Design, Synthesis, and Biological Evaluation of Quercetagetin Analogues as JNK1 Inhibitors

Article abstract of DOI:10.1002/chem.201502475

The recent discovery of c-Jun NH₂-terminal kinase JNK1 suppression by natural quercetagetin (1) is a promising lead for the development of novel anticancer agents. Using both X-ray structure and docking analyses we predicted that 5′-hydroxy- (2

Full text of DOI:10.1002/chem.201502475

DOI: 10.1002/chem.201502475
Full Paper
&
Kinase Inhibitors |Hot Paper|
Design, Synthesis, and Biological Evaluation of Quercetagetin
Analogues as JNK1 Inhibitors
Judith Hierold,[a] Sohee Baek ,[b, c, d] Rene Rieger,[a] Tae-Gyu Lim,[c] Saman Zakpur,[a]
Marcelino Arciniega,[b] Ki Won Lee,[c, e] Robert Huber ,*[b, f, g, h] and Lutz F. Tietze*[a]
Abstract: The recent discovery of c-Jun NH2-terminal kinase
JNK1 suppression by natural quercetagetin (1) is a promising
lead for the development of novel anticancer agents. Using
both X-ray structure and docking analyses we predicted that
5’-hydroxy- (2) and 5’-hydroxymethyl-quercetagetin (3)
would inhibit JNK1 more actively than the parent compound
1. Notably, our drug design was based on the active
enzyme–ligand complex as opposed to the enzyme’s
relatively open apo structure. In this paper we test our theo-
retical predictions, aided by docking-model experiments,
and report the first synthesis and biological evaluation of
quercetagetin analogues 2 and 3. As calculated, both
compounds strongly suppress JNK1 activity. The IC50 values
were determined to be 3.4 mM and 12.2 mM, respectively,
which shows that 2 surpasses the potency of the parent
compound 1 (IC50 =4.6 mM). Compound 2 was also shown
to suppress matrix metalloproteinase-1 expression with high
specificity after UV irradiation.
Introduction
have broad tissue distribution and are strongly activated by
various inflammatory signals and stress stimuli. Although some
debate exists regarding the exact role of JNKs in cancer, ex-
pression of JNK proteins is frequently altered in human tumors
and cancer cells, including liver,[2] prostate,[3] and breast
cancer.[4,5] In addition, JNKs are crucial mediators of obesity
and insulin resistance, thus they are potential targets in type II
diabetes.[6] Therefore, inhibition of JNKs is of considerable
pharmaceutical relevance for cancer and other chronic dis-
eases. Furthermore, matrix metalloproteinase-1 (MMP-1)—an
interstitial collagenase that degrades extracellular matrix com-
ponents[7]—is regulated by JNK signaling pathways[8] by modu-
lating its expression through AP-1 (activator protein) transacti-
vation.[9] JNK-activated AP-1 directly binds to human MMP-1
promoter and enhances MMP-1 transcription.[9] MMP-1 is an
important player in UV-induced skin aging, thus it is a relevant
pharmacological target and its activity may be suppressed by
JNK inhibitors.
The c-Jun NH2-terminal kinases (JNKs) are serine/threonine pro-
tein kinases belonging to the mitogen-activated protein kinase
family,[1] and they play a critical role in chronic diseases. JNKs
[a] Dr. J. Hierold, R. Rieger, S. Zakpur, Prof. L. F. Tietze
Institute for Organic and Biomolecular Chemistry
Georg-August-University Gçttingen
Tammannstrasse 2, 37077 Gçttingen (Germany)
Fax: (+49)551-39-9476
E-mail: ltietze@gwdg.de
[b] Dr. S. Baek , Dr. M. Arciniega, Prof. R. Huber
Max-Planck-Institute for Biochemistry, Am Klopferspitz 18
82152 Martinsried (Germany)
[c] Dr. S. Baek , Dr. T.-G. Lim, Prof. K. W. Lee
Advanced Institutes of Convergence Technology, Seoul National University
Suwon 443-270 (Republic of Korea)
[d] Dr. S. Baek
Proteros Biostructures GmbH
Bunsenstr. 7a, 82152 Martinsried (Germany)
In a study aimed at identifying a novel small-molecule inhib-
itor of JNKs, we recently reported that the natural flavonoid
quercetagetin[10] (1; Figure 1) strongly suppresses JNK1 activi-
ty.[11] Flavonoids are secondary plant metabolites showing
a wide range of biological activities, including anti-inflammato-
ry,[12] antiallergy, antifungal, antiviral, and antibacterial activi-
ties.[13] Consequently, both natural and synthetic flavonoids are
frequently encountered in drug discovery.[14–20] In our hands,
1 inhibited JNK1 in an in vitro immobilized-metal-ion affinity-
based fluorescence polarization (IMAP) assay with an IC50 value
of 4.6 mM (Figure 2a).[11] Thus, it showed a higher activity than
the pharmacological JNK1 inhibitor SP600125 (IC50 =5.2 mM)
(Figure 2b). Competitive-binding studies indicated that 1 occu-
pies the ATP-binding site of JNK1 because the ATP concentra-
tion affected the level of inhibition exhibited by 1. The func-
[e] Prof. K. W. Lee
WCU Biomodulation Major, Center for Food and Bioconvergence
Department of Agricultural Biotechnology, Seoul National University
Seoul (Republic of Korea)
[f] Prof. R. Huber
Department of Chemistry, Technical University of Munich
Lichtenbergstraße 4, 85748 Garching (Germany)
[g] Prof. R. Huber
Center for Medical Biotechnology
University of Duisburg-Essen, 45117 Essen (Germany)
[h] Prof. R. Huber
School of Biosciences, Cardiff University, Cardiff (Wales, UK)
Supporting information for this article is available on the WWW under
http://dx.doi.org/10.1002/chem.201502475. It contains NMR spectra of all
new compounds as well as a more detailed synthesis and characterization
section.
Chem. Eur. J. 2015, 21, 16887 – 16894
16887
ꢀ 2015 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Full Paper
Figure 1. Verified and predicted small-molecule JNK1 inhibitors.
tional significance of the binding of 1 to JNK1 was examined
in cell-based systems. The flavonoid was shown to strongly
suppress UVB-induced phosphorylation of c-Jun, AKT, and
GSK3b, as well as UVB-induced trans-activation of AP-1 and nu-
clear factor NF-kB. At a concentration of 5 mM, H-Ras-induced
neoplastic cell transformation was inhibited by 73%. A two-
stage mouse-skin-tumorgenesis model served to quantify the
pharmacological effects. The flavonoid reduced tumor inci-
dence in hairless mice by 46.7% following topical application
at 20 nmol concentration, which indicated that 1 might be
a suitable chemopreventive agent against UVB-mediated skin
cancer.
The molecular basis of the inhibition of JNK1 by 1 was char-
acterized by crystal-structure analysis of the JNK1–pepJIP1–
1 ternary complex (Figure 3).[11,21] Compound 1 is located in
the ATP-binding site and forms a number of stabilizing hydro-
gen bonds with the protein. The binding of 1 results in signifi-
cant structural changes of the apo-protein form through
a hinge-bending motion.[11,22,23] The whole N-terminal lobe
region is rotated towards the C-terminal lobe, which not only
narrows, but also effectively caps the binding site after docking
(Figure 3a). This feature improves compatibility of the ligand
with the binding site of JNK1; compound 1 reaches much
deeper into the ATP-binding site than SP600125.
A network of hydrogen bonds (Figure 3b) connects the cate-
chol moiety with the side chains Lys55, Asp169, and Glu73,
whereas the chromone portion forms hydrogen bonds with
Glu109 and Met111.[24] Additional hydrophobic interactions be-
tween the ligand and the binding cleft further stabilize the
structure. Notably, a water molecule trapped below the ligand
participates in the hydrogen-bonding network by connecting
the catechol moiety of 1 with Asp169 and Glu73. Its presence
suggests that replacement of the trapped water molecules by
an additional polar substituent on 1 might further increase the
ligand selectivity and affinity. To test this hypothesis, we mod-
eled the binding structure and affinity of 1 and appropriately
modified derivatives by using different docking analyses.[11] The
modeled docking poses of 5’-hydroxy- (2) and 5’-hydroxy-
methyl-quercetagetin (3) present an extra interaction with
Asp169Od compared with the experimental pose of 1 (Fig-
ure 3c and d versus Figure 3b). Additionally, 3 is predicted to
interact directly with Asp169N and Glu73Oe, in other words,
Figure 2. Inhibition of JNK1 (and the corresponding IC50 values) by A) 1,
B) SP600125, C) 2, and D) 3 determined in an IMAP assay. The assay was per-
formed as described in the Experimental Section. Each experiment was
performed in triplicate. The data points for the titrations of 1 and 2 above
100 mm are affected by limited solubility and are not included in the
calculation and curve fitting.
without mediation by the water molecule (Figure 3d). The pre-
dicted binding of 3 suggested that key interactions can be
made without coordination of the water molecule, although
the hydroxymethyl group does not occupy exactly the same
position.
The results of a competitive-docking analysis are represent-
ed as docking scores (Table 1),[11] which are semiquantitative
Chem. Eur. J. 2015, 21, 16887 – 16894
www.chemeurj.org
16888
ꢀ 2015 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Full Paper
Results and Discussion
The synthesis of 2 commenced
with a Claisen–Schmidt conden-
sation of commercially available
phenol 4 and benzaldehyde 5a
(Scheme 1) to give chalcone 6a
in 44% yield [76% based on
recovered
starting
material
(brsm)]. An intramolecular oxa-
Michael addition yielded chro-
manone 7a, which was subse-
quently oxidized to flavonol 9a
via the corresponding ketoxime.
Compound 7a was treated with
isoamyl nitrite under acidic con-
ditions to give ketoxime 8a,
which was hydrolyzed with sul-
furic acid in acetic acid at reflux
temperature to give to the de-
sired flavonol 9a in 68% over
two steps. The traditional Algar–
Flynn–Oyamada approach failed
to yield the flavonol directly
from chalcone 6a, instead a com-
Figure 3. A) Crystal structure of the ternary JNK1(olive)–pepJP1(orange)–quercetagetin(blue) complex.[11] Compari-
son of the hydrogen-bond network between the B) experimental pose of 1 and the predicted binding poses of
C) 2 and D) 3. Hydrogen bonds are depicted as blue dashed lines. In the model of 3 the crystallographic water
molecule is shown as a semitransparent sphere. It is unlikely to be present at this position and has not been
considered in the modeling due to steric clashes (red dashed line, 1.8 Š).
peting
a-addition
provided
aurone 10 in 13% yield. Global
deprotection of 9a with boron
tribromide smoothly yielded
Table 1. Evaluation of the binding affinity to JNK1 by docking analysis.[11]
Entry
Compound
H2O mimic
Docking score[a]
1
2
3
4
SP600125[b]
1
2
3
–
–
À12.1
À12.5
À13.2
À14.1
ÀOH
ÀCH2OH
[a] Docking score after incorporation into the conformation that JNK1
adopts after docking of 1 (activated form); [b] SP600125 is a pharmaco-
logical JNK1 inhibitor.
measures of the binding energy:[25,26] the lower the value, the
stronger the binding affinity. In agreement with the higher bio-
logical activity of 1, the docking score for 1 was lower than
the score for SP600125 (À12.5 and À12.1, respectively). The
calculations further indicated an improvement of the affinity
for 2 and 3 with docking scores of À13.2 and À14.1,
respectively.
In this paper we describe the first synthesis and biological
evaluation of quercetagetin analogues 2 and 3. As calculated,
both compounds strongly inhibit JNK1. Gratifyingly,
2
Scheme 1. Synthesis of 2: a) NaOEt, EtOH, RT, 20 h, 44% (76% brsm);
b) H2O2, NaOH (aq), MeOH, 08C to RT, 24 h, 13%; c) AcOH, 1208C, 72 h, 54%
(72% brsm); d) isoamyl nitrite, HCl, EtOH, 808C, 15 min; e) H2SO4, AcOH,
110 8C, 30 min, 68% (2 steps); f) BBr3, CH2Cl2, 08C to RT, 16 h, then MeOH,
658C, 2 h, 99%; g) Ac2O, cat. H2SO4, 08C to RT, 3 h, 98%; h) 6m HCl, CH3CN,
858C, 90 min, quant.
surpasses the activity of the parent compound 1 with an IC50
value of 3.4 mM.
Chem. Eur. J. 2015, 21, 16887 – 16894
www.chemeurj.org
16889
ꢀ 2015 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Full Paper
target compound 2. Overall, 2 was prepared in 16% yield
(37% brsm) over five consecutive steps.
The high polarity of 2 coupled with potential tautomerism
to the 1,2-diketone impeded product purification by traditional
methods. Whereas small quantities could be obtained by crys-
tallization from ethyl acetate, on a larger scale it was more
practical to purify the fully acetylated compound 11. When 11
was exposed to hydrochloric acid in acetonitrile at reflux
temperature the temporary acetyl groups were fully removed
and highly pure 2 precipitated from the reaction mixture.
Benzaldehyde 5b, required to access 2, was prepared from
commercially available 2,3-dihydroxybenzoic acid (12), which
was regioselectively brominated, methylated, and reduced to
give (5-bromo-2,3-dimethoxyphenyl)methanol (14) in 75%
yield over three steps (Scheme 2). Compound 14 was subse-
Scheme 3. Synthesis of 3: a) NaOEt, EtOH, RT, 20 h, 73%; b) AcOH, 1208C,
72 h, 56% (84% brsm); c) isoamyl nitrite, HCl, EtOH, 808C, 15 min, then
H2SO4, AcOH, 110 8C, 30 min, 20%; d) BBr3, CH2Cl2, 08C to RT, 16 h, then H2O,
1008C, 2 h, 97% (ꢀ90% purity).
With both target molecules in hand we evaluated their bio-
logical performance. Theoretical modeling and docking-score
calculations indicated an improvement of the affinity of 2 and
3 with docking scores of À13.2 and À14.1, respectively, com-
pared to À12.5 for 1. The IC50 values were calculated by using
the IMAP assay system.[27] The IC50 values for 2 and 3 were
determined to be 3.4 mM (Figure 2c) and 12.2 mM (Figure 2d),
respectively.[28]
The obtained values indicate strong inhibition of JNK1, with
2 surpassing the inhibition of both the natural product
(4.6 mM) and the pharmacological JNK1 inhibitor SP600125
(5.2 mM). Thus, our predictions based on the docking analysis
held true for 2, whereas hydroxymethyl-substituted inhibitor 3
proved to be less effective than predicted. These findings
show that although theoretical calculations are indeed helpful,
they must always be assessed by laboratory experimentation.
The docking scores were very similar throughout the flavonoid
series and a detailed ranking based solely on the scores must
be viewed cautiously. The lower activity exhibited by 3 might
indicate that the water molecule thought to be replaced by
the 5’-hydroxymethyl substituent may play a more-specific role
than expected.
Scheme 2. Synthesis of 5b: a) Br2, AcOH, RT, 16 h, 99%; b) Me2SO4, K2CO3,
acetone, 608C, 20 h, then NaOH (aq), MeOH, 708C, 2 h, 89%; c) BH3·Me2S,
THF, RT, 16 h, 85%; d) TBSCl, imidazole, DMF, RT, 16 h, 89%; e) (vinyl)Sn(nBu)3,
[Pd2(dba)3] (1.5 mol%), P(tBu)3 (6.0 mol%), CsF, dioxane, 1028C, 16 h, 86%;
f) OsO4, NaIO4, THF, H2O, RT, 1 h, 62%; g) tetrabutylammonium fluoride
(TBAF), THF, RT, 16 h, 98%.
quently silyl protected at the benzylic hydroxy group, which
was followed by Stille cross-coupling with palladium tris(diben-
zylideneacetone)dipalladium(0) ([Pd2(dba)3]), P(tBu)3, cesium
fluoride, and tributyl(vinyl)tin. The resultant styrene 15 then
underwent a Lemieux–Johnson oxidation (osmium tetroxide
and sodium periodate) to give the corresponding benzalde-
hyde. To ensure a high yield in the subsequent aldol condensa-
tion with phenol 4, the bulky silyl protecting group was
removed under standard conditions to yield benzaldehyde 5b.
Claisen–Schmidt condensation with phenol 4 resulted in
chalcone formation in 73% yield (Scheme 3). Intramolecular
oxa-Michael addition with concomitant acetylation of the ben-
zylic alcohol furnished chromanone 7b, which was oxidized
with isoamyl nitrite as detailed above. Lewis acid cleavage of
the ether CÀO bonds furnished the intermediate alkoxyborane,
which was hydrolyzed in water at reflux temperature to yield
the target compound 3. Performing the hydrolysis in metha-
nol, which was successful for 9a, led to methylation of the
benzylic hydroxy group in the case of 9b, therefore this
method was unsuitable for formation of 3. Purification of the
flavonoid again proved challenging, and only approximately
Because 2 performed satisfactorily in the IMAP assay we fur-
ther evaluated its biological profile. JNK1 plays a critical role
for UV-induced MMP-1 expression by activating AP-1, a crucial
transcription factor of MMP-1.[29] Thus, we investigated the ac-
tivity of 2 on solar UV light (sUV) induced MMP-1 expression
relative to the parent compound 1. At the concentrations in-
vestigated in this study neither compound showed any signifi-
cant cytotoxicity (Figure 4a). In line with previous reports,[30]
sUV enhances MMP-1 expression in normal human dermal fi-
broblasts (NHDFs). Interestingly, our studies showed that 2
dramatically suppressed sUV-induced MMP-1 expression with
a much-stronger activity than 1, which had almost no effect in
comparison (Figure 4b). However, because the JNK1 binding
affinities of 1 and 2 are very similar other components of the
JNK signaling pathway may play an essential role, which
demands further study.
90% purity of
processes.
3 was reached after repeated washing
Chem. Eur. J. 2015, 21, 16887 – 16894
www.chemeurj.org
16890
ꢀ 2015 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Full Paper
Experimental Section
Compound 2
BBr3 (1m in CH2Cl2, 3.0 mL, 3.0 mmol, 15 equiv) was added via
syringe pump over 15 min to a mechanically stirred solution of 9a
(84 mg, 0.20 mmol, 1.0 equiv) in dry CH2Cl2 (10 mL) at 08C. The re-
sultant clear, orange solution was warmed to RT and stirred for
16 h during which time an orange solid precipitated. The reaction
mixture was cooled to 08C and dry MeOH (10 mL) was added
dropwise. The resultant clear solution was heated at 658C for 2 h,
then cooled to RT. The volatile compounds were removed under
reduced pressure. The resultant orange solid was suspended in
H2O (10 mL) and sonicated for 20 min. The suspension was left
standing overnight and then carefully decanted. Recrystallization
of the residue from EtOAc yielded 2 as a mustard-yellow solid
1
(66 mg, 99%). H NMR (600 MHz, [D6]DMSO): d=6.46 (s, 1H; 8-H),
7.20 (s, 2H; 2’-H), 8.68 (brs, 2H; OH), 9.11 (brs, 3H; OH), 10.41 (brs,
1H; OH), 12.23 ppm (s, 1H; 3-OH); 13C NMR (125 MHz, [D6]DMSO):
d=93.01 (C-8), 103.20, 107.09 (C-2’), 121.00, 128.45, 135.36, 135.66,
145.64 (C-3’), 145.84, 146.62, 148.73, 153.51, 175.68 ppm (C-4); IR:
n˜ =3351, 1666, 1607, 1561, 1538, 1504, 1474, 1312, 1274, 1207,
1167, 1100, 1045, 1023, 979, 821 cmÀ1; UV/Vis (DMSO): lmax (loge)=
378 nm (4.2098); MS (ESI+): m/z (%): 335.0 (25) [M+H]+, 357.0 (21)
[M+Na]+; MS (ESIÀ): m/z (%):333.0 (100) [MÀH]À; HRMS (ESIÀ):
m/z calcd for C15H10O9: 333.0252 [MÀH]À; found: 333.0249.
Figure 4. A) Inhibitors 2 and 1 show no significant cytotoxicity in NHDFs.
The cells were cultured to confluence in 96-well plates, then treated with
inhibitor for 24 h. Cytotoxicity was analyzed by using an MTS solution as
indicated in the Experimental Section. Data are representative of three
independent experiments that gave similar results. B) Inhibitor 2 shows
a much-stronger inhibitory effect than 1 on sUV-induced MMP-1 expression.
The cells were chemically treated for 1 h, then irradiated with sUV . After
36 h of exposure, the cells were disrupted with a lysis buffer. The MMP-1
protein was evaluated by using Western blot analysis as described in the
Experimental Section. Data are representative of three independent experi-
ments that gave similar results.
Compound 3
BBr3 (1m in CH2Cl2, 489 mL, 489 mmol, 15 equiv) was added via
syringe pump over 15 min to a mechanically stirred solution of 9b
(15.0 mg, 32.6 mmol, 1.0 equiv) in dry CH2Cl2 (2 mL) at 08C. The re-
sultant light-brown solution was warmed to RT and stirred for 16 h.
The reaction mixture was cooled to 08C and H2O (2 mL) was
added dropwise. The resultant clear solution was heated at 808C
for 2 h and then cooled to RT. The volatile compounds were re-
moved under reduced pressure. The resultant brown solid was sus-
pended in H2O (10 mL) and sonicated for 20 min. The suspension
was left to stand overnight and then carefully decanted to yield 3
as a beige solid (11 mg, 97%, purityꢀ90%). 1H NMR (300 MHz,
CD3OD): d=4.73 (s, 2H; CH2), 6.51 (s, 1H; 8-H), 7.73 ppm (s, 2H; 2’-
H and 6’-H); MS (ESI+): m/z (%): 349.1 (3) [M+H]+, 371.0 (8)
[M+Na]+; HRMS (ESIÀ): m/z calcd for C16H12O9: 347.0409 [MÀH]À;
found: 347.0420.
Conclusion
The development of clinically useful small-molecule kinase
inhibitors has been a seminal event in the world of chronic
diseases. Thus, the discovery of JNK1 suppression by natural
product 1 was a promising lead in the development of novel
anticancer agents. Based on X-ray structure and docking analy-
ses we predicted that 2 and 3 would inhibit JNK1 more active-
ly than the parent compound 1. In this paper we evaluated
our theoretical predictions and reported the first synthesis of
derivatives 2 and 3. Both analogues suppress JNK1 activity:
compound 2 (IC50 =3.4 mM) surpassed the potency of 1 (IC50 =
4.6 mM). Conversely, 3 showed a comparatively high IC50 value
(12.2 mM), despite having the lowest docking score of the
series in the theoretical calculations. The lower affinity of 3 in
the experimental binding assay suggests that the water
molecule present in the substrate–JNK1 complex may play a
more specific structural role than expected. Cell biological
experimentation showed that 2 exhibits very high specificity in
suppressing MMP-1 expression after UV irradiation, and thus
could impact skin aging.
Compound 9a
Conc. HCl (2 mL) was added via syringe pump over 10 min to a
mechanically stirred solution of 7a (781 mg, 1.9 mmol, 1.0 equiv)
and isoamyl nitrite (1.56 mL, 11.6 mmol, 6.0 equiv) in dry ethanol
(70 mL) at 08C. The reaction was stirred at 808C for 15 min. After
cooling to RT the volatile compounds were evaporated under re-
duced pressure. The residue was taken up in EtOAc (60 mL) and
washed with H2O (2”50 mL). The combined aqueous layers were
extracted with EtOAc (2”50 mL) and CHCl3 (2”50 mL). The com-
bined organic layers were washed with brine (80 mL), dried over
MgSO4, and concentrated under reduced pressure to yield the in-
termediate oxime as a yellow solid. The oxime was dissolved in
glacial acetic acid (80 mL), cooled to 08C, and treated with 10%
aqueous H2SO4 (16 mL). The reaction was stirred at 110 8C for
30 min, then cooled to RT, and concentrated under reduced pres-
sure. The residue was taken up in EtOAc (70 mL) and washed with
H2O (2”50 mL). The combined aqueous layers were extracted with
EtOAc (2”50 mL) and CHCl3 (2”50 mL). The combined organic
Chem. Eur. J. 2015, 21, 16887 – 16894
www.chemeurj.org
16891
ꢀ 2015 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Full Paper
layers were washed with brine (100 mL) and dried over MgSO4.
Evaporation of the volatile compounds gave the crude product,
which was first purified by flash column chromatography (MeOH/
CH2Cl2 =1:19). The resultant highly fluorescent beige solid was sus-
pended in EtOAc (5 mL) and carefully decanted to leave pure 9a
(543 mg, 68%) as a pale-beige solid. Rf =0.41 (MeOH/CH2Cl2 =
209 (4.2554), 223 (4.2263), 270 (3.9929), 308 nm (3.6603); MS
(ESI+): m/z (%): 197.1 (100) [M+H]+, 219.1 (91) [M+Na]+; HRMS
(ESI+): m/z calcd for C10H12O4: 197.0808 [M+H]+; found: 197.0808;
HRMS (ESI+): m/z calcd for C10H12O4Na: 219.0628 [M+Na]+; found:
219.0628.
1
1:19); H NMR (600 MHz, CDCl3): d=3.91 (s, 6H; 2”OMe), 3.95 (s,
Compound 6b
6H; 2”OMe), 3.99 (s, 3H; OMe), 4.02 (s, 3H; OMe), 6.76 (s, 1H; 8-
H), 7.39 (brs, 1H; 3-OH), 7.46 ppm (s, 2H; 2’-H); 13C NMR (125 MHz,
CDCl3): d=56.4, 56.5, 61.0, 61.5, 62.2, 96.0, 105.1, 109.8, 126.4,
137.9, 139.8, 140.1, 142.1, 151.9, 153.3, 153.8, 158.5, 171.8 ppm; IR:
n˜ =3256, 2938, 2839, 1606, 1579, 1507, 1482, 1467, 1428, 1416,
1389, 1365, 1349, 1262, 1244, 1216, 1182, 1163, 1126, 1081, 1040,
994, 926, 850, 809, 770, 738, 694, 635, 568, 549, 531 cmÀ1; UV/Vis
(CH3CN): lmax (loge)=214 (4.7864), 264 (4.2382), 346 nm (3.5254);
MS (ESI+): m/z (%): 419.1 (100) [M+H]+, 447.1 (14) [M+Na]+, 859.2
(66) [2M+Na]+, 1277.4 (21) [3M+Na]+; HRMS (ESI+): m/z calcd for
C21H22O9: 419.1342 [M+H]+; found: 419.1335.
A
mechanically stirred solution of acetophenone 4 (580 mg,
2.95 mmol, 1.0 equiv) in dry EtOH (6 mL) was treated with a solu-
tion of NaOEt (1.02 g, 14.9 mmol, 5.0 equiv) in EtOH (20 mL), and
the mixture was stirred at RT for 3 h. A solution of benzaldehyde
5b (670 mg, 2.96 mmol, 1.0 equiv) in EtOH (16 mL) was added
dropwise, and the resultant mixture was stirred at RT for 20 h. The
solution was treated with H2O (30 mL), acidified to pH 1 with 2.5m
aqueous HCl (20 mL), stirred for 10 min, and extracted with EtOAc
(5”50 mL). The combined organic layers were washed with brine
(1”100 mL), dried over MgSO4, and concentrated under reduced
pressure. Purification by flash column chromatography
(EtOAc/PE=3:7!1:1) yielded 6b as an orange solid (870 mg,
73%). Rf =0.2 (Et2O/PE=1:1); m.p. 114–1198C; 1H NMR (300 MHz,
CDCl3): d=2.25 (brt, J=4.8 Hz, 1H; CH2OH), 3.84 (s, 3H; OMe), 3.90
(s, 3H; OMe), 3.92 (s, 3H; OMe), 3.93 (s, 3H; OMe), 3.94 (s, 3H;
OMe), 4.73 (brd, J=4.8 Hz, 2H; CH2OH), 6.30 (s, 1H; 3’-H), 7.12 (d,
J=2.0 Hz, 1H; 2-H), 7.28 (d, J=2.0 Hz, 1H; 6-H), 7.77 (d, J=
15.6 Hz, 1H; Ha), 7.87 (d, J=15.6 Hz, 1H; Hb), 13.67 ppm (brs, 1H;
2’-OH); 13C NMR (125 MHz, CDCl3): d=55.9, 56.1, 60.9, 61.2, 61.3,
61.9, 96.6, 108.7, 111.9, 121.1, 125.7, 131.3, 134.9, 135.3, 142.9,
148.8, 152.5, 154.9, 160.1, 162.6, 192.7 ppm; IR: n˜ =2979, 2940,
1688, 1556, 1492, 1318, 1255, 1201, 1145, 1107, 1016, 987, 839,
Compound 11
A
mechanically stirred suspension of 2 (100 mg, 0.30 mmol,
1.0 equiv) in acetic anhydride (10 mL) was degassed, cooled to
08C, and carefully treated with three drops of conc. H2SO4. After
15 min the solution turned clear and stirring was continued at RT
for 3 h. H2O (15 mL) was added at 08C, the mixture was diluted
with EtOAc (20 mL), and the layers were thoroughly mixed by vigo-
rous stirring at RT for 15 min. The layers were separated and the
aqueous layer was extracted with EtOAc (2”20 mL). The combined
organic layers were washed with H2O (25 mL), dried over MgSO4,
and concentrated under reduced pressure. Purification by flash
column chromatography (MeOH/CH2Cl2 =1:49) yielded 11 as
a pale-yellow solid (184 mg, 98%). Rf =0.67 (MeOH/CH2Cl2 =1:19);
1H NMR (300 MHz, CDCl3): d=2.32 (s, 6H; 2”OAc), 2.33 (s, 6H; 2”
OAc), 2.34 (s, 3H; OAc), 2.36 (s, 3H; OAc), 2.43 (s, 3H; OAc), 7.48 (s,
1H; 8-H), 7.60 ppm (s, 2H; 2’-H and 6’-H); 13C NMR (125 MHz,
CDCl3): d=20.02, 20.15, 20.36, 20.65, 20.69, 20.77, 110.29, 115.38,
120.77, 127.04, 132.97, 133.98, 136.87, 142.25, 143.62, 147.29,
153.24, 153.33, 166.49, 166.77, 167.05, 167.41, 167.84, 168.05,
169.83 ppm; IR: n˜ =2930, 1770, 1656, 1628, 1499, 1461, 1423, 1370,
1354, 1166, 1141, 1051, 1012, 896, 870, 824, 693 cmÀ1; UV/Vis
(CH3CN): lmax (loge)=253 (4.3279), 294 nm (4.2779); MS (ESI+): m/z
(%): 629.1 (38) [M+H]+, 651.1 (75) [M+Na]+, 1279.2 (100)
[2M+Na]+; HRMS (ESI+): m/z calcd for C29H24O16: 629.1137 [M+H]+
811 cmÀ1
; UV/Vis (CH3CN): lmax (loge)=209 (4.6730), 362 nm
(4.3711); MS (ESI+): m/z (%): 405.2 (100) [M+H]+, 831.3 (89)
[2M+Na]+, 1235.5 (67) [3M+Na]+; HRMS (ESI+): m/z calcd for
C21H24O8: 405.1544 [M+H]+; found: 405.1542.
Compound 7b
A mechanically stirred solution of chalcone 6b (809 mg, 2.0 mmol,
1.0 equiv) in AcOH (120 mL) was heated at 1208C for 72 h. The re-
action was cooled to RT, treated with H2O (50 mL), and extracted
with EtOAc (3”100 mL). The combined organic layers were
washed with brine (1”150 mL), dried over MgSO4, and concentrat-
ed under reduced pressure. Purification by flash column chroma-
tography (EtOAc/PE=1:4!1:1) yielded 7b as
a yellow solid
1
;
found: 629.1115; HRMS (ESI+): m/z calcd for C29H24O16Na:
(500 mg, 56%). Rf =0.65 (EtOAc/PE=1:1); m.p. 114–1188C; H NMR
(300 MHz, CDCl3): d=2.11 (s, 3H; OAc), 2.77 (dd, J=16.6, 2.9 Hz,
1H; 3-Ha), 3.01 (dd, J=16.6, 13.4 Hz, 1H; 3-Hb), 3.83 (s, 3H; OMe),
3.877 (s, 3H; OMe), 3.880 (s, 3H; OMe), 3.91 (s, 3H; OMe), 3.94 (s,
3H; OMe), 5.17 (s, 2H; 3’-CH2), 5.35 (dd, J=13.4, 2.9 Hz, 1H; 2-H),
6.36 (s, 1H; 8-H), 7.00 (d, J=2.1 Hz, 1H; 2’-H), 7.03 ppm (d, J=
2.1 Hz, 1H; 6’-H); 13C NMR (125 MHz, CDCl3): d=21.1, 45.7, 56.0,
56.2, 61.1, 61.4, 61.5, 61.7, 79.3, 96.4, 109.1, 110.4, 119.3, 130.1,
134.4, 137.6, 147.7, 152.8, 154.2, 159.4, 159.5, 170.6, 188.9 ppm; IR:
n˜ =2940, 2832, 1723, 1682, 1596, 1486, 1455, 1253, 1239, 1202,
1104, 1019, 998, 981, 857 cmÀ1; UV/Vis (CH3CN): lmax (loge)=276
(3.8903), 320 nm (3.7777); MS (ESI+): m/z (%): 447.2 (100) [M+H]+,
469.2 (86) [M+Na]+, 915.3 (92) [2M+Na]+. HRMS (ESI+): m/z calcd
for C23H26O9: 447.1650 [M+H]+; found: 447.1648; HRMS (ESI+): m/z
calcd for C23H26O9Na: 469.1469 [M+Na]+; found: 469.1466.
651.0947 [M+Na]+; found: 651.0957.
Compound 5b
A solution of TBS-protected alcohol 16 (see the Supporting Infor-
mation; 310 mg, 1.0 mmol, 1.0 equiv) and TBAF (1m in THF, 2.0 mL,
2.0 mmol, 2.0 equiv) in dry THF (15 mL) was stirred at RT for 16 h.
1m aqueous HCl (5 mL) was added and the mixture extracted with
EtOAc (5”20 mL). The combined organic layers were washed with
brine (25 mL), dried over MgSO4, and concentrated under reduced
pressure. Purification by flash column chromatography (EtOAc/pe-
troleum ether (PE)=3:7!1:1) yielded 5b as a white solid (192 mg,
1
98%). Rf =0.14 (EtOAc/PE=3:7); H NMR (300 MHz, CDCl3): d=2.75
(brs, 1H; OH), 3.89 (s, 3H; OMe), 3.93 (s, 3H; OMe), 4.72 (s, 2H;
CH2), 7.36 (d, J=1.9 Hz, 1H; 6-H), 7.48 (d, J=1.9 Hz, 1H; 2-H),
9.83 ppm (s, 1H; CHO); 13C NMR (125 MHz, CDCl3): d=56.2 (OMe),
61.0 (CH2), 61.2 (OMe), 110.5 (C-6), 124.9 (C-2), 132.4 (C-1), 134.9 (C-
3), 152.2 (C-4), 152.9 (C-5), 191.1 ppm (CHO); IR: n˜ =3422, 2937,
1686, 1587, 1299, 1136, 999, 728 cmÀ1; UV/Vis (CH3CN): lmax (loge)=
Compound 9b
Conc. HCl (1 mL) was added via syringe pump over 10 min to a
mechanically stirred solution of 7b (500 mg, 1.12 mmol, 1.0 equiv)
Chem. Eur. J. 2015, 21, 16887 – 16894
www.chemeurj.org
16892
ꢀ 2015 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Full Paper
and isoamyl nitrite (905 mL, 6.73 mmol, 6.0 equiv) in dry ethanol
(35 mL) at 08C. The reaction was stirred at 808C for 15 min. After
cooling the reaction mixture to RT, the volatile compounds were
evaporated under reduced pressure. The residue was taken up in
EtOAc (40 mL) and washed with H2O (2”25 mL). The combined
aqueous layers were extracted with EtOAc (2”25 mL) and CHCl3
(2”25 mL). The combined organic layers were washed with brine
(50 mL), dried over MgSO4, and concentrated under reduced pres-
sure to yield the intermediate oxime as a yellow solid. The oxime
was dissolved in glacial acetic acid (40 mL), cooled to 08C, and
treated with 10% aqueous H2SO4 (8 mL). The reaction was stirred
at 110 8C for 30 min, then cooled to RT, and concentrated under re-
duced pressure. The residue was taken up in EtOAc (40 mL) and
washed with H2O (2”25 mL). The combined aqueous layers were
extracted with EtOAc (2”25 mL) and CHCl3 (2”25 mL). The com-
bined organic layers were washed with brine (50 mL) and dried
over MgSO4. Evaporation of the volatile compounds gave the
crude product, which was first purified by flash column chroma-
tography (MeOH/CH2Cl2 =1:19). The resultant dark-yellow solid
(120 mg) was suspended in EtOAc (1 mL) and carefully decanted to
leave pure 9b (105 mg, 20%) as a light-yellow solid. Rf =0.41
(MeOH/CH2Cl2 =1:19); 1H NMR (600 MHz, CDCl3): d=2.11 (s, 3H;
OAc), 3.91 (s, 3H; OMe), 3.93 (s, 3H; OMe), 3.96 (s, 3H; OMe), 3.99
(s, 3H; OMe), 4.01 (s, 3H; OMe), 5.22 (s, 2H; CH2), 6.78 (s, 1H; 8-H),
7.34 (brs, 1H; 3-OH), 7.74 (d, J=6.0 Hz; 1H), 7.85 ppm (d, J=
6.0 Hz; 1H); 13C NMR (125 MHz, CDCl3): d=21.26, 56.21, 56.60,
61.35, 61.73, 61.95, 62.37, 96.23, 109.87, 112.51, 121.27, 126.97,
129.98, 137.89, 140.16, 142.13, 149.17, 151.93, 152.73, 153.93,
158.54, 171.04, 171.82 ppm; IR: n˜ =3281, 2942, 1737, 1728, 1607,
1481, 1261, 1215, 1162, 1082, 1017, 997 cmÀ1; UV/Vis (CH3CN): lmax
(loge)=194 (4.6625), 252 (4.2698), 344 nm (4.2808); MS (ESI+): m/z
(%): 461.2 (100) [M+H]+, 921.3 (27) [2M+H]+; HRMS (ESI+): m/z
calcd for C23H24O10: 461.1442 [M+H]+; found: 461.1444.
cultured in 96-well plates. Then, the NHDFs were starved with
serum-free Dulbecco’s modified eagle’s medium (DMEM) for 24 h.
The chemicals were added at the indicated concentrations. After
24 h, the MTS solution (20 mL) was added to each well and the
plate was incubated for 1 h at 378C in the presence of 5% CO2.
The absorbance was analyzed at l=492 nm.
Western blot
The NHDFs were confluently cultured and starved with serum-free
DMEM for 24 h. Chemicals were added to the cells for 1 h at the in-
dicated concentrations. The cells were exposed to sUV light. After
36 h of sUV irradiation, the cells were lyzed with radio-immuno-
precipitation assay (RIPA) buffer. Equal amounts of proteins were
loaded onto 10% sodium dodecyl sulfate (SDS)/polyacrylamide
gels and transferred to Immobilon P membranes (Millipore). The
membrane was blocked with 5% fat-free milk for 1 h. The mem-
brane was incubated with a specific primary antibody at 48C over-
night. The proteins were hybridized with a horseradish peroxidase
(HRP) conjugated secondary antibody, and the band was analyzed
with a chemiluminescence detection kit (GE Healthcare, Pittsburgh,
PA, USA).
Acknowledgements
This work was supported by the Deutsche Forschungsgemein-
schaft (DFG), Land Niedersachsen, and Volkswagenstiftung. J.H.
thanks the Dorothea-Schlçzer Program for a postdoctoral
scholarship.
Keywords: drug design · flavonoids · inhibitors · JNK1 ·
synthetic drugs
IMAP assay
The IMAP assay was performed as previously described.[11] The
IMAP assay was carried out with recombinant JNK1 (100 nm) in re-
action buffer [3-(N-morpholino)propanesulfonic acid (MOPS)
(20 mm; pH 6.5), dithiothreitol (DTT; 1 mm), MgCl2 (10 mm), and
polyoxyethyleneglycol dodecyl ether (Brij35; 0.01%)] in a 384-well
black plate with serially diluted test compounds. A fluorescein-iso-
thiocyanate-labeled JNK1 peptide substrate (LVEPLTPSGEAPNQK-
5FAM-COOH) (400 nm) and ATP (7.46 mm) were added. After
60 min incubation with the potential inhibitor (5 mm) at RT, IMAP
binding buffer (1:1200 dilution of IMAP progressive binding
reagent in 65:35 IMAP progressive binding buffer A/IMAP
progressive binding buffer B; supplied by Molecular Devices) was
added and incubated for 60 min. The plate was read by using
a PHERAstar Plus microplate reader (BMG Labtech). The excitation
and emission wavelengths were l=485 nm (bandwidth=20 nm)
and l=530 nm (bandwidth=25 nm), respectively.
[1] B. DØrijard, M. Hibi, I.-H. Wu, T. Barrett, B. Su, T. Deng, M. Karin, R. J.
Davis, Cell 1994, 76, 1025–1037.
[2] Q. Chang, Y. Zhang, K. J. Beezhold, D. Bhatia, H. Zhao, J. Chen, V. Castra-
nova, X. Shi, F. Chen, J. Hepatol. 2009, 50, 323–333.
[3] X. Ouyang, W. J. Jessen, H. Al-Ahmadie, A. M. Serio, Y. Lin, W.-J. Shih,
V. E. Reuter, P. T. Scardino, M. M. Shen, B. J. Aronow, A. J. Vickers, W. L.
Gerald, C. Abate-Shen, Cancer Res. 2008, 68, 2132–2144.
[4] E. F. Wagner, A. R. Nebreda, Nat. Rev. Cancer 2009, 9, 537–549.
[5] T. Sakurai, S. Maeda, L. Chang, M. Karin, Proc. Natl. Acad. Sci. USA 2006,
103, 10544–10551.
[6] J. Hirosumi, G. Tuncman, L. Chang, C. Z. Gçrgün, L. T. Uysal, K. Maeda,
M. Karin, G. S. Hotamisligil, Nature 2002, 420, 333–336.
[7] R. Visse, H. Nagase, Circ. Res. 2003, 92, 827–839.
[8] H. Endo, T. Watanabe, Y. Sugioka, M. Niioka, Y. Inagaki, I. Okazaki, Int. J.
Oncol. 2009, 35, 1237–1245.
[9] C. Zugowski, F. Lieder, A. Mueller, J. Gasch, F. M. Corvinus, R. Moriggl, K.
Friedrich, Biol. Chem. 2011, 392, 449–459.
[10] E. C. Bate-Smith, J. B. Harborne, Phytochemistry 1969, 8, 1035–1037.
[11] S. Baek, N. J. Kang, G. M. Popowicz, M. Arciniega, S. K. Jung, S. Byun,
N. R. Song, Y.-S. Heo, B. Y. Kim, H. J. Lee, T. A. Holak, M. Augustin, A. M.
Bode, R. Huber, Z. Dong, K. W. Lee, J. Mol. Biol. 2013, 425, 411 –423.
[12] G.-J. Kang, S.-C. Han, J.-W. Ock, H.-K. Kang, E.-S. Yoo, Biomol. Ther. 2013,
21, 138–145.
[13] I. Duarte Silva, J. Gaspar, G. Gomes da Costa, A. S. Rodrigues, A. Laires, J.
Rueff, Chemico-Biol. Interactions 2000, 124, 29–51.
[14] S. Holder, M. Zemskova, C. Zhang, M. Tabrizizad, R. Bremer, J. W. Nei-
digh, M. B. Lilly, Mol. Cancer Ther. 2007, 6, 163–172.
Cell culture
NHDFs were purchased from the American Type Culture Collection
(Manassas, VA, USA). NHDFs were cultured in DMEM</M199
(4:1 v/v) supplemented with fetal bovine serum (FBS; 10% v/v) at
378C under a 5% CO2 atmosphere.
[15] D. Ichimatsu, M. Nomura, S. Nakamura, S. Moritani, K. Yokogawa, S. Ko-
bayashi, T. Nishioka, K.-i. Miyamoto, Mol. Carcinog. 2007, 46, 436–445.
[16] C. Tsimplouli, C. Demetzos, M. Hadzopoulou-Cladaras, P. Pantazis, K.
Dimas, Eur. J. Nutr. 2012, 51, 181–190.
Cell cytotoxicity
The cell cytotoxicity of the compounds was evaluated with Cell
Titer 96 AQueous One Solution (Promega). In brief, the cells were
Chem. Eur. J. 2015, 21, 16887 – 16894
www.chemeurj.org
16893
ꢀ 2015 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Full Paper
[17] G.-J. Kang, S.-C. Han, N.-J. Kang, D.-H. Koo, D.-B. Park, S.-Y. Eun, H.-K.
Kang, E.-S. Yoo, Br. J. Dermatol. 2014, 171, 512–523.
[18] Y. Jung, S. Y. Shin, Y. Yong, H. Jung, S. Ahn, Y. H. Lee, Y. Lim, Chem. Biol.
Drug Des. 2015, 85, 574–585.
[19] K. Amrutha, P. Nanjan, S. K. Shaji, D. Sunilkumar, K. Subhalakshmi, L. Ra-
jakrishna, A. Banerji, Bioorg. Med. Chem. Lett. 2014, 24, 4735–4742.
[20] E. Lee, K.-W. Jeong, H. N. Jnawali, A. Shin, Y.-S. Heo, Y. Kim, Molecules
2014, 19, 13200–13211.
[24] For the importance of hydrogen-bond networks for kinase selectivity
see: W. D. Jang, J.-T. Kim, N. S. Kang, J. Mol. Liq. 2014, 191, 37–41.
[25] R. A. Friesner, J. L. Banks, R. B. Murphy, T. A. Halgren, J. J. Klicic, D. T.
Mainz, M. P. Repasky, E. H. Knoll, M. Shelley, J. K. Perry, D. E. Shaw, P.
Francis, P. S. Shenkin, J. Med. Chem. 2004, 47, 1739–1749.
[26] R. A. Friesner, R. B. Murphy, M. P. Repasky, L. L. Frye, J. R. Greenwood,
T. A. Halgren, P. C. Sanschagrin, D. T. Mainz, J. Med. Chem. 2006, 49,
6177–6196.
[27] J. R. Sportsman, E. A. Gaudet, A. Boge, Assay Drug Dev. Technol. 2004, 2,
205–214.
[21] PepJIP1 is a docking-site peptide fragment of the scaffold protein JIP1.
[22] Y.-S. Heo, S.-K. Kim, C. I. Seo, Y. K. Kim, B.-J. Sung, H. S. Lee, J. I. Lee, S.-Y.
Park, J. H. Wim, K. Y. Hwang, Y.-L. Hyun, Y. H. Jeon, S. Ro, J. M. Cho, T. G.
Lee, C.-H. Yang, EMBO J. 2004, 23, 2185–2195.
[23] K. M. Comess, C. Sun, C. Abad-Zapatero, E. R. Goedken, R. J. Gum, D. W.
Borhani, M. Argiriadi, D. R. Groebe, Y. Jia, J. E. Clampit, D. L. Haasch, H. T.
Smith, S. Wang, D. Song, M. L. Coen, T. E. Cloutier, H. Tang, X. Cheng, C.
Quinn, B. Liu, Z. Xin, G. Liu, E. H. Fry, V. Stoll, T. I. Ng, D. Banach, D. Mar-
cotte, D. J. Burns, D. J. Calderwood, P. J. Hajduk, ACS Chem. Biol. 2011, 6,
234–244.
[28] For a theoretical calculation of the antioxidative properties of 2, see: E.
Estrada, J. A. Quincoces, G. Patlewicz, Mol. Divers. 2004, 8, 21–33.
[29] M. P. Vincenti, C. E. Brinckerhoff, Arthritis Res. 2002, 4, 157–164.
[30] T.-G. Lim, J.-E. Kim, S.-J. Lee, J. S. Park, M. H. Yeom, H. Chen, A. M. Bode,
Z. Dong, K. W. Lee, Int. J. Mol. Sci. 2014, 15, 21419–21432.
Received: June 25, 2015
Published online on October 7, 2015
Chem. Eur. J. 2015, 21, 16887 – 16894
www.chemeurj.org
16894
ꢀ 2015 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim