Kinetic resolution of α-bromoamides: Experimental and theoretical investigation of highly enantioselective reactions catalyzed by haloalkane dehalogenases

Article abstract of DOI:10.1002/adsc.201000726

Haloalkane dehalogenases from five sources were heterologously expressed in Escherichia coli, isolated, and tested for their ability to achieve kinetic resolution of racemic α-bromoamides, which are important intermediates used in the preparation of bioactive compounds. To explore the substrate scope, fourteen α-bromoamides, with different Cα- and N-substituents, were synthesized. Catalytic activity towards eight substrates was found, and for five of these compounds the conversion proceeded with a high enantioselectivity (E value >200). In all cases, the (R)-α-bromoamide is the preferred substrate. Conversions on a preparative scale with a catalytic amount of enzyme (enzyme:substrate ratio less 1:50 w/w) were all completed within 17-46 h and optically pure α-bromoamides and α-hydroxyamides were isolated with good yields (31-50%). Substrate docking followed by molecular dynamics simulations indicated that the high enantioselectivity results from differences in the percentage of the time in which the substrate enantiomers are bound favourably for catalysis. For the preferred (R)-substrates, the angle between the attacking aspartate oxygen atom of the enzyme, the attacked carbon atom of the substrate, and the displaced halogen atom, is more often in the optimal range (>157°) for reactivity. This can explain the observed enantioselectivity of LinB dehalogenase in a kinetic resolution experiment.

Full text of DOI:10.1002/adsc.201000726

FULL PAPERS
DOI: 10.1002/adsc.201000726
Kinetic Resolution of a-Bromoamides: Experimental and
Theoretical Investigation of Highly Enantioselective Reactions
Catalyzed by Haloalkane Dehalogenases
a
a,b
a
c
´
Alja Westerbeek, Wiktor Szymanski, Hein J. Wijma, Siewert J. Marrink,
Ben L. Feringa,b and Dick B. Janssena,*
Department of Biochemistry, Groningen Biomolecular Sciences and Biotechnology Institute, University of Groningen,
Nijenborgh 4, 9747 AG Groningen, The Netherlands
a
Fax : (+31)-50-363-4165; phone: (+31)-50-363-4209; e-mail: D.B.Janssen@rug.nl
Center for Systems Chemistry, Stratingh Institute for Chemistry, University of Groningen, Nijenborgh 4, 9747 AG
Groningen, The Netherlands
Groningen Biomolecular Sciences and Biotechnology Institute and Zernike Institute for Advanced Materials, University
of Groningen, Nijenborgh 4, 9747 AG Groningen, The Netherlands
b
c
Received: September 23, 2010; Revised: February 17, 2011; Published online: April 13, 2011
Supporting information for this article is available on the WWW under
http://dx.doi.org/10.1002/adsc.201000726.
Abstract: Haloalkane dehalogenases from five sour- lated with good yields (31–50%). Substrate docking
ces were heterologously expressed in Escherichia followed by molecular dynamics simulations indicat-
coli, isolated, and tested for their ability to achieve ed that the high enantioselectivity results from differ-
kinetic resolution of racemic a-bromoamides, which ences in the percentage of the time in which the sub-
are important intermediates used in the preparation strate enantiomers are bound favourably for cataly-
of bioactive compounds. To explore the substrate sis. For the preferred (R)-substrates, the angle be-
scope, fourteen a-bromoamides, with different Ca- tween the attacking aspartate oxygen atom of the
and N-substituents, were synthesized. Catalytic activ- enzyme, the attacked carbon atom of the substrate,
ity towards eight substrates was found, and for five and the displaced halogen atom, is more often in the
of these compounds the conversion proceeded with a optimal range (>1578) for reactivity. This can ex-
high enantioselectivity (E value >200). In all cases, plain the observed enantioselectivity of LinB dehalo-
the (R)-a-bromoamide is the preferred substrate. genase in a kinetic resolution experiment.
Conversions on a preparative scale with a catalytic
amount of enzyme (enzyme:substrate ratio less 1:50 Keywords: a-bromoamides; enantioselective bioca-
w/w) were all completed within 17–46 h and optically talysis; haloalkane dehalogenase; kinetic resolution;
pure a-bromoamides and a-hydroxyamides were iso- molecular dynamics
Introduction
in the expansion of the stereochemical scope of possi-
ble derivatives and overcomes the limitations caused
Enantiopure amides substituted at the a-position with by the homochirality of the enzymes.
a good leaving group are important precursors for the
Haloalkane dehalogenases consist of an a/b-hydro-
synthesis of non-natural peptides[1], diketopipera- lase domain and a cap domain, with the catalytic site
zines[2], a-alkylthioamides[3], and other compounds located in between. The nucleophilic aspartate of the
important in medicinal chemistry.[4] The use of haloal- catalytic pentad catalyzes the hydrolytic cleavage of
kane dehalogenases for the kinetic resolution of race- halogenated alkanes, yielding the corresponding alco-
mic a-bromoamides would present a biocatalytic hols. Two other residues, a tryptophan and a trypto-
route towards these valuable intermediates. Addition- phan/asparagine, are involved in stabilizing the halide.
al possibilities for the functionalization of a-bromo- The remaining catalytic residues are a histidine and
ACHTUNGTRENNUNGamides are offered by the fact that subsequent reac- an aspartate/glutamate, which are activating a water
tions with nucleophiles may be carried out with either molecule for hydrolysis of the alkyl-enzyme inter-
inversion or retention of configuration.[5] This results mediate and stabilizing the positive charge that devel-
Adv. Synth. Catal. 2011, 353, 931 – 944
ꢀ 2011 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
931

Alja Westerbeek et al.
FULL PAPERS
ops on the histidine, respectively.[6] Various haloal- tion towards enhanced catalytic activity for 1,2,3-tri-
kane dehalogenases were studied for biotechnological chloropropane.[13]
applications.[7] Enantioselective conversions using hal-
oalkane dehalogenases have been reported for esters
and b-bromoalkanes.[8] We report here the enantiose- Results and Discussion
lectivity and diastereoselectivity of the dehalogenase-
catalyzed conversion of a-bromoamides.
Cloning and Production of Haloalkane
In the current study, we have screened the follow- Dehalogenases
ing six haloalkane dehalogenases for catalytic activity
and enantioselectivity towards a series of a-bromo- The haloalkane dehalogenase genes were expressed
ACHTUNGTRENNUNGamides: DhlA, LinB, DbjA, DmlA, DhaA and its in E. coli (for further details, see Experimental Sec-
mutant DhaA31. Haloalkane dehalogenase from Xan- tion). Typically, approximately 50 mg of enzyme were
thobacter autotrophicus (DhlA) is involved in growth obtained per 1 L of overnight induced cultures after
on 1,2-dichloroethane as a sole carbon source.[9] purification with Ni-NTA columns. The purity of the
DhaA was originally discovered in a Rhodococcus isolated enzymes was >95% as judged by SDS-poly-
rhodochrous strain, growing on 1-chlorobutane.[10] acrylamide gel electrophoresis.
LinB from Sphingomonas paucimobilis UT26 plays a
role in the metabolism of hexachlorocyclohexane.[11]
DbjA from Bradyrhizobium japonicum USDA110 Catalytic Activity and Enantioselectivity/Diastereo-
and DmlA from Mesorhizobium loti MAFF303099 selectivity of Haloalkane Dehalogenases towards a-
were discovered through genome mining.[12] DhaA31 Bromoamides
is a DhaA mutant variant obtained by directed evolu-
The obtained enzymes were screened for catalytic ac-
tivity and enantio- or diastereoselectivity towards rac-
emic a-bromoamides 1 (Scheme 1).
In order to study the influence of different substitu-
ents in substrates 1, we decided to independently
change the Ca- and N-substituents (R and R’ respec-
tively, Scheme 1) and test for catalytic activity and se-
lectivity for compounds 1a–1n (Figure 1). Simple com-
Scheme 1. Kinetic resolution of a-bromoamides.
Figure 1. The a-bromoamides tested for dehalogenase reactions in this study.
932
asc.wiley-vch.de
ꢀ 2011 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Adv. Synth. Catal. 2011, 353, 931 – 944

Kinetic Resolution of a-Bromoamides: Experimental and Theoretical Investigation
Table 1. Enantioselectivity and diastereoselectivity (de) of haloalkane dehalogenases.[a]
Compound DbjA
LinB
DhaA
I[b]
DhaA31
DmlA
I
1a
1b
>200 (15%) 67 (43%)
>200 (50%) 28 (46%)
>200
(18%)
>200
(50%)
ND
I
>200
(28%)
I
I
I
I
I
1c
1e
1f
1g
ND[c]
74[d] (53%)
10 (65%)
ND
ND
4 (57%)
>200 (42%) 11[e] (67%)
I
I
ND
I
deBr >99%;[e]
deBr =32%;[e])
deOH >99%[e] (50%)
deBr >95%;[d]
deOH =88%[e] (27%)
1h
I
I
I
I
I
I
I
deOH >95%[d] (50%)
>200[e] (37%)
1k
I
[a]
Reported are the E values and for 1g and 1h the diasteriomeric excess (de) of the remaining substrate (Br) or product
(OH). Conversion is given in brackets. The E values and conversions were calculated from the enantiomeric excess of
substrate and product using the equations of Chen et al.[14] For further details see Experimental Section. A complete
table is available in the Supporting Information, Table S1.
I, Inactive, less than 10% conversion in 3 h. Compounds 1d, 1i, 1j, 1l, 1m and 1n are also not converted with detectable
catalytic activity.
ND, not determined since conversion was too low to obtain a reliable E value.
These values are taken from Table 2, as compounds were not detectable on small scale.
Incubations done overnight.
[b]
[c]
[d]
[e]
pounds (1a–1d) and C-protected dipeptide precursors butyl ester. The conversion rate for DmlA was low
(1e–1j, 1l, 1m) were chosen. Enantioselectivity or dia- for both 1e and 1f.
stereoselectivity for these compounds is displayed in
Table 1.
To determine the scope of possible amino acid sub-
stituents that can be incorporated into compound 1,
Compound 1a (Figure 1, R=Me, R’=Bn) was con- we tested an additional set of dipeptide precursors.
verted by DbjA, LinB and DhaA31, with E values of For the l-alanine derivative 1g a reasonable catalytic
>200 for DbjA and DhaA31. Compound 1b (R’= activity was observed only with LinB and DmlA. The
Ph) was converted by four enzymes, and enantioselec- diastereoselectivity with LinB was high, as the diaste-
tivities of >200 were found for DbjA, DhaA and reomeric excess (de) for both the bromoamide and
DhaA31. Replacement of the aromatic group by an the hydroxyamide product was over 99%. For DmlA
aliphatic chain (R’=n-Bu, compound 1c), resulted in the diastereoselectivity was not spectacular, as 32%
a loss of catalytic activity with almost all enzymes. de and 88% de were measured for the a-bromoamide
Conversion of 1c by DbjA and DmlA was measura- and a-hydroxyamide, respectively. Catalytic activity
ble, yet low. LinB converted compound 1c with an ac- with the l-phenylalanine derivative 1h was only ob-
ceptable E value of 74. When the R’ substituent is a served with LinB, where the diastereoselectivity was
tert-butyl group (compound 1d), none of the studied high, since a de of over 95% was obtained for both
haloalkane dehalogenases possesses detectable cata- the a-bromoamide and the a-hydroxyamide. For the
lytic activity.
compounds 1g and 1h, the (R,S)-a-bromoamide was
Encouraged by the high enantioselectivities ob- preferred over the (S,S)-a-bromoamide. Of the other
served in the enzymatic conversions of compounds 1a, tested compounds, the l-valine derivative 1i and l-as-
1b, and 1c, we decided to test the catalytic activity of paragine derivative 1j were not converted by any of
haloalkane dehalogenases with ester derivatives that the haloalkane dehalogenases.
are potential dipeptide precursors. Catalytic activity
Subsequently, we focused on the Ca-substituent (R,
towards a glycine ethyl ester derivative 1e was ob- Scheme 1), which was altered from a methyl group in
served with DbjA and LinB; albeit with low E values compound 1a to an ethyl group (compound 1k). Cata-
(Table 1). However, changing the ethyl ester (1e) into lytic activity towards compound 1k was only found
a tert-butyl ester (compound 1f), enhances the enan- with LinB, with an excellent E value of over 200.
tioselectivity tremendously to an E value of >200 DbjA and DhaA31 show no catalytic activity when R
with DbjA. For LinB, the E value with 1f did not im- is an ethyl group. Compound 1n, where the Ca-sub-
prove when the ethyl ester was replaced by a tert- stituent is changed into an n-butyl group, is not ac-
cepted as a substrate by any of the enzymes, which
Adv. Synth. Catal. 2011, 353, 931 – 944
ꢀ 2011 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
asc.wiley-vch.de
933

Alja Westerbeek et al.
FULL PAPERS
sets a limitation to the substituents at the Ca-position.
The absolute configuration of the products of the
Since LinB accepted both amino-acid derived com- kinetic resolutions was determined by comparison to
pounds (1g and 1h) and an ethyl substituent at the literature data or by correlation synthesis of products
Ca-position (compound 1k), we have prepared addi- 2 from commercially available enantiopure a-hydroxy
tional compounds 1l and 1m, which combine these acids. In all cases, the enzymes preferentially convert-
features. However, no detectable catalytic activity was ed the (R)-enantiomer of the a-bromoamide with in-
found with any of the enzymes. This indicates that the version of configuration to the (S)-enantiomer of the
Ca-substituent is at a sensitive position and a sub- a-hydroxyamide.
stituent larger than methyl is only accepted if the sub-
stituent R’ is small.
DhlA had no detectable catalytic activity for any of Preparative Scale Conversion
the compounds (1a–n). It has been reported that
DhlA, the 1,2-dichloroethane dehalogenase from X. Since the preparative-scale conversions are limited by
autotrophicus, only converts smaller substrates, mainly the low solubility of the substrates, we studied the
primary halides, which is in agreement with its smaller possibility of increasing the solubility by the addition
active site.[15]
of the water-miscible cosolvent dimethyl sulfoxide
In summary, LinB is the best enzyme in terms of (DMSO). Substrate conversion, measured after
catalytic activity, as it converts no less than eight of 150 min, was compared to conversion in reactions
the fourteen tested compounds. This corresponds well without cosolvent (Figure 2). The relative conversion
to literature reports, where LinB was shown to have was still about 50% when up to 25% (v/v) DMSO
the largest active site cavity of all haloalkane dehalo- was added to the substrate solution. Using these opti-
genases of which the structure is solved.[16] It has the mized conditions (25% DMSO), a larger preparative
best activity towards larger substrates, which is in scale experiment was performed in which 250 mg of
agreement with its physiological function in tetra-
chlorocyclohexadiene degradation.[17] With respect to
the enantioselectivity, DbjA is the best enzyme, as the
E value for three out of the four converted substrates
is >200. This is in agreement with experimental data
measured for conversion of a-bromo esters and b-bro-
moalkanes.[8b]
For all the compounds for which enzymatic activity
and good selectivity were found, preparative kinetic
resolution experiments were conducted with usually
80 mg of substrate. In a typical experiment, the sub-
strate was incubated with a small amount of enzyme
(enzyme:substrate ratio 1:21–1:124 w/w, Table 2). All
conversions were complete within 17–46 h of incuba-
tion at 308C. All products were obtained with high
optical purity (95–99% ee/de), except for 1c, for
which an ee of 87% was found for the a-hydroxy-
ACHTUNGTRENNUNG
ACHTUNGTRENNUNG
Figure 2. Conversion of 1a by DbjA at various concentra-
31% to an excellent yield of 50% (Table 2).
tions of dimethyl sulfoxide.
Table 2. Results of the preparative scale experiments.
Compound Enzyme Enzyme:substrate ratio (w:w) Time [h]
(S)-a-Bromoamide
(S)-a-Hydroxyamide
E value
yield
ee/de
yield
ee/de
1a
1b
1c
1f
1g
1h
1k
DbjA
DbjA
LinB
DbjA
LinB
LinB
LinB
1:58
1:124
1:110
1:95
1:67
1:21
1:42
46
17
20
20
43
20
44
41%
44%
47%
44%
48%
39%
40%
>99%
>99%
>99%
98%
>99%
>95%
>99%
50%
46%
47%
40%
31%
39%
31%
>99%
>99%
87%
>99%
97%
>200
>200
74
>200
NAa)
NA
>95%
97%
>200
[a]
NA, not applicable.
934
asc.wiley-vch.de
ꢀ 2011 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Adv. Synth. Catal. 2011, 353, 931 – 944

Kinetic Resolution of a-Bromoamides: Experimental and Theoretical Investigation
racemic 1a was subjected to kinetic resolution by
DbjA (2.8 wt% enzyme/substrate), yielding after 44 h
44% of 1a and 49% of 2a, both with an ee of 99%.
Computational Analysis of LinB Stereoselectivity
To explain the origin of the enantioselective conver-
sion of bromoamide 1f by LinB, we carried out dock-
ing and MD simulations. A crystal structure of the
enzyme is available.[18] Possible explanations for the
observed (R)-selectivity with different substrates are
(i) stronger binding of (R)-substrates, preventing the
(S)-substrates from binding and conversion, or (ii)
higher reactivity of the bound (R)-substrates, resulting
Scheme 2. Distances and angles relevant for nucleophilic
attack and leaving group stabilization in LinB-substrate
complexes.
in faster turnover than with the (S)-substrate. Two
substrates, for which the enzyme differs in enantiose-
lectivity, were selected: compound 1f (E value=11)
and 1k (E value>200). Both enantiomers of these
compounds were docked into LinB.
NACs during simulations was found to correspond to
The first possibility, preferential conversion of the a higher experimental catalytic rate.[21] The criteria to
(R)-substrate due to stronger binding as compared to define a NAC for the initial nucleophilic attack were
the (S)-substrate, was evaluated by predicting the dis- distances d1 <3.41 ꢁ, d2 <3.50 ꢁ, d3 <3.50 ꢁ, angles
sociation constants of the substrates to the enzyme by q1 =157–1808,
q2 =120–1808,
q3 =120–1808
AutoDock4.[19] The program predicts the binding (Scheme 2); all criteria need to be met at the same
energy by calculating the difference in energy be- time. The criteria for d1 and q1 are from Hur et al.[24]
tween a bound and unbound state based on separate while the criteria for d2, d3, q2, q3, which indicate
potentials for H-bonding, desolvation, Van der Waals whether there are H-bonds from the side-chains of
and electrostatic interactions, and for the entropy loss Asn38 and Trp109 that stabilize the development of a
by the ligand upon binding, each with its own empiri- negative charge on the bromide, were used addition-
cally determined weighing factor. The predicted bind- ally since these H-bonds enhance catalysis.[25]
ing energies [(R)-1f, DG=ꢀ24 kJmolꢀ1; (S)-1f, DG=
The obtained percentages of the time that an NAC
ꢀ25 kJmolꢀ1; (R)-1k, DG=ꢀ26 kJmolꢀ1; (S)-1k, was present during the simulation were for (R)-1f
DG=ꢀ29 kJmolꢀ1] for the different substrates 6.7ꢁ0.6%, for (S)-1f 0.41ꢁ0.09%, for (R)-1k 0.81ꢁ
docked into LinB are within the reported error of the 0.25%, and less than 0.001% for (S)-1k. The ratio of
method (12 kJmolꢀ1)[19] identical to each other. Thus, these NAC percentages for (R/S)-1f was 16, in good
there is no evidence that the enantioselective catalysis agreement with the experimentally observed E value.
is due to preferential binding of the (R)-enantiomers Furthermore, the ratio for (R/S)-1k, which is extreme-
by LinB.
ly high (>800) since no NACs were observed for the
The alternative explanation, that the observed (S)-enantiomer, is in excellent agreement with the ex-
enantioselectivity is due to the (R)-enantiomers bind- perimentally observed E value of >200. Thus, the
ing in an orientation more suitable for catalysis, was MD simulations indicate that the enantioselectivity
evaluated by MD simulations of the enzyme-substrate can indeed be due to differences in the reactivity of
complexes. Earlier work revealed that for poor sub- the bound substrate.
strates of a haloalkane dehalogenase the critical step
To explore the origin of the differences in reactivity
in the catalytic cycle is the first step, in which an as- of the bound substrates, the angles and distances that
partate side-chain nucleophilically attacks the sub- are important for a nucleophilic substitution reaction
strate with the halide as the leaving group.[20] There- to occur were analyzed (Scheme 2). The distribution
fore, the relative reactivities for this nucleophilic of q1 in the (R)-1f-LinB complex has the highest over-
attack were quantified as the percentage of the time lap with the range of angles that are optimal for nu-
during simulation that a near attack conformation cleophilic attack, more than with (R)-1k and (S)-1f,
(NAC) occurs. NACs are defined as having reacting while for (S)-1k only a few excursions into the catalyt-
atoms at Van der Waals distance and angles ꢁ15–208 ically necessary range were observed (Figure 3). For
of the angle of the forming bond in the TS.[21] NACs all other catalytically relevant distances and angles,
necessarily precede the transition state in the catalytic only minor differences in the overlap of the thermally
pathway[22], their presence corresponds to transition accessible range with the catalytically relevant range
state stabilization,[23] and a higher percentage of were observed for the different enantiomers (see Sup-
Adv. Synth. Catal. 2011, 353, 931 – 944
ꢀ 2011 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
asc.wiley-vch.de
935

Alja Westerbeek et al.
FULL PAPERS
donate H-bonds to the halogen. Another H-bond is
formed between the substrate amide hydrogen atom
and a carboxylate oxygen of the catalytic Asp108
(Figure 4). The (R)- and (S)-substrates bind in similar,
but mirrored, orientations in the active site [(R)- and
(S)-substrates are shown superimposed]. The R and
R’ substituents are bound in the same, mostly hydro-
phobic, substrate cavity. Most residues that line the
active site and interact with the (R)-enantiomer also
interact with the (S)-enantiomer (Figure 4). There-
fore, it was not possible to identify particular residues
that are responsible for the observed enantioselectivi-
ty. Thus, the preferred (R)-substrates have a binding
mode similar to that of (S)-substrates, but with a dis-
tribution of the q1 angle for nucleophilic attack that is
better for reaching the transition state, and the deci-
sive 10–258 difference is due to the way in which the
R and R’ substituents of the substrate fit in the irregu-
larly shaped, mostly hydrophobic substrate binding
pocket.
Kinetic Characterization of LinB Enantioselectivity
Substrate 1f was selected for further experimental
characterization of its steady-state kinetic preference
for the (R) enantiomer, since the lower E value
makes it possible to follow conversion of both enan-
tiomers over time. Since the pure (R)-enantiomer of
1f is not available, the kinetic parameters were deter-
mined from a time course depletion of racemic sub-
strate by LinB (Figure 5A). In a plot of substrate con-
centration versus enzyme concentrationꢂtime (Sel-
wynꢃs test of inactivation[26]), the points measured at
high and low LinB concentrations superimpose (Fig-
ure 5B), which shows that no significant inactivation
Figure 3. Nucleophilic attack angle q1 from an MD simula-
tion of LinB-substrate complexes versus time. The black
trace is the plot. Angles that fall within the range defined
for the NAC are indicated with a gray background and a
gray separation line. The displayed time range is 2000 to
7000 ps after start of the MD simulation. The labels on the
right indicate the substrate.
porting Information Figure S1–Figure S5). To verify of LinB occurs during the 7 h assay. The E value de-
that none of the other five criteria determined the re- termined by fitting with numerical integration of the
activity of the substrates, the analysis was redone Michaelis–Menten equation was 24ꢁ3, which is
omitting the q1 criterion, indeed resulting in similar higher than found in a single timepoint measurement
percentages of pseudo-reactive conformation [46.7ꢁ (Table 1). The obtained kinetic parameters were
1.4% for (R)-1f, 75.2ꢁ0.7% for (S)-1f, 54ꢁ1% for kcatS/KmS =26Mꢀ1 sꢀ1, Km ꢂ1 mM, the kcatR =0.082ꢁ
S
(R)-1k, and 38.0ꢁ0.7% for (S)-1k], which cannot ex- 0.04 sꢀ1, and the KmR =0.13ꢁ0.02 mM. For kcat no
S
plain the observed enantioselectivities. This suggests unique value could be obtained since there was no
that the main cause for differences in catalytic activity saturation at the highest concentration tested
S
of bound substrate enantiomers is the frequency of (500 mM, maximum solubility), and for Km only a
occurrence of conformations with a suitable angle q1 lower limit (1 mM) could be obtained for the same
for nucleophilic attack (Scheme 2).
reason.
To investigate whether these kinetic parameters,
To understand why the preferred (R)-substrates are
bound with a distribution of the angle q1 that leads to which indicate a lower Km for the preferred (R)-enan-
a higher NAC percentage, the trajectories of the MD tiomer, are compatible with a faster dehalogenation
simulations (see Supporting Information, Movies S1– reaction of the preferred (R)-enantiomer without
S4) were visually inspected. This revealed that the much stronger binding, as predicted by the docking
substrates are bound with a total of three H-bonds, energies and MD simulations, the steady-state kinetic
similar to what was found earlier for the DbjA deha- parameters were fitted to the minimal kinetic
logenase by modelling.[8b] The side-chain amide hy- scheme[27] for LinB (Scheme 3). In this scheme, k2 is
drogen of Asn38 and the indole hydrogen of Trp109 the rate of carbon-halogen bond cleavage, and k3 is
936
asc.wiley-vch.de
ꢀ 2011 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Adv. Synth. Catal. 2011, 353, 931 – 944

Kinetic Resolution of a-Bromoamides: Experimental and Theoretical Investigation
Figure 4. Stereo presentation of binding modes of 1f (upper panels) and 1k (lower panels) enantiomers in LinB during the
MD simulations. Light gray, LinB carbon atoms that bind (S)-substrate; purple, carbon atoms of (S)-substrate; dark gray,
LinB carbon atoms binding (R)-substrate; yellow, carbon atoms of (R)-substrate; blue, nitrogen atoms; green, bromide
atoms; turquoise, hydrogen atoms involved in H-bonds. The atom positions are the average over all snapshots after equili-
bration. Pictures were created with Yasara (www.yasara.org) and Povray (www.povray.org). Left-hand and middle pictures,
cross-eyed view; middle and right-hand pictures, linear view.
Figure 5. Time-course determination of kinetic parameters for the 1f conversion by LinB. (A): Concentration of substrate
enantiomers versus time at different LinB concentrations. See panel B for symbols. (B): Selwynꢃs plot to test for inactiva-
tion.
the rate of hydrolysis of the covalent alkyl-enzyme in- Michaelis complex) depending on the kinetic scheme
termediate.[6] In general, Km can be larger than, equal and the constants.[28,29] For the kinetic mechanism de-
to, or smaller than KD (dissociation constant of the picted in Scheme 3, Eq. (1) and Eq. (2) apply. We
S
R
S
R
S
fitted the values of KDR, KD , k2 , k2 , k3 , and k3 to
reproduce all the measured values for 1f conversion
described in the previous paragraph, while constrain-
R
S
ing the ratio of k2 /k2 to the 16 that is predicted by
the NAC analysis above. Under this constraint, it was
still possible to obtain solutions that within 0.1% ac-
Scheme 3. Minimal kinetic scheme for a dehalogenase. ES,
Michaelis complex; EI, covalent intermediate.
Adv. Synth. Catal. 2011, 353, 931 – 944
ꢀ 2011 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
asc.wiley-vch.de
937

Alja Westerbeek et al.
FULL PAPERS
curacy reproduced the steady-state kinetic values and 1k. THis influences the reaction rate of the bound
kinetic resolution pattern (Figure 5A). Inspection of substrates. The (R)- and (S)-substrates both in confor-
Eq. (1) and Eq. (2) reveals that for this agreement it mations in which Asn38 and Trp109 make H-bonds to
the halide leaving group, the substrate amide makes a
H-bond to Asp108, and the remainder of the sub-
strate binds in the mostly hydrophobic active site
cavity of LinB. How the remainder of the substrate
enantiomer fits into this asymmetric part of the sub-
strate binding pocket appears to control the angle for
nucleophilic attack, which ultimately determines
enantioselectivity. The obtained steady state kinetic
parameters for conversion of (R)-1f and (S)-1f were
is required that the hydrolysis step (k3) is slower than compatible with the results of MD simulations, and
the dehalogenation step (k2), which for LinB was suggested that the effect of the faster dehalogenation
indeed confirmed experimentally with different sub- step in the catalytic cycle is not that the kcat increases
strates.[27] In this situation (k2 @k3), the higher k2 for for the (R)-enantiomer but that its Km is lowered,
the preferred enantiomer does not result in a signifi- which ultimately leads to preferred conversion of the
cant increase of kcat but the Km is lowered, causing a (R)-enantiomer.
net increase of kcat/Km. Thus, the MD simulations that
predict a faster k2 for the best enantiomer are com-
patible with the measured steady state kinetics for the
enantiomers of 1f and the known kinetics of LinB, in
Experimental Section
which a faster k2 lowers the Km for the substrate. A
difference in reaction rate causing enantioselectivity
has also been observed in halohydrin dehalogenase,
General Procedure for the Synthesis of Racemic a-
Bromoamides (1)
where unproductive binding of an epoxide substrate
prevented conversion of (S)-halohydrins and unpro- were stirred in DCM (8 mL) at 58C. Triethylamine
2-Bromoacyl bromide (2.0 mmol) and amine (2.0 mmol)
(2.2 mmol or 5.0 mmol, when the amine hydrochloride was
used as reactant) in DCM (2 mL) was added dropwise, and
the cooling was stopped. After 2–4 h the volatiles were
evaporated and the product was purified by flash chroma-
tography (silica gel 40–63 mm, pentane/diethyl ether).
ductive binding also explains enantioselectivity of
human AMP, dTMP and UMP-CMP kinases.[29]
Conclusions
N-Benzyl-2-bromopropionamide (1a): yield: 99%; white
needles; Rf =0.30 (pentane/diethyl ether, 1:1, v/v); mp. 94.1–
94.38C (lit.[30] 93.5–94.58C); 1H NMR (400 MHz, CDCl3):
In this paper we report highly enantioselective trans-
formations of a-bromoamides using haloalkane deha-
logenases. The kinetic resolution of a-bromoamides
3
d=1.92 (d, J=6.8 Hz, 3H, CH3), 4.54–4.82 (m, 3H, CHBr,
CH2Ph), 6.66 (br s, 1H, NH), 7.27–7.36 (m, 5H, ArH),
was shown to reach E values of over 200. C-Protected 1H NMR consistent with literature data;[31] 13C NMR
(50 MHz, CDCl3): d=23.5, 44.4, 45.5, 127.9, 128.0, 129.1,
137.7, 169.4; chiral HPLC analysis (Chiralcel AD, heptane/i-
PrOH, 95/5, v/v): retention times (min): 11.3 (R) and 12.0
(S).
N-Phenyl-2-bromopropionamide (1b): purchased from
Sigma–Aldrich; chiral HPLC analysis (Chiralcel ADH, hep-
tane/i-PrOH, 95/5, v/v): retention times (min): 24.0 (R) and
31.1 (S).
dipeptide precursors (1f–1h) are converted with high
stereoselectivity by either DbjA or LinB, resulting in
enantiopure (ee or de>95%) precursors of peptido-
mimetics. In general, the enzymes show broader spe-
cificity with different N-substituents (R’) (as smaller,
aliphatic, hydrophilic groups are accepted) than with
Ca-substituents (R), where methyl is the universal
group and only one compound with an ethyl group is
converted (1k). Conversions on a preparative scale
N-Butyl-2-bromopropionamide (1c): yield: 96%; colorless
oil; Rf =0.35 (pentane/diethyl ether. 1:1, v/v); 1H NMR
3
with usually less than 2 wt% of enzyme over substrate (300 MHz, CDCl3): d=0.95 (t, J=7.2 Hz, 3H, CH3), 1.34–
3
1.59 (m, 4H, CH2), 1.90 (d, J=7.5 Hz, 3H, CH3), 3.29 (m,
2H, CH2), 4.43 (q, 3J=6.9 Hz, 2H, CH2) 6.43 (br, s, 1H,
NH), NMR consistent with literature data;[32] 13C NMR
(CDCl3): d=14.0, 20.2, 23.6, 31.6, 40.2, 46.0, 169.4; chiral
HPLC analysis (Chiralcel AD, heptane/i-PrOH, 98/2, v/v):
retention times (min): 10.4 (R) and 11.5 (S).
gave high yields for both a-bromoamides and a-hy-
droxyamides, within 1–2 days. The results presented
here add a new class of biocatalytic reactions to the
repertoire of highly enantioselective biotransforma-
tions suitable for synthesis.
The molecular dynamics simulations of the LinB-
substrate complexes indicate that differences in dy-
namic behavior of bound substrate, resulting in an un-
favourable angle for nucleophilic attack by Asp108 in
N-tert-Butyl-2-bromopropionamide (1d): yield: 26%;
white powder; Rf =0.54 (pentane/diethyl ether. 1:1, v/v); mp
1
119–1208C (lit.[33] 119–1218C); H NMR (400 MHz, CDCl3):
3
d=1.36 (s, 9H, 3CH3), 1.83 (d, J=7.2 Hz, 3H, CH3), 4.30
3
case of the non-preferred (S)-enantiomers of 1f and (q, J=7.2 Hz, 1H, CHBr), 6.16 (br, s, 1H, NH).
938
asc.wiley-vch.de
ꢀ 2011 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Adv. Synth. Catal. 2011, 353, 931 – 944

Kinetic Resolution of a-Bromoamides: Experimental and Theoretical Investigation
Ethyl 2-(2-bromopropanamido)acetate (1e): yield: 71%;
white crystals; Rf =0.20 (pentane/diethyl ether. 1:1, v/v); mp
54–558C; 1H NMR (400 MHz, CDCl3): d=1.30 (t, 3J=
7.2 Hz, 3H, CH3), 1.90 (d, 3J=7.2 Hz, 3H, CH3), 4.05 (d,
3J=4.8 Hz, 2H, CH2N), 4.24 (q, 3J=7.6 Hz, 2H, CH2O),
4.45 (q, 3J=6.8 Hz, 1H, CHBr), 6.84 (br, s, 1H, NH);
13C NMR (CDCl3): d=14.1, 23.0, 41.9, 44.3, 61.7, 169.3,
169.4; HR-MS (ESI+): m/z=259.9892, calcd. for
C7H12NO3Na: 259.9893; HPLC (Chiralcel AD, heptane/i-
PrOH, 9/1, v/v): retention times (min): 11.1 and 12.0 (abso-
lute configuration not assigned).
168.9; HR-MS (ESI+): m/z=302.0353, calcd. for
C10H18NO3BrNa: 302.0368.
tert-Butyl
2-[(2-bromobutanoyl)amino]-3-phenylpropa-
noate (1m): yield: 68%; clear oil; dr=1:1, the diastereoiso-
meric ratio was determined by 1H NMR spectroscopy, by
comparison of CH2CH3 signal (t, 0.98 ppm and 1.02 ppm,
relative configuration not determined); HR-MS (ESI+):
m/z=392.0816, calcd. for C17H24BrNO3Na: 392.0837.
N-Benzyl-2-bromohexanamide (1n): yield: 93%; white
solid; Rf =0.57 (pentane/diethyl ether. 1:1, v/v); mp 46–
1
3
478C; H NMR (400 MHz, CDCl3): d=0.92 (t, J=6.8 Hz,
3H, CH3), 1.32–1.56 (m, 4H, CH3CH2CH2), 1.98–2.22 (m,
2H, CH2CHBr), 4.36 (dd, 3J=8.0 Hz, 3J=4.8 Hz, 1H,
CHBr), 4.42–4.53 (m, 2H, CH2NH), 6.70 (br, s, 1H, NH),
7.26–7.38 (m, 5H, Ar); 13C NMR (CDCl3, 100 MHz): d=
14.1, 22.2, 29.6, 36.0, 44.4, 52.4, 127.9, 127.9, 129.0, 137.8,
168.9; HR-MS (ESI+): m/z=306.0454, calcd. for
C13H18BrNONa: 306.0469.
tert-Butyl 2-(2-bromopropanamido)acetate (1f): yield:
91%; white powder; Rf =0.40 (pentane/diethyl ether. 1:1,
1
v/v); mp 73–758C; H NMR (400 MHz, CDCl3): d=1.48 (s,
3
9H, 3CH3), 1.89 (d, J=7.2 Hz, 1H, CHBr), 3.95 (m, 2H,
3
CH2), 4.43 (q, J=6.8 Hz, 1H, CH2), 6.85 (br, s, 1H, NH);
13C NMR (CDCl3): d=23.3, 28.3, 42.8, 44.7, 83.0, 168.7,
169.5; HR-MS (ESI+): m/z=288.0208, calcd. for
C9H16NO3BrNa: 288.0206; chiral HPLC analysis (Chiralcel
OD, heptane/i-PrOH, 95/5, v/v): retention times (min): 7.4
(R) and 9.2 (S).
Catalytic Activity with Compounds 1a–1n
Catalytic activity was measured using a halide release assay
in which the optical absorptions were measured at 460 nm.[9]
Assays were carried out in 200 mL incubations, with 1–
1.5 mM substrate, in 50 mM Tris·SO4, pH 8.2 at room tem-
perature, using 15–35 mg enzyme.
tert-Butyl 2-[(2-bromopropanoyl)amino]-propanoate (1g):
yield: 94%; white solid; dr=1:1, the diastereoisomeric ratio
1
was determined by H NMR spectroscopy, by comparison of
the CH(Br)CH3 signal [d, 1.87 ppm for (S,S), 1.89 ppm for
(R,S)].
tert-Butyl 2-[(2-bromopropanoyl)amino]-3-phenylpropa-
noate (1h): yield: 85%; white solid; dr=1:1, the diastereo-
isomeric ratio was determined by H NMR spectroscopy, by
comparison of the CH(Br)CH3 signal [d, 1.81 ppm for (S,S),
1.85 ppm for (R,S)].
Stereoselectivity with Compounds 1a–1n
1
Typically 120–160 mg enzyme were added to a 5 mL solution
of racemic a-bromoamide (1a–1n, 1–1.5 mM) in 50 mM
Tris.SO4, pH 8.2, and incubated at 308C.[20] Depending on
the catalytic activity, 2 mL samples were taken, after 1–16 h,
extracted with 1.5 mL of diethyl ether or ethyl acetate and
dried. Solvents were evaporated and enantiomeric/diastereo-
meric excess of a-bromoamide and a-hydroxyamide were
determined using Chiral HPLC (compounds 1a–1c, 1e, 1f,
1k) or NMR (compounds 1g and 1h). The E values and con-
versions were calculated by substituting the enantiomeric
excess of substrate and product in the equations for enantio-
meric ratio.[14]
tert-Butyl
2-[(2-bromopropanoyl)amino]-3-methylbuta-
noate (1i): yield: 98%; clear oil; dr=1:1,the diastereoisomer-
1
ic ratio was determined by H NMR spectroscopy, by com-
parison of the CH(Br)CH3 signal (d, 1.87 ppm and
1.89 ppm, relative configuration not determined); HR-MS
(ESI+): m/z=330.0664, calcd. for C12H22NO3BrNa:
330.0681.
tert-Butyl 4-amino-2-[(2-bromopropanoyl)amino]-4-oxo-
butanoate (1j): yield: 71%; white solid; dr=1:1, the diaste-
1
reoisomeric ratio was determined by H NMR spectroscopy,
Preparative Kinetic Resolution Experiments
by comparison of CH(Br)CH3 signal (d, 1.85 ppm and
1.87 ppm, relative configuration not determined).
Kinetic resolution of compound 1a: DbjA (1.4 mg) was
added to a solution of N-benzyl-2-bromopropionamide (1a,
0.34 mmol, 81 mg) in 250 mL of 50 mM Tris·SO4 buffer,
pH 8.2, and incubated at 308C. After 46 h, the reaction mix-
ture was extracted with diethyl ether (2ꢂ150 mL) and ethyl
acetate (1ꢂ150 mL), dried with MgSO4 and solvent was
evaporated. The products were purified by flash chromatog-
raphy (silica gel 40–63 mm).
N-Benzyl-2-bromobutanamide (1k): yield: 99%; white
powder; Rf =0.25 (pentane/diethyl ether. 1:1, v/v); mp 75–
1
3
768C; H NMR (400 MHz, CDCl3): d=1.05 (t, J=7.2 Hz,
3H, CH3), 2.06–2.21 (m, 2H, CH2CHBr), 4.34 (dd, 3J=
3
7.6 Hz, J=4.8 Hz, 1H, CHBr), 4.44–4.52 (m, 2H, CH2NH),
6.77 (br, s, 1H, NH), 7.26–7.37 (m, 5H, Ar); 13C NMR
(CDCl3, 50 MHz): d=12.0, 29.6, 44.3, 53.9, 127.9, 127.9,
129.0, 137.8, 168.8; HR-MS (ESI+): m/z=278.0141, calcd.
for C11H14NOBrNa: 278.0151; chiral HPLC analysis (Chiral-
cel OD, heptane/i-PrOH, 95/5, v/v): retention times (min):
13.8 (R) and 15.2 (S).
(S)-N-Benzyl-2-bromopropionamide [(S)-1a]: eluted with
pentane/diethyl ether, 6:4, v/v; yield: 41% (33 mg,
0.14 mmol); white crystals; [a]2D0: ꢀ1.2 (c 1.0, CHCl3), lit.[34]
ꢀ1.2; enantiomeric excess: >99% (HPLC).
tert-Butyl 2-(2-bromobutananamido)acetate (1l): yield:
96%; white crystals; Rf =0.54 (pentane/diethyl ether. 1:1,
(S)-N-Benzyl-2-hydroxypropionamide [(S)-2a]: eluted
with pentane/ethyl acetate, 3:7, v/v; yield: 50% (30 mg,
0.17 mmol); colorless oil; Rf =0.12 (pentane/ethyl acetate,
4:6, v/v); [a]D20: ꢀ10.0 (c 1.0, CHCl3) lit.[35] ꢀ6.8 (c 0.69,
CHCl3); 1H NMR (400 MHz, CDCl3): d=1.42 (d, 3J=
1
v/v); mp 50–518C; H NMR (400 MHz, CDCl3): d=1.06 (t,
3J=7.2 Hz, 3H, CH3), 1.48 [s, 9H, (CH3)3], 2.01–2.23 (m,
3
2H, CH3CH2), 3.95 (d, 2H, J=5.2 Hz, CH2NH), 4.31 (dd,
3J=7.2 Hz, 3J=5.2 Hz, 1H, CHBr), 6.87 (br, s, 1H, NH);
13C NMR (CDCl3): d=11.9, 28.3, 29.5, 42.8, 53.3, 82.9, 168.7,
3
6.8 Hz, 3H, CH3), 3.25 (d, J=4.8 Hz, 1H, OH), 4.21–4.28
(m, 1H, CHOH), 4.42–4.44 (m, 2H, CH2Ph), 6.97 (br s, 1H,
Adv. Synth. Catal. 2011, 353, 931 – 944
ꢀ 2011 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
asc.wiley-vch.de
939

Alja Westerbeek et al.
FULL PAPERS
1
NH), 7.24–7.37 (m, 5H, ArH), H NMR consistent with lit-
erature data;[35] chiral HPLC analysis (Chiralcel AD, hep-
tane/i-PrOH, 95/5, v/v): retention times (min): 11.3 (R) and
12.2 (S); enantiomeric excess: >99% (HPLC).
buffer, pH 8.2, and incubated at 308C. After 20 h, the reac-
tion mixture was extracted with ethyl acetate (3ꢂ150 mL),
dried with MgSO4 and solvent was evaporated. The products
were purified by flash chromatography (silica gel 40–63 mm).
(S)-tert-Butyl 2-(2-bromopropanamido)acetate [(S)-1f]:
eluted with pentane/diethyl ether, 6:4, v/v; yield: 44%
(35 mg, 0.13 mmol); [a]2D0: ꢀ12.2 (c 1.0, CHCl3); enantiomer-
ic excess: 98% (HPLC).
Kinetic resolution of compound 1b: DbjA (0.70 mg) was
added to a solution of N-phenyl-2-bromopropionamide (1b,
0.38 mmol, 87 mg) in 325 mL of 50 mM Tris·SO4 buffer,
pH 8.2, and incubated at 308C. After 17 h, the reaction mix-
ture was extracted with ethyl acetate (3ꢂ150 mL), dried
with MgSO4 and the solvent was evaporated. The products
were purified by flash chromatography (silica gel 40–63 mm).
(S)-N-Phenyl-2-bromopropionamide [(S)-1b]: eluted with
pentane/diethyl ether, 7:3, v/v; yield: 45% (38 mg,
0.17 mmol); white crystals; [a]2D0: ꢀ29.0 (c 1.0, CHCl3) lit.[36]
ꢀ33.0; enantiomeric excess: >99% (HPLC); Rf =0.49 (pen-
tane/diethyl ether, 1:1, v/v); 1H NMR (300 MHz, CDCl3):
d=1.98 (d, J=5.1 Hz, 3H, CH3), 4.55 (q, J=5.7 Hz, 1H,
CHBr), 7.16–7.55 (m, 5H, ArH), 8.05 (br, 1H, CONH),
1H NMR consistent with literature data;[37] chiral HPLC
analysis (Chiralcel ADH, heptane/i-PrOH, 95/5, v/v): reten-
tion times (min): 24.0 (R) and 31.1 (S).
(S)-tert-Butyl 2-(2-hydroxypropanamido)acetate [(S)-2f]:
eluted with ethyl acetate; yield: 40% (24 mg, 0.12 mmol);
colorless oil; [a]2D0: ꢀ21.0 (c 1.0, CHCl3); enantiomeric
excess: >99.5% (HPLC); Rf =0.15 (pentane/ethyl acetate,
4:6, v/v); 1H NMR (400 MHz, CDCl3): d=1.42 (d, 3J=
6.8 Hz, 3H, CH3), 1.46 [s, 9H, (CH3)3] 3.87–4.01 (m, 3H,
3
CH2NH2, OH), 4.27 (q, J=6.8 Hz, 1H, CHOH), 7.14 (br s,
1H, NH); 13C NMR (100 MHz, CDCl3): d=21.2, 28.2, 41.8,
68.6, 82.7, 169.3, 175.4; HR-MS (ESI+): m/z=226.1053,
calcd. for C9H17NO4Na: 226.1050; chiral HPLC analysis
(Chiralcel OD, heptane/i-PrOH, 95/5, v/v): retention times
(min): 11.3 (S) and 12.6 (R). The configuration was deter-
mined by comparison of NMR data to the spectrum ob-
tained for a reference compound, which was synthesized
from l-lactic acid and alanine tert-butyl ester hydrochloride
by EDC/HOBt mediated coupling.
3
3
(S)-N-Phenyl-2-hydroxypropionamide [(S)-2b]: eluted
with pentane/diethyl ether, 4:6, v/v; yield: 46% (29 mg,
0.18 mmol); colorless oil; [a]2D0: ꢀ14.2 (c 1.0, CHCl3) lit.[38]
ꢀ15.6, enantiomeric excess: >99% (HPLC); Rf =0.12 (pen-
tane/diethyl ether, 1:1, v/v); 1H NMR (400 MHz, CDCl3):
Kinetic resolution of compound 1g: LinB (1.2 mg) was
added to a solution of tert-butyl 2-[(2-bromopropanoyl)ami-
no]propanoate (1g, 0.29 mmol, 80 mg) in 250 mL of 50 mM
Tris·SO4 buffer, pH 8.2, and incubated at 308C. After 43 h,
the reaction mixture was extracted with ethyl acetate (3ꢂ
150 mL), dried with MgSO4 and solvent was evaporated.
3
d=1.52 (d, J=6.4 Hz, 3H, CH3), 3.22 (br s, 1H, OH), 4.35
(q, J=6.4 Hz, 1H, CHOH), 7.12 (t, J=8.0 Hz, 1H, ArH),
3
3
3
3
7.32 (app t, J=8.0 Hz, 2H, ArH), 7.55 (d, J=8.0 Hz, 2H,
ArH), 8.51 (br s, 1H, NH); 13C NMR (CDCl3): d=21.4,
69.1, 120.0, 124.8, 129.3, 137.4, 172.8, NMR is consistent
with literature data;[39] chiral HPLC analysis (Chiralcel
ADH, heptane/i-PrOH, 95/5, v/v): retention times (min):
27.7 (R) and 35.5 (S).
Kinetic resolution of compound 1c: LinB (0.72 mg) was
added to a solution of N-butyl-2-bromopropionamide (1c,
0.38 mmol, 79 mg) in 250 mL of 50 mM Tris·SO4 buffer,
pH 8.2, and incubated at 308C. After 20 h, the reaction mix-
ture was extracted with ethyl acetate (3ꢂ150 mL), dried
with MgSO4 and the solvent was evaporated. The products
were purified by flash chromatography (silica gel 40–63 mm).
(S)-N-Butyl-2-bromopropionamide [(S)-1c]: eluted with
pentane/diethyl ether, 1:1, v/v; yield: 47% (37 mg,
0.18 mmol); colorless oil; [a]2D0: ꢀ16.4 (c 1.0, CHCl3) lit.[32]
ꢀ10.2 (c 0.8, CHCl3, for ee=90%); enantiomeric excess:
>99% (HPLC).
AHCTUNGTREG(NNUN S,S)-tert-Butyl 2-[(2-bromopropanoyl)amino]propanoate
[(S,S)-1g]: eluted with pentane/diethyl ether, 8:2, v/v; yield:
38% (30 mg, 0.11 mmol); white solid; diastereoisomeric
1
excess: >95% (NMR); [a]2D0: ꢀ1.0 (c 1.0, CHCl3); H NMR
3
(400 MHz, CDCl3): d=1.39 (d, J=6.8 Hz, 3H, CH3), 1.47
3
[s, 9H, C
G
(m, 2H, CHBr, CHNH), 6.91 (br s NH); 13C NMR (CDCl3):
d=18.5, 23.2, 28.2, 44.8, 49.5, 82.6, 168.9, 172.0; HR-MS
(ESI+): m/z=302.0380, calcd. for C10H18NO3BrNa:
302.0367.
A
2-[(2-hydroxypropanoyl)amino]propa-
noate [(S,S)-2g]: eluted with diethyl ether; yield: 32%
(20 mg, 0.09 mmol); clear oil; fiastereoisomeric excess:
>95% (NMR); Rf =0.51 (pentane/ethylacetate. 4:6, v/v);
[a]2D0: ꢀ25.8 (c 1.0, CHCl3); mp 137–1388C. 1H NMR
3
(400 MHz, CDCl3): d=1.39 (d, J=6.8 Hz, 3H, CH3), 1.43
3
3
(S)-N-Butyl-2-hydroxypropionamide [(S)-2c]: eluted with
ethyl acetate, yield: 47% (26 mg, 0.18 mmol); colorless oil;
[a]2D0: ꢀ14.2 (c 1.0, CHCl3) lit. ꢀ16.2;[40] enantiomeric excess:
87% (HPLC); Rf =0.18 (pentane/ethyl acetate, 4:6, v/v);
1H NMR (300 MHz, CDCl3): d=0.93 (t, 3J=7.2 Hz, 3H,
(d, J=6.8 Hz, 3H, CH3), 1.47 [s, 9H, CACHTNGUTERNU(NG CH3)3] 2.75 (d, J=
4.4 Hz, 1H, OH), 4.20–4.27 (m, 1H, CHNH), 4.43–4.50 (m,
1H, CHOH), 6.94 (br s, 1H, NH); 13C NMR (CDCl3): d=
18.8, 21.3, 28.2, 48.5, 68.5, 82.4, 172.5, 174.4; HR-MS (ESI+
): m/z=240.1198, calcd. for C10H19NO4Na: 240.1212. Diaste-
3
1
CH3CH2), 1.30–1.56 (m, 4H, CH2CH2), 1.43 (d, J=6.6 Hz,
3H, CH3), 2.94 (d, J=4.2 Hz, 1H, OH), 3.25–3.32 (m, 2H,
reoisomeric excess was determined using H NMR spectros-
3
copy, by the comparison of CH(OH)CH3 signal [d, 1.39 ppm
for (S,S), 1.41 ppm for (R,S)]. The configuration was deter-
mined by comparison of NMR data to the spectrum ob-
tained for reference compound, which was synthesized from
l-lactic acid and alanine tert-butyl ester hydrochloride by
EDC/HOBt mediated coupling.
Kinetic resolution of compound 1h: LinB (0.72 mg) was
added to a solution of tert-butyl 2-[(2-bromopropanoyl)ami-
no]-3-phenylpropanoate (1h, 0.043 mmol, 15 mg) in 50 mL
of 50 mM Tris·SO4 buffer, pH 8.2, and incubated at 308C.
CH2NH2), 4.20–4.24 (m, 1H, CHOH), 6.55 (br s, 1H, NH);
13C NMR (50 MHz, CDCl3): d=14.0, 20.3, 21.6, 31.8, 39.1,
68.6, 174.5; HRMS (ESI+): m/z=168.0991, calcd. for
C7H15NO2Na: 168.0995; chiral HPLC analysis (Chiralcel
AD, heptane/i-PrOH, 98/2, v/v): retention times (min): 25.0
(R) and 27.9 (S).
Kinetic resolution of compound 1f: DbjA (0.84 mg) was
added to a solution of tert-butyl 2-(2-bromopropionamido)-
ACHTUNGTRENNUNGacetate (1f, 0.30 mmol, 80 mg) in 200 mL of 50 mM Tris·SO4
940
asc.wiley-vch.de
ꢀ 2011 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Adv. Synth. Catal. 2011, 353, 931 – 944

Kinetic Resolution of a-Bromoamides: Experimental and Theoretical Investigation
3
3
After 20 h, the reaction mixture was extracted with ethyl
acetate (3ꢂ25 mL), dried with MgSO4 and solvent was
evaporated.
(d, J=7.2 Hz, 3H, CH3), 2.36 (br, NH), 4.06 (t, J=4.8 Hz,
3
3
2H, CH2), 4.23 (q, J=7.5 Hz, 2H, CH2), 4.31 (q, J=6.9 Hz
1H, CHOH), 7.02 (br, OH); 13C NMR: d=14.4, 21.3, 41.2,
61.9, 68.7, 170.2, 175.3; HR-MS (ESI+): m/z=198.0733,
calcd. for C7H13NO4Na: 198.0737; HPLC (Chiralcel AD,
heptane/i-PrOH, 95/5, v/v): retention times (min)=16.6,
18.0 (absolute configuration not assigned).
ACHTUNGTRENNUNG(S,S)-tert-Butyl 2-[(2-bromopropanoyl)amino]-3-phenyl-
propanoate [(S,S)-1h]: eluted with pentane/diethyl ether,
8:2, v/v; yield: 39% (6 mg, 0.017 mmol); white solid; diaste-
reoisomeric excess: >95% (NMR); [a]2D0: +39.5 (c 0.4,
1
CHCl3); mp 66–678C; H NMR (400 MHz, CDCl3): d=1.43
3
[s, 9H, C
G
Conversion of N-Benzyl-2-bromopropionamide (1a)
with Dimethyl Sulfoxide
3
(m, 2H, CH2Ar), 4.36 (q, J=6.4 Hz, 1H, CHBr), 4.65–4.75
(m, CHNH), 6.77 (br s, NH), 7.17–7.31 (m, 5H, ArH);
13C NMR (CDCl3): d=22.9, 28.0, 37.8, 44.5, 53.9, 82.8, 127.1,
128.4, 129.6, 135.7, 168.6, 170.1; HR-MS (ESI+): m/z=
378.0658, calcd. for C16H22NO3BrNa: 378.0681.
Activity of DbjA in a solvent mixture of buffer and dimeth-
yl sulfoxide: DbjA (160 mg) was added to 5 mL of 1.3 mM
solution of 1a in a mixture of 50 mM Tris·SO4, and DMSO
(0–50% v/v), pH 8.2. After 150 min of incubation at 308C, a
sample of 2 mL was extracted with 1.5 mL diethyl ether,
dried, evaporated and analyzed by Chiral HPLC, as de-
scribed before. Conversions were calculated from the enan-
tiomeric excess of both substrate and product.[14]
Preparative scale experiment in 25 vol% dimethyl sulfox-
ide: DbjA (7 mg) was added to a solution of N-benzyl-2-
bromopropionamide (1a, 0.97 mmol, 250 mg) in 240 mL of
50 mM Tris.SO4, pH 8.2 and 25% (v/v) dimethyl sulfoxide.
The mixture was incubated at 308C for 44 h and extracted
with 2ꢂ150 mL diethyl ether and 2ꢂ150 mL ethyl acetate.
The products were purified by flash chromatography as de-
scribed above.
ACHTUNGTRENNUNG(S,S)-tert-Butyl 2-[(2-hydroxypropanoyl)amino]-3-phenyl-
propanoate [(S,S)-2h]: eluted with diethyl ether; yield: 39%
(5 mg, 0.017 mmol); white powder; diastereoisomeric
excess: >95% (NMR); Rf =0.49 (pentane/ethyl acetate. 4:6,
v/v); [a]2D0: +36.8 (c 1.0, CHCl3); mp 66–678C; 1H NMR
3
(400 MHz, CDCl3): d=1.33 (d, J=6.8 Hz, 3H, CH3), 1.40
3
[s, 9H, C
N
6.4 Hz, 1H, CHOH), 4.47 (br s, 1H, OH), 4.73 (m, CHNH),
3
7.08 (d, J=8.0 Hz, 1H, NH); 13C NMR (CDCl3): d=21.3,
28.2, 38.3, 53.3, 68.5, 82.8, 127.2, 128.6, 129.7, 136.1, 170.8,
174.6; HR-MS (ESI+): m/z=294.1687, calcd. for
C16H24NO4: 294.1700. Diastereoisomeric excess was deter-
mined by the comparison of CH(OH)CH3 1H NMR signal
[d, 1.34 ppm for (S,S), 1.37 ppm for (R,S)]. The configura-
tion was determined by comparison of NMR data to the
spectrum obtained for a reference compound, which was
synthesized from l-lactic acid and phenylalanine tert-butyl
ester hydrochloride by EDC/HOBt-mediated coupling.
Kinetic resolution of compound 1k: LinB (1.9 mg) was
added to a solution of N-benzyl-2-bromobutanamide (1k,
0.31 mmol, 80 mg) in 350 mL of 50 mM Tris·SO4 buffer,
pH 8.2, and incubated at 308C. After 44 h, the reaction mix-
ture was extracted with ethyl acetate (3ꢂ200 mL), dried
with MgSO4, and solvent was evaporated.
Time Course Experiment for (S)-tert-Butyl 2-(2-
Bromopropanamido)acetate (1f)
To 30 mL solution of (S)-tert-butyl-2-(2-bromopropanami-
do)acetate (1f, 0.030 mmol, 8.0 mg) LinB (1 or 5 mg) was
added. At various times samples were taken. A sample of
200 mL was mixed with 50 mL 250 mM potassium phosphate
buffer, pH 2 and analyzed by reverse-phase HPLC (Nucleo-
sil 100 (C18) column, 50% acetonitrile and 50% 50 mM po-
tassium phosphate buffer, pH 2, flow 1 mLminꢀ1 [retention
times (min)=8.5 (2f); 29.5 (1f)]. Enantiomeric excesses of
1f and 2f were determined by chiral HPLC as described
above. The obtained time-course (Figure 5B) was fitted to
(S)-N-Benzyl-2-bromobutanamide [(S)-1k]: eluted with
pentane/diethyl ether, 8:2, v/v; yield: 40% (33 mg,
0.13 mmol); white solid; [a]2D0: ꢀ21.0 (c 1.0, CHCl3); enantio-
meric excess: >99% (HPLC).
d[R]/dt=ꢀkcat
R[LinB][R]/{[(1+[S]/KmS)ꢂKMR]+[R]},
ACHTUNGTRENNUNG
d[S]/dt=ꢀkcat
S[LinB][S]/{[(1+[R]/KmR)ꢂKMS]+[S]},[42] and
ACHTUNGTRENNUNG
the E value=(kcatR/KmR)/(kcatS/KmS) using Scientist 2.0 (Mi-
(S)-N-Benzyl-2-hydroxybutanamide [(S)-2k]: eluted with
diethyl ether; yield: 31% (20 mg, 0.10 mmol); white
powder; Rf =0.50 (pentane/ethyl acetate, 4:6, v/v); [a]2D0:
cromath). The reported errors are the standard deviations.
1
ꢀ43.8 (c 1.0, CHCl3), lit.[41] ꢀ42.0 (c 5.0, MeOH); H NMR
General Cloning Strategy
3
(400 MHz, CDCl3): d=0.98 (t, J=7.6 Hz, 3H, CH3), 1.62–
For the creation of the haloalkane dehalogenase constructs,
the genes were cloned into a modified pBADmycHISA
vector, which included an N-terminal hexahistidine tag. In
this vector, the histidine-tag was cloned between NcoI and
NdeI, resulting in the following sequence: 5’-CCATGGGCAG
CAGCCATCATCATCATCATCACAGCAGCGGCCTGGTGCCGCGCG
GCAGCCATATG-3’. The genes of interest, from various sour-
ces, as described in the introduction, were amplified by
PCR, from plasmid DNA (dhlA and dhaA) or from the
original organisms (linB, dbjA and dmlA) and cloned via
NdeI restriction site. The mutant dhaA31 gene was supplied
by J. Damborsky.
1.92 (m, 2H, CH3CH2), 2.90 (s, 1H, OH), 4.08–4.14 (m, 1H,
CHOH), 4.44–4.46 (m, 2H, CH2Ph), 6.88 (br s, 1H, NH),
7.28–7.36 (m, 5H, ArH); 13C NMR (CDCl3): d=9.4, 28.2,
43.4, 73.3, 127.8, 128.0, 129.0, 138.2, 173.8; chiral HPLC
analysis (Chiralcel AD, heptane/i-PrOH, 95/5, v/v): retention
times (min)=14.4 (R) and 19.2 (S); enantiomeric excess:
97% (HPLC).
Racemic samples of products (2) were obtained in EDC/
HOBt-mediated coupling of a-hydroxy acids with amines.
Analytical data for compound 2, of which the respective a-
bromoamides were not converted by the enzymes with suffi-
cient selectivity, are presented below.
Ethyl 2-(2-hydroxypropanamido)acetate (2e): colorless
oil; Rf =0.23 (pentane/ethyl acetate; 4:6, v/v); 1H NMR
3
(300 MHz, CDCl3): d=1.29 (t, J=7.5 Hz, 3H, CH3), 1.46
Adv. Synth. Catal. 2011, 353, 931 – 944
ꢀ 2011 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
asc.wiley-vch.de
941

Alja Westerbeek et al.
FULL PAPERS
Protein Expression, Purification and Analysis
MD Simulations
Precultures were started by inoculating 5 mL of LB-medium
with E. coli TOP10 transformants from 10% v/v glycerol
stocks and incubated overnight at 378C, while shaking. The
next day, 2 L of LB, supplemented with 50 mgmLꢀ1 ampicil-
lin, was inoculated with 1% of an overnight culture and in-
cubated at 378C under vigorous aeration. The enzyme ex-
pression was induced by adding 0.002% w/v l-arabinose at
OD600 ~0.6. Cultures were incubated overnight at 308C,
except for LinB which was incubated at 178C. Then, the
cells were harvested, lysed by sonication, and subsequently
the haloalkane dehalogenases were purified using three in
line his-trap Ni-NTA columns from GE healthcare attached
to an ꢄKTA purifier system. The purified enzymes were
concentrated using an Amicon ultrafiltration cell and stored
in TMG buffer (10 mM Tris·SO4, 1 mM b-mercaptoethanol,
10% v/v glycerol, and pH 7.5). Protein concentrations were
determined by Bradford reagent (Biorad) using bovine
serum albumin as a standard. The purified enzyme was ana-
lyzed using SDS-polyacrylamide gel electrophoresis.
The docked enzyme-substrate complexes were placed in a
rectangular simulation cell with periodic boundary condi-
tions and distances between protein and wall of at least
10 ꢁ. Water (TIP3P) and 13 sodium ions to neutralize the
simulation cell were positioned in the box by Yasara[47c]; this
software was also used for the MD simulations, with the
leap-frog integration scheme. Since the goal of the simula-
tions was to reliably predict the angles and distances in the
active site of the enzyme, the Yamber3 force-field was used;
Yamber3 is an Amber99-derived force-field that was
specifically parameterized to more accurately reproduce
the three-dimensional structure of proteins.[47a,49] Long-range
(>7.86 ꢁ) Coulomb interactions were calculated with a par-
ticle mesh Ewald algorithm[50] with 4th degree B-spline func-
tions. No constraints on bond lengths involving hydrogen
atoms were applied. Prior to the MD simulation, an energy
minimization was carried out to remove clashes and relieve
conformational stress. This energy minimization was ended
when the total energy improved by less than 0.05 kJmolꢀ1
per atom during 400 fs. During the simulations, the time
step was 1.25 fs, with the non-bonded interactions updated
once every two time steps.
Docking
Three-dimensional structures of the substrates (R/S)-1f and
(R/S)-1k were created with Chem3D (CambridgeSoft) and
subsequently geometry-optimized with Yasara (www.yasar-
a.org) by AM1 quantum mechanics[43] in combination with
the implicit solvent model COSMO.[44] Substrate partial
charges for docking and MD simulation were calculated
with the AM1-BCC method[45], which gives accuracy similar
to that of RESP[46], at a much lower computational cost.
Prior to docking the substrates into LinB [PDB file
1MJ5][18], the most likely protonation states (at pH 8.5) of
the different side-chains were predicted.[47] Of the predicted
side-chain protonation states, only one seemed in disagree-
ment with the surrounding H-bonding pattern; this His107
was manually corrected to be singly protonated at its Ne.
The catalytically important residue His272 was singly pro-
tonated at its Nd, as found earlier to be the relevant proto-
nation state for catalysis.[18] In the active site, two crystallo-
graphic waters (HOH 3508 and HOH 3050, these water
molecules are not part of the substrate binding site, both are
important for their H-bonds to active site residues) were
preserved while the halide and other waters in the active
site were manually removed. To minimize the chance of
missing the catalytically most relevant binding mode of the
substrate, 800 separate docking runs were carried out for
every substrate enantiomer with AutoDock 4.[48] This dock-
ing was carried out with the maximum number of energy
evaluations set to 2,500,000. The starting structures for the
MD simulations were chosen based on a geometric ranking
criterion that approximately describes the suitability of its
binding orientation for catalysis. For all docked substrate
orientations the distances d1, d2, and d3 (Scheme 2) were
measured and the sum S=(2ꢂd1)+d2 +d3 was calculated
for each orientation. The single orientation selected for each
enantiomer had the lowest S of all. Dissociation constants
were predicted by AutoDock 4.[19]
The production phase simulation of 5000 ps was preceded
by a slow warming (in steps of 5 K from 5 to 298 K in 30 ps)
and an equilibration phase (1970 ps). The temperature was
controlled by a Berendsen thermostat.[51] During the entire
simulation, snapshots were saved every 5 ps to analyze root
mean square deviation (RMSD) and energy; only those
snapshots obtained after equilibration were used for calcula-
tion of the root mean square fluctuations (RMSF) of the
protein. The RMSF of the crystal structure was calculated
p
from its B-factors via RMSF= (3/8B-factor/p2) [please
note that this equation is an approximation, B-factors are
not fully equivalent to RMSF]. To more densely sample the
catalytically important distances and angles during the simu-
lation, these were recorded every 100 fs during the produc-
tion phase (in total 50000 times). The commonly used alter-
native would be to obtain them from the 1000 snapshots.
For every recorded set of distances and angles it was deter-
mined whether they met all criteria for an NAC. The report-
ed errors were obtained by block averaging.[52]
Supporting Information
As an electronic supplement the following materials are
available: 4 movies of the substrate-enzyme complexes
during the MD simulations, Figure S1 to Figure S5 show
angles and distances during the MD simulation, and Fig-
ure S6 to Figure S8, show the simulation cell energy, the
RMSD from the original crystal structures, and the RMSF
from the MD simulation.
Acknowledgements
We thank Jiri Damborsky (University of Brno, Czech Repub-
lic) for supplying the mutant gene DhaA31, Philip Poole
(University of Reading, United Kingdom) for supplying bac-
terial strain Mesorhizobium loti MAFF303099 and Mark
OꢀBrian (University of New York at Buffalo, USA) for sup-
942
asc.wiley-vch.de
ꢀ 2011 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Adv. Synth. Catal. 2011, 353, 931 – 944

Kinetic Resolution of a-Bromoamides: Experimental and Theoretical Investigation
plying bacterial strain Bradyrhizobium japonicum USDA110.
We thank T. Tiemersma-Wegman for technical assistance
with HPLC and HRMS. This project is financially supported
by the B-Basic partner organisations (www.b-basic.nl)
through B-Basic, a public-private NWO-ACTS program. Part
of this research was funded by an NWO ECHO grant.
[19] R. Huey, G. M. Morris, A. J. Olson, D. S. Goodsell, J.
Comput. Chem. 2007, 28, 1145–1152.
[20] J. P. Schanstra, J. Kingma, D. B. Janssen, J. Biol. Chem.
1996, 271, 14747–14753.
[21] T. C. Bruice, Chem. Rev. 2006, 106, 3119–3139.
[22] a) T. C. Bruice, Acc. Chem. Res. 2002, 35, 139–148;
b) T. C. Bruice, S. J. Benkovic, Biochemistry 2000, 39,
6267–6274.
[23] A. Warshel, P. K. Sharma, M. Kato, Y. Xiang, H. B.
Liu, M. H. M. Olsson, Chem. Rev. 2006, 106, 3210–
3235.
References
[24] S. Hur, K. Kahn, T. C. Bruice, Proc. Natl. Acad. Sci.
USA 2003, 100, 2215–2219.
´
[1] W. Szymanski, M. Zwolinska, R. Ostaszewski, Tetrahe-
dron 2007, 63, 7647–7653.
[25] a) M. Bohac, Y. Nagata, Z. Prokop, M. Prokop, M.
Monincova, M. Tsuda, J. Koca, J. Damborsky. Bio-
chemistry 2002, 41, 14272–14280; b) C. Kennes, F.
Pries, G. H. Krooshof, E. Bokma, J. Kingma, D. B. Jans-
sen, Eur. J. Biochem. 1995, 228, 403–407; c) G. H.
Krooshof, I. S. Ridder, A. W. J. W. Tepper, G. J. Vos,
H. J. Rozeboom, K. H. Kalk, B. W. Dijkstra, D. B. Jans-
sen, Biochemistry 1998, 37, 15013–15023; d) K. Kahn,
T. C. Bruice, J. Phys. Chem. B: 2003, 107, 6876–6885;
e) E. Y. Lau, K. Kahn, P. A. Bash, T. C. Bruice, Proc.
Natl. Acad. Sci. USA 2000, 97, 9937–9942; f) F. C.
Lightstone, Y. J. Zheng, T. C. Bruice, J. Am. Chem. Soc.
1998, 120, 5611–5621; g) J. Damborsky, M. Kuty, M.
Nemec, J. Koca, J. Chem. Inf. Comput. Sci. 1997, 37,
562–568.
[26] A. Cornish-Bowden, Fundamentals of Enzyme Kinetics,
Butterworths, London-Boston, 1979, pp 49–51.
[27] Z. Prokop, M. Monincova, R. Chaloupkova, M.
Kivana, Y. Nagata, D. B. Janssen, J. Damborsky, J. Biol.
Chem. 2003, 278, 45094–45100.
[28] a) B. F. Cook, W. W. Cleland, Enzyme Kinetics and
Mechanism, Garland Science London, New York, 2007,
pp 15; b) A. Fersht, Enzyme Structure and Mechanism,
2nd edn., W. H. Freeman and Co, New York, 1985,
p 102.
[2] P. Marchetti, Tetrahedron Lett. 2003, 44, 4121–4123.
[3] M. Murphy, D. Lynch, M. Schaeffer, M. Kissane, J.
Chopra, E. OꢃBrien, A. Ford, G. Ferguson, A. R. Ma-
guire, Org. Biomol. Chem. 2007, 5, 1228–1241.
[4] W. M. Kazmierski, E. Furfine, A. Spaltenstein, L. L.
Wright, Bioorg. Med. Chem. Lett. 2006, 16, 5226–5230.
[5] F. Dangeli, P. Marchetti, V. Bertolasi, J. Org. Chem.
1995, 60, 4013–4016.
[6] D. B. Janssen, Curr. Opin. Chem. Biol. 2004, 8, 150–
159.
[7] a) G. Stucki, M. Thuer, Environ. Sci. Technol. 1995, 29,
2339–2345; b) G. V. Los, L. P. Encell, M. G. McDou-
gall, D. D. Hartzell, N. Karassina, C. Zimprich, M. G.
Wood, R. Learish, R. F. Ohana, M. Urh, D. Simpson, J.
Mendez, K. Zimmerman, P. Otto, G. Vidugiris, J. Zhu,
A. Darzins, D. H. Klaubert, R. F. Bulleit, K. V. Wood,
ACS Chem. Biol. 2008, 3, 373–382.
[8] a) R. J. Pieters, J. H. Lutje Spelberg, R. M. Kellogg,
D. B. Janssen, Tetrahedron Lett. 2001, 42, 469–471;
b) Z. Prokop, Y. Sato, R. Brezovsky, T. Mozga, R. Cha-
loupkova, T. Koudelakova, P. Jerabek, V. Stepankova,
R. Natsume, J. G. E. van Leeuwen, D. B. Janssen, J. Flo-
rian, Y. Nagata, T. Senda, J. Damborsky, Angew. Chem.
2010, 49, 6111–6115.
[9] S. Keuning, D. B. Janssen, B. Witholt, J. Bacteriol. 1985,
163, 635–639.
[10] A. N. Kulakova, M. J. Larkin, L. A. Kulakov, Microbi-
[29] a) R. M. de Jong, J. J. Tiesinga, A. Villa, L. Tang, D. B.
Janssen, B. W. Dijkstra, J. Am. Chem. Soc. 2005, 127,
13338–13343; b) J. A. Alexandre, B. Roy, D. Topalis, S.
Pochet, C. Pꢅrigaud, D. Deville-Bonne, Nucleic Acids
Res. 2007, 35, 4895–4904.
ology 1997, 143, 109–115.
[11] Y. Nagata, K. Miyauchi, J. Damborsky, K. Manova, A.
Ansorgova, M. Takagi, Appl. Environ. Microbiol. 1997,
63, 3707–3710.
[12] Y. Sato, M. Monincova, R. Chaloupkova, Z. Prokop, Y.
Ohtsubo, K. Minamisawa, M. Tsuda, J. Damborsky, Y.
Nagata, Appl. Environ. Microbiol. 2005, 71, 4372–
4379.
[13] M. Pavlova, M. Klvana, Z. Prokop, R. Chaloupkova, P.
Banas, M. Otyepka, R. C. Wade, M. Tsuda, Y. Nagata,
J. Damborsky, Nat. Chem. Biol. 2009, 5, 727–733.
[14] C. S. Chen, Y. Fujimoto, G. Girdaukas, C. J. Sih, J. Am.
Chem. Soc. 1982, 104, 7294–7299.
[15] K. H. Verschueren, F. Seljee, H. J. Rozeboom, K. H.
Kalk, B. W. Dijkstra, Nature 1993, 363, 693–698.
[16] J. Marek, J. Vevodova, I. K. Smatanova, Y. Nagata,
L. A. Svensson, J. Newman, M. Takagi, J. Damborsky,
Biochemistry 2000, 39, 14082–14086.
[30] S. Kushner, R. I. Cassell, J. Morton, J. H. Williams, J.
Org. Chem. 1951, 16, 1283–1288.
[31] G. Zanotti, F. Filira, A. Delpra, G. Cavicchioni, A. C.
Veronese, F. Dangeli, J. Chem. Soc. Perkin Trans. 1
1980, 2249–2253.
[32] M. Quiros, V. M. Sanchez, R. Brieva, F. Rebolledo, V.
Gotor, Tetrahedron: Asymmetry 1993, 4, 1105–1112.
[33] S. R. Safir, H. Dalalian, W. Fanshawe, K. Cyr, R. Lo-
presti, R. Williams, S. Upham, L. Goldman, S. Kushner,
J. Am. Chem. Soc. 1955, 77, 4840–4844.
[34] F. Dangeli, P. Marchetti, G. Cavicchioni, G. Catelani,
F. M. K. Nejad, Tetrahedron: Asymmetry 1990, 1, 155–
158.
[35] S. N. Savinov, D. J. Austin, Org. Lett. 2002, 4, 1415–
1418.
[17] Y. Nagata, T. Nariya, R. Ohtomo, M. Fukuda, K. Yano,
M. Takagi, J. Bacteriol. 1993, 175, 6403–6410.
[18] A. J. Oakley, M. Klvana, M. Otyepka, Y. Nagata, M. C.
Wilce, J. Damborsky, Biochemistry 2004, 43, 870–878.
[36] F. Dangeli, P. Marchetti, R. Rondanin, V. Bertolasi, J.
Org. Chem. 1996, 61, 1252–1255.
Adv. Synth. Catal. 2011, 353, 931 – 944
ꢀ 2011 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
asc.wiley-vch.de
943

Alja Westerbeek et al.
FULL PAPERS
[37] G. Cavicchioni, P. Scrimin, A. C. Veronese, G. Balboni,
F. Dangeli, J. Chem. Soc. Perkin Trans. 1 1982, 2969–
2972.
[38] F. Maran, J. Am. Chem. Soc. 1993, 115, 6557–6563.
[39] G. Z. Wang, T. Mallat, A. Baiker, Tetrahedron: Asym-
metry 1997, 8, 2133–2140.
[40] S. V. Rogozhin, Y. A. Davidovich, Bull. Acad. Sci.
USSR Div. Chem. Sci. (Engl. Transl.) 1970, 19, 2041.
[41] L. Gajewczyk, A. Zejc, Pol. J. Chem. 1990, 64, 567–
571.
[42] J. H. Lutje Spelberg, R. Rink, R. M. Kellogg, D. B.
Janssen, Tetrahedron: Asymmetry 1998, 9, 459–466.
[43] a) M. J. S. Dewar, E. G. Zoebisch, E. F. Healy, J. J. P.
Stewart, J. Am. Chem. Soc. 1985, 107, 3902–3909;
b) M. J. S. Dewar, E. G. Zoebisch, THEOCHEM – J.
Mol. Struc. 1988, 180, 1–21; c) J. J. P. Stewart, J. Comp.-
Aid. Mol. Des. 1990, 4, 1–45.
[46] C. I. Bayly, P. Cieplak, W. D. Cornell, P. A. Kollman, J.
Phys. Chem. 1993, 97, 10269–10280.
[47] a) E. Krieger, T. Darden, S. B. Nabuurs, A. Finkelstein,
G. Vriend, Proteins Struct. Funct. Bioinf. 2004, 57, 678–
683; b) E. Krieger, G. Koraimann, G. Vriend, Proteins
Struct. Funct. Bioinf. 2002, 47, 393–402; c) E. Krieger,
J. E. Nielsen, C. A. E. M. Spronk, G. Vriend, J. Mol.
Graphics Modell. 2006, 25, 481–486.
[48] G. M. Morris, R. Huey, W. Lindstrom, M. F. Sanner,
R. K. Belew, D. S. Goodsell, A. J. Olson, J. Comput.
Chem. 2009, 30, 2785–2791.
[49] J. M. Wang, P. Cieplak, P. A. Kollman, J. Comput.
Chem. 2000, 21, 1049–1074.
[50] U. Essmann, L. Perera, M. L. Berkowitz, T. Darden, H.
Lee, L. G. Pedersen, J. Chem. Phys. 1995, 103, 8577–
8593.
[51] H. J. C. Berendsen, J. P. M. Postma, W. F. Van Gunste-
ren, A. Dinola, J. R. Haak, J. Chem. Phys. 1984, 81,
3684–3690.
[44] A. Klamt, J. Phys. Chem. 1995, 99, 2224–2235.
[45] A. Jakalian, D. B. Jack, C. I. Bayly, J. Comput. Chem.
2002, 23, 1623–1641.
[52] H. Flyvberg, H. G. Petersen, J. Chem. Phys. 1989, 91,
461–466.
944
asc.wiley-vch.de
ꢀ 2011 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Adv. Synth. Catal. 2011, 353, 931 – 944