Hydroxamic Acids as Potent Inhibitors of FeII and MnII E. coli Methionine Aminopeptidase: Biological Activities and X-ray Structures of Oxazole Hydroxamate-EcMetAP-Mn Complexes

Article abstract of DOI:10.1002/cmdc.201200076

New series of acids and hydroxamic acids linked to five-membered heterocycles including furan, oxazole, 1,2,4- or 1,3,4-oxadiazole, and imidazole were synthesized and tested as inhibitors against the Fe^II, Co^II, and Mn^II f

Full text of DOI:10.1002/cmdc.201200076

DOI: 10.1002/cmdc.201200076
Hydroxamic Acids as Potent Inhibitors of FeII and MnII
E. coli Methionine Aminopeptidase: Biological Activities
and X-ray Structures of Oxazole Hydroxamate–EcMetAP-
Mn Complexes
Florian Huguet,[a] Armelle Melet,[a] Rodolphe Alves de Sousa,[a] Aurꢀlie Lieutaud,[b]
Jacqueline Chevalier,[b] Laure Maigre,[b] Patrick Deschamps,[c] Alain Tomas,[c] Nicolas Leulliot,[c]
Jean-Marie Pages,[b] and Isabelle Artaud*[a]
New series of acids and hydroxamic acids linked to five-mem-
bered heterocycles including furan, oxazole, 1,2,4- or 1,3,4-oxa-
diazole, and imidazole were synthesized and tested as inhibi-
tors against the FeII, CoII, and MnII forms of E. coli methionine
aminopeptidase (MetAP) and as antibacterial agents against
wild-type and acrAB E. coli strains. 2-Aryloxazol-4-ylcarboxylic
acids appeared as potent and selective inhibitors of the CoII
MetAP form, with IC50 values in the micromolar range, whereas
5-aryloxazol-2-ylcarboxylic acid regioisomers and 5-aryl-1,2,4-
oxadiazol-3-ylcarboxylic acids were shown to be inefficient
against all forms of EcMetAP. Regardless of the heterocycle, all
the hydroxamic acids are highly potent inhibitors and are se-
lective for the MnII and FeII forms, with IC50 values between
1 and 2 mm. One indole hydroxamic acid that we previously re-
ported as a potent inhibitor of E. coli peptide deformylase also
demonstrated efficiency against EcMetAP. To gain insight into
the positioning of the oxazole heterocycle with reversed sub-
stitutions at positions 2 and 5, X-ray crystal structures of EcMe-
tAP-Mn complexed with two such oxazole hydroxamic acids
were solved. Irrespective of the [metal]/[apo-MetAP] ratio, the
active site consistently contains a dinuclear manganese center,
with the hydroxamate as bridging ligand. Asp97, which adopts
a bidentate binding mode to the Mn2 site in the holoenzyme,
is twisted in both structures toward the hydroxamate bridging
ligand to favor the formation of a strong hydrogen bond. Most
of the compounds show weak antibacterial activity against
a wild-type E. coli strain. However, increased antibacterial activ-
ity was observed mainly for compounds with a 2-substituted
phenyl group in the presence of the nonapeptide polymyxin B
and phenylalanine–arginine–b-naphthylamide as permeabilizer
and efflux pump blocker, respectively, which boost the intracel-
lular uptake of the inhibitors.
Introduction
The spread and dissemination of multidrug-resistant bacteria is
a major limitation to the treatment of infectious diseases with
known antibiotics. One strategy for combating this is to focus
on new targets involved in biochemical processes that are es-
sential to bacterial growth, such as N-terminal methionine exci-
sion (NME).[1] In prokaryotes all nascent polypeptide chains are
initiated with an N-formylmethionine residue, and NME re-
quires the action of peptide deformylase (PDF) and methionine
aminopeptidase (MetAP) to successively cleave the formyl
group and the N-terminal methionine. Deletion of the gene
that codes for MetAP was found to be lethal for Escherichia
coli[2] and Salmonella typhymurium.[3] Therefore, MetAP has
been selected as a target of interest in the search for new anti-
bacterial agents. MetAPs belong to the large family of metallo-
proteases[4] and have been extensively characterized over the
last decade. However, unlike most enzymes of this family, the
nature of the metal cation at the active site in vivo as well as
the nuclearity of the metal center are still a subject of contro-
versy. These metalloproteases are activated by several divalent
metal cations such as CoII, MnII, NiII, ZnII or FeII. The enzymatic
activity of E. coli MetAP (EcMetAP) is maximal in the presence
of only 1 equivalent of CoII,[5] FeII,[5] or MnII,[6] suggesting that
MetAP is a mononuclear metalloprotease. However, MetAP can
bind two metal cations, yet the binding affinities for the first
and second metal cations differ by three orders of magnitude;
for example EcMetAP loaded with CoII shows Kd1 =300ꢀ50 nm
and Kd2 =2.5ꢀ0.5 mm.[5] The high-affinity binding site was as-
signed to the histidine-containing site on the basis of EXAFS
experiments with the iron,[7] cobalt,[7] and manganese forms,[6]
a hypothesis that has been confirmed by recent X-ray crystal
structures of monometallated MnII forms of EcMetAP com-
[a] Dr. F. Huguet,+ Dr. A. Melet,+ Dr. R. Alves de Sousa, Dr. I. Artaud
Laboratoire de Chimie et Biochimie Pharmacologiques et Toxicologiques
UMR 8601, CNRS Universitꢀ Paris Descartes
45 rue des Sts. Pꢁres, 75270 Paris Cedex 06 (France)
E-mail: isabelle.artaud@parisdescartes.fr
[b] Dr. A. Lieutaud, Dr. J. Chevalier, Dr. L. Maigre, Dr. J.-M. Pages
UMR-MD1, Transporteurs Membranaires Aix-Marseille Universitꢀ
IRBA, Marseille (France)
[c] Dr. P. Deschamps, Prof. A. Tomas, N. Leulliot
UMR 8015, CNRS Universitꢀ Paris Descartes
Laboratoire de Cristallographie et RMN Biologiques, Paris (France)
[+] These authors contributed equally to this work.
Supporting information for this article is available on the WWW under
http://dx.doi.org/10.1002/cmdc.201200076.
ChemMedChem 0000, 00, 1 – 12
ꢁ 2012 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
&
1
&
ÞÞ
These are not the final page numbers!

I. Artaud et al.
MED
plexed with norleucine phosphonate[8] and a hydrazide inhibi-
tor.[9] In addition to these results that support a mononuclear
form of MetAP, MCD and atomic absorption spectrometry
data[10] as well as EPR experiments[11,12] have led to the assump-
tion that even at low metal content, a small fraction of the
active site is dinuclear due to a weak cooperative binding of
metal cations. Regarding the identity of the metal cation
in vivo, EcMetAP is suggested to be an iron enzyme on the
basis of whole-cell metal analy-
Results and Discussion
Chemistry
Two regioisomers of oxazole derivatives were prepared in
which the oxazole ring was substituted by an aryl group and
a carboxylic acid methyl ester at positions 2 and 5, or 5 and 2,
as outlined in Scheme 1. 2-Aryloxazol-4-ylcarboxylic acid
ses combined with activity
measurements.[13–15] The trend is
now to find metal-specific inhibi-
tors[16,17] that can bind a mononu-
clear or dinuclear form of MetAP.
In this prospect, the hydroxamic
acid moiety is a versatile func-
tional group that can serve as
a bidentate ligand to chelate
metal ions at mono- or dinuclear
enzyme active sites. Accordingly,
hydroxamates are promiscuous
inhibitors of most metalloen-
Scheme 1. Synthesis of oxazole acids. Reagents and conditions: a) NH2CH(CH2OH)COOMe·HCl, Et3N, CH2Cl2, RT,
36 h; b) 1. Deoxo-Fluor, CH2Cl2, 2. BrCCl3, DBU, 08C, 4 h; c) 1. LiOH, H2O/THF, 2. HCl, RT, 1 h; d) ClCOCOOEt, Et3N,
CH2Cl2, RT, 72 h; e) POCl3, reflux, 3.5 h.
zymes containing divalent cat-
ions including ZnII, MnII, NiII, or
FeII. By far the most common
mode of metal binding is
through the deprotonated hydroxamate and carbonyl oxygen
atom, which can adopt an (O,O) or (O,mO) chelating mode, de-
pending if the active site binds one or two metal ions. Hy-
droxamic acids are strong inhibitors of metalloproteases,[18]
whether mononuclear as with matrix metalloproteases[19] and
peptide deformylases,[20] or dinuclear, as with aminopeptidas-
es[21] and ureases.[22]
methyl esters (3a, 3b, and 3c) were obtained by dehydrative
cyclization of the corresponding b-hydroxyamides (2a, 2b, and
2c) with bis(2-methoxyethyl)aminosulfur trifluoride (Deoxo-
Fluor) at ꢁ108C, followed by in situ oxidation of the adduct at
08C with bromotrichloromethane and 1,8-deazabicyclo-
[5.4.0]undec-7-ene (DBU).[25] The corresponding 5-phenyloxa-
diazol-2-yl- and 5-(2-chlorophenyl)oxadiazol-2-ylcarboxylic acid
ethyl esters 6a and 6b, respectively, were synthesized by cycli-
zation, in phosphoryl chloride, of the N-(2-oxo-2-arylethyl)oxa-
lamic acid ethyl esters 5a and 5b prepared by condensation
of 2-aminoacetophenone 4a or 2-amino-1-(2-chlorophenyl)e-
thanone 4b with ethyl 2-chloro-2-oxoacetate.[26] Finally, saponi-
fication of the esters 3a, 3b, and 3c, and 6a and 6b under
mild conditions with aqueous lithium hydroxide in tetrahydro-
furan[26,27] afforded the corresponding acids A3–A7.
Synthetic pathways of the various oxadiazole esters and
acids are shown in Scheme 2. 1,2,4-Oxadiazole (two regioisom-
ers) as well as 1,3,4-oxadiazole derivatives were synthesized.
Methyl 2-phenyllamino-3-dimethylaminopropenoate 7a was
treated in aqueous hydrochloric acid with sodium nitrite at
08C and left at room temperature to afford the 5-phenyl-1,2,4-
oxadiazol-3-ylcarboxylic acid methyl ester 9a.[28] Under the
same conditions, treatment of the 2-chlorophenyl analogue of
7a[28] failed to produce the corresponding oxadiazole. Howev-
er, 9b was readily obtained by condensation of ethyl-2-oxime
acetate[29,30] with 2-chlorobenzoyl chloride followed by subse-
quent cyclization of the adduct 8b in N,N-dimethylformamide
at 1508C.[31] The 3-aryl-1,2,4-oxadiazolyl-5-carboxylic acid ethyl
ester regioisomers were prepared in two steps. Condensation
of the arylnitriles 10a and 10b with hydroxylamine in a mixture
of methanol and water at reflux afforded the amidoximes 11 a
In continuation of our ongoing research on new antibacteri-
al agents that target the NME process in bacteria, we have ex-
tended our work from PDF inhibitors[23,24] to MetAP inhibitors.
We previously prepared several indole hydroxamic acids as
potent PDF inhibitors,[23,24] and herein we describe the synthe-
sis of a new series of such derivatives linked to five-membered
heterocycles including furan, oxazole, oxadiazole, and imida-
zole to favor interactions with the protein backbone of MetAP.
Their inhibitory potencies, as well as those of some of the
indole derivatives, were evaluated against CoII, MnII, and FeII
forms of EcMetAP; we then assessed their antibacterial activi-
ties against Gram-positive and Gram-negative bacteria. These
data were compared with those of the corresponding acids,
for compounds that were sufficiently stable for isolation. In the
furan series, 5-arylfuranyl-2-carboxylic acids and 5-(2-chlrophe-
nyl)furanylhydroxamic acid were previously prepared and
tested by Ye et al.[16,17] Two X-ray crystal structures of the MnII
form with oxazole hydroxamic acids were solved, highlighting
a specific interaction of the hydroxamic moiety with a dinuclear
active site, even at low metal content.
&2
&
www.chemmedchem.org
ꢁ 2012 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ChemMedChem 0000, 00, 1 – 12
ÝÝ
These are not the final page numbers!

Hydroxamic Acid EcMetAP Inhibitors
compound was the acid or the ester, as shown in Scheme 3.
Carboxylic acids were activated by isobutyl chloroformate to
form mixed anhydrides, which were then treated with O-(tert-
butyldimethylsilyl)hydroxylamine to give the corresponding O-
protected hydroxamic acids. Final deprotection with tetrabuty-
lammonium fluoride in tetrahydrofuran afforded the hydroxa-
mic acids[24] HA1–HA8 and H14. Alternatively, direct reaction of
esters 8b, 12a, 12b, 14a, and 14b with NH2OK in methanol at
room temperature[36] resulted in the clean formation of hy-
droxamic derivatives HA9–HA13.
In vitro activity toward the CoII, MnII, and FeII metalloforms
of EcMetAP
The inhibitory activity of all synthesized compounds was evalu-
ated toward EcMetAP loaded with CoII, MnII, and FeII; the result-
ing IC50 values are listed in Table 1. IC50 values were determined
at a [metal]/[apo-MetAP] ratio of 5 and a metal concentration
of 15 mm. Under these assay conditions, based on the very dif-
ferent binding constants of the two metal cations,[5,37] only the
higher-affinity metal binding site (assigned to the His-contain-
ing site) is expected to be occupied, and the active enzyme is
presumably monometallated.[7,8]
Scheme 2. Synthesis of oxadiazole esters and acids. Reagents and conditions:
a) HNO2, H2O, 08C, 1 h; b) EtOCOC(NH2)NOH, Et3N, CH2Cl2, 08C, 1 h; c) DMF,
1508C, 4 h; d) 1. LiOH, H2O/THF, 2. HCl, RT, 1 h; e) NH2OH·HCl, Na2CO3, H2O/
MeOH, reflux, 12 h; f) ClCOCOOEt, iPr2EtN, THF, reflux, 2.5 h; g) 1. ClCO-
COOMe, Et3N, CH2Cl2, 2. TsCl, CH2Cl2, 08C, 1 h.
In the oxazole acid series, 2-aryl-oxazol-4-ylcarboxylic acids
A4 and A5 appear as potent and selective inhibitors for the
and 11 b. Further reaction with
ethyl 2-chloro-2-oxoacetate in
tetrahydrofuran at reflux yielded
the 3-aryl-1,2,4-oxadiazolyl-5-car-
boxylic acid ethyl esters 12a and
12b.[29] 1,3,4-Oxadiazole deriva-
tives 14a and 14b were pre-
pared by treatment of the aryl
hydrazides 13a and 13b with
methyl 2-chloro-2-oxoacetate in
dichloromethane at 08C, fol-
lowed by cyclization at room
temperature of the methyl 2-(ar-
ylhydrazinyl)-2-oxoacetate inter-
Scheme 3. Synthesis of hydroxamic acids. Reagents and conditions: a) 1. iBuOCOCl, Et3N, THF, 08C, 1 h,
2. NH2OTBDMS, RT, 12 h, 3. TBAF, RT, 4 h; b) NH2OK, MeOH, RT, 12 h.
mediates with tosyl chloride.[32]
Saponification of the oxadiazole
esters 9a and 9b with aqueous
lithium hydroxide in methanol afforded the 5-aryl-1,2,4-oxadia-
zolyl-2-carboxylic acids A8 and A9. The same treatment ap-
plied to the esters 12a, 12b, 14a, and 14b, however, did not
provide the expected acids, but instead the decarboxylated
aryl oxadiazoles, as previously reported.[33]
MnII form of MetAP, with IC50 values in the micromolar range.
These values are similar to those obtained for 5-arylfuran-2-yl-
carboxylic acids used as standards in our assay conditions, as
they were previously assayed against EcMetAP by Ye
et al.[16,17,38] Similarly, the unsubstituted phenyl derivative A3 is
less efficient, and the introduction of a substituent at the ortho
position of the phenyl ring improves inhibitory potency. How-
ever, there is a significant influence of the relative positions of
the aryl and acid substituents on the oxazole heterocycle: the
5-aryloxazol-2-ylcarboxylic acid regioisomers A6 and A7 are in-
efficient inhibitors. Moreover, 5-aryl-1,2,4-oxadiazol-3-ylcarbox-
ylic acids A8 and A9 do not inhibit any form of EcMetAP. The
imidazole acid A14 is only moderately active against the MnII
form.
5-(2-Chlorophenyl)-1H-imidazole-2-carboxylic acid A14 was
prepared by following a previously reported method[34] by re-
action of 3,3,3-trifluoro-2-oxopropanal with 2-chlorobenzalde-
hyde and ammonia in methanol, followed by saponification. Fi-
nally, 5-(2-chlorophenyl)furan-2-yl-carboxylic acid A1 and 5-(2-
trifluoromethylphenyl)furan-2-ylcarboxylic acid A2 were pre-
pared as previously described by Huang et al.[17]
Hydroxamic acids (for a review, see Yang and Lou[35]) were
synthesized by following two routes, whether the starting
ChemMedChem 0000, 00, 1 – 12
ꢁ 2012 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemmedchem.org
&
3
&
ÞÞ
These are not the final page numbers!

I. Artaud et al.
MED
As expected, hydroxamic
acids are more potent inhibitors
than the corresponding acids,
regardless of the five-membered
heterocycle being a furan, an
oxazole, an oxadiazole, or an
imidazole. Whereas with the
most efficient molecules bearing
a chloro or a trifluoromethyl
group at the ortho position of
the phenyl ring a complete loss
of metalloform selectivity is ob-
served, compounds with an un-
substituted phenyl ring—HA6,
HA8, and HA10—display high
selectivity for MetAP loaded
with MnII and FeII versus CoII.
Meanwhile, the 5-phenyl-1,2,4-
oxadiazol-2-ylhydroxamic acid
HA12 is highly selective for the
FeII form over the MnII and CoII
forms, with an IC50 value for Ec-
MetAP-Fe of 1.85 mm, which is
50- and 100-fold lower than for
EcMetAP-Mn and EcMetAP-Co,
respectively. Usually, the design
of an inhibitor targeting a metal-
loprotease should involve a bal-
ance between the binding affin-
ity at the metal active site and
the strength of interactions of
the peptide or non-peptide
scaffold with amino acid resi-
dues of the protein backbone.
However, in accordance with
the specific involvement of side-
rophores in iron chelators,[39] the
hydroxamic moiety is a stronger
chelator of iron than it is of
Table 1. In vitro inhibition assays against EcMetAP loaded with CoII, FeII, and MnII.
Acids
IC50 [mm][a]
FeII
Hydroxamic acids
IC50 [mm][a]
CoII
MnII
CoII
FeII
MnII
1.3
>200
94
>200
>200
>200
>200
>200
>200
2.9
3.5
1
>200
>200
>200
>200
>200
>200
1.5
5.7
1.3
1.6
1.2
0.9
1.2
0.95
1.4
7.1
1.8
1.9
5.4
1.9
80
36
4.5
2.6
4.8
7.3
>200
171
>200
>200
>200
15.1
97
>200
>200
>200
>200
1.2
0.9
1.3
1.15
1.85
1
6.7
2
20.5
>200
30
36.5
7
manganese.
To
specifically
>200
105
8.8
4.9
target the iron metalloform,
compounds with this binding
group need not have very high
affinity for the S1’ subsite,
which interacts with the me-
thionine side chain of MetAP
substrates.
66
40
>200
>200
62
2.35
Finally, we tested some of our
indole hydroxamic acids known
as potent inhibitors of EcPDF-Ni.
One of them, 2-(3-benzyl-5-bro-
moindol-1-yl)-N-hydroxyaceta-
mide, HA15 (23c in Ref. [24];
IC50 =312 nm against EcPDF-Ni),
appeared as active against Ec-
MetAP as the most efficient
5.2
1.0
2.8
[a] IC50 values represent the mean of three independent determinations with standard deviations <20%.
hydroxamic
acid
described
&4
&
www.chemmedchem.org
ꢁ 2012 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ChemMedChem 0000, 00, 1 – 12
ÝÝ
These are not the final page numbers!

Hydroxamic Acid EcMetAP Inhibitors
above. This inhibitor is the first mixed inhibitor of EcPDF and
EcMetAP.
geometry. In addition, Asp97 rotates slightly, resulting in
a lower OꢁMn2 interaction (Mn2···Asp97Od2: 2.79 and 2.68 ꢂ for
HA5 and HA7, versus 2.4 ꢂ in the holoenzyme). The carboxyl-
ate group of the aspartate residue is twisted toward the bridg-
ing NHO of the hydroxamate to promote the formation of
a strong hydrogen bond with the NH moiety of this group. In-
terestingly, Asp97 adopts a binding mode between that ob-
served in the free binuclear MnII enzyme (pure bidentate bind-
ing to Mn2),[8] and in the mononuclear MnII form complexed
with a hydrazide (strong hydrogen bonding with the NHNH2 of
the hydrazide;[9] see figure S1 in the Supporting Information).
Finally, the overall structure of inhibitors is not planar; the
aromatic ring and the oxazole heterocycle are twisted (dihedral
angles between the two cycles: 1398 for HA5 and 1448 for
HA7). In the two regioisomers, oxazole heteroatoms are both
hydrogen bonded to the two histidines His178 and His79,
which are highly conserved in all MetAP sequences. They are
proposed to activate the bound water molecule and to partici-
pate in the orientation of the substrate.[40,41] In HA5 and HA9
structures, it is always the nitrogen atom of the oxazole that is
the closer to His178 (HA5: N···H178 3.43 ꢂ, O···H178 5.50 ꢂ;
HA7: N···H178 2.88 ꢂ, O···H178 4.52 ꢂ), and the oxygen atom to
X-ray crystal structures of EcMetAP-Mn in complex with HA5
and HA7 and structural model of EcMetAP–HA15
To gain insight into the positioning of the oxazole heterocycle
with reversed substitutions at positions 2 and 5, we tried to
solve the X-ray crystal structure of EcMetAP with two such re-
gioisomers bearing a hydroxamic acid as chelating group. We
obtained crystals of EcMetAP-Mn in complex with HA5 and
HA7. These crystallizations were carried out at an enzyme con-
centration of 0.4 mm and [metal]/[apo-MetAP] ratios of 1, 0.5,
and 0.25 to favor the binding of a single divalent metal cation
at the active site. Unexpectedly, the crystal structures are iden-
tical regardless of metal content, and the enzyme consistently
has a dinuclear active site (Mn1ꢁMn2 distance: 3.4 ꢂ) with
metal ion occupancies of 100% for both sites. Structures
solved at a [metal]/[apo-MetAP] ratio of 1 are shown in Fig-
ure 1a,b. The hydroxamate binds the two metal cations and
adopts an (O,O) chelating mode to Mn1 at the higher-affinity
site, while the oxygen atom of the hydroxamate bridges the
two metal cations. In the EcMe-
tAP–HA5 complex, the bridging
oxygen atom O1 is roughly equi-
distant from the two MnII ions
(Mn1···O1: 2.08 ꢂ and Mn2···O1:
1.95 ꢂ), whereas in the EcMe-
tAP–HA7 complex the O1ꢁMn2
distance is significantly shorter
than that of O1ꢁMn1 (Mn1···O1:
2.34 ꢂ and Mn2···O1: 1.88 ꢂ). In
both complexes the carbonyl
oxygen atom of the hydroxamic
acid interacts weakly with Mn1
(Mn1···O2: 2.36 and 2.33 ꢂ for
HA5 and HA7, respectively).
In the holoenzyme active site,
the coordination geometry of
Mn1 is distorted trigonal bipyra-
midal, whereas that of Mn2 is
distorted octahedral.[37] In this
active site, Asp108Od1 and
Glu204Oe2 occupy the axial posi-
tions of Mn1, while OwaterA
,
His171Ne2 and Glu235Oe1 define
the equatorial plane. The same
OwaterA atom together with
Asp108Od2 and bidentate As-
p97Od1,Od2 make up the equato-
rial plane of Mn2, while
Glu235Oe2 and another OwaterB
are in axial position. Here, in
both structures in the presence
of HA5 or HA7, binding of the
inhibitors expands the coordina-
Figure 1. Stereo views of a) the HA5–EcMetAP-Mn complex and b) the HA7–EcMetAP-Mn complex; these struc-
tion sphere of Mn1 in a distorted tures correspond to crystals grown at a metal/apo-protein ratio of 1:1.
ChemMedChem 0000, 00, 1 – 12
ꢁ 2012 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemmedchem.org
&
5
&
ÞÞ
These are not the final page numbers!

I. Artaud et al.
MED
His79 (HA5: N···H79 4.27 ꢂ, O···H79 3.45 ꢂ; HA7: N···H79 5.52 ꢂ,
O···H79 3.83 ꢂ). The selectivity of these hydrogen bonds might
partly explain the differences in interactions of the hydroxa-
mates HA5 and HA7 to the dinuclear active site. They could
force HA7 to interact more strongly with Mn2 than with Mn1.
The question that arises is the relevance of these structures
in relation to the enzyme assay conditions. The fact that same
structures are observed for both inhibitors at [metal]/[apo-
MetAP] ratios of 0.5 and 0.25 supports a positive answer. The
starting enzyme is likely to be mononuclear in the assays, but
the inhibited form is dinuclear. This result may be consistent
with a cooperative binding of the two metal cations, as previ-
ously proposed by Larrabee et al.[11] and Holz and colleagues[12]
on the basis of spectroscopic data analyses. One hypothesis
should be that hydroxamates bind a mononuclear form of
MetAP, promoting an increase in the binding affinity for the
second Mn center, and the final inhibited state of MetAP
would be always dinuclear. Thus, the cooperative metal bind-
ing appears to be inhibitor-dependent.
We did not succeed in obtaining crystals of EcMetAP-Mn in
complex with the indole hydroxamic derivative HA15 suitable
for X-ray analysis, so we performed a molecular modeling
study by replacing the oxazole HA7 with HA15 in the crystal
structure (PDB ID: 4A6W). Both the indole and the phenyl rings
are located deeply inside the S1’ pocket, as shown in Figure 2.
Antibacterial activities
Supporting Information table S1 lists the results obtained on
isogenic E. coli strains (AG100 and AG100A) in the presence
and absence of sub-inhibitory concentrations of polymyxin B
nonapeptide (PMBN used at 20% of its MIC value), a molecule
known to increase membrane permeability.[42] AG100 is a wild-
type E. coli strain, and AG100A is the isogenic derivative strain,
in which the efflux pump acrAB genes are deleted. Most of the
MIC values are greater than 64 mgmLꢁ1. A small increase in sus-
ceptibility was observed in the presence of PMBN as permea-
bilizer; the most significant effect was observed with furan
(A1, HA1, and HA2) and oxazole (HA5 and HA7) derivatives.
Additional experiments were conducted with HA1, HA2, HA5,
and HA7 (Table 2). To evaluate the role of other efflux pump
systems present in E. coli that can limit the intracellular uptake
of MetAP inhibitors, the activities of these compounds were
tested on AG100A in the presence of PAbN, an efflux pump
blocker, and of both PMBN and PAbN. A two- to eightfold de-
crease in MIC values was observed, indicating that as previous-
ly reported by us[42] and others,[14,43] both the low permeability
of the outer membrane and the efficiency of efflux pumps
limit the intracellular activity of such hydroxamic compounds
in E. coli.
Figure 2. a) Representation of the HA15–EcMetAP-Mn complex obtained
after molecular dynamics and energy minimization. b) Superimposition of
HA5 (purple), H15 (green), and HA7 (brown) inside the EcMetAP-Mn struc-
ture.
ther were susceptible to the hydroxamic acids at low concen-
tration (data not shown).
It is generally reported that antibacterial molecules are more
active toward Gram-positive than Gram-negative bacteria, as
the additional outer membrane of the latter organisms im-
pedes uptake.[44] We therefore decided to determine whether
these MetAP inhibitors are active toward Gram-positive bacte-
ria. All the compounds were tested against two Gram-positive
strains, Staphylococcus aureus and Enterococcus hirae, but nei-
Conclusions
In this study the inhibitory potencies of a new series of hy-
droxamic acids attached to five-membered heterocycles substi-
tuted by an aryl group were evaluated against EcMetAP
loaded with CoII, MnII, and FeII. In contrast to furan acids, which
are selective for the MnII form, furan, oxazole, oxadiazole, and
&6
&
www.chemmedchem.org
ꢁ 2012 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ChemMedChem 0000, 00, 1 – 12
ÝÝ
These are not the final page numbers!

Hydroxamic Acid EcMetAP Inhibitors
the Cochin Institute of Paris Descartes University (France). The vari-
ous synthetic pathways for the products are depicted in
Schemes 1–3.
Table 2. Activity of MetAP inhibitors on E. coli strains: effect of adjuvant-
s.[a]
Conditions
MIC [mgmLꢁ1
HA2 HA5
]
5-(2-Chlorophenyl)furan-2-carboxylic acid A1: An aqueous solu-
tion of NaNO2 (414 mg, 6 mmol, 3 mL) was added to an aqueous
solution of 2-chloroaniline (638 mg, 5 mmol) in concentrated aque-
ous HCl (4n, 3 mL) at 08C under stirring; 20 min later an aqueous
solution of 2-furanoic acid (560 mg, 5 mmol) in acetone (3 mL) and
an aqueous solution of CuCl2 (256 mg, 1.5 mmol, 3 mL) were
added. The precipitate that formed after one day of stirring at
room temperature was isolated to give A1 as a brown powder
(523 mg, 47%). 1H NMR (250 MHz, CDCl3): d=7.24 (d, 3J=3.6 Hz,
HA1
HA7
AG100
AG100+PMBN
AG100A
AG100A+PMBN
AG100A+PAbN
AG100A+PMBN+PAbN
>64
16–8
64
16
32
16
>64 >64
>64
64
64
16
16
8
64–32
64
64
64
64–32
16–8
8–4
64
32–16
32–16
[a] Values are the means of four independent assays; AG100 and AG100A
are wild-type and acrABꢁ E. coli strains, respectively; PMBN permeabilizer
used at 20% of its MIC value; PAbN: efflux pump blocker used at
20 mgmLꢁ1
3
3
4
1H), 7.31 (d, J=3.6 Hz, 1H), 7.35 (td, J=7.5 Hz, J=1.7, 1H), 7.42
(td, 3J=7.5 Hz, 4J=1.35 Hz, 1H), 7.52 (dd, 3J=7.5 Hz, 4J=1.35 Hz,
.
3
4
1H), 7.98 ppm (dd, J=7.5 Hz, J=1.35 Hz, 1H); 13C NMR (500 MHz,
CDCl3): d=113.6, 120.5, 128.4, 129.6, 130.9, 131.8, 145.6, 154.6,
161.6, 129.2, 132 ppm; MS (ESIꢁ, MeOH) m/z (%): 220.9 (100)
[MꢁH]ꢁ; Anal. calcd for C11H7ClO3: C 59.35, H 3.17, found: C 59.21,
H 3.18.
imidazole hydroxamic acids are selective for both the MnII and
FeII forms. One compound, 5-phenyl-1,2,4-oxadiazol-2-ylhy-
droxamic acid HA12, is highly selective for the FeII form, with
an IC50 value of 1.85ꢀ0.15 mm, which is 50- and 100-fold lower
than those for the MnII and CoII forms, respectively. However,
despite significant activity toward an iron form, only the furan-
yl- and oxazolylhydroxamic acid bearing a 2-chloro or 2-tri-
fluoromethylphenyl group as heterocycle substituent display
a slight antibacterial activity against acrABꢁ E. coli strains. To
get insight into the positioning of the heterocycle in the pro-
tein, X-ray crystal structures of EcMetAP-Mn in complex with
two regioisomers of oxazole hydroxamic acids HA5 and HA7
were solved. They indicate that the heterocycle is anchored in
the protein through strong hydrogen bonding interactions be-
tween the nitrogen and oxygen heteroatoms of the oxazole
ring and the conserved His198 and His79 residues, respectively.
Moreover, although the enzyme is likely to be mononuclear in
enzyme assays with a [metal]/[apo-protein] ratio of 5, for EcMe-
tAP-Mn–inhibitor complexes, the enzyme is consistently dinu-
clear when crystallized with Mn/apo-protein ratios of 1, 0.5, or
0.25, suggesting a cooperative metal binding mode. An indole
hydroxamic acid has been identified as the first mixed EcPDF/
EcMetAP inhibitor. All these results underscore the interest in
hydroxamic acids, which are known to be capable of efficiently
binding the mononuclear or dinuclear active sites of metallo-
proteins.
5-(2-(Trifluoromethyl)phenyl)furan-2-carboxylic acid A2: Follow-
ing a similar procedure as above, 256 mg of A2 (20%) were ob-
tained as a brown powder from 2-trifluoromethylaniline (0.805 g,
1
3
5 mmol). H NMR (250 MHz, CD3OD): d=6.83 (dq, J=3.7 Hz, JHF
=
3
3
0.7 Hz, 1H), 7.30 (d, J=3.7 Hz, 1H), 7.72 (t, J=7.8 Hz, 1H), 7.61 (t,
3J=7.8 Hz, 1H), 7.84 ppm (m, 2H); 13C NMR (500 MHz, CD3OD): d=
113, 120.5, 133.5, 132, 130.5, 127.5 (JCF =5.4 Hz), 146.5, 155.5, 161.5,
128.5 (JCF =32 Hz), 125.3 (JCF =273.4 Hz), 129.5 ppm; MS (ESIꢁ,
MeOH) m/z (%): 254.8 (100) [MꢁH]ꢁ; Anal. calcd for C12H7F3O3: C
56.26, H 2.75, found: C 56.07, H 2.71.
2-Phenyloxazole-4-carboxylic acid A3: An aqueous solution of
LiOH (2m, 1 mL) was added to a suspension of 3a in a mixture of
H2O (1 mL) and THF (200 mL). The mixture was stirred for 1 h, then
acidified to pH 2 with HCl. This solution was extracted with EtOAc,
and the organic layer was washed with brine and dried over
MgSO4. Evaporating the solvent yielded A3 as a white solid
1
(77 mg, 81%). H NMR (250 MHz, [D6]DMSO): d=7.59 (m, 3H), 8.02
(m, 2H), 8.86 (s, 1H), 13.24 ppm (bs, 1H); 13C NMR (500 MHz,
[D6]DMSO): d=126.5, 126.6, 129.6, 131.6, 134.8, 145.8, 161.5,
162.4 ppm; MS (ESIꢁ, MeOH) m/z (%): 188.1 (10) [MꢁH]ꢁ, 144.1
(100) [MꢁCO2ꢁH]ꢁ; Anal. calcd for C10H7NO3: C 63.49, H 3.73, N
7.40, found: C 63.33, H 3.71, N 7.32.
A4 and A5 were obtained by using a similar procedure as de-
scribed for A3.
2-(2-Chlorophenyl)oxazole-4-carboxylic acid A4: Compound 3a
(143 mg, 92%) was obtained as a white powder from 3b (165 mg,
1
0.69 mmol). H NMR (250 MHz, [D6]DMSO): d=7.88 (m, 2H), 8.0 (d,
3
3J=7.5 Hz, 1H), 8.05 (d, J=7.5 Hz, 1H), 9.0 (s, 1H), 13.3 ppm (bs,
Experimental Section
1H); 13C NMR (250 MHz, CDCl3): d=125.2, 126.9, 131.2, 131.6,
132.1, 133, 133.6, 145.2, 160.8, 164.9 ppm; Anal. calcd for
C10H6ClNO3: C 53.71, H 2.70, N 6.26, found: C 53.46, H 2.72, N 6.15.
Chemistry
All solvents and chemicals were purchased from SDS and Aldrich,
respectively. DMF, THF, and CH2Cl2 were dried according to stan-
dard procedures. 1H NMR and 13C NMR spectra were recorded on
ARX-250 or Bruker Avance 500 spectrometers. Chemical shifts (d)
are expressed in ppm downfield from TMS. IR spectra were ob-
2-(2-Trifluoromethylphenyl)oxazole-4-carboxylic acid A5: Com-
pound A5 (194 mg, 93%) was obtained as a brown powder from
1
3c (220 mg, 0.81 mmol). H NMR (250 MHz, [D6]DMSO): d=7.88 (m,
3
4
2H), 8.0 (m, 1H), 8.05 (dd, J=7.3 Hz, J=1.5 Hz, 1H), 8.96 ppm (s,
1H); 13C NMR (250 MHz, CDCl3): d=121.1, 124.7, 127.1, 128.6,
131.2, 132, 132.1, 133.8, 133.8, 145.8, 160.8, 164.7 ppm; Anal. calcd
for C11H6F3NO3·0.2LiOH: C 50.43, H 2.38, N 5.34, found: C 50.64, H
2.46, N 5.31.
tained with
a PerkinElmer Spectrum One FTIR spectrometer
equipped with a MIRacle single-reflection horizontal ATR unit (ger-
manium crystal). ESI mass spectrometry analyses of small mole-
cules were performed on a Thermo Finnigan LCD Advantage spec-
trometer. HRMS and elemental analyses were carried out by the
microanalysis services at Gif-sur-Yvette (CNRS, France). ESIMS analy-
ses of proteins were performed on an LTQ Orbitrap instrument at
5-Phenyloxazole-2-carboxylic acid A6: Compound A6 (76 mg,
94%) was obtained by saponification of 6a (93 mg, 0.43 mmol)
ChemMedChem 0000, 00, 1 – 12
ꢁ 2012 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemmedchem.org
&
7
&
ÞÞ
These are not the final page numbers!

I. Artaud et al.
MED
with LiOH in a THF/H2O mixture as described above for A3.
1H NMR (250 MHz, [D6]DMSO): d=7.50 (m, 3H), 7.81 (d, 3J=7 Hz,
2H), 7.95 ppm (s, 1H); 13C NMR (250 MHz, [D6]DMSO): d=117.3,
124.4, 129.0, 129.5, 131.5, 137.4, 151.3, 156.1 ppm; HRMS-ESIꢁ m/z
[MꢁH]ꢁ calcd for C10H6NO3ꢁ: 188.0348, found: 188.0350.
mated to 1.75m. NH2OK in MeOH (4.2 mmol, 2.39 mL) was added
to the ester (0.42 mmol) dissolved in MeOH (2 mL). The mixture
was stirred overnight then acidified to pH 3–4 with 1n HCl at 08C.
The solution was diluted with EtOAc, washed with brine, dried
over Na2SO4, filtered, and evaporated in vacuo.
5-(2-Chlorophenyl)oxazole-2-carboxylic acid A7: Compound A7
(104 mg, 82%) was obtained as a white powder from 6b (150 mg,
0.6 mmol). H NMR (250 MHz, [D6]DMSO): d=7.53 (m, 2H), 7.67 (m,
5-(2-Chlorophenyl)-N-hydroxyfuran-2-carboxamide HA1: Com-
pound HA1 (27 mg, 40%) was obtained as a red powder from the
acid A1 (100 mg, 0.45 mmol) after purification over silica gel
1
1H), 7.87 (m, 1H), 8.02 (s, 1H), 8.58 ppm (bs, 1H); 13C NMR
(250 MHz, [D6]DMSO): d=119.2, 133.7, 137.9, 138.3, 139.2, 140.0,
143.1, 168.1,ꢁ 172.4, 176.8 ppm; HRMS-ESIꢁ m/z [MꢁH]ꢁ calcd for
(eluent
CH2Cl2/MeOH,
100:0!95:5).
1H NMR
(250 MHz,
[D6]acetone): d=7.23 (d, 3J=3.5 Hz, 1H), 7.31 (d, 3J=3.5 Hz, 1H),
7.45 (m, 2H), 7.58 (dd, 3J=7.7 Hz, 4J=1.5 Hz, 1H), 8.15 (dd, 3J=
C10H535ClNO3
:
221.9958, found: 221.9953; Anal. calcd for
4
7.7 Hz, J=1.5 Hz, 1H), 8.36 (bs, 1H, NH), 10.8 ppm (bs, 1H, OH);
13C NMR (500 MHz, [D6]DMSO): d=112.9, 115.9, 127.8, 128.4, 129.1,
130.0, 130.6, 131.2, 146.3, 151.7, 156.7 ppm; MS (ESIꢁ, MeOH) m/z
(%): 236 (100) [MꢁH]ꢁ; Anal. calcd for C11H9ClNO3·0.4CH2Cl2·0.2H2O:
C 49.71, H 3.44, N 5.08, found: C 49.51, H 3.43, N 5.07.
C10H6ClNO3: C 53.71, H 2.70, N 6.26, found: C 53.81, H 3.15, N 6.51.
5-Phenyl-1,2,4-oxadiazole-3-carboxylic acid A8: Compound A8
(200 mg, 82%) was obtained as a white powder by saponification
with LiOH in a THF/H2O mixture of 9a (217 mg, 1.06 mmol).
1H NMR (500 MHz, [D6]DMSO): d=7.67 (m, 2H), 7.56 (m, 1H),
8.16 ppm (d, 3J=7.6 Hz, 1H); 13C NMR (500 MHz, [D6]DMSO): d=
124.9, 128.3, 129.7, 133.8, 154.7, 163.5, 176.9 ppm; MS (ESIꢁ,
MeOH) m/z (%): 145.1 (100) [MꢁHꢁCO2]ꢁ; Anal. calcd for
C9H6N2O3·0.5 LiCl·H2O: C 47.13, H 3.52, N 12.21, found: C 47.05, H
3.67, N 12.05.
N-Hydroxy-5-(2-(trifluoromethyl)phenyl)furan-2-carboxamide
HA2: Compound HA2 (76 mg, 55%) was obtained as white nee-
dles after recrystallization from acetone/CH2Cl2 from the acid A2
(130 mg, 0.51 mmol). 1H NMR (250 MHz, [D6]acetone): d=6.89 (d,
3
3
3J=3.5 Hz, 1H), 7.22 (d, J=3.5 Hz, 1H), 7.67 (t, J=8 Hz, 1H), 7.79
(t, 3J=7.5 Hz, 1H), 7.9 (d, 3J=7.5 Hz, 1H), 7.98 (d, 3J=8 Hz, 1H),
8.43 (bs, 1H, NH), 10.64 ppm (bs, 1H, OH); 13C NMR (500 MHz,
5-(2-Chlorophenyl)-1,2,4-oxadiazole-3-carboxylic acid A9: Com-
pound A9 (421 mg, 95%) was obtained as a white powder by
saponification of 9b (500 mg, 1.98 mmol) with LiOH/THF/H2O.
1H NMR (500 MHz, [D6]DMSO): d=7.61 (m, 1H), 7.75 (m, 2H), 8.15
[D6]acetone): d=113.2, 116.7, 127.8, 130.7, 133.8, 132.2, 128.0 (JCF
=
6.3 Hz), 131.2, 148.2, 152.9, 157.9 ppm; MS (ESIꢁ, MeOH) m/z (%):
270.1 (100) [MꢁH]ꢁ; Anal. calcd for C12H8F3NO3: C 53.15, H 2.97, N
5.16, found: C 52.85, H 2.99, N 5.08.
3
(d, J=7.6 Hz, 1H); 13C NMR (500 MHz, [D6]DMSO): d=124.1, 129.9,
133.3, 134.2, 136.5, 160.3, 164.6, 176.7 ppm; Anal. calcd for
C9H5ClN2O3·0.1H2O: C 47.74, H 2.31, N 12.37, found: C 48.03, H
2.24, N 12.22.
N-Hydroxy-2-phenyloxazole-4-carboxamide HA3: Starting from
the acid A3 (77 mg, 0.41 mmol), HA3 (23 mg, 28%) was obtained
as a white powder after recrystallization in a mixture of CH2Cl2/
1
MeOH/cyclohexane 3:1:6. H NMR (250 MHz, [D6]acetone): d=7.57
5-(2-Chlorophenyl)-1H-imidazole-2-carboxylic acid A14: Com-
pound A14 was isolated as the hydrochloride salt (pink powder,
(m, 3H), 8.05 (m, 2H), 8.35 (bs, 1H, NH), 8.48 (s, 1H), 10.61 ppm
(bs, 1H, OH); 13C NMR (500 MHz, [D6]acetone): d=127.7, 128, 130.4,
132.4, 137.2, 142.6, 159.3, 162.7. MS (ESIꢁ, MeOH) m/z (%): 203.1
(100) [MꢁH]ꢁ; Anal. calcd for C10H8N2O3: C 58.82, H 3.95, N 13.72,
found: C 58.39, H 4.01, N 13.56.
1
3
4
70% yield). H NMR (250 MHz, CD3OD): d=7.62 (td, J=7.0 Hz, J=
3
2.3 Hz, 1H), 7.74 (m, 2H), 7.79 (d, J=7.0 Hz, 1H), 8.28 ppm (s, 1H);
13C NMR (500 MHz, CD3OD): d=124.22, 125.87, 128.45, 129.05,
131.87, 133.39, 134.68, 135.25, 145.51, 160.44 ppm; MS (ESIꢁ,
MeOH) m/z (%): 221.4 (100) [MꢁH]ꢁ; HRMS-ESI+ m/z calcd for
C10H8N2O235Cl 223.0274, found: 223.0268; Anal. calcd for
C10H8Cl2N2O2·0.37KCl: C 41.90, H 2.81, N 9.77, found: C 41.91, H
2.73, N 9.88.
N-Hydroxy-2-(2-chlrophenyl)oxazole-4-carboxamide HA4: Com-
pound HA4 (85 mg, 57%) was obtained as a white powder from
the acid A4 (142 mg, 0.63 mmol) after recrystallization in acetone/
1
CH2Cl2. H NMR (250 MHz, [D6]acetone): d=7.85 (m, 2H), 7.98 (dd,
3J=6.8 Hz, 4J=2.0 Hz, 1H), 8.09 (d, 3J=6.8 Hz, 1H), 8.35 (bs, 1H,
NH), 8.62 (s, 1H), 10.54 ppm (bs, 1H, OH); 13C NMR (500 MHz,
[D6]acetone): d=122.5, 126.3, 128, 132.5, 133.6, 137, 143.5, 158.8,
160.4 ppm; Anal. calcd for C10H7ClN2O3: C 50.33, H 2.96, N 11.74,
found: C 50.15, H 2.77, N 11.30.
Preparation of hydroxamic acids
Procedure A from the acid: Et3N (0.54 mmol) and isobutylchloro-
formate (0.47 mmol) were added to
a solution of the acid
(0.45 mmol) in THF (10 mL) at 08C under argon. The solution was
stirred for 1 h, then filtered under argon. O-(tert-Butyldimethylsilyl)-
hydroxylamine (0.9 mmol) was added to the clear solution. The
mixture was allowed to warm to room temperature and stirred
overnight. A solution of TBAF in THF (1m, 0.9 mmol) was added,
and the solution was stirred for an additional 4 h. After evapora-
tion of the solvent, the residue was dissolved in EtOAc, and the or-
ganic layer was washed with saturated aqueous NaHCO3, then
with 0.1n HCl and brine. Once filtered, the solution was evaporat-
ed in vacuo.
N-Hydroxy-2-(2-trifluoromethylphenyl)oxazole-4-carboxamide
HA5: Compound HA5 (239 mg, 53%) was obtained as a beige
solid from the acid A5 (226.7 mg, 0.88 mmol) after recrystallization
in acetone/CH2Cl2. 1H NMR (250 MHz, [D6]acetone): d=7.44 (m,
3
4
3H), 8.05 (dd, J=7.3 Hz, JHF =2.0 Hz, 1H), 8.36 (bs, 1H, NH), 8.44
(s, 1H), 10.47 ppm (bs, 1H, OH); 13C NMR (500 MHz, [D6]acetone):
d=107.7, 109.4, 128, 129.7, 133.5, 133.7, 134.6, 138.1, 144.1, 160.2,
161.7 ppm; MS (ESIꢁ, MeOH) m/z (%): 271 (100) [MꢁH]ꢁ; Anal.
calcd for C11H7N2F3O3: C 48.54, H 2.59, N 10.29, found: C 48.22, H
2.43, N 10.11.
Procedure B from the ester: A solution of NH2OK was prepared as
described previously by Huang et al.[36] A suspension of NH2OH·HCl
(2.35 g, 34 mmol) in MeOH was held at reflux under argon. A solu-
tion of KOH (3.35 g, 50 mmol) in MeOH (7 mL) was added to the
resulting clear solution. After holding at reflux for 30 min, the mix-
ture was cooled and filtered to give a clear solution of NH2OK esti-
N-Hydroxy-5-phenyloxazole-2-carboxamide HA6: The final prod-
uct prepared from the acid A6 (54 mg, 0.29 mmol) was solubilized
in a minimal amount of acetone and isolated as a white powder
upon precipitation in pentane (22 mg, 38%). 1H NMR (250 MHz,
3
[D6]acetone): d=7.52 (m, 3H), 7.72 (s, 1H), 7.87 (d, J=6.8 Hz, 2H),
&8
&
www.chemmedchem.org
ꢁ 2012 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ChemMedChem 0000, 00, 1 – 12
ÝÝ
These are not the final page numbers!

Hydroxamic Acid EcMetAP Inhibitors
8.64 (s, 1H), 10.96 ppm (s, 1H); 13C NMR (500 MHz, [D6]acetone):
d=124.3, 126, 128.4, 130.3, 130.7, 154.3, 155, 155.5 ppm; MS (ESIꢁ,
MeOH) m/z (%): 203 (100) [MꢁH]ꢁ; Anal. calcd for for C10H8N2O3: C
58.82, H 3.95, N 13.72, found: C 58.85, H 4.08, N 13.01.
N-Hydroxy-5-(2-chlorophenyl)-1,3,4-oxadiazole-2-carboxamide
HA13: Compound HA13 (29 mg, 54%) prepared from the ester
14b (53.5 mg, 0.22 mmol) was isolated as a white powder upon
precipitation from CH2Cl2 into pentane. 1H NMR (250 MHz,
[D6]DMSO): d=7.7 (m, 3H), 8.07 (d 3J=7.5 Hz, 1H), 9.8 (s, 1H),
12.2 ppm (s, 1H); 13C NMR (250 MHz, [D6]acetone): d=122.6, 127.7,
131.3, 131.6, 132.9, 133.4, 151.1, 157.8, 163.6 ppm; MS (ESIꢁ,
MeOH) m/z (%): 238 (100) [MꢁH]ꢁ; Anal. calcd for
C9H6ClN3O3·0.15H2O: C 44.6, H 2.62, N 17.34, found: C 44.75, H
2.56, N 17.11.
N-Hydroxy-5-(2-chlorophenyl)oxazole-2-carboxamide HA7: The
final product prepared from the acid A7 (80 mg, 0.36 mmol) was
solubilized in a minimal amount of acetone and isolated as a white
powder upon precipitation in pentane (14 mg, 16%). 1H NMR
(250 MHz, [D6]DMSO): d=7.55 (m, 2H), 7.68 (m, 1H), 7.99 (s, 1H),
8.01 (m, 1H), 9.56 (s, 1H), 11.93 ppm (s, 1H); 13C NMR (500 MHz,
[D6]acetone): d=127.1, 128.7, 128.9, 129.9, 131.7, 132, 132.4, 151.5,
154.3, 155.3 ppm; MS (ESIꢁ, MeOH) m/z (%): 237 (100) [MꢁH]ꢁ;
Anal. calcd for C10H7ClN2O3: C 50.33, H 2.96, N 11.74, found: C
50.39, H 3.13, N 11.34.
N-Hydroxy-5-(2-chlorophenyl)-1H-imidazole-2-carboxamide
HA14: Compound HA14 (74 mg, 50%) prepared from the acid
A14 (80 mg, 0.36 mmol) was isolated as a pink powder. 1H NMR
(250 MHz, [D6]acetone): d=7.53 (t, 3J=7.5 Hz, 1H), 7.59 (t, 3J=
3
4
7.0 Hz, 1H), 7.66 (d, 3J=7.5 Hz, 1H), 7.79 (dd, J=7.5 Hz, J=1.5 Hz,
1H), 8.01 (s, 1H), 10.29 (bs, 1H), 10.95 ppm (bs, 1H); 13C NMR
(500 MHz CD3OD): d=121.5, 124.3, 128.9, 129.2, 131.9, 133.5, 134.7,
135.2, 145.1, 157.8 ppm; MS (ESIꢁ, MeOH) m/z (%): 236 (100)
[MꢁH]ꢁ; HRMS-ESI+ m/z calcd for C10H9N3O235Cl 238.0383, found:
238.0382; Anal. calcd for C10H8ClN3O2·1.4KCl·1.5H2O: C 29.62, H
2.98, N 10.36, found: C 29.85, H 2.74, N.10.58.
N-Hydroxy-5-phenyl-1,2,4-oxadiazole-3-carboxamide HA8: Com-
pound HA8 (24 mg, 25%) was obtained as a white powder from
the acid A8 (90 mg, 0.47 mmol) after recrystallization in CH2Cl2.
3
1H NMR (500 MHz, [D6]acetone): d=7.68 (t, J=7.5 Hz, 2H), 7.75 (t,
3J=7.5 Hz, 1H), 8.19 ppm (d, 3J=7.5 Hz, 1H); 13C NMR (500 MHz,
[D6]acetone): d=124.9, 129.5, 130.9, 134.9, 155.2, 164.6, 177.8 ppm;
MS (ESIꢁ, MeOH) m/z (%): 204.1 (100) [MꢁH]ꢁ; Anal. calcd for
C9H7N3O3: C 52.69, H 3.44, N 20.48, found: C 52.68, H 3.46, N 20.31.
Protein purification and enzymatic assays
N-Hydroxy-5-(2-chlorophenyl)-1,2,4-oxadiazole-3-carboxamide
HA9: The final product (88 mg, 62%) prepared from the ester 8b
(150 mg, 0.59 mmol) was isolated as a white powder upon precipi-
tation from acetone into cold pentane. 1H NMR (500 MHz,
Purification of apo-EcMetAP (pET28) for enzyme assays: The ex-
pression plasmid for EcMetAP (pET28b-EcMetAP) was a kind gift
from Prof. J. Larrabee (Middlebury College, VT, USA), Dr. W. T. Low-
ther (Howard Hughes Medical Institute, University of Oregon, USA),
and Prof. B. W. Matthews (Wake Forest University School of Medi-
cine, Winston-Salem, NC, USA). EcMetAP was purified as the apo-
enzyme using the previously described protocol of Lowther
et al.,[45] with some minor modifications as described, such as the
use of the detergent Bugbuster 1ꢃ instead of Triton X-100. The
purity of the purified apo-enzyme was assessed by capillary elec-
trophoresis (Experion, Bio-Rad) as ꢂ95%. The molecular mass of
purified apo-MetAP was 29634.27 Da as determined by ESIMS on
an LTQ Orbitrap instrument, in good agreement with the theoreti-
cal value.
3
[D6]DMSO): d=7.63 (m, 1H), 7.75 (m, 2H), 8.14 (d, J=7.6 Hz, 1H),
9.66 (s, 1H), 11.87 ppm (s, 1H); 13C NMR (500 MHz, [D6]DMSO): d=
124.1, 129.9, 133.2, 134.2, 134.3, 136.5, 155.4, 164.8, 176.3 ppm; MS
(ESIꢁ, MeOH) m/z (%): 237.9, (100) [MꢁH]ꢁ; Anal. calcd for
C9H6ClN3O3·0.1H2O: C 44.77, H 2.58, N 17.4, found: C 44.23, H 2.54,
N 16.95.
N-Hydroxy-3-phenyl-1,2,4-oxadiazole-5-carboxamide
HA10:
Starting from the ester 12a (120 mg, 0.55 mmol), HA10 (32 mg,
28%) was isolated as a white powder after purification over silica
gel (eluent CH2Cl2/MeOH/cyclohexane 3:1:6). 1H NMR (250 MHz,
3
4
[D6]DMSO): d=7.63 (m, 3H), 8.08 (dd, J=6.5 Hz, J=1.5 Hz, 2H),
9.94 (s, 1H), 12.35 ppm (s, 1H); 13C NMR (500 MHz, CD3OD): d=
127.4, 128.6, 130.3, 133.0, 153.1, 169.8, 170.2 ppm; MS (ESIꢁ,
MeOH) m/z (%): 204.1 (100) [MꢁH]ꢁ; HRMS-ESIꢁ m/z [MꢁH]ꢁ calcd
for C9H6N3O3ꢁ: 204.0409, found: 204.0414.
Purification of apo-EcMetAP (pET23) for crystallization: The open
reading frame for EcMetAP was amplified by PCR from the
pXL1071 plasmid,[46] kindly provided by Dr. Meinnel (ISV, CNRS Gif-
sur-Yvette, France) and subcloned into the Novagen pET23a vector
using NdeI and EcoRI restrictions sites. BL21(DE3) E. coli cells con-
taining the expression plasmid pET23a-EcMetAP were grown in
Luria–Bertani (LB) medium supplemented with ampicillin
(100 mgmLꢁ1) at 378C until OD600 reached 0.5–0.6. Expression of Ec-
MetAP was induced by adding 0.3 mm isopropyl-b-d-thiogalacto-
side (IPTG) for 3 h at 308C. Cells were pelleted and stored at
ꢁ408C until use. Cell pellets were thawed in buffer [50 mm HEPES
(pH 7.5), 150 mm KCl, 15 mm methionine, 5 mm EDTA, lysozyme
(0.2 mgmLꢁ1), 0.1% Triton X-100] at 10 mL(g pellet)ꢁ1 and sonicat-
ed with a Vibra-cell sonicator (Fisher Bioblock, Vibra-Cell 75115,
500W) at 30% amplitude by two cycles of 1 min 15 s (3 s on, 15 s
off) on ice. After centrifugation at 18000 rpm (37600 g) for 30 min,
the clarified supernatant was subjected to ammonium sulfate pre-
cipitation (0!55!80%) and centrifuged (10000 rpm; 12000 g) for
30 min at 48C. The C80 pellet was resuspended in 5 mL buffer
[20 mm Tris·HCl (pH 7.4), 15 mm methionine, 5 mm EDTA] and dia-
lyzed overnight against buffer A [20 mm Tris·HCl (pH 7.4), 15 mm
methionine, 0.1 mm EDTA] followed by two buffer changes. The
following day, the dialysate was loaded onto a HiLoad 16/10 Q-Se-
pharose column (GE Healthcare) equilibrated with buffer A, and
N-Hydroxy-3-(2-chlorophenyl)-1,2,4-oxadiazole-5-carboxamide
HA11: Starting from the ester 12b (100 mg, 0.40 mmol), HA11
(92 mg, 92%) was isolated as a white powder after precipitation
1
from acetone into cold pentane. H NMR (500 MHz, [D6]DMSO): d=
3
3
3
7.58 (t, J=7.8 Hz, 1H), 7.66 (t, J=7.8 Hz, 1H), 7.72 (d, J=7.8 Hz,
1H), 7.95 (d, 3J=7.8 Hz, 1H), 9.94 (s, 1H), 12.29 ppm (s, 1H);
13C NMR (500 MHz, [D6]DMSO): d=124.7, 127.7, 130.8, 131.8, 132.1,
132.9, 150.2, 166.7, 168 ppm; MS (ESIꢁ, MeOH) m/z (%): 238.2 (100)
[MꢁH]ꢁ, 476.9 (100) [2m-H]ꢁ; HRMS-ESIꢁ m/z [MꢁH]ꢁ calcd for
C9H535ClN3O3ꢁ: 238.0019, found: 238.0019.
N-Hydroxy-5-phenyl-1,3,4-oxadiazole-2-carboxamide
HA12:
Compound HA12 (27 mg, 85%) prepared from the ester 14a
(50 mg, 0.21 mmol) was isolated as a white powder upon precipita-
1
tion from CH2Cl2 into pentane. H NMR (500 MHz, [D6]DMSO): d=
3
7.67 (m, 3H), 8.09 (d, J=7, 2H), 9.74 (s, 1H), 12.15 ppm (s, 1H);
13C NMR (500 MHz, [D6]DMSO): d=122.7, 126.9, 129.4, 132.5, 150.7,
157.4, 164.7 ppm; MS (ESIꢁ, MeOH) m/z (%): 204.1 (100) [MꢁH]ꢁ;
HRMS-ESIꢁ m/z [MꢁH]ꢁ calcd for C9H6N3O3ꢁ: 204.0409, found:
204.0403.
ChemMedChem 0000, 00, 1 – 12
ꢁ 2012 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemmedchem.org
&
9
&
ÞÞ
These are not the final page numbers!

I. Artaud et al.
MED
the sample was eluted with a linear NaCl gradient (0.2mhꢁ1
,
Table 3. X-ray data collection and refinement statistics.
2.5 mLminꢁ1). After analysis by capillary electrophoresis (Experion,
Bio-Rad), fractions from the main peak (containing EcMetAP) were
pooled and concentrated to 5 mL by ultrafiltration (Amicon Ultra-
15, Millipore). The sample was then further fractionated with
a HiLoad 16/60 Superdex-75 column (GE Healthcare) equilibrated
with chelexed buffer [25 mm HEPES (pH 7.5), 150 mm KCl, 15 mm
methionine]. The purified apo-protein was dialyzed against the
Parameters
HA5
HA7
Data collection
Space group
PDB ID
P21
4A6V
MnII
P21
4A6W
MnII
Metal ion
Unit cell parameters
a [ꢂ]
chelexed buffer for two more days, concentrated to 48 mgmLꢁ1
,
53.2
39.3
frozen with liquid nitrogen, and stored at ꢁ808C. Typical yields
were 30 mgLꢁ1 of culture. The molecular mass of the purified apo-
EcMetAP was 29194.99 Da as determined by ESIMS on an LTQ Or-
bitrap mass spectrometer, in agreement with the theoretical value
of 29199.6 Da.
b [ꢂ]
c [ꢂ]
a, b, g [8]
61.2
77.8
62.9
52.6
90, 107.6, 90
37.06–1.46
1.54–1.46
323041, 80700
98.3 (97.3)
4.0 (4.0)
90, 109.4, 90
31.25–1.46
1.57–1.46
151827, 38522
97.9 (96.2)
3.9 (3.9)
8.5 (4.7)
8.1 (21.6)
Resolution [ꢂ]
Outer resolution shell [ꢂ]
No. reflections (obsd, unique)
Completeness [%][a]
Multiplicity[a]
I/s(I)[a]
Inhibition assays: The apo-protein was washed free of methionine
using chelexed activity buffer [25 mm HEPES (pH 7.5), 150 mm KCl]
prior to all kinetics assays and refrozen until use. Just before the
assay, the apo-enzyme was activated by the addition of divalent
metal chloride salt (5 equiv) to 60 mm apo-enzyme for 10 min at
room temperature. Activations with FeII were performed under
anaerobic conditions under an inert atmosphere (glove box). A
direct spectrophotometric assay was used as described by Mitra
et al.[47] The methionine para-nitroanilide substrate (Met–pNA) was
purchased from Bachem Bioscience. The hydrolysis of Met–pNA
was monitored continuously at 308C by following the increase in
absorbance of pNA at l 405 nm for 10–30 min on a 96-well UV-star
microplate (Greiner) and microplate spectrophotometer (Power-
wave XS, Biotek). A typical inhibition assay contained activity
buffer [25 mm HEPES (pH 7.5), 150 mm KCl, 0.4 mm Met-pNA,
15 mm metal(II) chloride (or 15 mm FeCl2 + 30 mm sodium ascor-
bate)], activated enzyme diluted 20-fold, inhibitor at various con-
centrations, and 1.5% DMSO. The enzyme assay with iron(II) was
performed aerobically in the presence of sodium ascorbate as de-
scribed by Wang et al.[43] IC50 values were calculated by nonlinear
regression analysis of the normalized activity (ratio of initial rates
with or without inhibitor) against log[inhibitor concentration].
7.1 (3.4)
9.3 (27.5)
Rmerge [%][a,b]
Refinement:
Reflections (working, test)
R [%][a]
Rfree [%][a]
76648, 5569
16.2
22.3
36995 ,2545
15.6
22.6
RMSD:
Bond length [ꢂ]
Bond angle [8]
B-factor [ꢂ2]
0.022
1.978
19.7
0.022
1.974
16.4
Ramachandran statistics:
Most favored [%]
Allowed [%]
96.6
3.4
95.8
4.28
[a] Outer shell. [b] Rmerge =ShSi jIhiꢁhIhij/ShSiIhi, for which Ihi is the ith ob-
servation of the reflection h, and hIhi is the mean intensity of reflection h.
[c] Rfactor =Sj jFo jꢁjFc j j/jFo j ; Rfree. was calculated with a set of randomly
selected reflections (5%).
tural solutions.[50] Structures were achieved with the MOLREP pro-
gram using an available E. coli MetAP structure (PDB ID: 2MAT)[51]
as the search model. The model was fully refined and completed
from the data using REFMAC software.[52] Refinement statistics are
listed in Table 3. All residues fall in favorable regions of the Rama-
chandran plot. Selected bond distances are listed in table S2 (Sup-
porting Information). All protein structure images were generated
with PyMOL.[52]
Crystallization and data collection
Initial crystallization conditions were determined using Crystal
Screen and Index kits in 96-well sitting-drop plates (Hampton Re-
search) at 293 K. Final crystals were obtained independently at
293 K by the sitting-drop vapor diffusion method from 2 mL drops.
Inhibitors (HA5 and HA7) were added to concentrated enzyme
(12.75 mgmLꢁ1) at an inhibitor/enzyme ratio of 5:1. Mixtures of
protein–inhibitor complexes were incubated for 5 min with MnCl2
at various metal/enzyme ratios (1:1; 0.5:1; 0.25:1). Drops enclosed
1 mL protein solution mixed with 1 mL precipitant solution contain-
ing either 17–25% PEG 8000 and 100 mm HEPES buffer (pH 7.5) for
samples involving HA5 inhibitor, or 15–22% PEG 20000, 100 mm
MES buffer (pH 6.5) for samples involving HA7. Crystals were trans-
ferred to the mother liquor containing 10% glycerol prior to flash
freezing in liquid nitrogen. X-ray diffraction data from different
crystals were collected on the ID29 ESRF (Grenoble, France) or
PROXIMA1 SOLEIL (St. Aubin, France) beamlines and were pro-
cessed using MOSFLM[48] and SCALA.[49] The crystals belong to the
P21 space group for crystals involving HA7 and HA5, with one and
two molecule(s) per asymmetric unit respectively. Cell parameters
and data collection statistics are reported in Table 3.
Bacterial strains
Strains used in this study: Briefly, E. coli K-12 AG100 strain (argE3
thi-1 rpsL xyl mtl D (gal-uvrB) supE44), AG100A (acrABꢁ) derivatives
have been described elsewhere.[42]
Antibiotic susceptibility tests: Bacteria were grown in Mꢄller–
Hinton (MH) broth at 378C. Susceptibilities to MetAP inhibitors,
polymyxin B nonapeptide (PMBN), phenylalanine–arginine–b-naph-
thylamide (PAbN), norfloxacin, and chloramphenicol were deter-
mined by the broth dilution method as previously described.[42]
Minimal inhibitory concentrations (MICs) were determined with an
inoculum of 106 CFU in 1 mL MH broth containing twofold serial
dilutions of each antibiotic. Isolates were classified as susceptible,
intermediately susceptible, or resistant to the antibiotics tested ac-
cording to the Antibiogram Committee of the French Society for
Microbiology ([http://www.sfm-microbiologie.org/pages/?page=
746&id_page=182]). For combination assays, serial dilutions of
Structure solution and refinement: Structures, at a metal/enzyme
ratio of 1:1, were solved at a resolution of 1.5 ꢂ by single anoma-
lous diffraction (SAD). The CCP4i program was used to find struc-
&10
&
www.chemmedchem.org
ꢁ 2012 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ChemMedChem 0000, 00, 1 – 12
ÝÝ
These are not the final page numbers!

Hydroxamic Acid EcMetAP Inhibitors
[17] Q.-Q. Huang, M. Huang, F.-J. Nan, Q.-Z. Ye, Bioorg. Med. Chem. Lett.
2005, 15, 5386–5391.
[18] C. J. Marmion, D. Griffith, K. D. Nolan, Eur. J. Inorg. Chem. 2004, 3003–
3016.
[19] E. M. F. Muri, M. J. Nieto, R. D. Sindelar, J. S. Williamson, Curr. Med. Chem.
2002, 9, 1631–1653.
tested molecules were incubated in the presence of the PMBN at
10 or 20% of MICs previously determined. Norfloxacin and chlor-
amphenicol belonging to unrelated structural antibiotic groups
were used as internal standards. The results were scored after 18 h
at 378C and are expressed as mgmLꢁ1 (MIC).
[20] A. Sharma, G. K. Khuller, S. Sharma, Expert Opin. Ther. Targets 2009, 13,
753–765.
[21] B. Chevrier, H. D’Orchymont, C. Schalk, C. Tarnus, D. Moras, Eur. J. Bio-
chem. 1996, 237, 393–398.
[22] S. Benini, W. R. Rypniewski, K. S. Wilson, S. Miletti, S. Ciurli, S. Mangani, J.
Biol. Inorg. Chem. 2000, 5, 110–118.
[23] A. Boularot, C. Giglione, S. Petit, Y. Duroc, R. Alves de Sousa, V. Larue, T.
Cresteil, F. Dardel, I. Artaud, T. Meinnel, J. Med. Chem. 2007, 50, 10–20.
[24] S. Petit, Y. Duroc, V. Larue, C. Giglione, C. Leon, C. Soulama, A. Denis, F.
Dardel, T. Meinnel, I. Artaud, ChemMedChem 2009, 4, 261–275.
[25] A. J. Phillips, Y. Uto, P. Wipf, M. J. Reno, D. R. Williams, Org. Lett. 2000, 2,
1165–1168.
Supporting Information
Synthesis and characterization of all intermediates are detailed:
table S1: activity of MetAP inhibitors on E. coli strains; table S2: se-
lected bond angles and distances in the EcMetAP-Mn–inhibitor
complexes; figure S1: comparison of ligands binding at the free Ec-
MetAP-Mn active site under its dinuclear (A) and mononuclear (D)
forms with those observed in EcMetAP-Mn complexed with HA5
(B) and HA7 (C); figure S2: surface view from X-ray structure of Ec-
MetAP–HA7.
[26] G. Ernst, W. Frietze, R. Jacobs, E. Phillips (AstraZeneca), Nicotinic Acetyl-
choline Receptor Ligands, Int. Pat. WO/2005061510 (A1), 2005.
[27] R. N. Patel, L. Chu, R. Chidambaram, J. Zhu, J. Kant, Tetrahedron: Asym-
metry 2002, 13, 349–355.
Acknowledgements
ˆ
[28] M. Kmetic, B. Stanovnik, J. Heterocycl. Chem. 1995, 32, 1563–1565.
We thank Prof. J. Larrabee (Middlebury College, VT, USA), Dr. W. T.
Lowther (Howard Hughes Medical Institute, University of Oregon,
USA), and Prof. B. W. Matthews (Wake Forest University School of
Medicine, Winston-Salem, NC, USA) for the kind gift of the expres-
sion plasmid for EcMetAP (pET28b-EcMetAP), Dr. T. Meinnel (ISV,
CNRS Gif-sur-Yvette, France) for the plasmid (pXL1071) used in
crystallization assays, and Dr. C. Lꢀon (UMR 8601, Universitꢀ Paris
Descartes PRES Paris Citꢀ, France) for molecular modeling studies.
This work was supported by the CNRS (France) and Paris Des-
cartes University, and by Aix-Marseille University and IRBA. F.H. re-
ceived a PhD fellowship from Research Ministry and L.M. a post-
doctoral fellowship from ANR-2010 BLAN METABACT.
[29] L. Leite, M. Ramos, J. Bosco, M. P. Da Silva, A. Miranda, C. Fraga, E. Bar-
reiro, Farmaco 1999, 54, 747–757.
[30] P. S. Branco, S. Prabhakar, A. M. Lobo, D. J. Williams, Tetrahedron 1992,
48, 6335–6360.
[31] R. Urbanek, D. Brown, G. Steelman, W. Blackwell, S. Wesolowki, X. Wang
(AstraZeneca), Int. Pat. WO/2007078251 (A1), 2007.
[32] D. Leung, W. Du, C. Hardouin, H. Cheng, I. Hwang, B. F. Cravatt, D. L.
Boger, Bioorg. Med. Chem. Lett. 2005, 15, 1423–1428.
[33] R. Huisgen, J. Sauer, H. J. Sturm, J. H. Markgraf, Chem. Ber. 1960, 93,
2106–2124.
[34] K. Choi, A. D. Hamilton, J. Am. Chem. Soc. 2003, 125, 10241–10249.
[35] K. Yang, B. Lou, Mini-Rev. Med. Chem. 2003, 3, 349–360.
[36] Q. Huang, J. Mao, B. Wan, Y. Wang, R. Brun, S. G. Franzblau, A. P. Kozi-
kowski, J. Med. Chem. 2009, 52, 6757–6767.
[37] W. T. Lowther, B. W. Matthews, Chem. Rev. 2002, 102, 4581–4607.
[38] S.-X. Xie, W.-J. Huang, Z.-Q. Ma, M. Huang, R. P. Hanzlikc, Q.-Z. Ye, Acta
Crystallogr. 2006, D62, 425–432.
[39] D. S. Kalinowski, D. R. Richardson, Pharmacol. Rev. 2005, 57, 547–583.
[40] A. J. Copik, S. J. Swierczek, W. T. Lowther, V. M. D’souza, B. W. Matthews,
R. C. Holz, Biochemistry 2003, 42, 6283–6292.
Keywords: antibacterial agents
·
hydroxamic acids
·
inhibitors · methionine aminopeptidase · X-ray structures
[1] C. Giglione, C. A. Boularot, T. Meinnel, Cell. Mol. Life Sci. 2004, 61, 1455–
1474.
[41] S. J. Watterson, S. Mitra, S. I. Swierczek, B. Bennett, R. C. Holz, Biochem-
istry 2008, 47, 11885–11893.
[2] S. Y. Chang, E. C. McGary, S. Chang, J. Bacteriol. 1989, 171, 4071–4072.
[3] C. G. Miller, A. M. Kukral, J. L. Miller, N. R. Movva, J. Bacteriol. 1989, 171,
5215–5217.
[4] C. T. Supuran, A. Scozzafava, B. W. Clare, Med. Res. Rev. 2002, 22, 329–
372.
[42] L. Mamelli, S. Petit, J. Chevalier, C. Giglione, A. Lieutaud, T. Meinnel, I.
Artaud, J.-M. Pagꢅs, PLoS ONE 2009, 4, e6443.
[43] W.-L. Wang, S. C. Chai, M. Huang, H.-Z. He, T. D. Hurley, Q.-Z. Ye, J. Med.
Chem. 2008, 51, 6110–6120.
[44] H. Nikaido, Microbiol. Mol. Biol. Rev. 2003, 67, 593–656.
[45] W. T. Lowther, D. A. McMillen, A. M. Orville, B. W. Matthews, Proc. Natl.
Acad. Sci. USA 1998, 95, 12153–12157.
[5] V. M. D’Souza, B. Bennett, A. J. Copik, R. C. Holz, Biochemistry 2000, 39,
3817–3826.
[6] V. M. D’Souza, S. I. Swierczek, N. J. Cosper, L. Meng, S. Ruebush, A. J.
Copik, R. A. Scott, R. C. Holz, Biochemistry 2002, 41, 13096–13105.
[7] N. J. Cosper, V. M. D’Souza, R. Scott, R. C. Holz, Biochemistry 2001, 40,
13302–13309.
[46] F. Frottin, A. Martinez, P. Peynot, S. Mitra, R. C. Holz, C. Giglione, T. Mein-
nel, Mol. Cell. Proteomics 2006, 5, 2336–2349.
[47] S. Mitra, A. M. Dygas-Holz, J. Jiracek, M. Zertova, L. Zakova, R. C. Holz,
Anal. Biochem. 2006, 357, 43–49.
[8] Q.-Z. Ye, S.-X. Xie, Z.-Q. Ma, M. Huang, R. P. Hanzlik, Proc. Natl. Acad. Sci.
USA 2006, 103, 9470–9475.
[48] A. G. W. Leslie, Newsletter Prot. Crystallogr. 1992, 26, 27–33.
[49] P. Evans, Acta Crystallogr. D 2006, 62, 72–82.
[9] M. Huang, S.-X. Xie, Z.-Q. Ma, Q.-Q. Huang, F.-J. Nan, Q.-Z. Ye, J. Med.
Chem. 2007, 50, 5735–5742.
[50] Collaborative Computational Project Number 4, Acta Crystallogr. D Biol.
Crystallogr. 1994, 50, 760–766.
[10] X. V. Hu, X. Chen, K. C. Han, A. S. Mildvan, O. Jun, J. O. Liu, Biochemistry
2007, 46, 12833–12843.
[51] W. T. Lowther, A. M. Orville, D. T. Madden, S. Lim, D. H. Rich, B. W. Mat-
thews, Biochemistry 1999, 38, 7678–7688.
[11] J. A. Larrabee, C. H. Leung, R. Moore, T. Thamrong-Nawasawat, B. H.
Wessler, J. Am. Chem. Soc. 2004, 126, 12316–12324.
[12] S. Mitra, K. M. Job, L. Meng, B. Bennett, R. C. Holz, FEBS J. 2008, 275,
6248–6259.
[52] W. L. DeLano, The PyMOL Molecular Graphics System, 2002, http://
www.pymol.org (accessed March 15, 2012).
[13] V. M. D’Souza, R. C. Holz, Biochemistry 1999, 38, 11079–11085.
[14] S. C. Chai, W. L. Wang, Q.-Z. Ye, J. Biol. Chem. 2008, 283, 26879–26885.
[15] S. C. Chai, Q.-Z. Ye, Bioorg. Med. Chem. Lett. 2010, 20, 2129–2132.
[16] Q.-Z. Ye, S.-X. Xie, M. Huang, W.-J. Huang, J.-P. Lu, Z.-Q. Ma, J. Am. Chem.
Soc. 2004, 126, 13940–13941.
Received: February 6, 2012
Revised: March 7, 2012
Published online on && &&, 0000
ChemMedChem 0000, 00, 1 – 12
ꢁ 2012 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemmedchem.org
&
11
&
ÞÞ
These are not the final page numbers!

FULL PAPERS
F. Huguet, A. Melet, R. Alves de Sousa,
A. Lieutaud, J. Chevalier, L. Maigre,
P. Deschamps, A. Tomas, N. Leulliot,
J.-M. Pages, I. Artaud*
Great rings of five! New hydroxamic
acids linked to furan, oxazole, oxadia-
zole, imidazole, and indole heterocycles
were prepared and assayed against
E. coli methionine aminopeptidase
(EcMetAP). They showed selectivity for
MnII and FeII metalloforms. Two X-ray
crystal structures of oxazole hydroxamic
acid–EcMetAP-Mn complexes were
solved, and a molecular model of the
indole derivative–EcMetAP-Mn complex
was built.
&& – &&
Hydroxamic Acids as Potent Inhibitors
of FeII and MnII E. coli Methionine
Aminopeptidase: Biological Activities
and X-ray Structures of Oxazole
Hydroxamate–EcMetAP-Mn Complexes
&12
&
www.chemmedchem.org
ꢁ 2012 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ChemMedChem 0000, 00, 1 – 12
ÝÝ
These are not the final page numbers!