Novel HldE-K inhibitors leading to attenuated gram negative bacterial virulence

Article abstract of DOI:10.1021/jm301499r

We report here the optimization of an HldE kinase inhibitor to low nanomolar potency, which resulted in the identification of the first reported compounds active on selected E. coli strains. One of the most interesting candidates, compound 86, was shown to inhibit specifically bacterial LPS heptosylation on efflux pump deleted E. coli strains. This compound did not interfere with E. coli bacterial growth (MIC > 32 μg/mL) but sensitized this pathogen to hydrophobic antibiotics like macrolides normally inactive on Gram-negative bacteria. In addition, 86 could sensitize E. coli to serum complement killing. These results demonstrate that HldE kinase is a suitable target for drug discovery. They also pave the way toward novel possibilities of treating or preventing bloodstream infections caused by pathogenic Gram negative bacteria by inhibiting specific virulence factors.

Full text of DOI:10.1021/jm301499r

Article
pubs.acs.org/jmc
Novel HldE‑K Inhibitors Leading to Attenuated Gram Negative
Bacterial Virulence
Nicolas Desroy,†,∥ Alexis Denis,†,⊥ Chrystelle Oliveira,† Dmytro Atamanyuk,† Sophia Briet,†
Fabien Faivre,† Geraldine LeFralliec,† Yannick Bonvin,† Mayalen Oxoby,† Sonia Escaich,‡,#
́
Stephanie Floquet,‡ Elodie Drocourt,‡ Vanida Vongsouthi,‡ Lionel Durant,‡,∞ Franco
̧
is Moreau,‡
́
Theodore B. Verhey,§ Ting-Wai Lee,§ Murray S. Junop,§ and Vincent Gerusz*,†
‡
†Medicinal Chemistry and Biology, Mutabilis, 102 Avenue Gaston Roussel, 93230 Romainville, France
§Department of Biochemistry and Biomedical Sciences and Michael G. DeGroote Institute for Infectious Disease Research, McMaster
University, 1280 Main Street West, Hamilton, Ontario, L8S 4K1, Canada
S
* Supporting Information
ABSTRACT: We report here the optimization of an HldE kinase inhibitor to low nanomolar potency, which resulted in the
identification of the first reported compounds active on selected E. coli strains. One of the most interesting candidates, compound
86, was shown to inhibit specifically bacterial LPS heptosylation on efflux pump deleted E. coli strains. This compound did not
interfere with E. coli bacterial growth (MIC > 32 μg/mL) but sensitized this pathogen to hydrophobic antibiotics like macrolides
normally inactive on Gram-negative bacteria. In addition, 86 could sensitize E. coli to serum complement killing. These results
demonstrate that HldE kinase is a suitable target for drug discovery. They also pave the way toward novel possibilities of treating
or preventing bloodstream infections caused by pathogenic Gram negative bacteria by inhibiting specific virulence factors.
few HldE inhibitors have recently been disclosed, although
without any reported bacterial data.9−11 This antivirulence
target remains therefore largely unexploited at this time, since
there are no drugs on the market or in advanced clinical phases
acting via this mode of action.
INTRODUCTION
■
During the course of a bacterial infection, virulence factors are
responsible for the pathogenicity exhibited toward the host.1,2
In Gram-negative bacteria, the full outer membrane lip-
opolysaccharide (LPS) is an important virulence factor, since
it confers the prokaryotic cell resistance toward innate immune
system components. When LPS lacks the inner core
oligosaccharide heptoses, the bacteria produce a so-called
deep-rough phenotype involving a truncated or “Re-LPS”. They
are still viable and capable of gut colonization; however, they
are unable to mount a productive infection because of their
increased sensitivity toward the bactericidal effect of the host
complement.3,4 Since the heptose synthetic pathway described
in Scheme 1 is conserved among Gram negative species and has
not been described in eukaryotic cells, it represents an attractive
target to attenuate bacterial virulence.5 In addition, this novel
anti-infective approach would not affect the commensal flora
and would therefore be less likely to generate resistance in this
environment than traditional antimicrobial chemotherapy.6
In Escherichia coli, HldE constitutes the second enzyme
involved in the heptose synthetic pathway. It is a bifunctional
cytoplasmic protein comprising both carbohydrate kinase and
adenylyltransferase domains.7 Well conserved among Gram-
negative bacteria with the notable exceptions of Acinetobacter,
Moraxella, and Chlamydia,8 HldE shares little similarity with
human enzymes (40% at the most with human ribokinase). A
At Mutabilis, a high-throughput screening campaign led to
the discovery of compound 1 as an inhibitor of the kinase
activity of HldE (HldE-K) with an IC50 of 51 μM (Figure 1).
Optimization efforts culminated in compound 2, displaying an
IC50 of 0.11 μM on HldE-K.10 However, no satisfactory
antivirulence activity was observed on bacterial strains, even in
the absence of efflux-pumps. A second screening campaign led
to the identification of compound 3, another HldE-K hit with
an IC50 of 6 μM. In contrast to the previous inhibitors 1 and 2,
this new compound is not negatively charged, therefore
affording a better potential for penetrating the neutral inner
membrane of Gram-negative pathogens. Herein we describe the
structure−activity relationship (SAR) of this novel series of
compounds active against HldE-K. We also report for the first
time specific inhibition of LPS heptosylation on selected E. coli
strains leading to attenuation of bacterial virulence.
Received: October 16, 2012
Published: February 14, 2013
© 2013 American Chemical Society
1418
dx.doi.org/10.1021/jm301499r | J. Med. Chem. 2013, 56, 1418−1430

Journal of Medicinal Chemistry
Article
Scheme 1. Synthesis of ADP-L-glycero-β-D-manno-heptose from Sedoheptulose 7-Phosphate and Incorporation into LPS in E.
coli
Figure 1. HldE-K inhibitors.
CHEMISTRY
Scheme 2. Preparation of Thiazol-2-ylmethylamine Building
Blocks
■
a
One of the strategies used to synthesize thiazol-2-ylmethyl-
amine analogues began with silylation of thiazole 4 followed by
formylation at position 2 and reductive amination, which
afforded amine 6 (Scheme 2). Another procedure resorted to
the oxidation of available 2-methylbenzothiazoles 7 and 8 by
selenium dioxide and subsequent reductive amination to deliver
amino intermediates 9 and 10. 5-Hydroxybenzothiazole
derivative 11 was prepared by boron tribromide demethylation
of 10. Palladium-promoted cyclocondensation12 of 12 with 13
followed by Boc-deprotection under acidic conditions yielded
14. Replacing 13 by 4-chloro-2-iodoaniline in a similar
condensation afforded (6-chloro-1,3-benzothiazol-2-yl)-
methylamine used in the preparation of 15.
One method used to build the triazine template was based on
bromination of 1-(2,6-dimethoxyphenyl)ethanone 16 followed
by formation of a dimorpholine ketoaminal able to cyclize with
aminoguanidine to yield 3-aminotriazine 18 (Scheme 3).13
A
subsequent Sandmeyer reaction afforded 3-chlorotriazine 19,
which served as the building block for nucleophilic
substitutions with the corresponding thiazol-2-ylmethylamine
derivatives to provide 15, 20, 22, 23, 26, and 28. Compounds 3
and 30−48 (Tables 1 and 2) and precursors of 49 (Table 1)
and 50 (Table 2) were obtained in similar conditions using
appropriate amines, either commercially or synthesized
according to known procedures.14 Compound 51 (Table 1)
was prepared similarly from 19 and 6 after acid-mediated
deprotection of its silylated precursor. Saponification of 20
a
Reagents and conditions: (a) (i) TBDMSCl, imidazole, CH2Cl2, rt;
(ii) n-BuLi, DMF, THF, −78 to −15 °C, 67% over two steps; (b) (i)
H2NOH·HCl, MeONa, EtOH, rt; (ii) Zn, AcOH, 0 °C to rt, 91% over
two steps; (c) (i) SeO2, 1,4-dioxane, 105 °C; (ii) H2NOH·HCl,
MeONa, EtOH, rt; (iii) Zn, AcOH, rt, 42% over three steps; (d) BBr3,
DCM, −78 °C to rt, 37%; (e) (i) Pd2(dba)3, dppf, CaO, acetonitrile,
60 °C; (ii) HCl (37% aq), EtOAc, 0 °C, 16% over two steps.
1419
dx.doi.org/10.1021/jm301499r | J. Med. Chem. 2013, 56, 1418−1430

Journal of Medicinal Chemistry
Article
a
Scheme 3. Synthesis of Compounds Represented in Tables 1 and 2
a
Reagents and conditions: (a) phenyltrimethylammonium tribromide, THF, rt, 86%; (b) (i) morpholine, THF, 67 °C; (ii) aminoguanidine
bicarbonate, MeOH, AcOH, 67 °C, 31% over two steps; (c) t-BuONO, CuCl2, ACN, DCM, 65 °C, 29%. (d) For 20: ethyl [(1,3-benzothiazol-2-
ylmethyl)amino]acetate,14 DIPEA, ACN, 80 °C, 8%. For 22: 10, DIPEA, ACN, 80 °C, 72%. For 23: 11, ACN, 85 °C, 70%. For 26: 9, ACN, 85 °C,
46%. For 28: 14, DIPEA, ACN, 80 °C, 43%. (e) BrCH2COOEt, K2CO3, THF, 50 °C, 27%; (f) LiOH, THF, H2O, rt. For 21: 44%. For 25: 65%. (g)
AcCl, pyridine, DCM, 0 °C, 32%; (h) DIBAL-H (1 M in toluene), DCM, −78 to −45 °C, 48%.
afforded 21, and compound 49 was obtained likewise.
Alkylation of 23 with ethyl bromoacetate followed by
hydrolysis under basic conditions afforded 25. Acetylation of
26 provided 27, while a similar reaction on the amino derivative
obtained by substitution of 19 with 6-aminobenzothiazolylme-
thylamine prepared as 9 yielded 50. Compound 29 was
synthesized by DIBAL-H reduction of 28.
nation of 6-azauracil16 followed by subsequent substitution with
cis-2,6-dimethylpiperidine afforded 58, which was transformed
into 60 under classical conditions with 10. Compound 61 was
obtained in the same manner as 60, using 6-azauracil and 2-
piperidinemethanol.
The preparation of some 5-aryl-1,2,4-triazine derivatives
involved another pathway using selenium dioxide oxidation of
acetophenones to glyoxal intermediates followed by cyclo-
condensation with thiosemicarbazide (Scheme 5).17 In this
manner, 1-(2-methoxynaphthalen-1-yl)ethanone 6218 was
transformed into triazin-3-ol, which was submitted to
chloration of position 3 with phosphoryl chloride to provide
63. Substitution under standard conditions with amine 10
afforded 64. Compounds 66 and 72−75 (Table 3) were
obtained similarly. Demethylation of 64 under boron
tribromide conditions yielded 65. The same procedure was
used to obtain 76. The meta-position of the left-part aryl group
of 66 was modified under palladium-catalyzed conditions:
triethylsilane reduction of the meta-bromide moiety yielded 67,
while Buchwald coupling with acetamide afforded 68 and
Suzuki coupling with the corresponding boronic acids provided
69 and 77 (Table 3). In a similar way, 74 was transformed to
78−80 (Table 3). Compound 70, which was prepared by
standard substitution of the corresponding chlorotriazine14 with
10, was reacted under basic conditions with 2-chloro-N,N-
dimethylethanamine to afford 71.
Compounds 54 and 59−61 (Table 3) incorporating a
nonaromatic left part were prepared according to Scheme 4.
Scheme 4. Synthesis of 5-(Alkyl/piperazinyl)-1,2,4-triazine
a
Derivatives Represented in Table 3
a
Reagents and conditions: (a) (i) i-PrLi (0.7 M in pentane), THF,
−78 °C; (ii) DDQ, toluene, rt, 26% over two steps; (b) (i) mCPBA,
DCM, rt; (ii) 10, DIPEA, ACN, 85 °C, 14% over two steps; (c) t-
BuONO, CuCl2, acetonitrile, 70 °C, 41%; (d) (i) POCl3, TEA,
dioxane, 110 °C; (ii) cis-2,6-dimethylpiperidine, ACN, 0 °C, 3% over
two steps.
The synthesis of the compounds represented in Table 4 is
illustrated in Scheme 6. O-Alkylation of 81, obtained by
standard demethylation of 72, afforded derivatives 82 and 83.
Reduction of ester 83 under hydride conditions provided
alcohol 84, while its hydrolysis with lithium hydroxide led to
carboxylic acid 85 that was derivatized to amides 86 and 95
(Table 4) under classical peptide coupling conditions.
Phthalimide protection of 10 followed by demethylation with
boron tribromide yielded intermediate 88, which was O-
Isopropyllithium addition at position 5 of triazine 52 with
subsequent 2,3-dichloro-5,6-dicyano-1,4-benzoquinone oxida-
tion afforded 53. Thioether oxidation followed by substitution
with amine 10 under basic conditions yielded 54. Chloro-
triazine 56 was prepared by Sandmeyer chlorination of
intermediate 5515 and was then transformed to 59 under
previously described substitution conditions with 10. Chlori-
1420
dx.doi.org/10.1021/jm301499r | J. Med. Chem. 2013, 56, 1418−1430

Journal of Medicinal Chemistry
Article
a
Scheme 5. Synthesis of 5-Aryl-1,2,4-triazine Derivatives Represented in Table 3
a
Reagents and conditions: (a) (i) SeO2, dioxane, H2O, 105 °C; (ii) semicarbazide·HCl, K2CO3, EtOH·H2O, 105 °C, 30% over two steps; (b) POCl3,
DMF, CHCl3, 50 °C, 60%; (c) 10, DIPEA, ACN, 85 °C, 56%; (d) BBr3, DCM, −78 °C to rt, 65%; (e) triethylsilane, PdCl2(dppf)2·DCM, CsF, 1,4-
dioxane, 19%; (f) CH3CONH2, CsF, xantphos, Pd2(dba)3, 1,4-dioxane, 100 °C, 14%; (g) 1H-pyrazol-3-ylboronic acid, Na2CO3, Pd(PPh3)4, DMF,
100 °C, 51%; (h) 2-chloro-N,N-dimethylethanamine hydrochloride, NaHCO3, NaI, BuOH, 110 °C, 27%.
alkylated with N-(2-chloroethyl)-N,N-dimethylamine, depro-
tected with methylhydrazine, and substituted with 3-chloro-5-
(2,6-dichlorophenyl)-1,2,4-triazine to provide 90. Intermediate
88 was also submitted to Mitsunobu conditions with tert-butyl
2-hydroxyethylcarbamate to afford after phthalimide depro-
tection amine 91, which was substituted with 3-chloro-5-(2,6-
dichlorophenyl)-1,2,4-triazine and deprotected under acidic
conditions to afford 92. Acylation of this amine under classical
peptide coupling conditions provided compounds 93 and 94
(Table 4).
thiazole moiety (32 and 51) were found detrimental to
potency, suggesting either a steric hindrance or a lipophilic
environment in this area. Cyclization of the two methyl groups
in benzothiazole or tetrahydrobenzothiazole substituents (34
and 35) afforded a 3-fold improvement in potency, confirming
the beneficial effect of lipophilic groups in this location.
Variation of the orientation of the hydrogen-bond acceptor
nitrogen such as in pyridyl 33 or quinolyl 36 or removal of this
hydrogen-bond acceptor in benzothiophenyl 37 resulted in a
complete loss of potency. Increasing the hydrophilic character
of the bicycle such as in benzoxazolyl 38 and in benzimidazolyl
39 or increasing the steric hindrance in 1-methyl-2-
benzimidazolyl 40 also proved detrimental. Replacing the
nitrogen-containing amine within the linker by an oxygen-
containing ether (41) led to a 7-fold loss in potency.
Introducing a short substituent to this nitrogen was tolerated
(42 vs 3 and 21 vs 34), although increasing its size proved
detrimental (49 and 43). Substitutions on the methylene linker
bridging the amine and the heterocycle also led to a decrease in
potency (44 and 45).
The mechanism of action of compound 34 was investigated
in order to rule out promiscuous inhibition.20 Inhibition
kinetics of 34 were linear and were not dependent on the order
of reagent addition or on enzyme/inhibitor preincubation. The
inhibition was also unaltered by Triton-X100 (100 μM to 1
mM) or D-manno-heptose 7-phosphate substrate (up to 15 km).
Finally, as shown in Figure 3, compound 34 behaved as a
competitive reversible inhibitor with respect to ATP (Ki = 2.5
μM). These data suggest that the inhibition specifically takes
place in the nucleotide-binding site, which was further
substantiated by the cocrystal structures of B. cenocepacia
HldA with two of the analogues from this series.19
RESULTS AND DISCUSSION
■
Optimization of compound 3 was initiated using classical SAR-
driven medicinal chemistry approaches to increase HldE-K
potency. Later in the project, the procurement of cocrystal
structures of some of these inhibitors with HldA of Burkholderia
cenocepacia, the homologous kinase sharing 49% sequence
identity and 62% sequence similarity with E. coli HldE-K,
allowed some rationalization of our results.19
To assess the importance of the 1,2,4-triazine template, a
number of other heterocyclic scaffolds were initially synthesized
(Figure 2). Since all of these rescaffolding attempts led to a 5-
to 20-fold loss of potency against HldE-K (data not shown), the
1,2,4-triazine core was maintained for further exploration.
Various right-part groups were then evaluated to establish
SAR in this area (Table 1). Replacing the thiazolyl group by a
furanyl moiety led to a complete loss of potency, while the
pyrazolyl analogue exhibited a 4-fold loss of potency (30 and
31 vs 3). This might suggest the importance of a hydrogen-
bond acceptor nitrogen in the right-part heterocycle, which was
subsequently confirmed by cocrystal structures.19 Substituents
with hydrogen-bond acceptors or donors at position 5 of the
1421
dx.doi.org/10.1021/jm301499r | J. Med. Chem. 2013, 56, 1418−1430

Journal of Medicinal Chemistry
Article
a
Scheme 6. Synthesis of Compounds Represented in Table 4
a
Reagents and conditions: (a) iodoacetamide, K2CO3, THF, 50 °C, 40%; (b) ethyl bromoacetate, K2CO3, acetone, 60 °C, 43%; (c) LiOH, THF,
H2O, rt, 77%; (d) DIBAL-H, DCM −78 to −30 °C, then NaBH4, MeOH, rt, 52%; (e) ethanolamine, EDAC, 4-DMAP, THF, DCM, rt, 31%; (f)
phthalic anhydride, DIPEA, pTSA, toluene, 70 °C, 69%; (g) BBr3, DCM, −78 °C to rt, 87%; (h) N-(2-chloroethyl)-N,N-dimethylamine·HCl,
K2CO3, acetone, 60 °C, 56%; (i) MeNHNH2, ethanol, 80 °C, 34%; (j) 3-chloro-5-(2,6-dichlorophenyl)-1,2,4-triazine, DIPEA, ACN, 85 °C; 31%;
(k) (i) tert-butyl 2-hydroxyethylcarbamate, DIAD, PPh3, THF, 0 °C to rt, 62%; (ii) MeNHNH2, EtOH, 80 °C, 62%; (l) 3-chloro-5-(2,6-
dichlorophenyl)-1,2,4-triazine, DIPEA, ACN, 85 °C; 49%; (m) HCl (4 N in dioxane), MeOH, 0 °C to rt, 58%; (n) (i) BocNHCH2COOH, EDAC,
4-DMAP, THF, DCM, 50 °C; (ii) HCl (4 M in dioxane), THF, MeOH, 0 °C to rt, 7% over two steps.
Figure 2. Central rescaffolding of 3.
Having determined a 3-fold improvement in potency
afforded by benzothiazole 34 over the initial hit 3, we set out
to study its substitution potential (Table 2). A variety of
different groups were tolerated at position 5, either lipophilic
(46, 47, and 22) or hydrophilic (23 and 25−27). The most
potent compounds 23 and 25 were 5−8 times more active than
the unsubstituted benzothiazole 34. This could be rationalized
by a hydrogen bond between the hydroxyl substituent of 23
and the side chain amide group of a conserved asparagine
(Asn202 in B. cenocepacia HldA and Asn196 in E. coli HldE-K).
Similarly, the cocrystal structure of B. cenocepacia HldA with 25
displays a favorable dipole−dipole interaction between its
benzothiazole carboxyl substituent and the side chain amide
group of the same conserved asparagine.19 Position 6 of
benzothiazole was less permissive, since groups bulkier than
hydroxyl were detrimental to potency (28, 29, 50 vs 48).
1422
dx.doi.org/10.1021/jm301499r | J. Med. Chem. 2013, 56, 1418−1430

Journal of Medicinal Chemistry
Article
Table 1. Influence of the Right-Part Substitution on E. coli HldE-K Inhibition
The left-part substitution was then investigated with 5-
(methoxy or hydroxy)benzothiazole as the right-part moiety
(Table 3). Alkyl groups (54 and 59) and nonaromatic
heterocycles (60 and 61) were inactive, reflecting the
importance of an aromatic moiety in this location. Of all the
aromatic groups evaluated, the naphthyl derivatives 64 and 65
proved the most active against the enzyme, reaching low
nanomolar potency. This suggests a significant π-stacking
contribution, supported by the proximity between the
benzothiazole and the left-part aromatic groups in the bent
inhibitor conformation observed within the cocrystal struc-
tures.19 High-throughput screening of the analogues of our hit
revealed that the aromatic group had to be ortho-substituted in
order to be active against the enzyme. An ortho, ortho prime
substitution is even more beneficial as shown by the 10-fold
gain in potency by dichloro-substituted 72 versus monochloro-
substituted 67. The ortho substitution adjusts the left-part
aromatic moiety in a perpendicular way to the triazine core,
1423
dx.doi.org/10.1021/jm301499r | J. Med. Chem. 2013, 56, 1418−1430

Journal of Medicinal Chemistry
Article
of the benzothiazole ring might also account for the relative
SAR permissiveness of substitution exhibited at this location.
The discovery of these low nanomolar inhibitors of HldE-K
prompted us to investigate their antivirulence effects. Bacteria
with heptose-deficient LPS (hldE-deleted strains) display
increased sensitivity to hydrophobic antibiotics (Table 5),
which is possibly due to a reduction in the hydrophilic barrier
presented by a Re-LPS versus an intact LPS. We therefore
investigated the potential sensitization of efflux-pump deleted
strains E. coli K12 Δ-acrAB, ΔtolC and E. coli K1 Δ-tolC to
erythromycin by our inhibitors (Table 6). In both strains,
deletion of the major efflux pump AcrAB/TolC (in part or
entirely) leads to an increased susceptibility toward amphiphilic
and lipophilic antibiotics.22 The concentration of erythromycin
chosen for this study was at one-fourth of its minimal inhibitory
concentration (MIC) on these selected strains. The intrinsic
antibacterial effects of these compounds were also assessed
without erythromycin. In the most promising cases where good
sensitization of E. coli K1 Δ-tolC to erythromycin could be
observed without any inhibitor-induced antibacterial effects,
further cytotoxicity studies and LPS electrophoresis assays were
performed to determine the capacities of those compounds to
inhibit specifically the biosynthesis of heptose-containing LPS.
Most of the nanomolar inhibitors in Table 3 were inactive in
the sensitization assays. As a notable exception, 65 sensitized E.
coli K1 Δ-tolC to erythromycin at 2 μg/mL (Table 6).
However, its intrinsic antibacterial activity against E. coli K12 Δ-
acrAB, Δ-tolC and its cytotoxicity on eukaryotic cells cast
doubts on a sensitization specifically achieved by heptosylation
inhibition, and this compound was therefore not investigated
any further.
Since the dichloro derivative 72 exhibited a 5-fold improve-
ment in potency over the dimethoxy derivative 22, we re-
examined the substitutional effects at position 5 of the
benzothiazole with the hope of increasing the antibacterial
activity (Table 4). As already observed with compounds 23 vs
22, the replacement of the terminal methoxyl group by a
hydroxy group afforded a 1.7-fold gain in potency (81 vs 72).
Chain extensions with a variety of hydrogen-bond donors and
acceptors, whether neutral (82, 84, 86, and 94) or charged (85,
90, 92, 93, and 95), were well tolerated with potencies
remaining in the nanomolar range. The most potent inhibitors
(84, 85, and 95) have neutral or negatively charged moieties,
which may interact favorably with the side chain amide group of
the conserved binding pocket asparagine of the enzyme. By
contrast, positively charged moieties as found in 90, 92, or 93
presented a 5- to 10-fold loss in potency. Despite their good
Figure 3. Competition between 34 and ATP for E. coli HldE-K
measured in the presence of 2 μM of the D-manno-heptose 7-
phosphate substrate. This double reciprocal plot (1/(initial velocity) vs
1/[ATP]) displays lines intersecting with the vertical axis at 1/Vmax
.
which could facilitate the π-stacking or Cl−π interactions21
between the left and right aromatic groups of the compound.
Substitutions at the meta position of the left-part phenyl group
with both hydrogen-bond donor (68 and 69) and acceptor
(77) are tolerated. The para position also proved permissive,
with a variety of substituents allowed (70, 71, 74, 75, 79, and
80). Cocrystal structures revealed an expansion of the
nucleotide-binding site of the enzyme in these directions,
which probably accounts for the observed permissiveness of the
meta and para substitutions.
The key points of our potency SAR for this series are
summarized in Figure 4 and were found to be in full agreement
with the previously reported cocrystal structures in B.
cenocepacia HldA.19 The triazine scaffold turned out to be
optimal, which may be accounted for by H-bonding of N1 and
N2 with a canonical water molecule itself bonded to a backbone
NH of Leu 258 and possibly a backbone carbonyl of Asn 294.
Significant π-stacking or Cl−π interactions between the
required left and right aromatic parts of the inhibitors
contribute to their folded conformations and improve their
shape complementarity with the ADP binding site. The
importance of the right-part thiazole nitrogen revealed by our
SAR was confirmed by a H-bond interaction with the backbone
NH donor group of Ser 240. The highly polar enzymatic
environment and the partial solvent exposure at the 5-position
Table 2. Influence of Benzothiazole Substitution on E. coli HldE-K Inhibition
compd
R1
R2
IC50 (μM)
compd
R1
R2
IC50 (μM)
46
47
22
23
25
26
Me
Cl
H
H
H
H
H
H
1.5
27
15
28
29
48
50
NHAc
H
H
7.6
27
89
15
3.4
49
1.6
Cl
OMe
0.67
0.39
0.25
1.4
H
CO2Me
CH2OH
OH
OH
H
OCH2CO2H
NH2
H
H
NHAc
1424
dx.doi.org/10.1021/jm301499r | J. Med. Chem. 2013, 56, 1418−1430

Journal of Medicinal Chemistry
Article
Table 3. Influence of the Left-Part Substitution on E. coli HldE-K Inhibition
potency against the enzyme, the negatively charged inhibitors
showed poor potency against bacteria (Table 6, 85 and 95),
which might reflect difficulty of these compounds crossing
Gram-negative membranes. However, the most potent neutral
1425
dx.doi.org/10.1021/jm301499r | J. Med. Chem. 2013, 56, 1418−1430

Journal of Medicinal Chemistry
Article
Figure 4. Summary of potency SAR for triazine HldE inhibitors.
Table 4. Influence of 5-Benzothiazolyl Substitution on E. coli HldE-K Inhibition
LPS electrophoresis assays were performed with inhibitors
alone to determine their capacities in specifically inhibiting the
biosynthesis of intact LPS. In this assay, 86 started to inhibit
LPS synthesis at 2 μg/mL in E. coli K12 Δ-acrAB, Δ-tolC and at
16 μg/mL in K1 Δ-tolC (Figure 5). These concentrations are in
line with the ones observed in Table 6 that afforded
sensitization effects to erythromycin, thereby demonstrating a
specific mode of action. Potential bacterial sensitization to
complemented serum by 86 was also assessed (Table 7), since
Gram-negative bacteria devoid of heptose-containing LPS are
known to be susceptible to the complement system.3,4 While, as
expected, no antibacterial effects were observed against K12 Δ-
acrAB, Δ-tolC in the absence of serum complement, its
presence inhibited bacterial growth with 4 μg/mL 86. This
assay provided evidence that our inhibitor attenuated the
virulence of this selected Gram-negative strain under simulated
physiological conditions through sensitizing bacteria to killing
by serum complement.
Table 5. Differential of the MIC in μg/mL between HldE E.
coli Mutant and Wild Type Strains
E. coli K1 wild type (μg/mL) E. coli K1 ΔhldE (μg/mL)
erythromycin
telithromycin
novobiocin
rifampicin
32
8
0.5
0.25
0.5
32
8
0.125
2
synercid
>32
and positively charged inhibitors (84, 86, 90, and 93) displayed
for the first time sensitization of E. coli K12 Δ-acrAB, Δ-tolC
and K1 Δ-tolC strains to erythromycin. Compound 93 was
cytotoxic at 18 μg/mL, possibly because of its amphipathic
structure, whereas 84 and 90 displayed antibacterial activities at
16−32 μg/mL on E. coli K12 Δ-acrAB, Δ-tolC by themselves.
Compound 86 was not by itself antibacterial or cytotoxic at 32
μg/mL but could senzitize K12 Δ-acrAB, Δ-tolC to
erythromycin at 2 μg/mL and K1 Δ-tolC to the same antibiotic
at 16 μg/mL. These inhibitors were nevertheless unable to
senzitize E. coli K1 wild-type strain to erythromycin at one-
fourth of its MIC (32 μg/mL).
The ability of these inhibitors to recognize the nucleotide-
binding sites of other closely related enzymes was investigated
to determine their Gram-negative potential spectra as well as
their selectivity (Table 8). Inhibition data for HldE-K enzymes
from K. pneumoniae, S. marcescens, or N. meningitidis revealed
1426
dx.doi.org/10.1021/jm301499r | J. Med. Chem. 2013, 56, 1418−1430

Journal of Medicinal Chemistry
Article
Table 6. In Vitro Evaluation of Selected Compounds
MIC (μg/mL), E. coli K12 Δ-acrAB, Δ-tolC
MIC (μg/mL), E. coli K1 Δ-tolC
IC50 (μM),
E. coli HldE-K
without
erythromycin
+1 μg/mL
erythromycin
without
erythromycin
+1 μg/mL
erythromycin
cytotoxicity,
HepG2 (μg/mL)
a
b
compd
64
65
72
78
76
69
79
71
81
82
85
84
90
95
86
94
93
0.051
0.020
0.130
0.360
0.069
0.053
0.240
0.082
0.075
0.067
0.019
0.018
0.082
0.016
0.049
0.071
0.100
>32
2
>32
0.063
>32
>32
8
>32
>32
>32
>32
>32
>32
>32
>32
>32
>32
>32
>32
>32
>32
>32
>32
>32
>32
2
20
>32
>32
>32
>32
>32
>32
>32
>32
>32
16
>32
>32
>32
>32
>32
>32
>32
>32
>32
32
>32
>32
>32
>32
>32
32
0.5
4
>32
>32
32
16
>32
>32
>32
32
>32
2
>32
16
>32
4
>32
32
4
18
a
b
Erythromycin concentration is at one-fourth of its own MIC against E. coli K12 Δ-acrAB, Δ-tolC. Erythromycin concentration is at one-fourth of
its own MIC against E. coli K1 Δ-tolC.
Table 8. Potency against Related Enzymes
IC50 (μM)
a
b
c
d
a
ECO
KPN
SMA
NM
ECO
e
compd
HldE-K
HldE-K
HldE-K
HldE-K
RK
76
72
65
85
86
0.069
0.130
0.020
0.019
0.430
2.3
0.160
1.2
9.8
13
>300
>300
>300
>300
>300
0.085
0.100
0.150
0.041
0.041
0.076
1.3
1.8
1.9
Figure 5. LPS electrophoresis assays. (A) Positive and negative
controls obtained with (1) LPS of E. coli K1-ΔhldE and (2) LPS of E.
coli K1 wild type. (B) LPS of E. coli K12 Δ-acrAB, Δ-tolC incubated for
5 h at 37 °C with different concentrations (in μg/mL) of 86. The
highlighted concentration corresponds to the lowest dose at which Re-
LPS was observed. (C) LPS of E. coli K1 Δ-tolC incubated for 5 h at
37 °C with different concentrations (in μg/mL) of 86. The highlighted
concentration corresponds to the lowest dose at which Re-LPS was
observed.
0.049
b
a
c
ECO: E. coli. KPN: Klebsiella pneumoniae. SMA: Serratia marcescens.
Neisseria meningitidis. Ribokinase.
d
e
CONCLUSION
■
We report here the optimization of an HldE-K inhibitor
resulting in the development of several compounds with low
nanomolar potency against this necessary LPS biosynthetic
enzyme. This study resulted in the identification of the first
HldE-K inhibitors active against selected E. coli strains. One of
the most interesting candidates, compound 86, was shown to
attenuate Gram-negative bacterial virulence with a specific
mode of action, without displaying intrinsic antibacterial
activity. By inhibiting LPS biosynthesis of E. coli K12 Δ-
acrAB, Δ-tolC, 86 could sensitize this strain to killing by serum
complement under simulated physiological conditions. Selec-
tivity and the potential for acting against a broad Gram-negative
spectrum were also demonstrated for this series of compounds.
These results pave the way toward novel possibilities of treating
or preventing bloodstream infections caused by pathogenic
Gram negative bacteria without affecting their commensal flora,
thus exerting potentially less selective pressure for antibiotic
resistance than conventional antibacterial chemotherapy. The
recently determined cocrystal structures of two of these
inhibitors with HldA of B. cenocepacia represent an attractive
perspective for further optimizing their inhibitory potential on
wild-type bacterial strains. Some subsequent developments are
reported in a following article.23
Table 7. E. coli K12 Δ-acrAB, Δ-tolC Sensitization to Serum
Complement by 86
MIC (μg/mL) against E. coli K12 Δ-acrAB, Δ-tolC
no
with 30% of decomplemented
with 30% of normal
a
a
a
compd FBS
FBS
FBS
86
>32
>32
4
a
Fetal bovine serum.
good conservation of the SAR observed for E. coli. Most potent
compounds against E. coli HldE-K such as 65, 85, or 86 were
also low nanomolar inhibitors against the related enzymes from
K. pneumoniae and S. marcescens but displayed a 100-fold
decrease in potency against N. meningitidis HldE-K. These
results indicate the potential for these inhibitors exhibiting
broad-spectrum potency against Gram-negative pathogens. The
absence of potency against E. coli ribokinase, the most
sequentially similar nonrelated enzyme of HldE-K (sequence
similarity, 38%) at 300 μM, demonstrates the high selectivities
of these inhibitors against ribokinase, which is present in both
prokaryotes and eukaryotes.
1427
dx.doi.org/10.1021/jm301499r | J. Med. Chem. 2013, 56, 1418−1430

Journal of Medicinal Chemistry
Article
HepG2 Cytotoxicity. HepG2 cells (LGC Promochem, ref HB-
8065) in EMEM medium supplemented with fetal bovine serum
(10%), ampicillin (100 μg/mL), and streptomycin (100 μg/mL) were
plated on cell culture plates (Costar 3596, 2 × 104 cells/well) and
incubated overnight at 37 °C in 5% CO2. The culture medium was
then aspirated and replaced by supplemented EMEM (210 μL
including 2% DMSO) containing dilutions of compounds. Plates were
incubated overnight at 37 °C in 5% CO2. Cell viability was determined
by the Celltiter Aqueous One cell proliferation assay (G3581-80,
Promega) according to the manufacturer’s recommendations, using a
Multiskan absorption reader (490 nm).
Inhibition of E. coli LPS Biosynthesis. The compounds being
tested were prepared in deionized water/DMSO (50/50) solutions,
and 25 μL of each solution was added to a sterile culture microtube.
The strains used in this study were E. coli K12 Δ-acrAB, Δ-tolC and E.
coli K1 (018:K1:H7) wild type or Δtolc. The bacteria were isolated on
tryptic soy agar (TSA) overnight. Isolated colonies were cultured in 10
mL of Luria−Bertani (LB) medium at 37 °C to an optical density of
typically 0.15. These exponentially growing bacteria were finally
diluted to 5 × 105 CFU/mL, and 225 μL of each diluted culture was
added to each tube for incubation with the compound at 37 °C for
approximately 5 h, to an optical density of 0.2−0.4.
LPS Extraction. Bacterial cultures were normalized by optical
densities, pelleted, and washed with 1 mL of phosphate buffered saline
(PBS). The pellets were then denatured for 10 min at 95−100 °C in
50 μL of 0.2% sodium dodecyl sulfate (SDS), 1% β-mercaptoethanol,
36% glycerol, 30 mM Tris, pH 7.4, and 0.001% bromophenol blue.
Samples were cooled to room temperature, supplemented with 1.5 μL
of proteinase K at 20 mg/mL, incubated for 1 h at 55 °C, and
centrifuged for 30 min at 13 000 rpm at 25 °C. The resulting
supernatant, which contained LPS, was finally analyzed by SDS−
PAGE.
EXPERIMENTAL SECTION
■
Plasmid Construction, Expression, and Purification of HldE
and RK. Genes encoding HldE were amplified by PCR from the
genomic DNA of E. coli O18:K1 C7 (Robert Debre Hospital, Paris),
́
N. meningitidis 2C43 (Necker Hospital, Paris), S. marcescens
NEM017811 (Necker Hospital, Paris), and K. pneumoniae
ATCC700603 (LGC Promochem) by using Pfu polymerase
(Promega). Similarly, a gene encoding ribokinase was amplified from
́
E. coli K12 MG1655 (Robert Debre Hospital, Paris). PCR fragments
were cloned into pET101 (except for hldE of E. coli: pET100)
expression vector (Invitrogen), and the resulting plasmids were used
to transform E. coli Top10 (Invitrogen). Finally, purified plasmids were
used to transform E. coli BL21 (Invitrogen). Nucleotide sequencing of
the various cloned fragments revealed no mutation. Recombinant
proteins were expressed and purified as described: Exponential-phase
culture (optical density at 600 nm, 0.5−0.7) was induced for
expression by the addition of isopropyl-β-D-glucopyranoside (IPTG)
at a final concentration of 0.5 mM and incubated for further 3 h.
Bacterial cells containing overexpressed recombinant proteins were
harvested by centrifugation, and cell lysis was performed by sonication.
After the removal of cell debris or unlysed cells by centrifugation, Ni-
NTA agarose (Sigma) was added to the supernatant (i.e., soluble
fraction). Binding of the recombinant protein with the gel matrix was
maximized by incubation at 4 °C. His-tagged proteins were eluted by
stepwise increasing concentrations of imidazole (Sigma). The various
fractions were analyzed by SDS−PAGE, and fractions containing the
recombinant protein were pooled and concentrated by ultrafiltration
(Amicon Ultra-15-Millipore). Protein concentration was determined
by Bradford Method (Invitrogen). Protein solutions contained 50%
glycerol and were stored at −20 °C.
HldE-K and Ribokinase Assays. The assay buffer “AB” contained
50 mM Hepes, pH 7.5, 1 mM MnCl2, 25 mM KCl, 0.012% Triton-
X100, 1 mM dithiothreitol (DTT), and 0.1 μM myelin basic protein
(MBP). The following components were added to each well of a white
polystyrene Costar plate up to a final volume of 30 μL: 10 μL of a
inhibitor dissolved in DMSO/water, 50/50, and 20 μL of E. coli HldE
in AB. After 30 min of preincubation at room temperature, 30 μL of
substrate mix in AB was added to each well to a final volume of 60 μL.
This reaction mixture was then composed of 3 nM HldE, 0.2 μM β-
heptose 7-phosphate (H7P, custom synthesis) unless otherwise stated,
and 0.2 μM ATP unless otherwise stated (Sigma) in AB. After 30 min
of incubation at room temperature, 200 μL of revelation mix was
added to each well to a final volume of 260 μL, including the following
constituents at respective final concentrations: 5000 light units/mL
luciferase (Sigma), 30 μM D-luciferin (Sigma), and 100 μM N-
acetylcysteamine (Aldrich). Luminescence intensity was immediately
measured on Luminoskan (Thermofischer) and converted into
percentage inhibition. For IC50 determinations, each inhibitor was
tested at 6−10 different concentrations, and the resulting inhibition
data were fitted to a classical Langmuir equilibrium model using
XLFIT (IDBS). The same procedure was applied to HldE of K.
pneumoniae, N. meningitidis, and S. marcescens at 20, 20, and 10 nM,
respectively. The ribokinase assay was basically identical to the HldE-K
assay except for the following: H7P was replaced by D-ribose (3 μM);
HldE was replaced by ribokinase of E. coli (0.6 nM); the concentration
of ATP was 3 μM instead of 0.2 μM.
Bacterial Sensitization to Erythromycin. All chemicals were
from Sigma unless otherwise stated. From early to log phase
preculture, bacterial suspension was prepared in ca-MHB (Becton-
Dickinson) to obtain a final inoculum of 5 × 105 CFU/mL. Strains
were E. coli K12 MG1655 Δ-acrAB-Δtolc, E. coli K1 (018:K1:H7) wild
type, Δtolc, and ΔhldE. Each strain was grown in two conditions, one
excluding erythromycin and one including erythromycin at a
concentration of 1, 8, and 1 μg/mL, respectively. The MIC of the
compounds being tested was determined against these strains. Growth
was visually inspected in polystyrene 96-well microplates containing
serial dilutions of compounds in 2% DMSO after overnight incubation
at 35 °C. Each well has a final volume of 100 μL.
LPS electrophoresis. Polyacrylamide gels (16% and 4% acrylamide
for separation and concentration, respectively) were prepared, loaded
with 8 μL of each LPS extract, and migrated.
Silver Staining. Gels were incubated overnight in 5% acetic acid/
40% ethanol/deionized water, treated with 1% periodic acid/5% acetic
acid for 15 min, washed with deionized water for 10 min for four
times, and finally incubated for 18 min in the dark with a silver nitrate
solution composed of 56 mL of 0.1 N NaOH, 4 mL of 33% ammoniac,
45 mL of 5% AgNO3 (Tsai and Frasch), and 195 mL of deionized
water. Gels were then washed extensively with deionized water for 30
min and incubated for 10−15 min (up to LPS band apparition) in a
revelation mix composed of 300 mL of deionized water, 300 μL of
36.5% formaldehyde (Fluka), and 100 μL of 2.3 M citric acid. The
revelation was stopped by incubating the gels with 10% acetic acid for
5 min. Gels were finally washed with deionized water, numerized with
a Samsung PL51 camera, and analyzed by the ImageJ software. The
percentage inhibition of LPS heptosylation was defined as the relative
area of the Re-LPS band compared to the total area of the Re-LPS and
core-LPS bands.
Sensitization to Killing by Complement. The Gram-negative
bacterial strain E. coli K12 Δ-acrAB, Δ-tolC was grown on Muller
̈
Hinton agar (MHA) overnight at 35 °C. Five well isolated colonies
from overnight agar growth were then added to 10 mL of cation-
adjusted Muller−Hinton broth (MH2) and incubated at 35 °C with
̈
constant shaking for 2 h to reach the exponential phase. The
suspension was then diluted to obtain an inoculum of 1.5 × 104 CFU/
mL. Viable counts were obtained by plating 10- to 1000-fold dilutions
of the suspension on MHA. A solution of decomplemented fetal
bovine serum (dFBS) was prepared by heating the normal FBS at 56
°C for 30 min. Dilutions of each compound being tested were
prepared in DMSO/water 20/80 in polypropylene plates. The
following components were added on a sterile polystyrene culture
plate to a final volume of 100 μL: 5 μL of a compound dilution in 20%
DMSO, 30 μL of FBS (30% final) or dFBS (30% final) or MH2 (0%
FBS), 65 μL of bacterial suspension at 1.5 × 104 CFU/mL (final
inoculum of 1 × 104 CFU/mL). After incubation for 18−22h at 35 °C,
the MIC of the compound in the presence of FBS 30%, dFBS 30%,
and FBS 0% was determined by visual inspection of bacterial pellets.
1428
dx.doi.org/10.1021/jm301499r | J. Med. Chem. 2013, 56, 1418−1430

Journal of Medicinal Chemistry
Article
General Experimental. All reactions were carried out under inert
(nitrogen or argon) atmosphere unless indicated otherwise. Reagents
and solvents were obtained from commercial sources and were used
without further purification. Celite is a filter aid composed of
diatomaceous silica and is a registered trademark of Celite
Corporation. Analtech silica gel GF and E. Merck silica gel 60 F-254
thin layer plates were used for thin layer chromatography. Flash
chromatography was carried out on Flashsmart Pack cartridge irregular
silica 40−60 μm or spherical silica 20−40 μm. Preparative thin layer
chromatography was carried out on Analtech silica gel GF 1000 μm,
20 cm × 20 cm. Yields refer to purified products and are not
(6) Alekshun, M. N.; Levy, S. B. Targeting virulence to prevent
infection: to kill or not to kill? Drug Discovery Today: Ther. Strategies
2004, 1 (4), 483−489.
(7) McArthur, F.; Andersson, C. E.; Loutet, S.; Mowbray, S. L.;
Valvano, M. A. Functional analysis of the glycero-manno-heptose 7-
phosphate kinase domain from the bifunctional HldE protein, which is
involved in ADP-L-glycero-D-manno-heptose biosynthesis. J. Bacteriol.
2005, 187, 5292−5300.
(8) Caroff, M.; Karibian, D. Structure of bacterial lipopolysaccharides.
Carbohydr. Res. 2003, 338, 2431−2447.
(9) De Leon, G. P.; Elowe, N. H.; Koteva, K. P.; Valvano, M. A.;
Wright, G. D. An in vitro screen of bacterial lipopolysaccharide
biosynthetic enzymes identifies an inhibitor of ADP-heptose biosyn-
thesis. Chem. Biol. 2006, 13, 437−441.
1
optimized. All new compounds gave satisfactory analytical data. H
NMR spectra were recorded at 300 or 400 MHz on a Bruker
̈
instrument, and chemical shifts are reported in parts per million (δ)
downfield from the internal standard tetramethylsilane. Mass spectra
were obtained using electrospray (ESI) ionization techniques on an
Agilent 1100 series LCMS. HPLC (analytical and preparative) was
performed on an Agilent 1100 HPLC instrument with diode array
detection. Preparative HPLC was performed at 0.7 mL/min on a
Thermo Electron, Hypersil BDS C-18 column (250 mm × 4.6 mm, 5
μm) using a gradient of acetonitrile and water with 0.1% TFA (50% in
acetonitrile to 100% and then back to 50%). The tested compounds
were determined to be >95% pure via HPLC.
(10) Desroy, N.; Moreau, F.; Briet, S.; Fralliec, G. L.; Floquet, S.;
Durant, L.; Vongsouthi, V.; Gerusz, V.; Denis, A.; Escaich, S. Towards
Gram-negative antivirulence drugs: new inhibitors of HldE kinase.
Bioorg. Med. Chem. 2009, 17, 1276−1289.
́
(11) Durka, M.; Tikad, A.; Perion, R.; Bosco, M.; Andaloussi, M.;
Floquet, S.; Malacain, E.; Moreau, F.; Oxoby, M.; Gerusz, V.; Vincent,
S. P. Systematic synthesis of inhibitors of the two first enzymes of the
bacterial heptose biosynthetic pathway: towards antivirulence
molecules targeting lipopolysaccharide biosynthesis. Chem.Eur. J.
2011, 17, 11305−11313.
ASSOCIATED CONTENT
* Supporting Information
Synthetic procedures and spectroscopic details for all final
compounds. This material is available free of charge via the
Internet at http://pubs.acs.org.
■
(12) Takagi, K.; Iwachido, T.; Hayama, N. Palladium(0)-catalyzed
synthesis of 2-alkylbenzothiazoles by a novel thiation of 1-amino-2-
iodoarenes with thioamides. Chem. Lett. 1987, 16, 839−840.
(13) Limanto, J.; Desmond, R. A.; Gauthier, D. R.; Devine, P. N.;
Reamer, R. A.; Volante, R. P. A regioselective approach to 5-
substituted-3-amino-1,2,4-triazines. Org. Lett. 2003, 5, 2271−2274.
(14) Oliveira, C.; Vongsouthi, V.; Moreau, F.; Denis, A.; Escaich, S.;
Gerusz, V.; Desroy, N. New 1,2,4-Triazine Derivatives and Biological
Applications Thereof. WO 2010001220, 2010.
S
AUTHOR INFORMATION
Corresponding Author
*Phone: +33-157140522. Fax: +33-157140524. E-mail: vincent.
gerusz@mutabilis.fr.
■
(15) Russell, M. G. N.; Carling, R. W.; Street, L. J.; Hallett, D. J.;
Goodacre, S.; Mezzogori, E.; Reader, M.; Cook, S. M.; Bromidge, F.
A.; Newman, R.; Smith, A. J.; Wafford, K. A.; Marshall, G. R.;
Reynolds, D. S.; Dias, R.; Ferris, P.; Stanley, J.; Lincoln, R.; Tye, S. J.;
Sheppard, W. F. A.; Sohal, B.; Pike, A.; Dominguez, M.; Atack, J. R.;
Castro, J. L. Discovery of imidazo[1,2-b][1,2,4]triazines as GABAA
α2/3 subtype selective agonists for the treatment of anxiety. J. Med.
Chem. 2006, 49, 1235−1238.
Present Addresses
∥Galapagos, 102 Avenue Gaston Roussel, 93230 Romainville,
France.
⊥GlaxoSmithKline, 25-27 Avenue du Queb
́
ec, 91951 Les Ulis,
France.
#ESE Conseil, 66 Boulevard Senard, 92210 Saint Cloud,
France.
(16) Grundmann, C.; Schroeder, H.; Ratz, R. New 1,2,4-triazine
̈
∞Alderys, 86 Rue de Paris, 91400 Orsay, France.
derivatives. J. Org. Chem. 1958, 23, 1522−1524.
Notes
(17) Kusumoto, T.; Sato, K.-i.; Hiyama, T.; Takehara, S.; Ito, K.
Synthesis and electro-optical properties of dihydrobenzofurans and
dihydrofuropyridines as chiral dopants for ferroelectric liquid crystals.
Chem. Lett. 1995, 24, 537−538.
The authors declare no competing financial interest.
ACKNOWLEDGMENTS
We are grateful to Roowin and Syngene for building block
syntheses, as well as to Dr. Franco
Nassif, and Dr. Patrice Renaut for stimulating discussions on
■
(18) Catalan, J.; Del Valle, J. C.; Claramunt, R. M.; Maria, M. D. S.;
Bobosik, V.; Mocelo, R.; Elguero, J. Toward the photostability
mechanism of intramolecular hydrogen bond systems. 4. 3(5)-(1′-
Hydroxy-2′-naphthyl)pyrazoles and 3(5)-(2′-hydroxy-1′-naphthyl)-
pyrazoles. J. Org. Chem. 1995, 60, 3427−3439.
̧
is Bellamy, Prof. Xavier
these topics.
(19) Lee, T.-W.; Verhey, T. B.; Antiperovitch, P. A.; Atamanyuk, D.;
Desroy, N.; Oliveira, C.; Gerusz, V.; Malacain, E.; Loutet, S. A.;
Hamad, M.; Stanetty, C.; Andres, S. N.; Sugiman-Marangos, S.;
Kosma, P.; Valvano, M. A.; Moreau, F.; Junop, M. S. Structure,
function and inhibition of D-glycero-β-D-manno-heptose 7-phosphate
kinase (HldA) from Burkholderia cenocepacia: insights for design of
antivirulence drugs or antibiotic adjuvants targeting LPS biosynthesis
in gram-negative bacteria. J. Med. Chem. [Online early access]. DOI:
10.1021/jm301483h. Published Dec 20, 2012.
(20) McGovern, S. L.; Helfand, B. T.; Feng, B.; Shoichet, B. K. A
specific mechanism of nonspecific inhibition. J. Med. Chem. 2003, 46,
4265−4272.
(21) Imai, Y. N.; Inoue, Y.; Nakanishi, I.; Kitaura, K. Cl−π
interactions in protein−ligand complexes. Protein Sci. 2008, 17, 1129−
1137.
REFERENCES
■
(1) Escaich, S. Novel agents to inhibit microbial virulence and
pathogenicity. Expert Opin. Ther. Pat. 2010, 20, 1401−1418.
(2) Cegelski, L.; Marshall, G. R.; Eldridge, G. R.; Hultgren, S. J. The
biology and future prospects of antivirulence therapies. Nat. Rev.
Microbiol. 2008, 6, 17−27.
(3) Gronow, S.; Brade, H. Lipopolysaccharide biosynthesis: which
steps do bacteria need to survive? J. Endotoxin Res. 2001, 7, 3−23.
(4) Coleman, W. G., Jr.; Leive, L. Two mutations which affect the
barrier function of the Escherichia coli K-12 outer membrane. J.
Bacteriol. 1979, 139, 899−910.
(5) Kneidinger, B.; Marolda, C.; Graninger, M.; Zamyatina, A.;
McArthur, F.; Kosma, P.; Valvano, M. A.; Messner, P. Biosynthesis
pathway of ADP-L-glycero-β-D-manno-heptose in Escherichia coli. J.
Bacteriol. 2002, 184, 363−369.
1429
dx.doi.org/10.1021/jm301499r | J. Med. Chem. 2013, 56, 1418−1430

Journal of Medicinal Chemistry
Article
(22) Elkins, C. A.; Nikaido, H. 3D structure of AcrAB: the archetypal
multidrug efflux transporter of E. coli likely captures substrates from
periplasm. Drug Resist. Updates 2003, 6, 9−13.
(23) Atamanyuk, D.; Faivre, F.; Oxoby, M.; Ledoussal, B.; Drocourt,
E.; Moreau, F.; Gerusz, V. Vectorization efforts to increase Gram-
negative intracellular drug concentration: a case study on HldE-K
inhibitors. J. Med. Chem., submitted.
1430
dx.doi.org/10.1021/jm301499r | J. Med. Chem. 2013, 56, 1418−1430