Kinetic studies of the asymmetric Henry reaction catalyzed by hydroxynitrile lyase from Hevea brasiliensis

Article abstract of DOI:10.3109/10242422.2010.530661

The asymmetric Henry reaction catalyzed by hydroxynitrile lyase from Hevea brasiliensis is an example of enzymatic catalytic promiscuity. It could be an attractive method to produce optically active β-nitroalcohols, but unfortunately the enzyme has very low activity in this unnatural reaction. To get an insight into the reaction mechanism, the enzyme kinetics of this promiscuous biotransformation were studied using the cleavage and synthesis of 2-nitro-1-phenylethanol as a model system. The results indicate that the kinetic behavior of the enzyme in the Henry reaction fits the classical Rapid Equilibrium Random Bi Uni/Uni Bi mechanistic model with independent substrate binding. The measured kinetic parameters imply that the bottleneck for this biotransformation is the very low turnover number of the enzyme, not the binding of the substrates.

Full text of DOI:10.3109/10242422.2010.530661

Biocatalysis and Biotransformation, September–December 2010; 28(5–6): 348–356
ORIGINAL ARTICLE
Kinetic studies of the asymmetric Henry reaction catalyzed by
hydroxynitrile lyase from Hevea brasiliensis
RUSLANYURYEV1, THOMAS PURKARTHOFER2, MANDANA GRUBER2,
HERFRIED GRIENGL2 & ANDREAS LIESE1
1Institute of Technical Biocatalysis, Hamburg University of Technology, Hamburg, Germany and 2Research Centre Applied
Biocatalysis, Graz, Austria
Abstract
The asymmetric Henry reaction catalyzed by hydroxynitrile lyase from Hevea brasiliensis is an example of enzymatic cata-
lytic promiscuity. It could be an attractive method to produce optically active β-nitroalcohols, but unfortunately the enzyme
has very low activity in this unnatural reaction. To get an insight into the reaction mechanism, the enzyme kinetics of this
promiscuous biotransformation were studied using the cleavage and synthesis of 2-nitro-1-phenylethanol as a model system.
The results indicate that the kinetic behavior of the enzyme in the Henry reaction fits the classical Rapid Equilibrium
Random Bi Uni/Uni Bi mechanistic model with independent substrate binding. The measured kinetic parameters imply
that the bottleneck for this biotransformation is the very low turnover number of the enzyme, not the binding of the
substrates.
Keywords: Enzyme kinetics, hydroxynitrile lyase, catalytic promiscuity, asymmetric Henry reaction
Introduction
Selmar et al. 1987b). However, the enzyme is more
attractive in the reverse process called the cyanohy-
drin reaction (Figure 1(a)), when HCN adds to alde-
hydes or ketones yielding cyanohydrins – valuable
precursors for various fine chemicals (Fechter &
Griengl 2004). In this biotransformation HbHNL
shows high enantioselectivity and has a broad sub-
strate range that makes it appealing for industrial
exploitation. The multi-ton DSM process for pro-
duction of the chiral cyanohydrin of (S)-3-phenoxy-
benzaldehyde, which is used as an intermediate
in the synthesis of pyrethroids (Poechlauer et al.
2004), shows the high commercial potential of this
biocatalyst.
The catalytic promiscuity of HbHNL is revealed
in the Henry reaction, where nitroalkanes are used
instead of HCN (Figure 1(b)) (Purkarthofer et al.
2006).The products of this reaction – optically active
β-nitroalcohols – similarly to the cyanohydrins have
broad synthetic utility, but they are more stable and
less toxic at the same time. For example, they can be
The phenomenon of enzyme promiscuity opens new
opportunities in the field of organic biotransforma-
tions (Bornscheuer & Kazlauskas 2004). It is very
attractive to use enzymes as a multifunctional ‘Swiss
knife’ tool, which can be easily ‘adjusted’ to different
applications, but this is still practically unrealizable.
The main challenge here is usually the very low
activity of enzymes in the promiscuous reactions
compared with that in the native reactions. Over
recent years many efforts have been made to over-
come this problem by applying enzyme, reaction or
substrate engineering. In some cases positive results
were reported (Kazlauskas 2005), especially when
the alterations were based on mechanistic informa-
tion about the enzyme activity.
One of these promiscuous enzymes is hydro-
xynitrile lyase from Hevea brasiliensis (HbHNL). Its
natural catalytic function is to cleave acetone cyano-
hydrin, released from a cyanogenic glycoside linamarin,
into acetone and HCN (Wajant & Förster 1996;
Correspondence: A. Liese, Institute of Technical Biocatalysis, Hamburg University of Technology, Denickestrasse 15, 21073 Hamburg, Germany. Tel:
ꢀ49-40-42878-3218. Fax: ꢀ49-40-42878-2127. E-mail: biocatalysis@tuhh.de
(Received 30 May 2010; revised 6 September 2010; accepted 6 October 2010)
ISSN 1024-2422 print/ISSN 1029-2446 online © 2010 Informa UK, Ltd.
DOI: 10.3109/10242422.2010.530661

Kinetic studies of HbHNL-catalyzed asymmetric Henry reaction 349
OH
CN
δ (ppm): 2.86 (br, 1H, –OH); 4.49 (dd, 1H, 2Jꢁ13.2
Hz, 3Jꢁ2.9 Hz, –CHHNO2); 4.58 (dd, 1H, 2Jꢁ13.2
Hz, 3Jꢁ9.8 Hz, –CHHNO2); 5.43 (dd, IH, 3Jꢁ9.8
Hz, 3Jꢁ2.9 Hz, –CHOH–); 7.33–7.39 (m, 5H).
13C-NMR (125 MHz, CDCl3), δ (ppm): 71.2, 81.4,
126.2, 129.2, 129.3, 138.3.
HCN
O
(a)
HbHNL
OH
(b)
The enzyme HbHNL was kindly provided by
DSM (Heerlen, the Netherlands) in the form of a
NO2
NO2
(c)
MeNO2
solution with the following specifications: 99%
purity, 73 mg mL–1, specific activity 62 U mg–1, sta-
Figure 1. The cyanohydrin (a) and the Henry (b) reactions
catalyzed by HbHNL; (c) the dehydration of NPE to 1-nitro-
2-phenylethylene.
bilized by 30 ppm sodium azide (here, one unit of
activity is defined as the amount of enzyme that
cleaves one μmole of mandelonitrile per minute).
The enzyme was stored at 4°C over the course of
this work without any loss of activity, measured by
the spectrophotometric assay as described (Bauer et
al. 1999a). In the biocatalytic Henry reaction, one
unit of activity is defined as the amount of enzyme
that cleaves one μmole of NPE per minute.
easily converted to the chiral β-aminoalcohols, which
are valuable chiral building blocks for a number of
pharmaceuticals (Luzzio 2001; Ono 2001). Beside
this the HbHNL-catalyzed synthesis of chiral
β-nitroalcohols has a major advantage over the bio-
catalytic cyanohydrin reaction in avoiding manipula-
tions with dangerous HCN. Therefore, the catalytic
promiscuity of HbHNL raises the enzyme's practical
value as a biocatalyst and establishes new perspec-
tives for its industrial use. However, there is one seri-
ous limitation: the activity of HbHNL in the
nitroaldol condensation is too low for a feasible
application (Gruber-Khadjawi et al. 2007).
In order to debottleneck the biocatalytic Henry
reaction information about its mechanism is useful.
Although it is believed that HbHNL acts similarly in
both promiscuous and native reactions, there is not
much experimental evidence to support this. Since
no systematic data on the kinetics of the biocatalytic
Henry reaction are available, it remains unclear
whether the extremely low specific activity of HbHNL
in the promiscuous biotransformation is due to very
weak binding of nitromethane and β-nitroalcohols to
the active site of the enzyme or to a very low turnover
number. The present study aimed to provide this
missing information.
Methods
The reactions were carried out in 1.5-mL glass vials
thermostated at 25°C. Liquid handling was automated
by using a Gilson 241–402 pipette robot (Middleton,
WI, USA). In a typical procedure for NPE cleavage
followed for 20 min, the reaction was started by injec-
tion of 100 μL of 0.5 M phosphate buffer pH 6.0 and
100 μL of HbHNL solution to 800 μL of NPE solu-
tion in 1 mM HCl. After specified periods of time, 50
μL of the reaction mixture were withdrawn and
quenched in 950 μL of 2 M HCl. The precipitated
protein was separated by centrifugation at 6000 rpm
for 20 min and the supernatant was analyzed by
HPLC. Removal of the enzyme from the reaction mix-
ture by quenching prior to HPLC analysis was neces-
sary to extend the lifetime of the HPLC column. To
monitor the reactions within a 1 min interval a modi-
fied procedure was applied; the reaction was started
by rapid injection of 80 μL of the substrate solution
in 1 mM HCl (NPE or benzaldehyde–nitromethane
mixture), 10 μL of 0.5 M phosphate buffer pH 6.0
and 10 μL of HbHNL solution into a 400-μL glass
vessel thermostated at 25°C. After specified times the
mixture was quenched by rapid injection of 350 μL of
2 M HCl, centrifuged and analyzed by HPLC. For
convenience, 1 mM HCl was used to prepare stock
solutions of the substrates (NPE or benzaldehyde–
nitromethane mixture), because it eliminated the non-
specific Henry reaction and, therefore, allowed use of
the solutions for several hours.
Experimental
Materials
The chemicals used in this work were commercially
available
except
(rac)-2-nitro-1-phenylethanol
(NPE), which was synthesized according to the
following procedure. Benzaldehyde (10 mmol),
nitromethane (100 mmol) and triethylamine
(10 mmol) were mixed and incubated overnight at
4°C.Then the mixture was dried in vacuo (20 mbar,
40°C), the yellow residue was purified by column
chromatography on silica gel using cyclohexane–
thylacetate mixture (16:1 v/v) as eluent. After
solvent evaporation NPE was isolated as colorless
Chromatography was performed on a Knauer
Smartline chromatographic system (Berlin,Germany)
equipped with a Merck RP-8 column (Darmstadt, Ger-
many) thermostated at 30°C using a 67 mM KH2PO4
1
oil in 73% yield. H-NMR (500 MHz, CDCl3),

350 R.Yuryev et al.
buffer–methanol mixture (65:35, v/v) as eluent.The
flow rate was 2 mL min–1; peaks were detected by
monitoring the optical absorption of the eluate at
225 nm. The chromatograms were recorded and
analyzed with ChromStar version 6 (SCPA, Stuhr,
Germany) software.
Fitting of experimental data to algebraic equa-
tions was done by using the non-linear least-square
Marquardt–Levenberg algorithm in GNUPLOT
version 4.3 (freeware). The progress curve analysis
and concentration profile simulation, which required
the solution of differential equations, were imple-
mented in MATLAB release 2007a (The MathWorks
Inc., Natick, MA, USA) utilizing lsqnonlin and
ode45 routines.
condensation significantly faster than the spontane-
ous reaction. The convenient on-line UV spectro-
scopic method, which was successfully applied to
follow the cyanohydrin reaction by measuring the
benzaldehyde concentration at 280 nm (Bauer et al.
1999a), was unsuitable in this case since the absorp-
tion of benzaldehyde was masked by HbHNL. The
problem was overcome by using off-line HPLC anal-
ysis, which enabled simultaneous determination of
the concentrations of NPE, benzaldehyde and
1-nitro-2-phenylethylene in the reaction mixture at
selected time points.
Before starting mechanistic studies of HbHNL,
it was important to estimate the equilibrium con-
stant Keq for the model reaction under the selected
reaction conditions:
k1
[NPE]
Results and discussion
Keq
=
=
k−1 [BA][MeNO2 ]
Selection of model system and methods
Since the Henry reaction is reversible, it can be used
either to synthesize or to cleave β-nitroalcohols. To
distinguish each reaction direction, the cleavage of
β-nitroalcohols is sometimes described as the retro-
Henry reaction. This term is also adopted in the
present work and used wherever the differentiation
between both reaction directions is required.
where k1 and kꢂ1 are rate constants of the forward
and backward reactions, respectively; [NPE], [BA]
and [MeNO2] are the equilibrium concentrations
of the reactants (BA, benzaldehyde; MeNO2,
nitromethane). This constant determines the posi-
tion of equilibrium in the system, where rates of
NPE cleavage and synthesis are equal.To assess Keq
one may either measure the concentrations of reac-
tants when the system has reached equilibrium or
determine the rate constants k1 and kꢂ1. The first
approach is not very useful in this particular case
because of the side dehydration reaction, which con-
stantly removes NPE from the reaction mixture and
thus causes a permanent shift of the equilibrium
position. Therefore, the second approach was used.
By measuring initial reaction rates of spontaneous
NPE cleavage and synthesis at various starting con-
centrations of NPE, benzaldehyde and nitromethane
under selected reaction conditions (data not shown),
Keqꢁ6 M–1 was assessed. It is three orders of mag-
nitude lower than the equilibrium constant for the
addition of HCN to benzaldehyde (Keqꢁ4 mM–1;
Ching & Kallen 1978) indicating that, in comparison
to the cyanohydrin reaction, the equilibrium position
in the Henry reaction is shifted to the side of the
carbonyl compound.
The mechanism of HbHNL in the native cyano-
hydrin reaction has already been elucidated by study-
ing the kinetics of cleavage and synthesis of
mandelonitrile (Figure 1(a)) (Bauer et al. 1999a).
For kinetic investigations of the biocatalytic Henry
reaction a similar model system was chosen, in which
nitromethane was used instead of HCN and 2-
nitro-1-phenylethanol (NPE) is the condensation
product (Figure 1(b)). In this system two possible
side reactions have to be considered besides the
enantioselective enzymatic reaction: the unspecific
spontaneous nitroaldol condensation, which runs
parallel with the biocatalytic reaction; and the
dehydration of NPE with formation of 1-nitro-
2-phenylethylene (Figure 1(c)). The spontaneous
Henry reaction could be suppressed by performing
the process at pH 6. The chosen pH value also hin-
dered the dehydration of NPE; in all performed
experiments the concentration of nitrostyrene was
negligible, and therefore was not considered in the
analysis of kinetic data. By contrast, to diminish the
chemical background reaction in the HbHNL-cata-
lyzed cyanohydrin condensation significantly lower
pH values (around 4.5–5.0) are necessary (Selmar
et al. 1987b).
Using the estimated Keq value it was possible to
predict the conversion at the equilibrium position
X
eq (if the dehydration of NPE to 1-nitro-2-phenyl-
ethylene is neglected) as a function of the concentra-
tionofstartingmaterials:eitherNPEfortheretro-Henry
reaction or benzaldehyde and nitromethane for the
Henry reaction (Figure 2).The prediction was done
for the concentration range limited by the solubility
of each substance in water: ∼0.1 M for NPE and
The main problem during the selection of an
appropriate method to monitor the progress of the
biocatalytic Henry reaction was the high concentration
of the enzyme required to make the enzymatic

Kinetic studies of HbHNL-catalyzed asymmetric Henry reaction 351
1
0.8
0.6
0.4
0.2
0
1
0.8
0.6
0.4
0.2
0
c
b
a
0
20
40
60
80 100
0
20
40
60
80 100
[NPE], mM
[BA], mM
Figure 2. Equilibrium conversion, predicted for Keqꢁ6 M–1, for the cleavage (left) and the synthesis (right) of NPE at 25°C in 50 mM phosphate
buffer as a function of the concentration of the starting materials. Concentration of nitromethane: (a) 0.1 M, (b) 0.5 M, (c) 2 M.
benzaldehyde, ∼2 M for nitromethane. The calcula-
tions revealed that, at low starting concentrations of
NPE (ꢃ10 mM), the retro-Henry reaction can be
treated as irreversible (Xeq≈1). The same is true for
the Henry reaction if the starting concentration of
nitromethane is close to its solubility limit (∼2 M).
applied conditions the rate of the spontaneous retro-
Henry reaction was at least tenfold lower than that
of the biocatalytic one. The cleavage of NPE was
monitored over 1 min, where the concentration pro-
file for benzaldehyde was still linear (Figure 4). The
calculated initial rates (mM minꢂ1) were corrected
for the spontaneous reaction, normalized by the
enzyme concentration [HbHNL] and plotted against
the concentration of NPE (Figure 5). The graph
obtained shows the saturation kinetics of NPE cleav-
age, which fits the simple one-substrate Michaelis–-
Menten model:
Cleavage of NPE
Since HbHNL is (S)-enantioselective in the Henry
reaction, in the selected model reaction system the
enzyme catalyzes only the synthesis and the cleavage
of (S)-NPE. Therefore, to study the kinetics of the
retro-Henry reaction, it is logical to use enantiopure
(S)-NPE as a substrate. However, this chiral sub-
stance is commercially unavailable and cannot be
easily synthesized in sufficient quantity.To overcome
this problem, instead of the pure (S)-enantiomer, the
racemic mixture of (S)- and (R)-enantiomers was
used (Figure 3). While (rac)-NPE also cannot be
purchased commercially, it can be easily obtained on
a gram scale by chemical synthesis, in contrast to
(S)-NPE.This approach was also applied in the work
of Bauer et al. (1999a), who used racemic rather
than (S)-mandelonitrile for the kinetic studies of
HbHNL in the cyanohydrin reaction under the
assumption that the (R)-enantiomer has a negligible
influence on the enzyme's kinetics in the studied
concentration range. In the present work the same
assumption regarding (R)-NPE was made.
V
max,r[NPE]
u
=
[HbHNL] 2KM,(S)-NPE ꢀ[NPE]
where [NPE] is the concentration of (rac)-NPE;
[HbHNL]ꢁ0.82 mg mL–1; Vmax,r is the maximum
reaction rate of the biocatalytic cleavage of NPE,
U mgꢂ1; and KM,(S)-NPE is the respective Michaelis
0.3
0.2
0.1
0
The concentration of (rac)-NPE was varied in the
range 5–106 mM (up to the solubility limit under
the reaction conditions), while HbHNL concentra-
tion was kept constant at 0.82 mg mL–1. Under the
0
20
40
60
Time, s
Figure 4. Initial conversion rates observed during NPE cleavage
catalyzed by HbHNL (0.82 mg mL–1) in 50 mM phosphate buffer
pH 6.0 at 25°C. Initial concentration of NPE: ( ) 14 mM, (•) 41
mM, (᭜) 82 mM.
Figure 3. Cleavage of (rac)-NPE as a model system to study the
kinetics of the biocatalytic retro-Henry reaction.

352 R.Yuryev et al.
where Mwꢁ30 kg molꢂ1 is the molecular weight of
HbHNL.Thus, the enzyme needs around 6 s to split
one molecule of NPE in the active site, compared
with only 0.02 s for mandelonitrile.
0.4
0.3
Product inhibition during NPE cleavage
0.2
0.1
Since HbHNL catalyzes the cleavage and synthesis
of NPE, the enzyme should be able to bind both
substances, NPE and benzaldehyde, in its active site.
Hence benzaldehyde, formed as a product in the
course of NPE cleavage, would compete with the
β-nitroalcohol for the active site of HbHNL and
thereby inhibit the reaction. In the cleavage of man-
delonitrile, HbHNL is strongly inhibited by the for-
mation of benzaldehyde, for which the enzyme has
a high affinity (Ki,BAꢁ1.2 mM; Bauer et al. 1999a).
The same product inhibition effect was also observed
in the cleavage of NPE; the rate of the enzymatic
reaction decreased when benzaldehyde was initially
present in the reaction mixture (Figure 6).
0
0
25
50
75
100
[NPE], mM
Figure 5.The Michaelis–Menten plot for NPE cleavage catalyzed
by HbHNL (0.82 mg mL–1) in 50 mM phosphate buffer pH 6.0
at 25°C.
constant for (S)-NPE. The kinetic parameters
determined by non-linear fitting are presented in
Table I.
The empirical reaction progress curves fit the
following kinetic model, which assumes that benzal-
dehyde is a competitive inhibitor:
The Michaelis constant for (S)-NPE is of the
same order of magnitude as the one for (S)-
mandelonitrile in the cyanohydrin reaction
(KMꢁ1.55 mM) reported by Bauer et al. (1999a),
showing that HbHNL has relatively good affinity
to this β-nitroalcohol. Unfortunately, for the NPE
cleavage it was not possible to determine precisely
the enzyme's turnover number kcat, which would
show how many molecules of NPE are converted
by one molecule of HbHNL per second. The rea-
son was that there is currently no suitable active
site titration protocol which would allow determi-
nation of the exact molar concentration of HbHNL
active sites in a solution. However, assuming that
100% of the enzyme was present in its active form,
i.e. there was no denatured or misfolded enzyme,
the kcat value could be approximated from Vmax,r
as follows:
d[BA]
ꢁ uꢀkb[NPE]
dt
d[NPE]
d[BA]
ꢁ ꢅ
dt
dt
d[(S)-NPE]
dt
ꢁ ꢅ u ꢅ kb[(S)-NPE]
V
max,r[(S)-NPE]
u
ꢁ
[Hb HNL]
KM,(S )-NPE 1ꢀ([BA] / K
) ꢀ[(S)-NPE]
(
)
i,BA
where [BA], [NPE] and [(S)-NPE] are molar con-
centrations of benzaldehyde, (rac)-NPE and (S)-NPE,
respectively; kb is the reaction rate constant for the
spontaneous cleava ge of NPE, kbꢁ5.3ꢄ10ꢂ4 minꢂ1
(data not shown); V is the reaction rate of the bio-
catalytic reaction,mM minꢂ1; KM,(S)-NPE is the Michae-
lis constant for (S)-NPE, KM,(S)-NPEꢁ2.55 mM; Ki,BA
Mw
kcat ≈ Vmax,r ꢄ
ꢁ 0.16 s−1
60
Table I. Estimated kinetic parameters for cleavage and synthesis of NPE catalyzed by HbHNLa.
Cleavage
Synthesis
Value
Parameter
Value
Unit
Parameter
Unit
0.32ꢆ0.02b
0.16
0.027ꢆ0.003
0.013
Vmax,r
kcat
U mg–1
s–1
Vmax,f
kcat
U mg–1
s–1
KM,(S)-NPE
Ki,BA
Ki,MeNO2
mM
mM
mM
2.55ꢆ0.65
0.37ꢆ0.04
103ꢆ7
KM,BA
KM,MeNO2
mM
mM
0.33ꢆ0.01
64ꢆ21
aReaction conditions: 50 mM phosphate buffer pH 6.0, 25°C.
b95% confidence intervals were estimated from numerical fitting of experimental data to kinetic equations.

Kinetic studies of HbHNL-catalyzed asymmetric Henry reaction 353
5
4
3
2
1
0
4
3
2
1
0
0
5
10
15
0
5
10
15
Time, min
Time, min
Figure 6. Inhibition of HbHNL (2.1 mg mL–1) by benzaldehyde in the cleavage of NPE in 50 mM phosphate buffer pH 6.0 at 25°C.
Initial concentration of benzaldehyde: (᭿) 0 mM, (•) 0.5 mM, (᭡) 1.0 mM, (᭢) 1.5 mM, (᭜) 2.5 mM, (∗) 3.0 mM. Solid lines correspond
to numerical simulation of reaction progresses.
The observed patterns for the relative inhibition
v/v0 of HbHNL by nitromethane fit the following
equation, which implies that nitromethane acts as a
competitive inhibitor:
is the inhibition constant for benzaldehyde; Vmax,r is
the maximum reaction rate of the biocatalytic cleav-
age of NPE, Vmax,rꢁ0.32 U mgꢂ1; and [HbHNL]
istheappliedHbHNLconcentration,[HbHNL]ꢁ2.1
mg mL–1.This model also implies that nitromethane
has negligible influence on the enzyme in the stud-
ied concentration range.
1ꢀ [NPE] /2K
(
)
u
0
M,(S )-NPE
=
u0
1ꢀ [NPE] /2K
ꢀ [MeNO ]/K
(
(
)
)
0
M,(S)-NPE
2
i,MeNO2
From this, Ki,BA remained the only unknown
parameter to be estimated. By non-linear fitting of
the concentration profile to the model it was found
that Ki,BAꢁ0.37 ꢆ 0.04 mM.This value is within the
same order of magnitude as the inhibition constant
for benzaldehyde determined by in the cleavage of
mandelonitrile (Bauer et al. 1999a). Such coinci-
dence supports the hypothesis that both the Henry
and the cyanohydrin reactions take place in the same
active site of HbHNL and was expected a priori.
The concentration profile shown on Figure 6 was
also fitted to the models assuming other types of inhi-
bition for benzaldehyde (uncompetitive and non-com-
petitive), but the corresponding Ki,BA values, obtained
from the non-linear fitting, were either physically mean-
ingless or the numerical procedure did not converge.
The second product formed during the NPE
cleavage – nitromethane – also inhibited the enzyme,
albeit not as strongly as benzaldehyde (Figure 7). Its
inhibition effect was assessed by measuring initial
rates of cleavage of 5.2 and 52 mM NPE in the pres-
ence of 0.1–0.4 M MeNO2. The inhibition was sig-
nificant only when the concentration of nitromethane
[MeNO2] was higher than 0.1 M. In this case the
progress curve analysis, which was used to determine
the inhibition constant for benzaldehyde, was rather
inappropriate, because the presence of nitromethane
at such a high concentration, by analogy to HCN
(Bauer et al. 1999b), could cause denaturation of the
enzyme. Furthermore, at a high concentration of
nitromethane the rate of the backward (S)-NPE syn-
thesis should be considered as well.
where v is the observed initial reaction rate
corrected for the spontaneous reaction; v0 is the
corrected initial rate in the absence of nitromethane;
[NPE]0 is the initial molar concentration of (rac)-
NPE, [NPE]0ꢁ5.2 or 52 mM; and KM,(S)-NPEꢁ2.55
mM is the Michaelis constant for (S)-NPE, as
determined from the NPE cleavage experiments.
The inhibition constant for nitromethane Ki,MeNO2
found by non-linear fitting was equal to 103ꢆ7 mM.
The attempt to fit the inhibition patterns to equa-
tions corresponding to other inhibitor types (uncom-
petitive, non-competitive or mixed) failed since the
1
0.8
0.6
0.4
0.2
0
0
0.1
0.2
0.3
0.4
[MeNO2], M
Figure 7. Relative inhibition pattern for nitromethane in NPE
cleavage catalyzed by HbHNL (1.4 mg/mL) in 50 mM phosphate
buffer pH 6.0 at 25°C. Initial concentration of NPE: (᭿) 52 mM,
(•) 5.2 mM.

354 R.Yuryev et al.
0.12
0.09
0.06
0.03
0
numerical procedure did not converge.The relatively
large value of Ki,MeNO2 implies that binding of nitro-
methane to HbHNL is weak, but comparable with
HCN in the cleavage of mandelonitrile (Ki,HCNꢁ150
mM; Bauer et al. 1999a).
Remarkably, in the cyanohydrin reaction, HCN
shows a mixed-type inhibition while benzaldehyde
acts as a competitive inhibitor (Bauer et al. 1999a)
that is consistent with the Ordered Bi Uni mecha-
nism proposed for this reaction. In contrast, in the
biocatalytic Henry reaction both nitromethane and
benzaldehyde appeared to be competitive inhibitors.
This indicates that, in the promiscuous biotransfor-
mation, the enzyme follows a Rapid Equilibrium
Random Bi Uni mechanism.
0
20
40
Time, s
60
80
Figure 8. Initial conversion rates observed during synthesis of
(S)-NPE catalyzed by HbHNL (4.0 mg/mL) in the presence of
160 mM MeNO2 in 50 mM phosphate buffer pH 6.0 at 25°C.
Initial concentration of benzaldehyde: (᭿) 0.4 mM, (•) 1.2 mM,
(᭜) 2.0 mM.
Synthesis of NPE
The kinetics of NPE synthesis were studied using
the initial rate method in the concentration range
0.3–3 mM for benzaldehyde and 80–320 mM
for nitromethane. In all of these experiments
the concentration of HbHNL was kept constant at
4.0 mg mL–1.The initial rates were determined from
the linear fit of the conversion rates within 1 min
(Figure 8). Under the reaction conditions chosen no
spontaneous formation of NPE could be detected,
thus no rate corrections were necessary.
The corresponding Michaelis–Menten plot
(Figure 9), which is based on the measured initial
rates V (mM minꢂ1) normalized by the enzyme con-
centration [HbHNL] (4.0 mg mL–1), shows that the
synthesis of NPE also follows saturation kinetics
meaning that HbHNL can be saturated by both sub-
strates – benzaldehyde and nitromethane. The
empirical kinetic data fit the Rapid Equilibrium Ran-
dom Bi Uni mechanism with independent substrate
binding (Figure 10), the rate equation for which has
the following form:
preference of HbHNL for the retro-Henry reaction:
the enzyme requires approximately 1 min to synthe-
size one molecule of the nitroalcohol, but only
around 6 s to cleave it.
The Random mechanism assumes that all binding
(k2 to k4) and dissociation (k–2 to k–4) steps are very
rapid compared with the catalytic step (k1 and k–1),
which seems to be true for the biocatalytic Henry
reaction, and that the binding of one substrate does
not influence the affinity of the enzyme for the other
substrate (Segel 1975). The model also predicts that
both benzaldehyde and nitromethane should be com-
0.025
0.02
0.015
0.01
0.005
0
V
max,f [BA][MeNO2 ]
u
ꢁ
[HbHNL]
K
ꢀ[BA] K
ꢀ[MeNO ]
(
)(
)
M,BA
M,MeNO2
2
where Vmax,f is the maximum reaction rate of the
biocatalytic synthesis of NPE; [BA] and [MeNO2]
are molar concentrations of benzaldehyde and
nitromethane, respectively; and KM,BA and KM,MeNO2
are the corresponding Michaelis constants. The
kinetic parameters, determined from non-linear
fitting of the data to the model equation, are pre-
sented in Table I. Attempts to fit the data to the
Ordered Bi Uni mechanism were not successful,
since the numerical procedures did not converge.
The turnover number for NPE synthesis
kcat≈0.013 s–1, which was estimated from Vmax,f in
the same way as for NPE cleavage, reveals the kinetic
0
1
2
3
4
[BA], mM
Figure 9. The Michaelis–Menten plot for (S)-NPE synthesis
catalyzed by HbHNL (4.0 mg/mL) in 50 mM phosphate buffer
pH 6.0 at 25°C. Initial concentration of nitromethane: (᭿) 80 mM,
(•) 160 mM, (᭡) 240 mM, (᭜) 320 mM.

Kinetic studies of HbHNL-catalyzed asymmetric Henry reaction 355
to follow the same mechanism in the Henry reaction.
The observed differences in kinetic behavior might
indicate that, in the promiscuous reaction, HbHNL
reacts differently from the native reaction. However,
there is a plausible explanation which rationalizes the
observed differences in kinetic behavior of HbHNL
in the two reactions, even if they take place in the
same active site of the enzyme.
EA
k–3
k3
A
B
k–4
EAB
k–3
k4
k3
k–2
k2
k1
E
EP
E
P
k–1
B
A
k–4
EB
The Ordered Bi Uni mechanism can be derived
from the Random Bi Uni mechanism, when: (a)
substrate B binds to the enzyme E much more
strongly and much faster than substrateA (KB>>KA,
k4 >> k3); (b) the dissociation rate of the enzyme
complex with substrate B is much slower than the
rate of the catalytic step (k1 >> kꢂ4). In this case the
upper pathway in the random mechanism via the EA
complex would be non-operational (Figure 10).
Both conditions might be true for the biocatalytic
cyanohydrin reaction, and as a result HbHNL fol-
lows the Ordered Bi Uni mechanism in this
biotransformation. In contrast, in the biocatalytic
Henry reaction it has been shown that the rate of
the catalytic step (k1 and k–1) is very low in com-
parison to the cyanohydrin reaction.Therefore, the
condition (b) could be unsatisfied, and as a conse-
quence the upper pathway via the EA complex
would be operational, resulting in the Rapid
Equilibrium Random Bi Uni mechanism.
k4
Figure 10. Rapid Equilibrium Random Bi Uni mechanism
(including gray box) proposed in this work for the biocatalytic
Henry reaction and Ordered Bi Uni mechanism (excluding gray
box) suggested by Bauer et al. (1999a) for the biocatalytic
cyanohydrin reaction. E is HbHNL; B is benzaldehyde; A is either
nitromethane or HCN; P is either NPE or mandelonitrile; EA,
EB, EAB and EP denote the respective enzyme complexes.
petitive inhibitors for HbHNL in the NPE cleavage
reaction, which was confirmed by the inhibition stud-
ies. Moreover, according to this model, the inhibition
constants for benzaldehyde and nitromethane should
be equal to the respective Michaelis constants,because
in this particular case the inhibition and the Michae-
lis constants for each substrate represent the sub-
strate's true equilibrium binding constant (KAꢁk–3/k3
or KBꢁkꢂ4/k4, Figure 10).This expected equivalence
was obeyed here (Table I).
The consistency of the determined kinetic con-
stants was checked by using the Haldane relation-
ship (Segel 1975), which expresses the equilibrium
constant for the Henry reaction in terms of the
enzyme kinetic constants:
Conclusions
The kinetics of the HbHNL-catalyzed promiscuous
Henry reaction fits the classical mechanistic model –
Rapid Equilibrium Random Bi Uni with independent
substrate binding – which implies that the bottleneck
for this biotransformation is a very low turnover num-
ber of the enzyme, not the binding of the substrates.
This agrees with the hypothesis of Purkarthofer et al.
(2006) based on molecular docking simulations,
according to which NPE and mandelonitrile bind to
the active site of HbHNL in a similar manner. Unfor-
tunately, the answer to the question why the enzyme
shows such a low turnover number is beyond enzyme
kinetics. In order to debottleneck the biocatalytic
Henry reaction one needs a deeper insight into the
mechanism of this transformation at a molecular level.
In this respect the employment of other techniques,
like X-ray structural analysis combined with molecu-
lar modeling, should be considered.
Vmax,f K(S)-NPE
Keq
ꢁ
≅ 9 M−1
Vmax,r KBA KMeNO2
where Vmax,r and Vmax,f are the maximum reaction
rates for cleavage and synthesis of NPE, respectively;
BA, KMeNO2 and K(S)-NPE are the true equilibrium
K
binding constants for benzaldehyde, nitromethane
and (S)-NPE, which in this particular case are equal
to the respective Michaelis or inhibition constants.
On the other hand, Keqꢁ6 Mꢂ1 was calculated as a
ratio of the rate constants for the unspecific NPE
cleavage and synthesis. Both values are of the same
order of magnitude, indicating that the estimated
kinetic constants are consistent.
Interestingly, according to the Ordered Bi Uni
mechanism proposed by Bauer et al. (1999a) for the
HbHNL-catalyzed cyanohydrin reaction, the enzyme
binds benzaldehyde first and then HCN. Based
on the suggestion that HbHNL acts similarly in
both the Henry and the cyanohydrin reactions
(Purkarthofer et al. 2006), one would expect HbHNL
Declaration of interest: The authors report no
conflicts of interest. The authors alone are respon-
sible for the content and writing of the paper.

356 R.Yuryev et al.
Kazlauskas RJ. 2005. Enhancing catalytic promiscuity for bioca-
talysis. Curr Opin Chem Biol 9:195–201.
Luzzio FA. 2001. The Henry reaction: recent examples. Tetrahe-
dron 57:915–945.
References
Bauer M, Griengl H, Steiner W. 1999a. Kinetic studies on the
enzyme (S)-hydroxynitrile lyase from Hevea brasiliensis using
initial rate methods and progress curve analysis. Biotechnol
Bioeng 62:20–29.
Bauer M, Grieng H, Steiner W. 1999b. Parameters influencing
stability and activity of a S-hydroxynitrile lyase from Hevea
brasiliensis. Enzyme Microb Technol 24:514–522.
Bornscheuer UT, Kazlauskas RJ. 2004. Catalytic promiscuity in
biocatalysis: using old enzymes to form new bonds and follow
new pathways. Angew Chem Int Ed 43:6032–6040.
Ching WM, Kallen RG. 1978. Mechanism of carbanion addition
to carbonyl compounds. Equilibria and kinetics of substituted
cyanohydrin cleavage and formation in aqueous solution. Sub-
stituted cyanohydrin proton dissociation constants. J Am Chem
Soc 100:6119–6124.
Ono N. 2001. The nitro-aldol (Henry) reaction. In: The nitro
group in organic synthesis. Wiley-VCH NewYork. p 30.
Poechlauer P, Skranc W, Wubbolts M. 2004. The large-scale bio-
catalytic synthesis of enantiopure cyanohydrins. In: Schmidt E,
Blaser H-U, editors. Asymmetric catalysis on industrial scale.
Wiley-VCH Verlag GmbH & Co. KGaA Weinheim. p 149.
Purkarthofer T, Gruber K, Gruber-Khadjawi M, Waich K,
Skranc W, Mink K, Griengl H. 2006. A biocatalytic Henry
reaction – the hydroxynitrile lyase from Hevea brasiliensis also
catalyzes nitroaldol reactions. Angew Chem Int Ed 45:
3454–3456.
Segel IH. 1975. Enzyme kinetics. NewYork: Wiley.
Selmar D, Carvalho FJP, Eric E. Conn EE. 1987a. A colorimet-
ric assay for α-hydroxynitrile lyase. Anal Biochem 166:
208–211.
Selmar D, Lieberei R, Biehl B, Voigt J. 1987b. Hevea linamarase
– a nonspecific β-glycosidase. Plant Physiol 83:557–563.
Wajant H, Förster S. 1996. Purification and characterization of
hydroxynitrile lyase from Hevea brasiliensis. Plant Sci 115:25–31.
Fechter MH, Griengl H. 2004. Hydroxynitrile lyases: biological
sources and application as biocatalysts. Food Technol Biotech-
nol 42:287–294.
Gruber-Khadjawi M, PurkarthoferT, SkrancW, Griengl H. 2007.
Hydroxynitrile lyase-catalyzed enzymatic nitroaldol (Henry)
reaction. Adv Synth Catal 349:1445–1450.