Discovery of the first inhibitors of bacterial enzyme D-aspartate ligase from Enterococcus faecium (Aslfm)

Article abstract of DOI:10.1016/j.ejmech.2013.06.017

The D-aspartate ligase of Enterococcus faecium (Asl_fm) is an attractive target for the development of narrow-spectrum antibacterial agents that are active against multidrug-resistant E. faecium. Although there is currently little available information regarding the structural characteristics of Asl_fm, we exploited the knowledge that this enzyme belongs to the ATP-grasp superfamily to target its ATP binding site. In the first design stage, we synthesized and screened a small library of known ATP-competitive inhibitors of ATP-grasp enzymes. A series of amino-oxazoles derived from bacterial biotin carboxylase inhibitors showed low micromolar activity. The most potent inhibitor compound 12, inhibits Asl_fm with a K_i value of 2.9 μM. In the second design stage, a validated ligand-based pharmacophore modeling approach was used, taking the newly available inhibition data of an initial series of compounds into account. Experimental evaluation of the virtual screening hits identified two novel structural types of Asl_fm inhibitors with 7-amino-9H-purine (18) and 7-amino-1H-pyrazolo[3,4-d]pyrimidine (30 and 34) scaffolds, and also with K_i values in the low micromolar range. Investigation the inhibitors modes of action confirmed that these compounds are competitive with respect to the ATP molecule. The binding of inhibitors to the target enzyme was also studied using isothermal titration calorimetry (ITC). Compounds 6, 12, 18, 30 and 34 represent the first inhibitors of Asl_fm reported to date, and are an important step forward in combating infections due to E. faecium.

Full text of DOI:10.1016/j.ejmech.2013.06.017

European Journal of Medicinal Chemistry 67 (2013) 208e220
Contents lists available at SciVerse ScienceDirect
European Journal of Medicinal Chemistry
journal homepage: http://www.elsevier.com/locate/ejmech
Original article
Discovery of the first inhibitors of bacterial enzyme
ligase from Enterococcus faecium (Aslfm)
D-aspartate
a
b
,
b
ꢀ c
ꢀ
**
ꢀ
Veronika Skedelj , Andrej Perdih
, Matjaz Brvar , Ana Kroflic ,
,e,f,g,h,i,
Vincent Dubbée d
j, Victoria Savage m, Alex J. O’Neill m, Tom Solmajer b,
k
,e,f,g,h,i,j
ꢀ
ꢀ c
Marija Bester-Rogac , Didier Blanot ,l, Jean-Emmanuel Hugonnet d
,
,
e,
f
,
g
,
h,
i,
j
,
,e,f,g,h,i,j
Sophie Magnet d
1, Michel Arthur d
,
Jean-Luc Mainardi d
,j, Jure Stojan n, Anamarija Zega a
,
e
,
f,
g
,
h,
i
,
*
a
ꢀ
ꢀ
Faculty of Pharmacy, University of Ljubljana, Askerceva 7, 1000 Ljubljana, Slovenia
b National Institute of Chemistry, Hajdrihova 19, 1000 Ljubljana, Slovenia
c
ꢀ
ꢀ
Faculty of Chemistry and Chemical Technology, University of Ljubljana, Askerceva 5, 1000 Ljubljana, Slovenia
d Centre de Recherche des Cordeliers, LRMA, Equipe 12, Université Pierre et Marie Curie, Paris 6, UMR S 872, Paris F-75006 France
e INSERM, U872, Paris F-75006 France
f Université Paris Descartes, Sorbonne Paris Cité, UMR S 872, Paris F-75006 France
g Institut Parisien de Chimie Moléculaire, équipe GOBS, Université Pierre et Marie Curie, Paris 6, France
h Centre National de la Recherche Scientifique, UMR 7201, Paris F-75006 France
i AP-HP, Hôpital Européen Georges Pompidou, Paris F-75015 France
j Rhode Island Hospital, Brown University, Providence, RI 02903-4923, USA
k Laboratoire des Enveloppes Bactériennes et Antibiotiques, University Paris-Sud, IBBMC, UMR 8619, 91405 Orsay, France
l CNRS, 91405 Orsay, France
m Antimicrobial Research Centre and School of Molecular & Cellular Biology, University of Leeds, Leeds LS2 9JT, UK
n Institute of Biochemistry, Medical Faculty, University of Ljubljana, Vrazov trg 2, 1000 Ljubljana, Slovenia
a r t i c l e i n f o
a b s t r a c t
Article history:
The D-aspartate ligase of Enterococcus faecium (Aslfm) is an attractive target for the development of
Received 28 February 2013
Received in revised form
1 June 2013
Accepted 2 June 2013
Available online 28 June 2013
narrow-spectrum antibacterial agents that are active against multidrug-resistant E. faecium. Although
there is currently little available information regarding the structural characteristics of Aslfm, we
exploited the knowledge that this enzyme belongs to the ATP-grasp superfamily to target its ATP binding
site. In the first design stage, we synthesized and screened a small library of known ATP-competitive
inhibitors of ATP-grasp enzymes. A series of amino-oxazoles derived from bacterial biotin carboxylase
inhibitors showed low micromolar activity. The most potent inhibitor compound 12, inhibits Aslfm with a
Keywords:
Ki value of 2.9
mM. In the second design stage, a validated ligand-based pharmacophore modeling
D
-Aspartate ligase
approach was used, taking the newly available inhibition data of an initial series of compounds into
account. Experimental evaluation of the virtual screening hits identified two novel structural types of
Aslfm inhibitors with 7-amino-9H-purine (18) and 7-amino-1H-pyrazolo[3,4-d]pyrimidine (30 and 34)
scaffolds, and also with Ki values in the low micromolar range. Investigation the inhibitors modes of
action confirmed that these compounds are competitive with respect to the ATP molecule. The binding of
inhibitors to the target enzyme was also studied using isothermal titration calorimetry (ITC). Compounds
6, 12, 18, 30 and 34 represent the first inhibitors of Aslfm reported to date, and are an important step
forward in combating infections due to E. faecium.
Enterococcus faecium
ATP binding site
Pharmacophore modeling
Kinetic measurements
Isothermal titration calorimetry
Antibacterial agents
Drug design
Ó 2013 Elsevier Masson SAS. All rights reserved.
Abbreviations: ATP, adenosine-50-triphosphate; ITC, isothermal titration calorimetry; MIC, minimal inhibitory concentration; RA, residual activity; PDB, Protein databank;
SARs, structureeactivity relationships.
* Corresponding author. Tel.: þ386 1 4769 673; fax: þ386 1 4258031.
** Corresponding author. Tel.: þ386 1 4760 376; fax: þ386 1 4760300.
E-mail addresses: andrej.perdih@ki.si (A. Perdih), anamarija.zega@ffa.uni-lj.si (A. Zega).
Present address: INSERM U1070, Laboratoire de Pharmacologie des Antiinfectieux, Université de Poitiers, Pôle Biologie Santé, 1 rue Georges Bonnet, 86022 Poitiers,
1
BP633, France.
0223-5234/$ e see front matter Ó 2013 Elsevier Masson SAS. All rights reserved.
http://dx.doi.org/10.1016/j.ejmech.2013.06.017

ꢀ
V. Skedelj et al. / European Journal of Medicinal Chemistry 67 (2013) 208e220
209
1. Introduction
1972 (Fig. 1), activates
D-aspartic acid as
b
-aspartyl-phosphate in an
ATP-dependent reaction, and links the activated amino acid to the
ε-amino group of -Lys in the cytoplasmic precursor UDP-MurNAc-
pentapeptide (Fig. 2) [12,13]. Structural studies have also indicated
that in the mature peptidoglycan of E. faecium, the -carboxyl group
of the -aspartic acid is partly amidated ( -iso-asparagine) [5,9].
The gene encoding the -aspartate ligase in E. faecium (Aslfm
has been identified and enzyme activity was assessed following
purification of the protein overexpressed in Escherichia coli. The
structure of hexapeptide product of Aslfm was determined by tan-
dem mass spectroscopy [14]. The aslfm gene was also expressed in
Enterococcus faecalis to demonstrate the catalytic activity of Aslfm
in vivo. Analysis of the specifity of Aslfm for its amino acid substrate
Despite the relentless increase in resistance to antimicrobial
agents of the most important pathogens, the number of new anti-
biotics approved for clinical use has shown a remarkable decline
over the past two decades. Enterococci are one of the most
important pathogens for problems of resistance, and particularly
Enterococcus faecium, which belongs to the group of ’ESKAPE’
pathogens that currently cause the majority of nosocomial in-
fections [1,2].
Enterococci are among the most prevalent pathogens that have
developed resistance to vancomycin, the so-called antibiotic of last
resort. Vancomycin targets the bacterial cell wall by forming
complexes with peptidoglycan precursors, thereby inhibiting
peptidoglycan biosynthesis [3]. Since the first reported cases of
vancomycin-resistant E. faecium in 1988, enterococcal resistance to
vancomycin has dramatically increased [4].
L
a
D
D
D
)
shows that the enzyme is highly specific for
the length of the side chain or substitution of the
a hydroxyl group were not tolerated by the enzyme.
D
-Asp since increase in
-amino group by
-iso-aspara-
a
D
gine did not serve as a substrate for the enzyme, which indicates
Bacterial peptidoglycan is an attractive drug target, as it
completely surrounds the bacterial cytoplasmic membrane and
thus provides mechanical protection against the turgor pressure of
the cytoplasm. It also serves as an attachment site for various sur-
face polymers that interact with host cells and the immune system,
and it is involved in cell division [5,6]. Peptidoglycan is a polymeric
that its presence in the cross-linking of E. faecium originates from
amidation of the
a-carboxyl group after incorporation into pre-
cursors by Aslfm [14].
Close homologs of Aslfm can be found in the genomes of 10
bacterial species that can produce precursors substituted by
but there are none in bacteria that contain directly cross-linked
peptidoglycan or side-chains containing -amino acids and
D-Asp,
structure that is composed of
b
-linked N-acetylglucosamine
L
(GlcNAc) and N-acetylmuramic acid (MurNAc) polysaccharide
chains, which are cross-linked through variable peptide stems. The
polymerization process is performed by glycosyltransferases that
glycine. Low-level sequence similarity has indicated that Aslfm is a
member of the ATP-grasp superfamily [14], which is characterized
by an unusual nucleotide-binding fold: the ATP-grasp fold. En-
zymes of this family catalyze the formation of a carbonenitrogen
bond by a common reaction mechanism, which requires Mg2þ and
ATP for the formation of an acylphosphate intermediate. The pro-
tein superfamily includes biotin carboxylase, glutathione synthe-
tase, carbamoyl phosphate synthetase, ribosomal protein S6
modification enzyme (RimK), urea amidolyase, tubulin-tyrosine
ligase, and three enzymes of purine biosynthesis [15]. Additional
catalyze the formation of b-1,4 bonds, and D,D-transpeptidases that
cross-link the glycan strands [7]. The peptidoglycan assembly
pathway involves a wide range of other bacterial enzymes that act
at three distinct sites in the bacterial cell: the cytoplasm, and the
inner and outer sides of the cytoplasmic membrane.
The synthesis of the UDP-MurNAc-pentapeptide precursor of
peptidoglycan begins in the cytoplasm and involves six Mur en-
zymes (MurA to MurF). The first two of these (MurA and MurB)
catalyze the formation of UDP-MurNAc, onto which the following
four enzymes (MurC to MurF) subsequently add five amino acids,
members include
-Ala- -lactate-forming and
glycopeptide-resistant Gram-positive bacteria. Aside from Aslfm
D
-alanine:
D
-alanine ligase and the closely-related
D
D
D
-Ala- -Ser-forming enzymes found in
D
,
with alternating
-Ala [8]. The pathogenic Gram-positive bacteria of the genera
Staphylococcus, Streptococcus and Enterococcus have a conserved
alanyl- -glutamyl- -lysyl- -alanyl- -alanine peptide stem onto
which a variable side chain is linked to the ε-amino group of -Lys
[3,5,9]. The side chain consists of one to seven amino acids from
both the - and -series. The activation of -amino acids and glycine
for incorporation into the peptide side-chain is performed by
aminoacyl-tRNA synthetases, which form aminoacyl-tRNAs that
are then transferred to the precursors by the enzymes of the Fem
transferase family [10,11].
L
- and
D
-configurations except for terminal
D
-Ala-
these are the only known members of the family that are involved
in peptidoglycan synthesis [16].
D
L-
As Aslfm belongs to the ATP-grasp superfamily, this aided our
efforts in the development of what are, to the best of our knowl-
edge, the first inhibitors of this target. Our two-stage drug design
process and the biochemical and biophysical characterization are
depicted in Scheme 1. In the first stage, we designed and screened a
small library of known ATP-competitive inhibitors of bacterial
members of the ATP-grasp superfamily. Among these, there were
amino-oxazoles that inhibited the target enzyme Aslfm with Ki
values in the low micromolar range. In the next design phase of our
investigation, a ligand-based pharmacophore model was built us-
ing the available data that relate to the inhibitory action of the
initial series, in order to identify novel inhibitors of the target
enzyme that have different scaffolds. This approach enabled us to
g
-
D
L
D
D
L
L
D
L
While members of the Fem family have been extensively stud-
ied and characterized, the incorporation of
peptidoglycan precursors has not been well explored. The
E. faecium -aspartate ligase, Aslfm, which was partially purified in
D-amino acids into
D
Fig. 1. Function of Aslfm in peptidoglycan cross-bridge formation in E. faecium.

ꢀ
210
V. Skedelj et al. / European Journal of Medicinal Chemistry 67 (2013) 208e220
Fig. 2. Reaction catalyzed by Aslfm
.
discover novel classes of Aslfm inhibitors that were further studied
and characterized by isothermal titration calorimetry. For selected
compounds, we also evaluated the antimicrobial activity.
reductive amination of (1H-indol-5-yl)methanamine with 3,5-
dibromobenzaldehyde (Scheme 2).
All of these synthesized compounds (Table 1) were assayed for
in-vitro inhibitory activity toward recombinant Aslfm using a py-
ruvate kinaseelactate dehydrogenase enzymatic assay [19e21].
The results for the whole series are presented as residual activities
2. Results and discussion
(RAs) of the enzyme in the presence of 50 or 100 mM of the target
compounds, and as Ki values with respect to ATP for the most active
compounds (Table 1).
2.1. Design stage 1: design and synthesis of amino-oxazoles and
amino-thiazoles, starting from inhibitors of members of the ATP-
grasp enzyme superfamily
With a Ki of 44 mM, compound 6 represents the first known
inhibitor of our target enzyme Aslfm (Table 1), and it was therefore
Due to the currently limited knowledge relating to the structural
characteristics of our target enzyme, we took advantage of the only
known specific of the 3D structure of Aslfm: its affiliation to the ATP-
grasp superfamily of enzymes [14]. In many cases, although there is
no significant primary sequence homology between the enzymes,
the structures of their ATP-binding domains are largely superim-
posable [17].
Table 1
Inhibitory activities of amino-oxazoles (5e12) and amino-thiazoles (13e14) against
Aslfm
.
Based on this information, we first, synthesized and screened a
small library of known ATP-competitive inhibitors 5e7 of the
bacterial biotin carboxylase which is also a member of the same
ATP-grasp superfamily of enzymes (Table 1) [18]. Evaluation of
their biological activities against purified Aslfm using a pyruvate
kinaseelactate dehydrogenase detection system revealed a prom-
ising inhibitor of the target enzyme (compound 6, Fig. 3A, Table 1).
As the aromatic moiety turned out to be important for the inhibi-
tory activity, our next objective was the incorporation of more
electron-rich amines, to enable additional potential interactions
with the Aslfm active site. Thus, we synthesized seven target com-
pounds 8e12 (Table 1), with these characteristics. Finally, the
amino-oxazole moiety was replaced by two amino-thiazoles, as
these were commercially available (compounds 13 and 14, Table 1).
The synthesis of the compounds reported in the present study is
outlined in Scheme 2. In the first step, hydrolysis of ethyl esters
with 2e5.2 equivalents of sodium hydroxide provided oxazole (1)
and thiazole (2 and 3) acids. These were subsequently converted to
target compounds 5e14 using TBTU-promoted coupling with
commercially available primary or secondary amines. For com-
pound 12, the appropriate secondary amine 4 was obtained prior
to the coupling reaction, by sodium cyanoborohydride-effected
Cmpd
X
O
R1
H
R2
R3
RA [%]a
75
5
6
O
H
43
Ki ¼ 44
mM
7
8
O
O
H
H
94
74
9
O
O
H
H
98
90
10
H
H
11
12
13
O
O
H
H
79
44*
Ki ¼ 2.9
mM
S
S
H
80
82
14
CH3
a
Residual activity (RA, %) of the enzyme at 250
m
M of the tested compounds,
M due to limited
except for compound 12 marked with *, which was tested at 100
m
Scheme 1. Outline of the design steps and subsequent biochemical and biophysical
solubility. Data are means of two independent experiments. Standard deviations
characterization of the hit compounds used in this study.
were within ꢀ10% of the mean.

ꢀ
V. Skedelj et al. / European Journal of Medicinal Chemistry 67 (2013) 208e220
211
enzyme, we replaced both phenyl rings by the pyridine moiety
(compound 8). However, the enzymatic inhibitory activity was
significantly decreased or was lost. This was the case for N-(pyridin-
4-ylmethyl)ethanamine derivative 9, where only one pyridine ring
was introduced into the molecule. The introduction of two bromine
atoms on one phenyl ring and the replacement of another phenyl
moiety with indole led to the most potent inhibitor of Aslfm of this
series: compound 12, with a Ki of 2.9
mM. Finally, compounds
containing a thiazole ring (13 and 14) did not show any significant
inhibition of the target enzyme.
2.2. Design stage 2: ligand-based pharmacophore modeling
approach and virtual screening directed toward novel chemical
classes of Aslfm inhibitors
Pharmacophore modeling is a widely used method in modern
drug discovery [22,23], and it is particularly useful when the three-
dimensional (3D) structure of the target has not been described.
Several software tools are available to derive 3D pharmacophore
models in an automated fashion. LigandScout is an established
structure-based or ligand-based pharmacophore generator, and it
was used for the pharmacophore modeling part of the present
study [24].
To begin with, two active compounds, 6 and 12 (Fig. 3A), were
used as the starting point for the ligand-based drug design (see
Methods for full molecular modeling details). For the generation of
the initial pharmacophore models for each of 6 and 12, 500 unique
conformations were calculated and aligned [25]. This yielded 10
different ligand-based pharmacophore models, and the one with
the highest score (0.9311) was used in the large-scale virtual
screening investigation. The pharmacophore-derived model is
depicted in Fig. 3B, and this consists of two hydrophobic interaction
spheres, reflecting both of the lipophilic moieties of the chosen
compounds 6 and 12, two hydrogen bond acceptors (Fig. 3B, red
spheres), which describe the interactions with the carbonyl oxygen
and with the 5-membered ring nitrogen, and finally one hydrogen
bond donor (Fig. 3B, green sphere) relative to the interactions of the
amine group attached to the heterocyclic 5-membered ring. The
exclusion volume spheres were also created to approximate the
steric circumferences of the Aslfm binding site of these compounds
(see Supplementary material Fig. S1 for the full pharmacophore
model). The discriminatory performance of the derived pharma-
cophore model was critically assessed by screening it against both
active molecules 6 and 12 and 100 decoy molecules which were
generated from both active compounds (50 decoys per active
molecule) using the Decoyfinder software [26]. The pharmaco-
phore model successfully identified both active compounds in the
correct conformation and none of the available decoy molecules
were identified as virtual hits.
Fig. 3. (A) Chemical structures of the amino-oxazole compounds 6 and 12 used for the
construction of the ligand-based pharmacophore model. (B) Active compounds 6 and
12 aligned with the derived ligand-based pharmacophore model. The green sphere
represents the hydrogen bond donor, the two red spheres the hydrogen bond accep-
tors. The two yellow spheres denote hydrophobic pharmacophoric features (left), and
the derived pharmacophore model with exclusion volumes is depicted as gray spheres
(right). (For interpretation of the references to color in this figure legend, the reader is
referred to the web version of this article.)
an important starting point for the further design and optimization
of these Aslfm inhibitors. Based on compounds 5e7 [18], and
promising initial inhibition data, we developed and synthesized a
focused library of additional seven target compounds (compounds
8e14; Table 1), to establish further structureeactivity relationships
(SARs). The data from the enzymatic inhibition assay clearly
showed that for retention of inhibitory activity toward the target
enzyme, secondary amines with aromatic substituents must be
coupled to 2-amino-1,3-oxazole-5-carboxylic acid. If primary het-
erocyclic amines were coupled instead, the enzymatic inhibitory
activity significantly decreased (compound 11) or was even lost
(compound 10). Replacing one phenyl ring with the aliphatic butyl
group (compound 5) resulted in very modest inhibition of Aslfm
,
whereas dibutylamine derivative 7 was devoid of Aslfm inhibitory
activity. To retain aromatic substituents and possibly obtain an
additional hydrogen bond between the compound and target
Scheme 2. Reagents and conditions: (a) NaOH, EtOH or H2O or THF, room temperature. (b) NaCNBH3, MeOH, room temperature, overnight. (c) TBTU, NMM, DMF, room temperature,
overnight.

ꢀ
212
V. Skedelj et al. / European Journal of Medicinal Chemistry 67 (2013) 208e220
Subsequently, a large-scale virtual screening campaign was
the disubstituted compounds (15e17) showed any inhibition of the
target enzyme. 7-Amino-1H-pyrazolo[3,4-d]pyrimidine 19 was also
a moderate inhibitor of Aslfm, with a residual activity of 79% at
performed by screening a library of approximately 5.5 million
commercially available compounds. This library of virtual com-
pounds was prepared as described in the Experimental section. The
compounds were aligned with the investigated validated pharma-
cophore using the alignment method available in the LigandScout
program [21,22]. The virtual screening campaign yielded 3200 hit
compounds. Several chemical classes were identified as fitting the
screening pharmacophore pattern, including: triazines, 9H-purines,
pyrazolo[3,4-d]pyrimidines, 1H-benzo[d]imidazoles, imidazoles,
triazoles, tetrazoles and thiazoles (Scheme 3). The hit compounds
were visually inspected for their fit with the pharmacophore model,
and the most promising compounds (15e28) were purchased for
further experimental investigations. The results of the inhibition
100 mM. From the biological activities of the compounds from this
series, the importance of the amino group in position 7 can also be
established, as it is present in both active compounds 18 and 19.
The conformation of hit compound 18 that was obtained in the
virtual screening procedure is depicted in Fig. 4. The compound
contained a 7-amino-9H-purine heterocyclic moiety that enabled
the fulfillment of the required hydrogen bond acceptoredonore
donor pattern in the pharmacophore model derived for active
compounds 6 and 12. Two alternative positions of the 7-amino-9H-
purine moiety flipped by 180ꢁ in compound 18 were identified by
the screening software. Due to the lack of structural information
regarding the ATP binding site of Aslfm, it is currently not feasible to
differentiate between the derived orientations (Fig. 4). The
substituted benzene moiety attached to the 7-amino-9H-purine of
hit molecule 18 enabled interactions with both hydrophobic
spheres. The phenyl moiety was positioned between the hydro-
phobic spheres, and the chlorine atoms were located in both
spheres. It should be noted that since investigated compounds with
two lipophilic moieties and containing the same acceptoredonore
donor interaction pattern were inactive in the Aslfm inhibition as-
says (e.g. compounds 15e17), the potential for another favorable
binding placement of an alternative positioning of this hydrophobic
moiety within the Aslfm binding site cannot be ruled out. We
anticipate that when structural data for the Aslfm enzyme become
available, this will be investigated in more detail.
assay performed at 100
mM for selected commercially available
compounds (except for compound 16 and 17 measured at 50
mM)
are presented in Table 2 (see also Supplementary material, Table S1
for the origins, details and vendor quality-control procedures for
these compounds, and Section 2 for the analysis performed and the
purity characterization).
After careful analysis of the conformational alignment to the
pharmacophore model of the virtual screening hits obtained, we
focused our attention on nine different heterocyclic scaffolds that
might serve as replacements for the amino-oxazoles interaction
pattern: 7-amino-9H-purines (compounds 15e18), 7-amino-1H-
pyrazolo[3,4-d]pyrimidine 19,1,3,4-thiadiazol-2-amine 20, thiazole
21, imidazole 22, pyridine-2-ol 23, pyrazine-2-ol 24, tetrazoles 25
and 26 and (1,2,5-oxadiazol-3-yl)-1H-benzo[d]imidazoles 27 and
28 (Scheme 2 and Table 2).
Based on the promising results from the virtual screening
campaign, we selected 12 additional commercially available 7-
amino-9H-purines (compounds 29e33) and 7-amino-1H-pyrazolo
[3,4-d]pyrimidines (compounds 34e40), to provide more infor-
mation about the SARs of the two discovered compound classes of
Aslfm inhibitors. The inhibition of the selected compounds 29e40
Only compounds from the first two classes were found to be
active against the enzyme (Table 2). From the group of 7-amino-9H-
purines, only compound 18 showed promising inhibitory activity
(Ki ¼ 3.2
mM), which was comparable to the previously described
inhibitor 12 from the series of amino-oxazoles. We can clearly see
the importance of only one benzyl group at position 5 of the
7-amino-9H-purine scaffold for the inhibitory activity, as none of
was assayed at 100
mM and the results are presented in Table 3 (see
also Supplementary material, Table S1 for the origins, details and
Scheme 3. Different scaffolds identified by the ligand-based virtual screening campaign, as suitable replacements of the amino-oxazole 6 and 12.

ꢀ
V. Skedelj et al. / European Journal of Medicinal Chemistry 67 (2013) 208e220
213
Table 2
Inhibitory activities of the selected commercially available compounds derived from ligand-based virtual screening campaign against Aslfm
.
Cmpd
Structure
RA [%]a
Cmpd
Structure
RA [%]a
15
83
22
100
16
94*
23
93
17
110*
24
93
18
19
20
47
Ki ¼ 3.2
25
92
92
89
mM
79
26
95
27
21
91
28
116
a
Residual activity (RA, %) of the enzyme at 100
mM of the tested compound, except for compounds marked with *, which were tested at 50
m
M due to limited solubility. Data
are means of two independent experiments. Standard deviations were within ꢀ10% of the mean.

ꢀ
214
V. Skedelj et al. / European Journal of Medicinal Chemistry 67 (2013) 208e220
7-amino-9H-purine scaffold, the importance of a sulfur atom at
position 5 is illustrated by the good inhibitory activity of compound
30. Furthermore, we noted that the 2,6-disubstitution of the phenyl
ring is a good choice for inhibitory activity. This could be attributed
to the more favorable position of the lipophilic chlorine atoms, and
consequently better molecular recognition of the hydrophobic area
within the ATP binding site.
Novel active compounds 30 and 34 from the 7-amino-9H-pu-
rines and 7-amino-1H-pyrazolo[3,4-d]pyrimidine chemical classes,
respectively, were also aligned to the pharmacophore model. These
conformations are shown in Supplementary material Fig. S2.
Similar observations were made as already described for com-
pound 18. Again, two flipped positions of the heterocyclic moieties
were detected, along with analogous positions of the substituted
hydrophobic benzene moiety.
Fig. 4. Two of the alternative conformations obtained of the active hit compound 18
from the 7-amino-9H-purine chemical class identified in the virtual screening
campaign using the LigandScout ligand-based pharmacophore model (exclusion vol-
umes not shown). The green sphere represents the hydrogen bond donor, and the red
spheres hydrogen bond acceptors, while the yellow spheres denote hydrophobic
pharmacophoric features. (For interpretation of the references to color in this figure
legend, the reader is referred to the web version of this article.)
2.3. Kinetic determination of inhibitory constants
Catalysis by Aslfm was determined to proceed according to the
so-called ter-bi reaction mechanism, as summarized in Scheme 4
[27].
vendor quality-control procedures for these compounds, and Sec-
tion 2 for the analysis performed and purity characterization).
Among the selected analogs, compounds 30 and 34 were seen to
be promising inhibitors of the target enzyme. In the case of the
In the first step, the ATP molecule is bound to the enzyme
(EA), followed by
D
-aspartate (EAB). Next, the
D-aspartate is
Table 3
Inhibitory activities of commercially available 7-amino-9H-purines (29e33) and 7-amino-1H-pyrazolo[3,4-d]pyrimidines (34e40) against Aslfm
.
Cmpd
Y
R4
RA [%]a
85
Cmpd
R5
R6
RA [%]a
69
29
NH
34
H
H
Ki ¼ 22
mM
30
S
71
35
86
Ki ¼ 35
mM
31
32
33
NH
NH
NH
89
83
84
36
37
H
H
84
93
38
39
40
H
88
74
87
H
CH3
a
Residual activity (RA, %) of the enzyme at 100 mM of the tested compound (Cmpd). Data are means of two independent experiments. Standard deviations were within
ꢀ10% of the mean.

ꢀ
V. Skedelj et al. / European Journal of Medicinal Chemistry 67 (2013) 208e220
215
To decrease the number of measurements without sacrificing the
amount of information for reliable evaluation, we followed the
progress of the reaction in the absence and presence of putative
inhibitors for longer periods of time (10e15 min). Subsequently, we
analyzed the progress curves obtained with ENZO [28], a web tool
written in java, and designed for automatic generation and rapid
analysis of data by differential equations from different proposed
reaction schemes, with the purpose being to find a suitable one for
Aslfm. In the first step, we analyzed the data in the absence of in-
hibitors, to obtain characteristic dissociation and rate constants for
the Aslfm reaction itself [40]. In the second step, we fixed the
determined constants from the previous step and evaluated only
those characteristic for each individual tested inhibitor [41]. The
agreement between experimental curves and the theoretical ones
as depicted in Fig. 5 for compounds 6 and 18 strongly supports the
ATP-competitive effects of investigated compounds (6, 12, 18, 30,
34) all displaying analogous behavior. The estimated inhibition
constants are summarized in Tables 1e3.
Scheme 4. Putative reaction mechanism of Aslfm. In the scheme E represents free
enzyme, A is ATP, B is D-aspartic acid, C is phosphorylated D-aspartic acid, D is ADP, PP
is UDP-MurNAc-pentapeptide and I stands for the inhibitor. The substrates written on
the right side represents productive complexes while on the left denotes non-
productive complexes. All complexes with asterisk are assumed to evolve upon
isomerization.
2.4. Antimicrobial activity
The antimicrobial activities of compounds 6, 12, 18, 30 and 34
were evaluated against three enterococcal strains: E. faecium E1679,
E. faecium E1636 and E. faecalis ATTC 29212. For all of these com-
pounds, the minimum inhibitory concentrations (MICs) against all
phosphorylated (EDC) and an amide bond is formed upon the
approach of UDP-MurNAc-pentapeptide (EDCPP), by substitution of
the phosphate with the ε-amino group of the lysine at position 3 of
UDP-MurNAc-pentapeptide. The catalytic cycle is completed by an
ordered release of newly synthesized hexapeptide, ADP and inor-
ganic phosphate. The shape of the curves under continuous mea-
surement suggests that slow conformational changes in Aslfm go
with the binding and release of the participating peptides.
Accordingly, the nonproductive binding of UDP-MurNAc-pentape-
tide is unmasked when the binding order is violated. In the pres-
ence of ATP-competitive inhibitors 6, 12, 18, 30 and 34, the time
course of product formation shows the expected double character:
a moderate initial instantaneous decrease in activity is followed by
a slow, but substantially stronger, inhibition, which results in
sloped asymptotes (see Fig. 5 for selected compounds 6 and 18).
Kinetic analysis of such a three-substrate reaction, coupled with
two additional enzyme reactions for detection, is very complex.
three strains were >64
mg/mL. The antimicrobial activities of
compounds 5e14 were also tested against the clinical strain
E. faecium D344R in brain-heart infusion agar, using the disk
diffusion method. No zone of inhibition was observed. Activity was
further assessed in liquid medium using the macrodilution method.
No growth inhibition was observed up to the maximal tested
concentration (200
mg/mL), even in the presence of subinhibitory
concentrations of ampicillin (2
mg/mL) (Supplementary material
Table S3). The lack of activity of the compounds against these
tested bacteria might be due to problems associated with entry into
the bacterial cells. As our work presents the initial hit/lead phase of
the drug design process the physical properties of these com-
pounds were further assessed by comparing their Clog P and Clog D
(at pH ¼ 7.4) values (see Supplementary material Table S4). The
range of the observed values between ꢂ0.5 and 4 for all compounds
Fig. 5. Time courses of product formation in the enzyme reaction by Aslfm in the absence and presence of inhibitor 6 (A) and 18 (B). The concentrations of 6 (A) were: curve A, zero;
B, 50 M; C, 75 M; D, 100 M; and E, 150 M. The concentrations of 18 (B) were: curve A zero, B 10 M, C 25 M, D 50 M, E 75 M and 125 M. The starting concentrations of the
reactants were for both cases: 100 M ATP, 1 mM D-aspartate, 300 M UDP-MurNAc-pentapeptide, 2 mM phosphoenolpyruvate, 105 M NADH. The added concentration of Aslfm
was 0.75 M (6) or 1 M (18), and those of pyruvate kinase and lactate dehydrogenase were 0.3 mM and 0.02 mM, respectively. Solid lines indicate the calculated curves based on
the ter-bi reaction mechanism using ENZO program [28].
m
m
m
m
m
m
m
m
m
m
m
m
m
m

ꢀ
216
V. Skedelj et al. / European Journal of Medicinal Chemistry 67 (2013) 208e220
Table 4
compounds with comparable or preferably lower lipophilicity
character. The information that a single diclorophenyl substituent
in compound 18 instead of two phenyl rings in compound 6 yielded
comparable inhibitory activity is providing much optimism that
affinity of these compounds is not driven purely by lipophilicity. For
further rationally-driven optimization of these hits additional
structural insights into their molecular recognition by the Aslfm
enzyme will be of vital importance.
Determined standard thermodynamic parameters (enthalpy,
and Gibbs free energy,
GS0b) of the binding of 6 and 18 to Aslfm, and the corre-
sponding binding constants, Kb, obtained by isothermal titration calorimetry.a
Hb0; entropy, S0b;
D D
D
Cmpd
Hb0=Kcal molꢂ1 Kb/Mꢂ1 Ki/ Gb0=Kcal molꢂ1 S0b=Kcal molꢂ1
mM T
D D D
6
18
ꢂ4.2
2 ꢃ 105
2 ꢃ 105
5
5
ꢂ7.6
ꢂ7.6
3.4
ꢂ6.2
ꢂ13.8
a
Ki ¼ 1=Kb;
D
G0b ¼ RT⁡lnðKb=Mꢂ1Þ;
D
Gb0
¼
D
Hb0 ꢂ T S0b.
D
3. Conclusions
except compounds 15 and 17 indicated the drug-like nature of
these compounds. Furthermore, as interplay of several additional
factors (e.g. permeability, efflux) could be responsible for the lack of
activity of these compounds this aspect remains to be investigated
in the future.
By combining classical screening and ligand-based pharmaco-
phore modeling approaches in two drug-design stages, we have
successfully identified and experimentally characterized the first
inhibitors of the bacterial enzyme D-aspartate ligase from Entero-
coccus faecium. Due to the limited knowledge of the structure of
Aslfm, we focused on the ATP binding site of the enzyme and
screened a small collection of known ATP-competitive inhibitors of
the large family of the ATP-grasp enzymes. A series of amino-
oxazoles and thio-oxazoles derived from bacterial biotin carbox-
ylase inhibitors showed low micromolar activity, with the most
2.5. Biophysical characterization of the hit compounds 6 and 18 by
isothermal titration calorimetry
Two of the compounds, 6 and 18, were tested by isothermal
titration calorimetry to obtain thermodynamic parameters of their
binding into the Aslfm active site at 37 ꢁC. Heat changes upon
binding were measured directly, corrected for the heats of dilution,
and the model function was fitted to the experimental data points
(see Experimental section).
potent inhibitor (12) with a Ki of 2.9 mM with respect to ATP. In the
second design stage, ligand-based pharmacophore modeling ap-
proaches using newly acquired data for the inhibitory activities of
the initially discovered series, followed by virtual screening,
resulted in identification of new structural types of Aslfm inhibitors.
This identified novel ATP-competitive inhibitors of Aslfm based on
7-amino-9H-purines (18) and 7-amino-1H-pyrazolo[3,4-d]pyrimi-
dine (30 and 34) scaffolds with Ki values in the low micromolar
range. The binding of two selected inhibitors (6 and 18) was sub-
sequently confirmed and studied by isothermal titration calorim-
etry, with determination of the standard thermodynamic
parameters of the inhibitor binding to the enzyme active site.
Compounds 6, 12, 18, 30 and 34 have Ki values in the low micro-
molar range, and these represent the first inhibitors of Aslfm re-
ported to date. We believe that this set of inhibitors discovered here
will provide novel lead compounds in antibacterial drug design
efforts to combat E. faecium infections.
Experiments for the two investigated compounds (6 and 18)
yielded similar binding constants (Kb z 105
M
ꢂ1) that are in full
agreement with their Ki values (Table 2) obtained in the kinetic
analysis using ter-bi reaction mechanism. Interestingly, the ITC
results also revealed that the driving forces governing the binding
of these two different compound classes differ considerably
(Table 4 and Fig. 6). The binding of compound 6 to Aslfm results
from comparable and favorable enthalpic (
entropic (T
S ¼ 3.4 kcal/mol) contributions. On the other hand, the
binding of compound 18 is enthalpy driven (
H ¼ ꢂ13.8 kcal/mol)
at the temperature investigated, and accompanied by a negative
DH ¼ ꢂ4.2 kcal/mol) and
D
D
entropy change (T
of the opposing effects at 18 results in a binding affinity ð
D
S ¼ ꢂ6.2 kcal/mol). Nevertheless, compensation
D
G0bÞ
similar to that of 6, although there are presumably more non-
covalent bonds formed between 18 and the protein than for 6.
Compound 18 obtained in the second drug design stage appears to
be more suitable for further optimization, as the increased rigidity
of the molecule in its binding conformation should improve the
entropy contributions upon retaining the specific and oriented in-
teractions within the Aslfm active site [29,30]. Nevertheless,
comparing of the Clog P values for compound 6 (Clog P ¼ 2.4) and
compound 18 (Clog P ¼ 4.0) a caveat must be stated that subse-
quent optimization of this compound class must proceed toward
4. Experimental section
4.1. Chemistry
Chemicals were obtained from Acros, Aldrich Chemical Co.,
Apollo Scientific, and Fluka, and they were used without further
purification. Analytical thin-layer chromatography was performed
on silica gel Merck 60 F254 precoated plates (0.25 mm), visualized
with ultraviolet light, ninhydrin and 2,4-dinitrophenylhydrazine.
Fig. 6. Experimental titration curves and fits according to the single binding site model for binding of compounds 6 (a) and 18 (b) to the Aslfm active site at 37 ꢁC. Scheme Caption.

ꢀ
V. Skedelj et al. / European Journal of Medicinal Chemistry 67 (2013) 208e220
217
Flash column chromatography was carried out on silica gel 60
(particle size 0.040e0.063 mm; Merck, Germany). Melting points
were determined on a Reichert hot-stage microscope and are not
corrected. 1H NMR spectra were recorded on a Bruker Avance III
400 MHz spectrometer at 295 K and 400 MHz, and are reported in
ppm using solvent as internal standard (DMSO-d6 at 2.50 ppm,
CDCl3 at 7.26 ppm). 13C NMR spectra were recorded on a Bruker
Avance III 400 MHz spectrometer at 295 K and 100 MHz, and are
reported in ppm using solvent as the internal standard (DMSO-d6 at
39.5 ppm). Mass spectra data were recorded using a Q-Tof Premier
instrument (Waters-Micromass, Manchester, UK). HPLC analyses
were performed on an Agilent Technologies HP 1100 instrument,
with a G1365B UVevis detector, a G1316A thermostat, and a
J ¼ 8.4, 1.6 Hz, 1H), 7.25 (t, J ¼ 2.7 Hz, 1H), 7.29 (s, 1H, NHCH2), 7.40
(d, J ¼ 8.4 Hz, 1H, Ar-H), 7.48 (d, J ¼ 2.0 Hz, 2H, Ar-2,6), 7.57
(t, J ¼ 2.0 Hz, 1H, Ar-4-H), 7.59e7.65 (m, 1H, indole-H), 10.29 (br s,
1H, indole-H-1) ppm. HRMS (ESI): m/z [M þ H]þ calcd for
C16H15N2Br2, 395.1254; found, 395.1259.
4.1.5. Synthesis of compounds 5e14
The suspension of acid (2.00 mmol) in DMF (10 mL) was cooled
to 0 ꢁC in an ice bath. N-methylmorpholine (0.49 mL, 4.40 mmol)
and TBTU (0.84 g, 2.60 mmol) were added, and the reaction mixture
was stirred at 0 ꢁC for 0.5 h. Then it was allowed to reach room
temperature, and amine (2.00 mmol) was added to the solution.
The reaction mixture was stirred overnight at room temperature,
after which the solvent was evaporated under reduced pressure.
The residue was dissolved in ethyl acetate (40 mL) and washed with
a saturated aqueous solution of NaHCO3 (3 ꢃ 20 mL), 10% citric acid
(3 ꢃ 20 mL) and brine (20 mL). The organic phase was dried over
Na2SO4, and filtered, and the solvent was removed under reduced
pressure. The crude product was purified by crystallization or flash
column chromatography.
G1313A autosampler, using a Phenomenex Luna 5 mM C18 column
(4.6 mm ꢃ 150 mm) at a flow rate of 1.0 mL/min. The eluent con-
sisted of 0.1% trifluoroacetic acid in water (A) and methanol (B). The
gradient was 20% B to 80% B in 20 min. The purity of the tested
compounds was established to be ꢄ95%.
4.1.1. Synthesis of compound 1
To the suspension of the oxazole ethyl ester (3.13 g, 20.0 mmol)
in water, 1 M NaOH (0.04 mL, 40.0 mmol) was added. The mixture
was stirred at room temperature for 3 h, and afterward adjusted to
pH 2e3 with 1 M HCl. The resulting precipitate 1 was filtered,
washed with diethyl ether and dried overnight at 60 ꢁC. Yield: 93%;
white crystals, mp 210e212 ꢁC. 1H NMR (400 MHz, DMSO-d6): 7.01
(s, 2H, NH2), 7.13 (s. 1H, Ar-H), 12.11 (br s, 1H, COOH) ppm. HRMS
(ESI): m/z [M þ H]þ calcd for C4H5N2O3 129.1347; found 129.1537.
4.1.5.1. 2-Amino-N-benzyl-N-butyloxazole-5-carboxamide
(5).
The crude product was crystallized from CH2Cl2. Yield: 46%, off-
white crystals, mp 134e137 ꢁC. 1H NMR (400 MHz, DMSO-d6):
d
0.85 (t, J ¼ 7.4 Hz, 3H, CH3), 1.19e1.31 (m, 2H, CH2CH3), 1.45e1.58
(m, 2H, CH2CH2CH2), 3.29e3.43 (m, 2H, CH2CH2CH2), 4.69 (s, 2H,
Ar-CH2), 7.18 (br s, 2H, NH2), 7.21e7.40 (m, 6H, Ar-H þ oxazole-H)
ppm. 13C NMR (400 MHz, DMSO-d6):
d 13.69, 19.45, 29.80, 46.45,
49.28, 127.05, 128.53, 133.81, 137.32, 137.86, 158.04, 162.45 ppm.
HRMS (ESI): m/z [M þ H]þ calcd for C15H20N3O2, 274.1556; found,
274.1547. HPLC tR ¼ 15.965 min (96.22% at 220 nm, 98.78% at
254 nm).
4.1.2. Synthesis of compound 2
The thiazole ethyl ester (1.0 g, 5.81 mmol) was treated with
0.5 M ethanolic sodium hydroxide solution (60 mL), overnight at
room temperature. The reaction mixture was then cooled in an ice
bath, and neutralized with acetic acid. The precipitated acid 2 was
filtered, washed with diethyl ether, and dried overnight at 60 ꢁC.
Yield: 81%; off-white crystals, mp 225e227 ꢁC (lit. [31] 223 ꢁC). 1H
NMR (400 MHz, DMSO-d6): 2.35 (s, 3H, CH3), 7.61 (s, 2H, NH2),12.26
(br s, 1H, COOH) ppm. HRMS (ESI): m/z [M þ H]þ calcd for
C5H7N2O2S 159.1698; found 159.1663.
4.1.5.2. 2-Amino-N,N-dibenzyloxazole-5-carboxamide
crude product was crystallized from CH2Cl2. Yield: 57%, white
crystals, mp 216e219 ꢁC. 1H NMR (400 MHz, DMSO-d6):
5.77 (s,
4H, 2ꢃ CH2), 7.12 (s, 1H, oxazole-H), 7.18e7.42 (m, 12H, Ar-H þ NH2)
ppm. 13C NMR (400 MHz, DMSO-d6):
49.50, 127.03, 127.81, 128.63,
HRMS (ESI): m/z
(6). The
d
d
134.40, 136.89, 137.21, 158.47, 162.66 ppm.
d
[M þ H]þ calcd for C18H18N3O2, 308.1399; found, 308.1389. HPLC
tR ¼ 16.662 min (98.19% at 220 nm, 99.34% at 254 nm).
4.1.3. Synthesis of compound 3
To the solution of the thiazole ethyl ester (0.80 g, 4.32 mmol) in
THF/water (2:1), NaOH was added (0.65 g, 16.3 mmol), and the
reaction mixture was refluxed overnight. THF was removed under
reduced pressure and afterward the reaction mixture was
neutralized with 1 M HCl. The resulting precipitate 1 was filtered,
washed with diethyl ether and dried overnight at 60 ꢁC. Yield: 86%;
off-white crystals, mp 168e171 ꢁC (lit. [32] 172e173 ꢁC). 1H NMR
4.1.5.3. 2-Amino-N,N-dibutyloxazole-5-carboxamide (7). The crude
product was crystallized from CH2Cl2. Yield: 38%, yellow crystals,
mp 121e124 ꢁC. 1H NMR (400 MHz, DMSO-d6):
d
0.89 (t, J ¼ 7.4 Hz,
6H, 2ꢃ CH3), 1.22e1.34 (m, 4H, 2ꢃ CH2CH3), 1.45e1.56 (m, 4H,
CH2CH2CH2), 3.29e3.44 (m, 4H, CH2CH2CH2), 7.18 (br s, 2H, NH2),
7.23 (s, 1H, oxazole-H) ppm. 13C NMR (400 MHz, DMSO-d6): 13.74,
19.52, 30.09, 46.41, 133.14, 137.72, 157.52, 162.22 ppm. HRMS (ESI):
m/z [M þ H]þ calcd for C12H22N3O2, 240.1712; found, 240.1710.
HPLC tR ¼ 16.210 min (98.18% at 220 nm, 98.87% at 254 nm).
(400 MHz, DMSO-d6):
d 7.81 (s, 1H, oxazol-H), 8.99 (br s, 2H, NH2)
ppm. HRMS (ESI): m/z [M þ H]þ calcd for C4H5N2O2S 145.1475;
found 145.1502.
4.1.4. Synthesis of compound 4
4.1.5.4. 2-Amino-N,N-bis(pyridin-2-ylmethyl)oxazole-5-carboxamide
(8). The crude product was purified with flash column chroma-
tography using chloroform/methanol (9:1) as eluent. Yield: 21%,
yellow crystals, mp 125e128 ꢁC. 1H NMR (400 MHz, DMSO-d6):
To the solutions of (1H-indol-5-yl)methanamine (0.50 g,
3.42 mmol) and 3,5-dibromobenzaldehyde (0.90 g, 3.42 mmol) in
methanol, NaCNBH3 (0.43 g, 6.84 mmol) was added, and the
mixture was stirred overnight under an argon atmosphere. The
solvent was removed under reduced pressure and afterward the
residue was dissolved in ethyl acetate (50 mL) and washed with
water (3 ꢃ 20 mL). The organic phase was dried over Na2SO4, and
filtered, and the solvent was removed under reduced pressure. The
crude product was purified by flash column chromatography using
petrolether/ethyl acetate (3:1) as eluent. Yield: 70%, off-white
d
4.82 (s, 4H, 2ꢃ CH2), 7.20 (s, 1H, oxazole-H), 7.23e7.40 (m, 6H, pyr-
H þ NH2), 7.72e7.82 (m, 2H, pyr-H), 8.49e8.57 (m, 2H, pyr-H) ppm.
13C NMR (400 MHz, DMSO-d6):
2.28,121.56, 122.41, 134.32, 136.91,
d
137.01, 149.23, 156.98, 158.60, 162.63 ppm. HRMS (ESI): m/z
[M þ H]þ calcd for C16H16N5O2, 310.1304; found, 310.1309. HPLC
tR ¼ 15.965 min (96.22% at 220 nm, 98.78% at 254 nm).
crystals, mp 186e188 ꢁC. 1H NMR (400 MHz, CDCl3):
CH2), 3.91 (s, 2H, CH2), 6.54e6.59 (m, 1H, indole-H-3), 7.20 (dd,
d
3.79 (s, 2H,
4.1.5.5. 2-Amino-N-ethyl-N-(pyridin-4-ylmethyl)oxazole-5-
carboxamide (9). The crude product was purified with flash column

ꢀ
218
V. Skedelj et al. / European Journal of Medicinal Chemistry 67 (2013) 208e220
chromatography using chloroform/methanol (9:1) as eluent. Yield:
128.63, 132.32, 137.06, 151.04, 164.75, 167.77 ppm. HRMS (ESI): m/z
[M þ H]þ calcd for C19H20N3OS, 338.1327; found, 338.1321. HPLC
tR ¼ 15.613 min (100.0% at 220 nm, 100.0% at 254 nm).
39%, pale yellow crystals, mp 162e166 ꢁC. 1H NMR (400 MHz,
DMSO-d6):
d
1.15 (t, J ¼ 6.8 Hz, 3H, CH2CH3), 3.48 (q, J ¼ 6.8 Hz, 2H,
CH2CH3), 4.69 (s, 2H, CH2), 7.11 (s, 1H, oxazole-H), 7.24 (d, J ¼ 6.1 Hz,
pyr-H-3,5), 7.28 (br s, 2H, NH2), 8.52 (d, J ¼ 6.1 Hz, pyr-H-2,6) ppm.
4.2. Ligand-based pharmacophore modeling and virtual screening
procedures
13C NMR (400 MHz, DMSO-d6):
d 13.49, 42.35, 48.39, 122.04, 134.17,
137.04, 147.31, 149.68, 157.96, 162.60 ppm. HRMS (ESI): m/z
[M þ H]þ calcd for C12H15N4O2, 247.1195; found, 247.1203. HPLC
tR ¼ 9.277 min (97.09% at 220 nm, 98.99% at 254 nm).
Ligand-based pharmacophore modeling and virtual screening
were performed using the LigandScout software [21]. The initial
constructed conformations of the active compounds 6 and 12 were
minimized using the MMFF94 force field, and imported into the
ligand-based module available in Ligandscout. Five-hundred
unique conformations were calculated for each structure by
applying the following settings of the LigandScout conformer
generator coupled to the OMEGA software: maximum number of
output conformers per molecule ¼ 500; RMS threshold to duplicate
4.1.5.6. 2-Amino-N-(1H-benzo[d]imidazol-2-yl)oxazole-5-
carboxamide (10). The crude product was purified with flash col-
umn chromatography using ethyl acetate/methanol (8:1) as eluent.
Yield: 26%, off-white crystals, mp > 300 ꢁC. 1H NMR (400 MHz,
DMSO-d6):
d
7.11 (br s, 2H, NH2), 7.36e7.51 (m, 5H, Ar-H þ oxazole-
H), 7.77 (s, 1H, CONH) ppm. 13C NMR (400 MHz, DMSO-d6):
d
114.6,
ꢁ
123.7, 135.8, 136.2, 142.5, 147.4, 162.7, 166.2 ppm. HRMS (ESI): m/z
[M þ H]þ calcd for C11H10N5O2, 244.0834; found, 244.0838. HPLC
tR ¼ 18.464 min (98.65% at 220 nm, 99.09% at 254 nm).
conformers ¼ 0.4 A; maximum number of all generated conformers
per molecule ¼ 30,000; and maximum number of intermediate
conformers per molecule ¼ 4000. Subsequently, the conformers
were dynamically aligned [22] to yield 10 ligand-based merged
pharmacophore models. A scoring function that combined phar-
macophore fit and atom shape overlap was used to assess the
models produced. The ligand-based pharmacophore model with
the highest score of 0.9311 was selected for further use (see
Supplementary material, Fig. S1, for a complete merged Pharma-
cophore model). Visual inspection revealed that most of the
remaining nine derived pharmacophore models closely resembled
this model. The initial number of available merged pharmacophoric
features was reduced to obtain on the one hand sufficient molec-
ular recognition pattern required for the ligand binding, and to
increase on the other hand the chemical space identified by the
pharmacophore model [33]. The reduced ligand-based pharmaco-
phore model was used in the large-scale virtual screening
campaign. The derived pharmacophore model (see Fig. 3) consisted
of two hydrophobic interaction spheres, two hydrogen bond ac-
ceptors, and one hydrogen bond donor. Exclusion volume spheres
were also added, to approximate the steric circumference of the
Aslfm binding site. The discriminatory performance of the derived
pharmacophore model was validated by a screening experiment
against 100 decoy molecules generated for both active compounds
6 and 12 (50 decoys/molecule) using the Decoyfinder software [26].
Decoyfinder generates sets of decoy molecules for defined active
ligands. Decoys have similar number of rotational bonds, hydrogen
bond acceptors, hydrogen bond donors, log P value and molecular
weight as active molecules but are chemically different, which is
defined by a maximum Tanimoto value threshold between active
ligand and decoy molecule MACCS fingerprints [26]. The pharma-
cophore model successfully identified the both active compounds
in the correct orientation and none of the generated decoy mole-
cules were identified as potential hits.
4.1.5.7. N-((1H-Indol-5-yl)methyl)-2-aminooxazole-5-carboxamide
(11). The crude product was purified with flash column chroma-
tography using chloroform/methanol (9:1) as eluent. Yield: 46%,
light-brown crystals, mp 207e211 ꢁC. 1H NMR (400 MHz, DMSO-
d6):
d
4.44 (d, J ¼ 6.0 Hz, 2H, CH2), 6.36e6.40 (m, 1H, indole-H-3),
7.08 (dd, J ¼ 8.4, 1.6 Hz, 1H), 7.21 (br s, 2H, NH2), 7.33e7.42 (m,
2H, oxazole-H þ indole-H), 7.45 (s, 1H, indole-H), 7.49 (s, 1H,
indole-H), 8.46 (t, J ¼ 6.0 Hz, 1H, CONH), 11.09 (br s, 1H, indole-H-1)
ppm. 13C NMR (400 MHz, DMSO-d6):
d 42.19, 100.85, 111.16, 118.69,
120.97, 125.54, 127.51, 129.81, 131.32, 134.96, 138.36, 156.90,
162.43 ppm. HRMS (ESI): m/z [M þ H]þ calcd for C13H13N4O2,
257.1039; found, 257.1032. HPLC tR ¼ 8.482 min (98.97% at 220 nm,
99.54% at 254 nm).
4.1.5.8. N-((1H-Indol-5-yl)methyl)-2-amino-N-(3,5-dibromobenzyl)
oxazole-5-carboxamide (12). The crude product was purified with
flash column chromatography using chloroform/ethyl acetate
(1:10) as eluent. Yield: 68%, pink crystals, mp 187e191 ꢁC. 1H
NMR (400 MHz, DMSO-d6):
d 4.62 (s, 2H, CH2), 4.80 (s, 2H, CH2),
6.40 (m, 1H, indole-H-3), 6.98 (br s, 2H, NH2), 7.16 (s, 1H, oxazol-
H), 7.27e7.49, (m, 6H, Ar-H), 7.72 (s, 1H, Ar-40-H), 11.11 (br s, 1H,
indole-H-1) ppm. 13C NMR (400 MHz, DMSO-d6):
d 51.14, 59.75,
101.00, 111.70, 118.64, 120.49, 122.52, 125.83, 126.97, 127.75,
131.96, 134.69, 135.25, 136.83, 140.62, 142.72, 158.47, 162.71 ppm.
HRMS (ESI): m/z [M þ H]þ calcd for C20H17N4O2Br2, 502.9718;
found, 502.9712. HPLC tR ¼ 18.464 min (98.65% at 220 nm,
99.09% at 254 nm).
4.1.5.9. 2-Amino-N,N-dibenzylthiazole-5-carboxamide
(13). The
crude product was purified with flash column chromatography
using ethyl acetate/hexane (1:1) as eluent. Yield: 41%, yellow
The pharmacophore model was used to screen approximately
5.5 million commercially available compounds, all of which were
previously converted into multifunctional format (25 conformers
for each compound in the database) using the LigandScout
screening module. The conformers of the molecules in the
screening library were generated using the idbgen module that is
available in Ligandscout, coupled with the OMEGA software. The
default high-throughput settings were used for the library gener-
ation: maximum number of output conformers per molecule ¼ 25;
crystals, mp 114e118 ꢁC. 1H NMR (400 MHz, DMSO-d6):
d
4.67 (s,
4H, 2ꢃ CH2), 7.16 (s, 1H, thiazole-H), 7.21e7.43 (m, 10H, Ar-H), 7.54
(br s, 2H, NH2) ppm. 13C NMR (400 MHz, DMSO-d6):
50.24, 119.63,
d
126.99, 127.21, 128.68, 137.14, 142.68, 162.63, 171.61 ppm. HRMS
(ESI): m/z [M þ H]þ calcd for C18H18N3OS, 324.1171; found,
324.1176. HPLC tR ¼ 18.464 min (98.65% at 220 nm, 99.09% at
254 nm).
ꢁ
4.1.5.10. 2-Amino-N,N-dibenzyl-4-methylthiazole-5-carboxamide
(14). The crude product was purified with flash column chroma-
tography using ethyl acetate as eluent. Yield: 53%, yellow crystals,
RMS threshold to duplicate conformers ¼ 0.8 A; maximum number
of all generated conformers per molecule ¼ 30,000; and maximum
number of intermediate conformers per molecule ¼ 4000. In the
virtual screening experiments, each compound had to fulfill all of
the derived pharmacophore constraints to be identified as a virtual
hit. Pharmacophore Fit scoring function was used to score the
mp 125e128 ꢁC. 1H NMR (400 MHz, DMSO-d6):
d
2.21 (s, 3H, CH3),
4.56 (s, 4H, 2ꢃ CH2), 7.18e7.48 (m, 13H, Ar-H þ NH2 þ thiazole-H)
ppm. 13C NMR (400 MHz, DMSO-d6):
16.67, 49.51, 109.85, 127.31,
d

ꢀ
V. Skedelj et al. / European Journal of Medicinal Chemistry 67 (2013) 208e220
219
matching of the hit molecules to the pharmacophore model [32].
The LigandScout screening procedure retrieved 3200 hit com-
pounds that were carefully visually assessed to identify compound
classes, and afterward to select compounds for further inhibition
assays.
a stock solution of 6 or 18 in DMSO to the dialysis buffer to obtain
50 M solutions of the Aslfm inhibitor in 5% DMSO. Successive ali-
quots of the degassed ligand solutions (11e31 L) were injected at
5e8 min intervals by a motor driven syringe (300 L) into the
degassed Aslfm solution (approximately 1 M, 5% DMSO) in the
m
m
m
m
calorimeter cell (V ¼ 1.386 mL) with constant stirring (300 rpm) at
37 ꢁC. At the end of experiment, the final ligand:enzyme ratio in the
titration cell was about 12:1. The titration data was further cor-
rected for the heat changes observed in the control titration of the
ligand solution in the dialysis buffer alone (containing 5% DMSO).
The experimental data were analyzed using the Origin 7.0 software,
provided by MicroCal, and fitted using a home written program,
based on the non-linear LevenbergeMarquardt c2 regression pro-
cedure, assuming a single binding site model (two parameters fit)
[34]:
4.3. Protein expression and purification
DNA of E. faecium D359 was amplified with Pfu Turbo DNA po-
lymerase (Stratagene, La Jolla, CA) by PCR, using the primers Asl1
(50-GAGAGACCATGGTGAACAGTATTGAAAATGAAG-30) and Asl2 (50-
CTCCATGGCTAGGATCCTTCTTTCACATGAAAATACTTTG-30). The PCR
product was then digested with NcoI and BamHI (underlined) and
cloned into pET2818; the resulting plasmid (pSJL1) was introduced
into E. coli BL21 (
production, the overnight culture was grown at 37 ꢁC in braine
heart infusion (BHI) broth containing ampicillin (50 g/mL). Five
milliliter aliquots of overnight culture were used to inoculate 2 L
BHI broth containing ampicillin (50 g/mL) and bacteria grown at
37 ꢁC to an optical density at 600 nm (OD600) of 0.7e0.8. Isopropyl-
-D-thiogalactopyranoside (IPTG) was added (0.5 mM) and the
lDE3) harboring the pREP4 plasmid. For protein
P þ L#PL
m
0
B
1
ꢀ
ꢁ
ꢀ
ꢁ
rꢂ1þ
1
=
m
C
C
A
Kb½Pꢅtot
vnPL
vn2
Hb0
¼
12DHb0 1 ꢂ
B
rffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi
D
H ¼
D
ꢀ
ꢀ
ꢁ
ꢁ
@
1þrþ
1
2ꢂ4r
b
=
Kb½Pꢅtot
incubation was continued for 3.5 h at 37 ꢁC. The bacteria were
harvested, and lysed by sonication. The Aslfm was purified from
clarified lysate by affinity chromatography on Ni2þ-nitriloacetate
agarose resin (SigmaeAldrich) and by filtration (using a HiPrep 16/
60 Sephacryl S-200 column, GE Healthcare). Fractions containing
the recombinant enzyme were concentrated by ultrafiltration and
stored at ꢂ20 ꢁC in 50 mM TriseHCl (pH 8.0), 500 mM NaCl, and
50% glycerol [14].
where P, L and PL represent Aslfm, ligand, and the Aslfmeligand
complex, respectively, H and
Hb0 are the measured enthalpy
D
D
upon aliquot addition and the standard enthalpy change of the
binding reaction, respectively, n2 and nPL represent the total num-
ber of moles of ligand and Aslfmeligand complex in the titration
cell, r is the bound ligand to total protein concentration, [P]tot, ratio,
and Kb is the binding constant. A detailed explanation of the applied
model equation is given in the literature [35].
4.4. Inhibition assay
The target compounds were tested for their inhibition of the
Asp adding activity of -aspartate ligase, using the spectrophoto-
metric pyruvate kinaseelactate dehydrogenase enzymatic assay.
Each compound was tested in duplicate at 50, 100 and 250 M in a
reaction mixture with a final volume of 150 L, containing 100 mM
TriseHCl buffer, pH 8.0, 50 mM MgCl2, 2 mM phosphoenolpyruvate,
0.16 mM NADH, 0.3 mM UM5K, 1 mM ATP, 1 mM -Asp, Triton-114
D-
4.6. Antibacterial activity measurements
D
The antibacterial activities of the compounds were assessed
against E. faecalis ATCC 29212 [36], E. faecium E1679 [37] and
E. faecium E1636 [37]. Minimum inhibitory concentrations (MICs)
were determined by broth microdilution according to standard
methodologies [38].
m
m
D
(0.005%), 2.4 U pyruvate kinase, 3.0 U lactate dehydrogenase, pu-
rified His-tagged Aslfm and the test compound dissolved in DMSO.
The reaction mixture was first incubated at 37 ꢁC for 2 min, and
then initiated by the addition of UM5K (0.3 mM). The change in
absorbance at 340 nm, which corresponds to ADP formation in the
Aslfm-catalyzed reaction, was monitored at 37 ꢁC for 10 min using a
Perkin Elmer Lambda 45 UV/VIS spectrophotometer. Residual ac-
tivities were calculated relative to control assays without the
compounds and with DMSO. Ki determinations were performed
For the disk diffusion method, the compounds were dissolved in
100% DMSO to a concentration of 20 mg/ml, and 160
mg were
loaded onto paper disks, which were placed on BHI agar freshly
inoculated with E. faecium D344R [39]. Bacterial growth was
assessed after 18 h of incubation at 37 ꢁC. For the macrodilution
method, compounds were diluted to a final concentration of
200
In a second set of tubes, ampicillin (2
weyersheim, France) was added in order to detect putative syner-
mg/mL (400e800
mM) in glass tubes containing 2 mL BHI broth.
mg/mL; Euromedex, Souffel-
under similar conditions, using D-Asp (1 mM), ATP (50 and 100 mM)
gistic effects of the tested compounds with b-lactams. Each tube
and six inhibitor concentrations (depending on the range of its
inhibitory activity), with monitoring of the enzymatic reaction for
10 min.
was inoculated with 107 CFU/ml of exponentially growing
E. faecium D344R and incubated for 18 h at 37 ꢁC.
All commercially available compounds with inhibitory activity
on Aslfm (18, 30, and 34) were characterized using the HR-MS
technique, and their purities were determined by HPLC. The pu-
rities of the tested compounds were established as ꢄ95% (see
Supplementary material, Table S2).
Acknowledgments
This work was supported by the Slovenian Research Agency
(Grant: P1-0208) and by post-doctoral grant Z1-4111 (A.P.). A. Z.
was the recipient of a Bourse de Recherche de la Ville de Paris.
4.5. Isothermal titration calorimetry (ITC) measurements
The heat changes upon binding were measured using a VP-ITC
microcalorimeter from MicroCal, LLC (Northampton, MA, USA).
Purified Aslfm stock solution in 10% glycerol was dialyzed for 40 h at
Appendix A. Supplementary data
4
ꢁC prior to titration (100 mM TriseHCl, pH 8.0, 50 mM MgCl2,
Supplementary data related to this article can be found at http://
dx.doi.org/10.1016/j.ejmech.2013.06.017.
0.005% Triton X-114). The ligand solutions were prepared by adding

ꢀ
220
V. Skedelj et al. / European Journal of Medicinal Chemistry 67 (2013) 208e220
[22] T. Langer, R.D. Hoffmann, Pharmacophores and Pharmacophore Searches,
References
Wiley VCH, Weinheim, Germany, 2006.
[23] A. Perdih, A. Kovac, G. Wolber, D. Blanot, S. Gobec, T. Solmajer, Discovery of
novel benzene 1,3-dicarboxylic acid inhibitors of bacterial MurD and MurE
ligases by structure-based virtual screening approach, Bioorg. Med. Chem.
Lett. 19 (2009) 2668e2673.
[24] G. Wolber, T. Langer, LigandScout: 3-D pharmacophores derived from
protein-bound ligands and their use as virtual screening filters, J. Chem. Inf.
Model 45 (2005) 160e169.
[1] H.W. Boucher, G.H. Talbot, J.S. Bradley, J.E. Edwards, D. Gilbert, L.B. Rice, et al.,
Bad bugs, no drugs: no ESKAPE! an update from the Infectious Diseases So-
ciety of America, Clin. Infect. Dis. 48 (2009) 1e12.
[2] R.C. Moellering Jr., Advances in antibacterial therapy, Transplant. Proc. 43
(2011) 2441e2442.
[3] G.J. Patti, S.J. Kim, J. Schaefer, Characterization of the peptidoglycan of
vancomycin-susceptible Enterococcus faecium, Biochemistry 47 (2008) 8378e
8385.
[4] R. Leclercq, E. Derlot, J. Duval, P. Courvalin, Plasmid-mediated resistance to
vancomycin and teicoplanin in Enterococcus faecium, N. Engl. J. Med. 319
(1988) 157e161.
[5] W. Vollmer, D. Blanot, M.A. de Pedro, Peptidoglycan structure and architec-
ture, FEMS Microbiol. Rev. 32 (2008) 149e167.
[6] D. Mengin-Lecreulx, B. Lemaitre, Structure and metabolism of peptidoglycan
and molecular requirements allowing its detection by the Drosophila innate
immune system, J. Endotoxin Res. 11 (2005) 105e111.
[7] P. Macheboeuf, C. Contreras-Martel, V. Job, O. Dideberg, A. Dessen, Penicillin
binding proteins: key players in bacterial cell cycle and drug resistance pro-
cesses, FEMS Microbiol. Rev. 30 (2006) 673e691.
[8] H. Barreteau, A. Kovac, A. Boniface, M. Sova, S. Gobec, D. Blanot, Cytoplasmic
steps of peptidoglycan biosynthesis, FEMS Microbiol. Rev. 32 (2008) 168e207.
[9] K.H. Schleifer, O. Kandler, Peptidoglycan types of bacterial cell walls and their
taxonomic implications, Bacteriol. Rev. 36 (1972) 407e477.
[10] T.E. Benson, D.B. Prince, V.T. Mutchler, K.A. Curry, A.M. Ho, R.W. Sarver, et al.,
X-ray crystal structure of Staphylococcus aureus FemA, Structure 10 (2002)
1107e1115.
[25] G. Wolber, A.A. Dornhofer, T. Langer, Efficient overlay of small organic molecules
using 3D pharmacophores, J. Comput. Aided Mol. Des. 20 (2006) 773e788.
[26] A. Cereto-Massague, L. Guasch, C. Valls, M. Mulero, G. Pujadas, S. Garcia-
Vallve, DecoyFinder: an easy-to-use python GUI application for building
target-specific decoy sets, Bioinformatics 28 (2012) 1661e1662.
ꢀ
[27] V. Skedelj, et al., in press.
[28] S. Bevc, J. Konc, J. Stojan, M. Hodoscek, M. Penca, M. Praprotnik, et al., ENZO: a
ꢀꢀ
web tool for derivation and evaluation of kinetic models of enzyme catalyzed
reactions, PLoS One 6 (2011) e22265.
[29] E. Edink, C. Jansen, R. Leurs, I.J.P. de Esch, The heat is on: thermodynamic
analysis in fragment-based drug discovery, Drug Discov. Today Technol. 7
(2010) 189e201.
[30] J.E. Ladbury, G. Klebe, E. Freire, Adding calorimetric data to decision making in
lead discovery: a hot tip, Nat. Rev. Drug Discov. 9 (2010) 23e27.
[31] W. Knauf, Chemotherapeutic nitro heterocycles. XXII. Synthesis and antimi-
crobial activity of 2-nitrothiazole-5-carboxamide, Eur. J. Med. Chem. (1975)
533e534.
[32] K.E. Rao, Y. Bathini, J.W. Lown, Synthesis of novel thiazole-containing DNA
minor groove binding oligopeptides related to the antibiotic distamycin,
J. Org. Chem. 55 (1990) 728e737.
[11] J.-L. Mainardi, R. Villet, T.D. Bugg, C. Mayer, M. Arthur, Evolution of peptido-
glycan biosynthesis under the selective pressure of antibiotics in gram-
positive bacteria, FEMS Microbiol. Rev. 32 (2008) 386e408.
[12] W. Staudenbauer, J.L. Strominger, Activation of D-aspartic acid for incorpora-
tion into peptidoglycan, J. Biol. Chem. 247 (1972) 5095e5102.
[13] W. Staudenbauer, E. Willoughby, J.L. Strominger, Further studies of the D-
aspartic acid-activating enzyme of Streptococcus faecalis and its attachment to
the membrane, J. Biol. Chem. 247 (1972) 5289e5296.
[14] S. Bellais, M. Arthur, L. Dubost, J.-E. Hugonnet, L. Gutmann, J. van Heijenoort,
[33] J. Kirchmair, P. Markt, S. Distinto, G. Wolber, T. Langer, Evaluation of the
performance of 3D virtual screening protocols: RMSD comparisons, enrich-
ment assessments, and decoy selectionewhat can we learn from earlier
mistakes? J. Comput. Aided Mol. Des. 22 (2008) 213e228.
[34] W.H. Press, B.P. Flannery, S.A. Teukolsky, W.T. Vetterling, Numerical Recipes,
Cambridge University Press, Oxford, 1992.
ꢀ
ꢀ ꢂ
ꢀ
[35] V. Skedelj, Emilija Arsovska, T. Tomasic, Ana Kroflic, Vesna Hodnik,
Martina Hrast, et al., 6-Arylpyrido[2,3-d]pyrimidines as novel ATP-
competitive inhibitors of bacterial
(2012) e39922.
D-Alanine:D-Alanine ligase, PLoS One
et al., Aslfm, the D-aspartate ligase responsible for the addition of D-aspartic
acid onto the peptidoglycan precursor of Enterococcus faecium, J. Biol. Chem.
281 (2006) 11586e11594.
[36] M.-T. Arias-Moliz, P. Baca, S. Ordóñez-Becerra, M.-P. González-Rodríguez, C.-
M. Ferrer-Luque, Eradication of enterococci biofilms by lactic acid alone and
combined with chlorhexidine and cetrimide, Med. Oral Patol. Oral Cir. Bucal
(2012).
[37] W. van Schaik, J. Top, D.R. Riley, J. Boekhorst, J.E.P. Vrijenhoek,
C.M.E. Schapendonk, et al., Pyrosequencing-based comparative genome
analysis of the nosocomial pathogen Enterococcus faecium and identification
of a large transferable pathogenicity island, BMC Genomics 11 (2010) 239.
[38] National Committee for Clinical Laboratory Standards (NCCLS), Methods for
Dilution Antimicrobial Susceptibility Tests for Bacteria that Grow Aerobically,
fourth ed., vol. 17, 1997.
[39] W. Zorzi, X.Y. Zhou, O. Dardenne, J. Lamotte, D. Raze, J. Pierre, et al., Structure
of the low-affinity penicillin-binding protein 5 PBP5fm in wild-type and
highly penicillin-resistant strains of Enterococcus faecium, J. Bacteriol. 178
(1996) 4948e4957.
[40] A typical run in the absence of inhibitors was implemented under the following
project number enzo.cmm.ki.si/kinetic.php?uwd¼120727444&load¼true (click
’Set Parameters’ and ’Start’).
[15] M.Y. Galperin, E.V. Koonin, A diverse superfamily of enzymes with ATP-
dependent carboxylate-amine/thiol ligase activity, Protein Sci.
6 (1997)
2639e2643.
[16] S. Dutka-Malen, C. Molinas, M. Arthur, P. Courvalin, Sequence of the vanC
gene of Enterococcus gallinarum BM4174 encoding -alanine: -alanine
a
D
D
ligase-related protein necessary for vancomycin resistance, Gene 112
(1992) 53e58.
ꢀ
ꢀ ꢂ
ꢀ ꢀ
[17] V. Skedelj, T. Tomasic, L.P. Masic, A. Zega, ATP-binding site of bacterial
enzymes as a target for antibacterial drug design, J. Med. Chem. 54 (2011)
915e929.
[18] I. Mochalkin, J.R. Miller, L. Narasimhan, V. Thanabal, P. Erdman, P.B. Cox, et al.,
Discovery of antibacterial biotin carboxylase inhibitors by virtual screening
and fragment-based approaches, ACS Chem. Biol. 4 (2009) 473e483.
[19] J.E. Lindsley, Use of a real-time, coupled assay to measure the ATPase activity
of DNA topoisomerase II, Methods Mol. Biol. 95 (2001) 57e64.
[20] W.R. McClure, A kinetic analysis of coupled enzyme assays, Biochemistry 8
(1969) 2782e2786.
[41] See enzo.cmm.ki.si/kinetic.php?uwd¼120326994&load¼true for compound
[21] B. Hess, B. Wurster, Transient time of the pyruvate kinaseelactate dehydro-
genase system of rabbit muscle in vitro, FEBS Lett. 9 (1970) 73e77.
6; click ’Set Parameters’ and ’Start’.