Catalytic activity of halohydrin dehalogenases towards spiroepoxides

Article abstract of DOI:10.1039/c2ob25470k

A novel activity of halohydrin dehalogenases towards spiroepoxides has been found. The enzyme from Arthrobacter sp. (HheA) catalysed highly regioselective azidolysis of spiroepoxides containing 5, 6 and 7-membered cycloalkane rings, while the enzyme from Agrobacterium radiobacter (HheC), besides high regioselectivity, also displayed moderate to high enantioselectivity (E up to >200) that can be applied for the kinetic resolution of chiral spiroepoxides. The orientations of spiroepoxides in the active site of halohydrin dehalogenases were studied by quantum-chemical calculations and docking simulations. Analyses of the complexes obtained revealed the origins of diastereoselectivity and enantioselectivity of the investigated biotransformations.

Full text of DOI:10.1039/c2ob25470k

View Article Online / Journal Homepage / Table of Contents for this issue
Dynamic Article Links
Organic &
Biomolecular
Chemistry
Cite this: Org. Biomol. Chem., 2012, 10, 5063
www.rsc.org/obc
PAPER
Catalytic activity of halohydrin dehalogenases towards spiroepoxides†
Maja Majerić Elenkov,*a Ines Primožič,b Tomica Hrenar,b Ana Smolko,a Irena Dokli,a
Branka Salopek-Sondia and Lixia Tangc
Received 3rd March 2012, Accepted 24th April 2012
DOI: 10.1039/c2ob25470k
A novel activity of halohydrin dehalogenases towards spiroepoxides has been found. The enzyme from
Arthrobacter sp. (HheA) catalysed highly regioselective azidolysis of spiroepoxides containing 5, 6 and
7-membered cycloalkane rings, while the enzyme from Agrobacterium radiobacter (HheC), besides high
regioselectivity, also displayed moderate to high enantioselectivity (E up to >200) that can be applied for
the kinetic resolution of chiral spiroepoxides. The orientations of spiroepoxides in the active site of
halohydrin dehalogenases were studied by quantum-chemical calculations and docking simulations.
Analyses of the complexes obtained revealed the origins of diastereoselectivity and enantioselectivity of
the investigated biotransformations.
dehalogenases is described. The importance of enantioselective
azidolysis of epoxides stems from the facile conversion of the
Introduction
The ability of enzymes to catalyse the ring-opening of epoxides
has been intensively studied over the past years.1 Epoxide hydro-
lases are popular catalysts for the enantioselective hydrolysis of
racemic epoxides,2 while halohydrin dehalogenases can accept
several anionic nucleophiles, such as halides, azide, cyanide,
cyanate and nitrite, affording a broad range of 1,2-difunctiona-
lised organic compounds.3 Substrates for halohydrin dehalo-
genases are terminal mono- and disubstituted epoxides, while
activity with non-terminal substrates has not been observed until
now.4 In this paper, we describe a newly discovered activity of
halohydrin dehalogenases for the nucleophilic ring-opening of
spiroepoxides. Spiroepoxides are compounds with a three-mem-
bered oxygen-containing heterocycle joined to an alicyclic frag-
ment by a common carbon.5 This scaffold can be found in many
natural products,6 as well as synthetic bioactive compounds.7
To date, several examples of epoxide hydrolase catalysed
hydrolysis of spiroepoxides have been reported.8,9 As far as
racemic substrates are concerned, some kinetic resolutions were
successful and afforded enantiopure products.8,9 Here, for the
first time, azidolysis of spiroepoxides catalysed by halohydrin
1,2-azido alcohol products into valuable, highly enantiomerically
enriched 1,2-amino alcohols in just one step. In order to investi-
gate the nature of the interactions of spiroepoxides with the
enzymes and to predict required structural changes for obtaining
better selectivity, quantum-chemical calculations and docking
simulations were performed as well.
Results and discussion
Regioselective azidolysis of spiroepoxides catalysed by HheA
Studying the various substrates of halohydrin dehalogenase
HheA (an enzyme from Arthrobacter sp. AD2) we noticed
very rapid conversion of spiroepoxide 1b to azido alcohol
2b (Scheme 1). The observed initial enzyme activity
(9.8 μmol min−1 mg−1) was several fold higher compared to the
reaction of monosubstituted aliphatic epoxides, e.g. 1,2-epoxy-
butane.10 To expand a substrate scope of HheA we have chosen
to evaluate spiroepoxides containing 5- and 7-membered
cycloalkane ring (1a and 1c) as well (Scheme 1).
HheA-catalysed transformations of epoxides 1a–1c were
carried out in Tris-SO4 buffer (pH 7.0) containing 0.5% DMSO
to increase the solubility of the substrate, epoxide (50 mM) and
aRuđer Bošković Institute, Bijenička c. 54, 10000 Zagreb, Croatia.
E-mail: majeric@irb.hr; Fax: +385 1 4680-108; Tel: +3851 4560965
bFaculty of Science, University of Zagreb, Zagreb, Croatia.
E-mail: ines.primozic@chem.pmf.hr; Fax: +3851 4606 401;
Tel: +385 1 4606 400
cUniversity of Electronic Science and Technology, No.4, Section 2,
North Jianshe Road, Chengdu, China. E-mail: lixiatang@uestc.edu.cn;
Fax: +8628 83208238; Tel: +86 28 83208232
†Electronic supplementary information (ESI) available: Conditions of
GC analysis and NMR spectra of epoxides and azido alcohols, further
details about quantum-chemical calculations and docking simulations.
See DOI: 10.1039/c2ob25470k
Scheme 1 Azidolysis of spiroepoxides 1a–1c catalysed by HheA.
Org. Biomol. Chem., 2012, 10, 5063–5072 | 5063
This journal is © The Royal Society of Chemistry 2012

View Article Online
Table 1 HheA-catalysed (A) and non-catalysed (B) ring-opening of
epoxides 1a–1c
Table 2 Non-catalysed ring-opening of epoxides 1a–1c: calculated
standard Gibbs energies of activation (B3LYP/6-311++g(d,p) in H2O,
ε = 78.3553) for the transition states of α and β azidolysis
Transition
state
Δ‡G°/
ΔΔ‡G°/
Ratio
(2 : 3)
Comp.
1a
kcal mol−1
kcal mol−1
TSα
TSβ
TSα
TSβ
TSα
TSβ
31.65
31.32
35.11
30.71
33.26
32.18
0.33
4.40
1.08
64 : 36
100 : 0
86 : 14
Substrate Conditionsa t/h
Conv. b/ % Yield / % Ratio (2 : 3)
1b
1ae
1b
1c
A
B
A
B
A
B
1.5 >99
75c
28d
95c
61d
96c
26d
98 : 2
1c
20
2
20
3
100
>99
77
>99
38
59 : 41
100 : 0
97 : 3
99 : 1
57 : 43
20
ring-opening was modelled using reactants and one molecule of
water. The energy barriers, optimized reactant and transition
state structures for the attack on α (TSα) and β (TSβ) epoxide
carbon atoms are presented in Table 2 and in the ESI Fig. S1–
S3.† Reactions proceeding via transition states TSβ for all three
compounds may be considered as a very simple simulation of the
biocatalysed reaction as well, where the water molecule mimics
catalytic triad Tyr hydroxyl group. While the structures of TSβ are
very similar, Gibbs energies of activation for TSα are dependent
on unfavourable steric interactions between the cycloalkane ring
and the approaching nucleophile (Table 2). The calculated barriers
for the attack of azide at Cα and Cβ atom of 1a and 1b were
in a very good agreement with the experimentally determined
regioselectivity (Table 1), whereas in the case of 1c the
calculated differences did not meet completely the observed
regioselectivity (probably due to the greater flexibility of the
7-membered ring).
a Conditions A: Epoxide (1.0 mmol, 50 mM), NaN3 (2.0 eq), 0.5 mg
purified HheA in Tris-SO4 buffer (20 mL, 0.5 M, pH 7.0). Conditions
B:12 Epoxide (1.0 mmol, 0.5 M), NaN3 (5.0 eq) in 2 mL water at 30 °C,
pH 9.5. b Determined by GC. c Yield of 2. d Total yield 2 + 3. e Unstable
compound in water.
NaN3 (100 mM). Reactions were performed with purified
enzyme at room temperature and monitored by GC analysis. In
the control experiment performed in the absence of the enzyme
the non-catalysed formation of azido alcohols was found to be
slow. The stability of the substrates was also checked under the
experimental conditions. Significant chemical hydrolysis has
been observed only in the case of substrate 1a containing
5-membered ring (ca. 15% per hour). All reactions proceeded
with the high enzyme activity and almost complete regioselectiv-
ity. At 1.0 mmol scale quantitative conversions of epoxides were
achieved after 1.5–3 h and azido alcohols 2a–2c were isolated in
high yields (75–96%, Table 1, conditions A). It was observed
that the size of the ring had minor influence on the enzyme
activity. Somewhat lower isolated yield of 2a can be explained
with the above mentioned chemical instability of the starting
epoxide 1a.
A common method for the synthesis of 1,2-azido alcohols is
the ring-opening reaction of epoxides employing NaN3.11 Azido-
lysis of epoxides is typically performed in aqueous medium and
requires relatively long reaction times (12–48 h).12 It was found
that under acidic conditions these reactions can be strongly
accelerated, although with compromised regioselectivity.12 To
evaluate our results of the enzyme-catalysed azidolysis, we have
performed chemical azidolysis of 1a–1c following the
protocol of Fringuelli et al.12 The results are summarized in
Table 1 (conditions B). Comparing the HheA-catalysed (A) and
non-catalysed (B) azidolysis it can be concluded that the bio-
transformation proceeds with the much higher rate and almost
complete regioselectivity. In the case of substrates 1a and 1c, a
minor azido alcohol regioisomer 3a (2%) and 3c (1%) appear
as a result of slow non-catalysed reaction. The high β-regioselec-
tivity of the enzymatic reaction is remarkable and stands in
contrast to the majority of ring-opening reactions involving
spiroepoxides.
ONIOM calculations of HheA-spiroepoxide and azide complexes
To comprehend the high regioselectivity of HheA-catalysed azi-
dolysis and define the important interactions within the active
site, a hybrid approach, combining quantum chemical (DFT) and
semiempirical (PM3MM) calculations within the ONIOM
scheme was chosen. A model of the enzyme active site was built
by selecting the most important amino acids (26 amino acids)
with several atoms’ positions fixed, and placing the substrate
1a–1c and azide ion inside of the model site. The results are pre-
sented in Fig. 1.
Overlay of the obtained geometries showed that the inter-
actions with the amino acids of the active site, are pretty much
alike for all compounds. The azide ion is bound in the halide
binding site where it forms one very strong hydrogen bond with
a water molecule and several weak H-bonds with the backbone
amides in the surroundings. The epoxide ring is positioned in
such a way that the oxygen atom forms two strong hydrogen
bonds with catalytic Ser134 and Tyr147. The atoms of the
hydrophobic aliphatic rings are in close contacts with the resi-
dues of Asn87, Ser135, Leu141, Asn144, Tyr147 (aromatic ring)
and Tyr185. The optimal position of the oxygen atom of the sub-
strate as well as the spiro substituent enforced the position of the
Cβ atom as the closest to the azide nucleophile, thus resulting in
a high regioselectivity of the attack. The observed differences in
the enzyme activity of 1a–1c (Table 1) can be related to the
slight increase in distance of the Cβ atom from the nucleophile
(from 2.64 for 1a to 2.84 Å for 1c).
Regioselectivity of the non-enzyme ring opening reaction
To understand the different regioselectivity ratio in the azidolysis
of tested spiroepoxides 1a–1c without an enzyme, the epoxide
5064 | Org. Biomol. Chem., 2012, 10, 5063–5072
This journal is © The Royal Society of Chemistry 2012

View Article Online
quantitative, a considerable loss of the material during isolation
of 1a–1c occurred due to the low boiling points of the products.
The stereochemistry of sulfoxonium ylide epoxidation of cyclo-
hexanones are known to be controlled in such a manner that
racemic 1-oxaspiro[2.5]octanes 1d–1f were formed almost exclu-
sively as O-axial epimers (>97%, Scheme 2).9
Racemic methyl substituted 1-oxaspiro[2,5]octanes 1d–1f,
were tested in kinetic resolution catalysed by HheA. Enantio-
selectivities were low (E = 1–3, Table 3) which was not surpris-
ing knowing the low enantiodiscrimination of this enzyme.4
Apparently, both enantiomers of 1d–1f can be accommodated in
the HheA active site with less difference in spatial requirements.
Conversion of the substrate 1e was the fastest and the reaction
was slightly more selective compared to the other two chiral sub-
strates. The slow conversion of 1d and 1f can be assumed to be
due to the steric reasons, see Docking studies (vide infra). The
highest activity among all substrates was observed with non-
substituted 1b.
The absolute configuration has been ascertained by comparing
the chiral GC traces of the remaining substrate to standards of
known configuration. For substrates 1d and 1e, a slight prefer-
ence for the (3S)-enantiomer (spiro C-atom) was observed, oppo-
site to the 1f where inverted preference in favour of
(R)-enantiomer is shown.
Fig. 1 Overlay of 1a (white), 1b (grey), 1c (black) and the azide ion in
the HheA active site model represented by 26 selected amino acids
obtained by ONIOM calculations. Interatomic distances are given in Å
and hydrogen atoms are omitted for clarity.
Enantioselective azidolysis of chiral spiroepoxides catalysed by
HheA and HheC
It was shown previously that HheC (an enzyme from Agrobac-
terium radiobacter AD1) is a very powerful and versatile
biocatalyst offering high activity and enantioselectivity.14
According to the known high enantioselectivity of the HheC
towards 2,2-disubstituted epoxides, e.g. 2-methyl-1,2-epoxy-
butane15 low activity of HheC towards cyclic 2,2-disubstituted
epoxides was therefore expected. It has been shown that
(R)-enantiomer of 2-methyl-1,2-epoxybutane fits rather tightly
into the active site pocket with the methyl group pointing into a
small bulge bordered by the amino acids Trp139, Asn176 and
Thr134.15 A bulkier group at this position would lead to steric
conflict with those amino acids and thus the (S)-enantiomer
binds with great difficulties in a productive manner, resulting in
a high enantioselectivity of the process. To our great surprise
HheC not only catalysed azidolysis of spiroepoxides 1b and 1e
but reaction proceeded with the high rate, even higher than the
one catalysed by HheA (Table 3). Azidolyses were observed
with 1d and 1f as well, but both reacted much slower (ca. 40
times slower than 1e).
Investigations were further extended towards chiral spiro-
epoxides. Spiroepoxide substrates with different substitution pat-
terns were prepared. All substrates 1a–1f were prepared from the
corresponding cyclohexanones by using trimethylsulfoxonium
iodide in 30–80% yield.13 Although epoxide formations were
Scheme 2 Synthesis of racemic 1-oxaspiro[2.5]octanes 1d–1f.
Table 3 Results of halohydrin dehalogenase catalysed azidolysis of spiroepoxides 1b, 1d–1fa
Comp.
1b
Enzyme
Initial activityd
t/h
Conv.e/%
ees/%
eep/%
Config.f
E
HheAb
HheCb
HheAc
HheCc
HheAb
HheCb
HheAc
HheCc
9.8
13.9
0.2
0.4
4.8
20.4
0.5
0.5
3
3
23
23
3
1
23
22
60
63
25
31
39
39
57
39
—
—
1
39
27
58
16
63
—
—
3
87
42
90
12
>99
—
—
—
—
1
21
3
34
1.5
>200
1d
3S,4R
3S,4R
3S,5S
3R,5R
R
1e
1f
R
a Conditions: Epoxide (5.0 mM), NaN3 (1 eq), purified enzyme in Tris-SO4 buffer (20 mL, 50 mM, pH 7.0). b 56 μg enzyme. c 150 μg enzyme.
d μmol min−1 mg−1 e The conversion ratio could be determined on the basis of the two enantiomeric excesses [c = ees/(ees + eep)].34 f Faster reacting
epoxide enantiomer.
.
This journal is © The Royal Society of Chemistry 2012
Org. Biomol. Chem., 2012, 10, 5063–5072 | 5065

View Article Online
For all tested spiroepoxides HheC showed much higher
enantioselectivity (E = 21–200), compared to HheA (E = 1–3).
For 4-methyl substituted spiroepoxide 1d moderate enantioselec-
tivity (E = 21) and an enantioselective preference in favour of
(3S,4R)-1d enantiomer has been observed. This result was not in
agreement with previous studies which pointed out that HheC is
(R)-selective (absolute configuration at oxirane ring) for most
substrates.4,15 Geometry of the spiro substrate and conformation-
al flexibility of the 6-membered ring play a crucial role in this
inverted stereopreference as described further in the molecular
modelling and docking study. As mentioned earlier HheC cata-
lysed kinetic resolution of 1e with the highest rate, but also with
moderate enantioselectivity (E = 34) in favor of the (3R,5R)-1e
enantiomer. The same enantiopreference was found in 1f where
an additional axial methyl substituent was introduced on the
cyclohexane ring. This compound was resolved with excellent
selectivity (E > 200), although with significantly reduced
specific activity (Table 3). Small changes in the substrate struc-
ture and the enzyme itself had showed a remarkable influence on
both activity and enantioselectivity.
the unfavourable 1,3-diaxial interactions of the epoxide methyl-
ene group and 5-methyl substituent. Therefore, for the flexible
ligand docking both chair conformers of all tested compounds
were used with the exception of 1e where only largely predomi-
nate O-ax conformer was studied.
Overlay of geometries obtained by docking studies for 1b and
enantiomers of 1d–1f in the active site of HheA are presented in
Fig. 2. Generally, there are two main ways in which an epoxide
oxygen atom can efficiently bind in the active site: I) productive
binding: making two hydrogen bonds with the hydroxyl groups
of Ser134 and Tyr147 (two amino acids of the catalytic triad)
and II) non-productive binding: forming two hydrogen bonds
with the side chain amides of Asn87 and Asn144, as shown for
conformers of 1b (Fig. 2A). Analysis of the geometries obtained
for 1d enantiomers showed productive binding for both confor-
mers of (3S)-1d (O-eq being in a somewhat better position for a
nucleophilic attack), as well as (3R)-1dO-ax (Fig. 2B). In the case
of 1e both enantiomers can bind productively but optimal posi-
tioning of the methyl group resulted in the closer distance of the
Cβ atom of the (3S)-enantiomer to the azide, which is in agree-
ment with the observed enantiopreference (Fig. 2C).
Due to the bigger size (additional methyl group), conformers
of (S)- and (R)-1f bind differently than those of 1d. (R)-1fO-eq
makes strong H-bonds with the amino acids of the catalytic
triad, but Cβ atom is shifted away from the optimal position for
the reaction to occur. In a complex with (S)-1fO-eq, the epoxide
ring is positioned toward Asn87 and Asn144 (Fig. 2D). The
docking studies showed that the topology of the HheA active site
permits the binding of axial and equatorial conformers of
both enantiomers of the investigated compounds, resulting in
low E-values.
Docking studies of 1-oxaspiro[2,5]octane compounds 1b and
1d–1f in the active site of HheA and HheC
To explain enantioselectivities observed in the halohydrin de-
halogenase catalysed azidolysis of racemic spiroepoxides 1d–1f,
molecular docking simulations were performed (AutoDock 4.2
suite of programs).16 A molecular-level interpretation of the
experimentally obtained results is based on the X-ray crystal
structure of halohydrin dehalogenase HheA (PDB code:
1ZMO)17 and HheC (PDB code: 1ZMT).18 In the 1-oxaspiro
[2,5]octane compounds the cyclohexane ring can adopt a confor-
mation in which the position of the epoxide oxygen is pseudo-
axial (O-ax) or pseudo-equatorial (O-eq) (Scheme 3). A rapid
interconversion between the two chair conformers of the cyclo-
hexane ring occur and the position of equilibria is influenced by
the substitution pattern.8
HheC is known to possess a smaller active site compared to
HheA17 therefore different binding modes were expected.
Overlay of geometries obtained by docking studies for confor-
mers of 1b and enantiomers of 1d–1f in the active site of HheC
are presented in Fig. 3. In the active site of HheC, asparagine
side chain amides are missing compared to HheA, thus the best
position for binding the epoxide ring is the one stabilized with
catalytic Ser132 and Tyr145 (Fig. 3A). Additional non-
productive binding is possible through single H-bond stabiliz-
ation with hydroxyl group of Tyr187 found for the O-ax confor-
mer of (3R)-1dO-ax (Fig. S9†). As in HheA complexes, O-ax
confomer of 1b is productively and O-eq non-productively
bound. Although epoxide oxygen atom is stabilized through H-
bonds with the Ser132 and Tyr145, the optimized position of
cyclohexane ring is resulting with Cβ shifting from the azide
ion. Analysis of the geometries for 1d enantiomers showed that
(3S)-1dO-ax binds better then (3R)-1dO-eq which is in agreement
with the experimentally determined unexpected stereopreference
towards (3S)-enantiomer (Fig. 3B). On the other hand,
(3S)-enantiomers of 1e and 1f cannot be easily productively
bound resulting in a greater enantiomeric ratio of that reactions
towards (3R)-enantiomer. In the case of (3R)-1eO-ax, the atoms
of the epoxide and cyclohexane ring are completely overlapped
with those of 1bO-ax (Fig. 3C) and the activity of HheC is even
higher toward this substrate than to non-substituted 1b due to the
additional favorable interactions of the 5-methyl substituent
within the active site. (R)-1fO-ax fits tightly into the active site
and although the most sterically demanding substrate occupies
Conformational analyses were performed for all substrates and
results are presented in Table 4.
DFT calculations for all compounds confirmed that O-ax con-
former prevails. 1e showed the greatest O-ax preference due to
Scheme 3 Two conformers of the cyclohexane ring: oxygen atom in
the pseudo-axial (1bO-ax) and pseudo-equatorial (1bO-eq) position.
Table 4 Conformational analyses of cyclohexane spiroepoxides 1b,
1d–1f (calculated at B3LYP/6-311++g(d,p) level)
Comp.
ΔrGeq–ΔrGax/kcal mol−1
Ratio (O-ax : O-eq)
1b
1d
1e
1f
0.22
0.59
2.53
0.25
59 : 41
73 : 27
99 : 1
60 : 40
5066 | Org. Biomol. Chem., 2012, 10, 5063–5072
This journal is © The Royal Society of Chemistry 2012

View Article Online
Fig. 2 Epoxide-HheA complexes derived from docking studies. The
HheA active site is represented by some structurally important amino
acids. Only polar hydrogen atoms are presented. Interatomic distances
are given in Å. (A) overlay of 1bO-ax (grey) and 1bO-eq (white); (B)
overlay of 1bO-ax (dark yellow), (3R,4S)-1dO-ax (grey) and (3S,4R)-
1dO-eq (white); (C) overlay of 1bO-ax (dark yellow), (3R,5R)-1eO-ax
(stick model) and (3S,5S)-1eO-ax (ball and stick model); (D) overlay of
1bO-ax (dark yellow), (R)-1fO-ax (stick model) and (S)-1fO-ax (ball and
stick model).
Fig. 3 Epoxide-HheC complexes derived from docking studies. The
HheC active site is represented by some structurally important amino
acids. Only polar hydrogen atoms are presented. Interatomic distances
are given in Å. (A) overlay of 1bO-ax (grey) and 1bO-eq (white); (B)
overlay of 1bO-ax (dark yellow), (3R,4S)-1dO-eq (white) and (3S,4R)-
1dO-ax (grey); (C) overlay of (3R,5R)-1eO-ax (stick model), (3S,5S)-1eO-ax
(ball and stick model) and 1bO-ax (dark yellow, not seen because of the
complete overlap with (3R,5R)-1eO-ax); (D) overlay of 1bO-ax (dark
yellow), (R)-1fO-ax (grey), and (S)-1fO-eq (white).
This journal is © The Royal Society of Chemistry 2012
Org. Biomol. Chem., 2012, 10, 5063–5072 | 5067

View Article Online
analogous space as 1bO-ax. On the other hand, the optimal position-
ing of the methyl substituents orient the epoxide ring of (S)-1fO-eq
to a wrong position and thus prevents the enzymatic azidolysis
reaction. Only axial conformers of all tested spiroepoxides can
bind productively in the HheC active site, which results in a higher
enantioselectivity of the reactions compared to the HheA catalyses.
3,3-dimethylcyclohexanone (assay 90%), trimethylsulfoxonium
iodide, absolute DMSO, arabinose, ampicillin sodium salt,
β-mercaptoethanol, sorbitol, and EDTA were purchased from
Sigma-Aldrich. Bacto-tryptone, yeast extract, and bacto-agar
were purchased from Difco. Complete Protease Inhibitor Cock-
tail Tablets were supplied by Roche, while glycerol was obtained
from Kemika.
Conclusions
Production and purification of the enzyme
The results presented in this paper demonstrate that halohydrin
dehalogenase catalysed reactions are simple, convenient and
highly regioselective for preparing variety of β-substituted tert-
azido alcohols (achiral or chiral) starting from spiroepoxides.
Biocatalysts are environmental friendly, inexpensive and can be
easily produced by fermentation. The HheC was confirmed to be
more enantioselective enzyme among halohydrin dehalogenases
that can be applied also for the kinetic resolution of racemic sub-
strates. With the molecular modelling and docking studies the
conformational flexibility and binding of spiroepoxides were
explored and the preferable interactions within the enzymes
active sites determined giving the better insight into the regio-
selectivity and enantioselectivity of both enzymes. To the best
of our knowledge, this is the first study of the enzyme catalysed
azidolysis of spiroepoxides.
Enzymes were prepared according to the modified literature pro-
cedure.19 In brief, two constructs of the plasmid pBAD contain-
ing the gene hheA from Arthrobacter species AD2, and the gene
hheC from Agrobacterium radiobacter AD1, were used for
protein expression of respective halohydrin dehalogenases.
Each of the constructs was transformed into E. coli, strain
MC1061, and grown on Luria broth (LB) media supplemented
with 50 μg mL−1 ampicillin, at 37 °C, over night. Several colo-
nies of respective constructs were inoculated into 50 mL of
Terrific broth media (TB) supplemented with 50 μg mL−1 ampi-
cillin and grown over night at 30 °C with shaking at 250 rpm.
The initial culture was then added into 0.5 L of modified TB
media (containing double phosphate buffer concentration and
0.5 M sorbitol). To induce protein expression, arabinose was
added to final concentration of 0.02%, and the culture was incu-
bated for 72 h at 30 °C. Bacterial cells were pelleted by centrifu-
gation at 5000 g, at 4 °C, for 15 min, and stored at −20 °C until
further protein purification.
Experimental
General methods
In order to purify Hhe proteins, the pellet was resuspended in
TEMG buffer (10 mM Tris-SO4, 1 mM EDTA, 1 mM β-mercapto-
ethanol, 10% glycerol, pH 7.5) containing a protease inhibitor
1H and 13C NMR spectra were recorded on a Bruker AV 300
(1H 300 or 600 MHz and 13C 75 or 150 MHz) spectrometer in
CDCl3. Chemical shifts (δ) are given in ppm downfield from
TMS as the internal standard. Coupling constants are given in
Hz. IR spectra were recorded as thin films on NaCl plates on a
Perkin-Elmer Spectrum Two FT-IR and the peaks reported in
cm−1. Elemental analyses were carried out on a Perkin-Elmer
Series II 2400 CHNS/O analyser. Enzymatic reactions were
monitored by gas chromatography (GC) using β-DEX 225
column. GC analyses was performed on a Hewlett Packard
HP 5890 Series II gas chromatograph equipped with FID detec-
tor (set at 300 °C) and a split injector (set at 250 °C), using N2
as a carrier gas. Determination of enantiomeric excesses (ee’s)
was performed by GC using different chiral columns: β-DEX 225
(30 m × 0.25 mm × 0.25 μm, Supelco), Chiraldex G-TA (30 m ×
0.25 mm × 0.12 μm, Supelco), Lipodex E (25 m × 0.25 mm ×
0.25 μm, Machery–Nagel) and CP Chirasil-DEX CB (25 m ×
0.25 mm × 0.25 μm, Varian). Column chromatography was done
using silica gel (Fluka, 0.063–0.2 mm). TLC was performed on
0.25 mm silica gel 60-F plates (Merck). Spots were visualised
after dipping the TLC plate in a mixture of phosphomolybdic acid
(25 g): cerium(IV)sulphate (7.5 g): sulphuric acid (25 mL): water
(495 mL) and subsequent heating. Kugelrohr distillation was per-
formed using glass oven B-585 instrument, Büchi.
cocktail (Complete Protease Inhibitor Cocktail Tablets), 10 mL g−1
.
Following sonication (10 × 10 s, Labsonic M), extract was cen-
trifuged (10 000 g, 1 h, 4 °C). Supernatant was filtered through
filter (0.45 μm, Millipore) and subjected to a HiPrep Q FF 16/10
– Sepharose anion exchange column (GE Healthcare, Sweden)
connected to the ÄKTAFPLC system (GE Healthcare, Sweden).
The elution of the Hhe proteins was carried out with gradient of
0–0.45 (NH4)2SO4 in TEMG, at flow rate of 1–3 mL min−1. The
collected fractions were screened by SDS-PAGE. Fractions con-
taining purified protein of interest were pooled, concentrated by
using Centriprep YM-10 (Millipore, Ireland), and protein con-
centration was identified according to Bradford.20 The enzymes
were stored at −20 °C until further use.
Synthesis of substrates and reference compounds
General procedure for the synthesis of epoxides. Epoxides
1a–1f and (3R)-1e were prepared according to a literature pro-
cedure.13 DMSO (20 mL) was added dropwise to a mixture of
trimethylsulfoxonium iodide (5.1 g, 23.1 mmol) and sodium
hydride (a 60% disp. in mineral oil; 0.94 g, 23.1 mmol) cooled
to 0 °C, under argon. After stirring for 30 min at room tempera-
ture, a solution of the corresponding ketone (17.8 mmol) in
DMSO (10 mL) was added dropwise. Reaction mixture was
stirred for 20 h. Water was added (20 mL) and the mixture
extracted with diethyl ether (3 × 30 mL). The combined organic
layers were washed with water (2 × 30 mL) and brine (20 mL),
Materials
The commercially available, cyclohexanone and cycloheptanone
were supplied by Merck. Cyclopentanone, 2-methylcyclo-
hexanone, 3-methylcyclohexanone, (R)-3-methylcyclohexanone,
5068 | Org. Biomol. Chem., 2012, 10, 5063–5072
This journal is © The Royal Society of Chemistry 2012

View Article Online
(R,S)-5,5-Dimethyl-1-oxaspiro[2,5]octane (1f). Racemic com-
pound rac-1f was prepared from 3,3-dimethylcyclohexanone
(4.98 g of 90% purity, 35.6 mmol), according to general pro-
cedure. Epoxide was obtained as a colorless liquid. Yield: 3.7 g,
74%; δH (600 MHz; CDCl3; Me4Si) 0.94 (3H, s), 1.00 (3H, s),
1.31–1.33 (3H, m), 1.41 (1H, d, J = 13.5 Hz), 1.46–1.53 (2H,
m), 1.58–1.64 (1H, m), 1.72–1.78 (1H, m), 2.54 (1H, d, J = 5.0
Hz), 2.58 (1H, d, J = 5.0 Hz); δC (150 MHz; CDCl3; Me4Si)
20.9, 28.4, 29.5, 32.7, 33.1, 38.3, 45.9, 53.8, 57.6; νmax(film)/
cm−1 3033, 2947, 2868, 1457, 1365, 1268, 917, 818, 798;
El. an. calcd for C9H16O: C, 77.1; H, 11.5. Found: C, 77.3;
H, 11.4%.
dried over Na2SO4, filtered and evaporated under reduced
pressure.
1-Oxaspiro[2.4]heptane (1a). Compound 1a was prepared from
cyclopentanone (1.5 g, 17.8 mmol), according to general pro-
cedure, and obtained as a colorless liquid after Kugelrohr distil-
lation (70 °C, 23 mmHg). Yield: 0.52 g, 30%; δH (300 MHz;
CDCl3; Me4Si) 1.60–1.74 (4H, m), 1.77–1.95 (4H, m), 2.83
(2H, s); δC (150 MHz; CDCl3; Me4Si) 25.2, 32.4, 52.1, 65.2;
νmax(film)/cm−1 3032, 2961, 2872, 1485, 1438, 1389, 1208,
947, 907; El. an. calcd for C6H10O: C, 73.4; H, 10.3. Found:
C, 73.55; H, 10.6%.
1-Oxaspiro[2.5]octane (1b)21 Compound 1b was prepared
from cyclohexanone (1.75 g, 17.8 mmol), according to the
general procedure. The crude product was first purified by
column chromatography (SiO2; pentane–diethyl ether, 85 : 15)
followed by Kugelrohr distillation (65 °C, 2 mmHg). Epoxide
1b was obtained as a colorless liquid. Yield: 0.84 g, 42%; δH
(300 MHz; CDCl3; Me4Si) 1.48–1.59 (8H, m), 1.69–1.81 (2H,
m), 2.58 (2H, s); δC (75 MHz; CDCl3; Me4Si) 24.9, 25.2, 33.6,
53.3, 54.4. νmax(film)/cm−1 3033, 2933, 2858, 1495, 1484,
1448, 1249, 911, 821.
General procedure for the synthesis of β-azido alcohols.
Azido alcohols were prepared according to a literature procedure
starting from 1.0 mmol of epoxide.12 By following the protocol
at pH 9.5 a mixture of regioisomers 2a/3a and 2c/3c as well as
single 2b, 2d, 2e and 2f were obtained. At pH 4.2 both regio-
isomers 2b/3b were obtained. Column chromatography (SiO2;
n-hexane–ethyl acetate, 9 : 1) furnished the pure azido alcohols.
1-(Azidomethyl)cyclopentan-1-ol (2a). δH (600 MHz; CDCl3;
Me4Si) 1.60–1.70 (6H, m), 1.80–1.85 (2H, m), 1.86 (1H, s),
3.38 (2H, s); δC (150 MHz; CDCl3; Me4Si) 23.5, 37.4, 60.0,
81.6; El. an. calcd for C6H11N3O: C, 51.1; H, 7.9; N 29.8.
Found: C, 49.9; H, 7.85; N, 29.95%.
1-Azido-1-hydroxymethylcyclopentane (3a)22 δH (600 MHz;
CDCl3; Me4Si) 1.68–1.72 (4H, m), 1.77–1.84 (5H, m), 3.58
(2H, s); δC (150 MHz; CDCl3; Me4Si) 24.2, 33.9, 68.4, 73.9.
1-Oxaspiro[2.6]nonane (1c)21 Compound 1c was prepared
from cycloheptanone (1.99 g, 17.8 mmol), according to general
procedure. The crude product was first purified by column
chromatography (SiO2; pentane–diethyl ether, 85 : 15) followed
by Kugelrohr distillation (75 °C, 2 mmHg). Epoxide 1c was
obtained as a colorless liquid. Yield: 1.36 g, 61%; δH (300 MHz;
CDCl3; Me4Si) 1.46–1.78 (12H, m), 2.59 (2H, s); δC (75 MHz;
CDCl3; Me4Si) 24.6, 28.9, 35.9, 55.6, 60.5. νmax(film)/cm−1
3029, 2928, 2856, 1455, 1264, 944, 886.
1-(Azidomethyl)cyclohexan-1-ol (2b). δH (300 MHz; CDCl3;
Me4Si) 1.46–1.62 (10H, m), 1.71 (1H, s), 3.25 (2H, s); δC
(75 MHz; CDCl3; Me4Si) 21.7, 25.6, 35.0, 61.5, 71.6; El. an.
calcd for C7H13N3O: C, 54.2; H, 8.4; N 27.1. Found: C, 54.4;
H, 8.5; N, 27.0%.
(3R,4S),(3S,4R)-4-Methyl-1-oxaspiro[2,5]octane(1d)9 Racemic
compound 1d was prepared from rac-2-methylcyclohexanone
(1.0 g, 8.9 mmol), according to general procedure. The crude
product was first purified by column chromatography (SiO2;
n-hexane–ethyl acetate, 9 : 1). Epoxide 1d was obtained as a col-
orless liquid. Yield: 0.95 g, 85%; δH (300 MHz; CDCl3; Me4Si)
0.83 (3H, d, J = 7.0 Hz), 1.39–1.76 (9H, m), 2.48 (1H, d, J =
5.0 Hz), 2.75 (1H, d, J = 5.0 Hz); δC (75 MHz; CDCl3; Me4Si)
14.3, 24.0, 24.6, 32.6, 32.9, 34.5, 52.5, 60.9. νmax(film)/cm−1
3042, 2930, 2855, 1447, 1369, 1232, 958, 828.
1-Azido-1-hydroxymethylcyclohexane (3b). δH (300 MHz;
CDCl3; Me4Si) 1.29–1.82 (11H, m), 3.55 (2H, s); δC (75 MHz;
CDCl3; Me4Si) 21.8, 25.4, 31.1, 64.7, 69.3; El. an. calcd for
C7H13N3O: C, 54.2; H, 8.4; N 27.1. Found: C, 54.35; H, 8.3;
N, 27.2%.
1-(Azidomethyl)cycloheptan-1-ol (2c)23 δH (300 MHz; CDCl3;
Me4Si) 1.42–1.74 (12H, m), 1.74 (1H, s), 3.26 (2H, s); δC
(75 MHz; CDCl3; Me4Si) 22.2, 29.7, 38.5, 61.7, 75.6.
(3R,5R),(3S,5S)-5-Methyl-1-oxaspiro[2,5]octane (1e)9 Racemic
compound 1e was prepared from rac-3-methylcyclohexanone
(1.0 g, 8.9 mmol), according to general procedure. Epoxide was
obtained as a colorless liquid. Yield: 0.90 g, 80%; δH (300 MHz;
CDCl3; Me4Si) 0.92 (3H, d, J = 6.5 Hz), 0.95–1.03 (1H, m),
1.20–1.29 (2H, m), 1.48 (1H, dd, J1 = 13.5 Hz, J2 = 11.5 Hz),
1.59–1.81 (5H, m), 2.60 (1H, d, J = 4.7 Hz), 2.62 (1H, d, J =
4.7 Hz); δC (75 MHz; CDCl3; Me4Si) 22.2, 23.4, 30.4, 32.6,
33.7, 41.5, 53.9, 58.4. νmax(film)/cm−1 3032, 2977, 2868, 1491,
1457, 1289, 914, 866.
1-Azido-1-hydroxymethylcycloheptane (3c). δH (300 MHz;
CDCl3; Me4Si) 1.45–1.85 (12H, m), 3.50 (2H, d, J = 6 Hz); δC
(150 MHz; CDCl3; Me4Si) 22.5, 29.8, 34.4, 68.5, 69.0; El. an.
calcd for C8H15N3O: C, 56.8; H, 8.9; N 24.8. Found: C, 56.7;
H, 9.0; N, 24.85%.
1-Azidomethyl-2-methylcyclohexan-1-ol (2d). δH (300 MHz;
CDCl3; Me4Si) 0.90 (3H, d, J = 6.5 Hz), 1.21–1.77 (9H, m),
3.21 (1H, d, J = 12.0 Hz), 3.40 (1H, d, J = 12.0 Hz); δC
(75 MHz; CDCl3; Me4Si) 21.7, 22.9, 29.9, 32.1, 33.2, 41.0,
53.4; El. an. calcd for C8H15N3O: C, 56.8; H, 8.9; N 24.8.
Found: C, 56.6; H, 8.85; N, 24.9%.
(3R,5R)-5-Methyl-1-oxaspiro[2,5]octane
((3R,5R)-1e). Refer-
ence compound (3R,5R)-1e was prepared in 80% yield (0.90 g),
starting from (R)-3-methylcyclohexanone (1.0 g, 8.9 mmol). δH
(600 MHz; CDCl3; Me4Si) 0.92 (3H, d, J = 6.5 Hz), 0.93–1.00
(1H, m), 1.22–1.28 (2H, m), 1.48 (1H, dd, J1 = 13.5 Hz, J2 =
11.5 Hz), 1.62–1.67 (1H, m), 1.72–1.82 (4H, m), 2.59 (1H, d,
J = 4.5 Hz), 2.61 (1H, d, J = 4.5 Hz); δC (150 MHz; CDCl3;
Me4Si) 21.7, 22.9, 29.9, 32.1, 33.3, 41.0, 53.4, 57.9; El. an.
calcd for C8H14O: C, 76.1; H, 11.2. Found: C, 76.0; H, 11.2%.
1-Azidomethyl-3-methylcyclohexan-1-ol (2e). δH (300 MHz;
CDCl3; Me4Si) 0.91 (3H, d, J = 6.0 Hz), 1.16–1.27 (1H, m),
1.59–1.81 (8H, m), 3.23 (2H, s); δC (75 MHz; CDCl3; Me4Si)
21.0, 22.4, 27.4, 34.2, 34.5, 43.2, 62.9, 72.1; El. an. calcd for
C8H15N3O: C, 56.8; H, 8.9; N 24.8. Found: C, 56.95; H, 9.0;
N, 24.7%.
This journal is © The Royal Society of Chemistry 2012
Org. Biomol. Chem., 2012, 10, 5063–5072 | 5069

View Article Online
1-Azidomethyl-3,3-dimethylcyclohexan-1-ol (2f). δH (600 MHz;
CDCl3; Me4Si) 0.92 (3H, s), 1.10 (3H, s), 1.13–1.26 (3H, m),
1.44–1.52 (3H, m), 1.63 (1H, s), 1.64–1.67 (1H, m), 1.75–1.82
(1H, m), 3.19 (1H, d, J = 12.0 Hz), 3.21 (1H, d, J = 12.0 Hz);
δC(150 MHz; CDCl3; Me4Si) 18.0, 27.1, 30.5, 33.6, 34.9, 39.0,
46.6, 63.5, 72.6; El. an. calcd for C9H17N3O: C, 59.0; H, 9.4;
N 22.9. Found: C, 58.9; H, 9.3; N, 22.9%.
0.6 mL TMEG buffer). The reaction was performed at room
temperature, followed by GC and stopped at >99% conversion.
The reaction mixture was saturated with NaCl and extracted with
CH2Cl2. The combined organic extracts were dried on filtered
and evaporated. The organic phase was dried over Na2SO4.
The crude product was purified by column chromatography
(SiO2; n-hexane–ethyl acetate, 9 : 1).
1-(Azidomethyl)cyclopentan-1-ol (2a). Obtained as colorless
liquid (107 mg, 75%) after column chromatography. The NMR
data were identical to 2a prepared by standard chemical pro-
cedure. δH (300 MHz; CDCl3; Me4Si) 1.59–1.72 (6H, m),
1.82–1.86 (2H, m), 3.39 (2H, s); δC (75 MHz; CDCl3; Me4Si)
24.0, 37.9, 60.5, 82.1; νmax(film)/cm−1 3401, 2961, 2101, 1284;
El. an. calcd for C6H11N3O: C, 51.1; H, 7.9; N 29.8. Found:
C, 50.0; H, 7.85; N, 29.9%.
Regioselectivity of the ring opening of epoxides 1a–1c with
sodium azide in water. Epoxide (0.5 mmol) was added to 1 mL
of water containing sodium azide (2.5 mmol) at room tempera-
ture. After 20 h reaction mixtures were extracted with CH2Cl2,
extracts were dried and analyzed by GC under conditions
described in ESI.†
Determination of absolute configuration. Absolute configur-
ations were assigned by chiral GC analysis using reference com-
pounds. In the case of epoxide 1d assignment was based on
previously reported data.9 (3R,5R)-1e was prepared from opti-
cally pure (R)-3-methylcyclohexanone (Aldrich). The enantio-
merically enriched epoxide (R)-1f (97% ee) was prepared by
(R,R)-(salen)CrCl catalysed ring opening with TMSN3 according
to Label and Jacobsen.24
1-(Azidomethyl)cyclohexan-1-ol (2b). Obtained as colorless
liquid (147 mg, 95%) after column chromatography. The NMR
data were identical to 2b prepared by standard chemical pro-
cedure. δH (300 MHz; CDCl3; Me4Si) 1.24–1.73 (11H, m), 3.26
(2H, s); δC (75 MHz; CDCl3; Me4Si) 21.7, 25.6, 35.0, 61.5,
71.6; νmax(film)/cm−1 3422, 2934, 2859, 2103, 1447, 1286;
El. an. calcd for C7H13N3O: C, 54.2; H, 8.4; N 27.1. Found:
C, 54.3; H, 8.4; N, 27.0%.
Calculation of enantioselectivity. E-values were calculated
from eep and ees according to formula E = ln[(1 − ees)/(1 + ees/
eep)]/ln[(1 + ees)/(1 + ees/eep)].14
1-(Azidomethyl)cycloheptan-1-ol (2c). Obtained as colorless
liquid (163 mg, 96%) after column chromatography. The NMR
data were identical to 2c prepared by standard chemical pro-
cedure. δH (300 MHz; CDCl3; Me4Si) 1.38–1.73 (13H, m), 3.25
(2H, s); δC (75 MHz; CDCl3; Me4Si) 22.2, 29.7, 38.5, 61.7,
75.6. νmax(film)/cm−1 3428, 2927, 2856, 2102, 1461, 1280;
El. an. calcd for C8H15N3O: C, 56.8; H, 8.9; N 24.8. Found:
C, 56.9; H, 8.95; N, 24.7%.
General procedure for the enzymatic ring-opening of epoxides
on analytical scale
To 20 mL of Tris-SO4 buffer (50 mM, pH 7.0) at room tempera-
ture, 100 μL of a stock solution of substrate in DMSO was
added (final concentration 5.0 mM), followed by addition of a
stock solution in water of NaN3 (1.0 mL, final concentration
5.0 mM). Reactions were initiated by addition of enzyme HheA
or HheC (56–150 μg) in TEMG buffer. The progress of the reac-
tion was followed by periodically taking samples (0.5 mL) from
reaction mixture. Samples were extracted with MTBE (1.0 mL)
containing chlorobenzene as internal standard and analysed by
GC on a β-DEX 225 column. Non-enzymatic ring-opening reac-
tion was determined by monitoring product formation in the
absence of enzyme.
Quantum chemical calculations
Optimization of reactants, transition states and products were
performed at the B3LYP/6-311++g(d,p) level of the theory25,26
using the Gaussian 09 program package.27 For all optimized
structures harmonic frequencies were calculated to insure that
obtained geometries correspond to the local minimum
(or maximum) on the potential energy surface. The Gibbs ener-
gies were calculated at T = 298.15 K and p = 1 atm (Table S6†).
Solvation effects were incorporated in the calculations using the
reformulation of polarizable continuum model (PCM)28,29
known as integral equation formalism (IEFPCM) of Tomasi and
co-workers.30–33 Contents of obtained products were estimated
from the Boltzmann distribution using the differences in Gibbs
energies of activation. For a hybrid approach, combination of
quantum chemical (B3LYP/6-31+g(d)) and semiempirical
(PM3MM) calculations within the ONIOM34 scheme was
chosen. A model for enzyme placed in the low layer was made
of 26 amino acids surrounding the active site. The residues
included in the model were Phe12, Asp80, Arg84, Asn87,
Arg88, Thr133, Ser134, Ser135, Val136, Gly137, Leu141,
Tyr143, Asn144, Tyr147, Arg151, Gly176, Pro177, Asn178,
Phe179, Phe180, Asn182, Thr184, Tyr185, Phe186, Thr239,
Tyr242 and two molecules of water. To ensure the consistency of
the model, several atoms positions were kept fixed to their
Tests for hydrolytic stability of epoxides
Typically, reactions were performed analogous to the procedure
described for the enzymatic ring opening of epoxides without
adding enzyme and nucleophile. To 21 mL of Tris-SO4 buffer
(50 mM, pH 7.0) at room temperature, 100 μL of a stock solu-
tion of substrate in DMSO was added (final concentration
5.0 mM). Concentration of epoxide was followed by periodically
taking samples (0.5 mL) from reaction mixture. Samples were
extracted with MTBE (1.0 mL) containing chlorobenzene as
internal standard and analysed by GC on a β-DEX 225 column.
Enzymatic preparation of β-azido alcohols 2a–2c. The
epoxide (1.0 mmol, 50 mM) was dissolved in a 19.4 mL of Tris-
SO4 buffer (0.5 M, pH 7.0) followed by addition of 130 mg of
NaN3 (2.0 mmol, 100 mM) and purified HheA (0.5 mg in
5070 | Org. Biomol. Chem., 2012, 10, 5063–5072
This journal is © The Royal Society of Chemistry 2012

View Article Online
4 M. Majerić Elenkov, B. Hauer and D. B. Janssen, Adv. Synth. Catal.,
2006, 348, 579–585.
5 L. I. Kas’yan, A. O. Kas’yan and I. N. Tarabara, Russ. J. Org. Chem.,
2001, 37, 1431–1473.
crystallographic position during the optimizations. In high layer
an azide, ligand and one water molecule were placed.
6 T. E. Eble and F. R. Hanson, J. Bacteriol., 1949, 58, 527–529;
H. Hatanaka, T. Kino, M. Hashimoto, Y. Tsurumi, A. Kuroda, H. Tanaka,
T. Goto and M. Okuhara, J. Antibiot., 1988, 41, 999–1008; H. Nakajima,
Y. Hori, H. Terano, M. Okahura, T. Manda, S. Matsumoto and
K. Shimomura, J. Antibiot., 1996, 49, 1204–1211; T. Wakabayashi,
R. Kageyama−Kawase, N. Naruse, Y. Funahashi and K. Yoshimatsu,
J. Antibiot., 2000, 53, 591–596.
7 T. Yanase, M. Tamura, K. Fujita, S. Kodama and K. Tanaka, Cancer Res.,
1993, 53, 2566–2570; J. E. Baldvin, P. G. Bulger and R. Marquez, Tetra-
hedron, 2002, 58, 5441–5452; M. J. Evans, G. M. Morris, J. Wu,
A. J. Olson, E. J. Sorensen and B. F. Cravatt, Mol. BioSyst., 2007, 3,
495–506; C. Arico-Muendel, P. A. Centrella, B. D. Contonio,
B. A. Morgan, G. O’Donovan, C. L. Paradise, S. R. Skinner, B. Sluboski,
J. L. Svendsen, K. F. White, A. Debnath, J. Gut, N. Wilson,
J. H. McKerrow, J. L. DeRisi, P. J. Rosenthal and P. K. Chiang, Bioorg.
Med. Chem. Lett., 2009, 19, 5128–5131.
8 C. A. G. M. Weijers, P. M. Könst, M. C. R. Franssen and
E. J. R. Sudhölter, Org. Biomol. Chem., 2007, 5, 3106–3114;
A.-L. Bottalla, M. Ibrahim-Ouali, M. Santelli, R. Furstoss and
A. Archelas, Adv. Synth. Catal., 2007, 349, 1102–1110; G. Bellucci,
C. Chiappe, G. Ingrosso and C. Rosini, Tetrahedron: Asymmetry, 1995, 6,
1911–1918.
9 C. A. G. M. Weijers, P. Meeuwse, R. L. J. M. Herpers, M. C. R. Franssen
and E. J. R. Sudhölter, J. Org. Chem., 2005, 70, 6639–6646.
10 Unpublished results. HheA catalyses azidolysis of 1,2-epoxybutane with
a rate of 1.6 μmol min−1 mg−1 and 1,2-epoxyhexane with a rate of
1.8 μmol min−1 mg−1 when substrate concentration was 5.0 mM and
1.5 eq NaN3.
Flexible ligand docking
Docking studies were performed by using the AutoDock 4.2.2
suite of programs.16 AutoDock requires a precalculated electro-
static grid map for each atom type present in the substrate mole-
cule. These electrostatic maps were calculated using the
AutoGrid part of the suite with 0.2 Å spacing between grid
points and the centre of the grid was placed in the γ-oxygen
atom of catalytic Ser200. Dimensions of the active site box were
set to 20 Å × 20 Å × 20 Å thus ensuring appropriate size of the
ligand-accessible space. Consistencies of electrostatic maps were
ascertained by checking maximum and minimum values of van
der Waals energies and electrostatic potentials for each calculated
grid map.
Flexible ligand docking for all compounds were carried out
using the Lamarckian genetic algorithm and all parameters were
the same for each docking run. We used initially a population of
5000 random individuals, a maximum number of 2 × 107 energy
evaluations, a maximum number of generations of 1 × 106,
elitism value of 1, mutation rate of 0.02 and crossover rate of
0.08. For the local search, the pseudo-Solis and Wets method
was used with maximum of 10 000 iterations per local search,
the probability of performing a local search on an individual in
the same population was 0.06, the maximum number of consecu-
tive successes or failures before changing the rho was 4 in both
cases, size of local search space to sample was 1.0 and the lower
bound on rho was 0.01. To ensure the validity of results, the
docking procedure for each substrate consisted of 100 indepen-
dent docking runs. The resulting positions were clustered accord-
ing to r.m.s. criterion of 0.5 Å. Obtained structures were
analyzed visually and evaluated based on their interactions with
the amino acids within the active site of the enzyme.
11 B. Das, V. S. Reddy, M. Krishnaiah and Y. K. Rao, J. Mol. Catal. A:
Chem., 2007, 270, 89–92; B. Das, V. S. Reddy, F. Tehseen and
M. Krishnaiah, Synthesis, 2007, 5, 666–668; J. Boruwa, J. C. Borah,
B. Kalita and N. C. Barua, Tetrahedron Lett., 2004, 45, 7355–7358;
N. Iranpoor, H. Firouzabadi and M. Shekarize, Org. Biomol. Chem.,
2003, 1, 724–727; G. Sabitha, R. S. Babu, M. S. K. Reddy and
J. S. Yadav, Synthesis, 2002, 15, 2254–2258; G. Sabitha, R. S. Babu,
M. Rajkumar and J. S. Yadav, Org. Lett., 2002, 4, 343–345.
12 F. Fringuelli, O. Piermatti, F. Pizzo and L. Vaccaro, J. Org. Chem., 1999,
64, 6094–6096.
13 E. J. Corey and M. Chaykovsky, J. Am. Chem. Soc., 1965, 87, 1353–
1364.
14 C.-S. Chen, Y. Fujimoto, G. Girdaukas and C. J. Sih, J. Am. Chem. Soc.,
1982, 104, 7294–7299.
15 M. Majerić Elenkov, L. Tang, B. Hauer and D. B. Janssen, Adv. Synth.
Catal., 2007, 349, 2279–2285.
Acknowledgements
16 G. M. Morris, D. S. Goodsell, R. S. Halliday, R. Huey, W. E. Hart,
R. K. Belew and A. J. Olson, J. Comput. Chem., 1998, 19, 1639–1662.
17 R. M. de Jong, K. H. Kalk, L. Tang, D. B. Janssen and B. W. Dijkstra,
J. Bacteriol., 2006, 188, 4051–4056.
18 R. M. de Jong, J. J. W. Tiesinga, A. Villa, L. Tang, D. B. Janssen and
B. W. Dijkstra, J. Am. Chem. Soc., 2005, 127, 13338–13343.
19 J. E. T. van Hylckama Vlieg, L. Tang, J. H. L. Spelberg, T. Smilda,
G. J. Poelarends, T. Bosma, A. E. J. van Merode, M. W. Fraaije and
D. B. Janssen, J. Bacteriol., 2001, 183, 5058–5066.
Authors are grateful to Professor Dick B. Janssen for kindly
providing halohydrin dehalogenase plasmid pBADHheA and
pBADHheC. We thank Dr I. J. Elenkov for helpful discussions,
and Mrs Mirjana Čičak for valuable technical assistance. Finan-
cial support by the Unity Through Knowledge Fund (UKF grant
no. 51) is gratefully acknowledged.
20 M. M. Bradford, Anal. Biochem., 1976, 72, 248–254.
21 Y. Peng, J.-H. Yang and W.-D. Z. Li, Tetrahedron, 2006, 62, 1209–1215.
22 M. Ahmadian, N. K. Khare, J. M. Riordan, A. E. Klon and
D. W. Borhani, Tetrahedron, 2001, 57, 9899–9909.
Notes and references
1 E. J. de Vries and D. B. Janssen, Curr. Opin. Biotechnol., 2003, 14,
414–420; D. B. Janssen, M. Majerić Elenkov, G. Hasnaoui, B. Hauer and
J. H. L. Spelberg, Biochem. Soc. Trans., 2006, 34, 291–295; A. Archelas
and R. Furstoss, Curr. Opin. Chem. Biol., 2001, 5, 112–119.
23 M. Lorenzin, A. Guerriero and F. Pietra, J. Org. Chem., 1980, 45, 1704–
1705.
24 H. Label and E. N. Jacobsen, Tetrahedron Lett., 1999, 40, 7303–7306.
25 A. D. Becke, J. Chem. Phys., 1993, 98, 5648–5652.
2 J. Deregnaucourt, A. Archelas, F. Barbirato, J.-M. Paris and R. Furstoss,
Adv. Synth. Catal., 2007, 349, 1405–1417; A. Steinreiber, I. Osprian,
S. F. Mayer, R. V. A. Orru and K. Faber, Eur. J. Org. Chem., 2000, 3703–
3711; A. L. Botes, C. A. G. M. Weijers, P. J. Botes and M. S. van Dyk,
Tetrahedron: Asymmetry, 1999, 10, 3327–3336; M. T. Reetz, C. Torre,
A. Eipper, R. Lohmer, M. Hermes, B. Brunner, A. Maichele, M. Bocola,
M. Arand, A. Cronin, Y. Genzel, A. Archelas and R. Furstoss, Org. Lett.,
2004, 6, 177–180.
26 C. Lee, W. Yang and R. G. Parr, Phys. Rev. B, 1988, 37, 785–789.
27 M. J. Frisch, G. W. Trucks, H. B. Schlegel, G. E. Scuseria, M. A. Robb,
J. R. Cheeseman, G. Scalmani, V. Barone, B. Mennucci, G. A. Petersson,
H. Nakatsuji, M. Caricato, X. Li, H. P. Hratchian, A. F. Izmaylov,
J. Bloino, G. Zheng, J. L. Sonnenberg, M. Hada, M. Ehara, K. Toyota,
R. Fukuda, J. Hasegawa, M. Ishida, T. Nakajima, Y. Honda, O. Kitao,
H. Nakai, T. Vreven, J. A. Montgomery, Jr., J. E. Peralta, F. Ogliaro,
M. Bearpark, J. J. Heyd, E. Brothers, K. N. Kudin, V. N. Staroverov,
R. Kobayashi, J. Normand, K. Raghavachari, A. Rendell, J. C. Burant,
S. S. Iyengar, J. Tomasi, M. Cossi, N. Rega, J. M. Millam, M. Klene,
3 G. Hasnaoui-Dijoux, M. Majerić Elenkov, J. H. Lutje Spelberg, B. Hauer
and D. B. Janssen, ChemBioChem, 2008, 7, 1048–1051.
This journal is © The Royal Society of Chemistry 2012
Org. Biomol. Chem., 2012, 10, 5063–5072 | 5071

View Article Online
J. E. Knox, J. B. Cross, V. Bakken, C. Adamo, J. Jaramillo, R. Gomperts,
R. E. Stratmann, O. Yazyev, A. J. Austin, R. Cammi, C. Pomelli,
J. W. Ochterski, R. L. Martin, K. Morokuma, V. G. Zakrzewski,
G. A. Voth, P. Salvador, J. J. Dannenberg, S. Dapprich, A. D. Daniels,
O. Farkas, J. B. Foresman, J. V. Ortiz, J. Cioslowski and D. J. Fox,
Gaussian 09, Gaussian, Inc., Wallingford CT, 2009.
30 E. Cancès and B. Mennucci, J. Math. Chem., 1998, 23, 309–326.
31 E. Cancès, B. Mennucci and J. Tomasi, J. Chem. Phys., 1997, 107,
3032–3041.
32 B. Mennucci, E. Cancès and J. Tomasi, J. Phys. Chem. B, 1997, 101,
10506–10517.
33 B. Mennucci, R. Cammi and J. Tomasi, J. Chem. Phys., 1998, 109,
2798–2807.
28 S. Miertuš, E. Scrocco and J. Tomasi, Chem. Phys., 1981, 55, 117–
129.
34 S. Dapprich, I. Komáromi, K. S. Byun, K. Morokuma and M. J. Frisch,
THEOCHEM, 1999, 462, 1–21.
29 J. Tomasi and M. Persico, Chem. Rev., 1994, 94, 2027–2094.
5072 | Org. Biomol. Chem., 2012, 10, 5063–5072
This journal is © The Royal Society of Chemistry 2012