (CH ), 17.7 (C-6). HR-ESI-TOF-
MS m/z 341.1542 (Calcd for C H O Na, 341.1576).
Allyl (3,4,6-tri-O-acetyl-2-deoxy-2-trichloroacetamido-β-D-
glucopyranosyl)-(1→2)-3,4-O-(2′,3′-dimethoxybutane-2′,3′-
diyl)-α-L-rhamnopyranoside (6). TMSOTf (200 μL, 1.1 mmol,
0.1 equiv) was added to a solution of acceptor 4 (3.66 g, 11.5
mmol) and of the known donor 5 (7.52 g, 12.6 mmol, 1.1
equiv) in anhydr. DCM (180 mL) containing 4Å MS (10 g),
3
3
3
3
15
26
7
(
2) the calculation of the Shannon entropy using in house
software, called Sequester, to measure the position-dependent
amino acid residue variability in multiple sequence alignment of
GH13_4 subfamily sequences.
40
From these computations, the amino acid types and the
combinations of amino acid residues considered in the final
designed libraries were selected as follows: (1) at each enzyme
designed position, the wild-type amino acid residues with a
frequency of occurrence ≥ 5% were maintained; (2) the wild-
type amino acid residues with a frequency of occurrence < 5% or
absent in the design results and with a Shannon entropy < 0.2%
were introduced; (3) the other amino acid types were kept if only
their position-dependent frequency was higher than 10%; (4) for
correlated amino acid pairs, only the combination-pairs with a
frequency of occurrence ≥ 5% were preserved, while all
combinations were considered for noncorrelated amino-acid
residues.
The mutant F3 of NpAS was constructed starting from the
model of the wild-type enzyme docked with ED′A′ in which
mutations were introduced using the Biopolymer module of the
InsightII software package (Accelrys, San Diego, CA, U.S.A.).
The model was subsequently minimized following the same
energy minimization procedure described above.
stirred at −15 °C under an Ar atmosphere. After 30 min, Et N
3
was added, and the suspension was filtered over a pad of Celite.
The filtrate was concentrated to dryness. Crystallization of the
crude material from EtOAc/pentane gave disaccharide 6 (7.17 g,
9.54 mmol) as white crystals in a non optimized 83% yield. The
fully protected compound 6 had R = 0.31 (cHex/EtOAc 3:2);
f
2
4
[α] = −92.0° (c 1.0, CHCl ); mp = 186 °C (Et O/pentane);
D
3
2
1
H NMR (CDCl ): δ 6.83 (d, J
= 7.2 Hz, 1H, NH), 5.91−
3
2,NH
5.82 (m, 1H, CHCH ), 5.31−5.22 (m, J
= 17.2 Hz, 2H,
2
trans
CHCH , H-3 ), 5.17−5.13 (m, J = 10.4 Hz, 1H, CH
2
D′
cis
1
CH ), 5.08 (pt, J = 9.5 Hz, 1H, H-4 ), 4.98 (d, J = 8.2 Hz, J
2
D′
1,2
CH
= 164.4 Hz, 1H, H-1 ), 4.86 (d, J = 1.2 Hz, 1H, H-1 ), 4.24
D′
1,2
A′
(dd, J6a,6b = 12.1 Hz, J5,6a = 5.1 Hz, 1H, H-6a ), 4.18−4.10 (m,
D′
2H, H-6b , H ), 4.08−4.00 (m, 2H, H-2 , H-3 ), 3.98−3.92
D′
All
D′
A′
(m, 2H, H , H-2 ), 3.76−3.67 (m, 2H, H-5 , H-5 ), 3.56 (pt, J
All
A′
A′
D′
= 9.9 Hz, 1H, H-4 ), 3.23 (s, 3H, -OCH ), 3.17 (s, 3H, -OCH ),
A′
3
3
2.08 (s, 3H, CH3Ac), 2.02 (s, 3H, CH3Ac), 2.01 (s, 3H, CH3Ac),
1.25 (s, 3H, CH ), 1.24 (s, 3H, CH ), 1.21 (d, J = 6.2 Hz, 3H,
3
3
5,6
13
H-6 ); C NMR (CDCl ): δ 170.8, 170.7, 169.4 (3C, CO ),
A′
3
Ac
Chemical Synthesis of Disaccharide 1 (D′A′). General
conditions and NMR spectra of all compounds are provided in
Supporting Information.
161.9 (NHCO), 134.0 (CHCH ), 117.3 (CHCH ), 100.5
(C ), 100.2 (C ), 99.7 (C-1 , J = 164.1 Hz), 99.0 (C-
quat. quat. D′ C,H
2
2
1
1
1 , J = 173.8 Hz), 92.4 (CCl ), 76.1 (C-2 ), 72.2 (2C, C-3 ,
A′ C,H
3
A′
D′
Allyl 3,4-O-(2′,3′-dimethoxybutane-2′,3′-diyl)-α-L-rham-
nopyranoside (4). To a solution of butane-2,3-dione (6.0 mL,
C-5 ), 68.7 (2C, C-4 , C-4 ), 68.1 (2C, C-3 , C ), 66.8 (C-
D′ A′ D′ A′ All
5 ), 62.2 (C-6 ), 56.0 (C-2 ), 48.0 (-OCH ), 47.8 (-OCH ),
A′
D′
D′
3
3
68.4 mmol) in MeOH (25 mL) were added trimethylortho-
20.8 (CH3Ac), 20.7 (2C, CH3Ac), 18.0 (2C, CH ), 16.7 (C-6 ).
3 A′
formate (22.4 mL, 204 mmol, 3.0 equiv) and three drops of
concentrated sulfuric acid. The solution was refluxed for 20 h.
After complete conversion of the starting dione, as attested by
HR-ESI-TOF-MS m/z 772.1539 (Calcd for C H Cl NO Na,
29 42 3 15
772.1518).
Allyl (3,4,6-tri-O-acetyl-2-deoxy-2-trichloroacetamido-β-D-
glucopyranosyl)-(1→2)-α-L-rhamnopyranoside (7). Disac-
charide 6 (1.94 g, 2.58 mmol) was dissolved in 90% aq. TFA
(50 mL), and the solution was stirred for 50 min at rt, then
repeatedly coevaporated with toluene. The residue was purified
by column chromatography (cHex/EtOAc 50:50 → 40:60) to
give diol 7 (1.41 g, 2.21 mmol, 86%) as a white solid.
NMR analysis, solid NaHCO (5.2 g) was added, and the
3
suspension was concentrated to dryness under reduced pressure.
The residue was dissolved in Et O (150 mL), washed with satd
2
aq. NaHCO (50 mL) and brine (50 mL), then dried over
3
sodium sulfate. Volatiles were removed and the resulting material
was distilled under reduce pressure to afford 2,2,3,3-tetrame-
thoxybutane as a colorless oil (bp 35−37 °C, 4 mbar).
2
4
Disaccharide 7 had R = 0.38 (cHex/EtOAc 2:3); [α]
=
f
D
1
Distilled 2,2,3,3-tetramethoxybutane (4.15 g, 23.3 mmol, 1.2
equiv) and CSA (209 mg, 0.90 mmol, 0.05 equiv.) in anhydr.
MeCN (3 mL) were added to a solution of the known
−24.4° (c 1.0, CHCl ); H NMR (CDCl ): δ 7.40 (d, J
= 8.4
3
3
2,NH
Hz, 1H, NH), 5.92−5.83 (m, 1H, CHCH ), 5.29−5.24 (m,
2
J
= 17.2 Hz, 2H, CHCH , H-3 ), 5.19−5.16 (m, J = 10.4
trans
2
D′
cis
3
6
rhamnoside 3 (3.93 g, 19.2 mmol) in anhydr. MeCN (90 mL).
Hz, 1H, CHCH ), 5.09 (pt, J = 9.8 Hz, 1H, H-4 ), 4.94 (d, J
= 8.4 Hz, 1H, H-1 ), 4.89 (d, J = 1.3 Hz, 1H, H-1 ), 4.24 (dd,
D′ 1,2 A′
2 D′ 1,2
The mixture was refluxed for 30 min. Et N was added and
3
solvents were removed under reduced pressure. The crude oil
was purified by column chromatography (Tol/EtOAc, 80:20 →
J6a,6b = 12.4 Hz, J5,6a = 5.0 Hz, 1H, H-6a ), 4.17−4.06 (m, 3H, H-
D′
6b , H , H-2 ), 3.97−3.93 (m, 2H, H , H-2 ), 3.85 (dd, J =
D′ D′ A′ 3,4
All
All
60:40) to give alcohol 4 (5.37 g, 88%) as a colorless oil. Alcohol 4
9.5 Hz, J2 = 3.3 Hz, 1H, H-3 ), 3.76−3.72 (m, 1H, H-5 ),
,3
A′
D′
1
195
DOI: 10.1021/cs501288r
ACS Catal. 2015, 5, 1186−1198

ACS Catalysis
.56−3.36 (dq, J4 = 9.4 Hz, 1H, H-5 ), 3.41 (pt, J = 9.5 Hz, 1H,
Research Article
3
fourth step, 10 μL of each mixture coming from the third step
were mixed together to perform 1 cycle at 98 °C for 30 s and 10
cycles at 98 °C for 10 s and 72 °C for 1 min. No Phusion
polymerase was added in steps 3 and 4. In a fifth step, the gene
library obtained was finally amplified by adding outer primers
(see Supporting Information).
,5
A′
H-4 ), 2.11 (s, 3H, CH ), 2.06 (s, 3H, CH3Ac), 2.05 (s, 3H,
A′
3Ac
1
3
CH3Ac), 1.30 (d, J5,6 = 6.2 Hz, 3H, H-6 ); C NMR (CDCl ): δ
A′
3
1
71.0, 170.8, 169.5 (3C, CO ), 162.4 (NHCO), 133.9 (CH
Ac
1
CH ), 117.4 (CHCH ), 102.0 (C-1 , J = 163.4 Hz), 98.3
2
2
D′
C,H
1
(
C-1 , J = 173.2 Hz), 92.5 (CCl ), 79.2 (C-2 ), 73.7 (C-4 ),
A′ C,H 3 A′ A′
7
6
2.3 (2C, C-3 , C-5 ), 71.8 (C-3 ), 68.5 (C-4 ), 68.2 (C-5 ),
Amplified products were purified by a gel extraction kit.
Amylosucrase libraries were cloned into pGEX-6P-3 (GE
Healthcare Biosciences) using EcoRI HF and NotI HF restriction
sites. Ligation products were transformed in E.coli TOP 10
electrocompetent cells (Invitrogen, Carlsbad, U.S.A.). Freshly
transformed cells were plated on lysogeny broth (LB) agar
supplemented with 100 μg/mL ampicillin for 24 h at 37 °C.
Library Screening on Sucrose in Solid Media. A pH-
D′
D′
A′
D′
A′
8.1 (C ), 62.2 (C-6 ), 56.2 (C-2 ), 20.9, 20.7, 20.6 (3C,
All
D′
D′
CH3Ac), 17.5 (C-6 ). HR-ESI-TOF-MS m/z 658.0876 (Calcd
for C H Cl NO Na, 658.0837).
A′
23
32
3
13
Allyl (2-deoxy-2-trichloroacetamido-β-D-glucopyranosyl)-
(
1→2)-α-L-rhamnopyranoside (1 or D′A′). To a solution of
diol 7 (1.27 g, 2.0 mmol) in anhydr. MeOH (30 mL) stirred
under an Ar atmosphere, was added MeONa (25 wt % in MeOH,
2
35
75 μL, 1.20 mmol, 0.6 equiv). The reaction mixture was stirred
based high-throughput screening assay earlier described was
used to isolate sucrose active clones (see Supporting Information
and Methods provided in Supporting Information).
at rt for 2 h, then Dowex 50Wx8-200 resin (H+ form) was added.
The suspension was filtered over a pad of Celite. The solvent was
removed under vacuum and the residue was purified by column
chromatography (DCM/MeOH, 90:10 → 80:20) to give
pentaol 1 (937 mg, 1.83 mmol, 92%) as a white solid.
Expression of Active Clones in Microplate Format.
Active mutants were replicated to inoculate a 96-well microplates
−1
containing LB medium (200 μL), ampicillin (100 μg/mL ), and
glycerol (12% w/v) and incubated at 30 °C for 24h under
agitation at 450 rpm (orbital diameter of 3 mm). Plates were then
duplicated into 96-deep-well plates containing in each well 1 mL
2
4
Disaccharide 1 had R = 0.11 (DCM/MeOH 9:1); [α]
=
f
D
1
−
29.3° (c 1.0, water); H NMR (D O): δ 6.08−5.98 (m, 1H,
2
CHCH ), 5.45−5.39 (m, J
= 17.4 Hz, 1H, CHCH ),
2
trans
2
4
8
5
.37−5.33 (m, J = 10.4 Hz, 1H, CHCH ), 5.06 (d, J = 1.7
of autoinducing media ZYM-5052
supplemented with
cis
2
1,2
Hz, 1H, H-1 ), 4.98 (d, J = 8.1 Hz, 1H, H-1 ), 4.30−4.24 (m,
ampicillin (100 μg/mL) for culture and gene expression. These
plates were incubated 24 h at 30 °C with agitation at 700 rpm. To
evaluate the growth homogeneity, absorbance was measured at
600 nm for each culture diluted 10 times in water (Sunrise
A′
1,2
D′
1
H, H ), 4.16−4.11 (m, 2H, H , H-2 ), 4.00 (bd, J
= 12.5
All
All
A′
6a,6b
Hz, 1H, H-6a ), 3.91 (dd, J = 9.6 Hz, J = 3.1 Hz, 1H, H-3A′),
D′
3,4
2,3
3
.88−3.80 (m, 3H, H-6b , H-2 , H-3 ), 3.80−3.72 (m, 1H, H-
D′ D′ D′
5A′), 3.57−3.55 (m, 2H, H-4 , H-5 ), 3.50 (pt, J = 9.6 Hz, 1H,
H-4 ), 1.33 (d, J = 6.3 Hz, 1H, H-6 ). C NMR (D O): δ
̈
spectrophotometer, Tecan, Mannedorf, Switzerland). Plates
D′
D′
1
3
were then centrifuged (20 min, 3000g, 4 °C), supernatants
were removed, and the cell pellets were resuspended in 200 μL of
2× PBS buffer containing lysozyme (0.5 mg/mL) and DNase
and frozen at −80 °C overnight. After thawing at room
temperature, microplates were centrifuged (20 min, 3000g, 4
°C); soluble fractions were then transferred in new microplates
for reaction assays.
A′
5,6
A′
2
1
64.9 (NHCO), 133.3 (CHCH ), 118.5 (CHCH ), 101.4
2
2
1
1
(
(
7
C-1 , J = 164.4 Hz), 97.9 (C-1 , J = 172.9 Hz), 91.6
CCl ), 78.0 (C-2 ), 75.9 (C-4 ), 73.2 (C-3 ), 72.4 (C-4 ),
D′ C,H A′ C,H
3
A′
D′
D′
A′
0.1 (C-3 ), 70.0 (C-5 ), 68.8 (C-5 ), 68.3 (C ), 60.6 (C-
A′
D′
A′
All
6D′), 57.8 (C-2 ), 16.6 (C-6 ); HR-ESI-TOF-MS m/z
5
D′
A′
32.0504 (Calcd for C H Cl NO Na, 532.0520).
17 26 3 10
Bacterial Strains, Plasmids, and Chemicals. Details are
provided in Supporting Information and Methods.
Reaction assays in the presence of sucrose were carried out as
1
9,20
previously described.
All 55 positive hits were sequenced on
Library Construction. The gene library was constructed in
five PCR steps using PhusionR polymerase (New England
Biolabs, Ipswich, MA, U.S.A.) (Figure S2 of the Supporting
Information). In the first step, the six DNA wild-type fragments
surrounding the five regions containing designed mutations were
amplified (Figure S2 of the Supporting Information).
Amplified products were purified by a gel extraction kit. In a
second step, PCR reactions were divided in 10 pools (Figure S2
of the Supporting Information). In this way, in pool 1, the
fragment F1 (10 ng) was mixed with 226−229_rev (0.5 μM), the
reverse primer containing mutated codons for 226, 227, 228, and
the entire gene by GATC Biotech GA (Constance, Germany).
Reaction Assays with Sucrose and D′A′ Acceptor in
Microplate Format. Reactions were carried out in microplates
with 100 μL of soluble fractions and 50 μL of sucrose solution
(73 mM final concentration) and 50 μL of D′A′ acceptor
solution (73 mM final concentration). The enzymatic reaction
was incubated at 30 °C for 24 h under agitation (450 rpm). The
reactions were then stopped by heating at 95 °C for 5 min,
centrifugated (20 min at 3700 rpm), filtered and analyzed by
HPLC using conditions described hereafter.
Carbohydrate Analysis. Sucrose consumption was eval-
uated by ion exchange chromatography at 65 °C using an Aminex
HP87K column (300 × 7.8 mm, Biorad) with ultrapure water at
0.6 mL/min as eluent. The detection was performed using a
shodex RI 101 series refractometer. D′A′ consumption and
ED′A′ and EED′A′ production were identified using a reversed
C18 phase column (Venusil AQ 50 mm) with MeCN/water
(17:83, v/v) as eluent. The detection was performed using a UV
detector (Dionex, UVD 340) at 210 and 220 nm. The
glucosylated molecule was identified by LC-MS with the above
conditions. Using a MSQ Plus Mass Spectrometer (Dionex -
Thermoscientific equipment), molecules were ionised by
electrospray ionization technique with a source at 450 °C.
Separation was achieved with a quadripole and detected using the
negative mode.
2
2
2
29 positions. In pool 2, the fragment F2 was mixed with 226−
29_for, the forward primer containing mutated codons for 226,
27, 228, and 229 positions. In pool 3, the fragment F2 was
mixed with 289−300_rev, the reverse primer containing mutated
codons for 289, 290, and 300 positions, and so on for the 10
pools. PCR was conducted in a total volume of 20 μL in the
following conditions: one step at 98 °C for 30 s, followed by 20
cycles at 98 °C-10 s, 52 °C-30 s (pool 1 to 10), 72 °C-20 s. DNA
fragments and primers used in each pool contained overlapping
regions of at least 20 nucleotides. The primers containing
degenerated codons were specific to each library (Table S1 of the
Supporting Information). In a third step, amplicons obtained
from step 2 were mixed two by two to perform 1 cycle at 98 °C
for 30 s and 10 cycles at 98 °C for 10 s and 72 °C for 40 s. In a
1
196
DOI: 10.1021/cs501288r
ACS Catal. 2015, 5, 1186−1198

ACS Catalysis
Research Article
Production and Purification of Isolated Mutants. E. coli
6b , H-4 ), 3.66−3.63 (m, 2H, H-3 , H-5 ), 3.60−3.55 (m,
E′ A′ po 2,3 E
(pt , 1H, J = J4,5 = 9.3 Hz, H-4 ), 3.41 (pt , 1H, J = J4,5 = 9.2
po 3,4 E′ po 3,4
E′
D′
E′
A′
TOP10 electrocompetent cells containing pGST-AS mutated
2H, H-2 , H-5 ), 3.51 (dd , 1H, J = 10.4 Hz, 1H, H-2 ), 3.48
49
gene were produced as previously described but with a
preculture step at 30 °C overnight instead of 37 °C and then
cultures at 23 °C. After sonication, the crude extract was
centrifugated (10 000 rpm, 20 min, 4 °C) and the supernatant
was harvested for the purification step. NpAS mutants were
purified on an affinity chromatography column via affinity of the
GST tag to the Glutathion Sepharose 4B support (Dutscher,
Brumath, France) into Tris buffer (50 mM Tris, 150 mM NaCl, 1
mM EDTA, 1 mM DTT, pH 7.0). GST Tag was removed by
PreScission protease (Dutscher, Brumath, France). Protein
concentration was estimated using Nanodrop ND-1000
spectrophotometer, and specific activity was determined at 30
Hz, H-4 ), 3.40 (pt , 1H, J = J4,5 = 9.6 Hz, H-4 ), 1.23 (d, J =
E
po
3,4
A′
5,6
1
3
6.5 Hz, 1H, H-6 ); C NMR (D O): δ 167.5 (NHCO), 136.0
A′
2
1
(CHCH ), 121.1 (CHCH ), 104.0 (C-1 , J = 164.1
2
2
D′
C,H
1
1
Hz), 102.1 (C-1 , J = 172.7 Hz), 100.7 (C-1 , J = 171.8
E′
C,H
E
C,H
1
Hz), 100.5 (C-1 , J = 173.7 Hz), 94.2 (CCl ), 80.7 (C-2 ),
A′
C,H
3
A′
79.3 (C-3 ), 77.2 (C-5 ), 76.3 (C-3 ), 75.8 (C-3 ), 75.7 (C-
E
D′
D′
A′
4 ), 75.0 (C-4 ), 74.5 (C-3 ), 74.1 (2C, C-2 , C-2 ), 74.0 (C-
D′
A′
E
E′
E
5 ), 72.7 (C-5 ), 72.2 (C-4 ), 72.0 (C-4 ), 71.4 (C-5 ), 70.9
E′
E
E′
E
A′
(CAll), 68.5 (C-6E′), 63.3 (C-6D′), 63.1 (C-6
(C-6A′). m/z (ES-) for EED′A′, m/z 832 [M - H ].
), 60.4 (C-2D′), 19.2
E
+
°
C, in 50 mM Tris-HCl buffer pH 7.0, using 50 g/L sucrose
ASSOCIATED CONTENT
Supporting Information
The following file is available free of charge on the ACS
■
during 20 min by dinitrosalicylic acid assay. One unit of
amylosucrase activity corresponds to the amount of enzyme that
catalyzes the release of 1 μmol of reducing sugars per minute in
the assay conditions.
Characterization of Product Profiles of the Active
Mutant on D′A′. Reactions were performed at 30 °C during 24
h with 146 mM sucrose alone or supplemented with 146 mM
acceptor molecule. Wild-type enzyme and mutants were
employed at 1 U/mL, and reaction mixes were stopped by
heating at 95 °C during 5 min. Product profiles were analyzed by
HPAEC-PAD using a Dionex Carbo-Pack PA100 column at 30
S
*
Publications website at DOI: 10.1021/cs501288r.
Supplementary tables and figures showing details of library
design, construction, screening and analysis; general
1
13
methods for chemical synthesis; H and C NMR spectra
for all novel compounds (PDF)
AUTHOR INFORMATION
Corresponding Authors
*E-mail: isabelle.andre@insa-toulouse.fr.
*E-mail: remaud@insa-toulouse.fr.
■
°
C. Soluble fraction was analyzed with a gradient of sodium
acetate (from 6 to 300 mM in 28 min) in 150 mM NaOH and the
crude extract was solubilized in 1 M aqueous KOH. Detection
was performed using a Dionex ED40 module with a gold working
electrode and Ag/AgCl reference electrode.
Kinetic Parameters Determination of Active Mutant.
Enzyme kinetics were carried out with pure enzymes (0.2 and 0.1
U/ml) at 30 °C, under agitation. Catalytic efficiency of wild-type
NpAS and mutants was determined with sucrose alone from 0 to
Present Address
□
(
̀
For S.S.) Oxeltis, Cap Gamma − 1682, rue de la Valsiere, CS
2
7384, 34189 Montpellier, France
Author Contributions
#
A.V., E.C., and S.B. contributed equally.
Notes
The authors declare no competing financial interest.
6
00 mM and with both sucrose (146 mM) and acceptor
molecule (from 5 mM to 500 mM) as variable substrates.
Catalytic efficiencies were determined by measuring the release
ACKNOWLEDGMENTS
■
of fructose monitored for sucrose alone (k /Km, suc), and by
The authors are grateful to C. Topham for providing his help
with Sequester analysis. They also thank G. Cioci and S. Tranier
in using the PICT facility for protein biophysical character-
ization, N. Monties for her help in the product analysis and
purification by HPAEC-PAD and LC-MS as well as S. Bozonnet
and S. Pizzut-Serin in using the ICEO high-throughput facility,
devoted to the engineering and screening of new and original
cat
measuring the rate of formation (k /K ; k /Km,D′A′) of the
cat
m,suc cat
monoglucosylated disaccharide. After heating at 95 °C, and
centrifugation, aliquots were analyzed by HPLC RI with a Biorad
HPX-87K column.
Semipreparative Synthesis and Identification of Allyl
α-D-glucopyranosyl-(1→6)-α-D-glucopyranosyl-(1→4)-2-
deoxy-2-trichloroacetamido-β-D-glucopyranosyl-(1→2)-
α-L-rhamnopyranoside (E′ED′A′). A mixture of 146 mM
D′A′, 292 mM sucrose, and 1 U/mL mutant F3 (2 mL) was
gently stirred at 30 °C. After 24 h, the reaction mix was heated at
enzymes, of the Laboratoire d’Ingen
Biologiques et des Procedes (Toulouse, France). The authors
wish to thank C. Guerreiro (Unite de Chimie des Biomolecules,
́ ̀
ierie des Systemes
́
́
́
́
Institut Pasteur, Paris, France) for her technical contribution to
the synthetic aspects of this work, and F. Bonhomme (CNRS,
UMR 3523, Institut Pasteur, Paris, France) for providing the
HRMS data. This work was granted access to the HPC resources
95 °C for 5 min, and then centrifugated (6000 rpm, 4 °C, 20
min). The glucosylated products were separated by reverse-
phase semipreparative chromatography using a C18 column
(
Bishoff Chromatography, 250 mm × 8 mm), eluting with 17%
acetonitrile in pure water at a constant flow rate of 2 mL/min at
0 °C. The major product was lyophilized repeatedly to give
́ ́
of the Computing Center of Region Midi-Pyrenees (CALMIP,
Toulouse, France). This work was supported by the French
National Research Agency (ANR Project GLUCODESIGN,
ANR-08-PCVI-002-02) and by the European Commission
Seventh Framework Program (FP7/2007−2013) under Grant
agreement No. 261472-STOPENTERICS
3
1
tetrasaccharide EED′A′. This product had H NMR (D O): δ
5
5
J = 10.5 Hz, 1H, CHCH ), 4.98 (bs, 1H, H-1 ), 4.94 (d, J
cis
=
1
2
6
2
.99−5.89 (m, 1H, CHCH ), 5.43 (d, J = 3.8 Hz, 1H, H-1 ),
2
1,2
E
.35−5.31 (m, J
= 17.3 Hz, 1H, CHCH ), 5.27−5.25 (m,
trans
2
2
A′
1,2
3.6 Hz, 1H, H-1 ), 4.90 (d, J = 8.4 Hz, 1H, H-1 ), 4.19 (m,
REFERENCES
E′
1,2
D′
■
H, H ), 4.07−4.01 (m, 3H, H-2 , H-3 , H ), 3.99−3.91 (m,
All
A′ D′
All
(1) Turner, N. J.; Truppo, M. D. Curr. Opin. Chem. Biol. 2013, 17,
212−214.
(2) Turner, N. J.; O’Reilly, E. Nat. Chem. Biol. 2013, 9, 285−288.
H, H-6a , H-6a ), 5.88 (m, 1H, H-5 ), 3.84−3.77 (m, 3H, H-
E′
D′
E′
a , H-3 , H-5 ), 3.75−3.68 (m, 5H, H-6a , H-3 , H-6b , H-
D′
A′
E
E
E
E
1
197
DOI: 10.1021/cs501288r
ACS Catal. 2015, 5, 1186−1198

ACS Catalysis
Research Article
(
3) Davids, T.; Schmidt, M.; Bo
Opin. Chem. Biol. 2013, 17, 215−220.
4) Bornscheuer, U. T.; Huisman, G. W.; Kazlauskas, R. J.; Lutz, S.;
Moore, J. C.; Robins, K. Nature 2012, 485, 185−194.
5) Hohne, M.; Schatzle, S.; Jochens, H.; Robins, K.; Bornscheuer, U.
T. Nat. Chem. Biol. 2010, 6, 807−813.
̈
ttcher, D.; Bornscheuer, U. T. Curr.
(33) Reetz, M. T.; Kahakeaw, D.; Lohmer, R. ChemBioChem 2008, 9,
1797−1804.
(
(34) Stam, M. R.; Danchin, E. G. J.; Rancurel, C.; Coutinho, P. M.;
Henrissat, B. Protein Eng. Des. Sel. 2006, 19, 555−562.
(35) Champion, E.; Moulis, C.; Morel, S.; Mulard, L. A.; Monsan, P.;
(
̈
̈
Remaud-Simeo
(36) Pinto, B. M.; Morisette, D. G.; B., D. R. J. Chem. Soc., Perkin Trans.
987, 1, 9−14.
37) Ley, S. V.; Priepke, H. W. M.; Warriner, S. L. Angew. Chem., Int. Ed.
́ ́
n, M.; Andre, I. ChemCatChem 2010, 2, 969−975.
̈
̈
6) Kiss, G.; Çelebi-Olc u̧ m, N.; Moretti, R.; Baker, D.; Houk, K. N.
(
1
(
Angew. Chem., Int. Ed. 2013, 52, 5700−5725.
(
(
7) Hilvert, D. Annu. Rev. Biochem. 2013, 82, 447−470.
8) Kries, H.; Blomberg, R.; Hilvert, D. Curr. Opin. Chem. Biol. 2013,
Engl. 1994, 33, 2290−2292.
(38) Ley, S. V.; Polara, A. J. Org. Chem. 2007, 72, 5943−5959.
(39) Montchamp, J.-L.; Tian, F.; Hart, M. E.; Frost, J. W. J. Org. Chem.
1
(
7, 221−228.
9) Wijma, H. J.; Janssen, D. B. FEBS J. 2013, 280, 2948−2960.
10) Voigt, C. A.; Mayo, S. L.; Arnold, F. H.; Wang, Z.-G. Proc. Natl.
1
(
996, 61, 3897−3899.
40) Boutet, J.; Kim, T. H.; Guerreiro, C.; Mulard, L. A. Tetrahedron
Lett. 2008, 49, 5339−5342.
41) Mirza, O.; Skov, L. K.; Remaud-Simeon, M.; Potocki de Montalk,
G. P.; Albenne, C.; Monsan, P.; Gajhede, M. Biochemistry 2001, 40,
032−9039.
42) Malbert, Y.; Pizzut-Serin, S.; Massou, S.; Cambon, E.; Laguerre,
S.; Monsan, P.; Lefoulon, F.; Morel, S.; Andre, I.; Remaud-Simeon, M.
ChemCatChem 2014, 16, 2282−2291.
43) Cambon, E.; Barbe, S.; Pizzut-Serin, S.; Remaud-Simeon, M.;
Andre, I. Biotechnol. Bioeng. 2014, 111, 1719−1728.
44) Hornak, V.; Abel, R.; Okur, A.; Strockbine, B.; Roitberg, A.;
Simmerling, C. Proteins Struct. Funct. Genet 2006, 65, 712−725.
45) Kirschner, K. N.; Yongye, A. B.; Tschampel, S. M.; Gonzalez-
Outeirino, J.; Daniels, C. R.; Foley, B. L.; Woods, R. J. J. Comput. Chem.
008, 29, 622−655.
46) Wang, J.; Wolf, R. M.; Caldwell, J. W.; Kollman, P. A.; Case, D. A.
J. Comput. Chem. 2004, 25, 1157−1174.
47) Case, D. A.; Darde, T. A.; Cheatham III, T. E.; Simmerling, C. L.;
(
Acad. Sci. U.S.A. 2001, 98, 3778−3783.
11) Voigt, C. A.; Mayo, S. L.; Arnold, F. H.; Wang, Z. J. Cell. Biochem.
Suppl. 2002, 37, 58−63.
12) Wijma, H. J.; Floor, R. J.; Jekel, P. A.; Baker, D.; Marrink, S. J.;
Janssen, D. B. Protein Eng. Des. Sel. 2014, 27, 49−58.
13) Goldsmith, M.; Tawfik, D. S. Enzyme Engineering by Targeted
(
(
(
9
(
(
́
Libraries. In Methods in Enzymology; Keating, A. E., Ed.; Academic Press:
New York, 2013; Vol. 523, pp 257−283.
(
(
14) Allen, B. D.; Nisthal, A.; Mayo, S. L. Proc. Natl. Acad. Sci. U.S.A.
́
2
010, 107, 19838−19843.
15) Lippow, S. M.; Moon, T. S.; Basu, S.; Yoon, S.; Li, X.; Chapman, B.
A. Chem. Biol. 2010, 17, 1306−1315.
16) Chaparro-riggers, J. F.; Polizzi, K. M.; Bommarius, A. S. Biotechnol.
J. 2007, 2, 180−191.
17) Floor, R. J.; Wijma, H. J.; Colpa, D. I.; Ramos-Silva, A.; Jekel, P. A.;
Szymanski, W.; Feringa, B. L.; Marrink, S. J.; Janssen, D. B.
ChemBioChem 2014, 15, 1660−1672.
18) Andre, I.; Potocki-Veronese, G.; Barbe, S.; Moulis, C.; Remaud-
Simeon, M. Curr. Opin. Chem. Biol. 2014, 19, 17−24.
19) Champion, E.; Andre, I.; Moulis, C.; Boutet, J.; Descroix, K.;
Morel, S.; Monsan, P.; Mulard, L. A.; Remaud-Simeo
(
(
(
(
2
(
(
́
(
(
́
́
̀
Wang, J.; Duke, R. E.; Luo, R.; Merz, K. M.; Pearlman, D. A.; Crowley,
M.; Walker, R. C.; Zhang, W.; Wang, B.; Hayik, S.; Roitberg, A.; Seabra,
G.; Wong, K. F.; Paesani, F.; Wu, X.; Brozell, S.; Tsui, V.; Gohlke, H.;
Yang, L.; Tan, C.; Mongan, J.; Hornak, V.; Cui, G.; Beroza, P.; Mathews,
D. H.; Schafmeister, C.; Ross, W. S.; Kollmann, P. A. AMBER 9;
University of California: San Francisco, CA, 2006
́
(
́
́
n, M. J. Am. Chem.
Soc. 2009, 131, 7379−7389.
(
20) Champion, E.; Guer
́
in, F.; Moulis, C.; Barbe, S.; Tran, T. H.;
Morel, S.; Descroix, K.; Monsan, P.; Mourey, L.; Mulard, L. A.; Tranier,
S.; Remaud-Simeon, M.; Andre, I. J. Am. Chem. Soc. 2012, 134, 18677−
8688.
21) Perepelov, A. V.; Shekht, M. E.; Liu, B.; Shevelev, S. D.; Ledov, V.
A.; Senchenkova, S. N.; L’vov, V. L.; Shashkov, A. S.; Feng, L.; Aparin, P.
G.; Wang, L.; Knirel, Y. A. FEMS Immunol. Med. Microbiol. 2012, 66,
(
(
48) Studier, F. W. Protein Expr. Purif. 2005, 41, 207−234.
́
́
49) Potocki de Montalk, G.; Remaud-Simeon, M.; Willemot, R. M.;
́
1
(
Planchot, V.; Monsan, P. J. Bacteriol. 1999, 181, 375−381.
(50) Lemanski, G.; Ziegler, T. Eur. J. Org. Chem. 2006, 2618−2630.
(51) Dhara, D.; Kar, R. K.; Bhunia, A.; Misra, A. K. Eur. J. Org. Chem.
2014, 4577−4584.
(52) Hargreaves, J.; Le Guen, Y.; Guerreiro, C.; Descroix, K.; Mulard,
L. A. Org. Biomol. Chem. 2014, 12, 7728−7749.
2
01−210.
22) Shashkov, A. S.; Senchenkova, S. N.; Sun, Q.; Lan, R.; Wang, J.;
Perepelov, A. V.; Knirel, Y. A.; Xu, J. Carbohydr. Res. 2013, 24, 93−96.
23) Skov, L. K.; Mirza, O.; Henriksen, A.; Potocki de Montalk, G.;
Remaud-Simeon, M.; Sarcabal, P.; Willemot, R.; Monsan, P.; Gajhede,
M. J. Biol. Chem. 2001, 276, 25273−25278.
24) Skov, L. K.; Mirza, O.; Gajhede, M.; Feller, G.; D’Amico, S.;
Andre, G.; Potocki-Veronese, G.; van der Veen, B.; Monsan, P.;
Remaud-Simeon, M. J. Biol. Chem. 2004, 279, 726−734.
25) Skov, L. K.; Mirza, O.; Sprogøe, D.; Dar, I.; Remaud-Simeon, M.;
Albenne, C.; Monsan, P.; Gajhede, M. J. Biol. Chem. 2002, 277, 47741−
7747.
26) Hanson, S.; Best, M.; Bryan, M. C.; Wong, C-H. Trends Biochem.
Sci. 2004, 29, 656−663.
27) Wang, Z.; Chinoy, Z.; Ambre, S.; Peng, W; McBride, R.; de Vries,
R. P.; Glushka, J.; Paulson, J. C.; Boons, G. J. Science 2013, 341, 379−
83.
(
(
́
̧
(
́
(
4
(
(
3
(
(
28) Armstrong, Z.; Withers, S. G. Biopolymers 2009, 99, 666−674.
29) Muthana, S.; Cao, H.; Chen, X. Curr. Opin. Chem. Biol. 2009, 13,
5
73−581.
30) Lucas, R.; Hamza, D.; Lubineau, A.; Bonnaffe,
Chem. 2004, 10, 2107−2117.
31) Boutet, J.; Guerreiro, C.; Mulard, L. A. J. Org. Chem. 2009, 74,
651−2670.
32) Chassagne, P.; Raibaut, L.; Guerreiro, C.; Mulard, L. A.
Tetrahedron 2013, 69, 10337−10350.
(
́
D. Eur. J. Org.
(
2
(
1
198
DOI: 10.1021/cs501288r
ACS Catal. 2015, 5, 1186−1198