Oxidative dechlorination of halogenated phenols catalyzed by two distinct enzymes: Horseradish peroxidase and dehaloperoxidase

Article abstract of DOI:10.1016/j.abb.2010.09.018

The mechanism of the dehalogenation step catalyzed by dehaloperoxidase (DHP) from Amphitrite ornata, an unusual heme-containing protein with a globin fold and peroxidase activity, has remarkable similarity with that of the classical heme peroxidase, horseradish peroxidase (HRP). Based on quantum mechanical/molecular mechanical (QM/MM) modeling and experimentally determined chlorine kinetic isotope effects, we have concluded that two sequential one electron oxidations of the halogenated phenol substrate leads to a cationic intermediate that strongly resembles a Meisenheimer intermediate - a commonly formed reactive complex during nucleophilic aromatic substitution reactions especially in the case of arenes carrying electron withdrawing groups.

Full text of DOI:10.1016/j.abb.2010.09.018

Archives of Biochemistry and Biophysics 505 (2011) 22–32
Contents lists available at ScienceDirect
Archives of Biochemistry and Biophysics
journal homepage: www.elsevier.com/locate/yabbi
Oxidative dechlorination of halogenated phenols catalyzed by two distinct enzymes:
Horseradish peroxidase and dehaloperoxidase
Lukasz Szatkowski a, Matthew K. Thompson b, Rafal Kaminski a, Stefan Franzen b,
Agnieszka Dybala-Defratyka a,
⇑
a Technical University of Lodz, Faculty of Chemistry, Institute of Applied Radiation Chemistry, Zeromskiego 116, 90-924 Lodz, Poland
b North Carolina State University, Department of Chemistry, Raleigh, NC 27695, USA
a r t i c l e i n f o
a b s t r a c t
Article history:
Available online 24 September 2010
The mechanism of the dehalogenation step catalyzed by dehaloperoxidase (DHP) from Amphitrite ornata,
an unusual heme-containing protein with a globin fold and peroxidase activity, has remarkable similarity
with that of the classical heme peroxidase, horseradish peroxidase (HRP). Based on quantum mechanical/
molecular mechanical (QM/MM) modeling and experimentally determined chlorine kinetic isotope
effects, we have concluded that two sequential one electron oxidations of the halogenated phenol sub-
strate leads to a cationic intermediate that strongly resembles a Meisenheimer intermediate – a com-
monly formed reactive complex during nucleophilic aromatic substitution reactions especially in the
case of arenes carrying electron withdrawing groups.
Keywords:
Dehaloperoxidase
Horseradish peroxidase
Chlorine kinetic isotope effects
QM/MM
Halophenols
Oxidation
Ó 2010 Elsevier Inc. All rights reserved.
Introduction
groups of enzymes are capable of degrading the halogenated com-
pounds [13–15]. For example, hydrolytic dehalogenases replace
Application of enzymes in order to remove environmental and
industrial contaminants has been studied for the last 20 years with
increasing intensity as novel mechanisms are discovered in living
organisms [1–7]. The reactions carried out using enzyme-based
techniques are characterized by high efficiency and selectivity
and are significantly more benign for the environment [8–10]
when compared to purely chemical methods of environmental
remediation.
One of the most important groups of pollutants in industrial
wastewaters are chlorinated phenols. They originate primarily
from the pulp and paper industries and are toxic carcinogens. In
addition to their artificial origin, about 4000 organohalogenated
compounds are produced biologically or by natural abiogenic pro-
cesses such as volcanic eruptions, forest fires, and other geother-
mal processes [11]. Marine organisms such as seaweeds, sponges,
corals, tunicates, and bacteria constitute the major biological pro-
ducers of organohalogenated compounds, while terrestrial plants,
fungi, lichen, bacteria, insects, and even some higher animals
including humans produce a small fraction [12].
the halogen with a hydroxyl group, reductive dehalogenases re-
place the halogen with a hydrogen, and oxidative dehalogenases
exchange the halogen for oxygen with concomitant substrate oxi-
dation (Scheme 1) [12].
Many mechanistic studies on selected hydrolytic dehalogenases
have been presented [16–18], and even such a narrow class of en-
zymes shows a huge variety of possible pathways. Thus, research
leading to detailed analysis of biological reactions is necessary in
order to implement them more universally into detoxification pro-
cesses and employ them on an industrial scale. However, enzymes
as biocatalysts suffer from certain disadvantages (e.g. high costs of
producing enzymes in large amounts, inactivation under specific
conditions, possibility of utilizing them under very toxic environ-
ment, full characterization of all reaction products, etc.). Therefore,
before they can be used efficiently, a number of studies are
required in order to tune their performance accordingly. The enzy-
matic processes such as peroxidase- and phenol oxidase-catalyzed
treatment of phenols and aromatic amines are likely the most com-
prehensively studied systems for waste treatment so far. Peroxi-
dases present in plants, such as horseradish and soybean, and
fungi, such as white-rot fungi, as well as phenol oxidases from
mushrooms (tyrosinase) and white-rot fungi (laccase) have all
been found to have some potential for environmental applications,
and some studies toward this direction have been undertaken
[1–7,19,10]. Species living in environments where halogenated
contaminants are found have been forced to develop appropriate
Thus taking into account the vast number of different sources of
halogenated compounds, Nature has developed several mecha-
nisms in order to reduce the levels of these toxic molecules. Many
⇑
Corresponding author.
E-mail address: olaf@p.lodz.pl (A. Dybala-Defratyka).
0003-9861/$ - see front matter Ó 2010 Elsevier Inc. All rights reserved.
doi:10.1016/j.abb.2010.09.018

L. Szatkowski et al. / Archives of Biochemistry and Biophysics 505 (2011) 22–32
23
Scheme 1. Selected examples of enzymatic dehalogenation: (a) hydrolytic, (b) reductive, and (c) oxidative.
protection pathways. One example of such a species is the terebel-
lid polychaete Amphitrite ornata, a marine worm found to contain
the dehalogenating enzyme, dehaloperoxidase (DHP)1. DHP is a di-
meric hemoglobin that was identified as the active agent in the oxi-
dation of the naturally occurring repellent, 2,4,6-tribromophenol
(TBP). The repellents TBP, 2,4-dibromophenol (DBP) and 4-bromo-
phenol (4-BP) are synthesized by a host of organisms in benthic eco-
systems. The prevalence of TBP, DBP, and 4-BP in coastal wetland
environments suggests that a natural regulation of their catabolism
must exist.
Like horseradish peroxidase (HRP), DHP possesses a heme
cofactor with a distal and proximal histidine on either side and cat-
alyzes the oxidative dehalogenation of halogenated phenols to
their corresponding quinones when activated by H2O2 (Scheme 2)
[20–22]. The mechanism involves two consecutive one-electron
transfers in order to oxidize organic haloderivatives similar to
other heme-containing peroxidases including HRP. However,
although the net reaction is considered to proceed according to
the same mechanism, the enzyme intermediates (oxidizing spe-
cies) formed during the reaction differ. In HRP, the reaction pro-
ceeds via the two enzyme intermediates, compound
I and
compound II (Scheme 3), whereas in DHP, the formation of a com-
pound ES intermediate similar to the tryptophan cation radical
found in cytochrome c peroxidase (CcP) has been reported [23].
While compound I of HRP is characterized by an oxoferryl heme
with a porphyrin cation radical, compound ES of DHP is character-
ized by an oxoferryl heme with a tyrosine cation radical [23].
Therefore, the two enzymes, a prototypical heme-containing per-
oxidase (HRP) and hemoglobin with significantly enhanced perox-
idase activity (DHP), share a similar mechanism of oxidation but
differ in the details of the electron transfer pathway between the
substrate and heme iron.
The distal side of the heme near the active site of HRP is consid-
ered to be too narrow for organic substrates to enter and directly
interact with the heme iron center. Instead, they bind at the heme
edge where they are in contact with surrounding solvent water
molecules that aid in dehalogenation. Unlike HRP, DHP has been
shown to possess a well-defined binding pocket on the distal side
of the heme large enough to accommodate mono-halogenated or-
ganic substrates [24–28]. This led to the hypothesis that oxidation
of halogenated phenols occurred in the interior of DHP. While HRP
can catalyze the oxidation of all known derivatives of halophenols
Scheme 2. The catalytic cycle of DHP with 2,4,6-trichlorophenol (TCP) as
substrate.
a
1
Abbreviations used: DHP, dehaloperoxidase; TBP, 2,4,6-tribromophenol; DBP,
2,4-dibromophenol; 4-BP, 4-bromophenol; HRP, horseradish peroxidase; CcP,
cytochrome c peroxidase; DMCP, 2,6-dimethyl-4-chlorophenol; TCP, 2,4,6-trichloro-
phenol; 4-CP, 4-chlorophenol; QM/MM, quantum mechanical/molecular mechanical;
EMc, enzyme–Meisenheimer complex.
(mono-, di-, and tri-), DHP is primarily active toward trihalophe-
nols with only moderate activity toward dihalophenols and no

24
L. Szatkowski et al. / Archives of Biochemistry and Biophysics 505 (2011) 22–32
Aldrich, toluene (pure p.a.), n-hexane (pure p.a.) from Chempur
were used without further purification.
Dehaloperoxidase – protein expression, purification and oxidation
Recombinant his-tagged wild type protein was expressed in
E. coli and purified as previously described [27,32]. The isolated
His-DHP was oxidized with excess of potassium ferricyanide,
K3[Fe(CN)6] and buffer exchanged into 20 mM KH2PO4, pH 6 solu-
tion using a Sephadex G-25 column. The oxidized protein was
loaded onto a CM 52 ion exchange column for further purification.
It was washed on the column with 20 mM KH2PO4, pH-6 buffer
and eluted from the column with 100 mM KH2PO4, pH 7 buffer.
The concentration of the protein was determined using intensity
of the Soret absorption at 406 nm and a molar absorptivity of
116400 Mꢁ1 cmꢁ1
.
Dehalogenation of TCP catalyzed by DHP and HRP
We have used a 5 mM solution of TCP in 100 mM potassium
phosphate buffer at pH 7. Because of the low solubility of DMCP
in pure water, a small amount of ethanol (1% v:v) was added to
Scheme 3. The catalytic cycle of HRP with TCP as a substrate.
the solution. The protein concentrations were 500 and ꢂ260
lM
for DHP and HRP, respectively. The reactions were initiated by
addition of a calculated amount of H2O2 (assuming 1/1 stoichiom-
etry for H2O2/substrate turnover) required to make the reaction
proceed to ꢂ20% and 100% for KIE analysis.
activity toward monohalophenols. Further investigation of internal
binding of monohalophenols led to the discovery that internal
binding inhibits the peroxidase function of DHP [29]. Therefore,
DHP must have at least two binding sites, the internal site on the
distal side of the heme [30,31] and an external site similar to that
of HRP where substrate oxidation occurs. Two possible external
sites have been proposed on the basis of 1H/15N HSQC experiments
on 13C/15N labeled DHP; one at the external heme edge near the
distal His55 similar to HRP, and one near the dimer interface
[32]. Thus in the line with the previous studies of dehalogenating
enzymes [16,17], we were very interested to check whether the
dehalogenation step itself is accompanied by interactions with
these enzymes in any way. And if so, what are their mechanisms?
For this purpose we have employed kinetic isotope effect analysis
as their values that can be determined both experimentally and
theoretically have already proven many times to be a successful
tool in elucidating mechanisms of enzymatic reactions [34–36].
In this particular case, we have focused on chlorine kinetic isotope
effects as they are a good probe to differentiate between alterna-
tive reaction pathways (i.e. water attack on a dissociable phenoxy
radical vs. a cationic intermediate). Herein we present our efforts
in describing the role of the two distinctive enzymes in the degra-
dation process of organic pollutants like halogenated phenols. Our
QM/MM results suggest that the reaction pathway involving the
cationic intermediate is preferred for the dehalogenation reaction.
To the best of our knowledge these results represent the only chlo-
rine isotopic fractionation data available for dehalogenation reac-
tion catalyzed by peroxidases.
The progress of the reaction was monitored spectrophotometri-
cally as shown in Fig. 1.
Dehalogenation of 4-CP catalyzed by HRP
A 7.5 mM solution of 4-CP in 100 mM potassium phosphate
buffer, pH 5.5 was used in the reaction with ꢂ380
lM solution
of HRP. We changed the pH conditions for the reaction with 4-
CP based on study showing that optimum pH for p-chlorophenol
is about 6 [33]. The reaction was initiated and controlled by
accurate dosage of hydrogen peroxide (about 143 lmoles per
100 ml of the reaction mixture) via the reaction apparatus de-
scribed in Supporting data (Supplementary Fig. S1). Reaction
progress to ꢂ20% and 100% for KIE analysis was also monitored
spectrophotometrically.
Dehalogenation of DMCP catalyzed by DHP
Reaction of DHP with DMCP was conducted in the same way as
that for DHP with TCP. We have used a 1.9 mM solution of 2,6-
dimethyl-4-chlorophenol (DMCP) as an analog in 100 mM potas-
sium phosphate buffer, pH 7. The concentration of DHP in each
of the reactions was 300 lM. Again, the reactions were initiated
by addition of a calculated amount of H2O2 required to make the
reaction proceed to ꢂ20% and 100% for KIE analysis. Preparation
of this solution also required addition of ethanol due to low solu-
bility in pure water.
Experimental section
In order to avoid accidental quinone product degradation, as
was observed at higher temperatures [37], all reactions were car-
ried out at 4 °C.
Reagents
Nitric acid (pure p.a.), silver nitrite (pure p.a.), ethanol (special
pure), from POCh Gliwice, mercuric (II) thiocyanate (pure p.a.,
Fluka), ferric ammonium sulfate (L.P.P-H ‘‘OCh” Lublin, Poland),
concentrated sulphuric acid (suprapur, Merck), 2,6-dimethyl-4-
chlorophenol (DMCP) (97%, Acros Organics), 2,4,6-trichlorophenol
(TCP) (98%), 4-chlorophenol (4-CP) (99 + %), potassium phosphate
dibasic (Reagent Grade ꢀ98%), potassium phosphate monobasic
(Reagent Grade ꢀ99%), HRP, Type IV, from horseradish, sodium
biphenyl reagent, solution in 2-ethoxyethyl ether, from Sigma–
Samples preparation for Cl KIEs measurements
After stopping the reaction at the desired reaction progress (19–
20%), the mixture was filtered and silver chloride was precipitated
from the solutions with 0.2 M AgNO3 at room temperature. Next,
the precipitates of AgCl were washed with water three times and
were left to dry in a desiccator in the dark. Chlorine kinetic isotope
effect was calculated from the following equation [38]:

L. Szatkowski et al. / Archives of Biochemistry and Biophysics 505 (2011) 22–32
25
Fig. 1. The example of UV–Vis monitoring of the substrate conversion catalyzed by DHP. Panel A shows the spectrum of TCP disappearance and panel B presents the
dependence of the absorbance at 313 nm on the degree of substrate conversion. The reactions were conducted in 100 mM potassium phosphate buffer (pH 7.0) at 4 °C.
k35
k37
lnð1 ꢁ fÞ
where u = hm/kBT, h and kB are Planck and Boltzmann constants,
¼
ð1Þ
Rf
respectively, and T is the absolute temperature. Also, n is the
number of atoms, and vi are the frequencies of normal modes of
vibrations. The superscript ‘‘–” indicates the properties of the tran-
sition state.
lnð1 ꢁ f 1Þ
R
where f is the reaction progress, Rf is the chlorine isotopic ratio of
the product at the reaction progress f, and R is the chlorine isoto-
1
pic ratio of the completely converted chlorophenol. In the case of 4-
CP, this ratio was determined using Lassigne method [39] based on
the reaction of organic compound with metallic sodium, whereas in
the case of TCP, the complete conversion by the enzyme was used.
For DMCP, the Liggett method based on the reaction with sodium
biphenyl [40] was applied. Chlorine isotopic ratios were measured
using a hybrid FAB-IR mass spectrometer as described previously
[41] with a modified sample support that results in reduced sample
requirement. All of the measurements were repeated independently
three times.
To recognize the role of enzyme in the dechlorination step,
three different forms of substrate (phenolate, phenoxy radical
and cation) were considered, and the respective reactions in the
water environment using the same QM/MM scheme with TIP3P
water molecules as MM part representatives were modeled. Open
shell species were treated using unrestricted Hartree–Fock (UHF)
method [51]. Molecular geometries of all species were fully opti-
mized and vibrational analysis was carried out to confirm identity
of the stationary points (3n-6 real vibrations in the case of reac-
tants or products and one imaginary frequency corresponding to
the desired reaction coordinate in the case of transition state struc-
tures). The transition state structures have been located along the
lowest Hessian eigenvector using mixed Murtagh–Saragant/Powell
method [53] as implemented in Jaguar program [54]. The radical
intermediates, especially when 4-CP binds to enzymes, have a ten-
dency to form dimers or polymerize and either shut the reaction
down or escape from the reaction site. So although this route as
the alternative to water molecule attack on a cationic form of the
substrate sounds rather unlikely, we have decided to test this path-
way in our attempts on the mechanism elucidation.
Computational methods
General approach
In order to approach the mechanistic details of the dehalogen-
ation step catalyzed by DHP and HRP, the quantum mechanical/
molecular mechanical (QM/MM) method implemented in QSite
[42] has been used. The relaxed potential energy surfaces were
modeled. The QM parts of these systems consisting of substrate
molecule, either 4-chlorophenol (4-CP) or 2,4,6-trichlorophenol
(TCP), and selected water molecules were treated with B3LYP
[43,44] functional and lacvp* [45] basis set. Our choice of method
for modeling the reactions of interest was the compromise between
the methods which were found to be appropriate for heme-
containing enzymes on one hand and able to predict chlorine
isotope effects correctly on the other. Based on the data available
in the literature [46,47] the combination of B3LYP with lacvp basis
set seemed to be a good choice including open shell species [48]. On
the other hand the previous model studies of Fang et al. on chlorine
KIEs on SN2 reactions showed that both B1LYP and B3LYP functional
can predict the values of KIEs correctly [49] and hence can be suc-
cessfully used for modeling dehalogenation reaction. The MM part
was simulated using the OPLS-AA force field. All QM/MM simula-
tions have been performed in QSite. In order to include solvent
initial models have been soaked in 40 ꢃ 40 ꢃ 40 Å water box and
optimized to a minimum using the default convergence and optimi-
zation criteria. Calculations of KIEs were performed using ISOEFF98
program [50]. KIEs were obtained from the complete Bigeleisen
equation [38] at 300 K for the transition from proximity complexes
of both reactants to the corresponding transition states.
Modeling TCP dehalogenation in water environment
Each model studied consisted of a substrate molecule and 201
surrounding water molecules. Altogether the model comprised
615 atoms. The solvent molecules were included by soaking the
solute molecule into the 18 ꢃ 18 ꢃ 18 Å water box. After initial
minimization the solute and the four nearest water molecules
were defined as a QM part and were modeled using B3LYP/lacvp*
method in the subsequent QM/MM energy profile calculations.
During each step of the calculations, atoms beyond 8 Å from the
solute molecule were frozen. The reaction coordinate along which
the energy scan was performed was defined as a decreasing dis-
tance between the oxygen atom of the attacking water molecule
and the para position carbon of the substrate to which chlorine
was attached. For the case of phenoxy radical and cation pathways,
attack of a water molecule at ortho position was also considered.
The default convergence and optimization criteria were applied.
HRP-catalyzed reaction
3nsꢁ6
3n–ꢁ7
A model for the dehalogenation reaction catalyzed by HRP was
prepared based on the 1HCH X-ray structure [52]. The substrate
molecule (4-CP) was docked to the active site based on the position
s
h
B
h
vsi;l=2Þ
v
v
–i;lsinh k
ðv
–i;h=2Þ
–
Y
Y
v
i;hsinh k
ð
k1 v1
vsi;lsinh TTðvsi;h=2Þ
–i;hsinh T ð
v
–i;l=2Þ
BT
¼
ð2Þ
–
h
kB
h
kh vh
i
i
kB

26
L. Szatkowski et al. / Archives of Biochemistry and Biophysics 505 (2011) 22–32
of the first structure of a peroxidase–substrate complex demon-
strating the existence of an aromatic binding pocket [55]. This par-
ticular enzyme–substrate complex seemed to be a good candidate
for initial substrate position as the substrate, benzhydroamic acid
is a similar size molecule. However, the 1HCH structure is of better
resolution (1.57 vs. 2.00 Å in case of this reference structure) thus
we have decided to use it for our model preparation. After docking,
the entire structure was pre-minimized using OPLS-AA force field
[56–58] implemented in the Impact program. Then, in order to
evaluate which position (whether chlorine substituent or hydroxyl
group facing the heme iron) is favorable, the interactions analysis
between substrate molecule and surrounding residues has been
performed. The following amino acids have been chosen: Phe179,
Gln176, Gly69, Ala140, Leu138, Pro139, His42, Arg38, Phe142,
Arg178, Phe143, Ser73, Phe41 and Pro141. Those were the residues
found within 7 Å sphere from the substrate moiety. The interaction
energies were calculated based on fully optimized structures
where the substrate molecule, in either of the positions considered,
was treated quantum mechanically in the surrounding of MM pro-
tein and water atoms. Next, the optimized geometries of respective
substrate–amino acid pairs were taken out from the system and
the single point gas phase energies using the same level of theory
were calculated for a pair and separated components (100 Å away
to assure the total decay of interactions). The more negative values,
the larger the contribution to the stabilization of the substrate.
The pre-minimized system, with 4-CP in the proper orientation,
was soaked into the 40 ꢃ 40 ꢃ 40 Å water box. The total model
consisted of 7102 atoms, out of which 30 (4-CP plus 6 water mol-
ecules) were treated quantum mechanically. In an attempt to de-
scribe the dehalogenation reaction catalyzed by HRP at the
molecular level, a reaction coordinate has been defined as the dis-
tance between the chlorinated carbon and oxygen from the attack-
ing water molecule. Two alternative pathways have been taken
into consideration: via radical and via cation. During the energy
scan along the reaction coordinate, atoms beyond 15 Å from the
substrate molecule were frozen.
Fig. 2. Two possible external binding sites of DHP: panel A – near the heme edge,
panel B – dimer interface.
binding sites on DHP with the internal distal site above the heme,
we used the 1EWA X-ray structure of DHP [64] that has a 4-
iodophenol molecule bound internally so we could estimate
binding energies for both 4-CP and TCP. Then, using the optimized
structures, the substrate molecule was removed from each of them
by gradually elongating the distance between it and the rest of the
system until the total energy of the system stopped changing. The
binding energy was then estimated as the difference between
the substrate–enzyme complex and the separated species, where
the substrate was already separated by 100 Å.
Results and discussion
The values for chlorine kinetic isotope effects determined for 4-
chlorophenol (4-CP) and 2,4,6-trichlorophenol (TCP) dehalogen-
ation catalyzed by HRP and for TCP and 2,4-dimethyl-4-chlorophe-
nol (DMCP) catalyzed by DHP are shown in Table 1. Reaction of
DHP with 4-CP was not possible because 4-CP acts as an inhibitor
to DHP peroxidase activity (vida infra). Preliminary reactions of
DHP with TCP suggested that all chlorine atoms could be removed
which could affect the KIE results. For this reason, DMCP was se-
lected as a second substrate for DHP in order to ensure only the
para-halogen was extracted for KIE measurements.
The values clearly indicate that there is no kinetic isotope effect
on the reaction rate during the studied reactions. Several factors
may lead to such observations: (1) there is no isotopic fraction-
ation on this particular step of the reaction, (2) there is isotopic
fractionation but the dehalogenation step is masked by other steps
of the overall reaction, (3) the mechanism of this step is different
than one would expect for the reaction which seems to be a nucle-
ophilic attack on aromatic carbon atom, or (4) dehalogenation oc-
curs during a polymerization reaction as has been reported for
other reactions catalyzed by peroxidases [65].
DHP-catalyzed reaction
For the case of DHP, although there are enzyme crystal struc-
tures available, the task of preparing the appropriate enzyme mod-
el for the reaction was not so straightforward, as the binding site of
trihalophenols or any tri-substituted derivatives of phenol is not
known and internally bound monohalogenated phenols were
shown to act as inhibitors. Thus we have decided to explore the
idea of external binding site which has been recently proposed
[32]. For this purpose we have used the 2QFK structure [30] and
the two possible binding sites from Davis et al. have been taken
into consideration as shown in Fig. 2. One of them presents a heme
edge binding site whereas the other is located in the dimer inter-
face. Models with the bound substrate molecule were prepared
using the QSite program. A water molecule was coordinated to
the iron ion of heme. The resulting models comprised 2693 (heme
edge) and 5308 (dimer interface) atoms, respectively. In order to
estimate the binding energy of the substrate in each of the struc-
tures, we have also added solvent water molecules. We have used
Poisson–Boltzmann (PBF) [59–61] continuum model, and in each
case, the system was optimized using the default convergence
and optimization criteria. Only the TCP molecule in its anionic form
was defined as a QM part in the QM/MM simulation. DFT func-
tional M06-2X [62,63] with lacvp* basis set was applied for QM
calculations whereas the rest of the system was treated with
OPLS-AA force field. We have decided to use M06-2X functional in-
stead of B3LYP because it has been shown recently to be superior
to the latter for the study non-covalent interactions and calculating
interaction energies [63]. In order to compare the two external
In order to accommodate different mechanistic hypotheses we
have combined the experimentally determined values with
Table 1
Experimental Cl KIEs determined for different chlorophenols degradation catalyzed
by HRP and DHP at 277 K.
HRP
DHP
MCP
TCP
DMCP
1.0006 0.0005
0.9996 0.0002
nd
nd
1.0005 0.0001
0.9995 0.0004

L. Szatkowski et al. / Archives of Biochemistry and Biophysics 505 (2011) 22–32
27
theoretical prediction of isotope effects. Molecular modeling stud-
ies allowed us, on one hand, to get more detailed insight into the
reaction mechanism and, on the other, helped us to estimate the
role of the protein in the oxidation process under study.
TCP dehalogenation in water environment
The energy profiles for water attack at the aromatic carbon have
been modeled, and transition state structures have been deter-
mined for the phenolate and radical species. The geometries of
the saddle points and imaginary frequencies, as well as energetics
and isotope effects, are presented in the Table 2.
We have not calculated isotope effects for water attack at the
ortho aromatic carbon position for the radical species, but judging
from the geometry of the transition state, the KIE should not be dif-
ferent from the results obtained for the para position. For both the
anion and radical species, significant water deprotonation has been
observed during water attack which is reflected in both O–H dis-
tances and the inverse oxygen KIE. The high barrier heights for
the phenolate and phenoxy radical pathways suggest that these
two pathways are likely unfavorable for dechlorination. In fact,
the unrealistic value of 57.3 kcal/mol for the phenolate pathway
is expected since oxidation of the phenolate form occurs prior to
the dechlorination step.
Fig. 3. Overlaid structures of reactants shown in magenta and a Meisenheimer
complex (Mc) in yellow. Numbers indicate O–H bond lengths in Å in an attacking
water molecule.
However, in the case of the cationic form of the oxidized sub-
strate, the overall picture is quite different. No transition state
(TS) was found for the water attack as there was no barrier for this
process at either the ortho or para positions. The system forms a
kind of Meisenheimer complex (Mc) with a C–Cl bond length of
1.833/1.846 Å and C–O distance of 1.417/1.411 Å for ortho/para
positions, respectively. The QM/MM energies for this process are
ꢁ48.3/ꢁ28.7 kcal/mol for the ortho and para positions, and the lack
of barrier for this process indicates that it is spontaneous. Compar-
ison of the initial state with the formed stable complex reveals sig-
nificant water deprotonation as shown in the example of the
reaction at the para position in Fig. 3. This deprotonation seems
to occur just before the complex formation and may play a stabiliz-
ing role in the reaction.
Analysis of atomic charges derived from the electrostatic poten-
tial with monopoles located at the atomic centers [66–68] shows
significant charge migration from the benzene ring of the substrate
to the water molecules. In the reactants state, 0.87e was distrib-
uted within the aromatic ring whereas after Mc formation, this
charge is redistributed among water molecules. The biggest charge
changes are observed for C1 (from ꢁ0.44e in the reactant state to
0.00e in the Mc intermediate), C2 (from ꢁ0.44e to ꢁ0.28e), C5
(from 0.41e to 0.12e), and C6 (from ꢁ0.30e to 0.10e). The oxygen
atom from the attacking water molecule gained 0.32e and all of
the chlorine atoms show negative charge buildup, in particular
the departing one in the para position (Cl12) (atom numbering
shown in Fig. 4). This observed charge migration is accompanied
by single and double bond changes within the benzene ring toward
a quinone-like structure: bonds C1–C2 and C4–C5 get shortened by
0.03 and 0.02 Å, respectively, and bonds C6–C1 and C6–C5 get
elongated by 0.08 and 0.07 Å, respectively. Thus, the calculated
cationic pathway yields a charge redistribution and bond length
changes that closely resemble the quinone product structure.
Starting from the Meisenheimer complex, we have further elon-
gated the C–Cl bond to explore the nature of chlorine ion expulsion
in order to elucidate the mechanistic pathway that leads to re-
moval of the halogen substituents responsible for toxicity. In a
water environment, two energetic maxima were observed. The
first maximum, at about 7.6 kcal/mol, was recorded at a C–Cl bond
length of 2.34 Å, but it could not be optimized into any transition
state. Therefore, we concluded that it is most likely not related
with any bond breaking/forming act or such act proceeds via either
very flat barrier or without any at all, as the majority of this barrier
is in the MM term (4.4 kcal/mol) – Fig. 5, panel A. The second max-
imum, with 14.9 kcal/mol at a C–Cl bond length of 3.5 Å, can likely
be attributed to Clꢁ complexation with one of the water molecules
as illustrated in Fig. 5, panel B, green structure.
Within the QM/MM scheme implemented in QSite the total en-
ergy of the system is calculated based on the following equation:
Table 2
Calculated kinetic isotope effects and saddle point characteristics for dehalogenation
of TCP in anionic and radical forms in water at 300 K.
Phenolate
Phenoxy radical
p-position
p-position
o-position
37Cl KIE
1.0037
0.9998
1.0226
1.0085
0.9875
1.0364
18O KIE
13C KIE
Saddle point characteristics:
Geometry (Å):
C–Cl
1.866
1.981
1.937
C–O
1.740
1.666
1.589
O–H
1.021/0.974
ꢁ229
1.025/1.017
ꢁ353
1.065/0.974
ꢁ220
Frequency (cmꢁ1
Barrier (kcal/mol):
QM/MM
)
57.3
31.3
32.9
Fig. 4. Atom numbering in TCP molecule.

28
L. Szatkowski et al. / Archives of Biochemistry and Biophysics 505 (2011) 22–32
In order to provide mechanistic details of the dehalogenation
reaction of 4-CP catalyzed by HRP, we have computed the energy
profiles for the phenoxy radical and cationic pathways. Results
for the phenoxy radical pathway of 4-CP catalyzed by HRP are pre-
sented in Table 4. They are very similar to the results for the phen-
oxy radical pathway of TCP in a water environment. The chlorine
KIE is significant for the phenoxy radical pathway, which is in dis-
agreement with the experimentally observed value of 1.0006. The
calculated value of 1.0074 is within the typical range of 1.006–
1.008 [17] for chlorine KIEs with a moderately elongated C–Cl bond
length of about 0.3 Å. Although the experimental values were
determined at lower temperature, no significant change is ex-
pected at room temperature since no isotope effect was observed.
Also KIEs calculated at 277 K (TCP in water and HRP-catalyzed 4-
CP) do not show any significant difference. For the water mediated
reaction, we obtained a KIE of 1.0092 at 4 °C and for the HRP-
catalyzed reaction, we obtained a KIE of 1.0080. Moreover, the
energy barrier for the phenoxy radical pathway catalyzed by HRP
is high when compared to the uncatalyzed reaction (31.3 kcal/
mol). Thus we suspect that the disagreement of the KIEs and the
high energy barrier are further indications that the phenoxy radical
oxidation route is probably not adopted by peroxidases.
The formation of an enzyme–Meisenheimer complex (EMc) has
also been observed for the cationic pathway (Fig. 7). A stabilization
of ꢁ52 kcal/mol was observed for the cationic intermediate bound
to the enzyme model relative to that in a water environment
(ꢁ29 kcal/mol for TCP in water vs. ꢁ81 for 4-CP in the HRP active
site). However, this additional stabilization did not introduce a bar-
rier for this process, nor was there a calculated chlorine kinetic iso-
tope effect. Thus, the calculation for the cationic pathway agrees
with experiment.
We have also analyzed the interactions of the Meisenheimer
complex with the neighboring residues in the active site and com-
pared them with those seen for the enzyme–substrate complex.
The results of this comparison are shown in Fig. 8. It is evident that
the enzyme stabilizes the EMc complex and such stabilization, by
about 30 kcal/mol, is pronounced within both increasing positive
interaction and decreasing negative interaction with residues that
are interacting with 4-CP in a similar manner. While we are aware
that such an analysis should be treated with care and only qualita-
tively, it definitely indicates to some important patterns during
transformation of the ES into the EMc complex. The key geometri-
cal parameters for the enzyme–substrate complex as well as the
Meisenheimer complex are shown in Fig. 7.
Further analysis of the EMc formation reveals similar observa-
tions to the ones for the Mc intermediate in water. An analogous
charge migration has been observed from the aromatic ring to
the surrounding water molecules. The only difference is the charge
redistribution among the accompanying water molecules. For the
uncatalyzed reaction, most of the charge is transferred to the oxy-
gen of the attacking water, whereas for the catalyzed reaction, the
charge from the benzene ring is moved and redistributed on at
least three water molecules, forming a hydrogen-bonded, proton
donor–acceptor chain.
Fig. 5. Snapshots of the intermediate structures obtained during elongating C–Cl
bond in water. Panel A: three major structures related to the first observed
maximum on the energy profile, panel B: three major structures related to the
second observed maximum on the energy profile; in both cases, the starting point is
shown in magenta, the maximum in orange and the first point right after it in green.
EðQM=MMÞtot ¼ EðQMÞ þ EðMMÞ þ EðQM=MMÞint
ð3Þ
Where E(QM) represents the gas phase energy of the QM sub-
system, E(MM) is the classical energy of the remaining part of
the entire system, and E(QM/MM)int is the interaction energy be-
tween these two defined subsystems. Since only the total and
the MM energy values are reported during QM/MM calculations
in QSite, we have decided to explore the origin of the above find-
ings. For this purpose the QM subsystem has been extracted, and
its gas phase single point energy has been calculated using the
same level of theory as in the full QM/MM calculations. Based on
the obtained values, we could calculate the QM/MM interaction
energy. All energy values obtained within this analysis for the
two located maxima with respect to the initial state, which was
the Meisenheimer complex, are listed in Table 3. The QM gas phase
energy differences should not be treated as the exact barrier
heights for the theoretical gas phase process since they were not
obtained based on optimization but rather on single point energies
of the structures optimized using full QM/MM calculation. How-
ever, they can be treated along with derived interaction energies
as a clear indication of strong stabilization of the formed product.
These QM gas phase single point energies are heavily suppressed
by the QM/MM interactions which, as a result, give almost barrier-
less processes or processes with a very small, flat barrier.
HRP-catalyzed dehalogenation
In order to evaluate which binding orientation (whether chlo-
rine substituent or hydroxyl group facing the heme iron) is favor-
able, an interaction analysis between the 4-chlorophenol molecule
and surrounding amino acid residues has been performed. The fol-
lowing amino acids have been chosen: Phe179, Gln176, Gly69,
Ala140, Leu138, Pro139, His42, Arg38, Phe142, Arg178, Phe143,
Ser73, Phe41 and Pro141. The calculated interaction energy for
each substrate–AA pair is shown in the Fig. 6. On the basis of the
obtained values (ꢁ41.0 vs. ꢁ32.9 kcal/mol for O ? Fe and Cl ? Fe
position, respectively) it has been accepted that the orientation
with the oxygen facing the heme iron will be more favorable and
should be used in further mechanistic studies.
The subsequent exploration of Clꢁ departure revealed a lack of a
barrier for this process as well, which suggests that it may either
happen spontaneously or via a very small barrier. The most impor-
tant structures related to chlorine expulsion are shown in Fig. 9. By
analogy to the reaction studied in water environment, we also per-
formed a similar analysis of all the energy terms for this observed
maximum. With respect to the Meisenheimer complex, the maxi-
mum at 2.89 Å is characterized by the total QM/MM energy of
22 kcal/mol out of which 15.5 kcal/mol is the classical energy,
19.8 kcal/mol is the QM gas phase single point energy and the
interaction energy between the QM and MM subsystems is
ꢁ13.3 kcal/mol.
Table 3
Relative energies (kcal/mol) of maxima located during C–Cl elongation in a water
environment with respect to the initial Meisenheimer complex.
QM/MMtot
MM
QM
QM/MMint
1st maximum
2nd maximum
7.7
14.9
4.4
8.9
15.7
45.1
ꢁ12.4
ꢁ39.1

L. Szatkowski et al. / Archives of Biochemistry and Biophysics 505 (2011) 22–32
29
Fig. 6. Interaction energy between bound 4-chlorophenol and neighboring residues in the active site of horseradish peroxidase. Red and blue bars denote substrate–
aminoacid interaction energy for chlorine- and oxygen-facing the heme iron orientation, respectively.
Table 4
Calculated kinetic isotope effects and saddle point character-
istics for dehalogenation of 4-CP in radical form catalyzed by
HRP at 300 K.
Phenoxy radical
37Cl KIE
1.0074
0.9860
1.0379
18O KIE
13C KIE
Saddle point characteristics:
Geometry (Å):
C–Cl
1.951
C–O
1.664
O–H
1.032/1.018
ꢁ370
Frequency (cmꢁ1
Barrier (kcal/mol):
QM/MM
)
47.0
DHP-catalyzed dehalogenation
Recently, Thompson et al. [29] reported a unique two-site
competitive binding mechanism in DHP in which 4-CP and TCP
compete for an internal and an external binding site in/on DHP,
respectively. Binding of each molecule to its respective site pro-
duces a conformational change in the distal His55 amino acid.
Internal binding of 4-CP expels the iron-coordinated water mole-
cule and pushes His55 to a solvent exposed open conformation
resulting in a 5-coordinate high spin heme iron. Thus, internal
binding of 4-CP blocks access to the heme iron and disrupts the
peroxidase acid/base catalyst. This is in fact the mechanism by
which 4-CP inhibits the peroxidase activity of DHP and why we
were unable to use 4-CP for KIE analysis. External binding of TCP
pushes His55 into a closed conformation resulting in a predomi-
nately 6-coordinate high spin heme and positioning His55 to serve
as the peroxidase acid/base catalyst. Thus external binding primes
peroxidase function.
Fig. 7. The geometries of the enzyme–substrate (ES) and the enzyme–Meisenhei-
mer complex (EMc) formed during the attack of water molecule on the cationic
form of 4-CP in the active site of HRP. The numbers indicate the key geometrical
parameters.
In order to explore the idea of an external binding site, the 2QFK
structure was used with His55 in its closed confirmation. Two
models comprising two possible external binding sites, obtained
from locations presented by Davis et al. [32], were prepared and
compared to internal binding in the distal pocket of the heme.
Using the M06-2X/OPLS-AA calculation scheme implemented in
QSite, we have estimated the binding energy for each of the sites.
We decided to perform the analysis using a newly (at that time)
developed functional (M06-2X) as it was recommended to be more
appropriate for non-covalent interactions [63]. The results are pre-
sented in Fig. 10. The discrimination between monohalophenols
and trihalophenols in the distal pocket is clearly evident, which

30
L. Szatkowski et al. / Archives of Biochemistry and Biophysics 505 (2011) 22–32
Fig. 8. The comparison of interaction energies between bound 4-chlorophenol (S) and formed Meisenheimer complex (Mc) and the neighboring residues in the active site of
horseradish peroxidase. Red and blue bars denote substrate– and Meisenheimer complex–aminoacid interaction energy, respectively.
pressed in both the QM and QM/MMint energies whereas in the
case of DHP, although the sum of these two terms gives almost
the same value, it differs significantly in the contribution of each
of the terms as illustrated in Fig. 11. However, this observation
should be treated with care and can only be discussed at the qual-
itative level since in the case of DHP it is a preliminary result, and it
is based on a speculative binding site. Despite these obvious limi-
tations, this finding, if confirmed, can be another indication of the
pattern observed so far for oxidoreductase-catalyzed dehalogen-
ation. Two studied enzymes, a classical peroxidase and a globin
with enhanced peroxidase activity, show the same action on halo-
genated substrates. During the initial steps of their reaction via two
single electron transfer events, they prepare the organic substrate
to be ready for chloride ion expulsion which then apparently oc-
curs without any energetic cost to the protein. The intermediate
formed along the reaction pathway resembles the so called
Meisenheimer complex which has already been reported for
aliphatic dehalogenases, e.g. for the reaction catalyzed by 4-
chlorobenzoyl-CoA dehalogenase [69,70].
The QM/MM calculations for chlorophenol dehalogenation
strongly suggest that the cationic pathway is the most favorable
for chlorine expulsion. The results are in reasonable agreement
with a van’t Hoff analysis of experimental Michaelis–Menten
kinetics of DHP catalyzed TCP oxidation that showed an overall
activation energy of 13.4 kcal/mol [37]. Although the calculated
energetics cannot be compared directly to the experimental activa-
tion energy as the latter one comprises all the steps from substrate
binding to the product release, the mechanistic description ob-
tained with modeling should easily be accommodated within the
scenario drawn based on this experimental kinetic data.
Fig. 9. Overlaid structures obtained on C–Cl elongation pathway for MCP catalyzed
by HRP. EMc complex is shown in green, the maximum on the energy profile in
orange and the final – product structure in magenta.
is likely a partial confirmation of the inhibiting properties of the
former.
Although, on the basis of the obtained data, it is not possible to
differentiate between the two speculated external binding sites as
this issue requires further exploration using different techniques,
we have decided to use the heme edge site, in agreement with
the HRP mechanism, in order to model the solvent water attack
on the cationic form of the substrate. We have prepared the appro-
priate model from the 2QFK structure consisting of almost 14,000
atoms in an analogous way as we did for the case of HRP. 21 atoms
were assigned to the QM part of the system: TCP plus three water
molecules. The reaction coordinate was defined in the same way as
previously. Again, a stable cationic intermediate formation was ob-
served without any barrier at a C–O distance of 1.4 Å and a C–Cl
bond length of 1.81 Å. The QM/MM energy of this complex is
ꢁ68.5 kcal/mol compared to the starting enzyme–substrate com-
plex. Further exploration of the nature of this stabilization as com-
pared to uncatalyzed and HRP-catalyzed reactions revealed some
differences between the two enzymes. An analysis of all energy
terms, in case of HRP, the stabilization of the QM subsystem is ex-
Conclusions
In the present study we have investigated the catalytic power of
two enzymes of different origin but having very similar activity
toward halophenols. Using a complementary approach of kinetic
isotope effects and QM/MM calculations, a picture of the mecha-
nism of removing halogens by this group of enzymes, peroxidases,
can be drawn. The obtained results allow us to make a working
hypothesis that the peroxidase activity of the presented enzymes
is mostly focused on a transformation of the toxic halogenated
substrate into a cationic intermediate form which can easily expel
a halogen substituent to form a lesser toxic compound. For both

L. Szatkowski et al. / Archives of Biochemistry and Biophysics 505 (2011) 22–32
31
Fig. 10. The estimated binding energies for MCP and TCP bound in the different sites of DHP.
Committee for Scientific Research, Poland grant (3 T09B 007 29). SF
and MKT acknowledge the US Army Research Office (Grant#
52278-LS) and the National Science Foundation (grant OISE
0651876).
Appendix A. Supplementary data
Supplementary data associated with this article can be found, in
the online version, at doi:10.1016/j.abb.2010.09.018.
References
[1] A.M. Klibanov, B.N. Alberti, E.D. Morris, L.M. Felshin, J. Appl. Biochem. 2 (5)
(1980) 414–421.
[2] A.M. Klibanov, T.M. Tu, K.P. Scott, Science 221 (4607) (1983) 259–261.
[3] A.M. Klibanov, E.D. Morris, Enzyme Microb. Technol. 3 (2) (1981) 119–122.
[4] N. Singh, J. Singh, Prep. Biochem. Biotechnol. 32 (2) (2002) 127–133.
[5] V.A. Cooper, J.A. Nicell, Water Res. 30 (4) (1996) 954–961.
[6] E. Miland, R.S. Malcolm, O.F. Ciaran, J. Chem. Technol. Biotechnol. 67 (3) (1996)
227–236.
[7] M.D. Aitken, Chem. Eng. J. 52 (2) (1993) B49–58.
[8] M.D. Aitken, Chem. Eng. J. 52 (2) (1993) B49–B58.
[9] J. Karam, J.A. Nicell, J. Chem. Technol. Biotechnol. 69 (2) (1997) 141–153.
[10] J.A. Nicell, Interdiscip. Environ. Rev. 3 (1) (2001) 14–41.
[11] G.W. Gribble, Chemosphere 52 (2) (2003) 289–297.
[12] M.M. Haggblom, I.D. Bossert (Eds.), Dehalogenation, Microbial processes and
Environmental Applications, Kluwer Academic Publishers, 2003.
[13] D.B. Janssen, I.J. Dinkla, G.J. Poelarends, P. Terpstra, Environ. Microbiol. 7 (12)
(2005) 1868–1882.
[14] K. Furukawa, Biosci. Biotechnol. Biochem. 70 (10) (2006) 2335–2348.
[15] M.H. Gold, D.K. Joshi, K. Valli, H. Wariishi, Bioremediation Chlorinated
Polycyclic Aromat. Hydrocarbon Compd. (1994) 231–238.
[16] P. Paneth, Acc. Chem. Res. 36 (2) (2003) 120–126.
[17] P. Paneth, Chlorine Isotope Effects in Biological Systems, in: A. Kohen, H.H.
Limbach (Eds.), Isotope Effects in Chemistry and Biology, CRC Press, 2006.
[18] T. Kurihara, N. Esaki, Chem. Rec. 8 (2) (2008) 67–74.
[19] N. Duran, E. Esposito, Appl. Catal. B: Environ. 28 (2) (2000) 83–99.
[20] Y.-P. Chen, S.A. Woodin, D.E. Lincoln, C.R. Lovell, J. Biol. Chem. 271 (9) (1996)
4609–4612.
[21] J. Belyea, L.B. Gilvey, M.F. Davis, M. Godek, T.L. Sit, S.A. Lommel, S. Franzen,
Biochemistry 44 (2005) 15637–15644.
[22] S. Franzen, L.B. Gilvey, J.L. Belyea, Biochimica et Biophysica Acta 1774 (2007)
121–130.
[23] J. Feducia, R. Dumarieh, L.B.G. Gilvey, T. Smirnova, S. Franzen, G.A. Ghiladi,
Biochemistry 48 (5) (2009) 995–1005.
[24] L. Lebioda, M.W. LaCount, E. Zhang, Y.P. Chen, K. Han, M.M. Whitton, D.E.
Lincoln, S.A. Woodin, Nature 401 (6752) (1999) 445.
Fig. 11. Relative energies (kcal/mol) of Mc complex formed in water, DHP and HRP
with respect to the corresponding substrate (in cation form) state.
HRP and DHP, the dehalogenation reaction takes place at an exter-
nal site, and the QM/MM calculations suggest that the HRP- and
DHP-catalyzed dehalogenation reactions happen in a very similar
fashion as the same reaction in solution without the enzyme. The
methodology used in the presented study clearly indicates that
the dehalogenation reactions catalyzed by these enzymes are side
reactions or other reaction channels initiated by electron transfer
that do not demand much effort from the enzyme.
Biocatalysts functioning in this way might be useful tools in the
fight against halogenated contaminants, especially when their
range of reactivity has a tendency to expand, and there are exam-
ples of multifunctional enzymes such as DHP, C. fumago chloroper-
oxidase [71,72], and lignin and manganese peroxidase [15] that
have dehalogenating activity in addition to other basic functions.
[25] K. Nienhaus, E. Nickel, M.F. Davis, S. Franzen, G.U. Nienhaus, Biochemistry 47
(49) (2008) 12985–12994.
[26] K. Nienhaus, P.C. Deng, J. Belyea, S. Franzen, G.U. Nienhaus, J. Phys. Chem. B
110 (26) (2006) 13264–13276.
[27] M.F. Davis, H. Gracz, F.A.P. Vendeix, V. de Serrano, A. Somasundaram, S.M.
Decatur, S. Franzen, Biochemistry 48 (10) (2009) 2164–2172.
Acknowledgments
Thanks to Piotr Paneth and Vicente Moliner for helpful com-
ments and fruitful discussions. ADD, LS and RK acknowledge State

32
L. Szatkowski et al. / Archives of Biochemistry and Biophysics 505 (2011) 22–32
[28] E. Zhang, Y.P. Chen, M.P. Roach, D.E. Lincoln, C.R. Lovell, S.A. Woodin, J.H.
Dawson, L. Lebioda, Acta Crystallogr. Sect. D 52 (Pt 6) (1996) 1191–1193.
[29] M.K. Thompson, M.F. Davis, V. de Serrano, F.P. Nicoletti, B.D. Howes, G.
Smulevich, S. Franzen, Biophys. J. 99 (5) (2010) 1586–1595.
[30] V. de Serrano, Z. Chen, M.F. Davis, S. Franzen, Acta Cryst. Sect. D 63 (Pt 10)
(2007) 1094–1101.
[49] Y. Fang, Y. Gao, P. Ryberg, J. Eriksson, M. Kolodziejska-Huben, A. Dybala-
Defratyka, S. Madhavan, R. Danielsson, P. Paneth, O. Matsson, K.C. Westaway,
Chem. Eur. J. 9 (2003) 2696–2709.
[50] V. Anisimov, P. Paneth, J. Mathem. Chem. 26 (1–3) (1999) 75–86.
[51] J.A. Pople, R.K. Nesbet, J. Chem. Phys. 22 (1954) 571–572.
[52] G.I. Berglund, G.H. Carlsson, A.T. Smith, H. Szöke, A. Henriksen, J. Hajdu, Nature
417 (6887) (2002) 463–468.
[53] B.A. Murtagh, R.W.H. Sargent, Comp. J. 13 (1970) 185.
[54] Jaguar, version 7.5, Schrödinger, LLC, New York, NY, 2008
[55] A. Henriksen, D.J. Schuller, K. Meno, K.G. Welinder, A.T. Smith, M. Gajhede,
Biochemistry 37 (22) (1998) 8054–8060.
[56] W.L. Jorgensen, D.S. Maxwell, J. Tirado-Rives, J. Am. Chem. Soc. 118 (45) (1996)
11225–11236.
[31] C. Zuxu, V. de Serrano, L. Betts, S. Franzen, Acta Cryst. Sect. D 65 (Pt 1) (2009)
34–40.
[32] M.F. Davis, B.G. Bobay, S. Franzen, Biochemistry 49 (6) (2010) 1199–1206.
[33] K. Yamada, T. Shibuya, M. Noda, N. Uchiyama, A. Kashiwada, K. Matsuda, M.
Hirata, Biosci. Biotechnol. Biochem. 71 (2007) 2503–2510.
[34] A. Kohen, H.H. Limbach, Isotope Effects in Chemistry and Biology, CRC Press,
2006.
[35] A. Dybala-Defratyka, M. Rostkowski, P. Paneth, Arch. Biochem. Biophys. 474
(2) (2008) 274–282.
[57] W.L. Jorgensen, J. Chandrasekhar, J.D. Madura, J. Chem. Phys. 79 (2) (1983)
926–935.
[36] A. Dybala-Defratyka, R.A. Kwiecien, D. Sicinska, P Paneth, Intrinsic isotope
effects – the Holy Graal of studies of enzyme-catalyzed reactions in modelling
molecular structure and reactivity in biological systems, in: J. Brady, J. Gao, K.J.
Naidoo, M. Hann, M. Field (Eds.), Royal Society of Chemistry, (UK) Special
Publication S. No. 303, USA, 2006, pp. 163–173.
[37] H. Ma, M.K. Thompson, J. Gaff, S. Franzen, Kinetic analysis of a naturally
occurring bioremediation enzyme: Dehaloperoxidase-hemoglobin from
Amphitrite ornate, J. Phys. Chem. A, accepted.
[58] N.A. McDonald, W.L. Jorgensen, J. Phys. Chem. B 102 (41) (1998) 8049–8059.
[59] D.J. Tannor, B. Marten, R.B. Murphy, R.A. Friesner, D. Sitkoff, A. Nicholls, M.N.
Ringnalda, W.A. Goddard III, B. Honig, J. Am. Chem. Soc. 116 (26) (1994)
11875–11882.
[60] B. Marten, K. Kim, C. Cortis, R.A. Friesner, R.B. Murphy, M.N. Ringnalda, D.
Sitkoff, B. Honig, J. Phys. Chem. 100 (28) (1996) 11775–11788.
[61] M. Friedrichs, R.H. Zhou, S.R. Edinger, R.A. Friesner, J. Phys. Chem. B 103 (16)
(1999) 3057–3061.
[38] L. Melander, W.H. Saunders Jr., Reaction Rates of Isotopic Molecules, Wiley and
Sons, New York, 1980.
[39] G. Widmark, Acta Chem. Scand. 7 (1953) 1395–1406.
[40] M. Numata, N. Nakamura, H. Koshikawa, Y. Terashima, Anal. Chim. Acta 455
(1) (2002) 1–9.
[62] Y. Zhao, D.G. Truhlar, Acc. Chem. Res. 41 (2) (2008) 157–167.
[63] Y. Zhao, D.G. Truhlar, Theor. Chem. Acc. 120 (1–3) (2008) 215–241.
[64] M.W. LaCount, E. Zhang, Y.P. Chen, K. Han, M.M. Whitton, D.E. Lincoln, S.A.
Woodin, L. Lebioda, J. Biol. Chem. 275 (25) (2000) 18712–18716.
[65] J. Dec, J.-M. Bollag, Environ. Sci. Technol. 28 (3) (1994) 484–490.
[66] L.E. Chirlian, M.M. Francl, J. Comput. Chem. 8 (1987) 894.
[67] R.J. Woods, M. Khalil, W. Pell, S.H. Moffat, V.H. Smith Jr., J. Comput. Chem. 11
(1990) 297.
[68] C.M. Breneman, K.B. Wiberg, J. Comput. Chem. 11 (1990) 361.
[69] D. Xu, Y. Wei, J. Wu, D. Dunaway-Mariano, H. Guo, Q. Cui, J. Gao, J. Am. Chem.
Soc. 126 (42) (2004) 13649–13658.
[70] J. Wu, D. Xu, X. Lu, C. Wang, H. Guo, D. Dunaway-Mariano, Biochemistry 45 (1)
(2006) 102–112.
[41] K.C. Westaway, T. Koerner, Y.-R. Fang, J. Rudzinski, P. Paneth, Anal. Chem. 70
(17) (1998) 3548–3552.
[42] QSite, version 5.0, Schrödinger, LLC, New York, 2007.
[43] C. Lee, W. Yang, R.G. Parr, Phys. Rev. B 37 (2) (1988) 785–789.
[44] A.D. Becke, J. Chem. Phys. 98 (7) (1993) 5648–5652.
[45] P.J. Hay, W.R. Wadt, J. Chem. Phys. 82 (1) (1985) 299–310.
[46] D. Fishelovitch, C. Hazan, H. Hirao, H.J. Wolfson, R. Nussinov, S. Shaik, J. Phys.
Chem. B 111 (49) (2007) 13822–13832.
[47] W. Lai, H. Chen, K.-B. Cho, S. Shaik, J. Phys. Chem. A 113 (43) (2009) 11763–
11771.
[71] R.L. Osborne, G.M. Raner, L.P. Hager, J.H. Dawson, J. Am. Chem. Soc. 128 (4)
(2006) 1036–1037.
[48] K. Kühnel, E. Derat, J. Terner, S. Shaik, I. Schlichting, Proc. Natl. Acad. Sci. USA
104 (2007) 99–104.
[72] R.L. Osborne, M.K. Coggins, J. Terner, J.H. Dawson, J. Am. Chem. Soc. 129 (48)
(2007) 14838–143839.