Discovery and biophysical characterization of 2-amino-oxadiazoles as novel antagonists of PqsR, an important regulator of Pseudomonas aeruginosa virulence

Article abstract of DOI:10.1021/jm400830r

The human pathogen Pseudomonas aeruginosa employs alkyl quinolones for cell-to-cell communication. The Pseudomonas quinolone signal (PQS) regulates various virulence factors via interaction with the transcriptional regulator PqsR. Therefore, we consider t

Full text of DOI:10.1021/jm400830r

Subscriber access provided by University of Glasgow Library
Article
Discovery and Biophysical Characterization of 2-Amino-
Oxadiazoles as Novel Antagonists of PqsR, an Important
Regulator of Pseudomonas aeruginosa Virulence
Michael Zender, Tobias Klein, Claudia Henn, Benjamin Kirsch, Christine K. Maurer,
Dagmar Kail, Christiane Ritter, Olan Dolezal, Anke Steinbach, and Rolf W. Hartmann
J. Med. Chem., Just Accepted Manuscript • DOI: 10.1021/jm400830r • Publication Date (Web): 06 Aug 2013
Downloaded from http://pubs.acs.org on August 9, 2013
Just Accepted
“Just Accepted” manuscripts have been peer-reviewed and accepted for publication. They are posted
online prior to technical editing, formatting for publication and author proofing. The American Chemical
Society provides “Just Accepted” as a free service to the research community to expedite the
dissemination of scientific material as soon as possible after acceptance. “Just Accepted” manuscripts
appear in full in PDF format accompanied by an HTML abstract. “Just Accepted” manuscripts have been
fully peer reviewed, but should not be considered the official version of record. They are accessible to all
readers and citable by the Digital Object Identifier (DOI®). “Just Accepted” is an optional service offered
to authors. Therefore, the “Just Accepted” Web site may not include all articles that will be published
in the journal. After a manuscript is technically edited and formatted, it will be removed from the “Just
Accepted” Web site and published as an ASAP article. Note that technical editing may introduce minor
changes to the manuscript text and/or graphics which could affect content, and all legal disclaimers
and ethical guidelines that apply to the journal pertain. ACS cannot be held responsible for errors
or consequences arising from the use of information contained in these “Just Accepted” manuscripts.
Journal of Medicinal Chemistry is published by the American Chemical Society. 1155
Sixteenth Street N.W., Washington, DC 20036
Published by American Chemical Society. Copyright © American Chemical Society.
However, no copyright claim is made to original U.S. Government works, or works
produced by employees of any Commonwealth realm Crown government in the course
of their duties.

Page 1 of 51
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
Discovery and Biophysical Characterization of 2ꢀ
AminoꢀOxadiazoles as Novel Antagonists of PqsR,
an Important Regulator of Pseudomonas aeruginosa
Virulence
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
AUTHOR NAMES
1
1,†
1,‡
1
1
Michael Zender, Tobias Klein, Claudia Henn, Benjamin Kirsch, Christine K. Maurer,
2
3
4
1
1,5*
Dagmar Kail, Christiane Ritter, Olan Dolezal, Anke Steinbach, Rolf W. Hartmann
AUTHOR ADDRESS
1
HelmholtzꢀInstitute for Pharmaceutical Research Saarland, Department of Drug Design and
Optimization, Campus C2.3, 66123 Saarbrücken, Germany
2
PharmBioTec GmbH, Campus C2.2, 66123 Saarbrücken, Germany
3
Helmholtz Centre for Infection Research, Department of Macromolecular Interactions,
Inhoffenstraße 7, 38124 Braunschweig, Germany
4
Commonwealth Scientific and Industrial Research Organization (CSIRO), Materials Science
and Engineering, 343 Royal Parade, Parkville 3052 Victoria, Australia
1
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 2 of 51
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
5
Pharmaceutical and Medicinal Chemistry, Saarland University, Campus C2.3, 66123
Saarbrücken, Germany
ABSTRACT
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
The human pathogen Pseudomonas aeruginosa employs alkyl quinolones for cellꢀtoꢀcell
communication. The Pseudomonas Quinolone Signal (PQS) regulates various virulence factors
via interaction with the transcriptional regulator PqsR. Therefore, we consider the development
of PqsR antagonists a novel strategy to limit the pathogenicity of P. aeruginosa. A fragment
identification approach using surface plasmon resonance screening led to the discovery of
chemically diverse PqsR ligands. The optimization of the most promising hit (5) resulted in the
oxadiazoleꢀ2ꢀamine 37 showing pure antagonistic activity in E. coli (EC50 7.5 ꢁM) and P.
aeruginosa (38.5 ꢁM) reporter gene assays. 37 was able to diminish the production of the PQS
precursor HHQ in a PqsH deficient P. aeruginosa mutant. The level of the major virulence factor
pyocyanin was significantly reduced in wildꢀtype P. aeruginosa. In addition, siteꢀdirected
mutagenesis in combination with isothermal titration calorimetry and NMR INPHARMA
experiments revealed that the identified ligands bind to the same site of PqsR by adopting
different binding modes. These findings will be utilized in a future fragment growing approach
aiming at novel therapeutic options for the treatment of P. aeruginosa infections.
2
ACS Paragon Plus Environment

Page 3 of 51
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
INTRODUCTION
Our view of bacterial lifestyle has changed drastically in the last few years. In the past, bacteria
were deemed to be autonomous unicellular organisms. Nowadays, it is increasingly recognized
that they have the ability to communicate with each other. A particular form of bacterial
communication is referred to as quorum sensing (QS). QS functions via the release of signaling
molecules. The concentration of these molecules increases as a function of cell density. After a
threshold concentration has been reached they regulate gene expression in bacteria by activation
of corresponding transcriptional regulators. QS cellꢀtoꢀcell communication enables bacteria to
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
1
,2
monitor population density and to orchestrate their behavior on a populationꢀwide scale.
Pseudomonas aeruginosa, a Gramꢀnegative opportunistic pathogenic bacterium is an important
3,4
causative agent of nosocomial infections and is involved in several acute and chronic
infections. It is one of the major pathogens causing pneumonia e.g. in people suffering from
5
cystic fibrosis. The ability of P. aeruginosa to form biofilms contributes to its low antibiotic
susceptibility mainly due to lowered metabolic activity and decreased growth rate of biofilm
6
bacteria. Moreover, during the last decades, P. aeruginosa strains resistant against nearly all
7
,8
established antibiotics have emerged, necessitating alternative therapeutic options.
P. aeruginosa coordinates group behaviours via three interconnected QS circuitries. The
9
,10
11,12
las and rhl
communication systems operate via Nꢀacylated homoserine lactones (AHLs),
which are prevalent signaling molecules in Gramꢀnegative bacteria. In contrast, the pqs
1
3,14
15
system
is restricted to particular Pseudomonas and Burkholderia strains. It functions via the
2
ꢀalkylꢀ4(1H)ꢀquinolones (HAQ) signal molecules PQS (Pseudomonas Quinolone Signal; 2ꢀ
3
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 4 of 51
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
heptylꢀ3ꢀhydroxyꢀ4(1H)ꢀquinolone) and its precursor HHQ (2ꢀheptylꢀ4(1H)ꢀquinolone; Figure
1).
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
Figure 1: Structure of the P. aeruginosa signaling molecules PQS and HHQ as well as recently
published PqsR ligands
1
6,17
Both interact with their receptor PqsR
to control many genes involved in the production of
1
8
various virulence factors like elastase, pyocyanin and lectins. Moreover, PQS enhances its own
16,19,20
14
biosynthetic pathway by positively regulating the pqsABCD(E)
and phnAB operons as
1
8,21
well as it influences biofilm formation.
provides novel options for inhibiting the virulence without interfering with bacterial viability.
The discovery of QS systems in pathogenic bacteria
2
2
23
The interference with QS systems by synthetic inhibitors and the reduction of virulence factors
2
4
by flucytosine resulted in diminished P. aeruginosa pathogenicity in mouse pulmonary
infection models demonstrating the inꢀvivo relevance of these antiꢀvirulence strategies.
Regarding P. aeruginosa, we consider PqsR ꢀ a key player for its pathogenicity ꢀ an attractive
drug target. PqsR, also known as multiple virulence factor regulator (MvfR), belongs to the
17
family of LysRꢀtype transcriptional regulators (LTTR). The Nꢀterminal region contains a helixꢀ
turnꢀhelix DNA binding motif and the Cꢀterminus between amino acids 92ꢀ298 encodes for a
1
7
predicted ligand binding domain (LBD).
4
ACS Paragon Plus Environment

Page 5 of 51
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
Recently, we have reported on the discovery of two different classes of antagonists which
2
5
target PqsR (Figure 1). The HHQ derivative 1 is a potent PqsR antagonist, but shows
insufficient drugꢀlike properties such as low aqueous solubility. Given that it is very difficult to
strongly improve the physicoꢀchemical profile of lead compounds with high logP values without
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
2
6
27,28
losing activity, a fragment screening was initiated as an alternative
initially leading to the
29
benzamide 2. 2 showed mixed agonisticꢀantagonistic properties. The introduction of a
2
9
hydroxamic acid motif (3) provided a pure antagonistic profile. It is noteworthy that
3
0
hydroxamic acids are reported as chelating agents and accordingly have been used as inhibitors
31
of various metalloproteins. Furthermore, aromatic hydroxamic acids are described as potential
3
2
mutagenic agents. In order to overcome these shortcomings and to gain chemical diversity, it
was of major interest to discover PqsR antagonist with alternative scaffolds.
In this work, we present the discovery of structurally diverse, low molecular weight PqsR
ligands employing surface plasmon resonance (SPR) biosensor experiments. Selected hits
displaying sufficient affinity were functionally and thermodynamically characterized.
Furthermore, the most promising hit compound 5 was structurally optimized and transformed
into the potent PqsR antagonist 37 which was able to reduce the production of HHQ in a PqsH
deficient mutant and the major virulence factor pyocyanin in wildꢀtype P. aeruginosa.
5
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 6 of 51
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
RESULTS AND DISCUSSION
Chemistry. The top screening hits 4ꢀ9 (Figure 2) were commercially available. The synthesis
of the thiadiazole and oxadiazole derivatives was achieved by the use of different synthetic
routes as described in the following.
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
a
Scheme 1. Synthesis of Oxadiazoleꢀ2ꢀamine and Thiadiazoleꢀ2ꢀamine Derivatives
a
Reagents and Conditions: (a) oxalyl chloride, DMF cat, DCM, reflux; (b) thiosemicarbazide,
THF, 0 °C to RT; (c) thiosemicarbazide, POCl , 80°C; (d) H SO , RT; (e) 1,3ꢀDibromoꢀ5,5ꢀ
dimethylhydantoin, KI, NaOH, 2ꢀpropanol, 4 °C.
3
2
4
The synthesis of 1,3,4ꢀoxadiazoleꢀ2ꢀamine and 1,3,4ꢀthiadiazoleꢀ2ꢀamine derivatives is shown
in Scheme 1. Thiadiazoles 5 and 33 were synthesized in a oneꢀstep procedure from commercially
33
available benzoic acid derivatives and hydrazinecarbothioamide in POCl . A classic coupling
3
6
ACS Paragon Plus Environment

Page 7 of 51
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
reaction of benzoic acid chlorides and hydrazinecarbothioamide afforded the
3
4
benzoylthiosemicarbazides 23-30.
31 was obtained by treating the respective
3
5
benzoylthiosemicarbazide 23 with concentrated sulfuric acid as dehydrating agent. The ring
closure which gives the respective 2ꢀaminoꢀ1,3,4ꢀoxadiazoles 32 and 35-42 was achieved under
mild oxidative conditions using compounds 23-30 and employing 1,3ꢀdibromoꢀ5,5ꢀ
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
3
4
dimethylhydantoin as primary oxidant in the presence of catalytic amounts of potassium iodide.
a
Scheme 2. Synthesis of 2ꢀSubstituted Oxadiazoles
N
O
O
N
H
N
OMe
O
O
b
c
N
H
h
h
O
CF3
CF3
CF3
44
49
N
N
H
N
CF3
CF3
N
H
O
O
CF3
50
45
O
N
N
O
NH2
R1
d
Cl
N
O
a
H
R1 = H 47
= Me 48
CF3
CF3
CF3
17
43
N
N
N
O
N
O
f
e
O
i
O
OEt
NH2
N
N
NHMe
CF3
CF3
O
4
6
52
NH
N
CF3
g
O
51
O
CF3
53
a
Reagents and Conditions: (a) methanol, Et N, RT, then hydrazine hydrate, water, 120 °C; (b)
3
methyl chloroformate, NaHCO , water, dioxane, 0 °C; (c) trifluoroacetic anhydride, Et N, DCM,
3
3
RT; (d) R C(OEt) 150 °C; (e) ethyl oxalyl monochloride, Et N, DCM, 0 °C then tosyl chloride,
1
3,
3
RT; (f) Nꢀmethylhydrazinecarbothioamide, DCM, RT then EDC, 40 °C; (g)
carbonyldiimidazole, Et N, DCM, RT; (h) Burgess reagent, THF, MW radiation, 100 W, 150 °C;
3
(i) ammonia in methanol, RT.
7
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 8 of 51
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
The synthesis of compounds containing variations in the 2ꢀposition of the oxadiazole
heterocycle
(47–53)
is
summarized
in
Scheme
2.
The
precursor
3ꢀ
(
trifluoromethyl)benzohydrazide (43) was obtained from the corresponding benzoyl chloride 17
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
in a twoꢀstep procedure. Coupling of 43 with methyl chloroformate and trifluoroacetic anhydride
afforded 44 and 45. Ringꢀclosure was carried out under microwave radiation using Burgess
3
6
reagent as selective dehydrating agent yielding compounds 49 and 50. Heating of 43 in triethyl
3
7
orthoformate or triethyl orthoacetate provided the oxadiazoles 47 and 48, respectively. 46 was
synthesized in a oneꢀpot twoꢀstep procedure, following the coupling of 43 with ethyl oxalyl
38
monochloride and ring closure with tosyl chloride. 46 was treated with ammonia in methanol
3
9
to directly transform the ethyl ester in the corresponding carboxamide 52. The Nꢀmethylated
analog 51 was obtained in a oneꢀpot twoꢀstep protocol: the benzoylchloride was coupled with Nꢀ
40
methylhydrazinecarbothioamide and ring closure was carried out by employing EDC.
4
1
Oxadiazolꢀ2ꢀone 53 was prepared by conversion of 43 with carbonyldiimidazole.
8
ACS Paragon Plus Environment

Page 9 of 51
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
Identification of PqsR Ligands by Fragment Screening. The detection and analysis of
biomolecular interactions is one of the key issues in the hit identification phase of a drug
discovery project. Surface plasmon resonance spectroscopy represents a powerful tool to monitor
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
42
and characterize binding events in realꢀtime. We have applied this technology to screen a
library containing 720 compounds with a molecular weight ranging from 100 to 350 g/mol (for
details see SI; Table S2). PqsR was immobilized on the sensor chip surface via biotinꢀ
streptavidin specific interaction and compounds were initially screened at a single concentration
of 100 ꢁM. Approximately 40 binding compounds were selected based on affinity and selectivity
against negative control (bovine carbonic anhydrase II). Compounds showing undesirable SPR
43
binding characteristics as described by Giannetti et al. were not taken into consideration. These
included fragments displaying binding irregularities, fragments with nonꢀtypical slow
dissociation and compounds producing superꢀstoichiometric binding responses. The obtained hit
rate of 5ꢀ6% can be regarded as excellent compared to the screening of more complex
2
8
structures. The best fragments were further subjected to SPR doseꢀranging experiments to
accurately determine dissociation constants (K values). The five top candidates displayed KD
D
values below 11 ꢁM (4ꢀ8, Figure 2 and Table 1). To further evaluate binding of the best
screening hits we performed SPR competition experiments similar to those described
4
4
previously. Dose binding experiments with mixtures consisting of various concentrations of
either compound 4 or 5 and constant concentrations of PQS showed no overall additivity in the
SPR response (SI, Figure S2). Increasing the PQS concentration did not exceed the maximal
binding capacity of the chip surface (Rmax; experimentally determined for the different
fragments). These results show that both fragments compete with PQS for binding to PqsR.
9
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 10 of 51
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
Figure 2: Chemical structures of highest affinity compounds 4-8 and compound 9
As evaluation of the K values alone gives no clear indication of which ligands to select for
D
4
5
further progression, the thermodynamic parameters, which allow a more detailed understanding
4
6
of biomolecular interactions, were determined (Table 1) using isothermal titration calorimetry
ITC).
Table 1: Biophysical characterization of PqsR ligands.
(
a
Comp.
K
[ꢁM]
KD (ITC)
[ꢁM]
ꢂG
[kcal mol ]
ꢂH
[kcal
ꢀTꢂS
LE
D (SPR)
b
ꢀ1
ꢀ
[kcal
[kcal mol
ꢀ
1
ꢀ1
1
mol ]
ꢀ9.7±0.3
ꢀ8.6±0.2
ꢀ8.2±0.9
ꢀ6.9±0.1
ꢀ9.3±0.4
ꢀ8.6±0.0
ꢀ
mol ]
]
2
4
ꢀ
0.9±0.0
1.8±0.2
3.7±0.1
7.6±0.7
ꢀ8.3±0.0
ꢀ7.9±0.1
ꢀ7,4±0.0
ꢀ7.0±0.1
ꢀ6.8±0.1
ꢀ6.9±0.0
ꢀ
1.5±0.3
0.7±0.2
0.7±0.9
0.1±0.1
2.5±0.4
1.6±0.1
ꢀ
0.63
0.60
0.57
0.70
0.85
0.49
ꢀ
0.8±0.2
4.5±0.0
6.8±3.5
5
6
7
6.8±3.8 10.0±1.3
10.7±6.5 8.5±0.7
8
9
200ꢀ500
ꢀ
37
41
ꢀ
ꢀ
ꢀ
ꢀ
ꢀ
1.3±0.3
12.4±1.0
3.1±0.6
6.5±0.5
12.9±2.4
ꢀ8.0±0.1
ꢀ6.7±0.0
ꢀ7.5±0.1
ꢀ7.1±0.0
ꢀ6.7±0.1
ꢀ8.5±0.5
0.4±0.5
0.50
0.51
0.47
0.47
0.42
ꢀ4.0±0.1 ꢀ2.6±0.1
4
4
4
2
7
8
ꢀ7.6±0.0
ꢀ9.0±0.5
0.0±0.1
1.9±0.3
ꢀ3.8±0.2 ꢀ2.8±0.3
a
SPR measurements and ITC titrations were performed at 20 °C and 25 °C, respectively. Data
b
represent mean ± SD from at least two independent experiments. KD (SPR), PQS: 0.21±0.0 ꢁM
The ITC derived K values were in good accordance with the ones determined by SPR
D
experiments. The naphthaleneꢀ2ꢀcarboxamide 4 and the thiadiazole derivative 5 displayed the
1
0
ACS Paragon Plus Environment

Page 11 of 51
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
highest affinities (Table 1). Analysis of the thermodynamic signatures revealed that 4-8 are
enthalpyꢀdriven binders (Table 1) indicating a good nonꢀcovalent bond complementarity between
45
the protein site and the compounds. Given that it is harder to improve the ꢁH contribution than
to improve the ꢁS contribution, enthalpy driven binders are usually preferred for further
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
4
5
optimization. Moreover, we determined the ratio of Gibb’s free binding energy to the number
of nonꢀhydrogen atoms termed ligand efficiency (LE) that enables a better comparison of hits
4
7
with widely differing structures and activities. Compounds 4-8 showed LEs ranging from 0.49
to 0.85 kcal/mol and are above the threshold value of 0.33 kcal/mol, which should be exceeded
47
for an efficient optimization. In order to evaluate the agonistic and antagonistic activities of 4ꢀ
8
, we examined the PqsRꢀmediated transcriptional effect in a reporter gene assay by measuring
the βꢀgalactosidase activity in E. coli the pEAL08ꢀ2 plasmid containing the tac ꢀpqsR and
P
48
pqsA ꢀlacZ fusion genes (Table 2). E. coli was used, as it provides a system to characterize the
P
functionality of PqsR ligands independent of the entire pqs system and the low outer membrane
4
9
permeability characteristic for P. aeruginosa. While 6-8 showed no or only weak antagonistic
activity, compounds 4 and 5 exhibited marked antagonistic properties. In contrast to compound
4, compound 5 provided only moderate agonistic activity. Overall, the screening process
afforded thiadiazole 5 which combines a remarkably high affinity with regard to its low
molecular weight with superior functional properties in the E. coli reporter gene assay and was
therefore selected for further optimization.
Fragment Optimization. Another hit compound sharing a high degree of similarity with 5,
compound 9 (Figure 2), which had shown lower affinity in the SPR screen, but pure antagonistic
activity (Table 2) was included in the optimization strategy (Figure 3).
1
1
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 12 of 51
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
Figure 3: Fragment modification strategy
In the first modification step the position of the Cl substituent was shifted into the meta and
ortho position (Table 2), as we had observed that the antagonistic properties in the quinolone
25
class of PqsR antagonists strongly depended on the position of the aromatic substituents. The
compounds with Cl in meta position showed an opposed functional profile in case of both
heterocycles: the thiadiazole 31 displayed pure agonistic properties, while the oxadiazole 32
turned out to be a pure antagonist. The Cl substituent in ortho position (33, 34) led to a moderate
agonistic activity without antagonistic properties for both heterocycles.
Compound 32 was used for further optimization. The Cl group was exchanged for substituents
providing different electronic properties (35-42, Table 2). Omitting the Cl substituent (35) led to
a moderate agonistic activity. The bioisosteric exchange of Cl for electronꢀwithdrawing moieties
in meta position (36-39) resulted in pure antagonists with the Br compound 36 and the CF3
derivative 37 as the most potent ones (EC 7.5 ꢁM). Interestingly, in the class of HHQ derived
50
2
5
antagonist a CF substituted quinolone also showed the strongest antagonistic properties. The
3
introduction of a methoxy group (40) was accompanied by a complete loss of activity. In
contrast, the introduction of a methyl substituent resulted in a partial antagonist (41, SI, Figure
S5). In the case of the benzamide scaffold reported by Klein et al. (2, Figure 1) the exchange of
2
9
Cl for tertꢀbutyl resulted in an increased affinity due to an enthalpic gain. Assuming a similar
1
2
ACS Paragon Plus Environment

Page 13 of 51
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
binding mode of 2 and 37 (discussed in detail below) this structural modification was deemed to
have a similar effect in the oxadiazole class. Compound 42, however, showed a slightly reduced
affinity according to ITC analysis (Table 1) and more importantly a diminished antagonistic
activity. Overall, 37 showing a favorable logP of 2.12 (ACD Perspecta 14.0.0 consenus logP)
and a better aqueous solubility compared to 36 emerged as the most promising compound.
Accordingly, 37 was tested in the P. aeruginosa reporter gene assay. As the outer membrane is
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
4
9
much less permeable in P. aeruginosa the observed antagonistic activity being only 5ꢀfold
reduced (EC s: 7.5 ꢁM in E. coli, 38.5 ꢁM in P. aeruginosa; SI Figure S6) was remarkably
5
0
high.
1
3
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 14 of 51
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
a
Table 2: Agonistic and antagonistic activities of PqsR ligands
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
b
Compd.
Pos
X
R1
R2
Agonistic
activity
[%]
Antagonistic
activity
[%]
EC50
[ꢁM]
4
5
6
7
8
9
ꢀ
ꢀ
ꢀ
Cl
NH2
NH2
NH2
NH2
NH2
NH2
NH2
NH2
NH2
NH2
NH2
NH2
NH2
NH2
NH2
NH2
NH2
NH2
H
52±12
41±12
ꢀ
4
ꢀ
S
28±1
n.a.
43±16
n.i.
ꢀ
ꢀ
ꢀ
ꢀ
ꢀ
ꢀ
ꢀ
n.a.
22±17
n.i.
ꢀ
ꢀ
ꢀ
ꢀ
n.a.
ꢀ
4
3
3
2
2
3
3
3
3
3
3
3
3
3
3
3
3
3
3
ꢀ
O
S
Cl
n.a.
40±9
n.i.
ꢀ
ꢀ
3
1
Cl
105±11
n.a.
32
33
O
S
Cl
76±12
n.i.
26.7
ꢀ
Cl
48±1
22±5
28±3
n.a.
3
3
3
3
3
3
4
5
6
7
8
9
O
O
O
O
O
O
O
O
O
O
O
O
O
O
O
ꢀ
Cl
n.i.
ꢀ
H
n.i.
ꢀ
Br
105±6
111±13
84±1
69±6
n.i.
7.5
7.5
17.8
46.5
ꢀ
CF3
CN
OCF3
OMe
Me
tꢀButyl
CF3
CF3
CF3
CF3
CF3
CF3
ꢀ
n.a.
n.a.
n.a.
40
41
n.a.
c
n.a
55±20
42±11
49±22
n.i.
57.8
4
4
4
4
5
5
2
7
8
9
0
1
n.a.
ꢀ
n.a.
101.2
Me
n.a.
ꢀ
ꢀ
ꢀ
ꢀ
ꢀ
ꢀ
OMe
CF3
n.a.
n.i.
n.a.
n.i.
NHMe
CONH2
ꢀ
n.a.
n.i.
52
53
39±6
41±4
n.i.
n.i.
a
Agonistic and Antagonistic properties were evaluated in an E. coli reporter gene assay.
Agonistic activity was determined by measuring the PqsR stimulation induced by 100 ꢁM of the
test compound compared to 50 nM PQS (= 100%); n.a. = no agonism (agonistic activity ≤ 20%).
Antagonistic activity was determined by measuring the inhibition of the PqsR stimulation
induced by 50 nM PQS in the presence of 100 ꢁM test compound (full inhibition = 100%); n.i. =
no inhibition (antagonistic activity ≤ 20%). Mean values of at least two independent experiments
b
with n = 4. Concentration of the half maximal antagonistic activity. Mean values of at least two
independent experiments with compounds tested in at least 6 different concentrations and n = 4.
c
partial antagonist
1
4
ACS Paragon Plus Environment

Page 15 of 51
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
50,51
Characterization of Fragment Binding Mode. Despite many efforts,
the 3D structure of
the PqsR protein is not available. The amino acids Q194 and F221 are located within the LBD of
2
9
PqsR and are potential partners for nonꢀcovalent interactions. We could clearly show that Q194
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
and F221 are involved in the binding of benzamide 2 employing siteꢀdirected mutagenesis in
2
9
combination with ITC analysis. Accordingly, the optimized compound 37 and the screening hit
naphthaleneꢀ2ꢀcarboxamid 4 showing highest affinity were evaluated for their binding to the
Q194A and F221A PqsR mutants by ITC (Table 3).
Table 3: Effect of siteꢀdirected mutations on the thermodynamic parameters of PqsR ligands.
Q194A
ꢂꢂH
F221A
Ligand
ꢂꢂG
[kcal mol ]
ꢀTꢂꢂS
[kcal mol ]
ꢂꢂG
[kcal mol ]
ꢂꢂH
[kcal mol ]
ꢀTꢂꢂS
[kcal mol ]
ꢀ
1
ꢀ1
ꢀ1
ꢀ1
ꢀ1
ꢀ1
[kcal mol ]
☆
☆
☆
☆
☆
☆
2
4
ꢀ1.3±0
ꢀ1.1±0.1
ꢀ4.4±0.3
3.1±0.3
2.1±0.2
ꢀ1.0±0.1
ꢀ0.4±0.1
ꢀ7.6±0.3
6.6±0.4
6.7±0.2
☆
☆
☆
★
☆
☆
ꢀ3.2±0.2
ꢀ7.1±0.2
☆
☆
☆
☆
3
7
ꢀ0.6±0.1
ꢀ0.3±0.6
ꢀ0.3±0.7
ꢀ0.6±0.1
ꢀ7.1±0.5
6.5±0.6
ꢂꢂG, ꢂꢂH, and –TꢂꢂS are ꢂGWT – ꢂGmutant, ꢂHWT – ꢂHmutant, and –T(ꢂSWT – ꢂSmutant),
respectively. Negative values indicate a loss, positive values a gain compared to wildꢀtype.
Errors indicate SD calculated via Gaussian error propagation. Significance: effect of the point
☆
mutation on the thermodynamic parameters of ligand binding compared to the wildꢀtype. p <
★
0
.003; p < 0.05.
Comparing the thermodynamic signatures of 4 binding to Q194A mutant and wildꢀtype PqsR a
loss in ꢂH of 3.2 kcal/mol was observed. Given that a wellꢀplaced Hꢀbond makes a favorable
52
enthalpic contribution in a range from 4 to 5 kcal/mol, this value indicates the presence of a
weak Hꢀbond between 4 and Gln194. The enthalpic contribution of 4 binding to the F221A
mutant was significantly reduced compared to the wildꢀtype (ꢂꢂH = ꢀ7.1 kcal/mol). F221 might
be involved in a πꢀπ interaction with the naphthalene core of 4. However, theoretical
investigations revealed an interaction energy of only 4.0 to 4.2 kcal/mol for the benzeneꢀ
1
5
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 16 of 51
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
53
naphthalene complex. In all cases the loss in enthalpy is partly compensated by a gain in
entropy. This might be due to a higher conformational flexibility of the complex in the absence
of specific nonꢀcovalent interactions. Overall, the measured data indicate that napthaleneꢀ2ꢀ
carboxamide 4 adopts the same binding mode as the benzamide 2.
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
In contrast, there was no significant difference in the enthalpic contribution observed for the
oxadiazole 37 when binding to Q194A mutant and wildꢀtype PqsR. This indicates that the sideꢀ
chain of Gln194 is not directly involved in the binding of 37. As observed for 4, the binding of
37 was also affected by the F221A mutation (ꢂꢂH = ꢀ7.1 kcal/mol). The phenyl moiety of 37
might be involved in a πꢀstacking interaction with F221. Overall, these data lead to the
conclusion that aromatic carboxamides 2 and 4 as well as the oxadiazole 37 interact with
different amino acids within the LBD by adopting different binding modes. ITC analysis was
applied to analyze whether both fragments bind to the same site. PqsR was saturated with an
excess of 2 and 37 was titrated subsequently into the same solution (Figure 4a). No additional
heat release was detected which gives strong evidence that both compounds compete for binding
to PqsR. Additionally, the vice versa experiment was performed to exclude that the binding of 2
induces a conformational change which hinders the binding of 37. This experiment led to the
same result (SI Figure S4) confirming that both fragments bind to the same site. In order to
54
further validate the competitive binding (2, 37) we applied NMR INPHARMA experiments.
This technique allows to detect proteinꢀmediated interligand NOEs, if both compounds bind
competitively to the same site of the protein. When we conducted a 2D NOESY experiment of a
mixture of 2 and 37 in the presence of PqsR, clear interligand NOEs between the tertꢀbutyl group
of 2 and the phenyl ring of 37 were observed (Figure 4b, upper panel). Additionally, weaker
interligand NOEs between the two phenyl moieties of 2 and 37 were detected (Figure 4b, lower
1
6
ACS Paragon Plus Environment

Page 17 of 51
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
panel). These findings corroborated our ITC data that both ligands bind competitively to the
same site of PqsR. Given that both fragments interact with F221 and compete with PQS as
shown for the structural analogs 4 and 5 in SPR competition experiments, we conclude that 2 and
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
37 have overlapping binding sites within the LBD.
1
7
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 18 of 51
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
Figure 4. Competitive binding of 2 and 37 to PqsR (a) Raw ITC data (top) and integrated
normalized data (bottom) for titrations of 43 ꢁM PqsR with 500 ꢁM 37 in the absence of 2 (■)
and in the presence of 500 ꢁM compound 2 (ꢂ). (b) NOESY spectrum of a mixture of 500 ꢁM 2
and 400 ꢁM 37 in the presence of 10 ꢁM PqsR; interligand NOEs between the tertꢀbutyl moiety
of 2 and the phenyl ring of 37 (top), between the two phenyl rings (bottom); ( ) indicates
*
interligand INPHARMA NOEs mediated by PqsR; ( ) indicates intramolecular transferred
**
NOEs. Signals marked with ( ) or ( ) were not detected in a control experiment recorded under
*
**
identical conditions in the absence of PqsR.
1
8
ACS Paragon Plus Environment

Page 19 of 51
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
Biophysically Guided Structural Modifications. Structural modifications of 37 in
combination with thermodynamic analysis were performed to evaluate the relative contributions
of the different functional groups. The CF moiety was replaced by a methyl group and the
3
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
thermodynamic signature of the resulting compound 41 was compared to that of 37 (Figure 5).
Figure 5. Relative thermodynamic contributions of functional groups. Negative values indicate a
☆
★
loss; positive values a gain compared to compound 37; p < 0.003; p < 0.05
Compound 41 showed a reduced affinity for PqsR compared to 37 (Table 1), which was due to
a significant loss in ꢂH of 4.4 kcal/mol (Figure 5). This corroborates the proposed πꢀstacking
between F221 and the phenyl moiety of the oxadiazole 37 as electronꢀwithdrawing groups like
5
5,56
CF can intensify such interactions either through enhanced electrostatic
or direct effects of
3
57
the substituents. Considering the loss in enthalpy that is caused by the F221A mutation (7.1
kcal/mol) and the loss in enthalpy that is due to the exchange of the CF by CH (4.4 kcal/mol),
3
3
the enthalpic contribution of a possible πꢀπ interaction between 37 and Phe221 might be 2.7
1
9
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 20 of 51
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
kcal/mol. Remarkably, this is consistent with the binding energy of the gasꢀphase benzene dimer
5
8
ranging from 2.0 to 3.0 kcal/mol.
The thermodynamic profile of compound 37 binding to Q194A suggested that the amino group
of 37 does not make an Hꢀbond with the polar sideꢀchain of Gln194. To investigate whether the
amino group is involved in the formation of an Hꢀbond with polar backbone atoms or other sideꢀ
chains in the LBD of PqsR it was replaced by hydrogen. The binding affinity of the resulting
compound 47 was reduced (Table 1). The same tendency was observed for the functional
properties (Table 2). But even more interesting, there was no significant difference in the
enthalpic contributions (Figure 5). These findings initially led us to the conclusion that the amino
group is not involved in a specific nonꢀcovalent interaction (e.g. Hꢀbond) with PqsR. Hence, we
decided to replace the amino function for apolar moieties with different electronic properties like
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
methyl (48), methoxy (49) and CF (50) in order to overcome the enthalpy penalty associated
3
5
2
with desolvation of the polar amino group. In contrast to our assumption, compound 48
displayed a reduced affinity toward PqsR (Table 1), which results from a significant decrease in
52
ꢂH of 4.7 kcal/mol (Figure 5) typical for a loss of a wellꢀplaced Hꢀbond. Furthermore, these
modifications (48-50) completely abolished the antagonistic activity (Table 2). These findings
led us reconsider our hypothesis of a nonꢀinteracting amino function: the CꢀH bond in 2ꢀposition
of the 1,3,4ꢀoxadiazole (47) is highly polarized and can interact through a nonꢀclassical hydrogen
5
9
bond. In our case such an interaction could be mediated by a water molecule that is displaced in
the presence of an amino function (37). This explanation is sustained by the loss in entropy (ꢀ1.5
kcal/mol) due to the exchange of NH by H. Theoretical investigations revealed an interaction
2
energy for highly similar 5ꢀmembered heterocycles like oxazoles and isoxazoles in the range of
2
0
ACS Paragon Plus Environment

Page 21 of 51
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
59
3
.0 to 3.5 kcal/mol. Such an interaction might compensate the loss in enthalpy by omitting the
amino group.
In order to evaluate whether both hydrogen atoms of the amino substituent (37) are involved in
ligandꢀreceptor interaction we synthesized the Nꢀmethylated analogue 51. The latter showed no
antagonistic activity (Table 2) as did the carboxamide (52) and oxadiazolꢀ2ꢀon (53) indicating
that the amino group is essential for full antagonistic activity even if the exact function in the
ligandꢀreceptor interaction remains speculative.
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Effects on alkylquinolone signaling and virulence factor pyocyanin formation in P.
aeruginosa. Following our strategy of inhibiting P. aeruginosa pathogenicity, we evaluated the
effect of different oxadiazole derivatives on the production of alkylquinolones and the major
virulence factor pyocyanin in the highly virulent clinical isolate P. aeruginosa PA14. The
activation of PqsR upregulates the pqsABCD(E) operon which is essential for HHQ and PQS
1
6,19,20
21
biosynthesis.
Furthermore, PQS is reported to influence biofilm shape. Accordingly, we
examined the effect of 37 on extracellular levels of PQS in wildꢀtype P. aeruginosa. As we
25
reported for the HHQ derived antagonist no remarkable reduction of PQS levels was observed
(data not shown). We assumed that the autoinductive effect of high intrinsic PQS levels ranging
from 30 to 60 ꢁM in the culture supernatant might compensate the antagonistic activity of 37.
Therefore, we further investigated the effect of compound 37 on HHQ formation in the absence
17
of PQS. We employed a P. aeruginosa isogenic strain lacking PqsH, a monooxygenase
19,20
catalyzing the conversion of HHQ to PQS.
Interestingly, we observed a reduction of
extracellular HHQ (42±0% at 250 ꢁM of 37). The level of HQNO (2ꢀheptylꢀ4ꢀhydroxy
20
quinolineꢀNꢀoxide), another alkylquinolone under the control of PqsR, was also clearly
decreased (43±3% at 250 ꢁM of 37). The reduction of both alkylquinolones demonstrates that
2
1
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 22 of 51
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
PqsR antagonists block the QS system in P. aeruginosa. Future experiments will extend the
fragment structure aiming at antagonists with improved affinity. Such compounds would
hopefully disrupt the selfꢀinductive effect of high intrinsic PQS levels and potentially affect
biofilm shape in wildꢀtype P. aeruginosa.
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
The biological effects on the major virulence factor pyocyanin were examined subsequently.
Pyocyanin is a redoxꢀactive secondary metabolite released by P. aeruginosa that plays a critical
role during acute infections by provoking oxidative stress and inflammatory responses in host
6
0
cell tissue. Biological studies indicate a connection between the PQS system and pyocyanin
18,61
ꢀ
production,
pyocyanin.
supported by the fact that a PqsR deficient mutant (pqsR ) is incapable to produce
2
0,62
25
29
Analogously, the HHQ derived antagonist 1 and the hydroxamic acid 3
diminish pyocyanin in P. aeruginosa. The PqsR antagonist 37 likewise clearly reduced the
pyocyanin level (46±9% at 250 ꢁM). The weaker antagonists 38 and 41 showed a moderate or
only little reduction, respectively (32±3% at 250 ꢁM of 38; 18±3% at 250 ꢁM of 41). In contrast
the inactive compound 40 showed no effect. The correlation of these data with the antagonistic
properties indicates that the reduction of pyocyanin is caused by PqsR antagonism. A growth
curve of P. aeruginosa incubated with 37 confirmed that the observed effects are not due to an
antibacterial activity of the compound (SI, Figure S6). While in the absence of functional PqsR
ꢀ
20
(
pqsR ) no significant levels of pyocyanin and PQS were detected, our results show that in the
presence of weaker antagonists pyocyanin production is more sensitive to PqsR antagonism than
PQS biosynthesis. This finding might facilitate a more detailed understanding of the underlying
mechanisms. Overall these results demonstrate that the interference with the PQS system by the
oxadiazole type antagonists is a suitable approach to reduce P. aeruginosa pathogenicity.
2
2
ACS Paragon Plus Environment

Page 23 of 51
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
CONCLUSION
Considering the hydrophobic character of the natural effectors HHQ and PQS, the LBD of
PqsR is expected to be a mainly lipophilic pocket. This renders the transcriptional regulator PqsR
a quite challenging drug target. Nevertheless, we identified chemically diverse PqsR ligands by
employing an SPR fragment screening. The thermodynamic profiles of the top five hit
compounds were evaluated using ITC methodology. The discovered hits showed enthalpy driven
binding and excellent LEs. Furthermore, the functional properties of the best fragments were
analyzed in a ßꢀgalactosidase reporter gene assay as an additional guidepost for hit selection. The
thiadiazoleꢀ2ꢀamine derivative 5 displayed a mixed agonistic/antagonistic profile and was chosen
as starting point for structural modification. A functionalityꢀguided optimization process
afforded the pure antagonist 37 showing improved affinity and potency with consistently high
LE (5, LE = 0.57 kcal/mol vs. 37, LE = 0.50 kcal/mol). In contrast to the HHQ derived
antagonists (for example 1, Figure 1), 37 showed a more favorable logP and higher aqueous
solubility. The 2ꢀaminoꢀ1,3,4ꢀoxadiazole moiety present in the optimized ligand 37 has not been
reported as a chelating agent for metal ions, contrary to the hydroxamic acid motif of fragment 3
(Figure 1). Furthermore, an inꢀsilico toxicology prediction (Derek Nexus, Version
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
6
3
2
.0.3.201206121715) of compound 37 revealed no safety issues. Taken together these data
emphasize the high potential of the optimized fragment 37.
To shed light on the binding mode of different fragment classes, siteꢀdirected mutagenesis in
combination with thermodynamic analysis was applied. The oxadiazole 37 and the carboxamides
2
and 4 bind competitively to PqsR but do not cover exactly the same space in the LBD. This
provides the potential for hitꢀtoꢀlead optimization using fragment growing strategies.
Additionally, structural modifications of 37 together with calorimetric evaluation were
2
3
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 24 of 51
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
performed to evaluate the relative importance of the functional groups in order to guide future
optimization strategies. Furthermore, compound 37 is able to reduce the virulence factor
pyocyanin in wildꢀtype P. aeruginosa without affecting bacterial growth. Thus, the identified
PqsR antagonists represent an interesting starting point for the development of novel antiꢀ
infectives.
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
Future experiments will aim at merging the best features of the different fragments to develop
drugꢀlike molecules which may provide novel therapeutic options to counter the burden of P.
aeruginosa infections.
2
4
ACS Paragon Plus Environment

Page 25 of 51
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
EXPERIMENTAL SECTION
Surface Plasmon Resonance. SPR binding studies were performed using a Biacore T100
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
instrument optical biosensor (GE Healthcare). CSIRO fragment library collection consisted of
720 compounds with a molecular weight range of 100 to 350 Da. The expression, purification
C87
and minimal biotinylation of H SUMOꢀPqsR
as well as the immobilization onto the
6
streptavidin chip surface and is described in detail in the Supporting Information.
SPR Library Screening. 10 mM DMSO stock solutions of compounds were diluted in 1.05 x
DMSO free screening buffer (52.5 mM HEPES, pH 7.4, 157.5 mM NaCl, 0.0525% Tweenꢀ20
(v/v)) to 500 ꢁM. Final ligand concentrations (100 ꢃM) were achieved by diluting 1:5 (v/v) in
the instrument running buffer (50 mM HEPES, pH 7.4, 150 mM NaCl, 0.05% Tweenꢀ20 (v/v)
and 5% DMSO (v/v)). Two separate fragment screens were performed each time screening 360
fragments against protein targets immobilized on chips I and II. Fragment solutions (100 ꢁM)
were injected over the protein sensor surface for 30 s at 60 ꢁL/min at 4 °C. After every 120
injections, a positive control compound (PQS in the first 360 compound screen and fragment 4 in
the second screen) was injected for control binding purposes. Regeneration of the surfaces
between subsequent binding experiments was achieved by washing the surface at 60 ꢁL/min for
30 s with the instrument running buffer as well as by using a “carryꢀover injection” of running
buffer at 40 ꢁL/min for further 30 s. To determine binding affinity, candidate hit fragments were
analyzed using dosage experiments. Fresh 100 mM DMSO fragment solutions were diluted
directly into instrument running buffer to a final concentration of 81 ꢁM and then diluted 3ꢀfold
down to 1 ꢁM aiming for a 5ꢀpoint concentration series range. Each compound was injected for
30 s associationꢀ and 60 s dissociation time at 20 °C at a constant flow rate of 60 ꢁL/min.
2
5
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 26 of 51
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
Scrubber 2 (www.biologic.com.au) and Microsoft Excel software packages were utilized for
data processing and analysis. Using Scrubber software, SPR signals were referenced against the
blank surface (streptavidin + Dꢀbiotin) and further corrected for DMSO refractive index change
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
64
43
as previously described. A normalization scheme of Giannetti et al. was applied to the
processed data based on the maximal binding response (Rmax) that has been determined from
experiments with the control compound (initially PQS and later fragment 4). For the initial
ranking of the best hits, the K values were estimated using equation derived from the Langmuir
D
adsorption isotherm (1):
ꢄꢅꢆꢇ∗ꢈ
ꢀꢁ ꢂꢃ
ꢉ ꢊꢃ (1)
ꢄ
Rmax, R, and C correspond to the normalized saturation response of the compound, the
normalized response of the test compound, and the concentration of the test solution,
respectively.
To determine binding affinities (K ) from dosage experiments, binding responses at
D
equilibrium were fit to a 1:1 steady state affinity model available within Scrubber. As described
above, normalized saturation response (Rmax), derived using reference compound(s), was applied
to response curves obtained with fragment hits at nonꢀsaturating concentration. Equilibrium
isotherm containing the assumed Rmax was used to determine K values (SI, Figure S1).
D
2
9
Isothermal titration calorimetry ITC titrations were carried out as previously reported and
a detailed protocol as well as representative titration curves are included in the Supporting
Information.
2
6
ACS Paragon Plus Environment

Page 27 of 51
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
C87
INPHARMA NMR Experiments To detect interligand NOEs, 10 ꢀM H SUMOꢀPqsR was
6
mixed with 500 ꢀM 2 and 400 ꢀM 37 in PBS buffer (50 mM potassium phosphate, pH 7.4, 150
mM NaCl, 5% DMSOꢀd (v/v), 5 % D O (v/v)). Spectra were acquired at 25 °C on a 700 MHz
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
6
2
Bruker Avance III spectrometer equipped with a cryoprobe. Twoꢀdimensional NOESY spectra
were recorded with a mixing time of 600 ms, 16 scans for each of 400 complex points in the
indirect dimension and 512 complex points in the direct dimension, a sweep width of 9 ppm and
a recycle delay of 1.5 sec. Dephasing of water signals was achieved by a WATERGATE
C87
sequence. A control experiment under identical conditions without H SUMOꢀPqsR
was
6
carried out. Data were processed using Bruker TOPSPIN 3.1 software using squared sine bell
functions and zero filling to 2048 and 1024 complex points in the direct and indirect dimension,
respectively, for apodization.
Reporter Gene Assay in E. coli. The ability of the compounds to either stimulate or
25
antagonize the PqsRꢀdependent transcription was analysed as previously described using a βꢀ
galactosidase reporter gene assay in E. coli expressing PqsR with some modifications to enable a
6
5
48
higher throughput. Briefly, a culture of E. coli DH5α cells containing the plasmid pEAL08ꢀ2
which encodes PqsR under the control of the tac promoter and the βꢀgalactosidase reporter gene
lacZ controlled by the pqsA promoter were co incubated with test compound. Antagonistic
effects of compounds were assayed in the presence of 50 nM PQS. After incubation
galactosidase activity was measured photometrically using POLARstar Omega (BMG Labtech)
and expressed as ratio of controls. For the determination of EC value, compounds were tested
5
0
at least in six different concentrations. The given data represent mean of at least two experiments
2
7
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 28 of 51
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
with n = 4. The log (inhibitor) vs. response model (Prism 5.0) was applied for nonlinear
regression and determination of EC ꢀvalues.
50
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
Reporter Gene Assay in P. aeruginosa. In order to evaluate the antagonistic properties of
2
5
compound 37, the PqsRꢀdependent transcription was evaluated as previously described using a
4
8
βꢀgalactosidase reporter gene assay system in P. aeruginosa. In brief, a P. aeruginosa PA14
strain carrying a nonfunctional pqsA gene was used to eliminate intrinsic HHQ and PQS
4
8
production. A culture of P.aeruginosa PA14ꢁpqsA cells containing the plasmid pEAL08ꢀ2
were co incubated with 37 and 50 nM PQS. After incubation galactosidase activity was
measured photometrically using POLARstar Omega (BMG Labtech) and expressed as ratio of
controls. For the determination of EC value, 37 was tested at nine different concentrations. The
50
given data represent mean of two experiments with n = 4. The log (inhibitor) vs. response model
(
Prism 5.0) was applied for nonlinear regression and determination of EC ꢀvalues.
50
Pyocyanin Assay. Pyocyanin produced by P. aeruginosa PA14 was determined as described
6
6
67
previously according to the method of Essar et al. In short, cultures inoculated with a starting
OD600 of 0.02 were grown in the presence of inhibitors or DMSO as a control in PPGAS medium
at 37 °C, 200 rpm and a humidity of 75% for 16 h. For pyocyanin determination, cultures were
extracted with chloroform and reꢀextracted with 0.2 M HCl. The OD520 was determined using
FLUOstar Omega (BMG Labtech) and normalized to cell growth measured as OD . For each
6
00
sample, cultivation and extraction were performed in triplicates.
2
8
ACS Paragon Plus Environment

Page 29 of 51
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
Determination of Extracellular HHQ and HQNO Levels. Extracellular levels of HHQ and
6
8,69
HQNO were determined according to the methods of Lépine et al.
Briefly, cultures of P. aeruginosa PA14ΔpqsH mutant were inoculated with a starting OD600 of
.02 and grown in LB medium in the presence of DMSO as a control or DMSO solutions of
inhibitors at 37 °C, 200 rpm and a humidity of 75 % for 17 h. For HHQ and HQNO analysis, 500
l of bacterial cultures were supplemented with 50 ꢁL of a 10 ꢁM methanolic solution of the
with some modifications:
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
0
ꢁ
internal standard (IS) 5,6,7,8ꢀtetradeuteroꢀ2ꢀheptylꢀ4(1H)ꢀquinolone (HHQꢀd ) and extracted
4
with 1 mL of ethyl acetate by vigorous shaking. After centrifugation (14, 000 rpm, 12 min), 400
ꢁ
L of the organic phase were evaporated to dryness in LC glass vials and dissolved in methanol.
For each sample, cultivation and extraction were performed in triplicates. UHPLCꢀMS/MS
67
analysis was performed as described in detail by Storz et al. The following ions were
monitored (mother ion [m/z], product ion [m/z], scan time [s], scan width [m/z], collision energy
[V], tube lens offset [V]): HHQ: 244, 159, 0.5, 0.01, 30, 106; HQNO: 260, 159, 0.1, 0.01, 25, 88;
HHQꢀd4 (IS): 248, 163, 0.1, 0.01, 32, 113. Xcalibur software was used for data acquisition and
quantification with the use of a calibration curve relative to the area of the IS.
1
13
Chemical and Analytical Methods. H and C NMR spectra were recorded on a Bruker
DRXꢀ500 instrument. Chemical shifts are given in parts per million (ppm), and referenced
against the residual solvent peak. Coupling constants (J) are given in hertz. Purity control of final
compounds was carried out using a SpectraSystemsꢀMSQ LCMS system (Thermo Fisher
Scientific) consisting of a pump, an autosampler, VWD detector and a ESI quadrupole mass
spectrometer by determination of the relative peak area in the UV trace. Purities were greater
than 95%. Mass spectrometry was performed on an MSQ electro spray mass spectrometer
2
9
ACS Paragon Plus Environment