Amphipathic benzoic acid derivatives: Synthesis and binding in the hydrophobic tunnel of the zinc deacetylase LpxC

Article abstract of DOI:10.1016/j.bmc.2007.01.044

The first committed step in lipid A biosynthesis is catalyzed by uridine diphosphate-(3-O-(R-3-hydroxymyristoyl))-N-acetylglucosamine deacetylase (LpxC), a zinc-dependent deacetylase, and inhibitors of LpxC may be useful in the development of antibacteria

Full text of DOI:10.1016/j.bmc.2007.01.044

Bioorganic & Medicinal Chemistry 15 (2007) 2617–2623
Amphipathic benzoic acid derivatives: Synthesis and binding
in the hydrophobic tunnel of the zinc deacetylase LpxC
Hyunshun Shin,a Heather A. Gennadios,b
Douglas A. Whittingtonb, and David W. Christiansonb,*
aDepartment of Chemistry, University of San Francisco, 2130 Fulton Street, San Francisco, CA 94117-1080, USA
bRoy and Diana Vagelos Laboratories, Department of Chemistry, University of Pennsylvania, Philadelphia, PA 19104-6323, USA
Received 29 November 2006; revised 15 January 2007; accepted 26 January 2007
Available online 31 January 2007
Abstract—The first committed step in lipid A biosynthesis is catalyzed by uridine diphosphate-(3-O-(R-3-hydroxymyristoyl))-N-
acetylglucosamine deacetylase (LpxC), a zinc-dependent deacetylase, and inhibitors of LpxC may be useful in the development
of antibacterial agents targeting a broad spectrum of Gram-negative bacteria. Here, we report the design of amphipathic benzoic
acid derivatives that bind in the hydrophobic tunnel in the active site of LpxC. The hydrophobic tunnel accounts for the specificity
of LpxC toward substrates and substrate analogues bearing a 3-O-myristoyl substituent. Simple benzoic acid derivatives bearing an
aliphatic ‘tail’ bind in the hydrophobic tunnel with micromolar affinity despite the lack of a glucosamine ring like that of the sub-
strate. However, although these benzoic acid derivatives each contain a negatively charged carboxylate ‘warhead’ intended to coor-
˚
dinate to the active site zinc ion, the 2.25 A resolution X-ray crystal structure of LpxC complexed with 3-(heptyloxy)benzoate reveals
‘backward’ binding in the hydrophobic tunnel, such that the benzoate moiety does not coordinate to zinc. Instead, it binds at the
outer end of the hydrophobic tunnel. Interestingly, these ligands bind with affinities comparable to those measured for more com-
plicated substrate analogue inhibitors containing glucosamine ring analogues and hydroxamate ‘warheads’ that coordinate to the
active site zinc ion. We conclude that the intermolecular interactions in the hydrophobic tunnel dominate enzyme affinity in this
series of benzoic acid derivatives.
Published by Elsevier Ltd.
1. Introduction
toxic lipid A concentrations shed by dying bacteria. This
therapeutic strategy could allow for improved manage-
ment of Gram-negative sepsis and has inspired the
search for potent inhibitors of UDP-(3-O-(R-3-hydroxy-
myristoyl))-N-acetylglucosamine deacetylase (LpxC), a
zinc enzyme7 that catalyzes the first committed step of
lipid A biosynthesis.8–11 To date, a variety of inhibitors
have been developed against LpxC that contain
hydroxamate or phosphonate ‘warheads’ for zinc
coordination.12–20
Nearly 600,000 cases of sepsis or septic shock are diag-
nosed annually in the US, causing an estimated
100,000 deaths per year.1,2 Gram-negative bacterial sep-
sis arises from the systemic response to infection, mainly
the overexpression of cytokines and inflammatory medi-
ators in response to macrophage activation by lipopoly-
saccharide (LPS), which comprises the outer leaflet of
the outer membrane of Gram-negative bacteria.3–6 The
toxic component of LPS is lipid A, which serves as the
hydrophobic anchor of LPS and is essential for bacterial
survival.3–6 As a therapeutic strategy, the inhibition of
enzymes in the lipid A biosynthetic pathway will not
only kill Gram-negative bacteria, but will also reduce
The X-ray crystal structure of LpxC from Aquifex
aeolicus reveals a catalytic zinc ion ðZn2þ Þ coordinated
A
by H79, H238, and D242, and a solvent molecule at the
21
˚
base of a ꢀ20 A-deep active site cleft. The native en-
zyme structure was determined in the presence of excess
zinc, which resulted in the binding of an inhibitory zinc
Keywords: Lipid A biosynthesis; Zinc enzyme; Hydrophobic tunnel;
Enzyme–inhibitor complex.
ion ðZn2þ Þ to catalytic residues E78 and H265. Also
B
coordinated to Zn2þ was a fatty acid interpretable as
*
B
Corresponding author. Tel.: +1 215 898 5714; fax: +1 215 573
2201; e-mail: chris@sas.upenn.edu
either myristate or disordered palmitate, and the
aliphatic portion of the fatty acid was bound in a
hydrophobic tunnel separate from the main active site
Present address: Amgen Cambridge Research Center, One Kendall
Square, Building 1000, Cambridge, MA 02139, USA.
0968-0896/$ - see front matter Published by Elsevier Ltd.
doi:10.1016/j.bmc.2007.01.044

2618
H. Shin et al. / Bioorg. Med. Chem. 15 (2007) 2617–2623
cleft (Fig. 1). It was hypothesized that the 3-O-(R-3-
hydroxymyristoyl) group of the substrate occupied this
tunnel during catalysis,21 and subsequently determined
NMR and X-ray crystal structures of A. aeolicus LpxC
complexed with the substrate analogue inhibitor16,19
TU-514 were consistent with this hypothesis.22–24
in the hydrophobic tunnel of LpxC. Surprisingly, the
X-ray crystal structure of LpxC complexed with one of
these derivatives reveals a ‘backward’ binding mode in
the hydrophobic tunnel. Therefore, intermolecular inter-
actions in the hydrophobic tunnel, and not carboxylate-
zinc coordination interactions, appear to be the primary
affinity determinants for this series of benzoic acid
derivatives.
Importantly, the inhibitory Zn2þB ion as well as its bound
fatty acid are easily dialyzed out of the crystalline LpxC
active site, which facilitates the X-ray crystallographic
structure determinations of enzyme–inhibitor complex-
es.24,25 However, dialysis of Zn2þB out of the LpxC active
site in the absence of an inhibitor causes the fatty acid to
2. Results and discussion
2.1. Chemistry
shift such that it coordinates to Zn2þ with bidendate
A
coordination geometry,25 thereby completing a square
pyramidal zinc coordination polyhedron.24 Since lauric
acid (dodecanoic acid) binds to LpxC with Kd = 1 lM,21
we hypothesized that an amphipathic carboxylate deriva-
tive could serve as a lead for the design of inhibitors.
Moreover, since simple fatty acids bind with affinities
comparable to those reported for substrate analogues of
more complicated design and synthesis (e.g., the substrate
analogue inhibitor TU-514 binds with IC50 = 7.0 lM and
Ki = 0.65 lM),16,19 these results strongly suggest that the
primary determinants of LpxC-inhibitor affinity are a
suitable ‘warhead’ that completes a square pyramidal zinc
coordination polyhedron and a pendant ‘tail’ that binds
in the hydrophobic tunnel. That the N-aroyl-L-threonine
hydroxamate CHIR-090 lacks a hexose ring, yet is a slow,
tight-binding inhibitor with Ki ꢀ 1 nM,20 is consistent
with this view of structure–affinity relationships in the
LpxC active site.
Compounds 1–6 and 8 were obtained from commercial
sources. The 4-(heptyloxy)-benzoic acid derivatives
(compounds 10–13) were prepared with moderate to
good yields (60–80%, unoptimized) from commercially
available methyl 4-(hydroxy)benzoates 10a–13a by
treatment with potassium carbonate in acetone, fol-
lowed by hydrolysis with lithium hydroxide in tetrahy-
drofuran (THF) (Scheme 1). Compounds 9 and 14
were prepared in similar manner starting with commer-
cially available methyl 2-(4-(hydroxy)phenyl)acetate or
methyl 3-(hydroxy)benzoate, respectively. Compound 7
was prepared by the Wittig reaction of heptyltriphenyl-
phosphonium bromide and methyl 4-(formyl)benzoate
followed by hydrolysis with lithium hydroxide in THF.
2.2. Binding affinities of amphipathic benzoic acid
derivatives
Amphipathic benzoic acid derivatives bind to LpxC
from A. aeolicus with the affinities reported in Table 1
as determined by isothermal titration calorimetry.
Here, we report the synthesis and evaluation of simple
amphipathic benzoic acid derivatives designed to bind
Figure 1. Active site cleft of LpxC; for clarity, only Zn2þ and Zn2þ are shown, and the metal-bridging solvent molecule appears as a small red
A
B
sphere. Myristic acid (green) occupies the hydrophobic tunnel adjacent to Zn2þ
.
B
R
R
O
OCH3
O
i)
ii)
CH3(CH2)6 -I
+
HO
O
OH
10a: R = OH
10: R = OH
11a: R = OCH3
12a: R = NO2
13a: R = Cl
11: R = OCH3
12: R = NO2
13: R = Cl
Scheme 1. Synthesis of 4-(heptyloxy)benzoic acid derivatives. Reagents: (i) K2CO3/acetone; (ii) LiOH/THF.

H. Shin et al. / Bioorg. Med. Chem. 15 (2007) 2617–2623
Table 1. Amphipathic benzoic acid derivatives and affinities for LpxC
2619
unsaturated ligands can achieve a comparable binding
conformation and zinc coordination geometry to that
expected for a saturated fatty acid. Benzoic acid deriva-
tives lacking an aliphatic tail do not exhibit measurable
binding to LpxC (compounds 4 and 5). Moreover, the
introduction of a para-substituted aromatic ring into
the pendant tail of a longer fatty acid results in a ꢀ35-
fold affinity decrease (compound 6) compared with com-
pound 3. At first glance, this might suggest that the
introduction of a bulky, conformationally rigid aromat-
ic ring into an aliphatic carboxylate will not enhance en-
zyme–ligand affinity. However, the introduction of an
additional degree of unsaturation into the pendant tail
(compound 7), or the introduction of a polar oxygen
atom to generate an aliphatic–aromatic ether (com-
pound 8), compensates for the observed affinity loss of
compound 6. Lengthening the point of carboxylate
attachment to the aromatic ring by a single methylene
group has a negligible effect on affinity (compound 9).
The substitution of a single bulky substituent (hydroxy,
methoxy, nitro) in the meta-position relative to the car-
boxylate moiety causes only modest (3- to 4-fold)
decreases in affinity of the aliphatic–aromatic ether
derivative (compounds 10–12). However, the substitu-
tion of a meta-chloro substituent decreases affinity
14-fold (compound 13; since the methoxy and nitro sub-
stituents are larger than the chloro substituent, this nota-
ble affinity loss is probably not due to a steric effect). A
meta-substituted aliphatic–aromatic ether (compound
14) exhibits comparable potency to that of the para-
substituted isomer (compound 8). Taken together, these
results suggest that substituents at the para- and meta-
positions of the aromatic ring might yield a ‘two-pronged’
inhibitor that could achieve additional binding interac-
tions, for example, as recently demonstrated for certain
inhibitors of the zinc enzyme carbonic anhydrase.26,27
Compound
Kd (lM)
>100a
1
2
3
4
5
6
7
8
9
7.0a
0.9a
No binding
No binding
35
6.2
2.3
2.6
10
11
12
13
14
9.0
8.6
6.4
30
2.3. Crystal structure of the LpxC–3-(heptyloxy)benzoate
(14) complex
The X-ray crystal structure of LpxC complexed with
3-(heptyloxy)benzoate (14) (Table 1) determined at
˚
2.25 A resolution reveals 1:1 enzyme–ligand stoichiome-
try. An electron density map of the enzyme–ligand com-
plex is found in Figure 2. Ligand binding does not
perturb the overall structure of LpxC, and the rms devi-
ations of 267 Ca atoms between the zinc-inhibited struc-
2.0
ture21 and the complex with 14 are 0.215 A. As initially
˚
designed, the aliphatic ‘tail’ of 14 is inserted into the
active site hydrophobic tunnel. Surprisingly, however,
the ligand binds in a ‘backward’ fashion, such that the
benzoate carboxylate moiety does not coordinate to
Zn2þA. Instead, the carboxylate group binds at the outer
end of the hydrophobic tunnel and accepts hydrogen
bonds from K2O2 and water molecule #168/#120
(monomer A/B). In monomer A, the carboxylate group
also accepts a hydrogen bond from the backbone NH
group of L212 from monomer B (Fig. 2).
a Ref. 21.
Micromolar affinity is consistently achieved for com-
pounds bearing a saturated aliphatic ‘tail’ of 10 carbons
or longer (compounds 2 and 3).21 Based on the binding
of palmitate in the LpxC–palmitate complex,25 the
carboxylate group of these ligands had been expected
to complete a square pyramidal Zn2þ coordination
A
polyhedron in the enzyme active site.
The introduction of rigid C@C double bonds in the
ligand skeleton should enhance enzyme–inhibitor
affinity by decreasing the entropic cost of freezing out
C–C bond rotations required for ligand binding, if the
Curiously, a ‘Y’-shaped peak of electron density corre-
sponds to a nonprotein zinc ligand in the crystal struc-
ture of the LpxC complex with 3-(heptyloxy)benzoate,
and this electron density cannot be confidently

2620
H. Shin et al. / Bioorg. Med. Chem. 15 (2007) 2617–2623
Figure 2. LpxC–3-(heptyloxy)benzoate (14) complex (monomer A). (a) Simulated annealing omit electron density map contoured at 4r (magenta);
selected active site residues are indicated. A ‘Y’-shaped electron density peak corresponds to an uninterpretable nonprotein ligand to Zn2þ
.
A
(b) Hydrogen bond (orange) and metal coordination interactions (cyan) are indicated by dashed lines. Note that 3-(heptyloxy)benzoate binds
backward relative to the binding of palmitate,25 which is superimposed in light green. Atoms are color coded as follows: main chain ribbon and side
chain carbon atoms, gray (LpxC–3-(heptyloxy)benzoate complex); 3-(heptyloxy)benzoate carbon atoms, black; main chain ribbon side chain carbon
and palmitate carbon atomsm, green (LpxC–palmitate complex); N, blue; O, red; S, orange; zinc ions appear as gray (LpxC–3-(heptyloxy)benzoate
(14) complex) or green (LpxC–palmitate complex) spheres; and water molecules appear as small red spheres.
interpreted. This electron density could correspond to
an acetate anion; the carboxylate group of an otherwise
disordered fatty acid; or, perhaps, a disordered sulfate
ion. This electron density is unlikely to correspond to
a molecule of 3-(heptyloxy)benzoate bound with its car-
boxylate coordinated to zinc due to the significant break
between the ‘Y’-shaped peak and the peak correspond-
ing to backward-bound 3-(heptyloxy)benzoate (Fig. 2).
Nearby, a sulfate ion binds in the ‘basic patch’ of the en-
zyme active site: in monomer A, sulfate interacts with
the quality of the electron density in omit maps: discon-
nected electron density persisted in refinement due to a
substantial negative electron density peak between the
palmitate C2 and C4 atoms (data not shown). Thus, given
that the simulated annealing omit map of 3-(heptyl-
oxy)benzoate in Figure 2 clearly indicates the position
of the benzoic acid moiety, and given that the calculation
of a simulated omit map minimizes the effects of model
bias, we are confident in the interpretation of ligand bind-
ing in the LpxC active site as being comprised of one mol-
ecule of 3-(heptyloxy)benzoate bound backward in the
hydrophobic tunnel, and a separate and possibly disor-
˚
K239 (3.3 A) and also hydrogen bonds with water mol-
ecule #40; in monomer B, sulfate interacts with K239
(3.5 A), and the backbone NH groups of H265 and H58.
dered carboxylate-containing molecule bound to Zn2þ
.
˚
A
It should be noted that the ‘Y’-shaped peak of electron
density corresponding to the nonprotein zinc ligand was
significantly disconnected from the extended electron
density in the hydrophobic tunnel even in the initial stages
of map calculation and refinement. Although we consid-
ered that this peak might correspond to an undisplaced
palmitate molecule with its carboxylate group bound to
zinc as previously described,25 the interpretation and
refinement of palmitate in this structure did not improve
There is precedent for the backward binding of amphi-
pathic carboxylates in the hydrophobic tunnel of crys-
talline LpxC: in the complex with imidazole, the
carboxylate group of a fatty acid bound with backward
orientation makes no interactions with solvent or with
the adjacent monomer in the asymmetric unit;24 in the
complex with UDP, the carboxylate group of a fatty
acid bound with backward orientation hydrogen bonds
with solvent.28 Interestingly, in the complex with 3-(hep-

H. Shin et al. / Bioorg. Med. Chem. 15 (2007) 2617–2623
2621
tyloxy)benzoate, a salt link between the carboxylate and
K2O2 may stabilize the backward orientation of the li-
gand (Fig. 2b). No binding is measurable for 4-heptylan-
iline (data not shown), which is incapable of making a
salt link with K2O2 in a backward orientation or coor-
dinating to zinc in a forward orientation. Therefore, we
conclude that a negatively charged carboxylate group is
important for binding regardless of the ligand orienta-
tion in the hydrophobic tunnel.
3.1.1. n-Heptylation of methyl hydroxybenzoate deriva-
tives. A solution of 3-substituted-methyl 4-(hydroxy)-
benzoate (3-substituent = –OH, –OCH3, –NO2, or
–Cl) or methyl-3-(hydroxy)benzoate, or methyl 2-(4-
(hydroxy)phenyl)acetate (5.46 mmol) and anhydrous
potassium carbonate (10.7 mmol) in n-heptyl iodide
(10.7 mmol) and acetone (25 mL) was refluxed for
4–16 h. Evaporation of the solvent and flash chromatog-
raphy of the residue over silica gel, using 2:8 EtOAc/
hexane, gave each desired product as a colorless oil.
While it could be argued that the backward orientation
of 3-(heptyloxy)benzoate is an artifact of packing inter-
actions in the crystal lattice, the minimal number of
carboxylate interactions observed for ligands bound
with backward orientation might suggest that crystal
lattice interactions do not play a critical role in govern-
ing ligand orientation in the hydrophobic tunnel. If this
3.1.2. Hydrolysis of methyl heptyloxybenzoate deriva-
tives. Aqueous LiOH (2 M, 2 mL) was added to a stirred
solution of 3-substituted-methyl 4-(heptyloxy)benzoate
or methyl 3-(heptyloxy)benzoate or methyl 2-(4-(heptyl-
oxy)phenyl)acetate (1.13 mmol) in THF (2 mL). Stirring
was continued for 2–6 h at 70 °C, and the mixture was
acidified with concentrated hydrochloric acid. The mix-
ture was extracted with EtOAc. The combined organic ex-
tracts were dried (MgSO4) and evaporated. Flash
chromatography of the residue over silica gel, using 2:8
EtOAc/hexane, gave each of compounds 9–14 as a white
solid. Characterization data for each compound are
recorded in Sections 3.1.4–3.1.9.
is the case, then a carboxylate-Zn2þ interaction is not
A
a critical driving force for ligand association. Instead,
it follows that binding in the hydrophobic
tunnel is the primary affinity determinant for the series
of amphipathic benzoic acid derivatives shown in
Table 1. Given that a substrate analogue inhibitor bear-
ing the longer 3-O-myristic acid substituent is more po-
tent (IC50 = 7.0 lM) than an inhibitor bearing the
shorter 3-O-hexanoic acid substituent or no substituent
at all (IC50 > 3100 lM),16 van der Waals interactions be-
tween the long-chain aliphatic group of an inhibitor and
the hydrophobic tunnel appear to make a critical contri-
bution to inhibitor binding that surpasses that of zinc
coordination in certain cases.
3.1.3. 4-(Oct-1-enyl)benzoic acid (7). To a solution of
heptyltriphenylphosphonium bromide (18.3 mmol) in
anhydrous THF (30 mL), n-BuLi (2.5 M in hexanes,
18.3 mmol) was added dropwise via syringe at room
temperature. The reaction mixture was stirred for
15 min and then cooled to 0–4 °C. A solution of methyl
4-formylbenzoate (12.2 mmol) in anhydrous THF
(15 mL) was then added dropwise, and the reaction mix-
ture was warmed to ambient temperature and stirred for
1.45 h. The reaction was quenched with saturated aque-
ous ammonium chloride (25 mL) and extracted with ethyl
acetate (EtOAc). Combined organic extracts were dried
over magnesium sulfate and evaporated. Flash chroma-
tography of the residue over silica gel, using 2:8 EtOAc/
hexane, provided methyl 4-(oct-1-enyl)benzoate (2.1 g,
85%) as a light yellow oil. Hydrolysis of methyl 4-(oct-
1-enyl)benzoate was achieved by using the general meth-
In conclusion, amphipathic benzoic acid derivatives do
not contain an optimal ‘warhead’ for zinc coordination
in the active site of LpxC, based on the observed binding
mode of 3-(heptyloxy)benzoate. However, the results
reported herein illustrate the importance of interactions
in the hydrophobic tunnel and the substantial degree to
which the hydrophobic effect drives these interactions to
achieve impressive affinity for the binding of relatively
simple ligands.
1
od outlined above in Section 3.1.2. H NMR (CDCl3,
500 MHz) d 0.84–0.89 (m, 3H), 1.25–1.36 (m, 6H),
1.36–1.49 (m, 2H), 2.20–2.34 (m, 2H), 5.75–5.80 (m,
1H), 7.33–7.41(dd, J = 8.2, 8.2 Hz, 2H), 8.00–8.05 (dd,
J = 8.3, 8.2 Hz, 2H), 8.10 (d, J = 2.2 Hz, 1H); high reso-
lution mass spectrum (HRMS) m/z calcd for C14H20O2
232.1538; found: 232.1538.
3. Materials and methods
3.1. General synthetic procedures
Unless otherwise specified, materials were purchased
from commercial suppliers and used without further
purification. Flash chromatography was carried out on
Kieselgel 60 (230–400 mesh) and Biotage (Silica Car-
1
3.1.4. 2-(4-(Heptyloxy)phenyl)acetic acid (9). H NMR
(CDCl3, 500 MHz) d 0.85 (t, J = 8.4 Hz, 3H), 1.26–
1.38 (m, 8H), 1.63–1.68 (m, 2H), 3.03 (s, 2H), 3.87 (t,
J = 6.3 Hz, 2H), 6.71 (d, J = 8.1 Hz, 2H), 7.04 (d,
J = 8.1 Hz, 2H); MS (ES+) m/z calcd for C15H22O3
250.1563; found: 250.30.
TM
tridge FLASH 40 + S), and analytical thin-layer chro-
matography was performed on precoated silica gel (60
F254). H and 13C spectra were recorded at 500 MHz
1
in CDCl3, and chemical shifts (d) are expressed in parts
per million relative to residual CHCl3 at d = 7.24 ppm
for 1H and to CDCl3 at d = 77.0 ppm for 13C. 1H
NMR J values are reported in Hertz. Mass spectra were
recorded at low resolution using Micromass Platform
LC spectrometer in electrospray mode and at high reso-
lution with either a VG Micromass 70/70H or VG ZAB-
E spectrometer at the University of Pennsylvania.
3.1.5. 4-(Heptyloxy)-3-hydroxybenzoic acid (10). 1H
NMR (CDCl3, 500 MHz) d 0.86–0.88 (m, 3H), 1.22–
1.46 (m, 8H), 1.80–1.84 (m, 2H), 4.09 (t, J = 6.4 Hz,
2H), 5.63 (s, 1H), 6.86 (d, J = 8.4, 1H), 7.62–7.65 (m,
2H); HRMS m/z calcd for C14H20O4 252.1357; found:
252.1260.

2622
H. Shin et al. / Bioorg. Med. Chem. 15 (2007) 2617–2623
3.1.6. 4-(Heptyloxy)-3-methoxybenzoic acid (11). 1H
NMR (CDCl3, 500 MHz) d 0.87 (t, J = 7.0 Hz, 3H),
1.29–1.39 (m, 6H), 1.48–1.51 (m, 2H), 1.83–1.89 (m,
2H), 3.90 (s, 3H), 4.09 (t, J = 6.5 Hz, 2H), 7.11–7.13
(m, 1H), 6.94 (d, J = 8.73, 1H), 7.95–7.97 (dd, 1H),
8.10 (d, J = 1.0 Hz, 1H); 13C NMR (CDCl3, 500 MHz)
d 14.0, 22.6, 25.9, 29.0, 31.7, 56.1, 69.12, 111.5, 112.8,
121.2, 124.5, 149.0, 153.5, 170.5; HRMS m/z calcd for
C15H22O4 266.1512; found: 266.1518.
3.3. Crystallography
The C193A/DD284-L294 variant of LpxC from
A. aeolicus (henceforth ‘LpxC’) was overexpressed in
Escherichia coli and purified as described7,21 and crystal-
lized using previously reported conditions.21 To prepare
the enzyme–ligand complex, crystals were gradually
transferred to a stabilization buffer containing 100 mM
Bis–Tris (pH 6.0), 180 mM NaCl, 14–16% PEG 3350,
and 0.5 mM ZnSO4. This facilitated the removal of the
inhibitory zinc ion, Zn2þB, which is coordinated by the
side chains of E78, H265, and a solvent molecule that
1
3.1.7. 4-(Heptyloxy)-3-nitrobenzoic acid (12). H NMR
(CDCl3, 500 MHz) d 0.87 (t, J = 6.8 Hz, 3H), 1.28–
1.36 (m, 6H), 1.44–1.44 (m, 2H), 1.82–1.86 (m, 2H),
4.17 (t, J = 6.5 Hz, 2H), 7.10–7.12 (d, J = 8.9, 1H),
8.21–8.23 (dd, 1H), 8.53 (d, J = 2.2 Hz, 1H); 13C
NMR (CDCl3, 500 MHz) d 14.0, 22.6, 25.7, 28.7, 28.9,
31.7, 70.2, 113.9, 121.1, 127.8, 135.7, 156,4, 169.6;
HRMS m/z calcd for C14H19O5N 281.1341; found:
281.1334.
bridges the catalytic zinc ion ðZn2þ Þ in the structure
B
of the zinc-inhibited enzyme.21 Crystals were subse-
quently transferred to a similar stabilization buffer
containing 100 mM Hepes (pH 7.0) and 2 mM 3-(hep-
tyloxy)benzoate, and soaked for approximately 16 h.
Diffraction data were measured at the National
Synchrotron Light Source at Brookhaven National
Laboratories (NSLS, beamline X29A, Upton, NY).
Crystals of the enzyme–ligand complex were isomor-
phous with those of the zinc-inhibited enzyme21 and
belonged to space group P61 with unit cell parameters
1
3.1.8. 3-Chloro-4-(heptyloxy)benzoic acid (13). H NMR
(CDCl3, 500 MHz) d 0.89 (t, J = 6.9 Hz, 3H), 1.29–1.39
(m, 6H), 1.48–1.51 (m, 2H), 1.83–1.89 (m, 2H), 4.09 (t,
J = 6.5 Hz, 2H), 7.11–7.13 (m, 1H), 6.94 (d, J = 8.73,
1H), 7.95–7.97 (dd, 1H), 8.10 (d, J = 1.0 Hz, 1H); HRMS
m/z calcd for C14H19O3Cl 270.1101; found: 270.1102.
˚
˚
a = b = 101.0 A, c = 122.7 A (two monomers in the
asymmetric unit). Data were indexed and merged using
HKL2000.30 The structure of zinc-inhibited LpxC (PDB
entry 1P42),21 excluding all zinc ions, solvent, and fatty
acid molecules, was used as a search probe in molecular
replacement calculations using AMoRe.31 Initial elec-
tron density maps showed that the ligand 3-(heptyl-
oxy)benzoate was bound in the hydrophobic tunnel.
Iterative cycles of refinement and model building were
performed with CNS32 and O,33 respectively, to improve
the structure as monitored by Rfree. Atomic coordinates
of solvent molecules and 3-(heptyloxy)benzoate were
added during the last stages of refinement of the
enzyme–ligand complex. Data collection and refinement
3.1.9. 3-(Heptyloxy)benzoic acid (14). 1H NMR (CDCl3,
500 MHz) d 0.88 (t, 3H), 1.28–1.36 (m, 8H), 1.77–1.80
(m, 2H), 3.99 (t, 2H), 7.11–7.13 (m, 1H), 7.34 (t, 1H),
7.58–7.59 (d, 1H), 7.66–7.67 (dd, 1H); 13C NMR
(CDCl3, 500 MHz) d 14.1, 22.6, 26.0, 29.0, 29.2, 31.8,
68.3, 115.1, 120.9, 122.4, 129.5, 130.3, 159.5, 170.73;
HRMS m/z calcd for C14H20O3 236.1412; found:
236.1413.
3.2. Ligand affinity measurements
Ligand binding to A. aeolicus LpxC was assayed by iso-
thermal titration calorimetry29 performed at 30 °C on an
isothermal microcalorimeter from Microcal, Inc.
(Northampton, MA). The enzyme was stripped of all
metal ions by dialysis against 1.0 mM EDTA in
25 mM Hepes (pH 7.0), 0.1 M NaCl at room tempera-
ture for P4 h. The EDTA was then removed by exten-
sive dialysis against EDTA-free buffer and the enzyme
was reconstituted to a 1:1 Zn2+/LpxC ratio by the addi-
tion of ZnSO4. A colorimetric assay employing 4-(2-
pyridylazo)-resorcinol was used to determine Zn2+
concentrations and to verify the preparation of apo
and 1:1-reconstituted LpxC as described by Jackman
and colleagues.7 The calorimeter cell contained either
ꢀ40 or ꢀ60 lM enzyme, and the syringe contained
250 or 400 lM of ligand. A series of 30 injections of
8 lL each were performed at 180-s intervals. Titrations
of aliphatic compounds into buffer were also performed
as control experiments using identical conditions.
Data were fit to a single binding site model using
Origin (v. 2.9, Microcal, Inc.). In cases where DMSO
was necessary as a carrier solvent to facilitate solubiliza-
tion of the aliphatic compound of interest, equal
amounts of DMSO (vol %) were included in the protein
solution.
Table 2. Data collection and refinement statistics
Complex
LpxC–compound 14
˚
Resolution range (A)
50.0–2.25
99618/33203
98.8/99.8
0.119/0.508
10.4/2.5
4298
Reflections (measured/unique)
Completeness (%) (overall/outer shell)
Rmergea(overall/outer shell)
hI/ri (overall/outer shell)
Protein atoms (no.)b
Solvent atoms (no.)b
179
Metal ions (no.)b
5
34
Ligand atoms (no.)b
Reflections used in refinement (work/free) 31553/1650
R/Rfree(overall)c
R/Rfree(outer shell)c
rms deviations
0.198/0.227
0.242/0.301
˚
Bonds (A)
0.006
1.2
Angles (deg)
Proper dihedral angles (deg)
Improper dihedral angles (deg)
23.4
0.8
a Rmerge = R|Ij ꢁ hIji|/RIj, where Ij is the observed intensity for reflec-
tion j and hIji is the average intensity calculated for reflection j from
replicate data.
b Per asymmetric unit.
c R = R| |Fo| ꢁ |Fc | |/ R|Fo |, where R and Rfree are calculated using the
working and test reflection sets, respectively.

H. Shin et al. / Bioorg. Med. Chem. 15 (2007) 2617–2623
2623
Raetz, C. R. H.; Stover, C. K.; Witte, P. R.;
Yabannavar, A.; Zhu, S. J. Med. Chem. 2002, 45,
3112–3129.
statistics are reported in Table 2. Atomic coordinates of
the LpxC–3-(heptyloxy)benzoate complex have been
deposited in the Protein Data Bank with accession code
2O3Z.
15. Clements, J. M.; Coignard, F.; Johnson, I.; Chandler, S.;
Palan, S.; Waller, A.; Wijkmans, J.; Hunter, M. G.
Antimicrob. Agents Chemother. 2002, 46, 1793–1799.
16. Jackman, J. E.; Fierke, C. A.; Tumey, L. N.; Pirrung, M.;
Uchiyama, T.; Tahir, S. H.; Hindsgaul, O.; Raetz, C. R.
H. J. Biol. Chem. 2000, 275, 11002–11009.
Acknowledgments
We thank the National Synchrotron Light Source
(NSLS) for beamline access. We also thank Dr.
Ronen Marmorstein for the use of his instrument
for isothermal titration calorimetry experiments with
compound 14 and 4-heptylaniline. This work was
supported by the National Institutes of Health Grant
GM49758.
17. Pirrung, M. C.; Tumey, L. N.; Raetz, C. R. H.; Jackman,
J. E.; Snehalatha, K.; McClerren, A. L.; Fierke, C. A.;
Gantt, S. L.; Rusche, K. M. J. Med. Chem. 2002, 45,
4359–4370.
18. Pirrung, M. C.; Tumey, L. N.; McClerren, A. L.; Raetz, C.
R. H. J. Am. Chem. Soc. 2003, 125, 1575–1586.
19. Li, X.; Uchiyama, T.; Raetz, C. R. H.; Hindsgaul, O. Org.
Lett. 2003, 5, 539–541.
20. McClerrin, A. L.; Endsley, S.; Bowman, J. L.; Andersen,
N. H.; Guan, Z.; Rudolph, J.; Raetz, C. R. H. Biochem-
istry 2005, 44, 16574–16583.
References and notes
21. Whittington, D. A.; Rusche, K.; Shin, H.; Fierke, C. A.;
Christianson, D. W. Proc. Natl. Acad. Sci. U.S.A. 2003,
100, 8146–8150.
22. Coggins, B. E.; Li, X.; McClerren, A. L.; Hindsgaul, O.;
Raetz, C. R. H.; Zhou, P. Nat. Struct. Biol. 2003, 10, 645–
651.
23. Coggins, B. E.; McClerren, A. L.; Jiang, L.; Li, X.;
Rudolph, J.; Hindsgaul, O.; Raetz, C. R. H.; Zhou, P.
Biochemistry 2005, 44, 1114–1126.
24. Gennadios, H. A.; Whittington, D. A.; Li, X.; Fierke,
C. A.; Christianson, D. W. Biochemistry 2006, 45, 7940–
7948.
1. Parrillo, J. E. Ann. Intern. Med. 1990, 113, 227–242.
2. Parrillo, J. E. N. Eng. J. Med. 1993, 328, 1471–1477.
3. Raetz, C. R. H. J. Bacteriol. 1993, 175, 5745–5753.
4. Rick, P. D.; Raetz, C. R. H. In Endotoxin in Health and
Disease; Brade, H., Opal, S. M., Vogel, S. N., Morrison,
D. C., Eds.; Marcel Dekker, Inc: New York, 1999; pp 283–
304.
5. Wyckoff, T. J. O.; Raetz, C. R. H.; Jackman, J. E. Trends
Microbiol. 1998, 6, 154–159.
6. Raetz, C. R. H.; Whitfield, C. Ann. Rev. Biochem. 2002,
71, 635–700.
7. Jackman, J. E.; Raetz, C. R. H.; Fierke, C. A. Biochem-
istry 1999, 38, 1902–1911.
8. Anderson, M. S.; Bulawa, C. E.; Raetz, C. R. H. J. Biol.
Chem. 1985, 260, 15536–15541.
9. Anderson, M. S.; Robertson, A. D.; Macher, I.; Raetz, C.
R. H. Biochemistry 1988, 27, 1908–1917.
25. Hernick, M.; Gennadios, H. A.; Whittington, D. A.;
Rusche, K. M.; Christianson, D. W.; Fierke, C. A. J. Biol.
Chem. 2005, 280, 16969–16978.
26. Roy, B. C.; Banerjee, A. L.; Swanson, M.; Xiao, G. J.;
Haldar, M. K.; Mallik, S.; Srivastava, D. K. J. Am. Chem.
Soc. 2004, 126, 13206–13207.
10. Anderson, M. S.; Bull, H. G.; Galloway, S. M.; Kelly, T.
M.; Mohan, S.; Radika, K.; Raetz, C. R. H. J. Biol. Chem.
1993, 268, 19858–19865.
11. Young, K.; Silver, L. L.; Bramhill, D.; Cameron, P.;
Eveland, S. S.; Raetz, C. R. H.; Hyland, S. A.; Anderson,
M. S. J. Biol. Chem. 1995, 270, 30384–30391.
12. Onishi, H. R.; Pelak, B. A.; Gerckens, L. S.; Silver, L. L.;
Kahan, F. M.; Chen, M.-H.; Patchett, A. A.; Galloway, S.
M.; Hyland, S. A.; Anderson, M. S.; Raetz, C. R. H.
Science 1996, 274, 980–982.
27. Jude, K. M.; Banerjee, A. L.; Haldar, M. K.; Manokaran,
S.; Roy, B. C.; Mallik, S.; Srivastava, D. K.; Christianson,
D. W. J. Am. Chem. Soc. 2006, 128, 3011–3018.
28. Gennadios, H. A.; Christianson, D. W. Biochemistry 2006,
45, 15216–15223.
29. Fisher, H. F.; Singh, N. Methods Enzymol. 1995, 259, 194–
221.
30. Otwinowski, Z.; Minor, W. Methods Enzymol. 1997, 276,
307–326.
31. Navaza, J. Acta Crystallogr. A 1994, 50, 157–163.
13. Chen, M.-H.; Steiner, M. G.; de Laszlo, S. E.; Patchett, A.
A.; Anderson, M. S.; Hyland, S. A.; Onishi, H. R.; Silver,
L. L.; Raetz, C. R. H. Bioorg. Med. Chem. Lett. 1999, 9,
313–318.
14. Kline, T.; Andersen, N. H.; Harwood, E. A.; Bowman,
J.; Malanda, A.; Endsley, S.; Erwin, A. L.; Doyle, M.;
Fong, S.; Harris, A. L.; Mendelsohn, B.; Mdluli, K.;
32. Brunger, A. T.; Adams, P. D.; Clore, G. M.; DeLano, W.
¨
L.; Gros, P.; Grosse-Kunstleve, R. W.; Jiang, J.-S.;
Kuszewski, J.; Nilges, M.; Pannu, N. S.; Read, R. J.;
Rice, L. M.; Simonson, T.; Warren, G. L. Acta Crystal-
logr. D 1998, 54, 905–921.
33. Jones, T. A.; Zou, J.-Y.; Cowan, S. W.; Kjeldgaard, M.
Acta Crystallogr. A 1991, 47, 110–119.