Evaluation of the published kinase inhibitor set to identify multiple inhibitors of bacterial ATP-dependent mur ligases

Article abstract of DOI:10.1080/14756366.2019.1608981

The Mur ligases form a series of consecutive enzymes that participate in the intracellular steps of bacterial peptidoglycan biosynthesis. They therefore represent interesting targets for antibacterial drug discovery. MurC, D, E and F are all ATP-dependent ligases. Accordingly, with the aim being to find multiple inhibitors of these enzymes, we screened a collection of ATP-competitive kinase inhibitors, on Escherichia coli MurC, D and F, and identified five promising scaffolds that inhibited at least two of these ligases. Compounds 1, 2, 4 and 5 are multiple inhibitors of the whole MurC to MurF cascade that act in the micromolar range (IC₅₀, 32–368 μM). NMR-assisted binding studies and steady-state kinetics studies performed on aza-stilbene derivative 1 showed, surprisingly, that it acts as a competitive inhibitor of MurD activity towards D-glutamic acid, and additionally, that its binding to the D-glutamic acid binding site is independent of the enzyme closure promoted by ATP.

Full text of DOI:10.1080/14756366.2019.1608981

Journal of Enzyme Inhibition and Medicinal Chemistry
ISSN: 1475-6366 (Print) 1475-6374 (Online) Journal homepage: https://www.tandfonline.com/loi/ienz20
Evaluation of the published kinase inhibitor set
to identify multiple inhibitors of bacterial ATP-
dependent mur ligases
Martina Hrast, Kaja Rožman, Iza Ogris, Veronika Škedelj, Delphine Patin,
Matej Sova, Hélène Barreteau, Stanislav Gobec, Simona Golič Grdadolnik &
Anamarija Zega
To cite this article: Martina Hrast, Kaja Rožman, Iza Ogris, Veronika Škedelj, Delphine Patin,
Matej Sova, Hélène Barreteau, Stanislav Gobec, Simona Golič Grdadolnik & Anamarija Zega
(2019) Evaluation of the published kinase inhibitor set to identify multiple inhibitors of bacterial ATP-
dependent mur ligases, Journal of Enzyme Inhibition and Medicinal Chemistry, 34:1, 1010-1017,
DOI: 10.1080/14756366.2019.1608981
To link to this article: https://doi.org/10.1080/14756366.2019.1608981
©
2019 The Author(s). Published by Informa
View supplementary material
Submit your article to this journal
View Crossmark data
UK Limited, trading as Taylor & Francis
Group.
Published online: 10 May 2019.
Article views: 108
Full Terms & Conditions of access and use can be found at
https://www.tandfonline.com/action/journalInformation?journalCode=ienz20

JOURNAL OF ENZYME INHIBITION AND MEDICINAL CHEMISTRY
019, VOL. 34, NO. 1, 1010–1017
2
https://doi.org/10.1080/14756366.2019.1608981
RESEARCH PAPER
Evaluation of the published kinase inhibitor set to identify multiple inhibitors of
bacterial ATP-dependent mur ligases
a
a,b
c
a
d
a
d
ꢀ
Martina Hrast , Kaja Ro ꢀz man , Iza Ogris , Veronika Skedelj , Delphine Patin , Matej Sova , H ꢁe l ꢂe ne Barreteau ,
a
c
a
Stanislav Gobec , Simona Goli ꢀc Grdadolnik and Anamarija Zega
a
b
Faculty of Pharmacy, University of Ljubljana, Ljubljana, Slovenia; Department of Medicinal Chemistry, University of Minnesota, Minneapolis,
c
d
MN, USA; Molecular Structural Dynamics, Theory Department, National Institute of Chemistry, Ljubljana, Slovenia; Institute for Integrative
Biology of the Cell (I2BC), CEA, CNRS, Univ Paris-Sud, Universit ꢁe Paris-Saclay, Gif-Sur-Yvette Cedex, France
ABSTRACT
ARTICLE HISTORY
The Mur ligases form a series of consecutive enzymes that participate in the intracellular steps of bacterial
peptidoglycan biosynthesis. They therefore represent interesting targets for antibacterial drug discovery.
MurC, D, E and F are all ATP-dependent ligases. Accordingly, with the aim being to find multiple inhibitors
of these enzymes, we screened a collection of ATP-competitive kinase inhibitors, on Escherichia coli MurC,
D and F, and identified five promising scaffolds that inhibited at least two of these ligases. Compounds 1,
Received 5 February 2019
Revised 3 April 2019
Accepted 13 April 2019
KEYWORDS
Bacterial Mur (MurC–MurF)
ligases; published kinase
inhibitor set; steady-state
kinetics measurements;
NMR studies;
2
, 4 and 5 are multiple inhibitors of the whole MurC to MurF cascade that act in the micromolar range
(IC50, 32–368mM). NMR-assisted binding studies and steady-state kinetics studies performed on aza-
stilbene derivative 1 showed, surprisingly, that it acts as a competitive inhibitor of MurD activity towards
D-glutamic acid, and additionally, that its binding to the D-glutamic acid binding site is independent of
the enzyme closure promoted by ATP.
antibacterial agents
Introduction
inorganic phosphate and concomitant formation of a peptide or
amide bond.
The increasing emergence of bacterial strains resistant to currently
available antibiotics has created medical needs for antibacterial
therapy that remain unmet today. It is broadly accepted that new-
class agents represent unique and valuable opportunities to
achieve significant advances against bacterial resistance, because
they should not be as susceptible to the pre-existing mechanisms
of resistance, as seen with established antibacterial classes; i.e.
they should not show cross-resistance. However, despite large
efforts in this area over recent decades, the introduction of mech-
anistically novel antibacterial agents into clinical practice has been
The crystal structures of the Mur ligases from different bacterial
strains show a similar three-domain topology, with the N-terminal
and central domains binding the UDP precursor and ATP, respect-
ively, while the C-terminal domain binds the condensing amino
acid or dipeptide residue. While the topologies of the central and
C-terminal domains are similar among the Mur ligases, those of
the N-terminal domains show differences, with MurC and MurD
more closely related to each other than to MurE and MurF. These
differences are associated with the lengths of the UDP precursor
substrates. The ATP binding site of the Mur ligases is a highly con-
served region, with sequence identities ranging from 22 to 26%.
All of the Mur ligases have a glycine-rich P-loop, along with gluta-
mates and histidines, which are responsible for coordinating the
1–4
a rarely achieved goal of antibacterial research
.
In this context, there has been increasing interest in exploiting
Mur ligases, the enzymes involved in the intracellular steps of the
biosynthesis of peptidoglycan, an essential cell-wall polymer
2þ
10
5–7
Mg ions .
unique to prokaryotic cells (Figure 1) . Mur ligases catalyse the
formation of a peptide or amide bond between a uridine diphos-
phate N-acetylglucosamine (UDP)-substrate and a condensing
amino acid. They operate through similar chemical mechanisms,
and as shown for MurC and MurF, by an ordered kinetic
The Mur ligase family provides an attractive collection of
emerging drug targets, and it has been heavily investigated over
the last decade, with novel and diverse classes of inhibitors
1
1–17
discovered
. As all of the Mur ligases considered here (i.e.
8
,9
MurC-MurF) contain the ATP binding site, we screened published
mechanism . This series of enzymatic reactions is initiated by
1
8
binding of ATP to the free MurA, followed by binding of the cor- kinase inhibitor sets (PKIS) I and II , which are two sets of small-
responding UDP substrate. The terminal carboxyl group of the molecule ATP-competitive kinase inhibitors made available by
UDP substrate is then activated by ATP-promoted phosphoryl- GlaxoSmithKline. Inhibitors that target the ATP binding site of
ation, which results in formation of an acylphosphate intermedi- MurC to MurF would have the potential to act as multitarget
ate. This is then attacked by the amino group of the incoming inhibitors . This constitutes a promising strategy to combat bac-
amino acid or dipeptide, after its binding to MurC. The resulting terial resistance, because target-mediated resistance to such com-
tetrahedral high-energy intermediate collapses, with elimination of pounds will be less likely to evolve, as mutations that confer
1
9
CONTACT Anamarija Zega
anamarija.zega@ffa.uni-lj.si
Faculty of Pharmacy, University of Ljubljana, A ꢀs ker ꢀc eva 7, Ljubljana 1000, Slovenia
Supplemental data for this article can be accessed here.
ß 2019 The Author(s). Published by Informa UK Limited, trading as Taylor & Francis Group.
This is an Open Access article distributed under the terms of the Creative Commons Attribution License (http://creativecommons.org/licenses/by/4.0/), which permits unrestricted use,
distribution, and reproduction in any medium, provided the original work is properly cited.

JOURNAL OF ENZYME INHIBITION AND MEDICINAL CHEMISTRY
1011
Figure 1. The cytoplasmic steps of peptidoglycan biosynthesis catalysed by the Mur ligases. The Mur ligases considered specifically in the present study are in red.
resistance would have to occur to at least two different target of 1 ml/min. In the method the eluent was a mixture of 0.1% TFA
20,21
genes during a single generation
.
in water (A) and acetonitrile (B). The gradient was 10% B to 90% B
However, targeting the ATP binding site of bacterial enzymes in 20 min.
is associated with several issues. First, inhibitor binding to the ATP
Procedure a: To a solution of Pd(PPh3)4 (5 mol%) and LiCl
binding site must be selective for the targeted bacterial enzyme (6.55 mmol) in anhydrous DMF (12 ml), 3-bromo-5-isocyanopyri-
over the human ATP-dependent enzymes, and particularly the kin- dine (5.46 mmol) and tributyl(vinyl)tin (6.55 mmol) were added
ꢁ
ases. Additionally, an ATP-competitive inhibitor must be able to under Ar atmosphere and stirred at 70 C for 16 h. The reaction
compete with the high ATP concentration in the bacterial cell
mixture was cooled to room temperature, followed by addition of
saturated solution of KF (20 ml). The formed precipitate was fil-
tered through the pad of Celite and washed with diethyl ether
2
2
(
(
0.6–18 mM) , which is remarkably similar to that in human cells
1–10 mM) . Nevertheless, it has been successfully shown that it
2
3
is possible to design competitive and selective inhibitors of the
ATP binding site in protein kinases. Most importantly, successful
examples of ATP-competitive bacterial enzyme inhibitors with
antibacterial activities and good selectivity profiles with respect to
(
30 ml). Water phase was extracted with diethyl ether (2 ꢀ 40 ml),
combined, washed with water (1 ꢀ 50 ml), and dried over Na SO .
2
4
The solvent was evaporated under reduced pressure and the
product was purified by flash chromatography.
2
0
human enzymes show that these challenges can be overcome .
Procedure b: To a solution of 5-vinylnicotinonitrile (5.76 mmol)
in anhydrous DMF (12 ml), 5-bromo-2-iodo-1,3-dimethylbenzene
(
(
6.92 mmol) was added. Then Pd
2 3 3
dba (10 mol%) and P(o-tol)
Materials and methods
10 mol%) were added, followed by triethylamine (1.6 ml). The
ꢁ
Synthesis of compound 1
reaction mixture was stirred at 95 C for 18 h. To a cooled solution,
EtOAc was added (100 ml) and washed with water (2 ꢀ 50 ml), 1 M
HCl (50 ml), brine (50 ml) and dried over Na SO . The solvent was
evaporated under reduced pressure and the formed product was
recrystallized from MeOH.
Procedure c: To a solution of (E)-5–(4-bromo-2,6-dimethylstyryl)-
nicotinonitrile (0.96 mmol) in THF (8 ml), water (4 ml) and furan-2-
ylboronic acid were added, followed by K
5
-bromonicotinonitrile and tributylvinyltin were used in Stille cou-
2
4
pling to generate the vinyl intermediate, followed by Heck reac-
tion forming the aza-stilbene derivative. Finally, the tetrazole was
synthesised using the NH Cl and NaN in DMF (Figure 2).
4 3
All of the chemicals used were obtained from commercial and
were used without further purification. Reactions were monitored
using analytical thin-layer chromatography plates (Merck, Silica
Gel 60 F254, 0.25 mm), and the compounds were visualised with
ultraviolet light and ninhydrin. Silica gel grade 60 (particle size
2 3
CO (4.78 mmol) and
Pd(PPh (10 mol%). The reaction mixture was stirred under reflux
for 18 h. After cooling to room temperature, EtOAc (20 ml) was
3 4
)
added. Organic phase was washed with water (2 ꢀ 20 ml) and
0
.040–0.063 mm; Merck, Germany) was used for flash column chro-
1
13
matography. H and C NMR spectra were recorded on a Bruker brine (20 ml) and dried over Na SO . The solvent was evaporated
2 4
AVANCE III 400 MHz spectrometer in methanol-d4,. ESI-MS mass under reduced pressure and the product was purified by flash
spectra were recorded on the Advion expression compact mass chromatography.
spectrometer (CMS) at the Faculty of Pharmacy of the University
of Ljubljana. HPLC analyses were performed on an Agilent styryl)nicotinonitrile (0.20 mmol) in anhydrous DMF (5 ml), NH
Technologies HP 1100 instrument with a G1365B UV–VIS detector (0.60 mmol) and NaN (0.60 mmol) were added. The reaction mix-
254 nm), using a Luna C18 column (4.6 ꢀ 250 mm) at a flow rate ture was stirred at 110 C for 48 h. Mixture was cooled to room
Procedure d: To a solution of (E)-5–(4-(furan-2-yl)-2,6-dimethyl-
Cl
4
3
ꢁ
(

1
012
M. HRAST ET AL.
ꢁ
Figure 2. Reaction scheme for synthesis of compound 1. Reagents and conditions: (a) tributylvinyl tin, LiCl, Pd(PPh
3
)
2
Cl
2
, DMF, 70 C; (b) 5-bromo-2-iodo-1,3-dimethyl-
ꢁ
ꢁ
ꢁ
2 3 3 2 3 3 4
benzene, Pd dba , triethylamine, P(o-tol) , DMF, 95 C; (c) furan2-ylboronic acid, K CO , Pd(PPh ) , water, THF, 100 C; (d) NH
4
Cl, NaN
3
, anhydrous DMF, 110 C.
temperature, then 1M HCl was added dropwise until the white Steady-state kinetic analysis of compound 1
precipitate was formed. Precipitate was filtered off and addition- For compound 1, K values were determined against MurD from E.
i
ally purified by flash chromatography.
i
coli. K determinations were performed under similar conditions to
those described for the MurD inhibition assay, where different
concentration of one substrate and a fixed concentration of the
other two were used. First, the concentration of UMA was varied
Characterisation of compound 1
Yield: 32%, pale yellow crystals, H NMR (400 MHz, methanol-d
1
4
): d
(10, 20, 40, 80 mM) at fixed ATP (400 mM) and D-Glu (100 mM), then
(
ppm) 2.34 (s, 6H, CH ), 6.39–6.40 (m, 1H, Ar-H), 6.64 (d, 1H,
3
the concentration of D-Glu was varied (50, 100, 200, 400 mM) at
fixed ATP (400 mM) and UMA (80 mM), and finally, the concentra-
tion of ATP was varied (100, 200, 400, 800 mM) at fixed UMA
J ¼ 3.4 Hz, Ar-H), 6.70 (d, 1H, J ¼ 16.8 Hz, CH), 7.15 (bs, 1H, NH),
7
.32 (s, 2H, Ar-H), 7.38 (d, 1H, J ¼ 16.6 Hz, CH), 7.42–7.44 (m, 1H,
Ar-H), 8.57–8.60 (m, 2H, Ar-H), 8.99–9.00 (m, 1H, Ar-H); 13 C NMR (80 mM) and D-Glu (100 mM). The concentrations of 1 were 0, 15,
100 MHz, DMSO-d ): d (ppm) 21.6, 106.8, 113.0, 122.4, 123.7, 30, 60, 90, 120, 180 and 250 mM. After a 15 min incubation, 100 mM
29.3, 129.7, 131.4, 133.0, 134.9, 135.3, 137.7, 143.6, 145.3, 149.1, Biomol green reagent was added, and the absorbance was read at
(
6
1
1
þ
53.5, 154.7; ESI-MS ([M–H ], (m/z)) calcd. for C H N O: 343.14, 650 nm after 5 min. All of the experiments were run in triplicate.
20 17 5
found 342.49; HPLC: t
R
¼12.333 min (99%).
The data were analysed using the SigmaPlot 12.0 software. The
initial velocities were fitted to competitive, non-competitive,
uncompetitive and mixed enzyme inhibition. The K and mode of
inhibition from the best ranking model were used, as provided by
the software.
i
Enzyme assays for inhibition of Mur ligases
Inhibition assay
Inhibition of the Mur ligases was determined using the malachite
24
green assay, with slight modifications . The mixtures for the
respective Mur ligase assays had a final volume of 50 mL, which
Antimicrobial testing
contained 100 mM of each tested compound dissolved in dime- Antimicrobial testing was performed by the broth microdilution
thylsulphoxide (DMSO), added to:
method in 96-well plates following the Clinical and Laboratory
2
9
MurC: 50 mM Hepes, pH 8.0, 5 mM MgCl
2
, 0.005% Triton X-114, Standards Institute guidelines
20 mM L-Ala, 120 mM uridine-5 -diphosphate-N-acetylmuramine, Antimicrobial Susceptibility Testing recommendations . Bacterial
and European Committee on
0
30
1
4
25
50 mM ATP, and purified E. coli MurC .
MurD: 50 mM Hepes, pH 8.0, 5 mM MgCl , 0.005% Triton X-114, diluted with cation-adjusted Mueller–Hinton broth with TES buf-
suspensions equivalent to 0.5 McFarland turbidity standard were
2
0
5
1
00 mM D-Glu, 80 mM uridine-5 -diphosphate-N-acetylmuramoyl-L- fering (ThermoFisher Scientific), for a final inoculum of 10 CFU/
2
6
alanine (UMA), 400 mM ATP, and purified MurD .
MurE: 50 mM Hepes, pH 8.0, 15 mM MgCl
1
mL. The compounds dissolved in DMSO and the inoculum were
ꢁ
, 0.005% Triton X- mixed together and incubated for 20 h at 37 C. After this incuba-
2
0
14, 60 mM meso-diaminopimelic acid, 100 mM uridine-5 -diphos- tion, the minimal inhibitory concentrations (MICs) were deter-
phate-N-acetylmuramoyl-L-alanine-D-glutamate, 1000 mM ATP, and mined by visual inspection, as the lowest dilution of the
purified E. coli MurE .
compounds that showed no turbidity. The MICs were determined
MurF: 50 mM Hepes, pH 8.0, 50 mM MgCl , 0.005% Triton X- against Staphylococcus aureus (ATCC 29213) and E. coli (ATCC
2
7
2
0
1
14, 600 mM D-Ala- D-Ala, 100 mM uridine-5 -diphosphate-N-acetyl- 25922) bacterial strains. Tetracycline was used as the positive con-
muramoyl-L-alanine-D-glutamate-2,6-diaminopimelic acid, 500 mM trol on every assay plate, with MICs of 0.5 and 1 mg/mL for S. aur-
28
ATP, and purified E. coli MurF .
eus and E. coli, respectively.
In all cases, the final concentration of DMSO was 5% (v/v).
ꢁ
After incubation for 15 min at 37 C, the enzyme reaction was ter-
Nuclear magnetic resonance studies of ligand binding
minated by addition of 100 mM Biomol green reagent, and the
absorbance was measured at 650 nm after 5 min. All of the experi- The nuclear magnetic resonance (NMR) spectra were recorded at
ꢁ
ments were run in duplicate. Residual activities were calculated 25 C on a spectrometer (DirectDrive 800 MHz; Varian, Slovenian
with respect to control assays without the tested compounds, but NMR centre at National Institute of Chemistry) equipped with a
with the 5% DMSO carrier. The IC50 values were determined by cryoprobe. The pulse sequences provided in the Varian BioPack
measuring the residual activities at seven different compound library of pulse programmes was used. The samples were pre-
concentrations, and they represent the concentration of the com- pared in 90% H O/10% DMSO-d buffer containing 20 mM HEPES,
2
6
pound at which the residual activity was 50%.
7 mM (NH
4 2
) SO
4
, 3.5 mM MgCl
2
, and 2 mM dithiothreitol, pH 7.2.

JOURNAL OF ENZYME INHIBITION AND MEDICINAL CHEMISTRY
24
1013
1
3
Here, 0.07 mM MurD that was selectively labelled with C at the during enzymatic reactions . To avoid non-specific inhibition due
methyl groups of Ile (d1 only), Val and Leu was used. The protein to aggregate formation, all of the compounds were tested in the
was titrated using compound 1 in MurD:ligand molar ratios of 0.5, presence of detergent (0.005% Triton-X114). The data are pre-
1
, 2, 4 and 8, without and with 2 mM b,c-methyleneadenosine 5’- sented as the residual activities of these Mur ligases in the pres-
triphosphate (AMP–PCP). At the MurD:ligand molar ratio of 1:8, ence of 100 mM of each test compound (Supporting Information
compound 1 was not completely dissolved. The final concentra- Table S1). Compounds that showed residual activity <50% on at
tion of DMSO-d was 12% (v/v). The variation of DMSO from 10 to least two of these Mur ligases were considered as ‘hits’
6
3
1
1
2% (v/v) affects the chemical shift perturbations by <0.02 ppm .
(Supporting Information Table S1, gray). The IC50 values against
The heteronuclear single quantum coherence (HSQC) spectra MurC, D, E and F were determined for the selected hit compounds
1
13
for H/ C were acquired with 1024 data points in t
2
, 32 scans, 64 – one representative compound from each structural class was
1 13
33
34
complex points in t , and relaxation delay of 1 s. The H and
sweep widths were 9470 and 3338 Hz, respectively. The spectra alkynyl pyrimidine; compound 3 , pyrazolo [1,5]-b]pyridazine;
were processed and analysed with the Felix 2007 software pack- compound 4 , phenoxypyrimidine; compound 5 , 4,6-bis-anilino-
age (Felix NMR Inc., Laboratory of Biomolecular Structure at 1H-pyrrolo[2,3-pyrimidine (Table 1). The four hits (1, 2, 4, 5) inhib-
C
chosen: compound 1 , aza-stilbene derivative; compound 2 ,
1
3
5
3
6
37
National Institute of Chemistry). The spectra were zero-filled twice
and apodised with a squared sine bell function shifted by p/2 in
both dimensions using a linear prediction of the data in the incre-
ited all four of these Mur ligases in the micromolar range.
Compound 3 only inhibited MurC, and was thus not investigated
further. Among the hits, compounds for further mechanistic stud-
ies were chosen on the basis of their ligand efficiency. Ligand effi-
ciency quantifies the molecular properties, particularly size and
lipophilicity and is used as a selection criterion for structurally dif-
ferent compounds as more ‘efficient’ ligands have the potential to
1
13
mented dimension. The combined H/ C chemical shift perturba-
1
13
tions (Dd) were calculated from the H and C chemical shift
32
perturbations using Equation (1) :
ꢀ
ꢁ
1=2
2
1
2
13
38
ð
Þ
ð
Þ
Dd ¼ Dd H þ 0:252 ꢀ Dd C
:
(1) control for the inflation of these properties Thus ligand efficien-
cies were calculated for the four hits (Supporting Information
3
9,40
Table S2)
, and compound 1 was selected as the most promis-
Results and discussion
ing for additional studies. Additionally, regarding the data of pro-
41
filed compounds in large panels of human kinase assays
Supporting Information Table S4), only hit compound 5 among
Biological activities
(
The published kinase inhibitor set compounds were evaluated for our hits showed appreciable activity at human kinases, thus com-
inhibition of MurC, D and F ligases from E. coli using the pound 1 was selected for the further evaluation. To provide suffi-
Malachite green assay, which detects orthophosphate generated cient amounts of compound 1 for kinetic and NMR studies, it was
Table 1. Results of in vitro biological assays of selected compounds 1–5 against E. coli MurC–MurF ligases.
MurC
MurD
MurE
IC50 [mM]
MurF
IC50 [mM]
Comp.
Name
Structure
N
RA [%]
70
IC50 [mM]
RA [%]
45
IC50 [mM]
RA [%]
49
RA [%]
26
1
GW458344X
368
62
104
104
79
59
H
N
N
N
N
O
2
3
GW659893X
GW827105X
NH
2
43
10
51
83
50
84
157
30
39
O
N
F
N
N
H
Cl
N
N
74
90
–
–
100
–
N
O
N
N
N
F
H
F
F
4
5
SB-242721
H
48
29
27
10
63
56
26
139
51
16
95
N
NC
N
N
O
N
N
F
GSK1173862A
O
62
32
58
66
H
N
H
N
N
N
F
NH
NH2
N
O

1
014
M. HRAST ET AL.
0
Figure 3. (a) Lineweaver–Burk plot of competitive inhibition model of compound 1 versus D-Glu at fixed ATP (400 mM) and uridine-5 -diphosphate-N-acetylmuramoyl-
0
L-alanine (80 mM). (b) Lineweaver–Burk plot of mixed inhibition model of compound 1 versus ATP at fixed D-Glu (100 mM) and uridine-5 -diphosphate-N-acetylmura-
moyl-L-alanine (80 mM). Data points are means ± standard deviations (all <10%) of triplicates.
chemical shift perturbations upon binding of various ligands.
Namely, the active site for the initial phosphorylation of substrate
UMA is located in the cleft between the central and C-terminal
domain. The reactive part of substrate UMA enters this cleft from
Table 2. Inhibitory properties of compound 1 against all the MurD substrates.
2
Substrate
Inhibition mechanism
Ki (mM)
R
D-Glu
ATP
UMA
Competitive (full)
Mixed (full)
Mixed (partial)
65.5 ± 4.9
18.7 ± 2.6
14.8 ± 11.2
0.970
0.960
0.817 the side closest to N-terminal domain, while the ATP from the
opposite site. The sulphonamide inhibitors approach the enzyme
from the same side as UMA interacting with all three domains.
resynthesized according to the synthetic procedures (Figure 2) Their D-Glu moiety occupies the same site in the C-terminal
33
described in the literature .
domain as observed for the D-Glu moiety of product UMAG, the
To obtain more in-depth insight into the inhibition mechanism naphthalene ring is positioned in the cleft between all three
of compound 1, steady-state kinetics and NMR studies of its bind- domains, and the functional groups at position 6 of naphthalene
ing were performed with MurD, the only one of these Mur ligases ring extends into the binding pocket of the UMA/UMAG uracil
considered here for which NMR assignation of the crucial methyl ring in the N-terminal domain. Consequently, binding of ATP ana-
1
13
groups was available. With implementation of steady-state kinetic logues exert distinct pattern of MurD H/ C chemical shift pertur-
44
measurements, the potential for competitive, non-competitive, bations in comparison to binding of UMA or sulphonamide
4
4,45
uncompetitive and mixed mechanisms of inhibition with respect inhibitors
. Several pronounced perturbations are observed
to each substrate was investigated. The best model obtained only upon binding of AMP-PCP. These can be explained by the
showed that compound 1 acts as a competitive inhibitor of MurD effect of its adenine moiety, the binding pocket for which is
with respect to D-Glu (Figure 3(a); Table 2). The K determined located in the central domain far away from the binding site of
i
(
(
65.5 ± 4.9 mM) was in good agreement with the IC50 for MurD UMA and sulphonamide inhibitors. In contrast, several perturba-
104 mM) obtained in the standard screening assay. The produced tions are observed only at binding of UMA or sulphonamide inhib-
best model of inhibition with respect to ATP showed mixed-type itors. The most indicative are the large perturbations of Ile74,
of inhibition (Figure 3(b); Table 2). The model of inhibition versus Leu57 methyl groups in the uracil binding pocket observed upon
UMA revealed mixed-type of inhibition, however with less reliable binding of UMA and sulphonamide derivatives and pronounced
statistics (Table 2; Supporting Information Figure S1).
The kinetic data were confirmed with NMR spectroscopic meas- observed upon binding of sulphonamide derivatives. These groups
perturbation of Leu416 methyl groups in the D-Glu binding site
1
13
42,43
urements using H/ C-HSQC two-dimensional NMR
.
The are in the range of 5 Å to the ligand. None of these groups are
H/ C chemical shift perturbations upon binding of 1 to MurD affected upon binding of AMP-PCP. Moreover, the signals of
1
13
13
selectively labelled with C at the methyl groups of Ile, Val and Leu416 methyl groups are the only ones that are affected upon
4
4
45
Leu were examined in relation to the binding of AMP–PCP , binding of D-Glu itself . These specific differences in patterns of
44
45
31,45
1
13
UMA and D-Glu , and of sulphonamide inhibitors
reported the H/ C chemical shift perturbations upon binding of various
previously. Assignments of the crucial methyl groups that indi- MurD ligands with known binding modes offer a solid basis for
cated binding to the D-Glu binding site (Leu417) in the MurD identification of binding sites of novel MurD inhibitors.
C-terminal domain, or to the uracil binding site (Ile74, Leu57) in
Compound 1 only perturbed the signals of the Leu416 methyl
the MurD N-terminal domain, were shown in our previous NMR groups, as shown in the expansion of the H/ C HSQC spectra
studies of the binding mode of second-generation sulphonamide with the signals of the crucial Leu methyl groups in Figure 4. The
derivatives . Moreover, we previously identified the signals of the signals of the Leu57 methyl groups that are indicative for the
methyl groups that are affected only during the binding of location of a ligand in the uracil binding pocket were unperturbed
1
13
3
1
4
4
AMP–PCP , which indicated ligand binding to the ATP binding (Figure 4). These observations indicated that the location of com-
1
13
site in the MurD central domain.
pound 1 was in the D-Glu binding pocket. The combined H/ C
4
5–47
The crystallographic structures of MurD complexes
offer chemical shift perturbations of the Leu416 methyl groups upon
the understanding of differences in patterns of the H/ C addition of 1 were similar in the absence and presence of
1
13

JOURNAL OF ENZYME INHIBITION AND MEDICINAL CHEMISTRY
1015
is independent of AMP-PCP indicates that the mechanism of MurD
4
8
regulation still needs to be defined.
The scaffold of 1 should represent a promising starting point
for development of potent MurD inhibitors. This also arises from
the possible negative influence of the MurD fast domain motion
on ligand binding. It was shown that binding of sulphonamide
3
1
inhibitors that span from the C-terminal to the N-terminal
domains is exposed to the rapid bending-like motion of the C-ter-
minal and N-terminal domains. Therefore, development of inhibi-
tors that interact with either the C-terminal or N-terminal domain
might represent a future perspective.
Selected hit compounds 1–5 were assayed in vitro for their
potential antibacterial activities against the Gram-negative E. coli
and Gram-positive S. aureus bacterial strains. These compounds
were inactive against all of the bacteria strains tested. This was,
however, as expected, due to their relatively low inhibi-
tory activities.
1
13
13
Figure 4. Overlaid expanded regions of the H/ C HSQC spectra for C select-
ively-labelled MurD without (blue) and with (ligand:MurD, 4:1; red) compound 1,
which indicates that compound 1 interacts with the D-Glu binding pocket. Both
spectra were recorded in the presence of 2 mM AMP-PCP in the NMR sample.
Additionally, on the basis of the structures of 11 analogues
(Supporting Information Table S3) of compound 1 from the PKIS
set, simple SAR can be established which could offer the basis for
the further development of more potent (and selective) aza stil-
bene inhibitors of Mur ligases.
Conclusions
In summary, screening of the PKIS yielded four new scaffolds that
show potential for development into potent inhibitors of bacterial
cell-wall biosynthesis. For one of the hits, the aza-stilbene deriva-
tive 1, the binding mode for MurD was defined, and these data
represent a good starting point for structural optimisation of Mur
ligase inhibitors which contain this scaffold. We showed that com-
pound 1 binds to the D-Glu binding site in the C-terminal domain
of MurD independent of AMP-PCP. This is an interesting example
that indicates that the conformational states of MurD need to be
additionally evaluated along with ligand binding. Also, as inhibitor
1
interacts with the single C-terminal domain, its binding is not
affected by rapid movement of MurD domains, which might aid
the process towards development of more potent compounds.
For an insight into the binding of other identified structural types
that would provide data for development of more potent Mur lig-
1
13
Figure 5. Combined H/ C chemical shift perturbations (Dd) of higher field ase inhibitors, further kinetic and NMR/crystallographic studies
1
(
black) and lower field (gray) Leu416 methyl groups, calculated from the H and
3
should be performed.
1
C chemical shifts in the absence and presence of compound 1 at a ligand:MurD
Here, it is worth noting that we are aware that the use of pro-
tein kinase inhibitors that target the ATP binding site as hit com-
pounds for bacterial enzymes is challenging because of their lack
of specificity. However compounds identified as hits (1,2,4)
ratio of 4:1. Dd of Leu416 methyl groups upon addition of compound 1 in the
presence of 2 mM AMP–PCP (right) and without any addition of AMP–PCP (left)
are presented.
4
1
AMP–PCP (Figure 5). Moreover, regardless of the AMP–PCP con- showed no appreciable activity at human kinases (Supporting
centration in the NMR samples, none of the signals of methyl Information Table S4) and therefore are a viable starting point for
groups that were affected only during the binding of AMP–PCP development of bacterial cell wall synthesis inhibitors. Moreover, a
were perturbed upon addition of 1. This indicates that 1 does not bioactivity search for all hit compounds (1–5) in ChEMBL data-
4
9
interact with the ATP binding site in the MurD central domain. It base was performed (Supporting Information Table S4). Aside
approaches the enzyme from the other side as ATP, in a manner from results of PKIS set profiling on panels of human kinase
similar to D-Glu or sulphonamide inhibitors. In contrast to sul- assays, these data show no activity on other human targets. An
phonamide inhibitors, it lacks strong interactions with the uracil approach that can also be used for the design of selective inhibi-
binding pocket in the N-terminal domain.
tors of the Mur ligases is the structure-based drug design of com-
The PKIS library consists of kinase inhibitors that target ATP pounds that not only occupy the binding site, but at the same
binding sites. Thus, considering that compounds 1, 2, 4 and 5 time exploit interactions with amino acids adjacent to the binding
show multiple inhibition of all four of these Mur ligases (i.e. domain that are unique to these target enzymes. Further studies
MurC–MurF), we anticipated these compounds bind to the ATP will also have to concentrate on the design of molecules that will
binding site of these Mur ligases. However, the steady-state kinetic on the one hand fulfil the pharmacophore pattern, and on the
measurements together with the NMR studies of ligand binding other hand have appropriate physical–chemical properties that
indicated that aza-stilbene derivative 1 binds to the D-Glu binding will allow efficient compound penetration into bacterial cells,
site of MurD ligase. Additionally, the observation that binding of 1 which will promote increased antibacterial activity.

1
016
M. HRAST ET AL.
Acknowledgments
synthesis, crystal structures, and biological evaluation. Eur J
Med Chem 2011;46:5512–23.
The authors thank Dr Chris Berrie for proof-reading of
the manuscript.
ꢀ
1
6. Toma ꢀs i ꢁc T, Sink R, Zidar N, et al. Dual inhibitor of MurD and
MurE ligases from Escherichia coli and Staphylococcus aureus.
ACS Med Chem Lett 2012;3:626–30.
1
7. Perdih A, Hrast M, Barreteau H, et al. Benzene-1,3-dicarbox-
ylic acid 2,5-dimethylpyrrole derivatives as multiple inhibi-
tors of bacterial Mur ligases (MurC–MurF). Bioorg Med Chem
Disclosure statement
No potential conflict of interest was reported by the authors.
2
014;22:4124–34.
1
8. Brown DG, Bostr o€ m J. Where do recent small molecule clin-
ical development candidates come from? J Med Chem 2018;
61:9442–68.
Funding
This work was supported by the Slovenian Research Agency
(
Grants No. P1-0208, P1-0010, J1-8145) and by the Centre National 19. Proschak E, Stark H, Merk D. Polypharmacology by design: a
de la Recherche Scientifique (CNRS, Projet International de
Recherche Scientifique (PICS) 7757). The PKIS was supplied by
medicinal chemist’s perspective on multitargeting com-
pounds. J Med Chem 2019;62:420–44.
ꢀ
GlaxoSmithKline, LLC and the Structural Genomics Consortium 20. Skedelj V, Toma ꢀs i ꢁc T, Ma ꢀs i ꢀc LP, Zega A. ATP-binding site of
under an open access Material Transfer and Trust Agreement:
http://www.sgc-unc.org.
bacterial enzymes as a target for antibacterial drug design.
J Med Chem 2011;54:915–29.
2
2
1. Wong KK, Kuo DW, Chabin RM, et al. Engineering a cell-free
murein biosynthetic pathway: combinatorial enzymology in
drug discovery. J Am Chem Soc 1998;120:13527–8.
2. Chappelle EW, Levin GV. Use of the firefly bioluminescent
reaction for rapid detection and counting of bacteria.
Biochem Med 1968;2:41–52.
References
1
.
Payne DJ, Gwynn MN, Holmes DJ, Pompliano DL. Drugs for
bad bugs: confronting the challenges of antibacterial discov-
ery. Nat Rev Drug Discov 2007;6:29–40.
2
2
3. Traut TW. Physiological concentrations of purines and pyri-
midines. Mol Cell Biochem 1994;140:1–22.
4. Lanzetta PA, Alvarez LJ, Reinach PS, Candia OA. An
improved assay for nanomole amounts of inorganic phos-
phate. Anal Biochem 1979;100:95–7.
2
.
.
Brown ED, Wright GD. Antibacterial drug discovery in the
resistance era. Nature 2016;529:336–43.
Kinch MS, Patridge E, Plummer M, Hoyer D. An analysis of
FDA-approved drugs for infectious disease: antibacterial
agents. Drug Discov Today 2014;19:1283–7.
3
2
5. Liger D, Masson A, Blanot D, et al. Over-production, purifica-
tion and properties of the uridine-diphosphate-N-acetylmur-
amate: L-alanine ligase from Escherichia coli. Eur J Biochem
4
.
Gwynn MN, Portnoy A, Rittenhouse SF, Payne DJ. Challenges
of antibacterial discovery revisited. Ann NY Acad Sci 2010;
1213:5–19.
1
995;230:80–7.
5
.
.
Smith CA. Structure, function and dynamics in the Mur fam-
ily of bacterial cell wall ligases. J Mol Biol 2006;362:640–55.
Mengin-Lecreulx D, Flouret B, van Heijenoort J. Cytoplasmic
2
6. Auger G, Martin L, Bertrand J, et al. Large-scale preparation,
purification, and cristallization of UDP-N-acetylmuramoyl-L-
alanine: D-glutamate ligase from Escherichia coli. Protein
Expr Purif 1998;13:23–9.
6
steps of peptidoglycan synthesis in Escherichia coli.
Bacteriol 1982;151:1109–17.
J
ꢀ
27. Gordon E, Flouret B, Chantalat L, et al. Crystal structure of
UDP-N-acetylmuramoyl-L-alanyl-D-glutamate: meso-diamino-
pimelate ligase from Escherichia coli. J Biol Chem 2001;276:
7.
8.
9.
Hrast M, Sosi ꢀc I, Sink R, Gobec S. Inhibitors of the peptido-
glycan biosynthesis enzymes MurA-F. Bioorg Chem 2014;55:
2–15.
1
0999–1006.
Barreteau H, Kova ꢀc A, Boniface A, et al. Cytoplasmic steps of
peptidoglycan biosynthesis. FEMS Microbiol Rev 2008;32:
2
2
3
8. Dementin S, Bouhss A, Auger G, et al. Evidence of a func-
tional requirement for a carbamoylated lysine residue in
MurD, MurE and MurF synthetases as established by chem-
ical rescue experiments. Eur J Biochem 2001;268:5800–7.
9. Wayne PA, CLSI. Methods for dilution antimicrobial suscepti-
bility tests for bacteria that grow aerobically. Approved
standard—Tenth edition. CLSI Document M07-A10. Clinical
and Laboratory Standards Institute, 2015.
0. The European Committee on Antimicrobial Susceptibility
Testing. Breakpoint tables for interpretation of MICs and
zone diameters. Version 5.0, 2015. Available from: http://
www.eucast.org.
168–207.
Kouidmi I, Levesque RC, Paradis-Bleau C. The biology of Mur
ligases as an antibacterial target. Mol Microbiol 2014;94:
242–53.
1
0. El Zoeiby A, Sanschagrin F, Levesque RC. Structure and func-
tion of the Mur enzymes: development of novel inhibitors.
Mol Microbiol 2003;47:1–12.
1
1. Hrast M, Turk S, Sosi ꢀc I, others, et al. Structure-activity rela-
tionships of new cyanothiophene inhibitors of the essential
peptidoglycan biosynthesis enzyme MurF. Eur J Med Chem
2
013;66:32–45.
ꢀ
31. Sim ꢀc i ꢀc M, Sosi ꢀc I, Hodo ꢀs ꢀc ek M, et al. The binding mode of
second-generation sulfonamide inhibitors of MurD: clues for
rational design of potent MurD inhibitors. PLoS ONE 2012;7:
e52817.
1
1
2. Sink R, Barreteau H, Patin D, et al. MurD enzymes: some
recent developments. Biomol Concepts 2013;4:539–56.
3. Barreteau H, Sosi ꢀc I, Turk S, et al. MurD enzymes from differ-
ent bacteria: evaluation of inhibitors. Biochem Pharmacol
2
012;84:625–32.
32. Cavanagh J, Fairbrother WJ, Palmer AG, et al. Protein NMR
spectroscopy: principles and practice. 2nd ed. San Diego,
USA: Academic Press; 2007.
33. McDonald O, Lackey K, Davis-Ward R, et al. Aza-stilbenes as
potent and selective c-RAF inhibitors. Bioorg Med Chem Lett
2006;16:5378–83.
1
4. Zidar N, Toma sꢀi cꢁ T, Sꢀi nk R, et al. Discovery of novel 5-benzy-
lidenerhodanine and 5-benzylidenethiazolidine-2,4-dione
inhibitors of MurD ligase. J Med Chem 2010;53:6584–94.
ꢀ
1
5. Zidar N, Toma ꢀs i ꢁc T, Sink R, et al. New 5-benzylidenethiazoli-
din-4-one inhibitors of bacterial MurD ligase: design,

JOURNAL OF ENZYME INHIBITION AND MEDICINAL CHEMISTRY
1017
3
3
4. Waterson AG, Stevens KL, Reno MJ, et al. Alkynyl pyrimi- 42. Zega A. NMR methods for identification of false positives in
dines as dual EGFR/ErbB2 kinase inhibitors. Bioorg Med
Chem Lett 2006;16:2419–22.
biochemical screens. J Med Chem 2017;60:9437–47.
43. Arsovska E, Mali G, Goli ꢀc Grdadolnik S, Zega A. NMR in
5. Stevens KL, Reno MJ, Alberti JB, et al. Synthesis and evalu-
ation of pyrazolo[1,5-b]pyridazines as selective cyclin-
dependent kinase inhibitors. Bioorg Med Chem Lett 2008;18:
the design of antibacterials. Curr Med Chem 2014;21:
4
328–46.
ꢀ ꢀ
4
4. Simcic M, Pureber K, Kristan K, et al. A novel 2-oxoindolinyli-
dene inhibitor of bacterial MurD ligase: enzyme kinetics,
protein-inhibitor binding by NMR and a molecular dynamics
study. Eur J Med Chem 2014;83:92–101.
5758–62.
3
6. Boehm JC, Bower MJ, Gallagher TF, et al. Phenoxypyrimidine
inhibitors of p38a kinase: synthesis and statistical evaluation
of the p38 inhibitory potencies of a series of 1-(piperidin-4-
yl)-4-(4-fluorophenyl)-5-(2-phenoxypyrimidin-4-yl) imidazoles.
Bioorg Med Chem Lett 2001;11:1123–6.
4
5. Sosi ꢀc I, Barreteau H, Sim ꢀc i ꢀc M, et al. Second-generation sul-
fonamide inhibitors of D-glutamic acid-adding enzyme: activ-
ity optimisation with conformationally rigid analogues of D-
glutamic acid. Eur J Med Chem 2011;46:2880–94.
3
7. Chamberlain SD, Wilson JW, Deanda F, et al. Discovery of
4
,6-bis-anilino-1H-pyrrolo[2,3-d]pyrimidines: potent inhibitors
4
4
4
6. Bertrand JA, Auger G, Fanchon E, et al. Crystal structure of
UDP-N-acetylmuramoyl-L-alanine: D-glutamate ligase from
Escherichia coli. Embo J 1997;16:3416–25.
7. Bertrand JA, Auger G, Martin L, et al. Determination of the
MurD mechanism through crystallographic analysis of
enzyme complexes. J Mol Biol 1999;289:579–90.
8. Bertrand JA, Fanchon E, Martin L, et al. “Open” structures of
MurD: domain movements and structural similarities with
folylpolyglutamate synthetase. J Mol Biol 2000;301:1257–66.
of the IGF-1R receptor tyrosine kinase. Bioorg Med Chem
Lett 2009;19:469–73.
3
3
4
4
8. Hopkins AL, Groom CR, Alex A. Ligand efficiency: a useful
metric for lead selection. Drug Discov Today 2004;9:430–1.
9. Murray CW, Erlanson DA, Hopkins AL, et al. Validity of ligand
efficiency metrics. ACS Med Chem Lett 2014;5:616–8.
0. Leeson PD, Young RJ. Molecular property design: does
everyone get it? ACS Med Chem Lett 2015;6:722–5.
1. Elkins JM, Fedele V, Szklarz M, et al. Comprehensive charac-
terization of the published kinase inhibitor set. Nat 49. Gaulton A, Hersey A, Nowotka M, et al. ’The ChEMBL data-
Biotechnol 2016;34:95–103. base in 2017.’ Nucleic Acids Res 2017;45:D945–54.