Fluorescence polarization-based assays for detecting compounds binding to inactive c-Jun N-terminal kinase 3 and p38α mitogen-activated protein kinase

Article abstract of DOI:10.1016/j.ab.2016.02.018

Two fluorescein-labeled pyridinylimidazoles were synthesized and evaluated as probes for the binding affinity determination of potential kinase inhibitors to the c-Jun N-terminal kinase 3 (JNK3) and p38α mitogen-activated protein kinase (MAPK). Fluorescen

Full text of DOI:10.1016/j.ab.2016.02.018

Accepted Manuscript
Fluorescence polarization-based assays for detecting compounds binding to inactive
JNK3 and p38α MAP kinase
Francesco Ansideri, Andreas Lange, Ahmed El-Gokha, Frank M. Boeckler, Pierre
Koch
PII:
S0003-2697(16)00096-8
10.1016/j.ab.2016.02.018
YABIO 12324
DOI:
Reference:
To appear in: Analytical Biochemistry
Received Date: 4 December 2015
Revised Date: 25 February 2016
Accepted Date: 26 February 2016
Please cite this article as: F. Ansideri, A. Lange, A. El-Gokha, F.M. Boeckler, P. Koch, Fluorescence
polarization-based assays for detecting compounds binding to inactive JNK3 and p38α MAP kinase,
Analytical Biochemistry (2016), doi: 10.1016/j.ab.2016.02.018.
This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to
our customers we are providing this early version of the manuscript. The manuscript will undergo
copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please
note that during the production process errors may be discovered which could affect the content, and all
legal disclaimers that apply to the journal pertain.

ACCEPTED MANUSCRIPT
Fluorescence polarization-based assays for detecting compounds binding to inactive
JNK3 and p38α MAP kinase
Francesco Ansideri a, Andreas Lange b, Ahmed El-Gokha a,c, Frank M. Boeckler b, Pierre
Kocha*
Address:
aInstitute of Pharmaceutical Sciences, Department of Medicinal and Pharmaceutical
Chemistry, Eberhard Karls Universität Tübingen, Auf der Morgenstelle 8, 72076 Tübingen,
Germany
bInstitute of Pharmaceutical Sciences, Molecular Design and Pharmaceutical Biophysics,
Eberhard Karls Universität Tübingen, Auf der Morgenstelle 8, 72076 Tübingen, Germany
cMenofia University, Chemistry Department, Faculty of Science, Menofia, Egypt
*Corresponding author:
Email: Pierre Koch*- pierre.koch@uni-tuebingen.de
Keywords:
Fluorescence polarization; binding assay; c-Jun N-terminal kinase 3; p38α MAP kinase
1Abbreviations used: MAPK, mitogen-activated protein kinase; JNK, c-Jun N-terminal kinase;
AD, Alzheimer’s disease; PD, Parkinson’s disease; ATP, adenosine triphosphate; HTS, high
throughput screening; FP, fluorescence polarization; FITC, fluorescein isothiocyanate;
1

ACCEPTED MANUSCRIPT
ELISA, enzyme-linked immunosorbent assay; MOE, molecular operating environment; rmsd,
root-mean-square deviation; PDB, protein data bank; TLC, thin layer chromatography; HPLC,
high performance liquid chromatography; DAD, diode array detector; NMR, nuclear magnetic
resonance; ppm, parts per million; ESI-MS, electrospray ionization mass spectrometer; r.t.,
room temperature; BME, β-mercaptoethanol; BSA, bovine serum albumin; rpm, revolutions
per minute; %CV, percent coefficient of variation; SD, standard deviation; ITC, isothermal
titration calorimetry.
Abstract
Two fluorescein-labelled pyridinylimidazoles were synthesized and evaluated as probes for
the binding affinity determination of potential kinase inhibitors to the c-Jun N-terminal kinase
3 (JNK3) and p38α mitogen-activated protein kinase (MAPK). Fluorescence polarization
(FP)-based competition binding assays were developed for both enzymes using 1-(3',6'-
dihydroxy-3-oxo-3H-spiro[isobenzofuran-1,9'-xanthen]-5-yl)-3-(4-((4-(4-(4-fluorophenyl)-2-
(methylthio)-1H-imidazol-5-yl)pyridin-2-yl)amino)phenyl)thiourea (5) as an FP-probe (JNK3:
Kd = 3.0 nM; p38α MAPK: Kd = 5.7 nM). The validation of the assays with known inhibitors of
JNK3 and p38α MAPK revealed that both FP-assays correlate very well with inhibition data
received by the activity assays. This, in addition to the viability of both FP-based binding
assays for the high throughput screening procedure makes the assays suitable as
inexpensive pre-screening protocols for JNK3 and p38α MAPK inhibitors.
1. Introduction
Within the mitogen-activated protein kinase (MAPK)1 family, the two serine/threonine kinases
c-Jun N-terminal kinase 3 (JNK3) and p38α MAPK have emerged in the last decades as
particularly attractive therapeutic targets due to their implication in several pathologic
conditions. JNK3, which is predominantly expressed in the brain, is associated with
2

ACCEPTED MANUSCRIPT
neurodegenerative diseases such as Alzheimer’s disease (AD), Parkinson’s disease (PD)
and Huntington’s disease [1, 2]. On the other hand the p38α MAPK is responsible for
different inflammatory responses and thus has become a well-established target for some
chronic inflammatory diseases as rheumatoid arthritis [3], inflammatory bowel disease [4]
and psoriasis [5]. Recently, several clinical trials investigating the benefit of p38α MAPK
inhibitors for the treatment of chronic obstructive pulmonary disease were terminated [6, 7].
In addition, some studies have also highlighted p38α MAPK as a key protein in the
pathogenesis of neurodegenerative diseases including AD, PD and multiple sclerosis [8-10].
As a consequence, great efforts have been focused in finding inhibitors for both JNK3 and
p38α MAPK [11-13], which led to the achievement of molecules endowed with a good
potency, although often lacking a satisfactory selectivity.
JNK3 and p38α MAPK are both activated in response to environmental stress stimuli as well
as pro-inflammatory cytokines and take part in different signaling pathways [14-17]. The
activation is carried out by upstream MAPK kinases MKK4/MKK7 and MKK3/MKK6 for JNK3
and p38α MAPK, respectively, via a concomitant phosphorylation of a threonine and a
tyrosine residues situated at the activation loop [15, 18]. Amongst others, the most evident
conformational modification following is the shift of a tripeptide Glu-Phe-Gly situated at the
beginning of the activation loop, known as DFG motif, from an “out” conformation on the
inactive enzyme to an “in” conformation [19]. Such modification has been well characterized
and widely described for p38α MAPK [20, 21] and recently, the possible adoption of the
DFG-out conformation has been reported for JNK3 as well [22]. The majority of the p38α
MAPK inhibitors and almost all of the reported JNK3 inhibitors discovered so far are
classified as type I inhibitors, binding at the adenosine triphosphate (ATP) binding site of the
enzyme in its active conformation and thus being ATP competitive [23]. However, most of
type I inhibitors are also able to bind to the inactive form of the enzyme, probably due to the
fact that the position of the DFG motif doesn’t affect the binding mode of these compounds
[21, 24]. On the other side, type II inhibitors bind the ATP cleft of the inactive enzyme,
exploiting a hydrophobic pocket known as the deep pocket, which is only present in the DFG-
3

ACCEPTED MANUSCRIPT
out conformation [25]. Since the inactive form of the kinase is not able to bind ATP, this
modality of binding avoids the competition with high intracellular ATP concentrations.
Furthermore, as the amino acid composition of the deep pocket is low conserved within the
catalytic site of related kinases, type II binding allows to achieve an improved selectivity
profile. Despite the high desirability of MAPK type II inhibitors, only a few of these inhibitors
have been characterized for p38α MAPK [26], whereas to date, no type II inhibitors have
been reported for JNK3.
The activity of type I inhibitors on JNK3 and p38α MAPK is generally evaluated through
biological assays quantifying the amount of MAPK substrate phosphorylated from the active
enzymes in presence and absence of the inhibitor [27, 28]. These activity assays provide an
exhaustive indication about the inhibitory activity of compounds but are generally expensive,
time consuming, and not amenable for high throughput screening (HTS) due to multiple
separation and washing steps. The evaluation of type II inhibitors is instead far more
challenging, albeit some screening assays for inactive p38α MAPK have been reported in the
last years [29, 30]. An additional strategy in the identification of new kinase inhibitors is the
detection of the binding affinity of novel compounds for the enzymes, which is a substantial
prerequisite for any inhibitory activity. With this scope, herein we report the development and
optimization of fluorescence polarization (FP)-based competition binding assays, which
employ a novel fluorescein-labelled probe binding to both the active and inactive
conformations of JNK3 and p38α MAPK. The FP is a technique increasingly used within the
last years in the field of drug discovery, due to its several advantages as versatility, speed,
easy handling and cost efficiency [31, 32]. It is even useful for detecting weak binding events
only limited by compound solubility [33]. The basic principle of this methodology is that after
excitation by plane polarized light the fluorophore’s emission is polarized as well due to
photoselection. However, the capability of the fluorophore to retain such polarization
depends on its molecular size. Small molecules, as the fluorophore alone, are prone to fast
tumbling due to Brownian motions. Therefore the light emitted loses easily the polarization
achieved and produces low FP signal. On the other hand, the FP signal dramatically
4

ACCEPTED MANUSCRIPT
increases when the fluorophore binds to a protein, due to a slower tumbling of the high
molecular weight complex. As a result, the displacement of a fluorescent probe from the
enzyme by a tested compound can be followed by measuring the FP signal directly after the
addition (mix and measure) without separation and washing steps. This feature, together with
a significantly improved safety, makes the FP-based binding assay superior in comparison to
traditional binding assays based on radioligands. [34, 35].
The herein presented assay, using slightly different conditions, has been recently employed
from our group on the wildtype form as well as on gatekeeper mutations of JNK3 [36] and
represents the first example of an FP-based binding assay on JNK3. On the other side
Munoz et al. reported in 2010 an FP competition assay for the inactive form of the p38α
MAPK [37]. The probe employed in this assay (compound 2, Fig. 1) is based on the
prototypical p38α MAPK inhibitor SB203580, which was modified on its side chain and
labelled with the 4’-isomer of fluorescein isothiocyanate (FITC). However, the FP-assay
reported by Munoz et al. does not show correlation with the p38α MAPK activity assay (R² = -
0.4005) [37].
Fig. 1. Reported probe for p38α MAPK FP assay based on the p38α MAPK prototype
inhibitor SB203580.
5

ACCEPTED MANUSCRIPT
Herein, we describe the development of two novel and wider applicable probes (compounds
5 and 6, Fig. 2) due to their derivation from dual JNK3/p38α MAPK inhibitors (compounds 3
and 4, Fig. 2). Two analogous binding assays employing the inactive form of both, JNK3 and
p38α MAPK, using the most promising probe 5 were optimized. The use of the inactive form
of both kinases enhances the application spectrum of these assays since it enables the
evaluation of binding affinity for type I as well as type II inhibitors. The results obtained from
the binding assays were correlated with the ones obtained from enzyme-linked
immunosorbent assay (ELISA) activity assays and the viability of the optimized protocols in a
HTS format was investigated.
Fig. 2. Novel probes for JNK3 and p38α MAPK FP assay based on potent dual JNK3/p38α
MAPK inhibitors 3 and 4.
2. Materials and methods
2.1 Computational docking analysis
2.1.1 Molecule and Binding Pocket Preparation for Docking
To prepare molecules for docking we used the Molecular Operating Environment (MOE)
suite’s "molecule wash” function to deprotonate strong acids and protonate strong bases
6

ACCEPTED MANUSCRIPT
(MOE 2013.08; Chemical Computing Group Inc., 1010 Sherbooke St. West, Suite #910,
Montreal, QC, Canada, H3A 2R7, 2015). Energy minimization of all molecules was then
performed by using the MMFF94x force field at a gradient of 0.0001 root-mean-square
deviation (rmsd). Existing chirality was preserved and partial charges were calculated
according to the parameters of the force field.
All crystal structures (3FI3 and 1OUK) [38, 39] were downloaded from the Protein Data Bank
(PDB) [40]. Subsequently, protonation of the protein–ligand complex was performed with the
MOE ‘Protonate 3D’ function at standard settings (T = 300 K, pH = 7.0, ionic strength I = 0.1
mol/L).
2.1.2 Docking Experiments
All docking experiments were performed using the docking program GOLD (GOLD, version
5.2.2; The Cambridge Crystallographic Data Centre, Cambridge, UK, 2013) in combination
with the scoring functions ChemPLP [41] and GoldScore [42] for 3FI3 and the combination of
GoldScore and ChemScore [43] for 1OUK. The search efficiency for the genetic algorithm
was increased from standard 100% to 200% at automatic mode. Binding site residues were
defined by specifying crystal structure ligand coordinates, the active site radius setting
remained at a default value of 6 Å with the „detect cavity“ option enabled. For each target, we
employed a test docking of the ligand from the crystal structure into its binding pocket. If
pose retrieval was unsatisfying, we optimized docking parameters and continued with the
best settings in respect of pose retrieval.
2.2 Probes Synthesis
2.2.1 General
All reagents and solvents were of commercial quality and utilized without further purification.
Thin layer chromatography (TLC) reaction controls were performed for all reactions using
fluorescent silica gel 60 F254 plates (Merck) and visualized under natural light and UV
illumination at 254 and 365 nm. The purity of all synthesized probes and tested compounds
7

ACCEPTED MANUSCRIPT
are >95% as determined via reverse phase high performance liquid chromatography (HPLC)
on a Hewlett Packard HP 1090 Series II LC equipped with a UV diode array detector (DAD,
detection at 230 nm and 254 nm). The chromatographic separation was performed on a
Phenomenex Luna 5u C8 column (150 mm x 4.6 mm, 5 ꢀm) at 35 °C oven temperature. The
injection volume was 5 ꢀL and the flow 1.5 mL / min using the following gradient: 0.01 M
KH2PO4, pH 2.3 (solvent A), methanol (solvent B), 40% B to 85% B in 8 min; 85% B for 5
min; 85% to 40% B in 1 min; 40% B for 2 min; stop time 16 min. Column chromatography
was performed on Davisil LC60A 20 - 45 µm silica from Grace Davison and Geduran Si60
63-200 µm silica from Merck for the pre-column using an Interchim PuriFlash 430 automated
flash chromatography system. Nuclear magnetic resonance (NMR) spectra were measured
on a Bruker Avance NMR spectrometer at 250 MHz in the Organic Chemistry Institute,
Eberhard Karls Universität Tübingen or on a Bruker Avance 400 NMR spectrometer in the
Institute of Pharmaceutical Sciences, Eberhard Karls Universität Tübingen. Chemical shifts
are reported in parts per million (ppm) relative to tetramethylsilane. All spectra were
calibrated against the (residual proton) peak of the deuterated solvent used. Mass spectra
were performed on an Advion Expression S electrospray ionization mass spectrometer (ESI-
MS) with TLC interface in the Institute of Pharmaceutical Sciences, Eberhard Karls
Universität Tübingen.
2.2.2 Experimental procedures
N1-(4-(4-(4-Fluorophenyl)-2-(methylthio)-1H-imidazol-5-yl)pyridin-2-yl)benzene-1,4-
diamine (3)
In a pressure tube compound 8 (500 mg, 1.65 mmol) and p-phenylendiamine (267 mg,
2.47 mmol) were suspended in n-butanol (10 mL) and then 1.25 M HCl in ethanol (1.32 mL,
1.65 mmol) was added. After sealing the tube, the mixture was stirred at 180 °C for 16 h. The
solvent was evaporated at reduced pressure and the residue was purified by flash column
chromatography (dichloromethane – ethanol 19:1 to 93:07) giving 568 mg of product (88%).
1H NMR (400 MHz, DMSO-d6) δ (ppm) 2.61 (s, 3H,), 4.74 (br s, 2H), 6.49 (d, J = 7.1 Hz, 2H),
8

ACCEPTED MANUSCRIPT
6.53 - 6.96 (m, 2H), 7.11 (d, J = 7.1 Hz, 2H), 7.16 - 7.37 (m, 2H), 7.50 (br s, 2H), 7.80 - 8.11
(m, 1H), 8.24 - 8.54 (m, 1H), 12.65 (br s, 1H); HPLC: t = 3.36 min, 98%; ESI-MS: [M+H]+
calculated 392.13, found 392.2.
1-(3',6'-Dihydroxy-3-oxo-3H-spiro[isobenzofuran-1,9'-xanthen]-5-yl)-3-(4-((4-(4-(4-
fluorophenyl)-2-(methylthio)-1H-imidazol-5-yl)pyridin-2-yl)amino)phenyl)thiourea (5)
A suspension of 3 (200 mg, 0.51 mmol) and FITC isomer 5’ (298 mg, 0.76 mmol) in acetone
(60 mL) was stirred at room temperature (r.t.) for 24 h covered from light. The solvent was
evaporated at reduced pressure and the residue was purified by flash column
chromatography (dichloromethane – ethanol 19:1 to 4:1) giving 207 mg of the desired
1
compound (52%). H-NMR (250 MHz, DMSO-d6) δ (ppm) 2.62 (s, 3H), 6.55 - 6.85 (m, 7H),
7.05 - 7.35 (m, 6H), 7.45 - 7.65 (m, 4H), 7.83 (d, J = 8.2 Hz, 1H), 7.95 - 8.15 (m, 1H), 8.20 (br
s, 1H,), 9.10 (br s, 1H), 9.95 - 10.25 (m, 4H), 12.70 (br s, 1H); HPLC: t = 6.19 min, 95%; ESI-
MS: [M-H]- calculated 779.85, found 779.5.
2-Fluoro-4-(4-(4-fluorophenyl)-1H-imidazol-5-yl)pyridine (11)
Compound 10 (2.00 g, 8.09 mmol) was dissolved in DMF (5 mL) and then formaldehyde
(243 mg, 8.09 mmol), ammonium acetate (6.23 g, 80.9 mmol) and glacial acetic acid (47 mg,
0.81 mmol) were added and the mixture was stirred at 70 °C for 1 h. The reaction mixture
was poured into water and the precipitate formed was filtered, washed with water and dried.
The solid was then purified by column chromatography (dichloromethane – ethanol 97:3 to
4:1) yielding 200 mg (10%) of 11.
Spectroscopic data were in agreement with those in the literature [44].
N1-(4-(4-(4-Fluorophenyl)-1H-imidazol-5-yl)pyridin-2-yl)benzene-1,4-diamine (4)
In a pressure tube, compound 11 (150 mg, 0.58 mmol) and p-phenylendiamine (315 mg,
2.91 mmol) were suspended in n-butanol (3 mL) and then 1.25 M HCl in ethanol (0.46 mL,
0.58 mmol) was added. After sealing the tube, the mixture was stirred at 180 °C for 16 h. The
9

ACCEPTED MANUSCRIPT
solvent was evaporated in vacuo and the residue was purified twice by column
chromatography (dichloromethane – ethanol 19:1 to 4:1) giving 124 mg of product (60%).
1H-NMR (250 MHz, DMSO-d6) δ (ppm) 4.75 (br s, 2H), 6.46 (d, J = 8.6 Hz, 2H), 6.60 (dd, J1 =
5.3 Hz, J2 = 1.2 Hz, 1H), 6.77 (br s, 1H), 7.06 (d, J = 8.6 Hz, 2H), 7.20 - 7.30 (m, 2H), 7.45 -
7.55 (m, 2H), 7.80 (s, 1H), 7.95 (br s, 1H), 8.35 (br s, 1H), 12.60 (br s, 1H); ESI-MS: [M+H]+
calculated 346.38, found 346.3; [M-H]- calculated 344.38, found 344.3.
1-(3',6'-Dihydroxy-3-oxo-3H-spiro[isobenzofuran-1,9'-xanthen]-5-yl)-3-(4-((4-(4-(4-
fluorophenyl)-1H-imidazol-5-yl)pyridin-2-yl)amino)phenyl)thiourea (6)
A solution of compound 4 (70 mg, 0.20 mmol) and FITC isomer 5’ (118 mg, 0.30 mmol) in
acetone (20 mL) was stirred at r.t. for 24 h. The reaction mixture was concentrated to
dryness and the residue was purified by column chromatography (dichloromethane – ethanol
1
9:1 to 1:1) to yield 60 mg (40%) of product 6. H-NMR (250 MHz, DMSO-d6) δ (ppm) 6.50 -
6.67 (m, 6H), 6.73 (d, J = 5.5 Hz, 1H), 7.05 - 7.35 (m, 7H), 7.45 - 7.60 (m, 4H), 7.80 - 7.87
(m, 2H), 8.05 (br, 1H), 8.26 (br, 1H), 9.05 (br, 1H), 10.00 - 10.60 (m, 4H), 12.52 (br, 1H);
HPLC: t = 3.77 min, 96%; ESI-MS: [M-H]- calculated 733.76, found 733.4.
2.3 Expression and purification of inactive JNK3 and inactive p38α MAPK
Inactive JNK3 and p38α MAPK were expressed and purified following the procedure
previously reported by Lange et al. [36].
2.4 Activity assay
Probes 5 and 6 as well as compounds 1, 3 and 12-22 were tested for their inhibitory activity
for JNK3 and/or p38α MAPK in previously reported ELISA activity assays [27, 28].
2.5 FP assay
2.5.1 General
All FP measurements were performed in black, non-binding 96-well plates (Greiner Bio-One)
and conducted through a CLARIOstar microplate reader (BMG Labtech GmbH) using
10

ACCEPTED MANUSCRIPT
excitation and emission filters of 480 nm and of 530 nm, respectively. For JNK3
measurements buffer A (25mM HEPES pH 7.0, 100 mM NaCl, 2 mM MgCl2, 10 mM β-
mercaptoethanol (BME) and 0.05 mg/ml bovine serum albumin (BSA)) was used, whereas
buffer B (25 mM Tris pH 7.1, 100 mM NaCl, 10 mM MgCl2, 5 mM DTT, 5% glycerol, 0.05
mg/ml BSA) was employed in the assay performed on p38α MAPK. The concentration of
DMSO used in every assay was 5% (v/v) and the final volume in every well was 200 µL.
Incubation was carried out for the respective time at r.t. on an Eppendorf MixMate at 400
revolutions per minute (rpm) including 2 min inside the plate reader in order to equilibrate at
the measurement temperature (25 °C for JNK3 and 28 °C for p38 α MAPK) which only
deviated marginally from the incubation temperature. Before every measurement, the focal
height of the excitation beam was tuned in order to achieve the optimal fluorescence intensity
and the gain value for both detectors was adjusted. All experiments were performed three
times in quadruplicate. Absolute polarization values were normalized to the average values
of first well and last well replicates and reported as percentage. Data were fitted to a four-
parameter logistic curve with variable slope using the software GraphPad Prism 4 (GraphPad
Prism 4, GraphPad Software Inc., San Diego, CA, USA). Ki values for the competition assays
were calculated from the measured IC50 values using a modified Cheng-Prusoff equation
from the group of Shaomeng Wang [45, 46].
2.5.2 Probes characterization and time stability
A final concentration of 10 nM of probes 5 and 6 was titrated with 3-fold increasing
concentrations of JNK3 and p38α MAPK, ranging from 0 to 600 nM. Concentrations used for
both kinases were 0 nM, 0.01 nM, 0.03 nM, 0.09 nM, 0.27 nM, 0.82 nM, 2.47 nM, 7.40 nM,
22.22 nM, 66.66 nM, 200 nM and 600 nM. Prior to the reading, plates were incubated for a
total time of 15, 30 and 60 min.
2.5.3 Assay conditions optimization
A final concentration of 5 nM of probe 5 was titrated with 3-fold increasing concentrations of
JNK3 and p38α MAPK, ranging from 0 to 300 nM. Concentrations used for both kinases
11

ACCEPTED MANUSCRIPT
were 0 nM, 0.005 nM, 0.015 nM, 0.045 nM, 0.13 nM, 0.41 nM, 1.23 nM, 3.70 nM, 11.11 nM,
33.33 nM, 100 nM and 300 nM. The FP was measured after a total time of incubation of 30
min.
2.5.4 Autofluorescence assay
Prior to the measurement, compounds 1, 3 and 12-23 were tested for their fluorescence
intensity at the excitation and emission wavelengths of 480 nm and 530 nm, respectively,
using a concentration of 1 ꢀM of the test compound in buffer A. Buffer A containing 5%
DMSO (v/v) served as a negative control whereas two different concentrations of probe 5 (5
nM and 1 ꢀM) in the same buffer were used as positive controls.
2.5.5 Competition binding assay
A final concentration of 10 nM of the respective protein kinase and 5 nM of probe 5 were
measured after addition of increasing concentrations of compounds 1, 3, 12-16, 18 and 20-
23 (11 different concentrations including a well without inhibitor). For each measurement, a
well containing a 5 nM concentration of probe 5 in buffer was included as a negative control.
The range of compounds concentrations was selected in order to achieve a full sigmoidal
curve. The FP signal was measured after an incubation time of 30 min. Linear regression on
plotted logarithmic results of the FP assay and the ELISA assay was calculated using
GraphPad Prism 4.
2.5.6 HTS assay validation
The suitability for HTS of the competition binding assays on JNK3 and p38α MAPK was
evaluated following NCBI guidelines manual [47]. The interleaved-signal format described
was employed. For the assays, a 10 nM concentration of the enzyme in the respective assay
buffer and a 5 nM concentration of probe 5 were employed in every well. 5% (v/v) of DMSO
was constantly used in all the assay plates. As a negative control for both enzymes probe 5
alone at the aforementioned concentration was chosen. In the JNK3 experiments, compound
3 at a concentration of 1 ꢀM was employed as a positive control whereas the same
12

ACCEPTED MANUSCRIPT
compound at 20 nM concentration was used as middle control. Regarding the p38α MAPK,
the same compound (3) was employed at concentrations of 2 ꢀM and 50 nM for the positive
and middle control, respectively. The experiment was performed three times in three different
days: controls were plated in a black, non-binding 96-well plate following the scheme
reported in literature [47] and the FP signal was read after 30 min incubation at r.t. The
measurement was carried out at 25 °C and the focus height and the gain values were
adjusted before every measurement. Percent coefficient of variation (%CV) for the negative,
positive and middle controls were calculated for each plate. Each single well signal of middle
control was normalized to the mean of negative and positive controls of the same plate in
order to achieve a percentage of activity. After that, mean and standard deviation (SD) for
the middle control were calculated on the normalized values. Finally, the Z-factor of every
plate was calculated. The single well signals of each day were plotted separately in graphs
“by row, then column” and “by column, then row”, using the software Prism 4, in order to
detect drift or edge effects.
2.6 Isothermal titration calorimetry (ITC)
A reverse ITC was performed due to the lack of solubility of compounds to be tested. 10 mM
solutions of probe 5 and 6 in DMSO were diluted 1:100 with pure DMSO and eventually
further diluted 1:20 with buffer C (50 mM Hepes pH 7.0, 100 mM NaCl, 2 mM MgCl2 and 2
mM tris(2-carboxyethyl)phosphine (TCEP)) for JNK3 and with buffer D (25 mM Tris pH 7.1,
100 mM NaCl, 10 mM MgCl2 and 1 mM BME) for p38α MAPK in order to achieve a 4 µM
concentration of compound with a 5% (v/v) final DMSO concentration. The aforementioned
solutions were loaded into the sample cell of a MicroCal ITC200 (former GE Healthcare, now
Malvern). A freshly concentrated solution of JNK3 or p38α MAPK in buffer C or D,
respectively, containing a 5% (v/v) DMSO concentration, was loaded in the titration syringe.
Measurements were performed at 25 °C. Protein solut ion was titrated into the sample cell in
20 injection steps of 2 µL each, at a rate of 0.5 µL/s with an interval of 120 s between the
steps. Stirring was applied at 1000 rpm. Before the first step, 0.5 µL of protein solution was
13

ACCEPTED MANUSCRIPT
injected into the sample cell, in order to correct for protein diffusion from the syringe during
experimental set up and equilibration. The data point of this pre-step was neglected during
data analysis. Data analysis was performed using MicroCal Origin software (Origin v. 7.0552,
OriginLab Corporation, Northampton MA, USA).
3. Results
3.1 Design, synthesis and characterization of probes 5 and 6
As a starting point for the design of the two probes, the two pyridinylimidazoles 3 and 4 were
selected as precursor molecules (Fig. 2). Both compounds were identified in a current
ongoing research program as potent dual inhibitors of JNK3 and p38α MAPK (Table 1). The
p-phenylendiamino group present in both compounds was chosen for the linking of the
fluorescein tag. Computational docking studies of 3 with both kinases (Fig. S1,
supplementary material) as well as comparison of the binding mode with similar
pyridinylimidazole and pyrimidinylimidazole-based kinase inhibitors [38, 48, 49] revealed that
this portion of the molecule occupies the hydrophobic region II of the enzyme. This is an area
of the ATP binding pocket that is solvent exposed and thus supposed to be able to tolerate
the bulkiness of the fluorescein moiety. In contrast to the previously reported probe 2 by
Munoz et al. (Fig. 1) [37], we chose the more easily available 5’-isomer of FITC as a starting
material to insert the fluorophore. The design hypothesis of our probes was supported by
computational docking studies of compound 5 with both enzymes showing the bulky
fluorescein group ranging outside the enzyme and therefore presumably not hampering the
binding (Fig. S2, supplementary material).
The first probe, compound 5, was synthesized as depicted in Scheme 1 starting from known
2-chloro-4-(4-(4-fluorophenyl)-2-(methylthio)-1H-imidazol-5-yl)pyridine (8). Pyridinylimidazole
8, which is synthetically accessible from the commercially available 2-chloro-4-methylpyridine
(7) in four synthetic steps [50], was subsequently reacted in a nucleophilic aromatic
substitution reaction with p-phenylendiamine to obtain compound 3. The precursor 3 was
14

ACCEPTED MANUSCRIPT
labeled with the fluorophore through the formation of a thiourea group between the primary
amino group and the 5’-isomer of FITC at r.t.
Scheme 1. Synthesis of the probe 5 starting from 2-chloro-4-methylpyridine (7). Reagents
and conditions: (i) p-phenylendiamine, HCl in ethanol 1.25 M, n-butanol, 180 °C, 16 h; (ii)
FITC isomer 5’, acetone, r.t., 24 h; ayield derived from Laufer et al. [50].
The second probe, compound 6 (PIT0105016 in our previously published work [36]) was
prepared starting from already reported ethane-1,2-dione 10, which was synthesized
according to our reported protocol in two steps from 2-fluoro-4-methylpyridine (9) (Scheme 2)
[51]. Cyclization under Radziszewski conditions [52] of dione 10 using ammonium acetate
and formaldehyde provided pyridinylimidazole 11. Finally, the substitution with
p-phenylendiamine as well as the introduction of the fluorophore yielding probe 6 were
performed using the same conditions as described for probe 5.
15

ACCEPTED MANUSCRIPT
Scheme 2. Synthesis of probe 6 starting from 2-fluoro-4-methylpyridine (9). Reagents and
conditions: (i) formaldehyde, NH4OAc, acetic acid, dimethylformamide (DMF), 70 °C, 1 h; (ii)
p-phenylendiamine, HCl in EtOH 1.25 M, n-butanol, 180 °C, 16 h; (iii) FITC isomer 5’,
acetone, r.t., 24 h; ayield derived from Koch et al. [51].
Probes 5 and 6 were also tested for their ability to inhibit JNK3 and p38α MAPK showing only
a negligible decrease in potency in comparison with the unlabeled precursors 3 and 4 (Table
1). This demonstrates the low impact of the fluorescent moiety on the binding affinity,
confirming our design hypothesis as well as the results of the docking studies.
The insertion of a further spacer between the fluorescein moiety and the p-phenylendiamino
group was therefore considered unnecessary and potentially detrimental, since a higher
flexibility of the fluorescein-labelled probe could cause a decrease in the width of the signal
window [53].
Table 1. Biological activity of precursors 3 and 4 and probes 5 and 6 tested in ELISA activity
assays.
16

ACCEPTED MANUSCRIPT
IC50 ± SEMa (nM)
Cpd.
JNK3
24 ± 1
p38α MAPK
3
17 ± 0
7.5 ± 0.3
26 ± 0
4
4.7 ± 0.4
48 ± 2
5
6
74 ± 1
29 ± 2
an = 3.
3.2 Kd determination of probes
With the aim to estimate the binding of the two probes to the inactive form of the protein
kinases, a fixed concentration of 10 nM of probes 5 and 6 was titrated with increasing
concentrations of JNK3 and p38α MAPK (0 – 600 nM). The FP signal was read after
incubation at r.t. for 15, 30 and 60 min, thus allowing also to evaluate the time dependence of
the experiment in terms of equilibrium onset and signal stability. Non-linear fit of the data
obtained in the titration experiment of the two probes (Fig. 3) resulted in very close Kd values
in the low single-digit nanomolar range for both kinases (Table 2). Due to a very similar
binding profile of the two compounds, the selection of probe 5 for the further optimization of
the assays relied mainly on a higher overall yield of the synthetic route (23% in case of probe
5 versus 1% in case of probe 6, compare Scheme 1 vs. Scheme 2). Regarding the time
dependence of the assays, both probes showed in both enzymes only a small variation in Kd
value (≤1.3 nM) between 15 and 60 min. This suggests a rapid onset of the binding
equilibrium of the probes that is already reached after 15 min. Nevertheless, since an
incubation time of 30 min was considered more suitable in binding competition assays of
inhibitors with unknown binding kinetic, the Kd values of the two probes after 30 min of
incubation was selected as the Kd value to be considered in the competition assays.
Table 2. Binding affinities of probes 5 and 6 for JNK3 and p38α MAPK at different times of
incubation tested in the FP assays.
17

ACCEPTED MANUSCRIPT
Kd ± SEMa (nM)
JNK3
p38α MAPK
30 min
Probe
15 min
2.7 ± 0.2
2.3 ± 0.1
30 min
3.0 ± 0.2
2.6 ± 0.1
60 min
4.0 ± 0.3
3.2 ± 0.1
15 min
5.4 ± 1.0
6.2 ± 0.1
60 min
6.2 ± 1.2
7.2 ± 0.3
5
6
5.7 ± 1.1
6.5 ± 0.2
an = 3; concentration of probe used was 10nM; total volume of DMSO was 5% (v/v).
Probe 5
p38α MAPK
JNK3
100
75
50
25
0
100
75
50
25
0
15 min
30 min
60 min
15 min
30 min
60 min
10-11
10-10
10-9
10-8
10-7
10-11
10-10
10-9
10-8
10-7
JNK3 concentration (M)
p38α MAPK concentration (M)
Probe 6
JNK3
p38α MAPK
100
75
50
25
0
100
75
50
25
0
15 min
30 min
60 min
15 min
30 min
60 min
10-11
10-10
10-9
10-8
10-7
10-11
10-10
10-9
10-8
10-7
JNK3 concentration (M)
p38α MAPK concentration (M)
Figure 3. Titration of probes 5 and 6 with inactive JNK3 and inactive p38α MAPK. Probe
concentration was 10 nM in 5% (v/v) DMSO. Data points represent mean values ± SD.
The Kd values of probes 5 and 6 were further determined using ITC as an additional
technique. The Kd values (Table 3 and Fig. S3, supplementary material) obtained by ITC
experiments for both probes are slightly higher compared to those obtained by direct titration
in the corresponding FP assays, probably due to different intrinsic characteristics of the two
18

ACCEPTED MANUSCRIPT
techniques and of the experimental protocols. Nevertheless, the high affinity of the probes for
the two enzymes was again confirmed.
Table 3. Binding affinities of probes 5 and 6 for JNK3 and p38α MAPK determined by ITC.
Kd ± SEMa (nM)
Probe
JNK3
25 ± 5
20 ± 1
p38α MAPK
35 ± 5
5
6
33 ± 3
an = 3; results were obtained through reverse ITC. Protein
concentration in the titration syringe was 55 µM and compound
concentration in the sample cell was 4 µM.
3.3 Assay conditions optimization
After determination of Kd values for probe 5 (Table 2), a concentration of 5 nM was selected
as a final concentration to be used in the competition assays for both enzymes. This is in
accordance with FP assay development guidelines [54], suggesting a probe concentration as
low as possible, preferably not significantly higher than twice the probe Kd value. Using this
probe concentration, a second titration with increasing concentrations of the two enzymes
was performed in order to select the optimal protein concentration (Fig. S4, supplementary
material). In line with practical considerations [54], such concentration should yield an
increase in the polarization signal included between 50 and 80% of the overall signal
window, thus giving a still significant width of signal and at the same time being far from
stoichiometric conditions, in order to provide the assay with a better sensitivity. A
concentration of 10 nM of both enzymes yielding signal windows of ≈140 mP and ≈165 mP
for JNK3 and p38α MAPK, respectively, was therefore selected as the standard
concentration employed in the competition assays.
3.4 Assay validation and correlation
19

ACCEPTED MANUSCRIPT
With the purpose of validating the optimized assays, as well as assessing their viability in
predicting the potency of type I inhibitors in reference activity assays, a series of 13
compounds (Fig. 4) was selected to be evaluated for JNK3 and/or p38α MAPK for their
capability to displace the fluorescein-labelled probe 5. This series consists of different
pyridinylimidazoles from our compound library [55, 56] including the non-labelled precursor of
probe 5 (compound 3), as well as the two p38α MAPK reference compounds SB203580 (1)
and ML3403 (12). In order to demonstrate the full capacity of the probe, examples of non
pyridinylimidazole-type kinase inhibitors were also selected to be investigated. The early
JNK3 inhibitor SP600125 (17) [57], aminopyrimidine 18 [36] and clinical candidate AS601245
(19) [58] were chosen for the assay on JNK3, whereas high quality kinase probe Skepinone-
L (21) [59] as well as clinical candidate SB681323 (22) [60] were tested on p38α MAPK.
Finally, pyridinylquinoxaline 20 [51] was tested on both enzymes. All compounds were
previously evaluated in ELISA activity assays (IC50 values are shown in Table 3) and were
then chosen in order to achieve, within the series, a range of different activities for both
enzymes.
20

ACCEPTED MANUSCRIPT
Fig. 4. Structures of tested compounds.
All selected compounds were primarily tested for autofluorescence at the absorption and
emission wavelengths used for the fluorescein-labelled probe, in order to ensure that the
measured FP signal could be exclusively attributable to the probe. One of the drawbacks of
FP assay indeed, is its unsuitability for the evaluation of compounds showing
autofluorescence at the same wavelengths employed for the fluorophore [53]. For this reason
SP600125 (17) and AS601245 (19), which displayed a fluorescence comparable to the
21

ACCEPTED MANUSCRIPT
positive control (5 nM probe), had to be excluded from the assay (data not shown). In order
to achieve for every measurement a complete displacement of the labelled probe,
compounds were tested at different ranges of concentrations depending on their ability to
bind the two protein kinases. The IC50 values resulting from the competition binding assays
were then converted into Ki values through a modified Cheng-Prusoff equation [45, 46]
(Table 4).
Table 4. Comparison of activity of tested compounds on the ELISA assays and the FP
assays.
JNK3
p38α MAPK
activity assay
Cpd. IC50 ± SEMa [nM]
FP binding assay
Ki ± SEMa [nM]
activity assay
FP binding assay
Ki ± SEMa [nM]
IC50 ± SEMa [nM]
50 ± 0.4c
17 ± 0.4
40 ± 5d
1
727 ± 28b
24 ± 1
167 ± 10
3 ± 0.2
40 ± 6
13 ± 1
49 ± 2
142 ± 18
5 ± 1
56 ± 1
14 ± 2
38 ± 1
4 ± 1
3
12
13
14
15
16
18
20
21
22
176 ± 2
181 ± 4
370 ± 13
742 ± 12
67 ± 2
11 ± 0.9e
53 ± 4
39 ± 9
141 ± 10
61 ± 7
n.t.
170 ± 10
97 ± 8
147 ± 5
3,950 ± 200g
n.t.
29 ± 2
1,041 ± 32
n.t.
n.t.f
81 ± 5g
350 ± 10
4.2 ± 0.4
7 ± 0.1
5 ± 2h
n.t.
n.t.
13 ± 0.1
b
c
an = 3. data previously reported by Goettert et al. [27]; data previously reported by
Goettert et al. [28]; ddata previously reported by Laufer et al. [61]; edata previously reported
f
g
h
by Koch et al. [56]; n.t., not tested; data previously reported by Koch et al. [51] data
previously reported by Fischer et al. [62].
22

ACCEPTED MANUSCRIPT
JNK3
p38α MAPK
100
75
50
25
0
100
Compound 1
Compound 3
Compound 12
Compound 1
75
Compound 3
Compound 12
50
25
0
10-10 10-9 10-8 10-7 10-6 10-5 10-4 10-3
10-10 10-9
10-8
10-7
10-6
10-5
10-4
compound concentration (M)
compound concentration (M)
100
75
50
25
0
100
75
50
25
0
Compound 13
Compound 14
Compound 15
Compound 13
Compound 14
Compound 15
10-10 10-9 10-8 10-7 10-6 10-5 10-4 10-3
10-10 10-9
10-8
10-7
10-6
10-5
10-4
compound concentration (M)
compound concentration (M)
100
75
50
25
0
100
75
50
25
0
Compound 16
Compound 18
Compound 20
Compound 16
Compound 21
Compound 22
10-10 10-9 10-8 10-7 10-6 10-5 10-4 10-3
10-9
10-8
10-7
10-6
10-5
10-4
compound concentration (M)
compound concentration (M)
Fig. 5. Results of competition binding assays for the selected compounds. Probe 5
concentration was 5 nM and the concentration of enzyme was 10 nM. Data points represent
mean values ± SD.
As displayed in Fig. 5, the assays are valid for both enzymes since the unlabeled precursor
of probe 5 (compound 3) as well as other tested inhibitors are able to displace the probe in a
dose-dependent manner. Subsequently to the assay validation, the correlation between the
herein presented FP-based binding competition assays and the ELISA activity assays was
investigated. A correlation between these two assays could make the binding assays a
useful tool for preliminary evaluation of new inhibitors of JNK3 and p38α MAPK in a fast and
23

ACCEPTED MANUSCRIPT
relatively inexpensive manner with the aim to promote only better binders to a further
characterization of the inhibition profile. Correlation was assessed through linear regression
of the plotted logarithmic data for both assay methods (Fig. 6). With regard to JNK3, the
enzyme activity assay and the FP-based competition binding assay proved to correlate
remarkably (R² = 0.9555). In case of p38α MAPK, a good correlation was observed (R² =
0.8447), representing an advantage of our optimized p38α MAPK FP-assay in comparison to
the existing one. These results display the suitability of both presented assays in the
prediction of the inhibitory activity of novel compounds.
JNK3
p38α MAPK
8.0
7.5
7.0
6.5
6.0
5.5
5.0
8.5
8.0
7.5
7.0
6.5
21
3
13
3
18
16
22
12
13
12
14
1
14
1
20
15
16
15
R2 = 0.9555
8.0 8.5 9.0
R2 = 0.8447
8.0 8.5 9.0
20
5.5
6.0
6.5
7.0
7.5
6.0
6.5
7.0
7.5
pKi on FP binding assay
pKi on FP binding assay
Fig. 6. Linear regression of plotted results for the ELISA activity assay and the FP
competition binding assay.
As it is possible to observe from Table 4, the Ki values of the tested compounds are between
low single-digit nanomolar range to low single-digit micromolar range. In order to estimate
the limitation of binding affinity detection, a non-optimized inhibitor, namely 1-(4-
fluorophenyl)-5-(pyridin-4-yl)imidazole (23), was tested as an example of a weak binder in
both optimized FP-assays. Compound 23 displayed Ki values of 2,168 ± 29 nM and 24,030 ±
1,589 nM for JNK3 and p38α MAPK, respectively (Fig. 7). Binders of JNK3 and p38α MAPK
with Ki values up to the low double-digit nanomolar range are detectable using the optimized
FP-assay protocols presented.
24

ACCEPTED MANUSCRIPT
100
75
JNK3
p38α MAPK
50
25
10-9
10-8
10-7
10-6
10-5
10-4
10-3
compound concentration (M)
Fig. 7. Structure and binding affinity of 1-(4-fluorophenyl)-5-(pyridin-4-yl)imidazole (23).
3.5 HTS assay validation
The outcome of the experiments performed in section 3.4 prompted us to exploit the
correlation between the binding and the activity assays on both enzymes for the rapid
estimation of the (potential) inhibitory activity for large compounds sets. One of the
advantages of FP assay is indeed the possibility to measure the signal directly after the
addition of the components without washing or separation steps (mix and measure), which
makes such assays suitable for the HTS format. In order to assess the viability of the FP
assays developed for HTS, we followed the procedure reported in NCBI guidelines [47]. As a
negative control probe 5 alone was selected. As a positive control the unlabeled probe
(compound 3), at a concentration known to displace the probe completely, was added to the
same concentration of probe 5. Finally, compound 3 was used as a middle control at a
concentration close to its IC50 value. Results obtained (Table 5 and 6) were in agreement
with the acceptance criteria described in the guidelines (%CV ≤ 20% for all controls,
normalized SD for the middle control ≤ 20, low inter-plate variability of the normalized middle
signal and Z-factor ≥ 0.4) and the plots of the single well values showed the absence of
significant drift and edge effects (Fig. S5-S8, supplementary material) highlighting the
suitability of these assays for HTS procedure.
25

ACCEPTED MANUSCRIPT
Table 5. HTS assay validation results for JNK3.
negative control
positive control
middle control
Normalized
Z-
factor
a
#
FP value (mP)
FP value (mP)
FP value (mP)
%CV
%CV
%CV
activity
(mean ± SD)
300.3 ± 2.2
300.5 ± 3.8
298.0 ± 3.8
(mean ± SD)
161.7 ± 2.8
151.3 ± 3.3
143.1 ± 3.2
(mean ± SD)
220.8 ± 4.0
210.7 ± 5.3
195.1 ± 5.0
(mean ± SD)
1
0.7%
1.2%
1.2%
1.7%
2.2%
2.2%
1.8% 42.6% ± 2.9%
2.5% 39.8% ± 3.5%
2.5% 33.5% ± 3.2%
0.89
0.86
0.86
2
3
aexperiment number.
Table 6. HTS assay validation results for p38α MAPK.
negative control
positive control
middle control
Z-
factor
a
Normalized
activity
(mean ± SD)
#
FP value (mP)
FP value (mP)
FP value (mP)
(mean ± SD)
%CV
%CV
%CV
(mean ± SD)
296.1 ± 2.7
304.9 ± 4.1
291.9 ± 6.9
(mean ± SD)
122.4 ± 4.3
134.9 ± 3.5
124.9 ± 4.6
1
0.9%
1.3%
2.3%
3.5%
2.6%
3.6%
256.9 ± 3.4
226.9 ± 2.9
214.6 ± 4.7
1.3% 77.4% ± 2.0%
1.3% 54.1% ± 1.7%
2.2% 53.7% ± 2.8%
0.88
0.86
0.79
2
3
aexperiment number.
4. Discussion
The research for novel inhibitors of the two protein kinases JNK3 and p38α MAPK is still
ongoing and aimed to the discovery of molecules displaying a high efficacy together with a
satisfactory selectivity profile. A key approach might be represented by the identification of
molecules, which bind preferentially the inactive conformation of these enzymes (type II
inhibitors) taking advantage of binding pockets that are less conserved within the family of
closely related kinases. With this scope, we developed two FP-based competition binding
assays employing the inactive form of the two kinases able to measure the binding affinity of
both type I and type II inhibitors. Classical activity assays using the active kinase exclude
type II inhibitors predominantly binding to the inactive enzyme. Extending a previously
reported assay for p38α MAPK, the herein presented assay broadens the range of
26

ACCEPTED MANUSCRIPT
applicability including JNK3 due to the use of a dual JNK3/p38α MAPK inhibitor as a
precursor of the probe employed. In addition, the comparison of the FP-assays presented
with well-known ELISA activity assays allowed to observe a correlation between the two
assays performed on both protein kinases. These results point out the potential of the assays
presented as rapid and relatively inexpensive (low concentrations of kinase needed, no
costly kinase substrates and antibodies required) alternatives to activity assays, which can
be used to predict the inhibitory activity of new compounds in an early stage. Finally, the
suitability of these assays to the HTS format allows the screening of wide libraries of
compounds and therefore represents an improvement in the research of new JNK3 and p38α
MAPK inhibitors.
5. Conclusion
The synthesis of two novel fluorescein-based pyridinylimidazoles and their evaluation for
their ability to act as probes for the determination of binding affinity of kinase inhibitors to
JNK3 and p38α MAPK is presented. Fast and inexpensive FP-assays for both enzymes
using probe 5 (JNK3: Kd = 3.0 nM; p38α MAPK: Kd = 5.7 nM) were developed and validated
with known inhibitors of JNK3 and p38α MAPK. The comparison of the results obtained from
the FP-assays with the results of an ELISA-based activity assays revealed a good to
excellent correlation, which makes the p38α MAPK FP-assay superior than the reported one.
Moreover, both FP-assays are easily adaptable to the HTS procedure and therefore suitable
as inexpensive pre-screening protocols for JNK3 and p38α MAPK inhibitors.
Acknowledgement
The authors thank Katharina Bauer and Daniela Müller who carried out the testing of the
compounds in the activity ELISA assays. We are also grateful to Prof. Dr. Stefan Laufer for
providing samples of SP600125, compound 18, AS601245, Skepinone-L and SB681323. We
thank Tamer Kocakaya and Philip Krause for synthesizing compounds 16 and 23,
respectively. Miriam Gisevius, Regina Jedig and Mark Kudolo (all students at the Institute of
27

ACCEPTED MANUSCRIPT
Pharmaceutical Sciences, Eberhard Karls Universität Tübingen) are gratefully acknowledged
for their assistance in the FP assay. Special thanks to Silke Bauer for helpful discussions.
Appendix A. Supplementary material
Supplementary material related to this article can be found at http://dx.doi.org...... .
References
[1] V. Perrin, N. Dufour, C. Raoul, R. Hassig, E. Brouillet, P. Aebischer, R. Luthi-Carter, N.
Deglon, Implication of the JNK pathway in a rat model of Huntington's disease, Exp. Neurol.,
215 (2009) 191-200.
[2] L. Resnick, M. Fennell, Targeting JNK3 for the treatment of neurodegenerative disorders,
Drug Discov. Today, 9 (2004) 932-939.
[3] S. Kumar, S.M. Blake, J.G. Emery, Intracellular signaling pathways as a target for the
treatment of rheumatoid arthritis, Curr. Opin. Pharmacol., 1 (2001) 307-313.
[4] Y.J. Feng, Y.Y. Li, The role of p38 mitogen-activated protein kinase in the pathogenesis of
inflammatory bowel disease, J. Dig. Dis., 12 (2011) 327-332.
[5] A. Mavropoulos, E.I. Rigopoulou, C. Liaskos, D.P. Bogdanos, L.I. Sakkas, The role of p38
MAPK in the aetiopathogenesis of psoriasis and psoriatic arthritis, Clin. Dev. Immunol., 2013
(2013) 569751.
[6] J.C. Betts, R.J. Mayer, R. Tal-Singer, L. Warnock, C. Clayton, S. Bates, B.E. Hoffman, C.
Larminie, D. Singh, Gene expression changes caused by the p38 MAPK inhibitor dilmapimod
in COPD patients: analysis of blood and sputum samples from a randomized, placebo-
controlled clinical trial, Pharmacol. Res. Perspect., 3 (2015) e00094.
[7] W. MacNee, R.J. Allan, I. Jones, M.C. De Salvo, L.F. Tan, Efficacy and safety of the oral
p38 inhibitor PH-797804 in chronic obstructive pulmonary disease: a randomised clinical trial,
Thorax, 68 (2013) 738-745.
28

ACCEPTED MANUSCRIPT
[8] S.A. Correa, K.L. Eales, The Role of p38 MAPK and Its Substrates in Neuronal Plasticity
and Neurodegenerative Disease, J. Signal Transduct., 2012 (2012) 649079.
[9] D.N. Krementsov, T.M. Thornton, C. Teuscher, M. Rincon, The emerging role of p38
mitogen-activated protein kinase in multiple sclerosis and its models, Mol. Cell. Biol., 33
(2013) 3728-3734.
[10] L. Munoz, A.J. Ammit, Targeting p38 MAPK pathway for the treatment of Alzheimer's
disease, Neuropharmacology, 58 (2010) 561-568.
[11] P. Koch, M. Gehringer, S.A. Laufer, Inhibitors of c-Jun N-terminal kinases: an update, J.
Med. Chem., 58 (2015) 72-95.
[12] M. Gehringer, F. Muth, P. Koch, S.A. Laufer, c-Jun N-terminal kinase inhibitors: a patent
review (2010-2014), Expert Opin. Ther. Pat., 25 (2015) 849-872.
[13] S. Bühler, S.A. Laufer, P38 MAPK inhibitors: a patent review (2012-2013), Expert Opin.
Ther. Pat., 24 (2014) 535-554.
[14] A.M. Manning, R.J. Davis, Targeting JNK for therapeutic benefit: from junk to gold?, Nat.
Rev. Drug Discov., 2 (2003) 554-565.
[15] A. Cuenda, S. Rousseau, p38 MAP-kinases pathway regulation, function and role in
human diseases, Biochim. Biophys. Acta, 1773 (2007) 1358-1375.
[16] B. Kaminska, MAPK signalling pathways as molecular targets for anti-inflammatory
therapy--from molecular mechanisms to therapeutic benefits, Biochim. Biophys. Acta, 1754
(2005) 253-262.
[17] S. Kumar, J. Boehm, J.C. Lee, p38 MAP kinases: key signalling molecules as
therapeutic targets for inflammatory diseases, Nat. Rev. Drug Discov., 2 (2003) 717-726.
[18] R.J. Davis, Signal transduction by the JNK group of MAP kinases, Cell, 103 (2000) 239-
252.
[19] Y. Liu, N.S. Gray, Rational design of inhibitors that bind to inactive kinase conformations,
Nat. Chem. Biol., 2 (2006) 358-364.
29