Functional annotation and three-dimensional structure of an incorrectly annotated dihydroorotase from cog3964 in the amidohydrolase superfamily

Article abstract of DOI:10.1021/bi301483z

The substrate specificities of two incorrectly annotated enzymes belonging to cog3964 from the amidohydrolase superfamily were determined. This group of enzymes are currently misannotated as either dihydroorotases or adenine deaminases. Atu3266 from Agrobacterium tumefaciens C58 and Oant2987 from Ochrobactrum anthropi ATCC 49188 were found to catalyze the hydrolysis of acetyl-(R)-mandelate and similar esters with values of k_cat/K _m that exceed 10⁵ M^-1 s^-1. These enzymes do not catalyze the deamination of adenine or the hydrolysis of dihydroorotate. Atu3266 was crystallized and the structure determined to a resolution of 2.62 ?. The protein folds as a distorted (β/α)₈ barrel and binds two zincs in the active site. The substrate profile was determined via a combination of computational docking to the three-dimensional structure of Atu3266 and screening of a highly focused library of potential substrates. The initial weak hit was the hydrolysis of N-acetyl-d-serine (k_cat/K_m = 4 M^-1 s ^-1). This was followed by the progressive identification of acetyl-(R)-glycerate (k_cat/K_m = 4 × 10² M^-1 s^-1), acetyl glycolate (k_cat/K_m = 1.3 × 10⁴ M^-1 s^-1), and ultimately acetyl-(R)-mandelate (k_cat/K_m = 2.8 × 10⁵ M^-1 s^-1).

Full text of DOI:10.1021/bi301483z

Article
pubs.acs.org/biochemistry
Functional Annotation and Three-Dimensional Structure of an
Incorrectly Annotated Dihydroorotase from cog3964 in the
Amidohydrolase Superfamily
Argentina Ornelas,† Magdalena Korczynska,§ Sugadev Ragumani,‡ Desigan Kumaran,‡
Tamari Narindoshvili,† Brian K. Shoichet,*,§ Subramanyam Swaminathan,*,‡ and Frank M. Raushel*,†
†Department of Chemistry, P.O. Box 30012, Texas A&M University, College Station, Texas 77842-3012, United States
‡Biology Department, Brookhaven National Laboratory, P.O. Box 5000, Upton, New York 11973-5000, United States
§Department of Pharmaceutical Chemistry, University of California, San Francisco, 1700 Fourth Street, San Francisco, California
94158-2330, United States
S
* Supporting Information
ABSTRACT: The substrate specificities of two incorrectly
annotated enzymes belonging to cog3964 from the amidohy-
drolase superfamily were determined. This group of enzymes
are currently misannotated as either dihydroorotases or
adenine deaminases. Atu3266 from Agrobacterium tumefaciens
C58 and Oant2987 from Ochrobactrum anthropi ATCC 49188
were found to catalyze the hydrolysis of acetyl-(R)-mandelate
and similar esters with values of kcat/Km that exceed 105 M−1
s−1. These enzymes do not catalyze the deamination of adenine or the hydrolysis of dihydroorotate. Atu3266 was crystallized and
the structure determined to a resolution of 2.62 Å. The protein folds as a distorted (β/α)8 barrel and binds two zincs in the active
site. The substrate profile was determined via a combination of computational docking to the three-dimensional structure of
Atu3266 and screening of a highly focused library of potential substrates. The initial weak hit was the hydrolysis of N-acetyl-D-
serine (kcat/Km = 4 M−1 s−1). This was followed by the progressive identification of acetyl-(R)-glycerate (kcat/Km = 4 × 102 M−1
s−1), acetyl glycolate (kcat/Km = 1.3 × 104 M−1 s−1), and ultimately acetyl-(R)-mandelate (kcat/Km = 2.8 × 105 M−1 s−1).
that contains a mono- or binuclear metal center embedded at
he development of a comprehensive strategy for the
Tfunctional annotation of proteins and enzymes whose the C-terminal end of a distorted (β/α)8 barrel protein fold.
Typically identified by an HxH motif after the C-terminal end
of β-strand 1, the enzymes from the AHS also contain other
coordinating ligands to the metal center at the ends of β-strands
4−6 and 8.14 The metal center serves in the activation of a
hydrolytic water molecule for nucleophilic attack and to
stabilize the transition state. Most of the reactions catalyzed
by this diverse superfamily involve the hydrolysis of C−O, C−
N, or P−O bonds. However, some members of this superfamily
also catalyze decarboxylation, hydration, or isomerization
reactions.15−17 Enzymes from the AHS catalyze reactions in
the metabolism of carbohydrates, amino acids, and nucleic acids
and the degradation of organophosphate esters.
sequences have been deposited in public databases has proven
to be a difficult and persistent challenge. Utilization of
homology-based sequence comparisons for functional annota-
tion of newly sequenced genes can often lead to the annotation
of the wrong function when unreasonable threshold values are
used.1−3 The end result is often a misrepresentation of the
potential metabolic transformations contained within a given
organism. In addition to the misannotation of enzyme function,
there is a significant fraction of the total gene inventory that is
simply not annotated.4−6 This observation suggests that a
substantial segment of the metabolic landscape remains to be
discovered.
The AHS has been organized into 24 Clusters of
Orthologous Groups (COG).18,19 One of these clusters,
cog3964, represents one of the smaller homologous groups of
amidohydrolases, with approximately 200 sequences identified
to date. A sequence similarity network for this COG is
presented in Figure 1 at an E value cutoff of 10−70. Some
members of cog3964 have been annotated as dihydroorotases,
In many cases, functional annotation of newly sequenced
genes can be strategically approached by the integration of
structural genomics,7,8 computational docking,9,10 high-
throughput screening,11 and genome context analysis.12 We
have focused our efforts toward the development of a simple
and integrated strategy for functional annotation that is based
on an assault on the amidohydrolase superfamily (AHS) of
enzymes. This superfamily was first identified and recognized
from the three-dimensional structural similarities among urease,
adenosine deaminase, and phosphotriesterase.13 The amidohy-
drolase superfamily is characterized structurally by an active site
Received: October 31, 2012
Revised: December 7, 2012
Published: December 10, 2012
© 2012 American Chemical Society
228
dx.doi.org/10.1021/bi301483z | Biochemistry 2013, 52, 228−238

Biochemistry
Article
purification, and screening of Atu3266 and Oant2987 were
obtained from Sigma-Aldrich, unless stated otherwise. The
preparation of the acylated compounds tested for substrate
activity is provided in the Supporting Information. The
structures for many of these compounds are provided in
Schemes 2 and 3. The pET-20b(+) expression vector and
Rosetta-gami B(DE3)pLysS competent cells were obtained
from Novagen. The acetate and formate detection kits were
purchased from Megazyme. N-Acetyl-DL-phenylglycine was
purchased from Chem-Impex International. 3-Bromomandelic
acid and 4-bromomandelic acid were purchased from Oakwood
Products, Inc.
Gene Cloning. The gene for Oant2987 (gi|153010310)
from O. anthropi was amplified by polymerase chain reaction
using the primer pair of 5′-ACAGGAGCCATATGATTTCC-
GGTGAACAGGCGAAGCCG-3′ containing an NdeI restric-
tion site and 5′-ACGCGAATTCCCAGCGCCACGAATAGC-
CATGGCTATGGC-3′ having an EcoRI restriction site.
Oant2987 was inserted into a pET-20b(+) vector that had
been previously digested with NdeI and EcoRI. The gene for
Atu3266 was initially cloned by the New York SGX Research
Consortium for Structural Genomics from A. tumefaciens (gi|
159185666). The poor overexpression and solubility of this
clone in BL21(DE3) cells led to the removal of the gene from
its original TOPO-isomerase vector, pSGX3(BC), using the
restriction sites CCGC↓GG and CA↓TATG of the construct
and removing the gene by digestion with SacII and NdeI
restriction enzymes. The gene was transferred to a pET-20b(+)
vector previously digested with the same restriction enzymes.
The modified clone contained a His6 C-terminal purification
tag.
Figure 1. Sequence similarity network for the organization of cog3964
within the amidohydrolase superfamily.18,19 Each node in the diagram
represents a sequence that has been annotated as belonging to
cog3964, and each edge represents the pairwise connection between
two sequences at a BLAST E value of better than 10−70. Atu3266 and
Oant2987 are found in group 6 and are colored purple.
which catalyze the interconversion of dihydroorotate and N-
carbamoyl aspartate,20,21 while other members have been
annotated as catalyzing the deamination of adenine.22
Experimentally annotated dihydroorotases are found in
cog0044 and cog0418. Structurally and experimentally
characterized adenine deaminases are found in cog100122 and
in cog1816.23 A straightforward comparison of the amino acid
sequences found within the proteins from cog3964 with the
experimentally annotated adenine deaminases and dihydroor-
otases clearly indicates that the residues required for substrate
recognition of adenine and/or dihydroorotate are not
conserved. This observation suggests that either these
annotations are clearly wrong for cog3964 or a previously
unrecognized novel constellation of amino acids has evolved for
the deamination of adenine or the hydrolysis of dihydroorotate.
In this paper, we report the three-dimensional crystal
structure and substrate profile for two enzymes from
cog3964: Atu3266 from Agrobacterium tumefaciens and
Oant2987 from Ochrobactrum anthropi. Neither of these
enzymes is able to catalyze the hydrolysis of dihydroorotate
or the deamination of adenine. We have utilized a focused
compound library and computational docking to discover that
these two enzymes from cog3964 actually catalyze the
hydrolysis of acetylated α-hydroxyl carboxylates as shown in
Scheme 1. The best substrate identified to date is acetyl-(R)-
mandelate. This compound is enzymatically hydrolyzed with a
kcat/Km value of 2.8 × 105 M−1 s−1.
Expression and Purification of Oant2987 and
Atu3266. The plasmid containing the gene for Oant2987
was transformed into BL21(DE3) and plated on LB agar/
ampicillin plates. A single colony was inoculated into overnight
cultures containing 5 mL of LB medium supplemented with
100 μg/mL ampicillin. These cultures were utilized to make 1 L
cultures containing the same concentration of ampicillin and
grown at 30 °C in a shaker-incubator. After an OD600 of 0.3 had
been reached, the cells were supplemented with 150 μM 2,2′-
bipyridyl to coordinate excess ferric iron from the medium. At
an OD600 of 0.6, the cells were induced with 200 μM IPTG and
the addition of 1.0 mM Zn(OAc)2. The temperature upon
induction was reduced to 20 °C, and the cultures were allowed
to continue to grow for an additional 12 h. The cells were
harvested by centrifugation and frozen at −80 °C. The frozen
cell pellet was thawed and resuspended in 60 mL of 50 mM
HEPES buffer (pH 7.6) (buffer A). The cells were
supplemented with 10 μg/mL phenylmethanesulfonyl fluoride
(PMSF) and lysed by sonication at 0 °C. The supernatant
solution was treated with 2% (w/v) protamine sulfate and
centrifuged. The protein was precipitated with ammonium
sulfate between 40 and 80% saturation. The protein pellet was
resuspended with a minimal amount of buffer A and loaded
onto a Superdex 200 gel filtration column. The fractions of
interest were collected and subsequently loaded onto a
Resource Q column. The protein was eluted from the column
with a gradient of 20 mM HEPES buffer (pH 7.6) containing 1
M NaCl (buffer B). The purity of Oant2987 was greater than
95% based on sodium dodecyl sulfate−polyacrylamide gel
electrophoresis (SDS−PAGE).
Scheme 1
MATERIALS AND METHODS
■
Materials. The genomic DNA from O. anthropi strain
ATCC-49188 was purchased from the American Type Culture
Collection (ATCC). DNA sequencing reactions were con-
ducted at the Gene Technology Laboratory at Texas A&M
University. The chemicals utilized in the expression,
The plasmid containing the gene for Atu3266 was trans-
formed into Rosetta-gami B(DE3)pLysS electrocompetent
229
dx.doi.org/10.1021/bi301483z | Biochemistry 2013, 52, 228−238

Biochemistry
Article
Scheme 2
Scheme 3
cells. A single colony containing the gene of interest was
inoculated into an overnight culture containing 5 mL of LB
medium, 100 μg/mL ampicillin, and 20 μg/mL chloramphe-
nicol. The 5 mL starter cultures were used to inoculate 1 L of
LB medium. The cells were incubated at 30 °C in a shaker-
incubator. When the OD600 reached 0.3, the cells were treated
230
dx.doi.org/10.1021/bi301483z | Biochemistry 2013, 52, 228−238

Biochemistry
Article
with 150 μM 2,2′-bipyridyl. The cells continued to grow to an
OD600 of ∼0.6, at which point 200 μM IPTG and 1.0 mM
ZnCl2 were added. The temperature was reduced to 20 °C, and
the cells were grown overnight for 16 h. The cells where
harvested by centrifugation at 8000g for 10 min and then frozen
at −80 °C. The cell pellet was resuspended in binding buffer
containing 20 mM HEPES, 500 mM NaCl, and 5.0 mM
imidazole (pH 7.6). The cells were supplemented with 10 μg/
mL PMSF and lysed by sonication at 0 °C. The cell debris was
removed by centrifugation, and then Atu3266 was purified
using a His tag affinity column. The cell lysate supernatant was
filtered through a 0.2 μm cellulose acetate sterile filter and
loaded onto a Ni2+-NTA column equilibrated with binding
buffer. The column was washed with binding buffer until the
A280 remained constant and below 0.1. The protein of interest
was eluted using a gradient from 0 to 500 mM imidazole in a
buffer solution containing 10 mM HEPES and 250 mM NaCl
(pH 7.6). After dialysis, the protein was judged to be >95%
pure by SDS−PAGE.
Protein Concentration and Metal Analysis. The protein
concentration was determined spectrophotometrically using a
SPECTRAmax-384 PLUS UV−vis spectrophotometer. The
concentration was obtained from the absorbance at 280 nm
using the extinction coefficients determined from the amino
acid sequence (http://web.expasy.org/protparam/). The ex-
tinction coefficients for Oant2987 and Atu3266 are 43605 and
43969 M−1 cm−1, respectively.24 The metal content of the
purified proteins was determined by inductively coupled plasma
emission mass spectrometry (ICP-MS) using a Perkin-Elmer
Analyst 700 atomic absorption spectrometer. The samples were
prepared by heating 1.0 μM enzyme with 1% (v/v) HNO3 for
30 min.
negative control was prepared without the addition of either
Atu3266 or Oant2987. The initial screening reactions were
conducted at 30 °C for 15 h. The formation of free amino acids
was detected using a modified Cd-ninhydrin assay.27 Each 70
μL reaction mixture was quenched with 280 μL of ninhydrin
reagent. The entire reaction mixture was developed heating at
85 °C for 15 min and then cooled. A 250 μL aliquot was
transferred to a 96-well UV−visible micro plate, and the extent
of total free amino acids was measured at 507 nm.
Screening Methods for Deacetylase Activity. All
compounds having a hydrolyzable acetyl moiety were screened
using the acetic acid kit, KACETAF, from Megazyme. The
catalytic activities of Atu3266 and Oant2987 were monitored
by the formation of acetate and the subsequent reduction of
NAD+ by the coupled activity of acetyl-coenzyme A synthetase,
citrate synthase, and L-malate dehydrogenase. The reaction was
monitored spectrophotometrically at 340 nm. The 250 μL
reaction mixture contained 1.0 mM test compound, 20 mM
HEPES (pH 7.6), and 75 μL of the coupling system containing
128 mM TEA buffer (pH 8.4), 5.0 mM NAD+, 3.1 mM ATP,
3.2 mM MgCl2, 0.15 mM CoA, 4 units of L-malate
dehydrogenase, 0.6 unit of citrate synthase, and 0.3 unit of
acetyl-CoA synthetase. Each compound was tested for ester
hydrolysis at 30 °C, and the reaction was initiated by the
addition of 1.0 μM Atu3266 or Oant2987.
Screening of O-Propionyl Compounds. The hydrolysis
of (R)-3-hydroxy-2-(propionyloxy) propanoate (21) and (R)-2-
(propionyloxy) butanoate (22) was assessed using a pH-
sensitive colorimetric assay.28 The hydrolysis of the ester bond
releases a proton that was detected using a pH indicator dye,
cresol purple. The screening reactions were conducted in 2.5
mM Bicine buffer (pH 8.3) containing 0.2 M NaCl and each
compound at 1.0 mM with up to 1.0 μM enzyme. Each reaction
mixture contained 0.1 mM cresol purple in 1% DMSO, and the
change in absorbance was monitored at 577 nm. The extinction
coefficient under these reaction conditions was determined to
be 1.51 × 103 M−1 cm−1 using acetic acid as a titrant.
Data Analysis. The kinetic constants kcat, Km, and kcat/Km
were determined for Atu3266 and Oant2987 for selected
substrates by fitting the initial velocity data to eq 1. Competitive
inhibition constants were determined using eq 2 for a
competitive inhibitor.
Synthesis of Dipeptide Libraries, N-Acyl Amino Acids,
and Methyl Phosphonates. Syntheses of dipeptide libraries
(L-Xaa-L-Xaa, D-Xaa-L-Xaa, and L-Xaa-D-Xaa) and N-acetyl, N-
formyl, and N-succinyl derivatives of D- and L-amino acids were
conducted as described previously.25,26 The preparation of the
O- and N-methyl phosphonate derivatives of (R)-mandelate
(57) and D-phenylglycine (58), respectively, was conducted as
previously described.25 The structures of these compounds are
presented in Scheme 4; their composition was verified by mass
spectrometry and nuclear magnetic resonance (NMR).
v/Et = (kcat[A])/(Km + [A])
(1)
Scheme 4
v/Et = (kcat[A])/[Km(1 + I/Kis) + [A]]
(2)
where v is the initial velocity, Et is the enzyme concentration,
kcat is the turnover number, [A] is the substrate concentration,
Km is the Michaelis−Menten constant, I is the inhibitor
concentration, and Kis is the slope inhibition constant.
Crystallization and Structure Determination of
Atu3266. Selenomethionine (SeMet)-substituted Atu3266
from Agrobacterium tumefaciens with the six-His tag intact was
crystallized by mixing 1.5 μL of protein [5 mg/mL in a buffer
consisting of 10 mM HEPES (pH 7.5), 150 mM NaCl, 10 mM
methionine, 10% glycerol, and 5 mM DTT] with 1.5 μL of a
reservoir solution containing 0.1 M HEPES (pH 7.5), 25%
PEG 3350, and 0.2 M ammonium sulfate and equilibrating
against the same reservoir solution by the hanging drop vapor
diffusion method. Crystals appeared after 3 days and were
flashed-cooled in liquid nitrogen for data collection.
Screening of N-Acyl and Dipeptide Libraries. The
preliminary substrate screening of Atu3266 and Oant2987 was
initiated by mixing each protein with the N-modified libraries of
the D- and L-amino acids (N-acetyl, N-formyl, and N-succinyl)
and three dipeptide libraries (L-Xaa-L-Xaa, D-Xaa-L-Xaa, and L-
Xaa-D-Xaa). The assays were conducted as previously
described.26 Each assay was buffered with 20 mM HEPES
(pH 7.6), and each library contained 17−19 modified amino
acids (L- and D-cysteine were not included). Each assay
contained each component at ∼100 μM, and 0−1000 nM
Atu3266 or Oant2987 was added to initiate the reaction. A
A native data set and a complete MAD data set from single
crystals were collected at 100 K on beamline 31-ID at APS
231
dx.doi.org/10.1021/bi301483z | Biochemistry 2013, 52, 228−238

Biochemistry
Article
using a Mar CCD 225 detector. Crystals diffracted to 2.62 Å
and belong to orthorhombic space group P212121 with six
molecules in the asymmetric unit. Data were indexed,
integrated, and scaled using HKL2000.29 The selenium
substructure was determined using SHELXD.30 The phase
refinement and density modifications were conducted using
SHAARP and SOLOMON.31,32 Model building was conducted
using ARP/wARP.33 Further model building and refinement of
the structure were conducted in iterative cycles using O and
CNS version 1.1.34,35 Extra residual density was observed near
the NZ atom of Lys-175, and the lysine residue was modeled as
a carboxylated lysine. In addition, two zinc ions and one
imidazole were located per molecule and included in the later
stages of refinement. The atomic coordinates and structure
factors for the Atu3266 structure have been deposited in the
Protein Data Bank as entry 2OGJ. Crystal, diffraction data, and
refinement details are given in Table 1.
transition state with variable substrate protonation states.9,36
This library contains 6440 unique metabolites that generated
57672 HEI molecules. This virtual HEI library (HEI KEGG)
was docked against Atu3266.
The Atu3266 structure was prepared for docking as
previously described.37 All six chains from the asymmetric
unit of Atu3266 [Protein Data Bank (PDB) entry 2OGJ] were
aligned against chain A, and the best rotamer positions for each
amino acid were chosen on the basis of 2FoFc density maps. For
example, the positions of residues 177, 267, and 268 near the
active site were chosen on the basis of chain D positions.
Histidine protonation states were manually defined in the
vicinity of the two zinc ions, directing polar hydrogens away
from the binuclear metal center. All other polar hydrogens were
automatically generated. The charges on the α- and β-zinc ions
were assigned as 1.4 and 1.3, respectively,37 and the remaining
charge difference was distributed to the metal ligating residues
[His-77, His-79, His-208, and His-231 (0.25 each); Asp-291
(−0.9); and Kcx-175 (−0.8)] to give each metal an apparent
charge of 2.0.
Table 1. Data Collection and Refinement Statistics for
Atu3266
Molecules were docked into the active site of Atu3266 with
DOCK version 3.638,39 (http://dock.compbio.ucsf.edu/).
Docking was performed using receptor and ligand bin sizes
of 0.4 Å, an overlap of 0.1−0.2 Å, and a distance tolerance of
1.5 Å with label matching turned off. The docked molecules
were subjected to 250 cycles of rigid-body minimization and
were scored on the basis of van der Waals, electrostatic, and
solvation terms. Each pose of the scored HEI metabolites was
filtered, based on a maximal distance of 4 Å between the
reactive center of the HEI molecule and the zinc metal center.
Finally, the top 500 scored molecules were manually inspected
to identify potential substrates. Some of the docked substrates
were further subjected to minimization with SZYBKI;39 this
was used to confirm substrate fit and to further investigate
hydrogen bond networks.
Because many molecules were synthesized de novo for this
project, they were not present in the KEGG library. To
investigate the stereospecificity of the compounds hydrolyzed
by Atu3266 and Oant2987, a dedicated library (HEI AC)
containing all of the experimentally tested compounds
(Schemes 2 and 3, compounds 1−56) was generated as
previously described for the HEI KEGG library.9,36 To further
identify new ligands, the HEI AC library was expanded to
contain other molecules with the acetylated α-hydroxyl
carboxylate motif [smiles string OC(O)C(R2)OC(O)C],
and molecules with insertions around the chiral carbon {smiles
string [OC(O)XnC(R2)XnOC(O)Xn]}. These new com-
pounds were docked and scored in the same manner as the HEI
KEGG library molecules.
Data Set and Unit Cell Parameters
beamline
a (Å)
NSLS X12C
90.65
b (Å)
139.2
c (Å)
206.13
90
α = β = γ (deg)
space group
P212121
Data Collection Statistics
peak
inflection
remote
wavelength (Å)
resolution range (Å)
redundancy
0.9793
50−2.6
5.8
0.9798
50−2.8
5.5
0.94
50−2.8
5.2
mean I/σ(I)
14.5
13.2
12.8
a
Rmerge
0.121
0.125
138603
0.118
138965
no. of unique reflections
139906
Phasing Statistics
b
phasing power (ano)
0.21/0.28
c
FOM (centric/acentric)
0.85
Refinement Statistics
resolution range (Å)
no. of reflections (work)
no. of reflections (test)
no. of protein atoms
no. of heterogen atoms
2.62−39.7
139800
4173
16450
22
no. of solvent atoms
113
d
R
factor, Rfree
0.256, 0.302
0.016
1.6
root-mean-square deviation for bond lengths (Å)
root-mean-square deviation for bond angles (deg)
⟨B values⟩ (Å2)
58.3
Protein Data Bank entry
2OGJ
RESULTS
a
■
Rmerge = ∑j|Ih − ⟨Ih⟩|/∑Ih where ⟨Ih⟩ is the average intensity over
b
c
Purification of Atu3266 and Oant2987. Atu3266 was
purified to homogeneity, and the identity of this protein was
confirmed by sequencing the first eight amino acids from the
N-terminus. Oant2987 was purified to homogeneity using gel
filtration and anion exchange chromatography. Each protein
was >95% pure based on SDS−PAGE analysis. ICP-MS
confirmed the presence of 2.0 0.2 equiv of Zn per subunit
of Atu3266 and 1.0 0.1 equiv of Zn per subunit of Oant2987
with <0.1 equiv of iron and manganese. The addition of
supplemental zinc to Oant2987 did not improve the catalytic
hydrolysis of acetyl-(R)-mandelate.
symmetry equivalents. As defined in SHARP. FOM is the figure of
d
merit as defined by SHARP. Rfactor = ∑|Fobs − Fcalc|/∑|Fobs|. Rfree is
the same as Rfactor but for the test set.
Computational Docking. A nonredundant virtual library
of compounds from the Kyoto Encyclopedia of Genes and
Genomes (KEGG) was filtered on the basis of possible
amidohydrolase reactions and subsequently prepared as high-
energy intermediates (HEIs). The HEI compounds are
generated by the addition of an activated water to the
metabolite molecules from KEGG, which approximates the
232
dx.doi.org/10.1021/bi301483z | Biochemistry 2013, 52, 228−238

Biochemistry
Article
Three-Dimensional Structure of Atu3266. The three-
dimensional structure of Atu3266 was determined to a
resolution of 2.62 Å (PDB entry 2OGJ). The six protomers
in the asymmetric unit form a dimer of trimers as illustrated in
Figure 2. The hexamer appears as two disks (made of three
Figure 3. Ribbon representation of the three-dimensional structure
domain of Atu3266. The metal-coordinating ligands are colored
purple. The central β-barrel is colored gray; the helices are colored
blue. Colored rose and tan are the N- and C-terminal domains of the
enzyme, respectively.
Figure 4. Structural model for the active site of Atu3266 showing the
residues that coordinate the binuclear metal center. The metal-ligating
residues are colored gray; the binuclear metal center is colored
magenta, and the coordinating water molecule to the β-metal is
colored red.
Figure 2. Representation of the hexameric arrangement of subunits for
the three-dimensional structure of Atu3266. Face-on (A) and edge-on
(B) views of the hexameric structure of PDB entry 2OGJ.
is at 3.1 Å from the β-metal ion. The two metal ions are 3.3 Å
apart.
Substrate Specificity of Atu3266 and Oant2987. D- and
L-dihydroorotate (56) were the first two compounds to be
tested as substrates for Atu3266, because this enzyme has been
annotated in various databases as a dihydroorotase. However,
no hydrolysis of these compounds could be observed, and thus,
this annotation is not correct. This experiment was followed by
the utilization of a focused library of N-acetyl, N-succinyl, and
N-formyl derivatives of the common D- and L-amino acids and
multiple libraries of L-Xaa-L-Xaa, L-Xaa-D-Xaa, and D-Xaa-L-Xaa
dipeptides. From the more than 1200 compounds tested in the
initial screening, only two compounds were found with very
weak rates of hydrolysis. The values of kcat/Km for N-acetyl-D-
serine (1) and N-acetyl-D-threonine (2) are 4.0 and 2.0 M−1
s−1, respectively. This finding prompted the separate synthesis
of N-acetyl-D-cysteine (3) and N-formyl-D-serine (4), but these
compounds were not detectable substrates for Atu3266.
Methylating the α-carboxylate (compound 5) or phosphorylat-
ing the side chain hydroxyl (compound 6) of the weak
substrate, N-acetyl-D-serine, abolished the catalytic activity. The
protomers) stacked against each other. Each protomer consists
of two domains: a TIM barrel domain and a second domain
consisting of two β-sheets formed by both N- and C-terminal
residues as shown in Figure 3. The structure reveals a distorted
(β/α)8 TIM barrel fold that is similar to those of other
structurally characterized enzymes from the AHS.14 The active
site is dominated by a binuclear divalent metal center that is
reminiscent of the metal centers found in phosphotriesterase40
and dihydroorotase20 as illustrated in Figure 4. The structure of
Atu3266 was determined in the presence of imidazole in the
active site that is coordinated to the β-metal at a distance of 2.2
Å (not shown). The α-metal is coordinated by His-77 and His-
79 at distances of 2.2 and 2.4 Å, respectively. These residues are
positioned at the C-terminal end of β-strand 1. This metal ion
is also coordinated to Asp-291 from β-strand 8 and a
carboxylated Lys-175 (Kcx-175) from β-strand 4. The β-metal
is coordinated to the carboxylated lysine from β-strand 4 and to
His-208 and His-231 from β-strands 5 and 6. A water molecule
233
dx.doi.org/10.1021/bi301483z | Biochemistry 2013, 52, 228−238

Biochemistry
Article
a
Table 2. Kinetic Parameters for Atu3266 and Oant2987
Atu3266
Oant2987
Km (mM)
nc
substrate
kcat (s−1
na
)
Km (mM)
kcat/Km (M−1 s−1
)
kcat (s−1
nc
)
kcat/Km (M−1 s−1
)
1
na
na
na
4
nc
2
na
2
nc
nc
na
nc
3
<0.001
na
<0.001
na
9
12
40
1
25
4
(4.8 0.1) × 102
(1.3 0.2) × 104
(1.6 0.2) × 103
50 10
0.6 0.1
4.2 0.2
0.4 0.1
na
2.1 0.5
1.1 0.2
0.6 0.1
na
(3 1) × 102
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
3
3.1 0.5
0.5 0.1
(4.2 0.8) × 103
(7 1) × 102
40
0.8 0.1
0.5 0.1
0.6 0.1
1.5 0.1
3.1 0.3
1.1 0.2
0.41 0.04
na
10
2
9.5 0.7
58
6
na
na
60
7
6
1
1
(2.1 0.3) × 102
(5.0 0.8) × 102
(1.2 0.4) × 102
(2.0 0.5) × 102
(1.1 0.6) × 102
(1.8 0.6) × 102
50 10
na
na
200
0.81 0.08
0.23 0.02
0.30 0.02
na
1.2 0.1
1.4 0.1
0.71 0.02
na
(8.1 0.8) × 103
(2.2 0.2) × 102
(4.3 0.4) × 102
30
9.5 3.0
2.1 0.5
na
na
na
na
na
30
na
na
nc
nc
nc
na
na
(2.4 0.5) × 102
(2.5 0.2) × 102
(6.6 1.0) × 102
67 22
2.0 0.1
<0.05
2.0 0.5
na
(1.0 0.2) × 103
na
0.10 0.03
0.21 0.01
2.3 0.2
0.4 0.1
0.3 0.1
30 10
1.0 0.1
1.1 0.1
<0.02
na
na
na
na
40
280
34
70
43
30
na
5
(2.8 0.2) × 105
(3.4 0.3) × 104
(1.0 0.1) × 104
(1.1 0.2) × 104
(3.0 0.9) × 103
(3.0 0.3) × 103
(5.0 0.4) × 103
(1.6 0.1) × 104
(2.2 0.2) × 103
(1.0 0.1) × 103
(5.0 0.5) × 103
(1.0 0.1) × 104
(7.0 0.2) × 103
(5.1 0.1) × 103
(1.3 0.1) × 104
(3.5 0.1) × 103
130
6.0 0.2
60
4.1 0.3
1
0.7 0.1
0.21 0.02
3.4 0.5
0.5 0.1
(1.8 0.2) × 105
(3.0 0.3) × 104
(1.7 0.2) × 104
(8.0 0.8) × 103
(3.3 0.3) × 103
(1.0 0.2) × 104
(4.5 0.5) × 103
(5 1) × 104
(7.5 0.9) × 103
(1.0 0.2) × 104
(3.6 0.9) × 103
(1.2 0.3) × 104
(1.0 0.2) × 104
(7.0 0.7) × 103
(2.7 0.2) × 104
(1 0.3) × 103
2
7
5
9
7
1
1
4
4
10
na
na
12
na
na
3
3
60
50
36
6
5
5
18
5
2
1
1
na
8
190 20
150 10
3.0 0.5
4.0 0.5
na
na
30
50
40
10
20
60
80
2
1
5
1
2
6
6
5
1
10
1
2.0 0.2
0.6 0.1
1.4 0.3
11
3
7.0 0.2
0.8 0.2
2.0 0.5
10
50
90
17
1
4
9
3
10
7
1
9
1
1
1
3
0.2
5
28 10
na
a
nc denotes a parameter that was not calculated, mainly because there was no detection of acetic acid by the coupling system, whereas na denotes a
parameter that was not available, in the event that the substrate did not saturate the enzyme and the rates obtained were fit to a linear curve to obtain
V/K for the observed activity.
N-carbamoyl (compound 7) and N-phosphoryl (compound 8)
derivatives of D-serine were not catalytically active. However,
the substitution of the α-amino group with an α-hydroxyl
group, as in 2-acetyl-(R)-glycerate (9), resulted in a 2 order of
magnitude improvement in kcat/Km, relative to that of N-acetyl-
D-serine.
The dramatic increase in the rate of hydrolysis of 2-acetyl-
(R)-glycerate (9) prompted the exploration of modifications to
this substructure. However, 2-phospho-(R)-glycerate (10) and
3-phospho-2-acetyl-(R)-glycerate (11) were not hydrolyzed,
and no activity could be observed with 2-acetylglycerol (12).
Surprisingly, excision of the hydroxymethyl substituent to form
acetyl glycolate (13) increased kcat/Km by a factor of 27, relative
to that of compound 9. Further substitutions to the
substructure of acetyl glycolate (13) did not result in any
further improvements (compounds 14−28) in catalytic activity
except for the addition of a phenyl group. Acetyl-(R)-mandelate
(28) is hydrolyzed with a kcat/Km of 2.8 × 105 M−1 s−1. This is a
factor of 20 better than that of acetyl glycolate (13) and nearly
5 orders of magnitude better than that of the initial hit, N-
acetyl-D-serine (1). No activity was observed with phosphoe-
nolpyruvate (30), the amide of acetyl-(R)-mandelate (31), 2-
phospho-(R)-mandelate (32), acetyl-D-phenylglycine (33), the
methyl ester of acetyl-(R)-mandelate (34), acetyl-(S)-mande-
late (35), or 2-acetoxyisobutyric acid (29). Further modifica-
tions to acetyl-(R)-mandelate did not improve the rate of
hydrolysis (see compounds 36−50), relative to that of acetyl-
(R)-mandelate (28). Other modifications, including p-nitro-
phenyl acetate (51), aspirin (52), and various lactones (53−
55), were not active. The kinetic constants for the catalytically
active substrates are listed in Table 2.
Inhibition by N-Methyl Phosphonate of D-Phenyl-
glycine. The N-methylphosphonate and O-methylphospho-
nate derivatives of D-phenylglycine (57) and R-mandelate (56),
respectively, were synthesized and tested as inhibitors for the
hydrolysis of acetyl-(R)-mandelate (28) by Atu3266 and
Oant2987. The O-methylphosphonate inhibitor of (R)-
mandelate was unstable. However, the N-methylphosphonate
analogue of D-phenylglycine (57) was found to be a
competitive inhibitor versus acetyl-(R)-mandelate. A fit of the
234
dx.doi.org/10.1021/bi301483z | Biochemistry 2013, 52, 228−238

Biochemistry
Article
data to eq 2 gave inhibition constants of 35
Atu3266 and 40 2 μM for Oant2987.
2 μM for
moiety is oriented toward Asp-291, thus facilitating the
enzymatic hydrolysis of the acetylated α-hydroxyl carboxylates.
The docking calculations indicate that the two remaining
fragments of the molecule, the carboxylate and the mostly
nonpolar substituent at the chiral carbon, are necessary for the
correct binding and orientation of the substrates in the active
site. Critical to the binding is the carboxylate, which forms a
hydrogen bond network with the loop formed between β-
strand 7 and α-helix 7 of the AHS domain. These hydrogen
bond interactions include the amide nitrogens of Gly-267 and
Ala-268, and the hydroxyl side chain of Ser-269. In some
instances, an additional hydrogen bond is formed between the
carboxylate of the substrate and the Nε hydrogen of His-293.
The second and more variable portions of the substrates for
Oant2987 and Atu3266 contain mostly nonpolar substitutions
at the chiral α-carbon of the acetylated α-hydroxyl carboxylates.
All molecules identified as substrates possess the R-stereo-
chemistry at C2 of the acetylated α-hydroxyl carboxylates,
including the best substrate, acetyl-(R)-mandelate (28). The
catalytically competent pose of the transition state mimic for
acetyl-(R)-mandelate (28) identified from the HEI AC library
had an energy score of −140.65. Thus, this molecule would
have ranked 221 out of 57672 HEI KEGG molecules, placing it
in the 99.6th percentile. In this pose, the aromatic phenyl ring
moiety of acetyl-(R)-mandelate (28) provides a π-stacking
interaction with the side chain nitrogen atom of Asn-144. The
distance between these two components varies depending on
the substrates docked; however, in the catalytically competent
poses, these molecules adopt conformations where the center
of the ring is 2.8−3.5 Å from the amino acid side chain.41
With the HEI AC virtual library, we have been able to screen
compounds with slightly longer backbones (insertions around
the chiral carbon). Docking results suggest that larger
molecules can be accommodated into the active site, such as
O-acetyl-(S)-carnitine, (S)-3-acetoxy-4-(dimethylamino)-
butanoic acid, acetyltropic acid, and 3-carbamoyloxy-2-phenyl-
propionic acid (Figure S3a of the Supporting Information). Of
these, (S)-3-acetoxy-4-(dimethylamino)butanoic acid (Figure
S3b of the Supporting Information) and racemic acetyltropic
acid (Figure S3c of the Supporting Information) were tested as
possible substrates for Atu3266. However, only acetyltropic
acid is active, albeit with a very poor kcat/Km of 200 M−1 s−1.
This indicates that a small extension between the chiral carbon
and the ester can be tolerated, but the simple acetylated
hydroxyl carboxylate chemotype is preferred.
Computational Docking to the Active Site of
Atu3266. Molecules from the HEI KEGG library consisted
of high-quality poses enriched for carbamoylated and acetylated
amino acids. Thus, acetylated amino acids such as N-acetyl-L-
lysine (rank of 40), N-acetyl-D-methionine (rank of 171), N-
acetyl-D-cysteine (3) (rank of 218), N-acetyl-L-leucine (rank of
313), and N-acetyl-D-phenylalanine (48) (rank of 440) (Figure
S1 of the Supporting Information) and other metabolites such
as N-carbamoyl-L-aspartate, N-acetyl-D-glucosaminate, and
oxalureate were found at the top of the docking list. Further
virtual docking experiments with the HEI AC library suggested
a common binding pose for the compounds that contain the
HEI acetylated α-hydroxyl carboxylate (Figure 5A). A detailed
Figure 5. Modeled structure of docked acetyl-(R)-mandelate into the
active site of Atu3266. On the basis of the docking results of the HEI
library, panel A is a model of the ground state acetylated α-hydroxyl
carboxylate substrates (green sticks) in the active site. The orange
spheres illustrate the possible orientation of the R- and S-enantiomers.
The active site is shown as a surface representation; hydrogen bonds
between the carboxylate of the substrate and loop 267−269 are shown
as dashed lines. In this orientation, the bridging oxygen of the ester
points toward Asp-291 and the carbonyl oxygen stacks over the
binuclear metal center (purple spheres), directing the methyl group of
the acetate into a hydrophobic pocket lined by Ile-87, Leu-140, Cys-
142, His-79, and His-293. (B) The HEI substrate, acetyl-(R)-
mandelate (28) (cyan sticks) (rank of 221), is shown in its best
docked pose. Here the hydrogen on the bridging oxygen is 1.9 Å from
Asp-291. The HEI intermediate is stabilized by Arg-177. The
carboxylate hydrogen bonds with loop 267−269, and the phenyl
group is positioned between the cavity walls, allowing for stacking
interactions with Asn-144.
DISCUSSION
■
Three-Dimensional Structure of Atu3266. The three-
dimensional structure of Atu3266 shows that this enzyme
possesses a binuclear metal center, reminiscent of other
enzymes in the amidohydrolase superfamily such as dihydroor-
otase, phosphotriesterase, and urease.13 The buried α-metal is
coordinated by His-77, His-79, Asp-291, and carboxylated Lys-
175. The more solvent-exposed β-metal is coordinated to His-
208, His-231, carboxylated Lys-175, and a water molecule. The
individual subunits of Atu3266 associate to form a hexameric
oligomer.
examination revealed that the acetylated α-hydroxyl carbox-
ylates direct the ester group toward the metal cluster, where the
carbonyl oxygen is positioned over the α- and β-zinc atoms and
the transition state is stabilized by Arg-177 (Figure 5B). In this
orientation, the methyl group of the acetyl moiety is
surrounded by the hydrophobic pocket lined by Ile-87, Leu-
140, Cys-142, and the plane of the imidazole side chain of His-
79 and His-293. This hydrophobic pocket is large enough that
it can accommodate an ethyl group such as that of (R)-
propionyloxy mandelate (rank of 241) (Figure S2i of the
Supporting Information). The bridging oxygen of the ester
cog3964. There are 24 clusters of orthologous groups in
the AHS, and cog3964 is one of the smallest, with ∼200
sequences (Figure 1). The most common annotation in NCBI
for this cluster of enzymes is dihydroorotase and, in some cases,
adenine deaminase. The basis for this annotation is not clear,
but the experiments reported here demonstrate that this
235
dx.doi.org/10.1021/bi301483z | Biochemistry 2013, 52, 228−238

Biochemistry
Article
annotation is incorrect. When the E value cutoff is set to 10−70
,
Comparisons to Other N-Acetyl D-Amino Acid
Deacetylases. Other enzymes within the amidohydrolase
superfamily have been shown to catalyze the hydrolysis of N-
acetyl D-amino acid derivatives. For example, enzymes within
cog3653 have been shown to be N-acetyl D-amino acid
deacetylases.26,42−44 In cog3653, the binuclear metal center is
not bridged by a carboxylated lysine residue and the α-
carboxylate group of the substrate is ion paired and hydrogen
bonded to conserved arginine and tyrosine residues.26 These
motifs are not conserved in the enzymes of cog3964, where
instead the carboxylate group of the substrate is recognized by
the backbone residues following β-strand 7.
At this point, it is not clear whether the enzymes identified in
this investigation from cog3964 are generic esterases or
whether there is a more specific substrate. Acetyl-(R)-
mandelate has apparently not been identified as a physiological
metabolite in bacteria. In those organisms that have been found
to use mandelate as a carbon source, there is no indication of a
requirement of a deacetylase for the hydrolysis of acetyl-(R)-
mandelate.45
Computational Docking for Identification of Sub-
strate−Enzyme Interactions. Docking of the HEI KEGG
virtual metabolite library to the Atu3266 structure was highly
enriched for carbamoylated or acetylated amino acids (acetyl-
Met, -Cys, -Phe, -Leu, and -Lys) (Figure S1 of the Supporting
Information). An examination of the KEGG database revealed
that the database included only 10 acetylated amino acids (Met,
Cys, Phe, Leu, Lys, His, Trp, Asp, Glu, and Gln). This indicates
a high degree of enrichment in the docking list for the N-acetyl
amino acids as five of them were found in the top 0.76% of the
overall pool of HEI KEGG compounds. As such, the initial
docking experiments with the HEI KEGG library did not
include the two catalytically competent acetylated amino acids,
N-acetyl-D-serine (1) and N-acetyl-D-threonine (2). These were
subsequently prepared and docked in the HEI AC library and
yielded scores that would rank them 321st and 265th,
respectively, out of 57672 HEI KEGG molecules. The poses
of the two amino acids are suboptimal as they either leave the
side chain hydroxyl unsatisfied or allow it to partially interact
with the carboxylate binding loop. However, other N-acetylated
amino acids are found to optimally occupy the active site, as
seen with N-acetyl-D-phenylglycine (33) (Figure S4 of the
Supporting Information).
On the basis of the analysis of the focused HEI AC library
docking results, molecules that are turned over by the enzyme
contain an acetylated α-hydroxyl carboxylate that positions the
acetyl moiety over the zinc atoms, directing the methyl group
into a small hydrophobic pocket. This pocket is large enough to
accommodate ethyl groups as seen by the substrate (R)-
propionyloxy mandelate (Figure S2i of the Supporting
Information) and (R)-2-(propionyloxy) butanoate (22) (Figure
S2b of the Supporting Information). In addition, this
catalytically competent pose orients the carboxylate of the
substrate toward the loop formed by Gly-267, Ala-268, and Ser-
269 (Figure 5A). In this pose, any substitutions or bulky
additions on the substrate carboxylate would change the
electrostatics or cause steric clashes, precluding binding. For
example, molecules with an addition of a methyl group to form
an ester (34) or molecules with a substitution of the
carboxylate with an organic amide (31) demonstrate no
enzymatic activity.
the sequence similarity network for cog3964 segregates into
eight groups with three or more sequences per group and
Atu3266 and Oant2987 belong to Group 6 of cog3964. An
amino acid sequence comparison between those enzymes in
Group 6 of cog3964 and the structurally characterized
dihydroorotases and adenine deaminases (not shown)
demonstrates that the key residues that have been implicated
in substrate binding and recognition in these enzymes are not
present in either Atu3266 or Oant2987. This supports the
conclusion that Atu3266 and Oant2987 do not catalyze the
deamination of adenine or the hydrolysis of dihydroorotate.
Determination of the Substrate Profile. The first
indication of catalytic activity by Atu3266 was observed in a
library of N-acetyl D-amino acids. After each component of the
library had been tested independently, it was observed that N-
acetyl-D-serine (4 M−1 s−1) and N-acetyl-D-threonine (2 M−1
s−1) were the only compounds in the N-acetyl-D-Xaa library to
be hydrolyzed. Screening for hydrolytic activity was also
conducted using a comprehensive library of dipeptides;
however, none of the dipeptide libraries showed detectable
catalytic activity. Subsequently, we focused on the activity
obtained from N-acetyl-D-serine and developed a library of
compounds with various modifications, including changes in
stereochemistry at the central α-chiral carbon, the side chain,
and carboxylate functionalities. None of these changes
generated a better substrate. However, when the amide linkage
was changed to an ester, a 100-fold increase in activity was
observed. Atu3266 hydrolyzes the ester linkage of acetyl-(R)-
glycerate with a rate constant of 400 M−1 s−1. When the
hydroxymethyl group was removed and replaced with a
hydrogen (acetyl glycolate), the rate of hydrolysis increased
by an additional 2 orders of magnitude (1.3 × 104 M−1 s−1).
The highest activity was achieved in the deacetylation of acetyl-
(R)-mandelate with a rate constant of 2.8 × 105 M−1 s−1.
Additional modifications to the acetyl-(R)-mandelate sub-
structure failed to improve the catalytic activity. Substitutions
and modifications to the C1 carboxylate eliminated deacetyla-
tion activity. These alterations included methyl ester and amide
formation, phosphorylation, and reduction to a hydroxy methyl
group. At C2, the most beneficial modification was the change
of the amide nitrogen to oxygen to form an ester. Modification
of the acetyl group to a formyl, succinyl, or carbamoyl group
did not reveal any of the three were a better substituent than
the acetyl group. The stereochemistry at C2 must be R. The
acetyl group could be extended to a propionyl group, but the
rate of hydrolysis is reduced. The presence of the phenyl group
at the side chain position proved to be the best substitution, but
addition of functionality to the aromatic ring decreased the rate
of hydrolysis.
The organisms containing Atu3266 and Oant2987 are devoid
of other genes that are obvious candidates for the metabolism
of mandelate. However, most of the organisms that contain a
protein from cog3964 also have an adjacent gene that is an
apparent homologue to selenocysteine synthase (SelA) that
belongs to cog1921. The genes for two of the proteins
(Atu3263 and Oant2990) were cloned and expressed, and the
proteins were purified to homogeneity (data not shown). The
purified enzymes contained pyridoxal 5′-phosphate, but we
were unable to identify any catalytic activity for these proteins
with a library of amino acids. The relationship between the
proteins from cog1921 and cog3964 is not obvious.
Further docking indicates a preference for the R-enantiomer
of the acetylated α-hydroxyl carboxylates, such as acetyl-(R)-
236
dx.doi.org/10.1021/bi301483z | Biochemistry 2013, 52, 228−238

Biochemistry
Article
Accession Codes
The X-ray coordinates and structure factors for Atu3266 have
been deposited in the Protein Data Bank as entry 2OGJ.
mandelate (28). The R-enantiomers direct the hydrogen of the
chiral carbon toward the floor of the active site cavity and
provide ample space for the substituent (Figure 5B) while still
orienting the bridging oxygen with ideal geometry and distance
from Asp-291 and thus facilitating catalysis. On the basis of the
well-positioned docking poses of the substrates, we can infer
from the HEI docked molecules that the ground state R-
enantiomers could easily enter the active site in the extended
conformation (Figure 5A). In this conformation, the ligands
could easily be subjected to nucleophilic attack by an activated
water molecule. Substrates that contain the R-enantiomer can
easily accommodate bulky substituents, such as the phenyl ring
of acetyl-(R)-mandelate (28), and even larger hydrophobic
groups (37, 43, and 46) (Figure S2d,g,h,k of the Supporting
Information), which fit into the active site between the cavity
walls. Furthermore, compounds with aromatic rings engage in
perpendicular π-stacking interactions with the side chain
nitrogen atom of Asn-144. This interaction is not present in
the small branched or cyclohexane-containing compounds (13,
14, 18, 19, 22, 24−27, and 48−50), which may explain the
difference in affinity parameters. This is illustrated by the
different binding modes and interactions of (R)-2-acetoxybu-
tanoic acid (18) and (R)-2-acetoxy-2-cyclohexylacetate acid
(27) (Figure S2a,j of the Supporting Information). Only a few
ionic interactions are observed for charged substituents on the
acetylated α-hydroxyl carboxylates; these charged interactions
are with Asn-144 (Figure S2f,l of the Supporting Information)
and Lys-236 (Figure S2c,e of the Supporting Information) for
some of the substrates.
Strategy for Functional Annotation. The functional
annotation of Atu3266 as an enzyme that is able to catalyze the
hydrolysis of an ester linkage from substituted α-hydroxyl
carboxylates with R-stereochemistry began with the identi-
fication of the weak substrate, N-acetyl-D-serine. The
combination of focused chemical libraries and computational
docking to the three-dimensional crystal structure resulted in
the identification of acetyl-(R)-mandelate as a substrate that
was 5 orders of magnitude better than the initial hit. In addition
to Oant2987, we predict that seven other enzymes from Group
6 of cog3964 will have the same substrate profile as Atu3266.
These proteins are listed in Table S1 of the Supporting
Information. The identification of function for Group 6 of
cog3964 will facilitate the functional annotation for the
additional groups of proteins in this COG. In preliminary
experiments, a protein from Group 2 of cog3964 has been
purified and crystallized (EF0837 from Enterococcus faecalis, gi|
29375425, PDB entry 2ICS). The sequence of this protein is
35% identical with that of Atu3266, and EF0837 catalyzes the
hydrolysis of acetyl-(R)-mandelate with a kcat/Km value that is 3
orders of magnitude smaller than the value observed for
Atu3266 and Oant2987. These results suggest that the
substrate profile for this enzyme will be different from that of
Atu3266. Efforts to identify the substrate for the enzymes of
Group 2 are currently underway.
AUTHOR INFORMATION
■
Corresponding Author
*S.S.: telephone, (631) 344-3187; fax, (631) 344-3407; e-mail,
swami@bnl.gov. B.K.S.: telephone, (415) 514-4126; fax, (415)
514-4260; e-mail, shoichet@cgl.ucsf.edu. F.M.R.: telephone,
(979) 845-3373; fax, (979) 845-9452; e-mail, raushel@chem.
tamu.edu.
Funding
This work was supported in part by the Robert A. Welch
Foundation (A-840) and the National Institutes of Health (GM
71790 and GM 74945).
Notes
The authors declare no competing financial interest.
ACKNOWLEDGMENTS
■
We thank Dr. Peter Kolb and Dr. Ryan G. Coleman for
technical discussions relating to computational techniques.
REFERENCES
■
(1) Roberts, R. J. (2004) Identifying protein function: A call for
community action. PLoS Biol. 2, E42.
(2) Rost, B., Liu, J., Nair, R., Wrezeszczynski, K. O., and Ofran, Y.
(2003) Automatic prediction of protein function. Cell. Mol. Life Sci. 60,
2637−2650.
(3) Freidberg, I. (2006) Automated protein function prediction: The
genomic challenge. Briefings Bioinf. 7, 225−242.
(4) Glasner, M., Gerlt, J. A., and Babbitt, P. C. (2006) Evolution of
enzyme superfamilies. Curr. Opin. Chem. Biol. 10, 492−497.
(5) Newton, R. J., Griffin, L. E., Bowles, K. M., Meile, C., Gifford, S.,
Givens, C. E., Howard, E. C., King, E., Oakley, C. A., Risch, C. R.,
Rinta-Kanto, J., Sharma, S., Sun, S., Varaljay, V., Vila-Costa, M.,
Westrich, J. R., and Moran, M. A. (2010) Genome characteristics of a
generalist marine bacterial lineage. ISME J. 4, 784−798.
(6) Sweetlove, L. J., and Fernie, A. R. (2005) Regulation of metabolic
networks: Understanding metabolic complexity in the systems biology
era. New Phytol. 168, 9−24.
(7) Pellegrini, M., Marcotte, E. M., Thompson, M. J., Eisenberg, D.,
and Yeates, T. O. (1999) Assigning protein functions by comparative
genome analysis: Protein phylogenetic profiles. Proc. Natl. Acad. Sci.
U.S.A. 96, 4285−4288.
(8) Kim, S.-H. (2000) Structural genomics of microbes: An objective.
Curr. Opin. Biol. 10, 380−383.
(9) Hermann, J. C., Ghanem, E., Li, Y., and Raushel, F. M. (2006)
Predicting substrates by docking high-energy intermediates to enzyme
structures. J. Am. Chem. Soc. 128, 15882−15891.
(10) Kolb, P., Ferreira, R. S., Irwin, J. J., and Shoichet, B. K. (2009)
Docking and chemoinformatic screens for new ligands and targets.
Curr. Opin. Biochem. 20, 429−436.
(11) Senisterra, G. A., Markin, E., Yamazaki, K., Hui, R., Vedadi, M.,
and Awrey, D. E. (2006) Screening for ligands using a generic and
high-throughput light-scattering-based assay. J. Biomol. Screening 11,
940−948.
(12) Galperin, M. Y., and Koonin, E. V. (2010) From complete
genome sequence to ‘complete’ understanding? Trends Biotechnol. 28,
398−406.
(13) Holm, L., and Sander, C. (1997) An evolutionary treasure:
Unification of a broad set of amidohydrolases related to urease.
Proteins 28, 72−82.
(14) Seibert, C. M., and Raushel, F. M. (2005) Structural and
catalytic diversity within the amidohydrolase superfamily. Biochemistry
44, 6383−6391.
ASSOCIATED CONTENT
■
S
* Supporting Information
Synthetic procedures for the preparation of the compounds
used in this investigation and additional figures and tables. This
material is available free of charge via the Internet at http://
pubs.acs.org.
237
dx.doi.org/10.1021/bi301483z | Biochemistry 2013, 52, 228−238

Biochemistry
Article
(15) Li, T., Iwaki, H., Fu, R., Hasegawa, Y., Zhang, H., and Liu, A.
(2006) α-Amino-β-carboxymuconic-ε-semialdehyde decarboxylase
(ACMSD) is a new member of the amidohydrolase superfamily.
Biochemistry 45, 6628−6634.
(16) Hara, H., Masai, E., Katayama, Y., and Fukuda, M. (2000) The
4-oxalomesaconate hydratase gene, involved in the protocatechuate
4,5-cleavage pathway, is essential to vanillate and syringate degradation
in Sphingomonas paucimolbilis SYK-6-hydration reaction. J. Bacteriol.
182, 6950−6957.
(17) Williams, L., Nguyen, T., Si, Y., Porter, T., and Raushel, F. M.
(2006) Uronate isomerase: A nonhydrolytic member of the
amidohydrolase superfamily with an ambivalent requirement for
divalent metal ion. Biochemistry 45, 7453−7462.
(18) Tatusov, R. L., Galperin, M. Y., Natale, D. A., and Koonin, E. V.
(2000) The COG database: A tool for genome-scale analysis of
protein functions and evolution. Nucleic Acids Res. 28, 33−36.
(19) Tatusov, R. L., Natale, D. A., Garkavtsev, I. V., Tatusova, T. A.,
Shankavaram, U. T., Rao, B. S., Kiryutin, B., Galperin, M. Y., Federova,
N. D., and Koonin, E. V. (2001) The COG database: New
developments in phylogenetic classification of proteins from complete
genomes. Nucleic Acids Res. 29, 22−28.
(20) Thoden, J. B., Phillips, G. N., Neal, T. M., Raushel, F. M., and
Holden, H. M. (2001) Molecular structure of dihydroorotase: A
paradigm for catalysis through the use of a binuclear metal center.
Biochemistry 40, 6989−6997.
(21) Porter, T. N., Li, Y., and Raushel, F. M. (2004) Mechanism of
the dihydroorotase reaction. Biochemistry 43, 16285−16292.
(22) Kamat, S. S., Bagaria, A., Kumaran, D., Holmes-Hampton, G. P.,
Fan, H., Sali, A., Sauder, M. J., Burley, S. K., Lindahl, P. A.,
Swaminathan, S., and Raushel., F. M. (2011) Catalytic mechanism and
three-dimensional structure of adenine deaminase. Biochemistry 50,
1917−1927.
(23) Goble, A. M., Zhang, Z., Sauder, M. J., Burley, S. K.,
Swaminathan, S., and Raushel, F. M. (2011) Pa0148 from Pseudomonas
aeruginosa catalyzes the deamination of adenine. Biochemistry 50,
6589−6597.
(24) Pace, C. N., Vajdos, F., Fee, L., Grimsley, G., and Gray, T.
(1995) How to measure and predict the molar absorption coefficient
for a protein. Protein Sci. 4, 2411−2423.
(34) Brunger, A. T., Adams, P. D., Clore, G. M., Delano, W. L., Gros,
P., Grosse-Kunstleve, R. W., Jiang, F. S., Kuszwewski, J., Nilges, M.,
Pannu, N. S., Read, R. J., Rice, L. M., Somonsom, T., and Warren, G.
L. (1998) Crystallography & NMR system: A new software suite for
macromolecular structure determination. Acta Crystallogr. D54, 905−
921.
(35) Jones, T. A., Zou, J.-Y., Cowan, S. W., and Kjeldgaard, M.
(1991) Improved methods in building protein models in electron
density map and the location of errors in these models. Acta
Crystallogr. A47, 110−119.
(36) Xiang, D. F., Kolg, B., Fedorov, A. A., Meier, M. M., Fedorov, L.
V., Nguyen, T. T., Sterner, R., Almo, S. C., Shoichet, B. K., and
Raushel, F. M. (2009) Functional annotation and three-dimensional
structure of Dr0930 from Deinococcus radiodurans, a close relative of
phosphotriesterase in the amidohydrolase superfamily. Biochemistry 48,
2237−2247.
(37) Irwin, J. J., Raushel, F. M., and Shoichet, B. K. (2005) Virtual
screening against metalloenzymes for inhibitors and substrates.
Biochemistry 44, 12316−12328.
(38) Irwin, J. J., Shoichet, B. K., Mysinger, M. M., Huang, N., Colizzi,
F., Wassam, P., and Cao, Y. (2009) Automated docking screens: A
feasibility study. J. Med. Chem. 52, 5712−5720.
(39) Wlodek, S., Skillman, A. G., and Nicholls, A. (2010) Ligand
entropy in gas-phase, upon salvation and protein complexation. Fast
estimation with quasi-Newton Hessian. J. Chem. Theory Comput. 6,
2140−2152.
(40) Benning, M. M., Kuo, J. M., Raushel, F. M., and Holden, H. M.
(1995) Three-dimensional structure of the binuclear center of
phosphotriesterase. Biochemistry 34, 7973−7978.
(41) Ma, J. C., and Dougherty, D. A. (1997) The cation-π interaction.
Chem. Rev. 97, 1303−1324.
(42) Liaw, S.-H., Chen, S.-J., Ko, T.-P., Hsu, C.-S., Chen, C.-J., Wang,
A. H.-J., and Tsai, Y.-C. (2003) Crystal structure of D-aminoacylase
from Alcaligenes faecalis DA1. J. Biol. Chem. 278, 4957−4962.
(43) Lai, W.-L., Chou, L.-Y., Ting, C.-Y., Kirby, R., Tsai, Y.-C., Wang,
A. H.-J., and Liaw, S.-H. (2004) The functional role of the binuclear
metal center in D-aminoacylase: One-metal activation and second-
metal attenuation. J. Biol. Chem. 279, 13962−13967.
(44) Liu, J., Asano, Y., Ikoma, K., Yamashita, S., Hirose, Y.,
Shimoyama, T., Takahashi, S., Nakayama, T., and Nishino, T. (2012)
Purification, characterization and primary structure of a novel N-acyl-
D-amino acid amidohydrolase from Microbacterium natoriense
TNJL143-2. J. Biosci. Bioeng. 114, 391−397.
(25) Xiang, D., Patskovsky, Y., Chengfu, X., Meyer, A. J., Sauder, J.
M., Burley, S. K., Almo, S. C., and Raushel, F. M. (2009) Functional
identification of incorrectly annotated prolidases from the amidohy-
drolase superfamily of enzymes. Biochemistry 48, 3730−3742.
(26) Cummings, J. A., Fedorov, A. A., Xu, C., Brown, S., Fedorov, E.,
Babbitt, P. C., Almo, S. C., and Raushel, F. M. (2009) Annotating
enzymes of uncertain function: The deacylation of D-amino acids by
members of the amidohydrolase superfamily. Biochemistry 48, 6469−
6481.
(45) Ju, X., Yu, H.-L., Pan, J., Wei, D.-Z., and Xu, J.-H. (2010)
Bioproduction of chiral mandelate by enantioselective deacylation of
α-acetoxyphenylacetic acid using whole cells of newly isolated
Pseudomonas sp. ECU1011. Appl. Microbiol. Biotechnol. 86, 83−91.
(27) Doi, E., Shibata, D., and Matoba, T. (1981) Modified
colorimetric ninhydrin methods for peptidase assay. Anal. Biochem.
118, 173−184.
(28) Chapman, E., and Wong, C. H. (2002) A pH sensitive
colorimetric assay for the high-throughput screening of enzyme
inhibitors and substrates: A case study using kinases. Bioorg. Med.
Chem. 10, 551−555.
(29) Otwinowski, Z., and Minor, W. (1997) Processing of X-ray
diffraction data collected in oscillation mode. Methods Enzymol. 276,
307−326.
(30) Schneider, T. R., and Sheldrick, G. M. (2002) Substructure
solution with SHELXD. Acta Crystallogr. D58, 1772−1779.
(31) Collaborative Computational Project, Number 4 (1994) CCP4
Suite: Programs for protein crystallography. Acta Crystallogr. D50,
760−763.
(32) De-La-Fortelle, E., and Bricogne, G. (1997) Maximum-
likelihood heavy atom parameter refinement in the MIR and MAD
methods. Methods Enzymol. 276, 472−493.
(33) Perrakis, A., Morris, R., and Lamzin, V. S. (1999) Automated
protein model building combined with iterative structure refinement.
Nat. Struct. Biol. 6, 458−463.
238
dx.doi.org/10.1021/bi301483z | Biochemistry 2013, 52, 228−238