Insights into the fatty acid chain length specificity of the carboxylesterase PA3859 from Pseudomonas aeruginosa: A combined structural, biochemical and computational study

Article abstract of DOI:10.1016/j.biochi.2010.09.001

The open reading frame PA3859 of Pseudomonas aeruginosa encodes an intracellular carboxylesterase belonging to a group of microbial enzymes (EC 3.1.1.1) that catalyze the hydrolysis of aliphatic and aromatic esters with a broad substrate specificity. With

Full text of DOI:10.1016/j.biochi.2010.09.001

Biochimie 92 (2010) 1787e1792
Contents lists available at ScienceDirect
Biochimie
journal homepage: www.elsevier.com/locate/biochi
Research paper
Insights into the fatty acid chain length specificity of the carboxylesterase
PA3859 from Pseudomonas aeruginosa: A combined structural, biochemical
and computational study
*
Alessandro Pesaresi , Doriano Lamba
Istituto di Cristallografia, Consiglio Nazionale delle Ricerche, Area Science Park e Basovizza, S.S. 14, Km 163.5, I-34149 Trieste, Italy
a r t i c l e i n f o
a b s t r a c t
Article history:
Received 3 May 2010
Accepted 1 September 2010
Available online 17 September 2010
The open reading frame PA3859 of Pseudomonas aeruginosa encodes an intracellular carboxylesterase
belonging to a group of microbial enzymes (EC 3.1.1.1) that catalyze the hydrolysis of aliphatic and
aromatic esters with a broad substrate specificity. With few exceptions, for this class of enzymes,
belonging to the
a/b-hydrolase fold superfamily, very little information is available regarding their
biochemical activity and in vivo function. The X-ray crystal structure of recombinant PA3859 has been
determined for two crystal forms (space groups P21 and P21212). The kinetic properties of the enzyme
were studied using p-nitrophenyl esters as substrates and data fitted to a surface dilution mixed micelle
kinetic model. Enzymatic assays and computational docking simulations, pinpointed the enzyme’s
preference for esters of palmitic and/or stearic acids and provided insights into the enzymeesubstrate
favorable binding modes.
Keywords:
X-ray crystal structure
Automated docking
Enzymatic kinetic
Mixed micelle catalysis
Ó 2010 Elsevier Masson SAS. All rights reserved.
1. Introduction
p-nitrophenyl esters and a-naphthyl esters of carboxylic acids up to
oleic acid while no activity was detected against triacylglycerols [7].
Among those proteins showing a significant sequence similarity
to PA3859, the in vivo function has been elucidated only for the
human acyl-protein thioesterase I (35% sequence identity) which is
Carboxylesterases (EC 3.1.1.1) are members of the serine esterase
superfamily that catalyze the hydrolysis of carboxylic acid ester
bonds [1]. They share a common a/b-hydrolase fold and a catalytic
triad composed of Ser, His and Asp or Glu [2e4] with the Ser
residue acting as the nucleophile. These enzymes occur in most
living organisms and have been the object of many studies in view
of their potential biotechnological applications. An unusual feature,
often displayed by microbial carboxylesterases, is their relatively
broad substrate specificity [5]. Indeed, many microbial carbox-
ylesterases show a wide substrate tolerance suggesting that they
have evolved to enable access, as carbon sources, to a large variety
of organic compounds or to be involved in catabolic pathways [6].
In the absence of an unambiguous substrate specificity however,
the assignment of a clear physiological function for these enzymes
has often proven to be a difficult task.
involved in the depalmitoylation of Ga and Ras proteins and whose
3D structure has been determined (PDB ID 1FJ2) [9e11].
Although the biochemical properties and the physiological
implications of PA3859 sequence-related bacterial enzymes remain
still poorly investigated, for these proteins a wealth of structural
information is available. Indeed, beside the carboxylesterase II from
Pseudomonas fluorescens (PDB ID 1AUO) [12], that shows a 62%
sequence identity to PA3859, in the Protein Data Bank [9] there are
to date three additional entries that display high sequence and
structural similarities to PA3859 [13,14], (PDB ID 2R8B, 3B5E and
2H1I). With the exception of the carboxylesterase II from P. fluo-
rescens, for which a biochemical characterization was reported
[15,16], the classification of the above mentioned proteins as car-
boxylesterases is not supported by experimental data and their
enzymatic properties still remain to be elucidated.
We present the structure of two crystal forms of PA3859 from
P. aeruginosa. With the ultimate aim of gaining insights into the
PA3859 in vivo function, the substrate specificity of the enzyme has
been explored by kinetic measurements and automated docking
simulations, providing a rationale with respect to the likely physi-
ological substrates.
The carboxylesterase PA3859 from Pseudomonas aeruginosa is
a monomeric enzyme of 24 kDa that we have previously cloned,
expressed in Escherichia coli and crystallized [7,8]. A preliminary
biochemical characterization proved this enzyme to hydrolyze
* Corresponding author. Tel.: þ39 040 226881; fax: þ39 040 9221126.
E-mail address: alessandro.pesaresi@ts.ic.cnr.it (A. Pesaresi).
0300-9084/$ e see front matter Ó 2010 Elsevier Masson SAS. All rights reserved.
doi:10.1016/j.biochi.2010.09.001

1788
A. Pesaresi, D. Lamba / Biochimie 92 (2010) 1787e1792
2. Materials and methods
X-100 (Tx-100), a nonionic surfactant. The reactions were started
by adding 10
l of a 0.2 mg mlꢀ1 PA3859 protein solution to the
m
2.1. Structure solution and refinement
reaction mixture.
Substrates were purchased from SigmaeAldrich, Tx-100 from
Crystallization, preliminary X-ray analysis and phasing by Pat-
terson search methods of PA3859 crystal forms I and II have been
previously reported [8]. The crystal structure of carboxylesterase
from P. fluorescens, PDB ID 1AUO [12] was used as a search model,
after removing all solvent molecules. Crystal forms I and II contain
four and two molecules per asymmetric unit respectively.
The structures were refined using the CNS version 1.2 program
package [17,18]. The Fourier maps were visualized with the
graphical program O [19]. Two covalently bound N-morpholino
ethane sulphonic acid (MES) molecules (crystal form I) and two
monomethyl ether polyethylene glycol (PEGMME 5000) fragments
(crystal form II) were built-in by inspecting electron density maps.
The stereochemical quality of the final models was assessed with
PROCHECK [20]. Figures were drawn with PyMOL [21]. The
refinement statistics are summarized in Table 1. The crystal struc-
ture coordinates of form I and II have been deposited in the Protein
Data Bank with accession codes 3CN7 and 3CN9 respectively.
Fluka.
2.3. Automated docking simulations
Docking simulations were performed using the automated
docking program suite AutoDock [22,23]. The coordinates of the
substrates were generated in PDB format using the Dundee
PRODRG2 Server (http://davapcl.bioch.dundee.ac.uk/programs/
prodrg/). Atomic partial charges for the substrates (Gasteiger
charges) [24] and the protein (United Kollman charges) [25] were
assigned using the Automated Docking Tool (ADT). Electrostatic
and van der Waals’ interaction energy grids were calculated by the
AutoGrid module of Autodock. In order to encompass the whole
cleft the grid was centered on the Cd carbon atom of His197,
2
positioned at the active center cleft, and the grid dimensions were
65, 60 and 100 points for the X, Y and Z axes respectively (390,000
overall points). Grid point spacing was set to 0.375 Å. The search
space was explored using the Lamarckian Genetic Algorithm (LGA).
The docking of each substrate is the result 50 LGA runs with 2 ꢁ 106
energy evaluations per run.
2.2. Enzymatic hydrolysis of p-nitrophenyl esters
The reactions were carried out at room temperature in 50 mM
TriseHCl, pH 9.0. The formation of p-nitrophenol, the reaction
product, was monitored by measuring the increase in the absor-
bance at 400 nm using an ULTRASPEC 3000 spectrophotometer
(Pharmacia Biotech).
3. Results
3.1. PA3859 3D structure
PA3859 crystal forms I and II contain two and one homodimer
per asymmetric unit respectively. However, gel filtration experi-
ments indicate that the PA3859 protein exists as a monomer in
solution and homodimers are not detected [7]. For the sake of
brevity molecule A from crystal form II is described in the following
discussion, unless otherwise stated.
Water-insoluble substrates were dissolved in chloroform and
added to a 50 mM TriseHCl pH 9.0 solution containing 5 mM Triton
Table 1
Crystallographic refinement statistics of PA3859 crystal forms I and II.
The PA3859 molecular structure shows a canonical
lase fold [2], which holds nine -strands ( 1eb9) and six
1ea6). The -strands form a predominantly parallel
aligned in the order b1 b3 b2 b6 b7 b8 b9. The -sheet composes
a hydrophobic core flanked by five -helices: a1 a6 on one side and
a3 a4 a5 on the other. With respect to the Prototypic Hydrolase
Fold [2] PA3859 lacks the first N-terminal strand but contains an
insertion after strand b3. This insertion is made up of a motif
composed of strand b4, the 310 helix a2 and strand b5 followed by
a long loop connecting the latter to the -helix a3 (Fig. 1A).
Among the seven strands in the central -sheet, b6 b8 and b9
a
/
b
-hydro-
b
b
a-helices
Crystal form I
Crystal form II
(a
b
b-sheet
Resolution Range (Å)
Number of reflections (F0 ꢂ 0)
aFinal R-factor
23.82e2.99
18803
0.204
29.05e2.09
36106
0.196
,
,
,
,
,
,
b
a
,
bFinal Rfree-factor
0.274
0.240
,
,
Number of atoms
Non-hydrogen protein
Water molecules
Non-hydrogen MES
Non-hydrogen PEGMME 5000
R.m.s. Deviations
6584
94
22
3293
299
e
beaeb
e
64
a
b
,
bond length (Å)
bond angles (ꢃ)
0.011
1.537
0.010
1.487
provide the framework onto which the catalytic residues Ser113,
His197 and Asp166, are placed.
The PA3859 active site is located within a long L-shaped cleft
that engraves the protein surface (Fig. 1B). The long branch of the
cRamachandran statistics (%)
Residues in most favored regions
Residues in additional allowed regions
Residues in generously allowed regions
Residues in disallowed regions
Average Temperature Factors (Å2)
Protein
86.3
12.7
0.5
88.8
10.7
0.5
e
cleft is formed by helices a1, a6 and the loop that precedes the helix
0.4
a6. This portion of the cleft is highly hydrophobic and is lined by
residues Phe29, Val32, Leu36, Val39, Phe112, Val199 and Ile204.
The short branch of the L-shaped cleft is a rather polar pocket
32.1
20.3
57.1
e
31.7
39.4
e
Water
MES
formed by the three loops after strands b2, b7, b8, the N-terminal
PEG
53.8
3.33
3CN9
part of helix a4 and the very winding loop after b5. These loops
often confer substrate specificity to the hydrolase or are responsible
for conformational changes during adsorption of the interfacial
enzymes to the lipid phase [26]. The catalytic Ser113 is located at
the bottom of the active-site cleft with its side chain being exposed
to the solvent.
A comparison of the uninhibited PA3859 active-site features
(crystal form II) with that of the MES inhibited structure (crystal
form I), allowed us to infer the location of the oxyanion hole
which appears to be formed by the amide main chain nitrogen
dR.m.s.d.
D
B (Å2)
7.26
3CN7
PDB ID code
P
P
a
R ¼ hðjFoj ꢀ jFcjÞ= h jFoj, where jFoj and jFcjare the observed and calculated
structure factor amplitudes for reflection h. The summation is extended overall
unique reflections to the specified resolution.
b
Rfree, R-factor calculated using 1849 (I), 3602 (II) randomly chosen reflections
[9.8% (I) 10.0% (II)] set aside from all stages of the refinement.
c
Residues A-Ser113 (I), C-Ser113 (I), D-Ser113 (I), and A-Ser113 (II), B-Ser113 (II),
are in the generously allowed region; B-Ser113 (I) is in the disallowed region.
d
Root mean square deviation of
D
B (Å2) e isotropic temperature factor e for all
bonded atoms.

A. Pesaresi, D. Lamba / Biochimie 92 (2010) 1787e1792
1789
Fig. 1. The three-dimensional structure and molecular surface representation of the carboxylesterase PA3859. (A) Ribbon diagram presenting the secondary structure elements of
the PA3859 protein monomer. Seven -strands (blue arrows) are numbered sequentially from 1 to 3 and from 6 to 9. Five helkices (in green) are labeled 1 and from 3 to 6. In
comparison to the structure of a canonical -hydrolase, there is an additional motif (in pink) between the b3 strand and the a3 helix. The catalytic residues (Ser113, Asp166,
b
a
a
/b
b-a-b
His197) are shown in red stick representation. (B) Molecular surface representation of PA3859. The catalytic Ser113, the residues forming the oxyanion hole and the regions defining
the hydrophobic cleft and the polar cavity are shown in blue, light blue, yellow and pink respectively. The carbon atoms of the bound PEGMME 5000 nonaethylene glycol and
diethylene glycol fragments (crystal form II) are drawn in cyan and black respectively; oxygen atoms in red. The covalently bound moiety of MES (crystal form I) is also shown
(carbon, nitrogen, oxygen and sulphur atoms in green, blue, red and orange respectively).
atoms of Leu22 and Gln114. In the MES inhibited PA3859 struc-
ture the sulphur atom is covalently bound to the O of the cata-
In this milieu, at first the enzyme (E) is attracted to the deter-
gent-substrate mixed micelle surface (M) and forms the EM
complex. Then the interaction with the substrate (S) takes place
and a productive structural arrangement of an active EMS complex
is accomplished that ends up with the hydrolysis of the substrate
and the release of the product (P).
Thus, the MichaeliseMenten equation was modified according
to Equation (1), where [M] represents the total molar concentration
of the mixed micelle in the bulk solution and cs is the molar frac-
tion of the substrate within the mixed micelle. The fitting param-
eter Ks represents the dissociation constant (kꢀ1/k1) for the mixed
micelleeenzyme complex, Km is the Michaelis constant (k3 þ kꢀ2)/
k2 on the mixed micelle surface and Vmax is the catalytic turnover
rate (k3).
g
lytic serine and the morpholino moiety is positioned in the polar
pocket. The water molecules W267 (subunit A) and W270
(subunit B) in the uninhibited PA3859 protein structure (crystal
form II) are displaced by one of the MES sulphate group oxygen
atoms in the structure of the PA3859-MES protein complex.
Hence, these water molecules could possibly be the candidates for
the nucleophilic attack on the acyl intermediate, leading to the
regeneration of the enzyme.
In crystal form II the enzyme is not covalently inhibited but
a nonaethylene glycol strand of PEGMME 5000 [8], originated from
the crystallization cocktail, is lining the hydrophobic portion of the
cleft and a shorter strand, roughly a diethylene glycol, is located
nearby the end of the cleft (Fig. 1B). Interaction between neigh-
boring dimers is mediated by these shorter PEGMME 5000 strands.
Vmax½Mꢄcs
V ¼
(1)
KmKs þ Km½Mꢄ þ cs½Mꢄ
3.2. Enzymatic assay
It is useful to consider the reciprocal form of Equation (1) so that
1/V is expressed as a function of 1/[M] (Equation (2)).
The PA3859 ability of hydrolyzing p-nitrophenyl (p-NP) ester of
acetic (C2) to stearic acid (C18) was assayed. While acetic, propionic
and butyric acids display a water-solubility good enough to allow
the kcat and Km determination in a homogeneous phase system, the
low solubility of substrates carrying a longer acyl chain required the
addition of 0.2% Tx-100 in the reaction mixture.
For the water-soluble substrates (C2 to C4), the obtained data
well fit the MichaeliseMenten equation. The increase of the acyl
chain length indeed gives rise to a decrease of the Km value (Fig. 2A).
On the contrary, the MichaeliseMenten model showed to be
inadequate to describe the reaction mechanism of the water-
insoluble substrates (C5 to C18). Therefore we exploited the surface
dilution mixed micelle kinetic model (SDMMK) [27,28](Scheme 1).
ꢀ
ꢁ
1
V
KmKs
1
1
Km
cs
¼
þ
1 þ
(2)
V
maxcs ½Mꢄ Vmax
Equation (2) shows that if the initial rate studies are performed
by varying the total mixed micelle concentration [M] while holding
constant the substrate molar fraction within the mixed micelle cs
the plot of 1/V versus 1/[M] gives linear curves. Re-plotting of the
slopes and intercepts of such curves as a function of cs allows the
determination of Vmax, Ks and Km parameters [18,19].
In the case of the water-soluble substrates (C2 to C4), the
presence in the reaction mixture of 0.2% Tx-100 does not affect
significantly the enzymeesubstrate affinity but rather it causes
a reduction of the reaction rate of roughly 70%, hence mimicking
a non competitive-like inhibition (Fig. 2B).
,
For the water-insoluble substrates (C5 to C18), the highest kcat
/
Km ratio value is achieved for p-NP C8, characterized by the
lowest Km and the highest kcat values respectively (Fig. 2C).
Overall, the Km values do not differ significantly for substrates
with chain lengths ranging from C8 to C18 while substrates with
alkyl chain lengths longer than C10 exhibit a dramatic reduction
Scheme 1.

1790
A. Pesaresi, D. Lamba / Biochimie 92 (2010) 1787e1792
Fig. 2. Kinetic parameters for the enzymatic hydrolysis of p-NP esters. The bars represent the data standard deviation. (A) Km
constants derived from the predicted docking energies (empty circles) are plotted as a function of the substrate acyl chain length. The turnover rate kcat (sꢀ1) (B), the specificity
constant kcat/Km (Mꢀ1 sꢀ1) (C) and the dissociation constant for the formation of the mixed micelleeenzyme complex Ks (
M) (D) are plotted as a function of the substrate acyl chain
length. The parameters concerning the hydrolysis of the soluble substrates (C2 to C4) that have been determined also in the absence of Tx-100 are shown as empty circles.
(mM) values (solid circles) and the theoretically binding
m
in the kcat values (Fig. 2B). The affinity of the enzyme for the Tx-
100 substrate mixed micelle, as estimated by the Ks parameter,
roughly corresponds to an overall mixed micelle concentration of
be too long (8 and 9 carbon atoms were located outside the cleft,
see Fig. 3C) and the corresponding torsional energy too high to
force the acyl chain to be accommodated within the binding cleft
(Table 2), pinpointing to stearic acid (C18) as the longest chain
binder.
340
critical micelle concentration (CMC) equal to 240
value is roughly 100 M (Fig. 2D). It is worthy of note that Ks
mM. However, being, in the assay condition, the Tx-100
mM, the likely Ks
m
values are independent of the fatty acid acyl chain length for all
the examined substrates.
4. Discussion
4.1. Enzyme specificity
3.3. Docking simulations
Under the assumption of the SDMMK model, the enzymatic
hydrolysis of p-NP esters with acyl chains ranging from C2 to C18
showed that the highest PA3859 catalytic efficiency, quantified by
the kcat/Km ratio, was achieved by the p-NP ester of octanoic acid. A
clear decrease of the catalytic efficiency has been observed instead
for substrates with acyl chains either shorter or longer than 8
carbon atoms (Fig. 2C). These findings seem to be in contrast to
a distinctive molecular feature observed in the structure of the
The finding, in crystal form II, of a PEGMME 5000 fragment
(nonaethylene glycol) lining the hydrophobic cleft (Fig. 1B) sug-
gested that this may represent the substrate acyl chain binding site.
To explore this hypothesis, a comparative docking study of the
kinetically characterized eleven p-NP carboxylic acid esters (Table
2) has been carried out using the AutoDock program suite [22,23].
A substrate was considered to be successfully docked if after
repeated dockings its carbonyl carbon atom was in the vicinity of
the MES sulphur atom covalently bound to the catalytic serine
(crystal form I).
Among the tested substrates only p-NP acetate could not be
successfully docked. In Table 2 are reported the interaction energies
and the theoretical binding constants for each docked substrate. In
all of the found solutions the p-NP moiety is positioned in the polar
cavity and the acyl chain lies, completely or partially, within the
hydrophobic crevice (Fig. 3). Distances between the substrate
carbonyl carbon atom and the MES sulphur atom ranges between
1.33 and1.48 Å and the carbonyl oxygen atoms always point toward
the oxyanion hole.
Table 2
Predicted automated molecular docking energies.
Substrate Intermolecular Ligand Internal Torsional Free Binding Free Binding
Energy
Energy
Energy
Energy
Constant
M)
(kcal molꢀ1
)
(kcal molꢀ1
)
(kcal molꢀ1
)
(kcal molꢀ1) (
m
p-NP C4
p-NP C6
p-NP C8
p-NP C10 ꢀ9.83
p-NP C12 ꢀ10.13
p-NP C14 ꢀ10.80
p-NP C16 ꢀ11.25
p-NP C18 ꢀ11.79
p-NP C20 ꢀ12.55
p-NP C22 ꢀ12.61
ꢀ6.41
ꢀ7.41
ꢀ8.32
ꢀ0.06
ꢀ0.16
ꢀ0.14
ꢀ0.31
ꢀ0.35
ꢀ0.22
ꢀ0.13
ꢀ0.10
0.13
1.56
2.18
2.80
3.42
4.05
4.67
5.29
5.91
6.54
7.16
ꢀ4.85
ꢀ7.56
ꢀ5.52
ꢀ5.96
ꢀ6.08
ꢀ6.13
ꢀ5.96
ꢀ5.88
ꢀ6.01
ꢀ5.45
276
148
89
43
35
32
43
49
39
For the C4 to C18 substrates an almost identical extended acyl
chain conformation was predicted that nicely fitted the cleft
(Fig. 3A, B). For the C20 and C22 substrates the acyl chain assumed
different conformations characterized by positive internal energies.
The C20 adopted a cork like conformation that allowed the
substrate to be entirely accommodated within the cleft. Instead, in
all of the sampled docking solutions, the C22 acyl chain resulted to
0.78
101
Interaction energies and binding constants for p-nitrophenyl esters of carboxylic
acids with different chain lengths (C4 to C22). The estimated Binding Free Energy is
the summation of the Intermolecular Energy and the Torsional Free Energy that
accounts for the loss of ligand conformational degree of freedom upon binding.

A. Pesaresi, D. Lamba / Biochimie 92 (2010) 1787e1792
1791
Fig. 3. Representative automated molecular docking of p-NP esters bound to PA3859. The best ranking docking solutions are shown: (A) p-NP C4 (red) and C6 (blue). (B) p-NP C16
(red) and C18 (blue). (C) p-NP C20 (red) and C22 (blue). In (D) a non-productive docking of the p-NP C6 is shown: the NO2 moiety, rather than the carbonyl moiety, has been placed
into the catalytic binding site. Similar non-productive dockings were obtained also for p-NP C2 and C4. The coloring scheme used for the molecular surface representation of the
PA3859 protein monomer is the same adopted in Fig. 1B.
crystal form II of PA3859, i.e. the presence of a PEGMME 5000 frag-
ment (nonaethylene glycol) lining an exposed hydrophobic crevice.
This specific structural feature led us to hypothesize that this cleft
could well represent the fatty acid ester acyl chain binding site. In
order to validate this hypothesis, a comparative computational
docking simulation study of the C2 to C22 p-NP esters has been
carried out. The simulations showed that the hydrophobic crevice
can accommodate fatty acid as long as stearic acid (C18) (Table 2
and Fig. 3B, C).
enzyme itself [30e34]. As a result, the kinetic parameters will
depend on the physicochemical properties of the substrates rather
than on the enzyme specificity. This explains why, when the
specificity of lipolytic enzymes has been exploited by using the
SDMMK model using p-NP esters as substrates, almost exclusively
an optimal acyl chain length of 6e8 carbon atoms has been found
[29e31,35,36]. Indeed, the SDMMK model does not take properly
into account the interaction between the substrates and the
surfactants.
Although these results seem incompatible with the enzyme
specificity suggested by the kinetic data, the following consider-
ations may explain the observed differences.
To further exploit this issue we attempted to measure the
hydrolysis of long chain p-NP esters in the absence of detergent
micelles by replacing water with an organic solvent. Enzymes in
organic solvents have been extensively studied [37]. Carbon tetra-
chloride, benzene and 1,4-dioxane are organic solvents commonly
used to dissolve freeze-dried lipases and to study their activity in
the absence of water. In these solvents the p-NP (C2 to C18) fatty
acids esters displayed a good solubility. PA3859 on the contrary,
unlike lipases and other reported enzymes [37], strictly required
the presence of water. PA3859 tolerated a 30% (V/V) water: 1,4-
dioxane mixture (data not shown) but the kcat/Km ratio could not be
determined since in these conditions the substrates with long acyl
chain showed a poor solubility.
Previous studies have reported that in the case of the enzymatic
hydrolysis of carboxylic esters, the structure and the internal
energy of the tetrahedral transition state do not depend on the acyl
chain length of the substrates [29]. Therefore, as long as the turn-
over rate only depends on the ability of the active site to reduce the
activation energy barrier by stabilizing the tetrahedral transition
state, the kcat values for the hydrolysis of p-NP esters of fatty acids
would not be expected to display dependence on the substrate
chain length. Thus, the enzyme preference for substrates with
different acyl chain length should only reflect the affinity of the
substrates for the enzyme i.e. the Km values.
The simple SDMMK model for the water-insoluble p-NP esters
as substrates has been exploited and this is likely the cause of the
disagreement between the above discussed expectation and the
observed kinetic data (Fig. 2B). Indeed it has been suggested that
one of the limitations of the SDMMK model for the study of lipolytic
enzymes is that if the substrate replenishment rate, defined as the
rate of the substrate to exchange between micelles, is not faster
than the enzyme turnover rate, the reaction kcat will be limited by
the replenishment rate rather than by the catalytic efficiency of the
4.2. Enzyme e p-NP esters affinity and mode of interaction
Docking simulations, using the AutoDock program suite,
provided a rationale for the substrate affinity shown by PA3859.
The predicted binding constants are in good agreement with the Km
values for the enzymatic hydrolysis of p-NP esters derived
according to the SDMMK model. As shown in Fig. 2A, the experi-
mentally derived Km values and the in vacuo AutoDock computed
binding constants, display a very similar substrate chain length

1792
A. Pesaresi, D. Lamba / Biochimie 92 (2010) 1787e1792
[11] J.A. Duncan, A.G. Gilman, A cytoplasmic acyl-protein thioestease that removes
palmitate from G protein alpha subunits and p21 (RAS), J. Biol. Chem. 273
(1998) 15830e15837.
dependence. Docking simulations suggested that the high Km
values obtained for p-NP C2 to C6 are a consequence of their
preferential non-productive interaction with the enzyme active site
(Fig. 3D).
[12] K.K. Kim, H.K. Song, D.H. Shin, K.Y. Hwang, S. Choe, O.J. Yoo, S.W. Sush, Crystal
structure of carboxylesterase from Pseudomonas fluorescens, an
with broad substrate specificity, Structure 5 (1997) 1571e1584.
[13] K. Murayama, M. Shirouzu, T. Terada, S. Kuramitsu, S. Yokoyama, Crystal
Structure of TT1662 from Thermus thermophilus HB8: a member of the
hydrolase fold enzymes, Proteins 58 (2005) 982e984.
a/b hydrolase
PA3859 displays similar affinity for p-NP substrates with acyl
chains from C8 to C18 (Fig. 2A). The data in Table 2 also show a lack
of dependence between substrate chain lengths and the predicted
binding constants: as the length of the substrate chain increases
(C8 to C22), the predicted decrease of the intermolecular energy,
due to the increased number of interacting atoms, is almost
perfectly balanced by the predicted increase of the torsional free
energy, due to the substrate undergoing from an unbound to
a bound state. As a consequence the overall binding free energy,
and thus the binding constant, does not vary for substrates C8 to
C20.
Overall, the results discussed above, demonstrate that PA3859 is
indeed a lipolytic enzyme. The apparent optimal substrate chain
length of 8 carbon atoms found for PA3859, rather than being
representative of real substrate specificity is likely to be merely the
consequence of the substrates intrinsic properties due to the limi-
tations of the SDMMK model adopted. Computational docking
simulations carried out in order to gain insights into the hydro-
phobic determinants of the binding site specificity support the
hypothesis that carboxylic esters derived from the condensation of
palmitic and/or stearic acid are the likely physiological substrates.
Furthermore, PA3859 is unable of hydrolyzing triacylglycerols [7].
Efforts are currently under way to infer the binding specificity of
the polar active-site pocket of the enzyme in order to unveil its
primary in vivo function.
a
/
b
[14] A. Dong, et al., In situ proteolysis for protein crystallization and structure
determination, Nat. Methods 4 (2007) 1019e1021.
[15] K.H. Hong, W.H. Jang, K.D. Choi, O.J. Yoo, Characterization of Pseudomonas
fluorescens carboxylesterase: cloning and expression of the esterase gene in
Escherichia coli, Agric. Biol. Chem. 55 (1991) 2839e2845.
[16] K.K. Kim, K.Y. Hwang, W.H. Jang, H. Lee, O.J. Yoo, M.U. Choi, S.W. Suh, Crys-
tallization and preliminary X-ray crystallographic analysis of carboxylesterase
from Pseudomonas fluorescens, Arch. Biochem. Biophys. 302 (1993) 417e419.
[17] A.T. Brünger, P.D. Adams, G.M. Clore, W.L. DeLano, P. Gros, R.W. Grosse-
Kunstleve, J.S. Jiang, J. Kuszewski, M. Nilges, N.S. Pannu, R.J. Read, L.M. Rice,
T. Simonson, G.L. Warren, Crystallography & NMR system: a new software
suite for macromolecular structure determination, Acta Crystallogr. D54
(1998) 905e921.
[18] A.T. Brünger, Version 1.2 of the crystallography and NMR system, Nat. Protoc.
2 (2007) 2728e2733.
[19] T.A. Jones, J.Y. Zou, S.W. Cowan, M. Kjeldgaard, Improved methods for building
protein models in electron density maps and the location of errors in these
models, Acta Crystallogr. A47 (1991) 110e119.
[20] R.A. Laskowski, M.W. MacArthur, D.S. Moss, J.M. Thornton, PROCHECK:
a program to check the stereochemical quality of protein structures, J. Appl.
Crystallogr. 26 (1993) 283e291.
[21] W.L. DeLano, The Pymol Molecular Graphics System. Delano Scientific, Palo
Alto, CA, USA, 2002.http://www.pymol.org.
[22] G.M. Morris, D.S. Goodsell, R.S. Halliday, R. Huey, W.E. Hart, R.K. Belew,
A.J. Olson, Automated docking using a Lamarckian genetic algorithm and
empirical binding free energy function, J. Comp. Chem. 19 (1998) 1639e1662.
[23] D.S. Goodsell, G.M. Morris, A.J. Olson, Docking of flexible ligands: applications
of AutoDock, J. Mol. Recognition 9 (1996) 1e5.
[24] J. Gasteiger, M. Marsili, Iterative partial equalization of orbital electronega-
tivity e a rapid access to atomic charges, Tetrahedron 36 (1980) 3219e3228.
[25] D.M. Hayes, P.A. Kollman, Electrostatic potentials of proteins. 1. Carboxy-
peptidase A, J. Am. Chem. 98 (1976) 3335e3345.
[26] H. van Tilbeurgh, M.P. Egloff, C. Martinez, N. Rugani, R. Verger, C. Cambillau,
Interfacial activation of the lipaseeprocolipase complex by mixed micelles
revealed by X-ray crystallography, Nature 362 (1993) 814e820.
[27] R.A. Deems, B.R. Eaton, E.A. Dennis, Kinetic analysis of phospholipase A2
activity toward mixed micelles and its implication for the study of lipolytic
enzymes, J. Biol. Chem. 250 (1975) 9013e9020.
Acknowledgments
We are grateful to Dr. Francesco Secundo (ICRM-CNR, Milano)
for critical reading of the manuscript and useful discussion.
[28] M.E. Haberland, J.A. Reynolds, Interaction of L-a-palmitoyl lysophosphati-
dylcholine with the AI polypeptide of high density lipoprotein, J. Biol. Chem.
250 (1975) 6636e6639.
References
[29] R.A. Burdette, D.M. Quinn, Interfacial reaction dynamics and acyl-enzyme
mechanism for lipoprotein lipase-catalyzed hydrolysis of lipid p-nitrophenyl
esters, J. Biol. Chem. 261 (1986) 12016e12021.
[30] O. Redondo, A. Herrero, J.F. Bello, M.G. Roig, M.V. Calvo, F.J. Plou, F.J. Burguillo,
Comparative kinetic study of lipase A and B from Candida rugosa in the
hydrolysis of lipid p-nitrophenyl esters in mixed micelles with Triton X-100,
Biochim. Biophys. Acta 1243 (1995) 15e24.
[31] A. Moraleda-Muñoz, L.J. Shimkets, Lipolytic enzymes in Myxococcus xanthus, J.
Bacteriol. 189 (2007) 3072e3080.
[32] O.G. Berg, M.K. Jain, Interfacial Enzyme Kinetics. J. Wiley & Sons Ltd., Chi-
chester, England, 2002.
[33] J.W. Nichols, Thermodynamics and kinetics of phospholipid monomer-vesicle
interaction, Biochemistry 24 (1975) 6390e6398.
[34] J.W. Nichols, R.E. Pagano, Use of resonance energy transfer to study the
kinetics of amphiphile transfer between vesicles, Biochemistry 21 (1982)
1720e1726.
[35] C. Alquati, L. De Gioia, G. Santarossa, L. Alberghina, P. Fantucci, M. Lotti, The
cold-active lipase of Pseudomonas fragi. Heterologous expression, biochemical
characterization and molecular modeling, Eur. J. Biochem. 269 (2002)
3321e3328.
[36] G. Santarossa, P.G. Lafranconi, C. Alquati, L. De Gioia, L. Alberghina, P. Fantucci,
M. Lotti, Mutations in the “lid” region affect chain length specificity and
[1] T. Hotelier, L. Renault, X. Cousin, V. Negre, P. Marchot, A. Chatonnet, ESTHER,
the database of the a/b hydrolase fold superfamily of proteins, Nucleic Acids
Res. 32 (2004) D145eD147.
[2] D.L. Ollis, E. Cheah, M. Cygler, B. Dijkstra, F. Frolow, S.M. Franken, M. Harel,
S.J. Remington, I. Silman, J. Schrag, J.L. Sussman, K.H.G. Verschueren,
A. Goldman, The
[3] P. Heikinheimo, A. Goldman, C. Jeffries, D.L. Ollis, Of barn owls and bankers:
lush variety of alpha/beta hydrolases, Struct. Fold. Des (R) (1999)
a/b hydrolase fold, Prot. Eng. 5 (1992) 197e211.
a
7
R141eR146.
[4] M. Nardini, B.W. Dijkstra,
a/b hydrolase fold enzymes: the family keeps
growing, Curr. Opin. Struct. Biol. 9 (1999) 732e737.
[5] L.M. Johnson, H.W. Talbot, Detoxification of pesticides by microbial enzymes,
Experientia 39 (1983) 1236e1246.
[6] U.T. Bornscheuer, Microbial carboxyl esterases: classification, properties and
application in biocatalysis, FEMS Microbiol. Rev. 26 (2002) 73e81.
[7] A. Pesaresi, G. Devescovi, D. Lamba, V. Venturi, G. Degrassi, Isolation, charac-
terization and heterologous expression of a carboxylesterase of Pseudomonas
aeruginosa PAO1, Curr. Microbiol. 50 (2005) 102e109.
[8] A. Pesaresi, D. Lamba, Crystallization, X-ray diffraction analysis and phasing of
carboxylesterase PA3859 from Pseudomonas aeruginosa, Biochim. Biophys.
Acta 1752 (2005) 197e201.
[9] H.M. Berman, J. Westbrook, Z. Feng, G. Gilliland, T.N. Bhat, H. Weissig,
I.N. Shindyalov, P.E. Bourne, The protein data bank, Nucleic Acids Res. 28
(2000) 235e242.
[10] Y. Devedjiev, Z. Dauter, S.R. Kuznetsov, T.L. Jones, Z.S. Derewenda, Crystal
structure of the human acyl protein thioesterase I from a single X-ray data set
to 1.5Å, Structure 8 (2000) 1137e1146.
thermostability of
a Pseudomonas fragi lipase, FEBS Lett. 579 (2005)
2383e2386.
[37] F. Secundo, S. Spadaro, G. Carrea, P.L. Overbeeke, Optimization of Pseudomonas
cepacia lipase preparations for catalysis in organic solvents, Biotechnol. Bio-
eng. 62 (1999) 554e561.