Revised molecular basis of the promiscuous carboxylic acid perhydrolase activity in serine hydrolases

Article abstract of DOI:10.1002/chem.201200052

Several serine hydrolases catalyze a promiscuous reaction: perhydrolysis of carboxylic acids to form peroxycarboxylic acids. The working hypothesis is that perhydrolases are more selective than esterases for hydrogen peroxide over water. In this study, we tested this hypothesis, and focused on L29P-PFE (Pseudomonas fluorescens esterase), which catalyzes perhydrolysis of acetic acid 43-fold faster than wild-type PFE. This hypothesis predicts that L29P-PFE should be approximately 43-fold more selective for hydrogen peroxide than wild-type PFE, but experiments show that L29P-PFE is less selective. The ratio of hydrolysis to perhydrolysis of methyl acetate at different concentrations of hydrogen peroxide fit a kinetic model for nucleophile selectivity. L29P-PFE (β₀=170 M^-1) is approximately half as selective for hydrogen peroxide over water than wild-type PFE (β₀=330 M ^-1), which contradicts the working hypothesis. An alternative hypothesis is that carboxylic acid perhydrolases increase perhydrolysis by forming the acyl-enzyme intermediate faster. Consistent with this hypothesis, the rate of acetyl-enzyme formation, measured by ¹⁸O-water exchange into acetic acid, was 25-fold faster with L29P-PFE than with wild-type PFE, which is similar to the 43-fold faster perhydrolysis with L29P-PFE. Molecular modeling of the first tetrahedral intermediate (T_d1) suggests that a closer carbonyl group found in perhydrolases accepts a hydrogen bond from the leaving group water. This revised understanding can help design more efficient enzymes for perhydrolysis and shows how subtle changes can create new, unnatural functions in enzymes. Copyright

Full text of DOI:10.1002/chem.201200052

DOI: 10.1002/chem.201200052
Revised Molecular Basis of the Promiscuous Carboxylic Acid Perhydrolase
Activity in Serine Hydrolases
[
a]
DeLu (Tyler) Yin and Romas J. Kazlauskas*
Abstract: Several serine hydrolases cat-
alyze a promiscuous reaction: perhy-
drolysis of carboxylic acids to form per-
oxycarboxylic acids. The working hy-
pothesis is that perhydrolases are more
selective than esterases for hydrogen
peroxide over water. In this study, we
tested this hypothesis, and focused on
L29P-PFE (Pseudomonas fluorescens
esterase), which catalyzes perhydrolysis
of acetic acid 43-fold faster than wild-
type PFE. This hypothesis predicts that
L29P-PFE should be approximately 43-
fold more selective for hydrogen per-
oxide than wild-type PFE, but experi-
ments show that L29P-PFE is less se-
lective. The ratio of hydrolysis to per-
hydrolysis of methyl acetate at differ-
ent concentrations of hydrogen perox-
ide fit a kinetic model for nucleophile
pothesis, the rate of acetyl-enzyme for-
1
8
mation, measured by O-water ex-
change into acetic acid, was 25-fold
faster with L29P-PFE than with wild-
type PFE, which is similar to the 43-
fold faster perhydrolysis with L29P-
PFE. Molecular modeling of the first
ꢀ
1
selectivity. L29P-PFE (b =170m ) is
0
approximately half as selective for hy-
drogen peroxide over water than wild-
ꢀ
1
type PFE (b =330m ), which contra-
0
dicts the working hypothesis. An alter-
native hypothesis is that carboxylic
acid perhydrolases increase perhydroly-
sis by forming the acyl-enzyme inter-
mediate faster. Consistent with this hy-
tetrahedral intermediate (T 1) suggests
d
that a closer carbonyl group found in
perhydrolases accepts a hydrogen bond
from the leaving group water. This re-
vised understanding can help design
more efficient enzymes for perhydroly-
sis and shows how subtle changes can
create new, unnatural functions in en-
zymes.
Keywords: enzyme catalysis · en-
zyme kinetics · enzymes · perhy-
drolysis · protein engineering
Introduction
Peroxycarboxylic acids, such as peracetic acid, are important
reagents to add oxygen to olefins, ketones, aldehydes, transi-
[1–4]
tion metals, and organic sulfur and phosphorus.
Peracids
are disinfectants because they kill microorganisms and inac-
[5,6]
tivate viruses.
Peracids also remove lignin from lignocel-
lulose to brighten pulp for paper making or pretreat biomass
[7–9]
for biofuel production.
The oxidizing power of peracids
is the same as for hydrogen peroxide; their advantage is that
they are much faster oxidants than hydrogen peroxide. The
hydroxide leaving group in hydrogen peroxide is replaced
by the better acetate leaving group in peracetic acid. One
disadvantage of peracids is that concentrated solutions
Figure 1. Perhydrolases catalyze the reaction of hydrogen peroxide with:
a) carboxylic acids, or b) carboxylic acid esters to make peroxycarboxylic
acids. This study focuses on the perhydrolysis of acetic acid to make per-
acetic acid.
[10]
(
>70 vol%) can explode. This hazard makes them hard to
sis of carboxylic acid esters have different requirements, so
perhydrolases are usually efficient for only one of these re-
actions. The focus of this study was perhydrolysis of carbox-
ylic acids. Perhydrolysis of carboxylic acids is reversible and
thermodynamically controlled. It is best suited for the con-
tinuous generation of a low steady state concentration of
peroxycarboxylic acids.
store and expensive to transport. In situ formation of pera-
cids avoids these problems.
Perhydrolysis is the formation of peracids from carboxylic
acids or carboxylic acid esters and hydrogen peroxide. Per-
[11]
hydrolases (initially called metal-free haloperoxidases)
are enzymes that catalyze either of these reactions
Figure 1). Perhydrolysis of carboxylic acids and perhydroly-
(
All perhydrolases identified so far resemble esterases
since they have an a/b hydrolase fold and contain a Ser-His-
Asp catalytic triad in their active site. Given the reactive
nature of peracids and low natural concentrations of hydro-
[
a] D. Yin, Prof. R. J. Kazlauskas
University of Minnesota, Department of Biochemistry
Molecular Biology & Biophysics and The Biotechnology Institute
[12]
gen peroxide, it is unlikely that perhydrolysis is the natu-
ral function of these enzymes. Instead, perhydrolysis is an
1
479 Gortner Avenue, Saint Paul, MN 55108 (USA)
[13]
E-mail: rjk@umn.edu
accidental (promiscuous) function of these enzymes. Un-
8130
ꢀ 2012 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2012, 18, 8130 – 8139

FULL PAPER
derstanding the molecular basis of perhydrolysis is an impor-
tant step to the design of more efficient enzymes for this re-
action and to better grasp how new, unnatural functions
occur in enzymes.
The catalytic mechanism for serine hydrolase catalyzed
perhydrolysis is similar to that for hydrolysis (Figure 2a).
The catalytic serine reacts with acetic acid to form an
[14]
acetyl-enzyme intermediate. This first step creates a tetra-
hedral intermediate, T 1, which loses water to form the
d
acetyl-enzyme intermediate. In the second step, hydrogen
peroxide reacts with acetyl-enzyme through a second tetra-
hedral intermediate, T 2, to form the product peracetic acid.
d
The overall kinetic mechanism is ping-pong bi-bi, in which
the first substrate, acetic acid, reacts with the enzyme to
form the acetyl-enzyme intermediate, and then the second
substrate, hydrogen peroxide, reacts with this intermediate
to form the product peracetic acid. The catalytic histidine is
alternately a base and acid during the reaction and the oxy-
anion hole (two main chain NꢀH groups) donates hydrogen
bonds to stabilize the tetrahedral intermediates and the
transition states.
The free energy profile of the perhydrolysis of acetic acid
involves multiple transition states, two tetrahedral inter-
mediates and an acetyl-enzyme intermediate (Figure 2b).
This free energy diagram applies under standard conditions
(
1m concentrations) for reactants and products. The similar
free energy levels of the starting materials and products re-
flects the equilibrium constant for perhydrolysis, which
slightly favors perhydrolysis (Keq ~3). At the start of the re-
action when the concentration of product is much lower
than at standard conditions, formation of product is strongly
favored and the free energy profile differs from that shown.
The free energy of the products at the start of the reaction
is much lower because their concentration is also lower than
at standard conditions. As the product accumulates, its free
energy increases until at equilibrium it is equal to that of
the remaining starting materials.
The rate determining step for the perhydrolysis of acetic
acid is the first step: formation of acetyl-enzyme intermedi-
ate. Both tetrahedral intermediates have similar free ener-
gies at standard conditions, but at the start of the reaction,
Figure 2. Serine-hydrolase-catalyzed perhydrolysis of acetic acid. a) Per-
hydrolysis of acetic acid follows the ping-pong bi-bi mechanism in which
the first substrate (acetic acid) binds and reacts with the enzyme to form
an acetyl-enzyme intermediate. Next, the second substrate (hydrogen
peroxide) reacts with acetyl-enzyme to form the product peracetic acid.
Formation of the acetyl-enzyme involves attack of S94-Og on acetic acid
T 2 is lower. The low concentration of products lowers the
d
free energy of the products and associated intermediates, in-
cluding T 2.
d
There are two ways to increase the rate of perhydrolysis.
The first is to increase the rate of formation of the acetyl-
enzyme intermediate. This formation is the rate-limiting
step for perhydrolysis of acetic acid because the acetyl-
enzyme ester link from acetic acid and the active site serine
hydroxyl is unfavorable. This first hypothesis predicts that
L29P-PFE will form the acetyl-enzyme approximately 43-
times faster than wild-type PFE.
to form a tetrahedral intermediate, T
the acetyl-enzyme involves a second tetrahedral intermediate, T
d
1, which releases water. Release of
2, which
d
adds hydrogen peroxide. The amino acid numbering corresponds to the
amino acids in PFE. The third member of the catalytic triad, Asp222, is
not shown for clarity. b) Approximate free energy diagram for perhydrol-
ysis of acetic acid involves an acetyl-enzyme intermediate (Ac-E) and
two high energy tetrahedral intermediates (T 1 and T 2). Nucleophilic
d d
attack of the active site serine on acetic acid leads to a high energy tran-
sition state (acetyl-H O TS) and a corresponding tetrahedral intermedi-
2
A second way is to increase the proportion of acetyl-
enzyme that goes on to form peracetic acid rather than hy-
drolyze to reform acetic acid. This second hypothesis pre-
dicts that L29P-PFE is approximately 43-fold more selective
than wild-type PFE for hydrogen peroxide over water be-
ate (T
enzyme intermediate. The binding of hydrogen peroxide forms another
high energy transition state (acetyl-H TS) and the corresponding tet-
rahedral intermediate (T 2). The release of peracetic acid results in
a slightly lower energy than the starting materials because the overall
equilibrium constant slightly favors peracetic acid (Keq ꢁ3).
d d
1). Collapse of T 1 and the release of water forms the acetyl-
2
O
2
d
Chem. Eur. J. 2012, 18, 8130 – 8139
ꢀ 2012 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
8131

R. J. Kazlauskas and D. Yin
cause L29P-PFE catalyzes perhydrolysis 43-fold faster than
wild-type PFE.
Table 1. Apparent steady state kinetic constant for perhydrolysis and hy-
drolysis (in parentheses) of methyl acetate catalyzed by wild-type PFE
and L29P-PFE.
In this study, we tested these predictions and ruled out
higher selectivity as the reason, and found support for faster
acetyl-enzyme formation. We propose a new molecular basis
for carboxylic acid perhydrolysis that is consistent with the
kinetics and with X-ray crystal structures.
app
cat
app
M
app
Enzyme
k
K
kcat/KM
ꢀ
1
ꢀ1
ꢀ1
A
H
U
G
R
N
N
]
[mm]
A[ s
C
H
T
U
N
G
T
R
E
N
N
U
N
G
m ]
wild-type PFE perhy-
60ꢂ6 (43ꢂ3) 300 (600)
[
a]
drolysis (hydroly-
[b]
sis)
L29P-PFE perhydroly- 8.7ꢂ0.5 (7.7ꢂ0.2) 420ꢂ70 (50ꢂ5) 20 (200)
sis (hydrolysis)
Results
[a] Perhydrolysis of methyl acetate was measured at 238C in citrate
buffer (100 mm, pH 6.5) by monitoring the absorbance at 290 nm. The
concentration of methyl acetate was varied whereas the concentration of
hydrogen peroxide was constant at 10 mm. [b] Hydrolysis of methyl ace-
tate was measured colorimetrically by using 0.45 mm p-nitrophenol as the
pH indicator at pH 7.2 with BES buffer (5 mm).
L29P-PFE is a tenfold better hydrolase than perhydrolase
for methyl acetate: Previous work established that L29P-
PFE catalyzes perhydrolysis of acetic acid 43-fold faster
[14]
(
kcat) than wild-type PFE. However, by using acetic acid
as the substrate hides the hydrolytic abilities of the enzymes.
To compare both the perhydrolysis and hydrolysis abilities
of L29P-PFE and wild-type PFE, we tested an ester sub-
strate, methyl acetate. Methyl esters are the only good ester
substrates for L29P-PFE because esters with larger alcohol
groups, like ethyl or phenyl, do not fit in the smaller active
Measuring the selectivity of the acetyl-enzyme intermedi-
ate for water versus hydrogen peroxide requires an ester
substrate. A carboxylic acid substrate hides the hydrolysis
rate since the hydrolysis product cannot be distinguished
from the starting material. Both ester perhydrolysis and acid
perhydrolysis involve the same acetyl-enzyme intermediate,
so an ester substrate reveals information about the second
step of acid perhydrolysis.
[14]
site of L29P-PFE.
The second hypothesis—that L29P-
PFE is more selective for hydrogen peroxide over water
than wild type—makes two predictions. First, L29P-PFE
should catalyze perhydrolysis of esters faster than wild-type
PFE. Second, L29P-PFE should catalyze perhydrolysis of an
ester faster than hydrolysis of the same ester. We tested
these predictions and found that both are incorrect.
To compare the abilities of L29P-PFE for perhydrolysis
and hydrolysis, we measured the rate of hydrolysis of methyl
acetate at pH 7.2 using a pH indicator, p-nitrophenol
[17]
(pNP).
Hydrolysis of methyl acetate forms acetic acid,
Kinetics were measured at the optimum pH for each reac-
tion, which ranged from 5.0 to 7.2. Reactions involving
esters (hydrolysis and perhydrolysis) used pH 6.5–7.2, which
which protonates the phenoxide and reduces its absorbance
at 404 nm. Wild-type PFE is threefold more efficient than
L29P-PFE for hydrolysis of methyl acetate ((k /K =600
cat
M
[15]
ꢀ1 ꢀ1
is the optimum pH
because the active site histidine is
vs. 200 s m ). The K values of methyl acetate are similar
M
fully deprotonated. Reactions involving acetic acid (perhy-
(43 and 50 mm), but the k is threefold higher for wild-type
PFE (25 vs. 7.7 s ).
cat
18
ꢀ1
drolysis, O water exchange) used pH 5.0, which is the pH
[14]
optimum because it balances deprotonation of the active
site histidine and protonation of acetate to acetic acid.
There is no evidence that the reaction mechanism changes
at these different pH values.
Comparison of these hydrolysis data to those of the per-
hydrolysis above shows that L29P-PFE catalyzes hydrolysis
tenfold better (k /K comparison) than perhydrolysis. The
cat
M
kcat for perhydrolysis versus hydrolysis are similar (8.7 and
[16]
ꢀ1
The monochlorodimedone (MCD) assay measures the
rate of perhydrolysis. Monochlorodimedone favors the enol
tautomer, which absorbs light at 290 nm. Peracid generated,
either chemically or enzymatically reacts with bromide ions
to form hypobromous acid, which brominates MCD. This
bromination prevents formation of the enol tautomer there-
by decreasing the absorbance at 290 nm. Fitting the initial
rate data to the Michaelis–Menten equation revealed the
apparent steady-state kinetic constants for wild-type and
L29P-PFE.
7.7 s ), but the K towards methyl acetate shows tighter
M
binding during hydrolysis than in perhydrolysis (50 vs.
420 mm). This result contradicts the prediction that L29P-
PFE should catalyze perhydrolysis of an ester faster than
hydrolysis. Wild-type PFE is twofold more efficient at hy-
drolysis than perhydrolysis, so the relative hydrolysis ability
is also higher for L29P-PFE.
These comparisons of hydrolysis and perhydrolysis of
methyl acetate suggest that L29P-PFE is not more selective
for hydrogen peroxide over water. To confirm this sugges-
tion, we measured the selectivity of both enzymes for hydro-
gen peroxide versus water.
In contrast to the first prediction, wild-type PFE (k /
cat
ꢀ
1
ꢀ1
K =300 s m ) is 15-fold more efficient than L29P-PFE
20 s m ) for perhydrolysis of methyl acetate (Table 1).
M
ꢀ
1
ꢀ1
(
The KM for methyl acetate contributes a factor of seven to
this difference (wild-type PFE: 60 mm; L29P: 420 mm),
Wild-type and L29P-PFE show similar selectivity for hydro-
gen peroxide over water: Selectivity for water and hydrogen
peroxide by wild-type and L29P-PFE were measured for
methyl acetate as the substrate. The release of protons by
a pHstat revealed the initial rates of hydrolysis and perhy-
drolysis (Figure 3). Hydrolysis of methyl acetate releases
whereas k contributes a factor of 1.7 (wild-type PFE:
cat
ꢀ
1
ꢀ1
1
5 s ; L29P-PFE: 8.7 s ). Thus, even though L29P-PFE is
a 43-fold better perhydrolase than PFE for acetic acid, it is
a 15-fold worse perhydrolase than PFE for methyl acetate.
8132
www.chemeurj.org
ꢀ 2012 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2012, 18, 8130 – 8139

Carboxylic Acid Perhydrolase Activity
FULL PAPER
perhydrolysis. The rate of perhydrolysis is the difference be-
tween the two rates, as shown in Equation (1):
Vperhydrolysis ¼ VHþðNaBrÞ ꢀ VH
þ
ð1Þ
A plot of the rate of perhydrolysis relative to hydrolysis (the
selectivity for perhydrolysis) as a function of hydrogen per-
oxide concentration shows a hyperbolic saturation curve
(
Figure 4). The relative amount of perhydrolysis increases
with increasing hydrogen peroxide concentration and then
reaches a plateau.
Figure 3. Measuring the relative amounts of perhydrolysis and hydrolysis
of methyl acetate by using a pHstat. a) Hydrolysis of methyl acetate at
pH 7.2 releases a proton, but the accompanying perhydrolysis of methyl
acetate does not release significant amounts of proton because peracetic
acid (pK =8.2) remains mainly protonated. b) In the presence of sodium
a
bromide, peracetic acid reacts with bromide to also generate a proton
after several steps. Figure 4a shows typical data.
one proton due to the ionization of acetic acid. Perhydroly-
sis of methyl acetate does not release significant amounts of
proton because peracetic acid remains mostly protonated at
1
pH 7.2. When the solution contains NaBr, additional reac-
tions of peracetic acid occur that result in the release of one
proton. Peracetic acid oxidizes bromide to bromonium,
which reacts with additional bromide to form bromine. (An
equivalent alternative is that bromonium may first react
with water to form hypobromous acid and next react with
bromide to form bromine.) The bromine oxidizes hydrogen
Figure 4. Relative initial rates of perhydrolysis to hydrolysis of methyl
acetate catalyzed by wild-type and L29P-PFE. a) Typical experimental
data by using the assay in Figure 3. The amount of base needed to main-
tain a constant pH of 7.2 during the PFE-catalyzed perhydrolysis of
methyl acetate (1.5m) in hydrogen peroxide (44 mm). The experiment
without sodium bromide (c) measures hydrolysis, whereas the experi-
ment with sodium bromide (c) measures both hydrolysis and perhy-
drolysis. b) The ratio of initial rates of perhydrolysis and hydrolysis in-
creased with increasing hydrogen peroxide concentration and then
reached a plateau. Selectivity parameters for the model in Figure 5b
[18]
peroxide to form oxygen and hydrogen bromide.
The
overall reaction releases one proton for each peracetic acid.
Two experiments validated this new assay for perhydroly-
sis. First, adding a known amount of peracetic acid to a solu-
tion of sodium bromide at neutral pH caused a decrease in
pH. One equivalent of sodium hydroxide returned the pH
to its original value. This experiment shows that one proton
is released for each molecule of peracetic acid. Second,
adding a known amount of bromine to a hydrogen peroxide
solution decolorized the bromine and decreased the pH.
Two equivalents of sodium hydroxide were needed to return
the pH to its original value. This experiment shows that the
last equation in Figure 3b is correct.
were determined by fitting data to the formula shown. The parameter b
0
is the perhydrolysis over hydrolysis ratio at low nucleophile concentra-
tions and 1/g is the perhydrolysis over hydrolysis ratio at saturating nu-
cleophile concentrations.
Measuring the rate of proton release without NaBr re-
veals the rates of hydrolysis, whereas measuring the rate of
proton release with NaBr reveals the sum of hydrolysis and
To interpret this behavior, we considered two kinetic
schemes for the hydrolysis and perhydrolysis reactions
(Figure 5). The simplest kinetic scheme involves formation
of the acetyl-enzyme followed by competition between
water and hydrogen peroxide for this acetyl-enzyme (Fig-
ure 5a). The product ratio depends on the relative reactivity
1
This approach ignores the small amount of ionization of peracetic acid
(
a
pK =8.2) at pH 7.2. Attempts to include this ionization led to impos-
sible values (negative rates) in some cases.
Chem. Eur. J. 2012, 18, 8130 – 8139
ꢀ 2012 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
8133

R. J. Kazlauskas and D. Yin
hydrolysis reaches a plateau once the enzyme is saturated
with hydrogen peroxide.
Using this more complex kinetic model, Youshko and co-
[21]
workers
derived three measures of selectivity, two of
which we discuss here. Their work focused on penicillin-acy-
lase-catalyzed aminolysis of esters, but the same kinetic
form applies to perhydrolysis. The parameter b measures
0
the selectivity for nucleophile over water at low nucleophile
concentration [Eq. (2)]. This parameter includes the KM for
the nucleophile (hydrogen peroxide) but not the KM for
water, which is impossible to measure in water, so the units
ꢀ
1
are m . Large values of b indicate higher selectivity for hy-
0
drogen peroxide over water, but because the KM of water is
not included, b is not the numeric value of the selectivity.
0
The parameter 1/g measures the selectivity for nucleophile
over water at saturating nucleophile concentrations [Eq. (3)]
and has no units. It gives the true numeric value for the se-
lectivity of the enzyme for hydrogen peroxide over water
under these conditions. Fitting the data in Figure 5 to Equa-
Figure 5. Kinetic schemes for enzyme-catalyzed perhydrolysis of methyl
acetate. a) Simple scheme involving direct competition of water and hy-
drogen peroxide for the acetyl-enzyme intermediate. This scheme pre-
dicts that ratio of peracid to acid formed depends linearly on the concen-
tration of hydrogen peroxide. b) More complex scheme in which acetyl-
enzyme–hydrogen peroxide complex may also undergo hydrolysis. At
low hydrogen peroxide concentrations, water and hydrogen peroxide
compete for the acetyl-enzyme intermediate and the relative amount of
perhydrolysis increases with increasing hydrogen peroxide concentra-
tions. At saturating hydrogen peroxide concentrations, higher hydrogen
peroxide concentrations will not increase the amount of perhydrolysis.
tion (4) reveals the values of b and 1/g.
0
ꢀ
ꢁ
k4
b0 ¼
ð2Þ
ð3Þ
ð4Þ
k3KH2O2
k4
k5
g ¼
ꢀV
ꢁ
ꢀ
ꢁ
3 5
The k and k steps involve water as a reactant, but the diagram omits
½H O ꢃ*b
perhydrolysis
Vhydrolysis
2
2
0
water because the kinetic equations omit water. Water concentration is
constant and the binding constants for water are implicitly included in
the other constants.
¼
1 þ b *g*½H O ꢃ
0
2
2
initial
The fit of Equation (4) to the experimental data is excellent
for both wild-type PFE and for L29P-PFE (Figure 4). The
fit indicates that L29P-PFE is not more selective for hydro-
gen peroxide than wild-type PFE. At low hydrogen peroxide
of the enzyme with each nucleophile and the relative con-
centrations of the two nucleophiles. The relative amount of
peracid should increase linearly with increasing concentra-
tions of hydrogen peroxide as water and hydrogen peroxide
compete for the acetyl-enzyme. This simple kinetic scheme
is analogous to that proposed by Bender and co-workers for
the competition of the acyl-enzyme in chymotrypsin for
concentration, wild-type PFE (b =330ꢂ11) is twice as se-
0
lective as L29P-PFE (b =170ꢂ20) for hydrogen peroxide
0
and at saturating hydrogen peroxide concentrations, wild-
type PFE (1/g=15.3ꢂ0.2) is 37% more selective than
L29P-PFE (1/g=11.2ꢂ0.1) for hydrogen peroxide. These
findings contradict the prediction from the second hypothe-
sis that L29P-PFE is more selective for hydrogen peroxide
than wild-type PFE.
[19]
water and an alcohol nucleophile. The data for both wild-
type PFE and L29P-PFE show a curve that flattens to a pla-
teau at high hydrogen peroxide concentration; thus this
simple model does not fit the experimental data. Previous
similar cases—competition between water and alcohol nu-
Acetyl-enzyme forms faster in L29P-PFE than in wild-type
PFE: The first hypothesis is that L29P-PFE catalyzes perhy-
drolysis faster because the acetyl-enzyme forms faster. To
test this hypothesis, we measured the initial rate of isotope
[20]
cleophile for the acyl-enzyme in papain and the competi-
tion between water and amine nucleophile in penicillin acy-
[
21]
18
lase —also show a similar plateau.
exchange of O-water with acetic acid at pH 5.0 using gas
Fink and Bender introduced a more complex model (Fig-
ure 5b) to account for poor fit of the data to the simple
model. The more complex kinetic model provides two hy-
drolysis paths: at low concentrations of hydrogen peroxide,
water competes directly with hydrogen peroxide for the
acetyl-enzyme, as above, whereas at high concentrations of
hydrogen peroxide, water can react with the acetyl-enzyme·-
chromatography combined with ion trap mass spectrometry.
The signals at 47 and 45 atomic mass units, which corre-
spond to the loss of a methyl radical in the ion trap, re-
vealed the relative abundance of O-acetate to O-acetate.
Isotope exchange involves formation of the acetyl-enzyme
intermediate, so faster isotope exchange indicates faster
acetyl-enzyme formation.
1
8
16
H O2 complex. This second path allows hydrolysis even
when the acetyl-enzyme is saturated with hydrogen perox-
ide. This model predicts that the ratio of perhydrolysis to
The initial rate of L29P-PFE catalyzed isotope exchange
of O-water with acetic acid was 25-fold faster than that for
wild-type PFE (Table 2). Wild-type PFE catalyzed the for-
2
1
8
8134
www.chemeurj.org
ꢀ 2012 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2012, 18, 8130 – 8139

Carboxylic Acid Perhydrolase Activity
FULL PAPER
1
8
Table 2. Rate of O-water exchange with acetic acid and rate of perhy-
[
a]
drolysis catalyzed by wild-type PFE and L29P-PFE.
Enzyme Specific activity for isotope
Specific activity for perhydrol-
ꢀ
1
ꢀ1
exchange [Umg
relative rate)
]
ysis of acetic acid [Umg
(relative rate)
]
(
wild-
type
3.4 (1)
0.23 (1)
PFE
L29P-
PFE
84 (25)
9.9 (43)
ꢀ
1
18
[
(
a] The reaction mixture contained enzyme (0.08–1.0 mgmL ), O-water
45 mol%), and acetate (1.0m, pH 5.0). Aliquots (1 mL) taken between 0
and 60 min of reaction were flash frozen with liquid nitrogen and later
thawed with phosphoric acid buffer (1.5 mL, 2.5 mm) and immediately
1
8
analyzed by GC/MS. The spontaneous rate of O-water exchange with
acetic acid was <3% of the slowest rate measured. The measured initial
1
8
rates were corrected to account for only 45 mol% O in the reaction
mixture.
1
8
ꢀ1
ꢀ1
mation of O-acetic acid at 3.4 mmolmin mg , whereas
ꢀ
1
ꢀ1
the L29P-PFE catalyzed rate was 84 mmolmin mg . A
control sample without enzyme showed no spontaneous iso-
tope exchange: <3% of the slowest rate. This lack of spon-
taneous exchange under these conditions is consistent with
[22]
previous measurements.
Molecular basis for faster acetyl-enzyme formation in L29P-
PFE: To identify the molecular basis of the faster acetyl-
enzyme formation, we built computer models of the tetrahe-
dral intermediate T 1 starting from the X-ray crystal struc-
d
[
23]
tures of wild-type PFE (PDB ID: 1VA4) and the L29P-
PFE-acetate complex (PDB ID: 3HI4). The geometry of
both models were optimized by using the OPLS 2005 force-
Figure 6. Computer models of T
d
1 in L29P-PFE and wild-type PFE sug-
[
14]
gest a molecular basis for the faster acetyl-enzyme formation in L29P-
PFE. a) The tetrahedral intermediate in L29P-PFE accepts two hydrogen
bonds from the catalytic histidine (H251), two from the NꢀH in the oxy-
[24]
field. The X-ray crystal structures of L29P-PFE and four
other perhydrolases for carboxylic acids all show that the
key difference between esterases and carboxylic acid perhy-
drolases is a proline in the oxyanion loop. The amide bond
between the amine of this proline and the preceding amino
acid carbonyl adopts a cis conformation, which moves the
carbonyl group (Trp28 carbonyl in the case of L29P-PFE)
closer to the active site.
anion hole (M95 and W28), and donates a hydrogen bond to the carbonyl
oxygen of W28. This hydrogen bond speeds formation of the acetyl-
enzyme by favoring protonation of the hydroxyl group by the catalytic
histidine. b) The tetrahedral intermediate in wild-type PFE accepts two
hydrogen bonds from the catalytic histidine (H251) and two from the
NꢀH in the oxyanion hole (M95 and W28). The carbonyl oxygen of W28
is too far from the hydroxyl group of the tetrahedral intermediate (OꢀO
distance is 4.4 ꢁ; not shown) to accept a hydrogen bond, but a water
molecule can bridge this gap. This model structure is less stable due to
the entropic cost of immobilizing a water molecule. The water molecule
may also slow formation of the acetyl-enzyme by donating a hydrogen
bond to the tetrahedral intermediate (not shown). The dotted lines indi-
cate hydrogen bonds.
This closer carbonyl group might speed the conversion of
the tetrahedral intermediate to the acetyl-enzyme intermedi-
ate by accepting a hydrogen bond from the hydroxyl group
of the tetrahedral intermediate, T 1 (Figure 6). The model
d
of L29P-PFE shows that the carbonyl oxygen of Trp28 ac-
cepts a hydrogen bond from the hydroxyl group of the tetra-
hedral intermediate (OꢀO distance is 2.9 ꢁ). This hydrogen
can bridge this gap by accepting a hydrogen bond from the
hydroxyl group of T 1 and donating a hydrogen bond to the
d
bond might speed up the formation of the acetyl-enzyme in-
termediate by promoting protonation of the leaving hydrox-
yl group by the active site histidine.
carbonyl oxygen of W28 (Figure 6b). This bridging structure
represents a less stable intermediate since it costs entropy to
immobilize the extra water molecule. Another problem with
the bridging structure is that the water molecule can, besides
accepting a hydrogen bond from the hydroxyl group of the
tetrahedral intermediate, also donate a hydrogen bond. This
donation would slow down formation of the acetyl-enzyme
intermediate. The higher energy structure with the water
molecule in wild-type PFE is unlikely to contribute to catal-
ysis.
Computer models of T 1 in wild-type PFE show no low
d
energy way to make a similar hydrogen bond. The carbonyl
oxygen of W28 is farther from the active site in wild-type
PFE, so the model of T 1 shows an OꢀO distance of 4.4 ꢁ
d
between the carbonyl oxygen of W28 and the hydroxyl
group of the tetrahedral intermediate (data not shown). This
distance is too far for a hydrogen bond. A water molecule
Chem. Eur. J. 2012, 18, 8130 – 8139
ꢀ 2012 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
8135

R. J. Kazlauskas and D. Yin
Molecular basis for high selectivity for hydrogen peroxide
over water in both wild-type and L29P-PFE: Previous mod-
eling suggested L29P-PFE is more selective for hydrogen
peroxide due to an extra hydrogen bond formed by the
chain carbonyl is longer (3.3 vs. 2.8 ꢁ), so a hydrogen bond
is weak or missing. However, the peroxy hydroxyl can form
an internal hydrogen bond to the oxyanion oxygen. Since
both L29P and wild-type PFE can form one hydrogen bond
with hydrogen peroxide, both favor hydrogen peroxide over
water by a similar amount.
second tetrahedral intermediate (T 2). Here we identify
d
a different hydrogen bond that can form with T 2 of wild-
d
type PFE. Since both L29P-PFE and wild-type PFE can
form an extra hydrogen bond with hydrogen peroxide, and
assuming both hydrogen bonds have similar strength, both
enzymes are similarly selective for hydrogen peroxide over
water.
Discussion
Perhydrolysis of acetic acid involves two steps: formation of
the acetyl-enzyme intermediate and cleavage of the acetyl-
enzyme with hydrogen peroxide. The rate determining step
is the first step, the formation of the acetyl-enzyme inter-
mediate. To increase the rate of perhydrolysis, one could
either increase the rate of formation of the acetyl-enzyme
intermediate (first step) or increase the selectivity of hydro-
gen peroxide over water in the second step. Our initial hy-
pothesis (hypothesis 2) was that carboxylic acid perhydrolas-
es, including L29P-PFE, are better perhydrolases than ester-
ases because they are more selective for hydrogen peroxide.
Two experiments indicate that this hypothesis is incorrect.
First, comparing the hydrolase and perhydrolase abilities
of L29P-PFE toward the same substrate shows it is a tenfold
better hydrolase. Previous comparisons concluded the re-
verse because they compared different substrates for hydrol-
ysis and perhydrolysis. Here we compare hydrolysis and per-
hydrolysis with the same substrate to reach the opposite
conclusion. We use methyl acetate as the acyl donor because
Models of T 2 containing hydrogen peroxide were built
d
starting with the crystal structures of wild-type PFE (PDB
[
23]
[14]
ID: 1VA4) or L29P-PFE (PDB ID: 3HI4) and their ge-
[24]
ometry was optimized by using the OPLS 2005 forcefield.
Similar to previous modeling,
[25,26]
the T 2 for L29P-PFE
d
can donate a hydrogen bond to the main chain carbonyl of
W28 (Figure 7a). This hydrogen is also near the oxyanion
oxygen, so the formation of an internal hydrogen bond is an
alternative. For wild-type PFE, the distance to the main
[14]
esters with a larger alcohol group are poor substrates,
whereas acetic acid as the substrate hides the hydrolase ac-
tivity. L29P-PFE catalyzed hydrolysis with a k /K of
cat
M
ꢀ
1
ꢀ1
ꢀ1 ꢀ1
2
00 s
m
versus perhydrolysis with a k /K of 20 s m .
cat M
In comparison, wild-type PFE showed a similar catalytic ef-
ꢀ
1
ꢀ1
ficiency for both hydrolysis (k /K =600 s m ) and perhy-
cat
M
ꢀ
1
ꢀ1
drolysis (k /K =300 s m ). Thus, the L29P mutation in
cat
M
PFE has decreased its ability to catalyze perhydrolysis of
methyl acetate 15-fold, but decreased its ability to catalyze
hydrolysis of methyl acetate only threefold. This observation
is not consistent with an increase in selectivity for hydrogen
peroxide in L29P-PFE.
Second, the selectivity of L29P-PFE for hydrogen perox-
ide over water is slightly lower than that for wild-type PFE.
A new pHstat assay measured hydrolysis and perhydrolysis
by using two experiments. With no additive, the amount of
acid released measured the hydrolysis contribution, whereas
with sodium bromide added, the amount of acid released
measured the sum of hydrolysis and perhydrolysis. Fitting
a kinetic model developed previously for acyl transfer to al-
cohols and amines revealed the selectivity for hydrogen per-
oxide versus water. This model defines two selectivities: one
Figure 7. Computer models of T
d
2 in L29P-PFE and wild-type PFE sug-
a low hydrogen peroxide concentration (b ), and another,
0
gest a molecular basis for the similar selectivity for hydrogen peroxide
over water. Both: a) L29P-PFE, and b) wild-type PFE show one hydro-
gen bond involving the peroxy hydroxyl. For L29P-PFE, the peroxy hy-
droxyl donates a hydrogen bond to main chain carbonyl of W28. For
wild-type PFE, the peroxy hydroxyl donates an internal hydrogen bond
to the oxyanion oxygen. The dotted lines indicate hydrogen bonds.
a high hydrogen peroxide concentration (1/g). The selectivi-
ty of L29P-PFE at low hydrogen peroxide concentrations,
b , was approximately twofold lower than that for wild type
0
ꢀ1
(
170 versus 330m ), whereas the selectivity at high hydro-
gen peroxide concentrations, 1/g, was 37% lower than that
8136
www.chemeurj.org
ꢀ 2012 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2012, 18, 8130 – 8139

Carboxylic Acid Perhydrolase Activity
FULL PAPER
18
for wild type (11.2 versus 15.3). Thus, the selectivity of
L29P-PFE is not 43-fold higher than for wild-type PFE, but
instead slightly lower.
of perhydrolysis and isotope exchange between H2 O and
acetic acid. Wild-type PFE can form a similar hydrogen
bond, but only by using a water molecule to bridge the gap
The other possibility for why L29P-PFE is a better perhy-
drolase is that the formation of the acyl-enzyme is faster for
L29P-PFE than for wild type. Consistent with this hypothe-
sis, L29P catalyzed the isotope exchange between acetic acid
and H2 O 25-times faster than the wild type. L29P also cat-
alyzed perhydrolysis of acetic acid 43-times faster (changes
between T 1-OH and the main chain carbonyl oxygen of
W28 (Figure 8b). The need to immobilize a water molecule
costs entropy and is less favorable.
Both L29P-PFE and wild-type PFE show similar selectivi-
ty for hydrogen peroxide because both acetyl-enzymes can
hydrogen bond to hydrogen peroxide (Figure 9). Previous
d
18
in k ) than wild-type PFE. Thus, the mutations that in-
cat
crease the rate of perhydrolysis of acetic acid also increase
the rate of formation of acyl-enzyme intermediate.
The X-ray crystal structure of L29P soaked with acetic
acid at pH 5.0 suggests a structural basis for the faster for-
[14]
mation of the acetyl-enzyme (Figure 2 from Yin et al.).
The structure shows acetic acid in a catalytically productive
orientation to form T 1 on the way to the acetyl-enzyme in-
d
termediate. The carbonyl oxygen accepts hydrogen bonds
from both NꢀH of the oxyanion hole and the carbonyl
carbon lies 2.8 ꢁ from Ser94-Og. The structure also shows
a water molecule hydrogen bonded between the acetic acid
Figure 9. Molecular mechanisms that promote selectivity for hydrogen
peroxide. Hydrogen peroxide can form a hydrogen bond to the acetyl-
enzyme in both wild-type PFE and L29P-PFE, but this extra hydrogen
bond differs. In wild-type PFE, the hydrogen bond acceptor is the acetyl-
enzyme carbonyl oxygen similar the intramolecular hydrogen bond in
peracetic acid. In L29P-PFE, the hydrogen bond acceptor is a main-chain
carbonyl oxygen. For clarity, the catalytic histidine, His251, is not shown.
This histidine accepts a hydrogen bond from the remaining hydroxyl of
hydrogen peroxide during the reaction.
and Trp28 carbonyl oxygen. A computer model of T 1 in
d
L29P built starting from this structure shows a direct dona-
tion of a hydrogen bond from the T 1-OH to the main chain
d
carbonyl oxygen of W28 (Figure 8a). This hydrogen bond in
L29P-PFE can increase the rate of formation of the acetyl-
enzyme intermediate. This increased rate increases the rate
X-ray structures and modeling of L29P-PFE identified a hy-
drogen bond between hydrogen peroxide and the carbonyl
oxygen of W28. Modeling indicates that wild-type PFE also
makes an extra hydrogen bond, but at a different location.
The hydrogen bond acceptor is the carbonyl oxygen of the
acetyl-enzyme intermediate, similar to the internal hydrogen
[27]
bond in peracetic acid.
Other hydrolases, such as chloroperoxidases (CPO-T,
[
28]
[29]
CPO-L, and BPO-A1),
coumarin hydrolase,
(ꢀ)-g-lactamase
and dihydro-
[30]
likely use the same mechanism as
L29P-PFE for perhydrolysis of acetic acid since all of these
catalysts contain a proline at the oxyanion loop.
Another group of perhydrolases are those that catalyze
perhydrolysis of esters. These perhydrolases are also serine
hydrolases, but do not have a conserved proline in the oxy-
anion loop. The rate determining steps and kinetics for
these perhydrolases differ from the ones in this study.
These kinetics and modeling based on X-ray crystal struc-
tures identify the details within a serine hydrolaseꢂs active
site that control the relative preference for hydrolysis versus
perhydrolysis. Surprisingly, the interaction of main chain
atoms with the substrate is a key part of this reaction selec-
tivity. This work also shows the subtle changes that allow en-
zymes to acquire new catalytic activities, including unnatural
ones like perhydrolysis. Enzyme-catalyzed formation of per-
acids is a useful reaction for synthesis and these details ex-
plain why these perhydrolases are more effective than hy-
drolases.
Figure 8. Molecular mechanisms to promote acetyl-enzyme formation.
a) Proton transfer from the catalytic histidine to serine Og of the tetrahe-
dral intermediate reforms the starting acid, whereas proton transfer to
the OH group leads to the acetyl-enzyme intermediate. In carboxylic
acid perhydrolases with a proline at the oxyanion loop, the main chain
carbonyl oxygen accepts a hydrogen bond from the T 1 OH group. This
d
hydrogen bond favors proton transfer to the OH and increases the frac-
tion of this tetrahedral intermediate that forms acetyl-enzyme. b) In es-
d
terases, the carbonyl group of Trp28 is further from T 1, so a direct hy-
drogen bond is not possible. A water molecule bridge shown can make
a similar interaction, but immobilizing a water molecule costs free energy
as entropy.
Chem. Eur. J. 2012, 18, 8130 – 8139
ꢀ 2012 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
8137

R. J. Kazlauskas and D. Yin
1
8
Isotope exchange of acetic acid with O-water: Initial rates for enzyme-
Experimental Section
catalyzed isotope exchange were measured by using a gas chromatograph
(
Varian Star 3400 CX with DB-FFAP column from Agilent Technologies,
General: Water with 18 W purity was prepared by using a Milli-Q system
Santa Clara, CA) with a quadrupole ion trap mass spectrometer (Varian
(
Millipore, Billerica, MA). Chemicals were from Sigma–Aldrich; E. coli
Saturn 3) to detect the relative abundance of acetic acid (m/z 45) and
DH5a-T1 competent cells were from Invitrogen (Carlsbad, CA). All pri-
mers were from Integrated DNA Technologies (Coralville, IA). Enzyme
activity was measured in 96-well plates by using a diode-array microplate
18
18
O-acetic acid (m/z 47). In a typical reaction, 45 mol% of O-water and
acetic acid (1m, pH 5.0) were mixed with enzyme (0.16 mg for L29P,
.0 mg for wild-type PFE) to a final volume of 0.10 mL. Aliquots
1.0 mL) were withdrawn at appropriate time intervals and flash frozen in
1
(
reader. Kinetic constants (Vmax and K
the specific activity as a function of substrate concentration. Data was fit
to the Michaelis–Menten equation: rate=Vmax ꢃ[S]/(K +[S]) by using
M
) were determined by measuring
liquid nitrogen. For analysis, the aliquots were diluted with phosphoric
acid (2.5 mm, 1.5 mL), and an aliquot (1.0 mL) was injected into the gas
chromatograph. The no-enzyme control showed 9.9CAHTUNGTRENNUNG
(
<0.05 mmolmin ) indicating no detectable spontaneous exchange under
these conditions. Previous researchers measured a spontaneous exchange
rate of 0.60 mmolmin at 1018C, which is consistent with our measure-
A
H
U
G
R
N
N
M
2
Origin v.8.0 software (Origin Lab, Northampton, MA). The R values
were >0.97 for both mutants and wild-type PFE.
18
[
14]
ꢀ1
Protein expression and purification: The plasmid, pL29P contained the
[
31]
L29P-PFE gene with an inducible rhamnose promoter. Typically, lyso-
ꢀ1
ꢀ1
[22]
A
C
H
T
U
N
G
T
R
E
N
N
U
N
G
geny broth (LB) medium (5 mL containing 0.1 mgmL ampicillin) was
ꢀ
1
inoculated from a single colony, then grown, overnight, at 378C. The
ment of <0.05 mmolmin at 238C. The L29P sample showed 50 mmol
1
8
overnight culture was diluted (1:100) with fresh LB media (100 mL con-
O at 3 min and this value increased linearly (measured at 3, 5, 7, 10,
ꢀ
1
ꢀ1
taining 0.1 mgmL ampicillin) and grown at 378C until an absorbance at
00 nm of 0.6 was reached. Filter-sterilized rhamnose (20% w/v) was
and 13 min) at 4.5 mmolmin to reach 97 mmol (3.1 mol%) at 13 min.
This rate corresponds to a specific activity of 38 mmolmin mg protein.
The wild-type PFE sample showed 47 mmol O at 21 min and this value
increased linearly (measured at 21, 28, 35, 42, 49, and 60 min) at 1.5 mmol
min to reach 60 mmol (3.3 mol%) at 60 min. This rate corresponds to
a specific activity of 1.5 mmolmin mg protein. These specific activities
were divided by 0.45 to account for only 45 mol% O in the solution to
yield the values listed in Table 2.
ꢀ
1
ꢀ1
6
1
8
added at a concentration of 2% (w/v) to induce protein expression and
the culture was incubated for an additional 3 h at 378C. The induced cul-
ture was centrifuged (4000g, 15 min) and the cell paste was resuspended
ꢀ
1
ꢀ1
ꢀ1
in buffer A (50 mm NaH
2
PO
4
, 300 mm NaCl, 20 mm imidazole) to a con-
1
8
centration of 20% (w/v). The resuspended culture was flash frozen in
liquid nitrogen, thawed to room temperature, lysozyme was added (final
ꢀ
1
concentration of 1 mgmL ) and incubated on ice for 30 min. The cell
lysate was centrifuged (10000g, 60 min) and the supernatant was poured
onto a column of Ni-NTA agarose resin (5 mL, Invitrogen) pre-equili-
brated with buffer A (25 mL). The column was washed with buffer B
Computer modeling: Enzyme complexes with the first or second tetrahe-
dral intermediate were modeled by using molecular mechanics with the
programs Maestro and Macromodel v9.9 (Schrçdinger, New York) and
the OPLS 2005 forcefield. Starting from the X-ray crystal structures of
the acetate complexes (L29P-PFE/acetate PDB ID: 3HI4) or wild-type
PFE (PDB ID: 1VA4), Protein Preparation Wizard software (Schrç-
dinger, New York) removed water molecules further than 5 ꢁ from the
nearest heteroatom, removed glycerol and sulfate, removed five out of
the six protein chains in the crystal structure, added hydrogen atoms to
all heteroatoms and optimized the geometry of the hydrogen atoms until
an rmsd of <0.05 ꢁ was reached. The entire protein structure was geom-
etry optimized by using the same forcefield until an rmsd of <0.3 ꢁ was
reached.
[
24]
[
14]
(
50 mL, 50 mm NaH
type or L29P-PFE was eluted with buffer C (10 mL, 50 mm NaH
00 mm NaCl, 250 mm imidazole). Protein concentrations were measured
by absorbance at 280 nm by using the calculated extinction coefficient of
2
PO
4
, 300 mm NaCl, 40 mm imidazole) and the wild-
[
23]
2
PO
4
,
3
ꢀ
1
ꢀ1 [32]
PFE (35410m cm ); typical yield was 10–15 mg protein.
Steady-state kinetic constants for perhydrolysis of acetic acid: Kinetic
constants for perhydrolysis were determined by using the monochlorodi-
[
16]
medone (MCD) assay, in which the amount of enzyme added was ad-
justed to give a linear dependence of the reaction rate to enzyme concen-
tration. All reactions contained MCD (0.0472 mm) and sodium bromide
The tetrahedral intermediates were built by using Maestro by attaching
acetic or peracetic acid to S94-Og in the active site. The geometry of the
intermediate along with any additional water (for the model in Fig-
ure 6b) was optimized until an rmsd of <0.05 ꢁ was reached. The entire
protein and water molecules were further geometry optimized until
a rmsd of 0.05 ꢁ was reached.
(
149 mm) and citrate buffer (100 mm, pH 6.5). The concentrations of hy-
drogen peroxide and acetic acid were varied to give evenly spaced data
points above and below the apparent K . When varying the concentra-
M
tion of hydrogen peroxide, the concentration of acetic acid was 2.00m;
when varying the concentration of acetic acid, the concentration of hy-
drogen peroxide was 10 mm.
Steady-state kinetic constants for perhydrolysis of methyl and ethyl ace-
tate: Perhydrolysis was measured as above by using the MCD assay. The
concentration of either ethyl acetate or methyl acetate were varied to
Acknowledgements
M
give a minimum of three data points above and below the apparent K ,
whereas hydrogen peroxide concentration was kept constant at 14.7 mm.
The kinetic constants were found by fitting the initial rates to the Mi-
chaelis–Menten equation by using nonlinear regression. The single sub-
strate Michaelis–Menten model is valid because the second substrate, hy-
drogen peroxide is completely saturated.
We thank the US National Science Foundation (CBET-0932762) and the
Korea Science and Engineering Foundation funded by the Ministry of
Education, Science and Technology (WCU program R32-2008-000-10213-
0
) for financial support. We thank the Minnesota Supercomputing Insti-
tute for access to computers and software for computer modeling. We
thank Tom Krick of the Mass Spectrometry Consortium of the Depart-
ment of Biochemistry, Molecular Biology and Biophysics for help with
the isotope exchange experiments.
Nucleophile competition between hydrogen peroxide and water: Initial
rates for perhydrolysis and hydrolysis were determined by using pHstat
at 238C with NaOH (0.01N) as the titrant. The amount of hydrogen per-
oxide was varied from 14.7 to 147 mm whereas the concentration of
methyl acetate was held constant at 1.5m. The reaction solution was ad-
justed with NaOH (0.01N) before the addition of enzyme to pH 7.2. The
[1] C. R. Harrison, P. Hodge, J. Chem. Soc. Perkin Trans. 1 1976, 2252–
2254.
[2] G. J. ten Brink, I. W. C. E. Arends, R. A. Sheldon, Chem. Rev. 2004,
104, 4105–4124.
[3] A. R. Doumaux, J. E. McKeon, D. J. Trecker, J. Am. Chem. Soc.
1969, 91, 3992–3993.
[4] H. Sugimoto, D. T. Sawyer, J. Am. Chem. Soc. 1985, 107, 5712–5716.
[5] M. S. Harakeh, FEMS Microbiol. Lett. 1984, 23, 27–30.
ꢀ
1
amount of protein added was 1.2 mgmL for wild-type PFE and L29P-
PFE. For perhydrolysis, the NaBr (166 mm) was added to react with per-
acetic acid. For hydrolysis in the presence of hydrogen peroxide, sodium
chloride (166 mm) replaced the NaBr to keep the ionic strength constant.
Subtraction of perhydrolysis+hydrolysis rate from rate of hydrolysis
gave the rate for perhydrolysis. The initial rate data were fit to Equa-
0
tion (4) by nonlinear regression to find the values of b and g.
8138
www.chemeurj.org
ꢀ 2012 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2012, 18, 8130 – 8139

Carboxylic Acid Perhydrolase Activity
FULL PAPER
[
[
6] J. Koivunen, H. Heinonen-Tanski, Water Res. 2005, 39, 4445–4453.
7] B. Saake, R. Lehnen, E. Schmekal, A. Neubauer, H. H. Nimz, Holz-
forschung 1998, 52, 643–650.
[21] M. I. Youshko, G. G. Chilov, T. A. Shcherbakov, V. K. Svedas, Bio-
chim. Biophys. Acta 2002, 1599, 134–140.
[22] D. R. Llewellyn, C. OꢂConnor (nꢅe Bishop), J. Chem. Soc. 1964, 110,
545–549.
[23] J. D. Cheeseman, A. Tocilj, K. Morley, S. Park, J. D. Schrag, R. J.
Kazlauskas, Acta Crystallogr. Sect. D 2004, 60, 1237–1243.
[24] W. L. Jorgensen, D. S. Maxwell, J. Tirado-Rives, J. Am. Chem. Soc.
1996, 118, 11225–11236.
[
8] S. Duncan, Q. Jing, W. W. AlDajani, A. Katona, J. Schilling, U.
Tschirner, R. J. Kazlauskas, Appl. Biochem. Biotechnol. 2010, 160,
1
637–1652.
[
9] D. L. Yin, Q. Jing, W. W. AlDajani, S. Duncan, U. Tschirner, J. Schil-
ling, R. J. Kazlauskas, Bioresour. Technol. 2011, 102, 5183–5192.
[
10] B. Phillips, P. S. Starcher, B. D. Ash, J. Org. Chem. 1958, 23, 1823–
826.
11] M. Picard, J. Gross, E. Lubbert, S. Tolzer, S. Krauss, K. H. van-Pee,
A. Berkessel, Angew. Chem. 1997, 109, 1245–1248; Angew. Chem.
Int. Ed. Engl. 1997, 36, 1196–1199.
[25] P. Bernhardt, K. Hult, R. J. Kazlauskas, Angew. Chem. 2005, 117,
2802–2806; Angew. Chem. Int. Ed. 2005, 44, 2742–2746.
[26] W. Lee, L. Vojcic, D. Despotovic, R. Prodanovic, K.-H. Maurer, U.
Schwaneberg, M. Zacharias, Theor. Chem. Acc. 2010, 125, 375–386.
[27] L. M. Hjelmeland, G. H. Loew, Chem. Phys. Lett. 1975, 32, 309–314.
[28] B. Hofmann, S. Tolzer, I. Pelletier, J. Altenbuchner, K. H. van Pee,
J. J. Hecht, J. Mol. Biol. 1998, 279, 889–900.
[29] K. Line, M. N. Isupov, J. A. Littlechild, J. Mol. Biol. 2004, 338, 519–
532.
[30] M. Kataoka, K. Honda, S. Shimizu, Eur. J. Biochem. 2000, 267, 3–
10.
1
[
[
[
[
12] W. M. Draper, D. G. Crosby, Arch. Environ. Contam. Toxicol. 1983,
1
2, 121–126.
13] O. Khersonsky, D. S. Tawfik, Annu. Rev. Biochem. 2010, 79, 471–
05.
5
14] D. L. Yin, P. Bernhardt, K. L. Morley, Y. Jiang, J. D. Cheeseman, V.
Purpero, J. D. Schrag, R. J. Kazlauskas, Biochemistry 2010, 49, 1931–
1
942.
[31] K. D. Choi, G. H. Jeohn, J. S. Rhee, O. J. Yoo, Agric. Biol. Chem.
1990, 54, 2039–2045.
[32] E. Gasteiger, C. Hoogland, A. Gattiker, S. Duvaud, M. R. Wilkins,
R. D. Appel, A. Bairoch in The Proteomics Protocols Handbook
(Ed.: J. M. Walker), Humana Press, Totowa, 2005, pp. 571–607.
[
15] N. Krebsfꢄnger, F. Zocher, J. Altenbuchner, U. T. Bornscheuer,
Enzyme Microb. Technol. 1998, 22, 641–646.
16] P. D. Shaw, L. P. Hager, J. Biol. Chem. 1959, 234, 2565–2569.
17] L. E. Janes, A. C. Lowendahl, R. J. Kazlauskas, Chem. Eur. J. 1998,
[
[
4
, 2324–2331.
18] W. C. Bray, R. S. Livingston, J. Am. Chem. Soc. 1923, 45, 1251–
271.
[
1
Received: January 6, 2012
Revised: March 9, 2012
Published online: May 22, 2012
[
[
19] M. L. Bender, B. Zerner, J. Am. Chem. Soc. 1964, 86, 3669–3674.
20] A. L. Fink, M. L. Bender, Biochemistry 1969, 8, 5109–5118.
Chem. Eur. J. 2012, 18, 8130 – 8139
ꢀ 2012 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
8139