Discovery of pyrrolidine-2,3-diones as novel inhibitors of p. Aeruginosa pbp3

Article abstract of DOI:10.3390/antibiotics10050529

The alarming threat of the spread of multidrug resistant bacteria currently leaves clinicians with very limited options to combat infections, especially those from Gram-negative bacteria. Hence, innovative strategies to deliver the next generation of antibacterials are urgently needed. Penicillin binding proteins (PBPs) are proven targets inhibited by β-lactam antibiotics. To discover novel, non-β-lactam inhibitors against PBP3 of Pseudomonas aeruginosa, we optimised a fluorescence assay based on a well-known thioester artificial substrate and performed a target screening using a focused protease-targeted library of 2455 compounds, which led to the identification of pyrrolidine-2,3-dione as a potential scaffold to inhibit the PBP3 target. Further chemical optimisation using a one-pot three-component reaction protocol delivered compounds with excellent target inhibition, initial antibacterial activities against P. aeruginosa and no apparent cytotoxicity. Our investigation revealed the key structural features; for instance, 3-hydroxyl group (R²) and a heteroaryl group (R¹) appended to the N-pyrroldine-2,3-dione via methylene linker required for target inhibition. Overall, the discovery of the pyrrolidine-2,3-dione class of inhibitors of PBP3 brings opportunities to target multidrug-resistant bacterial strains and calls for further optimisation to improve antibacterial activity against P. aeruginosa.

Full text of DOI:10.3390/antibiotics10050529

antibiotics
Article
Discovery of Pyrrolidine-2,3-diones as Novel Inhibitors of
P. aeruginosa PBP3
Arancha López-Pérez 1,2 , Stefan Freischem 3,4 , Immanuel Grimm 1, Oliver Weiergräber 3,
Andrew J. Dingley 3,4 , María Pascual López-Alberca 5, Herbert Waldmann 5, Waldemar Vollmer 2,
Kamal Kumar 1,5 and Cuong Vuong 1,
*
1
AiCuris Anti-Infectives Cures GmbH, 42117 Wuppertal, Germany; arancha.lopez@aicuris.com (A.L.-P.);
immanuel.grimm@Aicuris.com (I.G.); kamal.kumar@aicuris.com (K.K.)
Centre for Bacterial Cell Biology, Biosciences Institute, Newcastle University,
Newcastle upon Type NE2 4AX, UK; waldemar.vollmer@newcastle.ac.uk
Institute of Biological Information Processing, Structural Biochemistry, Forschungszentrum Jülich,
52428 Jülich, Germany; s.freischem@fz-juelich.de (S.F.); o.h.weiergraeber@fz-juelich.de (O.W.);
a.dingley@fz-juelich.de (A.J.D.)
Institut für Physikalische Biologie, Heinrich-Heine-Universität Düsseldorf, 40225 Düsseldorf, Germany
Max Plank Institute of Molecular Physiology, Department of Chemical Biology, 44227 Dortmund, Germany;
mpascualla@gmail.com (M.P.L.-A.); herbert.waldmann@mpi-dortmund.mpg.de (H.W.)
Correspondence: cuong.vuong@aicuris.com; Tel.: +49-202-31763-1163
2
3
4
5
*
Abstract: The alarming threat of the spread of multidrug resistant bacteria currently leaves clinicians
with very limited options to combat infections, especially those from Gram-negative bacteria. Hence,
innovative strategies to deliver the next generation of antibacterials are urgently needed. Penicillin
ꢀꢁꢂꢀꢃꢄꢅꢆꢇ
ꢀꢁꢂꢃꢄꢅꢆ
binding proteins (PBPs) are proven targets inhibited by
β-lactam antibiotics. To discover novel,
Citation: López-Pérez, A.; Freischem,
S.; Grimm, I.; Weiergräber, O.;
non- -lactam inhibitors against PBP3 of Pseudomonas aeruginosa, we optimised a fluorescence assay
β
based on a well-known thioester artificial substrate and performed a target screening using a focused
protease-targeted library of 2455 compounds, which led to the identification of pyrrolidine-2,3-
dione as a potential scaffold to inhibit the PBP3 target. Further chemical optimisation using a
one-pot three-component reaction protocol delivered compounds with excellent target inhibition,
initial antibacterial activities against P. aeruginosa and no apparent cytotoxicity. Our investigation
revealed the key structural features; for instance, 3-hydroxyl group (R2) and a heteroaryl group
(R1) appended to the N-pyrroldine-2,3-dione via methylene linker required for target inhibition.
Overall, the discovery of the pyrrolidine-2,3-dione class of inhibitors of PBP3 brings opportunities to
target multidrug-resistant bacterial strains and calls for further optimisation to improve antibacterial
activity against P. aeruginosa.
Dingley, A.J.; López-Alberca, M.P.;
Waldmann, H.; Vollmer, W.; Kumar,
K.; Vuong, C. Discovery of
Pyrrolidine-2,3-diones as Novel
Inhibitors of P. aeruginosa PBP3.
Antibiotics 2021, 10, 529. https://
doi.org/10.3390/antibiotics10050529
Academic Editor: Muriel Masi
Received: 16 April 2021
Accepted: 3 May 2021
Published: 4 May 2021
Keywords: PBPs; P. aeruginosa; penicillin binding proteins; HTS; β-lactams; multi-drug resistance;
antibiotics; antibacterial agents; drug discovery and screening
Publisher’s Note: MDPI stays neutral
with regard to jurisdictional claims in
published maps and institutional affil-
iations.
1. Introduction
Pseudomonas aeruginosa is a ubiquitous Gram-negative bacterium and an opportunistic
pathogen that causes both acute and chronic infections in humans [ ]. P. aeruginosa is a
1,2
major cause of nosocomial infections in immunocompromised and critically ill patients
treated in intensive care units (ICUs) and the main cause of infections with a high mortality
rate in cystic fibrosis patients. Furthermore, P. aeruginosa is a leading cause of infections that
arise from wounds and burn injuries, and community-acquired and ventilator-associated
pneumonias [3]. Notably, P. aeruginosa is intrinsically resistant to most antibiotics because
of several multidrug efflux pumps, the low permeability of the cell envelope and its ability
to adapt to unfavourable growth or environmental conditions (especially during infections)
Copyright:
© 2021 by the authors.
Licensee MDPI, Basel, Switzerland.
This article is an open access article
distributed under the terms and
conditions of the Creative Commons
Attribution (CC BY) license (https://
creativecommons.org/licenses/by/
4.0/).
by virtue of its large and versatile genome [4]. The World Health Organisation (WHO)
has recently categorised P. aeruginosa as a leading and critical bacterium in its priority
Antibiotics 2021, 10, 529. https://doi.org/10.3390/antibiotics10050529
https://www.mdpi.com/journal/antibiotics

Antibiotics 2021, 10, 529
2 of 18
pathogen list, in order to accelerate the development of innovative antibacterials against
this pathogen [5]. Thus, the discovery of new small molecules and potentially novel
mechanism of action on a target represents a key strategy to combat P. aeruginosa infections
and break the current resistance mechanisms against available antibiotics.
Bacterial cell shape is maintained through the essential peptidoglycan (PG) layer of
the cell wall. The PG plays a key role in the growth cycle and protects the bacterial cell
from bursting by turgor, which prevents cell lysis and, eventually, death [
possess sophisticated mechanisms to maintain the integrity and biogenesis of the PG layer
as well as coordinate the synthesis of the PG layer during the life cycle of the cell [ ].
PG is a net-like polymer consisting of glycan strands made of alternating, -1-4-linked N-
6–8]. Bacteria
6–8
β
acetylglucosamine and N-acetylmuramic acid residues, which are linked by short peptides.
Glycosyltransferases (GTases) polymerise the glycan strands from lipid II precursor. DD-
transpeptidases (TPases) belonging to the family of penicillin binding proteins (PBPs)
form the peptide cross-links [
6
–8]. Class A PBPs are bifunctional GTase/TPases, while the
class B PBPs are monofunctional PG transpeptidases [
6–
11]. Class C PBPs do not form
DD-crosslinks but display DD-endopeptidase or DD-carboxypeptidase activity. Additional
monofunctional GTases that associate with the monofunctional class B PBPs have been
described as member of the SEDS family (e.g., FtsW and RodA) [9,12]. The GTase and
TPase activities are essential for bacterial survival, and the inhibition of these activities
leads to cell death.
The TPase catalytic signature contains nine conserved residues in three motifs: S*XXK
in the N-terminal part of helix
3 [11 13]. The nucleophilic serine (S*) attacks the carbonyl carbon of the D-alanine at
position 4 of the donor pentapeptide substrate to transfer the tetrapeptide to the -amino
α2, SXN in the α4-α5 linker and KS/TG in the strand
β
,
ε
group of a meso-diaminopimelic acid (mDAP) residue at position 3 of an acceptor peptide.
The TPase domain of each PBP is highly conserved, but different bacteria have varying
numbers of PBPs, and these PBPs can have different roles. Gram-negative species, such as
Escherichia coli or P. aeruginosa (which have been studied extensively), harbour more than
eight PBPs, of which some are essential and validated antibacterial targets for β-lactam
antibiotics. Gram-negative rod-shaped bacteria require at least one class A bifunctional
TPase/GTase, PBP1A or PBP1B and the monofunctional TPases PBP2 and PBP3 to growth
and survive [11,13].
PBPs are targeted by
dramatically decreased with the emergence and spread of multidrug-resistant (MDR)
strains harbouring a vast range of genetically or plasmid-encoded -lactamases and addi-
tional resistance mechanisms (e.g., efflux transporters and target mutations). -lactamases
abolish the antibacterial activity of -lactam antibiotics by hydrolysing the -lactam ring.
A further mechanism to bypass the effect of -lactams is the acquisition of alternative
PBPs encoded in horizontal-genetic elements or by mutations in the chromosomal genes
encoding PBPs (e.g., the mecA gene encoding PBP2a in Staphylococcus aureus) [14 15]. In-
terestingly, -lactams are no longer the only chemical class capable of inhibiting PBPs.
β-lactam antibiotics. However, the efficacy of β-lactams has
β
β
β
β
β
,
β
The discovery of cyclic boronates and diazabicyclooctanes (DBOS) demonstrate that the
TPase domain can be inhibited by mechanisms distinct from the known binding mode
of β-lactam [16–18]. Therefore, the discovery of other non-β-lactam PBP-targeting antibi-
otics may lead to the development of new, urgently needed antibacterials that will not be
inactivated by β-lactamases.
In this study, we selected PBP3 of P. aeruginosa (PaPBP3) for a target-based inhibitor
screening to discover novel chemical scaffolds. Because PaPBP3 is essential for the survival
of P. aeruginosa, this protein represents a key target for antibacterial drug discovery and
development. It is a clinical validated target, and the absence of a human counterpart
reduces the risk of drug-related side effects. The catalytic domain of PBP3 is located in
the periplasm, making it accessible to potential small molecule inhibitors. Furthermore,
we report the optimisation of the previously developed S2d assay to measure the TPase
activity of PaPBP3. Using the optimised assay, we screened a library of 2455 compounds

Antibiotics 2021, 10, 529
3 of 18
and identified novel potent PBP3 inhibitors derived from the pyrrolidine-2,3-dione scaffold
that target Gram-negative bacterial, especially P. aeruginosa.
2. Results
2.1. Optimisation of the S2d Assay for Fluorescence-Based Screening
Thioesters of hippuric acids have been previously used in the presence of a carbonyl
acceptor (e.g., D-alanine) as an artificial substrate for measuring the TPase activity of several
PBPs, as illustrated in Figure 1A. The TPase reaction can be measured by quantifying the
transpeptidation product using LC-MS/MS or by quantifying the released mercaptoacetate
product using thiol dyes [19–23].
Traditional colorimetric dyes such us aldrithiol (2,2’dipyridyldisulfide) or Ellman’s
reagent (5,5’-dithiobis-(2-nitrobenzoic acid)) have been used to investigate PBPs kinetics (e.g.,
PBP1B [24], PBP3 [22] and PBP5 [25] of E. coli, PBP2B [22] and PBP2x [19
pneumoniae, PBP1 and PBP3 of Enterococcus hirae [22], the DD-peptidase R61 of Streptomyces [20
and the DD-peptidase R39 of Actinomadura sp. R39 [18 28]) and for quantifying the effect of
,25–27] of Streptococcus
]
,
protein partners on the activity of E. coli PBP1B [29]. However, colorimetric readouts limit
the use of the assay in drug discovery because the intrinsic spectra of many test compounds
overlap with that of aldrithiol or Ellman’s reagent. Therefore, we replaced the colorimetric
dye with a fluorescent thiol detecting dye, monobromobimane (mBBr) (Ex/Em: 394/490 nm).
The emission spectrum of mBBr does not interfere with the absorbance wavelength of most
common compounds, thus improving the quality of the read-out spectra.
Figure 1. Transpeptidation biochemical assay. (A) Enzymatic reaction. The thiol substrate is used
as an acyl donor for the transpeptidation reaction in the presence of carbonyl acceptors (i.e., D-
alanine). The reaction yields a transpeptidation product and a mercaptoacetate. The thiol group in
the mercaptoacetate enables the indirect measurement of the TPase reaction by its quantification with
a thiol dye. Figure adapted from [25]. (
determined using the average and SD of the positive (black dots: +PBP3) and negative (grey dots:
without (w/o) PBP3). (
) The Z’-factor of the fluorometric S2d-based assay for PaPBP3. Red dots
B
) Fluorometric S2d-based assay for PaPBP3. The Z’factor was
C
represent the Z’factor calculated for each plate measured and the Z’factor indicated and highlighted
in grey was calculated using the data from eight replicates.

Antibiotics 2021, 10, 529
4 of 18
Initial experiments were performed to determine the optimal conditions, which were
found to be 0.3 µ
M PaPBP3, 0.8 mM artificial substrate S2d, 1 mM D-alanine and 0.05 mM
thiol dye. The assays were carried out in the presence of 5% DMSO and 0.01% Triton X100
to avoid the formation of colloidal structures and adsorption of protein and/or compounds
to surfaces [30
,
31]. The Z’-factor determined for the assay was 0.73
±
0.28, indicating the
32] (Figure
suitability and robustness of the assay for high throughput screening (HTS) [29
,
1B,C). Additional statistical parameters, including the signal-to-noise ratio and the coefficient
of variation and their values, indicated that the assay was sensitive and robust (Table S1).
The newly modified fluorescence-based S2dfluo assay was then employed in a HTS
campaign using a focused library consisting of 2455 pre-selected cysteine protease-targeting
inhibitors (ChemDiv, San Diego, USA) to identify potential PaPBP3 inhibitors. Screening
of the library identified 27 hits that inhibited PBP3 over a range of 60% to 100% at a
concentration of 100 µM (Table 1 and Table S2). Interestingly, all potent PBP3 inhibitors
contained a pyrrolidine-2,3-dione core, indicating a potential conserved scaffold that could
be used as a new starting point for the further development of PBP3 TPase inhibitors. Three
proprietary Max Plank Institute of Molecular Physiology (MPI) compounds containing the
pyrrolidine-2,3-dione-like core were also added to the hit selection (Table 1).
Table 1. Biochemical characterisation of the pyrrolidine-2,3-dione cluster.
IC50 Determination with PaPBP3 a (µM)
Compound
Source
PBP3 Inhibition at 100 µM (%)
S2dfluo Assay
FP-Bocillin FL Assay
Br
OH
O
MPI b
ND b,c
O
1
28 ± 9
7 ± 3
N
F
Cl
OH
O
CH3
O
N
ChemDiv
8015-0105
2
101 ± 0.4
100 ± 0.8
4 ± 6
17 ± 9
O
H
3C
CH3
ChemDiv
8015-0104
3
4
19 ±1
>100
OH
N
O
S
O
O
ChemDiv
F232-0415
101
24 ± 20
61 ± 1
>100
H3C
O
CH3
O
H3
C
ND b,c
5
MPI
69 ± 13
OH
N
O
O
ChemDiv
8015-3001
NH
6
7
92 ± 7
63 ± 4
9 ± 6
H3C
O
H
3C
Cl
OH
O
C
H3
O
F
N
ChemDiv
C202-3660
N
62
65 ± 19
58 ± 21
HN
S
N
O
H
3C
H
3C
CH3
CH3
HO
O
N
O
MPI
[33]
ND b,c
8
9
66 ± 16
93 ± 18
O
CH3
O
CH3
ChemDiv
D159-0900
60
99 ± 1
18 ± 3
a
The IC50 values represent the average and the standard deviation of three independent measurements; b Max Plank Institute of Molecular
Physiology, Department of Chemical Biology, Dortmund, Germany; c ND, not determined.

Antibiotics 2021, 10, 529
5 of 18
To further characterise the inhibitory activity of the 30 hits, we performed concentration-
dependent experiments using the S2dfluo assay and compounds between 100
1.56 M. Table 1 shows nine compounds that inhibited the PBP3 enzyme in a concentration-
dependent manner. Notably, compound exhibited an excellent potency, with an IC50 of
M in the S2dfluo assay. The remaining 21 compounds did not show concentration-
dependent inhibition but had only an effect at the highest test concentrations of 100
(Table S2), suggesting nonspecific inhibition. The IC50 values of the test compounds in the
S2dfluo assay varied from 4 M to >100 M and only compounds and 4 inhibited the
enzyme with IC50 < 50 M. This concentration was the second highest concentration used
µM and
µ
2
4
±
6 µ
µM
µ
µ
1, 2, 3
µ
in the assay and was the selected threshold for hits selection.
Bocillin FL is a fluorescent penicillin V derivate used as a competitor for PBP in-
hibitors [32
(FP) of Bocillin FL in the presence of decreasing concentrations of compounds
and Table S2). Because of the low potency of these compounds when compared with that of
other covalent PBP inhibitors (such as -lactams with IC50 values in the nM concentration
,
34–
36]. To further validate the hits, we measured the fluorescence polarisation
1
to (Table 1
9
β
range), we incubated the inhibitors for 30 min prior to the addition of Bocillin FL. Seven of
the nine compounds competed for the TPase site and inhibited the binding of Bocillin FL in
a dose-dependent manner, indicating a potential occupation of the active site. In contrast,
compounds
but failed to inhibit Bocillin FL binding at test concentrations. These data suggested that
compounds and
were presumably bound to PaPBP3 at a site distinct from the TPase
3
and
4
showed IC50 values of 19
±
1
µM and 24
±
20 µM in the S2dfluo assay
3
4
pocket in a way that still affects the enzyme activity by an unknown mechanism.
Based on the observed target inhibition and supported by the in vitro biochemical
data, we selected compounds
synthesis of a pyrrolidine-2,3-dione focused library of small analogues. Compound
1
and
2
as promising starting points for the design and
was
1
used as the primary scaffold for optimisation because of the simplicity of its structure and
the relative weak activity in other molecules with N-aryl pyrrolidine-2,3-diones (2–5, Table
1). The design of this focused library aimed to: (i) preliminarily explore the structure–
activity relationship (SAR) of the compounds and (ii) confirm the inhibitory target activity
of this compound cluster. Notably, small molecules with a pyrrolidine-2,3-dione core have
been identified previously in a virtual screen as potential inhibitors of tyrosine phosphatase
B (MptpB) from Mycobacterium tuberculosis. However, the inhibition activities reported for
small molecules that are similar to the hits identified here were lower than 40% and thus
did not justify further studies [37]. Among other known biological activities, pyrrolidine-
2,3-dione molecules are reported as stabilisers of the potential 14-3-3-PMA2 protein–protein
interaction of Nicotiana tabacum with a Kd of 80 µM [38,39]. Thus, we discovered herein a
novel biological potential of the pyrrolidine-2,3-dione class of compounds and pursued
their further development as an innovative antibacterial class with a new mechanism of
action against PaPBP3.
2.2. Synthesis of Compound 1 Analogues and Generation of the Pyrrolidine-2,3-dione Cluster
Inspection of the structures of the initial list of inhibitory and non-inhibitory pyrrolidine-
2,3-dione compounds (Table 1) led us to conclude that the best active molecules embody a
3-hydroxy-2,5-dihydro-1H-pyrrol-2-one core with three substitutions. We observed that
halogen substituted benzoyl- and phenyl groups were required for activity at positions 4
and 5, respectively, of the core scaffolds (Figure 2). Therefore, the pyrrolidine-2,3-dione
core was modified at the nitrogen (R1) and at position 3 (R2), and, thus, 15 analogues were
designed and synthesised (Scheme 1 and Table 2).

Antibiotics 2021, 10, 529
6 of 18
Figure 2. PBP3 targeting substituted pyrrolidin-2,3-diones.The pyrrolidine-2,3-dione core was mod-
ified at the nitrogen (R1) and at position 3 (R2). The rest of the molecule was maintained in the
pyrrolidine-2,3-dion derivatives.
Scheme 1. Synthesis of pyrrolidin-2,3-diones. One-pot three-component reaction used to synthesise
the pyrrolidine-2,3-dione derivatives.
The desired pyrrolidine-2,3-diones were prepared using a one-pot three-component
reaction protocol reported previously (Scheme 1) [40]. In a typical reaction procedure
leading to pyrrolidine-2,3-diones with R1 substitution, a mixture of amine (R1NH2) (1.0
mmol) and aldehyde (1.0 mmol) in dry dioxane (2 mL) was stirred for 30 min at 50 ◦C.
Then acetic acid (0.5 mL, 8.74 mmol) and p-bromophenyl pyruvate (1.0 mmol) were added
and the resulting mixture was stirred for 2 h at 90 ◦C. After cooling the reaction mixture
to ambient temperature, the precipitates were filtered, washed with dioxane and dried in
vacuo to yield the target compounds with a 3-hydroxy group. Different N-substitutions
(heterocycles in particular) were bound to the pyrrolidone via alkyl linkers that were also
explored, considering the desired physical chemical properties of the final compounds. To
this end, fluoro-substituted benzyl groups, benzylic five-membered heterocycles (furan and
thiazole) and benzimidazole and indoles were linked to the core structure via methylene or
ethylene linkers (Figure 3).
Treatment of the 1,4,5-substituted 3-hydroxy-pyrrolidines with ethyl amine or mor-
pholine amine in an acidic solvent yielded the pyrrolidines with corresponding amine
substitutions at position 3 (Figure 3).

Antibiotics 2021, 10, 529
7 of 18
Figure 3. Pyrrolidine-2,3-derivatives. The pyrrolidine-2,3-dione core was modified at the nitrogen
(R1), whereas the hydroxyl substitution at position 3 (R2) was maintained to generate compounds 32
–
42. Position 3 (R2) was substituted in analogues 43
–45 to investigate the essentiality of the hydroxyl
group.
Next, heterocyclic groups were introduced at the N-benzylic position as R1 to generate
more polar compounds with better logP and hydrogen-bond donor/acceptor profiles
(Figure 3). Pyrrolidones attached to the indole and aza-indole ring via C2-connection
and by a methylene linker (34–36) or ethylene linker (37) were synthesised following the
above-mentioned method (Scheme 1). The indole ring was replaced by benzimidazole
and other heterocycles (38). More indole-rings linked via the C3-connection and through
methylene (39) and ethylene (40) linkers were also obtained. Small heterocycles, such as
furan (41) and thiazole (42), were introduced as R1 substitutions via a methylene linker. As
a control to study the role of N-substitution on the biological activity of this class of small
molecules, an unsubstituted pyrrolidone (32) and the simpler N-ethyl pyrrolidone (33
were synthesised. Finally, a subset of this class of pyrrolidones was synthesised wherein
the 3-hydroxyl function was replaced by either ethylamine or morpholine (43–45).
)
2.3. Inhibition of PaPBP3 by Pyrrolidine-2,3-dione Derivates
Initially, we investigated the inhibitory effect of pyrrolidine-2,3-derivates on the TPase
activity of PaPBP3 (Table 2) by determining the IC50 using the biochemical assays with
S2dfuo or FP-Bocillin FL. The results are summarised in Table 2.
Comparing the inhibitory activities of the new compounds revealed two key struc-
tural properties that are required for the inhibition of the PaPBP3: the 3-hydroxyl group
substitution (R2) and a bulky group such as a benzyl substitution or a heteroaryl group at
the R1 position. Replacement of the 3-fluorophenylgroup with a hydrogen atom (analogue
32) or an ethyl group (analogue 33) caused loss of the inhibitory effect, indicating that a
benzylic group engaged with PaPBP3, most likely through a hydrophobic interaction with
the active site of the target enzyme. In line with these results, when the benzyl group

Antibiotics 2021, 10, 529
8 of 18
was substituted with some benzylic heterocycles (for example, an indole in analogue 34),
we observed a two-fold increase in inhibition (IC50 = 14
µM) when compared with that
of the parental compound (IC50 = 28 M). Substitution of the methyl indole with ethyl
1
µ
thiazole in analogue 42 or a furan in analogue 41 abolished the inhibitory activity. Finally,
we explored whether the position wherein a heterocycle is linked to the N-pyrrolidinone
influences the activity by synthesizing and testing analogues 38 and 39. The additional
nitrogen group in the benzimidazole resulted in a drastic loss of the inhibitory activity
in by 38 and the C3-linkage of the indole moiety in 39 increases the IC50 values to 24
µM.
Interestingly, a similar C3-linked indole (40) attached via an ethylene linker displayed an
improved target inhibition of 8 µM, probably due to better fitting of the heterocycle into
the binding site of PaPBP3. The inhibitory potential of the corresponding C2-appended
indole (37) was almost reduced by two when compared with that of 40, indicating a narrow
structure-activity space to fit the ligand into binding site of the protein. In fact, the IC50
displayed by 37 was 16
pyrrolidone with 5-chloro substitution emerged as another interesting candidate with an
improved IC50 value of 3 M, suggesting that the chloride atom interacts with the protein.
µM and very similar to that of the analogue 34 (IC50 = 14 µM). A
µ
All pyrrolidones where R2 was not a hydroxyl group and was replaced by amines (43
–45)
lost their inhibitory effect on the activity of PaPBP3 and the Bocillin FL bound to the TPase
site. These results strongly suggest that the hydroxyl group plays an important role in
recognising and/or interacting with PaPBP3.
Table 2. Characterisation of the pyrrolidine-2,3-dione cluster. The biochemical characterisation of the pyrrolidine-2,3-dione
cluster was determined by S2dfluo, fluorescence polarisation (FP) with Bocillin FL and surface plasmon resonance (SPR).
The data represent the average and the standard deviations of three independent replicates. The minimal inhibitory
concentration (MIC) against two P. aeruginosa strains was determined in the presence of PMBN. The cytotoxicity of the
compounds was determined against the T-lymphoblast CCRF-CEM cell line, and the average and standard deviation from
three independent measurements is indicated.
MIC (µM) against
Biochemical Characterisation a
Cytotoxicity
CCRF-CEM
P. aeruginosa
IC50 (µM)
IC50 (µM)
FP-Assay
Kd (µM)
SPR
(µM)
PAO1 b
Nr
K28926 c
S2dfluo Assay
1
28 ± 9
7 ± 3
ND
ND
ND
6.44
9.82
ND
454
62.7
ND
22.9
ND
ND
ND
ND
ND
>100
>100
>100
3.13
3.13
>100
6.25
>100
6.25
25
>100
>100
>100
>100
>100
12.5
>100
>100
12.5
12.5
100
12.5
>100
25
>100
25
>100
>100
>100
>100
>100
94 ± 6
58 ± 6
>100
32
33
34
35
36
37
38
39
40
41
42
43
44
45
>100
>100
>100
>100
14 ± 9
3 ± 1
>100
16 ± 8
100 ± 11
24 ± 7
8 ± 4
>100
>100
>100
>100
>100
4 ± 4
0.6± 2
>100
17 ± 3
117 ± 4
11 ± 4
4 ± 2
>100
>100
>100
>100
>100
53 ± 5
1 ± 2
8 ± 6
>100
>100
93 ± 4
>100
>100
73 ± 9.2
52 ± 3
23 ± 11
PBP3 of P. aeruginosa; b PAO1 (wild-type strain) with 4
µ
g/mL PMBN; c K28926 (
∆
mexB;
∆
mexX,
∆
mexCD-oprJ) with 4
µ
g/mL PMBN;
a
FP, fluorescence polarisation; SPR, surface plasmon resonance; ND: not determined.
The IC50 values obtained by the FP assay (Table 2) correlated well with the IC50 values
obtained by the S2dfluor assay, indicating that the compounds competed for the same
catalytic TPase pocket site occupied by Bocillin FL as a representative
to exert target inhibition, which eventually causes bacterial cell death.
β-lactam analogue
Next, we confirmed the binding of the best inhibitors (analogues 34
, 35, 40, 38 and 37)
by surface plasmon resonance (SPR), using PaPBP3 coated on a CM5 chip as the ligand. The

Antibiotics 2021, 10, 529
9 of 18
sensorgrams (Figures A1C and A2C) indicated fast association and dissociation with all
compounds tested. Dissociation constants (Kd) were determined via steady-state analysis
using the maximum resonance signal obtained with a range of analyte concentrations. The
calculated Kd values of the investigated compounds ranged from 6.4 to 62.7 µM (Table 2).
These results indicate that, in general, the compounds bind to the protein at concentrations
on the order of the IC50. The Kd value obtained for compound 37 was too high to conclude
a specific binding, although the IC50 values calculated by both the S2dfluor and the Bocillin
FL assay indicated a good inhibition.
2.4. Antimicrobial Activities of Pyrrolidine-2,3-dione Derivates
The results presented above implied that seven pyrrolidine-2,3-dione derivatives
inhibited the activity of PaPBP3 in vitro. Consequently, we were very interested in inves-
tigating the potential of these compounds to affect the growth of P. aeruginosa. Two P.
aeruginosa bacterial strains (the wild-type PAO1 and the isogenic
oprJ efflux pump mutant) were used to determine the MIC and the potential role of the
main P. aeruginosa efflux systems in compound uptake [34 35]. To increase the antibacterial
sensitivity of the tested P. aeruginosa isolates to the potential inhibitors, we performed
MIC experiments in the presence and absence of a sub-MIC concentration (4 g/mL)
∆mexB ∆mexX ∆mexCD-
,
µ
of the outer membrane permeabilising polymyxin B nonapeptide (PMBN). The tested
compounds showed only antibacterial activities in the presence of PMBN, and no growth
inhibition was detected in the absence of PMBN at the highest compound concentration of
100 µM (data not shown). Thus, we concluded that the physicochemical properties of the
pyrrolidine-2,3-dione compounds were perhaps not yet favourable enough to cross the P.
aeruginosa cell envelope without a permeabiliser because of their high hydrophobicity.
The results in Table 2 indicated that analogues 34, 35, 37 and 39 inhibited the growth of
P. aeruginosa wild-type PAO1 and the efflux mutant in the presence of PMBN. Interestingly,
these results showed a direct correlation between target inhibition and, eventually, an
antibacterial effect. Surprisingly, the MICs of compounds 34, 35 and 39 were 2 to > 4-fold
lower against the wild type PAO1 when compare with the efflux mutant K2896. This effect
was even more pronounced with compound 40, which was not active against the mutant
but was active against the wild type strain. In contrast, compounds 1 and 41 showed MICs
against the P. aeruginosa K2896 of 12.5
µM and 25 µM, respectively, while no inhibitory
activity was detected against PAO1, suggesting that these two compounds were prone to
efflux by the Mex efflux system. Notably, compound 39 did not inhibit PaPBP3 in vitro
Thus, its antibacterial effect is likely caused by an alternative mechanism.
Additional bacterial strains were tested, and the MIC values are summarised in
Table S3. Overall, the antimicrobial activity of the most active compounds correlated
between the in vitro inhibition of PaPBP3 and the measured MIC values, especially against
the wild-type P. aeruginosa PAO1 strain.
.
2.5. Effect of the Pyrrolidine-2,3-diones on Eukaryotic Cells
We finally tested the cytotoxicity of the compounds against four eukaryotic cell lines
(Table 2 and Table S5), NRK-52E, an epithelial kidney cell line of Rattus norvegicus; Hep
G2, an human epithelial liver cell line; and two suspension cell lines: CCRF-CEM, a T
lymphoblast; and H9, a cutaneous T lymphocyte cell line. Most PaPBP3 inhibitors were
found not to be cytotoxic toward the tested concentrations, except for compounds 36, 37
and 45 that showed high cytotoxic potential at the tested cell lines, which is most relevant
for compound 37, with an MIC activity that is a similar concentration.
3. Discussion
β-lactam antibiotics have served their purpose for decades as a reliable and potent
class of antibacterial drugs. However, increasing -lactam resistance limits current treat-
β
ment options severely, especially against infections caused by Gram-negative pathogens,
such as P. aeruginosa. Many efforts have been directed towards the research and develop-

Antibiotics 2021, 10, 529
10 of 18
ment of improved
β-lactams or combinations of known
β-lactams with novel
β-lactamase
inhibitors. Interestingly, two novel PBP inhibitors (e.g., boronate acids and DBOs) that
inhibit the transpeptidation activity of PBPs and bypass -lactam resistance were recently
discovered and are currently in preclinical development [16 18 41 42]. Alternative ap-
proaches have been explored in the past to identify noncovalent PBP inhibitors. However,
the potency of the identified compounds was disappointingly low [25 32 43 46]. Nev-
β
,
, ,
,
, –
ertheless, targeting the essential TPases remains of great interest and discovering novel
entities capable of inhibiting these enzymes can contribute immensely to the fight against
the antimicrobial resistance crisis.
In this work, we have optimised a high-throughput assay for screening compound
libraries targeting PaPBP3 and used this assay to identify a novel scaffold class of small
molecules as PaPBP3 inhibitors. The absorbance S2d assay approach has been traditionally
used to investigate the TPase activities of several bacteria, but the assay has some draw-
backs that limit its use as a target-based screening and identification of PBP inhibitors.
Here, we provide a new fluorometric readout alternative that enabled us to eliminate
interferences previously reported with screening compounds. The assay was used to screen
a library of 2455 protease-targeted inhibitors. After eliminating compounds that did not
display dose-dependent target inhibition or were considered as weak inhibitors, a set of
nine hit compounds based on a pyrrolidine-2,3-dione core was identified. Fluorescence
polarization experiments indicated that these compounds are competitive inhibitors and
can displace Bocillin FL binding at the catalytic TPase site. A small-focused library of
pyrrolidine-2,3-dione-based derivates was designed and synthesised to investigate the SAR
of this cluster. Initial MIC testing against P. aeruginosa isolates in the presence of PMBN
showed antibacterial activity of this pyrrolidine-2,3-diones class with no obvious cytotoxic
effects. The results are therefore encouraging and provide an opportunity to explore this
chemical class for antibacterial drug discovery. However, without a clear understanding of
a distinct binding mode of the most active compounds at the active centre of PaPBP3, the
data presented here cannot rule out the possibility of a secondary off-target activity for this
compound class.
We generated an in-silico docking model (LeadIT version 2.3.2, BioSolveIT GmbH,
Sankt Augustin, Germany) illustrating the potential binding mode of compounds 34 and
35 to PaPBP3 (6HZR) [13]. Best-fit poses were defined using the LeadIT software when 50%
of the maximum allowed overlap volume and 0.5 intra-ligand clashed. These simulations
indicated that the essential hydroxyl group at position R2 interacts with the main-chain
atoms of Val333 located at the TPase site (Figure 4). The secondary amine of the indole
moiety and its connection to the pyrrolidone ring via a C2-linker in compounds 34 and
35 may facilitate an extra interaction with Thr487 that may explain the lower IC50 and
Kd values of these compounds. These docking models are supported by the fluorescence
polarisation experiments against Bocillin FL and strongly suggest that the compounds
interact with the TPase domain. However, a cocrystal structure with PaPBP3 would be
required to confirm this hypothesis and provide a binding mode for further optimisation of
the identified scaffold as potent PBP3 inhibitors. Moreover, further optimisation steps will
focus on improving the physicochemical properties of the molecules. For example, reducing
the hydrophobicity of the molecules may facilitate bacterial cell envelope penetration.
The work presented here underscores a novel class of PBP3 small molecule inhibitors,
which offer excellent starting points for further hit-to-lead optimisation and development.
The combination of promising PaPBP3 inhibition, initial antibacterial activities and safe
cytotoxic profile is highly encouraging. Further exploration of this chemical cluster will
focus on implementing the “new rules of permeation” to advance this class of molecules to
the periplasmic target for improved PBP3 inhibition and enforced antibacterial activity.

Antibiotics 2021, 10, 529
11 of 18
Figure 4. Docking model of pyrrolidine-2,3-dione compounds and PaPBP3 (PDB ID:6HZR) [13].
(A) Best-fit position of compound 34. (B) Best-fit position of compound 35. Dot lines indicate
hydrogen-bonds, green lines indicate potential hydrophobic interactions.
4. Materials and Methods
4.1. Chemicals
The SAR study of the pyrrolidine-2,3-dione cluster and the design of further analogons
was performed at AiCuris. Custom synthesis of analogues and the S2d artificial substrate
was subcontracted to Enamine Building Blocks Ltd. (Riga, LVA). The custom synthesised
compounds were delivered as powder in amber–glass bottles and purities higher than 98%.
Compound were stocked at 4 ◦C and 10 mM stock solutions in 96% DMSO were stored at
4 ◦C to avoid crystallisation.
Representative synthesis of pyrrolidine-2,3-diones [38,39]. A mixture of ethyl 4-(4-bromophenyl)-
2,4-dioxobutanoate (0.125 g, 1.0 mmol) and p-chlorobenzaldehyde (0.14 g, 1.0 mmol) in dry
dioxane (2 mL) was stirred for 30 min at 50 ◦C. Acetic acid (0.5 mL, 8.74 mmol) and m-fluoro-
◦
benzylamine (0.285 g, 1.0 mmol) were added. The resulting mixture was stirred for 2 h at 90 C.
After cooling the reaction mixture to r.t., the precipitate was filtered, washed with dioxane, and
dried in vacuum to give 0.208 g of target compound
MHz, DMSO-d6): 7.65 (br, 4H), 7.32 (br, 5H), 7.06-7.09 (m, 1H), 6.89–6.95 (m, 2H), 5.33 (s, 1H),
1
(0.415 mmol, 42% yield). 1H NMR (400
δ
4.75 (d, J = 12 Hz, 1H), 3.98 (d, J = 12 Hz, 1H); 13C NMR (100 MHz, DMSO-d6)
δ 188.3, 165.9,
163.2 (d, 1JCF = 139 Hz), 139.8, 139.8, 137.5, 135.4, 133.4, 133.3, 131.6, 131.1, 130.9, 130.9, 130.3,
129.0, 126.8, 124.1, 124.1, 119.2, 115.0, 114.8, 114.6, 114.5, 60.8, 44.2. APSI MS: M++1 503.0.
Analytical data of representative pyrroldine-2,3-diones. 4-(4-bromobenzoyl)-5-(4-chlorophenyl)-3-
hydroxy-1,5-dihydropyrrol-2-one (32). 1H NMR (400 MHz, DMSO-d6)
δ
9.43 (s, 1H), 7.73–7.51
(m, 4H), 7.43–7.26 (m, 4H), 5.46 (s, 1H), 3.17 (br s, 1H); 13C NMR (100 MHz, DMSO-d6):
δ
188.7, 167.4, 138.1, 137.6, 132.8, 131.7, 131.1, 129.6, 128.8, 126.8, 120.1, 56.6. HRMS (ESI): Calcl.
mass C17H11BrClNO3 390.9611 (M+), found 390.9609and 393.9663 (M++1).
4-(4-bromobenzoyl)-5-(4-chlorophenyl)-3-hydroxy-1-(1H-indol-2-ylmethyl)-5H-pyrrol-2-one (34). 1H
NMR (400 MHz, DMSO-d6):
6H), 7.02 (br, 2H), 6.16 (s, 1H), 5.30 (s, 1H), 5.05 (d, J = 15.2 Hz, 1H), 3.88 (d, J = 15.7 Hz, 1H);
13C NMR (100 MHz, DMSO-d6):
188.2, 165.5, 137.5, 136.8, 135.4, 135.4, 134.1, 133.3, 131.6,
δ 12.23 (s, 1H), 11.03 (s, 1H), 7.64-748 (br, 4H), 7.46–7.36 (br m,
δ

Antibiotics 2021, 10, 529
12 of 18
131.6, 131.1, 130.1, 130.1, 129.1, 128.2, 126.8, 121.5, 119.4, 111.6, 100.9, 60.3, 38.2. HRMS (ESI):
Calc. mass C26H18BrClN2O3 520.0189 [M+], found 520.0181 and 523.0234 (M++1).
1
4-(4-bromobenzoyl)-5-(4-chlorophenyl)-1-(furan-2-ylmethyl)-3-hydroxy-5H-pyrrol-2-one (41). H
NMR (400 MHz, DMSO-d6):
δ 7.78–7.49 (br, 5H), 7.48–7.18 (br, 4H), 6.31 (s, 1H), 6.19 (s,
1H), 5.25 (s, 1H), 4.81 (d, J = 16.0 Hz, 1H), 3.89 (d, J = 16.1 Hz, 1H); 13C NMR (100 MHz,
DMSO-d6):
δ 189.2, 165.3, 149.7, 143.3, 137.4, 135.2, 135.2, 133.2, 131.6, 131.6, 130.0, 129.0,
126.9, 119.3, 110.9, 109.1, 60.5, 37.4. HRMS (ESI): Calc. mass C22H15BrClNO4 470.9873 (M+),
found 470.9873 and 473.9923 (M++1).
4-(4-bromobenzoyl)-5-(4-chlorophenyl)-3-(ethylamino)-1-[(3-fluorophenyl)methyl]-5H-pyrrol-2-one
1
(
43). H NMR (400 MHz, DMSO-d6):
δ
8.93 (br s, 1H), 7.45 (d, J = 8.0 Hz, 2H), 7.23–7.35 (m,
3H), 7.12 (d, J = 8.0 Hz, 1H), 7.06 (t, J = 8.0 Hz, 2H), 6.72–6.96 (m, 4H), 5.50 (s, 1H), 4.64 (d, J
= 15.6 Hz, 1H), 3.83-3.88 (overlapping d with br s, 3H) 3.83-3.63 (br, 1H), 1.19 (t, J = 7.1 Hz,
3H); 13C NMR (100 MHz, DMSO-d6):
δ
189.0, 162.6, 162.5 (d, 1JCF = 194 Hz), 139.9, 139.8,
139.8, 136.3, 132.8, 131.5, 130.9, 130.8, 130.2, 130.1, 129.5, 128.6, 124.2, 124.2, 115.0, 114.8,
114.6, 114.5, 110.8, 62.0, 44.1. HRMS (ESI): Calc. mass C26H21BrClFN2O2 526.0459 (M+),
found 526.0453 and 529.0507 (M++1).
4-(4-bromobenzoyl)-5-(4-chlorophenyl)-1-[(3-fluorophenyl)methyl]-3-(morpholin-4-yl)-5H-pyrrol-
1
2-one (44). H NMR (400 MHz, DMSO-d6):
δ 7.65 (d, J = 8.5 Hz, 2H), 7.55 (d, J = 8.4 Hz,
2H), 7.30 (q, J = 7.4 Hz, 1H), 7.23 (d, J = 8.4 Hz, 2H), 7.15 (d, J = 8.2 Hz, 2H), 7.06 (t, J = 8.3
Hz, 1H), 6.80 -6.92 (m, 2H), 5.25 (s, 1H), 4.64 (d, J = 15.4 Hz, 1H), 3.99 (d, J = 15.5 Hz, 1H),
3.46-3.52 (m, 6H), 2.93-3.02 (m, 2H); 13C NMR (100 MHz, DMSO-d6):
δ 189.9, 166.6, 162.5
(d, 1JCF = 197 Hz), 144.6, 138.6, 138.5, 138.3, 134.4, 134.1, 131.9, 130.4, 130.3, 129.7, 129.1,
129.0, 127.7, 124.1, 124.0, 118.6, 115.4 (d, 2JCF = 17 Hz), 114.9 (d, 2JCF = 17 Hz), 66.8, 61.9,
50.0, 44.2, 44.1. HRMS (ESI): Calc. mass C28H23BrClFN2O3 568.0565 (M+), found 568.0557
and 571.0610 (M++1).
4.2. Protein Expression and Purification
The soluble wild-type P. aeruginosa PBP3 (amino acids 40 to 563) was overexpressed using
the plasmids pET41_PBP-3. The plasmid was transformed and expressed in the BL21 (DE3)
Rosetta strain. The expression strain was inoculated from glycerol stock in two sequential
LB medium precultures supplemented with kanamycin (50
µ
g/mL) and chloramphenicol
◦
(30
µg/mL) and incubated overnight at 37 C and 160 rpm. Expression cultures were
inoculated to OD578 0.1 and were grown to the exponential phase in baffled flasks to improve
aeration at 37 ◦C and 160 rpm. Induction of expression was done between OD578 0.3 to 0.6
by the addiction of 1 mM isopropyl-
period of 20 h at 21 ◦C and 120 rpm. The cultures were then harvested (5000
15 min), washed with 50 mL of water and stored over night at
80 ◦C. Lysis of cell pellets
β
-D- thiogalactopyranoside (IPTG) and an incubation
×
g, 4 ◦C, and
−
was done via the Qproteome Bacterial Lysis Buffer (Qiagen, Hilden, Germany), following the
manufacture’s protocol. Purification of the His-tagged PBP3 was done via immobilised metal
affinity chromatography (IMAC) using the ÄKTA chromatography system (GE Heathcare,
Freiburg, Germany) with HisTrap HP column (CV:2 mL). Washing and elution of the protein
was done by imidazole gradients with a mix of IMAC Buffer A (1x Sorensen Buffer, 200 mM
NaCl, 0.5 mM DDM, 10% glycerol, pH 7.5), and IMAC Buffer B (1
mM Na2HPO4 and 10 mM KH2PO4), 200 mM NaCl, 0.5 mM DDM, 10% glycerol, 300 mM
Imidazole, pH 7.5) in 1 mL fractions up to 20 CV. The eluted protein solution was dialysed
in 2 L of dialysis buffer (1 Sorensen Buffer, 200 mM, 20% glycerol, 0.5 mM DDM, pH 7.5).
×
Sorensen Buffer (57
×
×
Protein concentrations were determined via photometric measurement.
4.3. IC50 Determination by S2dfluo
The activity of soluble PaPBP3 (300 nM) was measured in the presence of 0.8 mM
S2d, 1 mM D-alanine and 0.5 mM mBBr for 25 min by monitoring the increase in the
fluorescence intensity of mBBr at Ex/Em: 394/490 nm using an Infinite M1000 microplate
reader (Tecan, Männedorf, Switzerland). To test the effect of inhibitors, 100
µM or 0.8
to 100 M of compounds diluted in 96% DMSO were incubated for 30 min and 450 rpm
µ

Antibiotics 2021, 10, 529
13 of 18
with 300 nM soluble PBP3. The incubation period was set to improve the binding of
non-covalent inhibitors and 0.01% Triton X100 (Sigma-Aldrich, Taufkirchen, Germany) was
added to prevent aggregation of compounds. The enzyme reaction was started by adding
a substrate mix of 0.8 mM S2d and 0.05 mM mBBr. The enzyme, compound and substrate
mix were mixed 1:1, pipetted with a Liquidator 96 (Melter Toledo, Gießen, Germany) and
the fluorescence intensity was directly monitored with an Infinite M1000 microplate reader
(Tecan) at Ex/Em: 394/490 nm. The assay was performed in black, flat-bottom 96-well
microplates (Greiner Bio-One, Kremsmünster, Austria) at 37 ◦C in Sorensen Phosphate
Buffer (57 mM Na2HPO4 and 10 mM KH2PO4, pH 7.5 + 0.01% Triton X-100). The velocity
of the PBP3 reaction was calculated in the presence of each concentration of compound
and the velocity values were used to estimate IC50 by sigmoidal nonlinear regression,
using Prism v.9 software (San Diego, CA, USA) and the velocity of the reaction at each
concentration.
4.4. IC50 Determination by the Bocillin FL Biochemical Assay
Fluorescence polarisation (FP) experiments were carried out as described before [32,35,36,47].
To determine the binding of Bocillin FL to PBP, 200 nM of PaPBP3 was incubated for 30 min
with 100 nM of Bocillin FL (Thermo Fisher Scientific, Waltham, MA, USA). The fluorescence
polarisation signal of free Bocillin FL was monitored in a Pherastar microplate reader
(BMG Labtech, Ortenberg, Germany) in the polarisation mode with Ex/Em: 485/520
nm excitation and emission filters. Millipolarisation (mP) units were calculated using
the MARS data analysis software v30.40.R2 (BMG Labtech, Ortenberg, Germany). All
the experiments were done in Sorensen buffer at 37 ◦C using flat-bottom, black 96-well
plates. To determine the IC50 values of the potential Bocillin FL competitors, 5
µL of serial
compound dilutions (100 M to 0.8 M compound) in 96% DMSO were incubated with
µ
µ
PBP3 of P. aeruginosa for 30 min to promote binding of the compounds. Bocillin FL was
diluted in Sorensen buffer and added to the reaction and directly measured for 30 min
at 37 ◦C. The FP was continuously measured in the presence and absence of compound
and enzyme. The differences between the mP values at 0 min and 17 min were used as
the parameter to estimate the IC50 values. The estimated IC50 values were established by
sigmoidal nonlinear regression, using Prims v.8 software.
4.5. SPR Experiments
◦
SPR experiments were performed at 25 C by using a Biacore T200 instrument from Cytiva
(Marlborough, MA, USA). Purified PaPBP3 was coated on a CM5 chip, that was activated
with EDC/NHS (1-ethyl-3-(3-dimethylaminopropyl) carbodiimide/N-hydroxysuccinimide)
beforehand and blocked with ethanolamine after protein attachment. We achieved a protein
amount of 11.820 RU. For every analyte, two concentration series between 3.25
µM and 200 µM
were prepared in the running buffer (13 mM NaCl, 257 M KCl, 171 M KH2PO4, 950 µM
µ
µ
Na2HPO4, 0.05% Tween-20 and 5% DMSO, pH 7.4) in a 96 well plate. Prior to the injection, the
baseline was determined using running buffer for 300 s. Then, the analytes were then associated
for 100 s and dissociated for 300 s with a flow rate of 30 µL/min. A 30 s washing step with
50% DMSO was used after every injection to regenerate the chip. In the final step, ten different
DMSO concentrations ranging from 4.0% to 5.8% were used as the internal solvent correction.
Steady state analysis was performed by using the internal Biacore software and figures were
prepared with Prism v.8 using the points of one concentration series, respectively.
4.6. Bacterial Strains
The wild-type P.aeruginosa PAO1 and the isogenic P. aeruginosa K2896 ∆mexB∆mexXY∆mex
CD-oprJ isolates were used to determine the MIC in the presence and absence of PMBN [48].
4.7. MIC Determination
MIC determination was performed via standard broth microdilution according to the
Clinical & Laboratory Standards institute guidelines [49,50]. Overnight colonies grown

Antibiotics 2021, 10, 529
14 of 18
on BD-Columbia agar with 5% Sheep Blood (Becton Dickinson, Franklin Lakes, NJ, USA),
were diluted in Mueller Hinton Broth 2 cation adjusted (MH+) medium (Thermo Fisher
Scientific) to a concentration of 5
×
105 CFU/mL. Serial dilutions (1:2) of test compounds
◦
were incubated at 37 C with 5
×
104 bacteria in a final volume of 100
µL. Sterile and growth
controls in the presence of 1% DMSO and PMBN if the case, and the positive controls,
aztreonam (AZT) and tigecycline (TIG) were included in the experiments (Table S4). The
MICs were determined visually after 18 h at 37 ◦C and confirmed by the addition of 10%
PrestoBlue (ThermoFisher Scientific) reagent according to the manufacture’s protocol. The
fluorescence signal of PrestoBlue reagent was measured at Ex/Em: 550/595 nm after 30 min
incubation at 37 ◦C using at a ClarioStar microplate reader (BMG Labtech). PrestoBlue
reduction occurs in the presence of living cells and consequently turns red in colour and is
highly fluorescent.
4.8. Cytotoxicity Assays
◦ ◦
120 C were thawed to 37 C and resuspended in 10 mL of
Cell lines stored at
−
medium pre-warmed to 37 ◦C. Table S6 summarises the culture media and conditions for
each cell line. Cells were pelleted at 300
×
g for 10 min, and then resuspended in fresh
◦
medium, seeded in a T-75 culture flask, and placed in a CO2 incubator (37 C with 5%
CO2 tension). Once the cells reached approximately 70–80% confluency, they were then
passaged regularly. The adherent cell lines NRK52E, HepG2 and Huh 7 were washed
once with 10 mL DPBS and detached with Trypsin-EDTA for 5 to 7 min and resuspended
in 15 mL of the appropriate medium. The suspension cell line CCRF-CEM were diluted
and resuspended in fresh RPMI 1640 PAN medium. To calculate the cell density, the cell
resuspensions were diluted 1:6 with DPBS. Equal volumes of cell suspension and Trypan
Blue were mixed and gently loaded into a haemocytometer (INCYTO, 31056 Republic of
Korea). In each experiment, cell concentrations were adjusted to reach a concentration
of 2.5
CCRF-CEM and 3
compounds (1
L in 96% DMSO) for 72 h at 37 ◦C with 5% CO2 tension. Cycloheximide,
×
103 cell/well for NRK52E, 5
×
103 cell/well for HepG2, 6.2
×
104 cell/well for
×
103 for H9. The cells were incubated with decreasing concentration of
µ
a cytotoxic compound that interferes with the translocation step in eukaryotic protein
synthesis was used as a control for the experiments. The assay was carried out in flat-
bottom, transparent 96-well plates. Cell viability was determined using the PrestoBlue
reagent (Thermo Fisher Scientific, Waltham, MA, USA) according to the manufacture’s
protocol. First, 10
×
PrestoBlue was added to each well to obtain a 1
×
solution, and
incubated at 37 ◦C for 45 min. Then, fluorescence was measured at Ex/Em: 550/595 nm
using a ClarioStar microplate reader (BMG Labtech). Cytotoxicity was determined as the
percentage of the difference in absorbance of the untreated sample to the treated sample.
CC50 was calculated using the values obtained in each concentration using the sigmoidal
dose-response by Prism v.8. The media used in each cell line is detailed in Table S6.
Supplementary Materials: The following are available online at https://www.mdpi.com/article/10
.3390/antibiotics10050529/s1, Table S1. Additional HTS statistical parameters; Table S2. Pyrrolidine-
2,3-dione screening hits without concentration-dependent inhibition; Table S3. Minimal Inhibitory
Concentrations of pyrrolidine-2,3-dione against wild-type strains; Table S4. MIC of control com-
pounds against selected strains Table S5. Detailed cytotoxicity of pyrrolidine-2,3-dione against several
cell lines; Table S6. Eukaryotic cell and culture media; Figure S1. NMR spectra of representative
pyrrolidine-2,3-diones.
Author Contributions: Conceptualisation, A.L.-P. and C.V.; methodology, A.L.-P., S.F., I.G., O.W. and
M.P.L.-A.; formal analysis, A.L.-P., S.F. and K.K.; investigation, C.V., A.L.-P. and K.K.; resources, H.W.
and C.V.; data curation, A.L.-P. and S.F.; writing—original draft preparation, A.L.-P., S.F. and K.K.;
writing—review and editing, A.L.-P., S.F., I.G., O.W., W.V., H.W., A.J.D., K.K. and C.V.; visualisation,
A.L.-P. and K.K.; supervision, C.V., A.J.D., O.W. and W.V.; project administration, C.V; funding
acquisition, C.V. All authors have read and agreed to the published version of the manuscript.

Antibiotics 2021, 10, 529
15 of 18
Funding: This project has received funding from the European Union’s Horizon 2020 Research and
Innovation Programme Train2Target under grant agreement No 721484 and Anti-Infective Cures GmbH.
Data Availability Statement: The data presented in this study are available on request from the
corresponding author.
Acknowledgments: We acknowledge Ute Sauer, Nabila Anwari, Gisela Bukow and Marc Svenich
for the technical contribution to this paper.
Conflicts of Interest: The funders had no role in the design of the study, in the collection, analyses,
or interpretation of data, in the writing of the manuscript, or in the decision to publish the results.
Appendix A
Figure A1. Biochemical characterisation compound 34 as PaPBP3 inhibitor. (
represent the PaPBP3 kinetic curves in the presence of decreasing concentration of compounds. The right panel shows the
IC50 normalised. ( ) Fluorescence Polarisation assay. The left panel represents the raw bocillin FL fluorescence polarisation
traces in the presence of decreasing concentration of compounds and the right panel the calculated IC50. ( ) SPR assay. The
left panel shows the sensorgrams signals in the presence of compound and the right panel the Kd determination.
A) Fluorescence
−
S2d based assay, the left panel
B
C

Antibiotics 2021, 10, 529
16 of 18
Figure A2. Biochemical characterisation compound 35 as PaPBP3 inhibitor
panel represent the PaPBP3 kinetic curves in the presence of decreasing concentration of compounds. The right panel
shows the IC50 normalised. ( ) Fluorescence Polarisation assay. The left panel represents the raw bocillin FL fluorescence
polarisation traces in the presence of decreasing concentration of compounds and the right panel the calculated IC50. ( ) SPR
assay. The left panel shows the sensorgrams signals in the presence of compound and the right panel the Kd determination.
.
(A
) Fluorescence
−
S2d based assay, the left
B
C
References
1.
2.
3.
Sandri, A.; Lleo, M.M.; Signoretto, C.; Boaretti, M.; Boschi, F. Protease inhibitors elicit anti-inflammatory effects in CF mice with
Pseudomonas aeruginosa acute lung infection. Clin. Exp. Immunol. 2021, 203, 87–95. [CrossRef]
Vonberg, R.; Gastmeier, P. Isolation of Infectious Cystic Fibrosis Patients: Results of a Systematic Review. Infect. Control Hosp.
Epidemiol. 2005, 26, 401–409. [CrossRef]
Malhotra, S.; Hayes, D.; Wozniak, D. Cystic Fibrosis and Pseudomonas aeruginosa: The Host-Microbe Interface. Clin. Microbiol. Rev.
2019, 32, e00138-00118. [CrossRef]

Antibiotics 2021, 10, 529
17 of 18
4.
5.
Klockgether, J.; Cramer, N.; Wiehlmann, L.; Davenport, C.; Tümmler, B. Pseudomonas aeruginosa Genomic Structure and Diversity.
Front. Microbiol. 2011, 2. [CrossRef]
Tacconelli, E.; Carrara, E.; Savoldi, A.; Harbarth, S.; Mendelson, M.; Monnet, D.; Pulcini, C.; Kahlmeter, G.; Kluytmans, J.; Carmeli,
Y.; et al. Discovery, research, and development of new antibiotics: The WHO priority list of antibiotic-resistant bacteria and
tuberculosis. Lancet Infect. Dis. 2018, 18, 318–327. [CrossRef]
6.
7.
8.
9.
Egan, A.; Errington, J.; Vollmer, W. Regulation of peptidoglycan synthesis and remodelling. Nat. Rev. Microbiol. 2020, 18, 446–460.
[CrossRef] [PubMed]
Vollmer, W.; Joris, B.; Charlier, P.; Foster, S. Bacterial peptidoglycan (murein) hydrolases. Fems Microbiol. Rev. 2008, 32, 259–286.
[CrossRef] [PubMed]
Vollmer, W.; Blanot, D.; de Pedro, M.A. Peptidoglycan structure and architecture. Fems Microbiol. Rev. 2008, 32, 149–167.
[CrossRef]
Taguchi, A.; Welsh, M.A.; Marmont, L.S.; Lee, W.; Sjodt, M.; Kruse, A.C.; Kahne, D.; Bernhardt, T.G.; Walker, S. FtsW is a
peptidoglycan polymerase that is functional only in complex with its cognate penicillin-binding protein. Nat Microbiol 2019, 4,
587–594. [CrossRef] [PubMed]
10. Meeske, A.J.; Riley, E.P.; Robins, W.P.; Uehara, T.; Mekalanos, J.J.; Kahne, D.; Walker, S.; Kruse, A.C.; Bernhardt, T.G.; Rudner, D.Z.
SEDS proteins are a widespread family of bacterial cell wall polymerases. Nature 2016, 537, 634–638. [CrossRef]
11. Sauvage, E.; Kerff, F.; Terrak, M.; Ayala, J.A.; Charlier, P. The penicillin-binding proteins: Structure and role in peptidoglycan
biosynthesis. FEMS Microbiol. Rev. 2008, 32, 234–258. [CrossRef]
12. Leclercq, S.; Derouaux, A.; Olatunji, S.; Fraipont, C.; Egan, A.J.; Vollmer, W.; Breukink, E.; Terrak, M. Interplay between
Penicillin-binding proteins and SEDS proteins promotes bacterial cell wall synthesis. Sci. Rep. 2017, 7, 43306. [CrossRef]
[PubMed]
13. Bellini, D.; Koekemoer, L.; Newman, H.; Dowson, C.G. Novel and Improved Crystal Structures of H. influenzae, E. coli and
P. aeruginosa Penicillin-Binding Protein 3 (PBP3) and N. gonorrhoeae PBP2: Toward a Better Understanding of beta-Lactam
Target-Mediated Resistance. J. Mol. Biol. 2019, 431, 3501–3519. [CrossRef] [PubMed]
14. Rolo, J.; Worning, P.; Boye Nielsen, J.; Sobral, R.; Bowden, R.; Bouchami, O.; Damborg, P.; Guardabassi, L.; Perreten, V.; Westh, H.;
et al. Evidence for the evolutionary steps leading to mecA-mediated
e1006674. [CrossRef] [PubMed]
β-lactam resistance in staphylococci. PLoS Genet. 2017, 13,
15. Katayama, Y.; Ito, T.; Hiramatsu, K. A New Class of Genetic Element, Staphylococcus Cassette Chromosome mec, Encodes
Methicillin Resistance in Staphylococcus aureus. Antimicrob. Agents Chemother. 2000, 44, 1549–1555. [CrossRef] [PubMed]
16. Rajavel, M.; Kumar, V.; Nguyen, H.; Wyatt, J.; Marshall, S.H.; Papp-Wallace, K.M.; Deshpande, P.; Bhavsar, S.; Yeole, R.; Bhagwat,
S.; et al. Structural Characterization of Diazabicyclooctane beta-Lactam “Enhancers” in Complex with Penicillin-Binding Proteins
PBP2 and PBP3 of Pseudomonas aeruginosa. mBio 2021, 12. [CrossRef]
17. Papp-Wallace, K.; Bonomo, R. New
[PubMed]
β-Lactamase Inhibitors in the Clinic. Infect. Dis. Clin. North Am. 2016, 30, 441–464. [CrossRef]
18. Zervosen, A.; Herman, R.; Kerff, F.; Herman, A.; Bouillez, A.; Prati, F.; Pratt, R.; Frère, J.M.; Joris, B.; Luxen, A.; et al. Unexpected
Tricovalent Binding Mode of Boronic Acids within the Active Site of a Penicillin-Binding Protein. J. Am. Chem. Soc. 2011, 133,
10839–10848. [CrossRef]
19. Thomas, B.; Wang, Y.; Stein, R. Kinetic and mechanistic studies of penicillin-binding protein 2x from Streptococcus Pneumoniae.
Biochem. 2001, 40, 15811–15823. [CrossRef]
20. Jamin, M.; Adam, M.; Damblon, C.; Christiaens, L. Accumulation of acyl-enzyme in DD-peptidase-catalysed reactions with
analogues of peptide substrates. Biochem. J. 1991, 280, 499–506. [CrossRef]
21. Adam, M.; Damblon, C.; Plaitin, B.; Christiaens, L.; Frère, J.M. Chromogenic depsipeptide substrates for beta-lactamases and
penicillin-sensitive DD-peptidases. Biochem. J. 1991, 270, 525–529. [CrossRef] [PubMed]
22. Adam, M.; Damblon, C.; Jamin, M.; Zorzi, W.; Dusart, V.; Galleni, M.; el Kharroubi, A.; Piras, G.; Spratt, B.G.; Keck, W.
Acyltransferase activities of the high-molecular-mass essential penicillin-binding proteins. Biochem. J. 1991, 279. [CrossRef]
23. Pratt, R.; Govardhan, C. Beta-Lactamase-catalyzed hydrolysis of acyclic depsipeptides and acyl transfer to specific amino acid
acceptors. Proc. Natl. Acad. Sci. USA 1984, 81, 1302–1306. [CrossRef]
24. Terrak, M.; Ghosh, T.; Van Heijenoort, J.; Van Beeumen, J.; Lampilas, M.; Aszodi, J.; Ayala, J.; Ghuysen, J.; Nguyen-Distèche, M.
The catalytic, glycosyl transferase and acyl transferase modules of the cell wall peptidoglycan-polymerizing penicillin-binding
protein 1b of Escherichia coli. Mol. Microbiol. 1999, 34, 350–364. [CrossRef] [PubMed]
25. Zervosen, A.; Lu, W.; Chen, Z.; White, R.; Demuth, T.; Frère, J.M. Interactions between Penicillin-Binding Proteins (PBPs) and
Two Novel Classes of PBP Inhibitors, Arylalkylidene Rhodanines and Arylalkylidene Iminothiazolidin-4-ones. Antimicrob. Agents
Chemother. 2004, 48, 961. [CrossRef]
26. Adediran, S.A.; Sarkar, K.S.; Pratt, R.F. Kinetic Evidence for a Second Ligand Binding Site on Streptococcus pneumoniae Penicillin-
Binding Protein 2x. Biochemistry 2018, 57, 1758–1766. [CrossRef] [PubMed]
27. Zhao, G.; Yeh, W.K.; Carnahan, R.H.; Flokowitsch, J.; Meier, T.I.; Alborn, W.E.; Becker, G.W.; Jaskunas, S.R. Biochemical characterization
of penicillin-resistant and -sensitive penicillin-binding protein 2x transpeptidase activities of Streptococcus pneumoniae and mechanistic
implications in bacterial resistance to beta-lactam antibiotics. J. Bacteriol. 1997, 179, 4901. [CrossRef] [PubMed]

Antibiotics 2021, 10, 529
18 of 18
28. Zervosen, A.; Zapun, A.; Frère, J.M. Inhibition of Streptococcus pneumoniae Penicillin-Binding Protein 2x and Actinomadura; R39
DD-Peptidase Activities by Ceftaroline. Antimicrob. Agents Chemother. 2013, 57, 661. [CrossRef]
29. Zhang, J.; Chung, T.; Oldenburg, K. A Simple Statistical Parameter for Use in Evaluation and Validation of High Throughput
Screening Assays. J. Biomol. Screen. 1999, 4, 67–73. [CrossRef]
30. Feng, B.Y.; Shoichet, B.K. A detergent-based assay for the detection of promiscuous inhibitors. Nat. Protoc. 2006, 1, 550–553.
[CrossRef]
31. Shoichet, B.K. Screening in a spirit haunted world. Drug Discov. Today 2006, 11, 607–615. [CrossRef]
32. Fedarovich, A.; Djordjevic, K.A.; Swanson, S.M.; Peterson, Y.K.; Nicholas, R.A.; Davies, C. High-throughput screening for novel
inhibitors of Neisseria gonorrhoeae penicillin-binding protein 2. PLoS ONE 2012, 7, e44918. [CrossRef]
33. Knoth, T.; Warburg, K.; Katzka, C.; Rai, A.; Wolf, A.; Brockmeyer, A.; Janning, P.; Reubold, T.; Eschenburg, S.; Manstein, D.; et al.
The Ras Pathway Modulator Melophlin A Targets Dynamins. Angew. Chem. Int. Ed. 2009, 48, 7240–7245. [CrossRef] [PubMed]
34. Park, M.; Sutherland, J.B.; Rafii, F. beta-Lactam resistance development affects binding of penicillin-binding proteins (PBPs) of
Clostridium perfringens to the fluorescent penicillin, Bocillin FL. Anaerobe 2020, 62, 102179. [CrossRef]
35. Shapiro, A.B.; Gu, R.F.; Gao, N.; Livchak, S.; Thresher, J. Continuous fluorescence anisotropy-based assay of Bocillin FL penicillin
reaction with penicillin binding protein 3. Anal. Biochem. 2013, 439, 37–43. [CrossRef] [PubMed]
36. Zhao, G.; Meier, T.; Kahl, S.; Gee, K.; Blaszczak, L. BOCILLIN FL, a sensitive and commercially available reagent for detection of
penicillin-binding proteins. Antimicrob Agents Chemother 1999, 43, 1124–1128. [CrossRef]
37. Zhang, D.; Lin, Y.; Chen, X.; Zhao, W.; Chen, D.; Gao, M.; Wang, Q.; Wang, B.; Huang, H.; Lu, Y.; et al. Docking- and
pharmacophore-based virtual screening for the identification of novel Mycobacterium tuberculosis protein tyrosine phosphatase B
(MptpB) inhibitor with a thiobarbiturate scaffold. Bioorg Chem 2019, 85, 229–239. [CrossRef]
38. Rose, R.; Erdmann, S.; Bovens, S.; Wolf, A.; Rose, M.; Hennig, S.; Waldmann, H.; Ottmann, C. Identification and structure of
small-molecule stabilizers of 14-3-3 protein-protein interactions. Angew. Chem. Int. Ed Engl. 2010, 49, 4129–4132. [CrossRef]
39. Richter, A.; Rose, R.; Hedberg, C.; Waldmann, H.; Ottmann, C. An optimised small-molecule stabiliser of the 14-3-3-PMA2
protein-protein interaction. Chemistry Eur. J. 2012, 18, 6520–6527. [CrossRef] [PubMed]
40. Ryabukhin, S.; Panov, D.; Plaskon, A.; Grygorenko, O. Approach to the Library of 3-Hydroxy-1,5-dihydro-2H-pyrrol-2-ones
through a Three-Component Condensation. ACS Comb. Sci. 2012, 14, 631–635. [CrossRef]
41. Moya, B.; Barcelo, I.; Bhagwat, S.; Patel, M.; Bou, G.; Papp-Wallace, K.; Bonomo, R.; Oliver, A. WCK 5107 (Zidebactam) and WCK
5153 Are Novel Inhibitors of PBP2 Showing Potent “
[CrossRef]
β-Lactam Enhancer” Antimicrob. Agents Chemother. 2017, 61, e02529-02516.
42. Papp-Wallace, K.M.; Nguyen, N.; Jacobs, M.; Bethel, C.; Barnes, M.; Kumar, V.; Bajaksouzian, S.; Rudin, S.; Rather, P.; Bhavsar,
S.; et al. Strategic Approaches to Overcome Resistance against Gram-Negative Pathogens Using -Lactamase Inhibitors and
-Lactam Enhancers: Activity of Three Novel Diazabicyclooctanes WCK 5153, Zidebactam (WCK 5107), and WCK 4234. J. Med.
Chem. 2018, 61, 4067–4086. [CrossRef]
β
β
43. Turk, S.; Verlaine, O.; Gerards, T.; Zivec, M.; Humljan, J.; Sosic, I.; Amoroso, A.; Zervosen, A.; Luxen, A.; Joris, B.; et al. New
noncovalent inhibitors of penicillin-binding proteins from penicillin-resistant bacteria. PLoS ONE 2011, 6, e19418. [CrossRef]
44. Sosicˇ, I.; Turk, S.; Sinreih, M.; Trošt, N.; Verlaine, O.; Amoroso, A.; Zervosen, A.; Luxen, A.; Joris, B.; Gobec, S. Exploration of the
chemical space of novel naphthalene-sulfonamide and anthranilic Acid-based inhibitors of penicillin-binding proteins. Acta Chim.
Slov. 2012, 59, 280–388. [PubMed]
45. Spink, E.; Ding, D.; Peng, Z.; Boudreau, M.; Leemans, E.; Lastochkin, E.; Song, W.; Lichtenwalter, K.; O’Daniel, P.; Testero, S.; et al.
Structure–Activity Relationship for the Oxadiazole Class of Antibiotics. J. Med. Chem. 2015, 58, 1380–1389. [CrossRef]
46. Toney, J.; Hammond, G.; Leiting, B.; Pryor, K.; Wu, J.; Cuca, G.; Pompliano, D. Soluble Penicillin-Binding Protein 2a: β-Lactam
Binding and Inhibition by Non-β-Lactams Using a 96-Well Format. Anal. Biochem. 1998, 255, 113–119. [CrossRef]
47. Shapiro, A.; Gao, N.; Gu, R.; Thresher, J. Fluorescence anisotropy-based measurement of Pseudomonas aeruginosa penicillin-binding
protein 2 transpeptidase inhibitor acylation rate constants. Anal. Biochem. 2014, 463, 15–22. [CrossRef] [PubMed]
48. Fraud, S.; Campigotto, A.J.; Chen, Z.; Poole, K. MexCD-OprJ Multidrug Efflux System of Pseudomonas aeruginosa: Involvement in
Chlorhexidine Resistance and Induction by Membrane-Damaging Agents Dependent upon the AlgU Stress Response Sigma
Factor. Antimicrob. Agents Chemother. 2008, 52, 4478. [CrossRef] [PubMed]
49. CLSI. Performance Standars for Antimicrobial Susceptibility Testing. In Clinical & Laboratory Standards Institute Guideline M100;
CLSI: Wayne, PA, USA, 2018; Volume 28.
50. CSLI. Methods for Dilution Antimicrobial Susceptibility Tests for Bacteria That Grow Aerobically. Approved Standard. In Clinical
and Laboratory Standards Institute M07-A9; CLSI: Wayne, PA, USA, 2012.