M. Hrast et al.
Bioorganic & Medicinal Chemistry Letters 40 (2021) 127966
Table 1 (continued)
1
2
c
Cpd
X
Y
R
R
MurC
MurD
MurE
MurF
MIC [mM]
a
b
a
b
a
b
a
b
%
IC50 [µM]
%
IC50 [µM]
%
IC50 [µM]
%
IC50 [µM]
SA
EC
O
O
O
N
N
N
O
3
1
2
N
N
C
C
42%
56%
151
82
47%
70%
160
85
10%
46%
/
40%
58%
201
71
0.031
0.25
NH
3
150
>0.25
>0.25
S
a
%
of inhibition of the enzyme activity at 100 µM of the tested compound. Data are means of two independent experiments. Standard deviations are within 10% of
the means.
b
Concentration of the inhibitor that reduces the activity of enzyme by 50%. The IC50 values were determined for the compounds with % of inhibition ≥ 40%.
c
Minimal concentration of an inhibitor to inhibit the growth of specific bacteria. SA – Staphylococcus aureus; EC – Escherichia coli.
several different aromatic or heterocyclic rings were introduced to
alternate the furan moiety. We then set to investigate the importance of
the nitrogen atom and its position in the pyridine core of the molecule.
Finally, a variety of compounds were prepared, where the tetrazole
fragment was replaced with bioisosteres and different heterocyclic rings.
A series of 20 stilbene derivatives was synthesized and evaluated for
inhibitory potency against Mur ligases (MurC–MurF) and antibacterial
activity against two representative bacterial strains: E. coli ATCC 25922
and S. aureus ATCC 29213.
basic conditions. The amide derivative 29 was prepared from 6a by
hydrolysis of the nitrile group under mild basic conditions with
hydrogen peroxide. 2-Oxazoline (30) and 2-imidazoline (31) derivatives
were obtained using ethanolamine and ethylenediamine, respectively,
in the presence of a catalytic amount of sulphur. Condensation of
cysteamine and nitrile was a straightforward method for the preparation
of a 2-thiazoline derivative 32 (Scheme 1).
First, compounds devoid of tetrazole (6a) or the furan fragment (5a)
were assayed in the Malachite green assay18 to confirm the importance
of both moieties. Both compounds were less potent inhibitors of Mur
ligases in comparison to compound 1, which demonstrates the signifi-
cance of these two fragments. The first set of compounds (17a–27a) with
One of the key features to be considered in the design of inhibitors
that act on multiple targets is the degree of structural similarity between
the enzymes. Mur ligases C–F reportedly share conserved amino acid
regions, structural features and the common kinetic mechanism.7,14
A
modifications of the furan ring (R ) generally afforded less potent in-
1
common claim found in the literature is that a high degree of similarity
between Mur ligases C–F should alleviate the quest for inhibitors that
target more than one enzyme.7 However, this has been only partially
hibitors of MurC–MurF. Nonetheless, compound 17a inhibited MurD
ligase in the same range as the starting compound 1, indicating the
importance of the oxygen atom at the o-position of the aromatic or
heterocyclic ring. Next, compounds 7b and 7c, where the central pyri-
dine ring was replaced with phenyl and o-substituted pyridine, respec-
tively, showed comparable inhibition of MurD as 1. However, while
compound 7b was inactive against the other three ligases, 7c had a
comparable inhibitory profile against all Mur ligases as compound 1.
Obviously, the nitrogen atom is important for the interaction with MurE
and MurF ligases. Next, replacing the tetrazole part of molecule 1 with
bioisosteres and other heterocycles resulted in multiple inhibitors
against all four ligases. The compound with the archetypal tetrazole
bioisostere - carboxylic acid (28), exhibited similar inhibitory potency
against MurD and MurF ligases as compound 1, and it was more potent
,15
1
0
validated by a small number of dual inhibitors and by our hit com-
pound 1.9 Importantly, the binding site of compound 1 has only been
confirmed by NMR on MurD. One of the questions that arise is whether
the increase in activity against one enzyme could lead to a proportional
increase in activity against another enzyme. With that in mind, a mul-
tiple sequence alignment (MSA) and 3D superposition of UMAG binding
sites for all four ligases MurC–F were made to check the similarity
among all four enzymes. The results are available among supporting
data (Fig. S1, Table S3). The MSA demonstrates that ligases MurC–F
show low sequence identity (26%). Furthermore, amino acids that were
identified by docking to be relevant for binding are only partially
conserved among all four ligases (Table S4). On the other hand, the
topologies of UMAG binding sites are similar in Mur ligases with an
average RMSD of 4.4 Å. Reasonable topological similarity and low
sequence identity indicate that it is theoretically possible on one hand i)
to design multiple inhibitors that target all four enzymes, but on the
other hand ii) it is uncertain whether an increase in potency against one
enzyme is followed by an increase in potency against the other ligase.
Commercially available bromides (5-bromonicotinonitrile (2a), 3-
bromobenzonitrile (2b) and 2-bromoisonicotinonitrile (2c)) and tribu-
tyl(vinyl)tin were used in a Stille coupling to produce the vinyl in-
termediates (3a–c), followed by a Heck reaction using 5-bromo-2-iodo-
,16
Table 2
Calculated ADME descriptors for active compounds with QikProp.
a
b
c
d
Cpd.
PSA
logP
logS
Pcaco
Oral Absorption (%)
1
80.5
67.4
67.6
80.5
76.6
63.3
69.1
47.6
50.4
63.7
3.5
3.1
4.4
3.9
4.3
4.2
3.3
4.7
4.9
5.6
ꢀ 5.7
ꢀ 5.1
ꢀ 6.4
ꢀ 6.0
ꢀ 6.5
ꢀ 5.4
ꢀ 5.1
ꢀ 6.0
ꢀ 6.6
ꢀ 6.9
310
92.1
89.7
100
96.7
96.6
90.3
94.2
100
100
100
5
a
310
7b
7c
565
428
1
2
2
3
7a
310
8
9
0
140
469
1
,3-dimethylbenzene to form the corresponding trans stilbene de-
3516
2359
3818
1
7
rivatives (4a–c) (Scheme 1). Then, different boronic acids were used in
Suzuki couplings to yield intermediates 6a–c and 7a–14a. Additionally,
compounds 15a and 16a were synthesized under Buchwald-Hartwig
31
3
2
a
Description and recommended values: Van der Waals surface area of polar
2
b
cross-coupling conditions using Pd(OAc)
2
and Xantphos (4,5-bis
nitrogen and oxygen in Å (7–200), Predicted octanol/water partition coeffi-
c
(
diphenylphosphino)-9,9-dimethylxanthene) as catalysts. Finally, the
cient (–2.0 – 6.5), Conformation-independent predicted aqueous solubility
d
tetrazole derivatives (5a, 7b, 7c, 17a–27a,) were synthesized using
ammonium chloride and sodium azide in DMF. The tetrazole bioisostere
of carboxylic acid 28 was synthesized from nitrile 6a under reflux using
(–6.5 – 0.5), Predicted apparent Caco-2 cell permeability in nm/sec (<25 poor,
>500 great).
4