Use of structure-based drug design approaches to obtain novel anthranilic acid acyl carrier protein synthase inhibitors

Article abstract of DOI:10.1021/jm050523n

Acyl carrier protein synthase (AcpS) catalyzes the transfer of the 4′-phosphopantetheinyl group from the coenzyme A to a serine residue in acyl carrier protein (ACP), thereby activating ACP, an important step in cell wall biosynthesis. The structure-based

Full text of DOI:10.1021/jm050523n

7960
J. Med. Chem. 2005, 48, 7960-7969
Use of Structure-Based Drug Design Approaches to Obtain Novel Anthranilic
Acid Acyl Carrier Protein Synthase Inhibitors
Diane Joseph-McCarthy,*,† Kevin Parris,† Adrian Huang,† Amedeo Failli,‡ Dominick Quagliato,‡
Elizabeth Glasfeld Dushin,§ Elena Novikova,§ Elena Severina,§ Margareta Tuckman,§ Peter J. Petersen,§
Charles Dean,§ Christian C. Fritz,| Tova Meshulam,| Maureen DeCenzo,| Larry Dick,| Iain J. McFadyen,†
William S. Somers,† Frank Lovering,† and Adam M. Gilbert§
Wyeth Research, 200 CambridgePark Drive, Cambridge, Massachusetts 01240, Wyeth Research, CN-8000,
Princeton, New Jersey 08852, Wyeth Research, 401 North Middleton Road, Pearl River, New York 10945, and
Millennium Pharmaceuticals, Inc., 75 Sidney Street, Cambridge, Massachusetts 02139
Received June 3, 2005
Acyl carrier protein synthase (AcpS) catalyzes the transfer of the 4′-phosphopantetheinyl group
from the coenzyme A to a serine residue in acyl carrier protein (ACP), thereby activating ACP,
an important step in cell wall biosynthesis. The structure-based design of novel anthranilic
acid inhibitors of AcpS, a potential antibacterial target, is presented. An initial high-throughput
screening lead and numerous analogues were modeled into the available AcpS X-ray structure,
opportunities for synthetic modification were identified, and an iterative process of synthetic
modification, X-ray complex structure determination with AcpS, biological testing, and further
modeling ultimately led to potent inhibitors of the enzyme. Four X-ray complex structures of
representative anthranilic acid ligands bound to AcpS are described in detail.
Introduction
This paper describes the structure-based design of
novel anthranilic acid AcpS inhibitors. We have previ-
ously reported on the combinatorial synthesis and
testing of a related series of inhibitors.8 The identifica-
tion and optimization of a high-throughput screening
(HTS) lead is presented. The optimization process
involved modeling of the initial lead into an available
AcpS X-ray structure, the design and synthesis of
proposed improved compounds, biological testing, and
further structure determination, ultimately leading to
potent inhibitors of the enzyme. The determination of
four X-ray complex structures of representative inhibi-
tors bound to AcpS is detailed, and an analysis of the
structures is discussed.
Acyl carrier proteins (ACP) play a vital role in the
biosynthesis of fatty acids and other biomolecules in all
organisms through the mediation of acyl group trans-
fers.1 Typically consisting of 80-100 residues, ACP are
converted from the inactive apo form to the active holo
form by posttranslation transfer of a 4′-phosphopan-
tetheinyl (4′-PP) moiety from coenzyme A (CoA) to a
conserved serine residue in the ACP (Ser36 in Escheri-
chia coli ACP).2 See Figure 1. This reaction is mediated
by holo-(acyl carrier protein) synthases (AcpS), which
typically contain ∼120 residues. High-resolution crystal
structures of AcpS, AcpS + CoA, and AcpS + ACP from
Bacillus subtilis3 as well as AcpS and AcpS complexed
with 3′5′-adenosine diphosphate (ADP) from Strepto-
coccus pneumoniae4 have shown that AcpS functions as
a trimer of three identical monomers, with three identi-
cal active sites at each of the monomer-monomer
interfaces.
ACP occurs in two main classes, based on the archi-
tecture of the system.5 Type I, or associated, ACP occurs
as an integrated domain in a multifunctional protein
and is found in mammals, fungi, and certain mycobac-
teria. Type II, or dissociated, ACP occurs as an inde-
pendent protein in a nonaggregated multienzyme sys-
tem and is found in plants and most bacteria. It has
been shown that the activation of ACP by AcpS is an
essential step in the bacterial biosynthesis of lipopolysac-
charides and membrane lipids through genetic studies
of E. coli6 and Streptococcus pneumoniae.7 Thus, AcpS
has been considered a promising antibacterial target.
Results and Discussion
Biological Data. An HTS for inhibitors of B. subtilis
AcpS identified a number of anthranilic acids as leads,
with 0 and 4 being representative of them. Both of these
compounds had a relatively weak IC50 against B. subtilis
AcpS (12.7 and 32.3 µM, respectively; see Table 1 and
Figure 2). 4 was taken as the initial lead because of its
relative novelty and perceived greater potential for
optimization.
4 subsequently also demonstrated measurable mini-
mum inhibitory concentrations (MICs) (32-64 µg/mL)
against a number of Gram-positive organisms (Table 2).
Two different assays were implemented to determine if
the observed antibacterial activity seen was linked to
the inhibition of AcpS. The first, a membrane-damage
assay, addressed whether antibacterial activity might
instead be attributed to the disruption of the membrane
rather than the inhibition of AcpS. 4 was shown to cause
rapid cell death of B. subtilis at its MIC (90% of cells
appeared dead after a 5-min incubation). Because we
presumed that specific inhibition of AcpS would not be
cidal in such a short time frame, the MIC of 4 most
* Author to whom correspondence should be addressed. E-mail:
djoseph@wyeth.com. Phone: 617-665-5627. Fax: 617-665-5682.
† Wyeth Research, Massachusetts.
‡ Wyeth Research, New Jersey.
§ Wyeth Research, New York.
| Millennium Pharmaceuticals, Inc.
10.1021/jm050523n CCC: $30.25 © 2005 American Chemical Society
Published on Web 11/16/2005

Design of Novel Anthranilic Acid AcpS Inhibitors
Journal of Medicinal Chemistry, 2005, Vol. 48, No. 25 7961
Figure 1. Reaction catalyzed by AcpS.
Table 1. Summary of IC50 Data for Anthranilic Acid
Derivatives
Figure 2. Concentration-response curves for select anthra-
nilic acid derivative inhibitors of AcpS. Assays were performed
as described in Methods. Data plotted are an average of two
replicates.
cessed, resulting in a mixed population of ACP.9 An
intense band approximately one-third down the gel is
consistent with the homodimer of holo-ACP, which
forms during sample preparation, identified in reference
13. In untreated cells, almost all of the ACP pool has
been phosphopantetheinylated and exists either as holo-
ACP or as holo-ACP with various extensions off of the
phosphopantetheine arm that arise during lipid me-
tabolism (lane 1, Figure 3). Under conditions of AcpS
underexpression (J. Hixon, unpublished data), apo-ACP
accumulates with concomitant reduction in the density
of holo-ACP. After incubation of B. subtilis PY79 with
as little as 4 µg/mL 16, there was a clear accumulation
of apo-ACP (Figure 3), and at 8 µg/mL 16, there is nearly
complete inhibition of AcpS. Given that the accumula-
tion of apo-ACP has previously been shown to be toxic
to cells,13 the apparent reduction in the total ACP pool
(unmodified and modified), seen in Figure 3, lane 5,
could possibly be the result of feedback control on ACP
translation. Slight membrane damage could also lead
to reduced visible levels of ACP.
Optimization of the Lead Series through Molec-
ular Modeling. Modeling suggested modification to the
initial HTS anthranilic acid lead, 4. More than 50
analogues of this lead were modeled into the site. The
results of the modeling are summarized in Figure 4.
Specifically, the models indicated that shifting the Cl
from the 5-position to the 4-position of the anthranilic
head piece should improve potency by better filling a
hydrophobic pocket, that adding a small donor group
at the 5-position would be expected to be favorable, and
that the Cl on the benzyl group interacts favorably with
Arg53 (largely through polar/electrostatic interactions)
(Figure 5). The modeling also showed that removing the
double bond was expected to improve the binding
affinity and that the R enantiomer was expected to bind
preferentially (see Figure 5A for a model of 10). Fur-
a Activities of compounds in the HTRF assay.
likely arises from a disruption of the cell membrane,
an undesirable property in the development of target-
specific antibacterial agents.
The structure-based approach described below led to
the design and synthesis of a number of derivatives of
4 with the aim of improving its potency and properties.
One of the more potent analogues, 16, had a 30-fold
lower IC50 in the AcpS biochemical assay (1.4 µM or 0.55
µg/mL) and did not display any of the membrane-
damaging liabilities of the initial lead, 4. 16 had MICs
equivalent to those of 4 (Table 2), but relatively little
membrane damage was detected after a 5-min incuba-
tion at its MIC (64 µg/mL) with B. subtilis PY79
(approximately 20% dead after 5 min. incubation com-
pared to 5% dead in an untreated culture).
For a direct detection of inhibition of AcpS within B.
subtilis, the relative populations of apo- and holo-ACP
were examined by Western blotting mid-logarithmic
growth phase cultures. The apo and holo forms of ACP
can be distinguished on a polyacrylamide gel containing
moderate levels of urea (Figure 3). In Figure 3, two
bands can be seen for the monomers of apo- and holo-
ACP. This is likely due to the fact that the N-terminal
methionine of B. subtilis ACP is not completely pro-

7962 Journal of Medicinal Chemistry, 2005, Vol. 48, No. 25
Table 2. Enzyme Activity and Antimicrobial Susceptibilitya
Joseph-McCarthy et al.
MIC (µg/mL)
E. faecalis VRE
cmpd
AcpS IC50 (µg/mL)
B. subtilis 168
S. aureus MRSA
S. pneumoniae penR
E. coli
C. albicans
4
16
1.3
0.55
32
64
32
64
64
128
32
64
>128
>128
>128
>128
a Activities of compounds in the HTRF and MIC assays. Strains used are Bacillus subtilis 168, methicillin-resistant Staphylococcus
aureus, vancomycin-resistant Enterococcus faecalis, penicillin-resistant Streptococcus pneumoniae, Escherichia coli, and Candida albicans.
Figure 3. Western blot of urea-PAGE with anti-ACP anti-
bodies. B. subtilis PY79 cultures were treated with 16,
fractionated on by urea-PAGE, and Western blotted with anti-
ACP antibodies. Lane 1 is the no compound control; lanes 2-5
are 1, 2, 4, and 8 µg/mL, respectively, of 16.
Figure 4. Overview of the interactions and opportunities for
synthesis determined through molecular modeling.
Figure 5. Models for proposed compounds bound in the AcpS
active site. (A) Model for 10. (B) Two binding modes for a close
thermore, replacing the methyl group with a methyl
thiol group or disulfide was predicted to have a small
favorable effect on potency and maintain the overall
binding mode. Modeling of a close analogue of 16 (shown
in Figure 5B) showed that only the R enantiomer docked
and that two alternate conformations for the chloro
benzyl group were likely. There was interest in poten-
tially introducing a disulfide at this position to allow
for dynamic combinatorial synthesis of a large number
of compounds to explore the unoccupied region of the
binding site;10,11 this approach, however, was not pur-
sued further. Modeling also predicted that cyclizing to
connect the methyl group to the benzyl ring (as in 7)
would be favorable and maintain the overall binding
mode. Synthesis and testing of 7, 10, and 16 confirmed
the predictions. Subsequently, the X-ray structure of 16
analogue of 16 with essentially equivalent energy scores. (C)
Proposed analogue with even greater predicted binding affin-
ity.
bound to AcpS confirmed the predicted binding modes;
in the structure, described below in detail, only the R
enantiomer of the compound was bound, and as pre-
dicted, two alternate conformations for the chloro benzyl
group were observed.
X-ray Crystal Structures of Ligand-AcpS Com-
plexes. The structures of AcpS in complex with the
small-molecule inhibitors 0, 4, 7, and 16 have been
determined (Table 3). As shown in prior work,3,4 AcpS
exists in a trimeric state with three active sites; each
active site is at a dimer interface (see Figure 6 for an
illustration of the 4-AcpS structure). In general, the

Design of Novel Anthranilic Acid AcpS Inhibitors
Journal of Medicinal Chemistry, 2005, Vol. 48, No. 25 7963
Table 3. Refinement Statistics
AcpS + Cpd 0
AcpS + 4
AcpS + 7
AcpS + 16
space group
R32
P212121
P3
P3
AcpS molecules per asymmetric unit
1
3
6
6
unit cell
a (Å)
4.297
84.297
140.211
62.553
83.661
85.979
83.475
83.475
112.105
83.674
83.674
106.719
b (Å)
c (Å)
R, â, γ (deg)
R ) â ) 90 γ ) 120 R ) â ) γ ) 90 R ) â ) 90 γ ) 120 R ) â ) 90 γ ) 120
1.9
18.8
22.9
0.027
maximum resolution (Å)
2.3
2.2
2.0
Rworka (%)
23.4
21.0
21.2
Rfreeb (%)
26.8
22.9
24.7
rms deviations from ideal geometry for bonds (Å)
0.007
1.23
0.011
1.62
0.007
1.27
rms deviations from ideal geometry for angles (deg) 1.81
AcpS residues observed
water molecules
other ions
1-119
115
1-119
87
Cl-, glycerol
1-119
330
citrate
1-119
370
citrate
b
a Rwork ) ∑||Fobs| - |Fcalc|/∑|Fobs|. Rfree is equivalent to Rwork but calculated for a randomly chosen 5% of reflections omitted from the
refinement process.
Figure 7. Superposition of the 4-AcpS structure and the
CoA-ACPS structure. 4 and CoA are shown colored by
element with carbons in cyan for 4 and magenta for CoA. Only
protein backbone atoms were superimposed.
between Lys86, Pro87 of the AcpS molecule and the side
Figure 6. Ribbon diagram of the overall structure of the
chain of Lys62 of a symmetry-related AcpS molecule.
4-AcpS complex, showing the catalytic trimer and the position
The carboxylic acid is within hydrogen-bond distance
to the Nú of Lys86, the Oγ of Ser102, and the main-
chain N-H of Ile103. Aside from a hydrogen bond to a
water molecule, the only interaction with the furan ring
is a van der Waals (VDW) interaction with Pro87. The
final model consists of molecule 0, a continuous chain
involving protein residues 1-119, and 115 water mol-
ecules.
4-AcpS Complex Structure. The structure of AcpS
in complex with 4 contains three molecules of AcpS in
the asymmetric unit arranged to form the catalytic
trimer. In addition to the three AcpS molecules, the final
model contains three molecules of 4, a glycerol molecule,
a chloride ion, and 87 waters. NMR analysis demon-
strated that the sample of 4 used for the crystallization
experiment was a mixture of cis and trans isomers. Most
interestingly, both isomers were found bound to AcpS
in this crystal structure. The cis isomer is found in two
of the three active sites in the anthranilate binding site.
This isomer, like 0, is held in place by VDW interactions
with Lys86, Pro87, and Lys62. Additionally, the same
interactions exist between the carboxylic acid of the
inhibitor and the protein: hydrogen bonds to the Nú of
of the bound ligand (colored by element with carbons in cyan)
at each dimer interface.
anthranilate compounds in the structures presented
bind in the AcpS active site where the adenine, ribose,
and 3′-phosphate of CoA reside. Hereafter, this site will
be referred to as the anthranilate binding site (Figure
7). The inhibitors adopt a U-shape that allows them to
sandwich around Pro87 (Figure 8). When referring to
residues in the structures, residues in normal text will
be from one molecule, whereas those in bold text will
be from a second (or symmetry-related) molecule.
0-AcpS Complex Structure. In the structure of 0
(a related HTS hit) in complex with AcpS, the asym-
metric unit contains only one molecule of AcpS. This
molecule of AcpS is positioned on a crystallographic axis
such that a R32 symmetry operator generates the
catalytic trimer. In this structure (Figure 8), the sul-
fonamide nitrogen is positioned to be within hydrogen-
bonding distance to the main-chain carbonyl of Asn84.
Additionally, the oxygen atoms of this sulfonamide are
adjacent to the main-chain amides of Thr66 and Gly67.
The phenyl ring of the anthranilate is sandwiched

7964 Journal of Medicinal Chemistry, 2005, Vol. 48, No. 25
Joseph-McCarthy et al.
Figure 8. Structure of the 4-AcpS complex is shown with unbiased 2Fo - Fc electron density contoured at 1σ around the ligand.
(A) Key protein interactions with the inhibitor are highlighted in the snapshot of the active site. (B) Similar electron density for
the ligands in the other three complex structures (7, 16, and 0).
Lys86, the Oγ of Ser102, and the main-chain N-H of
Ile103. Adjacent to only one of the molecules of cis-4 is
a molecule of glycerol that came from the cryopreser-
vation process prior to data collection. The 1-hydroxyl
is positioned to be within hydrogen-bonding distance of
the amide carbonyl of the cis-4 molecule. The 3-hydroxyl
is within hydrogen-bonding distance of Gly58 and
Ile68. The density was unambiguous, but to be certain,
we conducted an experiment where the cryopreservation
solvent was ethylene glycol rather than glycerol. Ex-
amination of the electron density from this experiment
showed that the electron density in the area in question
had changed and was now consistent with a water
molecule. The third active site has no compound in the
anthranilate binding pocket but instead has a molecule
of the trans isomer. This molecule is found in the active
site at the location where the â-mercaptoethylamine
unit of CoA is bound. Additionally, when this structure
is superimposed with the structure of the AcpS/ACP
complex, this trans molecule overlaps with the positions
of the Met44, Val39, and Leu37 side chains of ACP.
These ACP residues extend into pockets on AcpS when
ACP binds. Specifically, the chloro atom from the
anthranilic acid moiety of the trans molecule superim-
poses with the sulfur of Met44 (ACP), whereas the
Leu37 (ACP) side chain coincides with the location of
the phenyl ring of the chlorophenyl moiety. Despite
these alignments, this binding site is likely to be an
artifact of crystal packing as the trans isomer is held
into this position by a symmetry-related molecule of
AcpS. Supporting the hypothesis of an artifactual bind-
ing site is the fact that in the two active sites where
the cis isomer binds, the trans binding location is
available but is not occupied.
7-AcpS Complex Structure. The structure of AcpS
in complex with 7 contains six molecules of AcpS in the
asymmetric unit. Each of these molecules is placed on
a crystallographic three-fold axis, and thus, the catalytic
trimer for each is generated by symmetry. In addition
to the six AcpS molecules, the final model contains six
molecules of 7, six citric acid molecules, and 330 waters.
Each of the 7 molecules is bound in the anthranilate
binding pocket with the same contacts as detailed above.
In the active site of each AcpS there is also a molecule
of citric acid, a component of the crystallization buffer.
The citric acid moiety is found ∼3.5 Å from 7 and has
potential hydrogen bonds to Asp 8, Lys 62, and Thr 104.
16-AcpS Complex Structure. The structure of
AcpS in complex with 16 contains six molecules of AcpS
in the asymmetric unit. Like the costructure of AcpS
and 7, each of these molecules is located on a crystal-
lographic three-fold axis; thus, the catalytic trimers for
each are generated by symmetry. In addition to the six
AcpS molecules, the final model contains six molecules
of 16, six molecules of citric acid, and 370 waters. Each
molecule of 16 is bound in the anthranilate binding
pocket with the same contacts as detailed above for 4.
Although it is not involved in any substantial contacts
with the protein, the thio-methyl group is positioned in
the active site so that it coincides with 5′ position of the
ribose moiety of CoA. Only the R enantiomer of 16 (a
racemic mixture) was observed in the structure. Each
active site also indicates that there are electron densi-
ties for two alternate conformations for the chloro benzyl

Design of Novel Anthranilic Acid AcpS Inhibitors
Journal of Medicinal Chemistry, 2005, Vol. 48, No. 25 7965
Scheme 1a
a (a) 2-Amino-4-chlorobenzoic acid methyl ester, diisopropylcar-
bodiimide, cat. DMAP, CH2Cl2, 23 °C; (b) 4-chlorobenzaldehyde,
DBU, LiCl, THF, 23 °C; (c) 1 N LiOH, THF, H2O, 23 °C; (d) 2 N
HCl, 23 °C.
Figure 9. Superposition of the 16-AcpS structure and the
7-AcpS structure. Only protein backbone atoms (shown in a
ribbon representation for each structure) were superimposed.
7 (thick lines) and the corresponding citrate (thin lines) found
in the active site are shown colored by element with carbons
in cyan. The predicted lower energy mode of the R enantiomer
of 16 (thick lines) and the corresponding citrate (thin lines)
found near the active site are shown with carbons in magenta;
the other observed binding mode for the R enantiomer of 16
(thick lines) is shown with carbons in yellow.
Scheme 2a
group of 16 (Figure 8). Both binding modes for the chloro
benzyl group were expected on the basis of the molecular
modeling results. A carbon-carbon bond rotation results
in an alternate conformation that gives a more linear
molecule rather than one that is U-shaped. This linear
rotamer extends the chloro benzyl group into a pocket
formed by His105, Gly52, and Glu48.
a (a) (COCl)2, cat. DMF, PhMe, 23 °C; (b) 2-amino-5-chloroben-
zoic acid, toluene, 23 °C.
molecules to further improve their potencies while still
maintaining druglike properties.
Chemistry. 4. The 1H NMR spectrum of the histori-
cal sample of 4 showed a mixture of cis and trans
double-bond isomers. Thus, the presence of both isomers
in the 4-AcpS complex structure is most likely due to
this mixture. Pure cis-4 was prepared as outlined in
Scheme 1.
Comparison of the Various Complex Structures.
The binding modes for the anthranilic acid portion of
each inhibitor in the AcpS protein structure are es-
sentially identical. As expected from the modeling,
removing the double bond from 4 improves the binding
affinity by allowing the compound to adopt a conforma-
tion with a more complementary fit to the protein
surface. In the model of 10, an intermediate in the
design process, the Cl at the 5-position fills a small
hydrophobic pocket, and the less favorable interaction
of Cl at the 4-position with the backbone carbonyl
oxygen of Asn85 is removed. 7 and 16 have similar
potencies (0.82 and 1.4 µM, respectively), although the
R enantiomer of 16 is expected to have a greater potency
than either 7 or the racemic mixture that constitutes
16. The position of the terminal ring of the tricyclic of
7 overlaps fairly well with the position of the benzyl ring
in the predicted lower-energy mode of the R enantiomer
of 16 (Figure 9). The carbonyl oxygen on the tricyclic of
7 makes VDW contact with a backbone R-carbon but
does not make any polar interactions with the protein
and, as such, may not be optimal. The citrate molecule
in the active site of structure 7 superimposes with the
position of the phosphate group in CoA. 16 clearly could
be extended from the position of the sulfur (or 7 from
the 3- or 4-positions of the tricyclic) to fill the half of
the active site occupied by the phosphopantetheine
moiety of CoA. The compounds presented in this paper
fill the side of the AcpS active site occupied by the
adenine portion of CoA. Because 7 and 16 are both
below 385 MW, it should be possible to extend the
The diethoxyphosphoryl carboxylic acid 1, prepared
according to Krawczyk et al.,12 was coupled with 2-amino-
4-chlorobenzoic acid methyl ester in the presence of 1,3-
diisopropylcarbodiimide and catalytic DMAP to give
amide 2. Horner-Emmons coupling with 4-chloroben-
zaldehyde in the presence of 1,8-diazabicyclo[5.4.0]-
undec-7-ene and lithium chloride in THF gave the cis
olefin 3 (100% E selectivity determined by HPLC
analysis). Saponification of 3 provided the desired cis-
isomer of 4.
7. This compound was prepared by the coupling of
2-amino-5-chlorobenzoic acid with 9-fluorenone-4-car-
bonyl chloride (6), as shown in Scheme 2.
10. R-Methylhydrocinnamic acid 8 is converted to the
corresponding acid chloride 9 using oxalyl chloride. This
acid chloride is then coupled with methyl 2-amino-5-
chlorobenzoate to give the amide, which underwent
basic hydrolysis using LiOH‚H2O to give 10 (Scheme
3).
16 was prepared as shown in Scheme 4. Amino acid
11 was converted to the corresponding R-bromo acid 12
with NaNO2 and KBr in aqueous H2SO4. Conversion to
the amide 14 was effected by conversion to the acid
chloride 13 using oxalyl chloride, followed by reaction
with methyl 2-amino-5-chloro benzoate. The corre-

7966 Journal of Medicinal Chemistry, 2005, Vol. 48, No. 25
Joseph-McCarthy et al.
Scheme 3a
together have also clearly delineated opportunities for
further optimization of this series of AcpS inhibitors.
Experimental Section
Protein Purification for Biological Assays. The glu-
tathione S-transferase (GST) fusion of B. subtilis ACP was
purified as reported.3
Synthesis of Biotin-CoA. CoA (Sigma) was dissolved in
0.1 M potassium phosphate, pH 7.2, and mixed with a
stoichiometric amount of biocytin maleimide (Molecular Probes,
Eugene, OR). The reaction was complete within a few minutes,
and aliquots were stored at -20 °C.
a (a) (COCl)2, cat. DMF, CH2Cl2, 23°C; (b) methyl 2-amino-5-
chlorobenzoate, i-Pr2NEt, CH2Cl2, 23 °C; (c) LiOH‚H2O, THF/H2O,
23 °C.
Scheme 4a
Enzyme Assay. Enzyme activity was measured using a
homogeneous time-resolved fluorescence (HTRF) assay. Assays
were run in 384-well plates and consisted of compound, 15 ng/
mL AcpS, 1.9 µM biotin-CoA, and 1.25 µg/mL GST-ACP in
50 mM Tris‚HCl, pH 8.0, 7.5 mM MgCl2, 3.75 mM DTT,
0.0375% Tween-20, and 37.5 µg/mL BSA in a total volume 20
µL. After 3 h at room temperature, the reaction was termi-
nated by adding 60 µL of the stop/detection mix (82 ng/mL
europium cryptate conjugate of GST monoclonal antibody
(Packard, Meriden, CT), 28.7 µg/mL streptavidin-allophyco-
cyanin conjugate (ProZyme, San Leandro, CA), 410 mM KF,
4.1 mM EDTA in 0.5 × PBS). The mixture was incubated
overnight before reading in a Wallac Victor plate reader
(excitation at 340 nm, emission at 665 nm).
Membrane-Damage Assay. Molecular Probes’ LIVE/
DEAD BacLight Kit was used to examine the viability of B.
subtilis PY79 cultures after incubation with compound. Rapid
cell death (within 5 min of treatment) was interpreted as
arising from membrane damage induced by the compound
rather than due to specific inhibition of AcpS.
Urea Gel Assay. The urea-PAGE analysis of ACP is an
adaptation of the method found in reference 13. A saturated
culture of B. subtilis PY79 was diluted 1:200 into Luria-
Burtani (LB) medium with compound. The culture was grown
to midlogarithmic phase at 37 °C; 0.5 mL was spun down,
washed with PBS, and resuspended in 0.5 mL PBS. An equal
volume of 10% TCA was added, and proteins were precipitated
on ice for 1 h. The precipitate was collected, washed with 1%
TCA, and resuspended in 20 µL 1 M Tris. After adding 5 µL
loading buffer (100 mM Tris, pH 6.8, 30% glycerol, 0.5 M urea,
0.12 mg/mL bromophenol blue), the sample was fractionated
on an 18% polyacrylamide gel containing 1 M urea. Proteins
were transferred to an Immobilon-P membrane (Millipore,
Bedford, MA) and Western blotted using anti-ACP raised in
chicken, rabbit anti-chicken IgG horseradish peroxidase con-
jugate (Pierce, Rockford, IL), and were developed with Super-
Signal West Pico chemiluminescent substrate (Pierce).
Antimicrobial Susceptibility Testing. The in vitro ac-
tivities of the antibiotics were determined by the broth
microdilution method as recommended by the National Com-
mittee for Clinical Laboratory Standards (NCCLS).14 Mueller-
Hinton II broth (MHBII) (BBL Cockeysville, MD) was the
medium employed in the testing procedures. Microtiter plates
containing serial dilutions of each antimicrobial agent were
inoculated with each organism to yield the appropriate density
(105 CFU/mL) in a 100 µL-final volume. The plates were
incubated for 18-22 h at 35 °C in ambient air. The minimal
inhibitory concentration for all isolates was defined as the
lowest concentration of antimicrobial agent that completely
inhibits the growth of the organism as detected by the unaided
eye.
a (a) NaNO2, KBr, 1.25 M H2SO4, -8 to 23 °C; (b) (COCl)2, cat.
DMF, CH2Cl2, 0 to 45 °C; (c) methyl 2-amino-5-chlorobenzoate,
pyridine, CH2Cl2, 23 °C; (d) NaSMe, DMSO, 23 °C; (e) LiOH‚H2O,
THF/MeOH/H2O, 50 °C; (f) EDCI‚HCl, DMAP, CH2Cl2, 23 °C.
sponding thiomethyl analogue 15 was produced using
sodium methylsulfide, and hydrolysis of the methyl
ester with LiOH‚H2O gave 16.
The resolution of 16 by chiral HPLC could not be
carried out due to the poor solubility characteristics
of the compound. We then attempted the conversion of
16 to the corresponding diastereomeric esters using
1S,2S,3S,5R-(+)-isopinocampheol. Under the coupling
conditions, only the benzooxazin-4-one 17 was obtained.
Successful resolution of 17 was achieved using chiral
HPLC; however, subsequent hydrolysis of the individual
enantiomers of 17 resulted in complete racemization of
the acid formed.
Conclusions
Cocrystallization of AcpS-Inhibitor Complexes. His-
tidine-tagged AcpS was purified and concentrated as described
earlier.3 Cocrystals with 0 were obtained by using the hanging
drop method with a 10 mg/mL protein solution containing 2
mM 0 equilibrated at 18 °C against 0.2 M potassium dihydro-
gen phosphate and 20% PEG 3350. Similarly, 4 and 7 were
equilibrated against 0.2 M diammonium hydrogen citrate and
20% PEG 3350, whereas 16 was equilibrated against 0.2 M
ammonium chloride and 20% PEG 3350.
A series of novel anthranilic acid inhibitors of AcpS,
an infectious disease target, were obtained through the
use of a structure-based approach for the optimization
of a relatively weak HTS lead, 4. 4 had also been shown
to cause cell damage. On a fairly short time scale of less
than nine months, 16 was synthesized. The potency of
16 was improved 30-fold over the initial lead, and rapid
cell death was not observed in the membrane-damage
assay. Molecular modeling and X-ray crystallography
Data Collection and Refinement. Single-wavelength (1.1
Å) data for each inhibitor cocrystal was collected on beamline

Design of Novel Anthranilic Acid AcpS Inhibitors
Journal of Medicinal Chemistry, 2005, Vol. 48, No. 25 7967
5.0.2 at the ALS, Berkeley using an ADSC Quantum-4 CCD
detector. A single crystal, cooled to -180 °C, was used for each
data set. The data were processed using DENZO and Scale-
pack.15
yield) as a white solid, mp 105-106 °C. The sample was 98.3%
pure by LC/MS: tR 2.4 min; 362 [M - H]-, 100% pure by HPLC
at 254 nm (Keystone Aquasil C18 column; 2.5 min gradient
from 95% 10 mM ammonium acetate to 95% acetonitrile; UV
detection at 254 and 214 nm). 1H NMR (DMSO-d6, 400 MHz):
δ 2.16 (s, 3H), 3.89 (s, 3H), 7.29 (m, 1H), 7.51 (m, 5H), 8.01 (d,
1H), 8.68 (m, 1H), and 11.37 (s, 1H); MS [(+)ESI, m/z]: 363.9
[M + H]+.
The structure of the each AcpS/inhibitor complex was solved
by molecular replacement using the program AMORE.16 The
probe used in the molecular replacement search was an
aggregate of the six monomers of AcpS, as found in the AcpS
structure (PDB ID 1F7T), superimposed to yield one “mono-
mer”. Prior to refinement, 5% of the data were randomly
selected and designated as an Rfree test set to monitor the
progress of the refinement. The structures of each complex
were then rebuilt within QUANTA (Accelrys, San Diego, CA)
utilizing a series of omit maps. The statistics from refinement
are given in Table 3.
Molecular Modeling of Proposed Compounds. Con-
firmed hits from the HTS and proposed compounds were
modeled in the AcpS active site using a Monte Carlo docking
method. An active site is located at each of the three dimer
interfaces in the trimer structure. Starting with the AcpS-
CoA X-ray structure,3 the CoA was removed from each active
site, and polar hydrogens were added and optimized using the
program CHARMm (Accelrys, San Diego, CA, 200017). An
approximately 17 Å-radius spherical region of the protein
structure, centered around one of the active sites, was used
for the docking calculations. This site file was generated using
the program FLO99.18 The protein was held fixed during the
simulations except for the hydroxyl groups of Ser102 and
Ser61. Each ligand or potential ligand was docked into the
site by carrying out 1000 mcdock cycles with FLO99. The 25
best poses for each ligand based upon their overall score were
retained and viewed graphically. Once the AcpS-0 complex
structure was obtained, a site file for that protein structure
was similarly prepared and used for docking all proposed
compounds.
4-Chloro-2-{[(2E)-3-(4-chlorophenyl)-2-methylprop-2-
enoyl]amino}benzoic Acid (4). To a 0 °C solution of 3 (2.1
g, 5.78 mmol) in THF (75 mL) was added 1 N aqueous lithium
hydroxide (17.3 mL), and the mixture was stirred for 12 h at
room temperature. The solvent was removed in vacuo; the
residue was treated with water and acidified with 2 N
hydrochloric acid (to pH 2). The mixture was extracted with
dichloromethane (2×) and ethyl acetate (2×). The combined
extracts were dried over anhydrous magnesium sulfate and
evaporated to provide a solid (1.97 g). Recrystallization from
methanol yielded the title compound (1.83 g, a 91% yield) as
a white solid, mp 206-207 °C. HPLC: tR 2.803 min; purity
100% at 254 nm (Cromolith Monolith column, length 100 mm,
5 min gradient from 10 to 100% acetonitrile in water with 0.1%
trifluoroacetic acid; flow rate 4 mL/min; UV detection at 254
1
nm). H NMR (DMSO-d6, 400 MHz): δ 2.16 (s, 3H), 7.24 (m,
1H), 7.5 (m, 5H), 8.03 (d, 1H), 8.78 (s,1H), and 11.93 (s, 1H).
E stereochemistry supported by NOE data; MS [(-)ESI, m/z]:
348.01 [M - H]-. Anal. Calcd for C17H13Cl2NO3: C 58.31, H
3.74, N 4. Found: C 58.37, N 3.67, N 3.89.
5-Chloro-2-{[(9-oxo-9H-fluoren-4-yl)carbonyl]amino}-
benzoic Acid (7). To a stirred mixture of 9-fluorenone-4-
carboxylic acid 5 (500 mg, 2.2 mmol) and toluene (20 mL)
contaning a catalytic amount of DMF, was added oxalyl
chloride (312 mg, 2.5 mmol) in drops. After 1 h, a suspension
of 2-amino-5-chlorobenzoic acid (421 mg, 2.5 mmol) was added,
and the mixture was stirred for 12 h at room temperature.
The solvent was removed, and the residue was partitioned
between 0.1 N hydrochloric acid and ethyl acetate. The
insoluble material was collected, and the organic phase was
evaporated to dryness. The evaporated organics and the
insoluble solids were combined and triturated with methanol.
The resulting yellow solid was collected and dried under
vacuum (108 mg, a 13% yield). 1H NMR (DMSO-d6, 400
MHz): δ 7.41 (t, 1H), 7.56 (m, 2H), 7.67 (d, 1H), 7,78 (m, 3H),
7.85 (d, 1H), 7.96 (s, 1H), 8.58 (d, 1H), and 11.67 (s, 1H); MS
[(+)ESI, m/z]: 376 [M + H]+; RP-HPLC: H2O/MeCN/0.2%
TFA: 2.49 min; 97.4%; H2O/MeCN/0.2% Et3N: 3.10 min,
95.1%.
The lead compound, 4, was docked into the AcpS active site
as described above, and the binding mode was largely con-
firmed via X-ray crystallography. Modifications to 4 were
modeled, and those that preserved the original binding mode
and had improved estimated binding affinities were synthe-
sized.
4-Chloro-2-[2-(diethoxy-phosphoryl)-propionylamino]-
4-methyl-benzoic Acid Methyl Ester (2). To a 0 °C solution
of 2-(diethoxy-phosphoryl)-propionic acid12 1 (2.3 g, 10.9 mmol)
and 4-(dimethylamino)pyridine (130 mg, 1.09 mmol) in CH2-
Cl2 (26 mL) was added 1,3-diisopropylcarbodiimide (0.96 mL,
10.9 mol). After 30 min, methyl 4-chloro-2-amino benzoate
(3.04 g, 16.4 mmol) was added. The mixture was allowed to
warm to room temperature for 12 h with stirring. An ad-
ditional 0.2 equiv of 4-(dimethylamino)pyridine and 1,3-
diisopropylcarbodiimide were added, and the stirring contin-
ued for another 16 h. The mixture was diluted with EtOAc,
and the solution was washed with 2 N HCl, water, saturated
sodium carbonate, and brine, dried over anhydrous magnesium
sulfate, and evaporated to dryness. The residue was purified
by flash chromatography over silica Merck-60, eluting with a
gradient (from 1:1 to 8:2) of ethyl acetate in hexane to provide
2 (3.65 g, an 89% yield). 1H NMR (DMSO-d6, 400 MHz): δ
1.22 (m, 6H), 1.34 (m, 3H), 3.36 (m, 1H), 3.88 (s, 3H), 4.05 (m,
4H), 7.28 (d, 1H), 7.97 (d, 1H), 8.51 (s, 1H), and 10.93 (s, 1H);
MS [(+)ESI, m/z]: 377.9 [M + H]+.
4-Chloro-2-{[(2E)-3-(4-chlorophenyl)-2-methylprop-2-
enoyl]amino}benzoic Acid Methyl Ester (3). To a stirred
solution of 2 (3.65 g, 9.68 mmol) in THF (100 mL) was added
lithium chloride (0.47 g, 11.1 mmol) and 1,8-diazabicyclo[5.4.0]-
undec-7-ene (1.11 mL). After 30 min at room temperature,
4-chlorobenzaldehyde was added (1.9 g, 13.5 mmol), and the
stirring was continued for 12 h. The reaction mixture was then
diluted with water (80 mL) and extracted with Et2O (3 × 100
mL). The combined organics were dried over anhydrous
magnesium sulfate and evaporated to dryness. The residue,
preabsorbed onto a column of silica Merck-60, was flash
chromatographed using a gradient (from 2 to 10%) of ethyl
acetate in hexane to provide the title compound (1.7 g, a 48%
5-Chloro-2-[(2-methyl-3-phenylpropanoyl)amino]ben-
zoic Acid (10). To a solution of R-methylhydrocinnamic acid
8 (0.2820 g, 1.72 mmol), a few drops of DMF, and 50 mL CH2-
Cl2 was added oxalyl chloride (0.178 mL, 2.06 mmol) dropwise
at 23 °C. After stirring at room temperature for 1 h, the
mixture was concentrated to dryness. A solution of methyl
2-amino-5-chlorobenzoate (0.318 g, 1.72 mmol), i-Pr2Et (0.191
g, 1.90 mmol), and 15 mL CH2Cl2 was added dropwise, and
the reaction mixture was stirred at room temperature over-
night. After dilution with water and extraction with ethyl
acetate, the organic extracts were evaporated to provide a
yellow oil (0.357 g, 63%), which was treated with LiOH‚H2O
(0.45 g, 10.8 mmol) in 20 mL 1:1 THF/H2O. The resulting
mixture was stirred overnight at room temperature. The
reaction mixture was neutralized to pH 6 by the addition of 1
M aqueous sodium dihydrogenphosphate and then was ex-
tracted with EtOAc. The organic extracts were dried over
anhydrous Na2SO4 and evaporated to dryness. The resulting
solid was dissolved in a minimum amount of dichloromethane,
and the solution was diluted with hexane until a precipitate
was formed. The precipitate was collected by filtration to give
1
the title compound (0.345 g, 100%) as a white solid. H NMR
(CD3OD, 300 MHz): δ 1.19 (d, J ) 6.6 Hz, 3H), 2.71 (m, 2H),
2.94 (m, 1H), 7.14 (m, 5H), 7.44 (dd, J ) 9.0, 2.7 Hz, 1H), 7.92
(d, J ) 2.7 Hz, 1H), and 8.47 (d, J ) 9.0 Hz, 1H). MS [(-)ESI,
m/z]: 316.06 [M - H]-; [(+)ESI, m/z]: 318.11 [M + H]+. RP-

7968 Journal of Medicinal Chemistry, 2005, Vol. 48, No. 25
Joseph-McCarthy et al.
HPLC: H2O/CH3CN/0.1% formic acid: 5.87 min; 99.8%; H2O/
CH3OH/0.1% formic acid: 7.12 min, 100.0%.
(s, 3H); MS (ESI) m/z: 384 ([M + H]+); m/z: 382 ([M - H]-).
Anal. Calcd for C17H15NCl2O3S: C 53.13, H 3.93, N 3.64.
Found: C 53.23, H 4.00, N 3.51.
2-Bromo-3-(4-chloro-phenyl)propionic Acid (12). Into
1.25 M aqueous sulfuric acid (65 mL) at -8 °C was added KBr
(11.42 g, 96.0 mmol) followed by 4-chlorophenyl alanine 11
(5.99 g, 30.0 mmol). To this stirred suspension was added
sodium nitrite (2.07 g, 30.0 mmol) in portions over a period of
45 min. The mixture was allowed to warm slowly to room
temperature for 1 h. The reaction mixture was extracted with
EtOAc (3×), and the combined organic extracts were washed
with water and brine. The organic phase was dried (MgSO4),
filtered, and evaporated to a light-yellow oil that solidified
upon standing. The yield of the title compound was 72% (5.69
g). This material was carried on directly to the next step.
2-[2-Bromo-3-{4-chloro-phenyl)propionylamino]-5-chlo-
robenzoic Acid Methyl Ester (14). 2-Bromo-5-(4-chlorophe-
nyl)propionic acid 12 (5.31 g, 20.2 mmol), dichloromethane (30
mL), and DMF (2 drops), under a N2 atmosphere, were cooled
to 0 °C and treated with oxalyl chloride (12.8 mL, 0.101 mol).
After stirring and warming to room temperature over 2 h, the
mixture was placed in an oil bath at 45 °C for 3 h. TLC analysis
(aliquot quenched with methanol) showed complete conversion
to acid chloride 13. The mixture was evaporated and then
reconstituted with dichloromethane and re-evaporated. This
was repeated 3×. The evaporator was filled with dry nitrogen
gas when opened. The resulting yellow oil was dissolved in
dry dichloromethane and cooled to 0 °C, and pyridine (1.46 g,
18.4 mmol) was added. A solution of methyl 2-amino-5-
chlorobenzoate (3.41 g, 18.4 mmol) in dichloromethane (30 mL)
was rapidly added in drops, and the mixture was stirred at
room temperature for 12 h. The resulting mixture was washed
with 0.1 N aqueous HCl (2×), water, and brine. The organic
phase was dried (MgSO4), filtered, and evaporated. The title
compound was isolated pure by flash chromatography (silica
gel, 10% ethyl acetate in hexane) and crystallized from hot
hexane/ethyl acetate. The yield of the title compound was 68%
(5.90 g). 1H NMR (CDCl3, 400 MHz): δ 11.59 (br s) and 11.43
(br s) (rotamers, 1H), 8.64 (m, 1H), 8.00 (m, 1H), 7.51 (m, 1H),
7.24 (m, 2H), 7.18 (m, 2H), 4.61(m) and 4.52(m) (rotamers, 1H),
3.92 (s, 3H), and 3.58-3.24 (series of m, 2H).
6-Chloro-2-[2-(4-chloro-phenyl)-1-methylsulfanyl-ethyl]-
benzo[d][1,3]oxazin-4-one (17). A mixture of 5-chloro-2-[3(4-
chloro-phenyl)-2-methylsulfanyl proprionylamino) benzoic acid
16 (50 mg, 0.130 mmol) and 4-(dimethylamino)pyridine (0.016
g, 0.050 mmol) in CH2Cl2 (4 mL) at 0 °C was treated with 1-[-
3-(dimethylamino)propyl]ethylcarbodiimide HCl (28 mg, 0.148
mmol). The resulting mixture was stirred for 2 h at 23 °C.
The CH2Cl2 was evaporated, and the residue was partitioned
between ethyl acetate and water. The organic phase was then
washed with aqueous NaHCO3 (2×), water (2×), and brine
(1×). The organic phase was dried (Na2SO4), filtered, and
evaporated. The solid was purified using flash chromatography
(silica gel, 15-20% gradient of diethyl ether/hexane). Yield:
1
0.035 g (74%), mp 125 °C. H NMR (DMSO-d6, 400 MHz): δ
8.08 (d, J ) 2.45 Hz, 1H), 7.95 (dd, J ) 2.45 and 8.71 Hz, 1H),
7.64 (d, J ) 9.03 Hz, 1H), 7.33 (m, 4H), 4.03 (t, J ) 7.9 Hz,
1H), 3.33 (m,1H), 3.11 (m, 1H), and 2.12 (s, 3H); MS (ESI)
m/z: 366 ([M + H]+); m/z: 364 ([M - H]-); Anal. Calcd for
C17H13NCl2O2S: C 55.75, H 3.58, N 3.83. Found: C 55.66, H
3.22, N 3.27.
Chiral Separation of 17. Chiral separation was achieved
using a preparative Chiralcel AD column (25 cm × 2 cm) using
methanol as the mobile phase (12 mL/min). The separated
optical antipodes of 17 were then used in the following
experiment.
(()-5-Chloro-2-[3(4-chlorophenyl)-2-methylsulfanyl-
proprionylamino)benzoic Acid (16). A single optical isomer
of 6-chloro-2-[2-(4-chlorophenyl)-1-methylsulfanyl-ethyl]benzo-
[d][1,3]oxazin-4-one 17 (3.2 mg, 8.33 µmol) in 250 µL of THF/
water/MeOH, (5.0/2.5/1.0) was treated with LiOH‚H2O (1.1 mg,
24.78 µmol), and the resulting mixture was stirred at room
temperature. TLC showed complete conversion to product in
1.5 h. The mixture was adjusted to pH 6 using 0.1 N aq HCl,
and the mixture was extracted with dichloromethane (3 × 10
mL). The combined organic extracts were washed with water
(2×), dried (MgSO4), filtered, and evaporated. 1H NMR analy-
sis of this material confirmed that 16 was obtained, but chiral
HPLC (using a preparative Chiralcel AD column (25 cm × 2
cm) using methanol as the mobile phase (12 mL/min)) showed
that complete racemization had occurred.
5-Chloro-2-[3-(4-chloro-phenyl)-2-methylsulfanyl-pro-
pionylamino)benzoic Acid Methyl Ester (15). To a solution
of 2-[2-bromo-3-{4-chloro-phenyl)propionylamino]-5-chloroben-
zoic acid methyl ester 14 (2.00 g, 4.63 mmol) in DMSO (10
mL) was added sodium methyl sulfide (406 mg, 5.78 mmol).
The mixture was stirred at room temperature for 1 h. Water
(25 mL) was then added, and the resulting mixture was
extracted with ethyl acetate (2 × 50 mL). The combined
organic extracts were washed with water and brine, dried
(MgSO4), filtered, and evaporated to give a white solid.
Recrystallization from hot hexane/ethyl acetate gave 1.84 g
(4.62 mmol, a 99% yield) of the title compound as a white
Supporting Information Available: Elemental analysis
and HPLC results. This material is available free of charge
via the Internet at http://pubs.acs.org.
References
(1) Magnuson, K.; Jackowski, S.; Rock, C. O.; Cronan, J. E. J.
Regulation of fatty acid biosynthesis in Escherichia coli. Micro-
biol. Rev. 1993, 57, 522-542.
(2) Elovson, J.; Vagelos, P. R. Acyl carrier protein. X. Acyl carrier
protein synthase. J. Biol. Chem. 1968, 243, 3603-3611.
(3) Parris, K. D.; Lin, L.; Tam, A.; Mathew, R.; Hixon, J.; et al.
Crystal structures of substrate binding to Bacillus subtilis holo-
(acyl carrier protein) synthase reveal a novel trimeric arrange-
ment of molecules resulting in three active sites. Struct. Folding
Des. 2000, 8, 883-895.
1
crystalline solid, mp 141 °C. H NMR (DMSO-d6, 400 MHz):
δ 10.78 (br s, 1H), 8.17 (m, 1H), 7.82 (m, 1H), 7.62 (m, 1H),
7.26 (m, 4H), 3.80 (s, 3H), 3.77 (m,1H), 3.14 (m, 1H), 2.91 (m,
1H), and 2.07 (s, 3H). MS (ESI) m/z: 398 ([M + H]+); Anal.
Calcd for C18H17NCl2O3S: C 54.82, H 4.30, N 3.52. Found: C
54.34, H 4.24, N 3.17.
(4) Chirgadze, N. Y.; Briggs, S. L.; McAllister, K. A.; Fischl, A. S.;
Zhao, G. Crystal structure of Streptococcus pneumoniae acyl
carrier protein synthase: An essential enzyme in bacterial fatty
acid biosynthesis. EMBO J. 2000, 19, 5281-5287.
(5) Prescott, D. J.; Vagelos, P. R. Acyl Carrier Protein. Adv. Enzymol.
1971, 36, 269-311.
(6) Takiff, H. E.; Baker, T.; Copeland, T.; Chen, S. M.; Court, D. L.
Locating essential Escherichia coli genes by using mini-Tn10
transposons: The pdxJ operon. J. Bacteriol. 1992, 174, 1544-
1553.
(7) McAllister, K. A.; Peery, R. B.; Meier, T. I.; Fischl, A. S.; Zhao,
G. Biochemical and molecular analyses of the Streptococcus
pneumoniae acyl carrier protein synthase, an enzyme essential
for fatty acid biosynthesis. J. Biol. Chem. 2000, 275, 30864-
30872.
(8) Gilbert, A. M.; Kirisits, M.; Toy, P.; Nunn, D. S.; Failli, A.; et al.
Anthranilate 4H-oxazol-5-ones: Novel small molecule antibacte-
rial acyl carrier protein synthase (AcpS) inhibitors. Bioorg. Med.
Chem. Lett. 2004, 14, 37-41.
5-Chloro-2-[3(4-chloro-phenyl)-2-methylsulfanyl-pro-
prionylamino)benzoic Acid (16). To 5-chloro-2-[3-(4-chloro-
phenyl)-2-methylsulfanyl-propionylamino)benzoic acid methyl
ester 15 (1.23 g, 2.87 mmol) in 50 mL of THF/H2O/MeOH (5.0/
2.5/1.0) was added LiOH‚H2O (902 mg, 21.51 mmol). The
mixture was heated in an oil bath at 50 °C for 1 h. The organics
were removed by evaporation, water (30 mL) was added and,
the pH was adjusted to 4 using 2 N HCl. This mixture was
extracted using ethyl acetate (4 × 50 mL). The combined
organic extracts were dried (MgSO4), filtered, and evaporated
to a white solid. The solid was purified by crystallization from
ethanol/water to give 0.625 g (57% yield) of the title compound
as a white crystalline solid, mp 173-174 °C. 1H NMR (DMSO-
d6, 400 MHz): δ 11.37 (br s, 1H), 8.42 (d, J ) 9.03 Hz, 1H),
7.90 (d, J ) 2.56 Hz, 1H), 7.64 (dd, J ) 9.03 and 2.56 Hz, 1H),
7.31 (m, 4H), 3.79 (t, 1H), 3.18 (m,1H), 2.95 (m, 1H), and 2.09

Design of Novel Anthranilic Acid AcpS Inhibitors
Journal of Medicinal Chemistry, 2005, Vol. 48, No. 25 7969
(9) Morbidoni, H. R.; de Mendoza, D.; Cronan, J. E., Jr. Bacillus
subtilis acyl carrier protein is encoded in a cluster of lipid
biosynthesis genes. J. Bacteriol. 1996, 178, 4794-4800.
(10) Eliseev, A. V.; Lehn, J.-M. Dynamic combinatorial chemistry:
Evolutionary formation and screening of molecular libraries.
Curr. Top. Microbiol. Immunol. 1999, 243, 159-172.
(11) Erlanson, D. A.; Braisted, A. C.; Raphael, D. R.; Randal, M.;
Stroud, R. M.; et al. Site-directed ligand discovery. Proc. Natl.
Acad. Sci. U.S.A. 2000, 97, 9367-9372.
(15) Otwinowski, Z.; Minor, W. Processing of X-ray diffraction data
collected in oscillation mode. Methods Enzymol. 1997, 276, 307-
326.
(16) Navaza, J. AmoRe: An automated package for molecular
replacement. Acta Crystallogr., Sect. A: Found. Crystallogr.
1994, 40, 157-163.
(17) Brooks, B. R.; Bruccoleri, R. E.; Olafson, B. D.; States, D. J.;
Swaminathan, S.; et al. CHARMM: A program for macromo-
lecular energy, minimization, and dynamics calculations. J.
Comput. Chem. 1983, 4, 187-217.
(18) McMartin, C.; Bohacek, R. S. QXP: Powerful, rapid computer
algorithms for structure-based drug design. J. Comput.-Aided
Mol. Des. 1997, 11, 333-344.
(12) Krawczyk, H.; Koszuk, J.; Bodalski, R. A Useful Synthesis of
Diethyl 1-Substituted Vinylphosphonates. Pol. J. Chem. 2000,
74, 1123-1128.
(13) Keating, D. H.; Carey, M. R.; Cronan, J. E., Jr. The unmodified
(apo) form of Escherichia coli acyl carrier protein is a potent
inhibitor of cell growth. J. Biol. Chem. 1995, 270, 22229-22235.
(14) Methods for Dilution Antimicrobial Susceptibility Tests for
Bacteria That Grow Aerobically; Approved Standards, M7-A5;
National Committee for Clinical Laboratory Standards: Wayne,
PA, 2000.
(19) Figures 5, 9, and the TOC graphic were created using the
program PyMOL.; Delano, W. L. The PyMOL Molecular Graph-
ics System; DeLano Scientific LLC: San Carlos, CA. http://
www.pymol.org.
JM050523N