New selective cyclooxygenase-2 inhibitors from cyclocoumarol: Synthesis, characterization, biological evaluation and molecular modeling

Article abstract of DOI:10.1016/j.ejmech.2018.01.054

In this work, a serie of cyclocoumarol derivatives was designed, synthesized, characterized and studied for their potentialities as selective inhibitors of COX-2. All target compounds have been screened for their anti-inflammatory activity by the assay of PGE₂ production. Among them, compound 5d exhibited the most potent inhibitory activity with a PGE₂ inhibition compared to NS-398 (79% and 88% respectively) and showed non-inhibitory activity towards the COX-1 enzyme. Docking studies revealed the capacity of this compound to occupy the selective COX-2 cavity establishing additional hydrogen bonds between the oxygen of the methoxy group and the His90 and Arg513 of the binding site of the enzyme.

Full text of DOI:10.1016/j.ejmech.2018.01.054

Accepted Manuscript
New selective cyclooxygenase-2 inhibitors from cyclocoumarol: Synthesis,
characterization, biological evaluation and molecular modeling
Anita Rayar, Nathalie Lagarde, Frederique Martin, Florent Blanchard, Bertrand Liagre,
Clotilde Ferroud, Jean-François Zagury, Matthieu Montes, Maité Sylla-Iyarreta Veitía
PII:
S0223-5234(18)30067-9
10.1016/j.ejmech.2018.01.054
EJMECH 10131
DOI:
Reference:
To appear in: European Journal of Medicinal Chemistry
Received Date: 12 October 2017
Revised Date: 26 December 2017
Accepted Date: 16 January 2018
Please cite this article as: A. Rayar, N. Lagarde, F. Martin, F. Blanchard, B. Liagre, C. Ferroud, Jean.-
Franç. Zagury, M. Montes, Maité. Sylla-Iyarreta Veitía, New selective cyclooxygenase-2 inhibitors from
cyclocoumarol: Synthesis, characterization, biological evaluation and molecular modeling, European
Journal of Medicinal Chemistry (2018), doi: 10.1016/j.ejmech.2018.01.054.
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

European Journal of Medicinal Chemistry
ACCEPTED MANUSCRIPT
New selective cyclooxygenase-2 inhibitors from cyclocoumarol: synthesis,
characterization, biological evaluation and molecular modeling
c
Anita Marie Rayar a, b, Nathalie Lagarde b, Frederique Martinc, Florent Blanchardd, Bertrand Liagre ,
Clotilde Ferroud a, Jean-François Zaguryb, Matthieu Montes b, Maité Sylla-Iyarreta Veitíaa*
a. Equipe de Chimie Moléculaire du Laboratoire CMGPCE, EA 7341, Conservatoire national des arts et
métiers, 2 rue Conté, 75003, Paris
b. Laboratoire Génétique Bioinformatique et Applications, EA 4627, Conservatoire National des Arts et
Métiers, 2 rue Conté, 75003 Paris.
c. Laboratoire de Chimie des Substances Naturelles, EA1069, Faculté de Pharmacie, 2 rue du Dr
Marcland, 87025 Limoges.
d. Institut de Chimie des Substances Naturelles, CNRS UPR 2301, Université. Paris-Sud, Université
Paris-Saclay, Gif-sur-Yvette, France.
Corresponding author:
E-mail address : maite.sylla@lecnam.net (M. Sylla-Iyarreta Veítia)
ABSTRACT
In this work, a serie of cyclocoumarol derivatives was designed, synthesized, characterized and studied
for their potentialities as selective inhibitors of COX-2. All target compounds have been screened for
their anti-inflammatory activity by the assay of PGE2 production. Among them, compound 5d exhibited
the most potent inhibitory activity with a PGE2 inhibition compared to NS-398 (79% and 88%
respectively) and showed non-inhibitory activity towards the COX-1 enzyme. Docking studies revealed
the capacity of this compound to occupy the selective COX-2 cavity establishing additional hydrogen
bonds between the oxygen of the methoxy group and the His90 and Arg513 of the binding site of the
enzyme.
Keywords : Anti-inflammatory, COX-2 inhibitors, COX-1/COX-2 inhibition, Cyclocoumarol,
Cyclooxygenase, Docking study
1

European Journal of Medicinal Chemistry
ACCEPTED MANUSCRIPT
1- Introduction
Inflammation is a complex phenomenon affecting millions of people around the world and is involved
in the development of many human diseases. The inflammatory response is controlled by a wide variety
of mediators, including prostaglandins (PGs), which play a major role in the inflammatory response
mediated by cyclooxygenases. The inhibition of these inflammatory mediators can treat the disease,
reduce the inflammation but lead to undesirable side effects [1]. Therefore, the development of new
anti-inflammatory drugs represents a real challenge considering the benefit-risk balance as one of major
concerns in therapeutics.
Cyclooxygenase-2 (COX-2) is a key enzyme involved in the inflammation process particularly in the
conversion of arachidonic acid to prostanoids [2, 3]. Non-steroidal anti-inflammatory drugs (NSAID)
act as selective COX-2 inhibitors by specifically targeting COX-2. Several NSAID have been developed
and approved for commercial use (Fig. 1). Nonetheless clinical trials revealed that despite the reduction
of intestinal disorders these inhibitors cause significant cardiovascular, liver and allergic problems
resulting in the withdrawal of certain compounds from the market [4, 5].
Fig. 1. Commercially available COX-2 inhibitors used to treat inflammation
In a previous study, the cyclocoumarol, an anticoagulant drug, was identified as a potential anti-
inflammatory compound by TOMOCOMD-CARDD method. The in silico results were confirmed by
biological studies using a zebrafish model [6]. Based on these observations and in an attempt to pursue
the development of new anti-inflammatory agents potently COX-2 selective inhibitors, a serie of novel
cyclocoumarol derivatives was designed. We hereby report the synthesis and spectroscopic
characterization of these derivatives. The X-ray crystallographic structural analysis is also described.
The prepared compounds were evaluated for their inhibitory activity on COX-2 and compared with a
2

European Journal of Medicinal Chemistry
ACCEPTED MANUSCRIPT
recognized selective COX-2 inhibitor, NS-398. The selectivity for COX-1 and COX-2 isozymes has
also been evaluated. To our knowledge, the biological evaluation of COX-2 inhibitory activity of these
compounds is related herein for the first time. Additionally, molecular docking studies were carried out
to understand the binding modes of these new compounds to the active site of COX-2.
2. Results and discussion
2.1 Chemistry
The synthetic strategy used to prepare the target compounds 5a-g, 5a’-g’ is outlined in Scheme 1. The
key intermediates, benzalacetones 1a-g, were synthetized from acetone and the corresponding aldehyde
in the presence of aqueous NaOH solution under microwave activation. This optimized procedure has
been previously described [7]. The functionalized warfarin derivatives 4a-g were prepared by 1.4
Michael addition, coupling an equimolar of 4-hydroxycoumarin 3 and the corresponding benzalacetones
2a-g under reflux in water [8]. Inspired by Barker et al. protocol [9] the experimental conditions were
optimized by adding 0.05 equiv. of N, N diisopropylethylamine (DIPEA) as catalyst. Warfarin
intermediates were isolated in yields between 27% and 96% with reaction times among 4- 48 h. To our
knowledge, the use of DIPEA in water, has never been described in literature to prepare warfarin
analogues via 1.4 Michael addition (See Supplementary data). It should be noticed that the low yield of
27% obtained for compound 4e could be explained by a laborious workup.
3

European Journal of Medicinal Chemistry
ACCEPTED MANUSCRIPT
Scheme 1. Synthesis of cyclocoumarol derivatives (5-a-g), (5-a’-g’) (a): Aldehyde (1 equiv), acetone
(13.6 equiv), NaOH (0.6 g /cm3 of water), microwaves irradiation (40/50 °C, 5 W, 10-15 min.); (b):
Hydroxycoumarin (3) (1 equiv), DIPEA (5% mole), water [0,6 M], reflux, 4-48 h; (c): HCl 4% MeOH
(10 equiv), reflux, 22-66 h.
Compounds 5a-g and 5a’-g’ were prepared by cyclisation of corresponding functionalized warfarins 4a-
g. Heating at reflux of compounds 4a-g in methanolic HCl 4% afforded, the desired pyranocoumarins
5a-g and 5a’-g’ with reaction times among 22-66 h. Then, purified by flash chromatography major
diastereomers 5a-g were obtained in yields between 38% and 64% and minor diastereomers 5a’-g’ were
1
isolated in yields of 9-19%. The structures of stereoisomers were confirmed by H NMR and X-ray
crystallographic analysis.
2.2 1H NMR and X-ray diffraction analysis
The 1H NMR spectra of cyclocoumarol derivatives 5a-g, 5a’-g’ exhibited characteristic signals allowing
an easy distinction between the major and the minor diastereomers. To illustrate this, the proton
correlation have been confirmed by 2D NMR (Fig. 2 and 3).
4

European Journal of Medicinal Chemistry
ACCEPTED MANUSCRIPT
The major diastereomer 5a, displayed an ABX system where diastereotopic protons H13eq (2.52 ppm)
and H13ax (2.01ppm) were coupled with the benzylic proton in axial position, Hax (4.16 ppm). The
protons H13eq and H13ax were coupled with a geminal coupling constant value of 14.0 Hz and the
benzylic proton was coupled with protons H13eq and H13ax with a coupling constant of 6,9 and 11,8
Hz respectively, (Fig. 2).
Chemical Shift in ppm
Structural information
1.70 (s, 3H);
C21-H
2.01 (dd, 1H, , Jgem = 14.0 Hz, J13ax-14ax = 11.8 Hz ); C13ax -H
2.52 (dd, 1H, H13eq, Jgem = 14.0 Hz, Jeq-ax = 6.9 Hz ); C13eq -H
3.36 (s, 3H,);
C22-H
4.16 (dd ,1H, Jax-ax = 11.8 Hz, Jax-eq = 6.9 Hz );
7.24-7.26 (m, 2H);
C14ax -H
C16-H, C20-H
7.28-7.35 (m, 5H);
C6-H, C8-H, C17-H, C18-H, C19-H
7.53-7.58 (ddd, 1H, J = 9.4, 7.4, 1.7 Hz);
7.91 (dd, 1H, J = 8.0, 1.3 Hz);
C7-H
C5-H
Fig. 2. Protons correlations and chemical shift values of major diastereomer of cyclocoumarol 5a by 2D
NMR.
The minor diastereomer 5a’ exhibited a typical coupling constant of a benzylic proton in equatorial
position. The Heq at 4.19 ppm was coupled with the two diastereotopic protons H13eq and H13ax with
coupling constants of 3.7 and 7.2 Hz respectively. The proton H13ax (2.38 ppm) was coupled with
H13eq (2.52 ppm) with a geminal coupling constant of 14.1 Hz, (Fig. 3).
Chemical Shift in ppm
Structural information
1.69 (s, 3H);
C21-H
2.38 (dd, 1H, Jgem = 14.0 Hz, J13ax-14eq = 7.4 Hz );
C13ax -H
2.52 (dd, 1H, H13eq, Jgem = 14.1 Hz, Jeq-eq = 3.8 Hz ); C13eq -H
3.29 (s, 3H,);
C22-H
4.19 (dd ,1H, Jeq-ax = 7.2 Hz, Jeq-eq = 3.7 Hz );
7.21-7.32 (m, 5H);
C14eq -H
C16-H, C17-H, C18-H, C19-H, C20-
7.34-7.43 (m, 2H);
H
7.60-7.62 (m, 1H,);
C6-H, C8-H
C7-H
7.96 (dd, 1H, J = 7.8, 1.2 Hz);
C5-H
Fig.3. Protons correlations and chemical shift values of minor diastereomer of cyclocoumarol 5a’ by 2D
NMR.
5

European Journal of Medicinal Chemistry
ACCEPTED MANUSCRIPT
Crystallography studies of 2-methoxy-2-methyl-4-phenyl-3,4-dihydropyrano[3,2-c]chromen-5(2H)-one
(5a and 5a’) and 2-methoxy-2-methyl-(1-(4-methoxyphenyl))-3,4-dihydropyrano[3,2-c]chromen-5(2H)-
one (5d) were performed at room temperature on a Rigaku XtalabPro diffractometer equipped with a
microfocus source (MicroMax003_Mo) and multilayer confocal mirrors (Mo Kα radiation, λ = 0.71075
Å). The crystal and structure refinement data of compounds 5a, 5a’ and 5d are reported in the
supplementary data.
The crystallographic structure of the two diastereomers couples of cyclocoumarol 5a and 5a’ were in
agreement with the literature[10] and confirmed that cyclocoumarol could be exist in the form of a pair
of diastereomers: (12S,14S)/(12R,14R) and (12S,14R)/(12R,14S).
In a solid state, minor diastereomer 5a’ (12S,14R)/(12R,14S) form a preferred half-chair conformation in
which the methoxy and phenyl groups are axially and pseudoaxially positioned respectively. This
unusual spatial rearrangement could be explained by a close nonbonded contact between the carbone 15
and the oxygen atom of ketal function of 2.923(2) Å. The relative stability of this conformation reflects
the dipole interactions of the oxygens bonded to the ketal carbon and is comparable with the anomeric
effect observed in glycosides [10] (Fig. 4).
Fig. 4. Representation of the proposed spatial rearrangement of minor diastereomer of 5a.
For major diastereomer 5a, containing the enantiomers couples (12S,14S)/(12R,14R), only the half-chair
conformation, in which the methyl and phenyl groups are equatorial and pseudoequatorial respectively
1
is consistent with the values of the coupling constants determined by H NMR (See experimental
section). Since compound 5a possess the all-staggered conformation in which no obvious nonbonded
6

European Journal of Medicinal Chemistry
ACCEPTED MANUSCRIPT
interactions occur between the oxygen bonded to the ketal carbon and the carbone 15, it is the most
realistic conformation based on steric and dipole effects.
2.3 Biological evaluation of cyclocoumarol analogues
The inhibitory activity on COX-2 was evaluated by determination of the PGE2 production inhibition, the
prostaglandin predominantly overproduced in case of inflammation. The synthesis of PGE2 was
evaluated after pretreatment with the compounds at 10 µM for 2 h and then treatment with LPS (10
ng/mL) for 24 h. The Raw 264.7 cells were seeded at 2.105 cells / well for 24 h and then processed as
described in the experimental section. PGE2 levels were determined from the culture supernatants using
an EIA kit (Cayman Chemical). The results are expressed as the mean ± SD of independent
experiments. NS-398, a specific inhibitor of COX-2, was used as a reference [11] (Table 1 and Fig. 5).
All the synthesized cyclocoumarol derivatives exhibited an inhibitory activity of about 50% inhibition
(5a, 5b, 5c, 5e). The compound 5d bearing a methoxy group displayed the strongest inhibition of PGE2
synthesis compared to the reference NS-398 (79% and 88% inhibition respectively). However,
introduction of electro-withdrawing groups trifluoromethyl and nitro (compound 5f and 5g) led to a
significant decrease of inhibitory activity of COX-2 (31% and 0% inhibition respectively). Minor
diastereomers were also evaluated and displayed lower inhibiting activities compared to their
corresponding major isomers (Table 1 and Fig. 5).
Table 1: Inhibition of PGE2 production stimulated by LPS by pharmacological agents in the Raw 264.7
cells (compared with NS-398 used as reference, 88% of inhibition)
% inhibition of PGE2
Compounds
R
H
50 (27)
60 (30)
52 (45)
5a (5a’)
5b (5b’)
5c (5c’)
CH3
t-Bu
7

European Journal of Medicinal Chemistry
ACCEPTED MANUSCRIPT
OMe
F
79 (nd)
48 (14)
31 (4)
0
5d (5d’)
5e (5e’)
5f (5f’)
5g (5g’)
CF3
NO2
8

European Journal of Medicinal Chemistry
ACCEPTED MANUSCRIPT
Fig. 5. Evaluation of COX-2 activity inhibition of synthesized analogues of cyclocoumarol 5a-g, 5a’-
5g’. Synthesis of PGE2 after pretreatment with pharmacological agents; nd: no determined, A with
major diastereomers, 5a-g B with minor diastereomers, 5a’-5g’. PGE2 production was measured in
culture supernatants according to the manufacturer’s instructions (PGE2 EIA Kit, Cayman Chemical).
Values are expressed as mean±SEM (*P-value relative to control group, * p<0.05; (#P-value relative to
LPS group, #p<0.05).
Considering these results on PGE2 production, we evaluated the dose-response behavior of the most
active molecule 5d. Four experiments at different concentrations were conducted at 1 µM, 5 µM, 10 µM
and 20 µM using NS-398 as reference. At 1 µM, a low inhibition for 5d (3%) and the reference NS-398
(11%) was noted. At 5 µM 45% and 61% of inhibition was observed for 5d and NS-398 respectively.
An activity comparable to NS-398 was observed at 10 µM and 20 µM with 79% and 89% of inhibition
respectively. These results confirmed that the analogue 5d displays a dose-dependent inhibition profile
(Table 2, Fig. 6A). We then analyzed if the molecule 5d had an effect on COX-2 expression. With
regard to the dose-response used, we can say that the molecule 5d does not modify the expression of
COX-2 (Fig. 6B) but only its activity (Fig. 6A).
Table 2 Inhibition of PGE2 production stimulated by LPS by molecule 5d in the Raw 264.7 cells at
different doses.
% inhibition de PGE2
Dose (µM)
NS-398
5d
3
1
5
11
61
45
79
89
10
20
88
94
9

European Journal of Medicinal Chemistry
ACCEPTED MANUSCRIPT
Fig. 6. Dose-response of 5d compound of COX-2 activity (A) and expression (B) inhibition. (A) PGE2
production was measured in culture supernatants according to the manufacturer’s instructions (PGE2
EIA Kit, Cayman Chemical). Values are expressed as mean±SEM (*P-value relative to control group, *
p<0.05; (#P-value relative to LPS group, #p<0.05). (B) COX-2 expression was evaluated in the total
cellular pool using Western blot analysis (β-actin was used as a loading control and the blot shown is
representative of three separate experiments).
10

European Journal of Medicinal Chemistry
ACCEPTED MANUSCRIPT
COX-2/COX-1 inhibition of majors diastereomers 5a-5g was also studied to confirm the COX-2
selectivity. It has been extensively noted in the literature that non-selective molecules have side effects
affecting especially the gastrointestinal system. The COX-1 activity inhibition assay was performed
without LPS in the medium, measuring the production rate of PGE2 synthesized by COX-1 (Fig. 7A) .
The effect of 10 µM molecule 5d on COX-1 expression was also analyzed (Fig. 7B).
Fig. 7. Evaluation of COX-1 activity (A) and expression (B) inhibition of synthesized analogues of
cyclocoumarol. (A) PGE2 production was measured in culture supernatants according to the
manufacturer’s instructions (PGE2 EIA Kit, Cayman Chemical). Values are expressed as mean±SEM.
(B) Raw 264.7 cells were treated or not (C, control) with 10 µM molecule 5d for 24 h. COX-1
11

European Journal of Medicinal Chemistry
ACCEPTED MANUSCRIPT
expression was evaluated in the total cellular pool using Western blot analysis (β-actin was used as a
loading control and the blot shown is representative of three separate experiments).
The results obtained in the in vitro COX-1/COX-2 inhibition study showed that synthesized molecules
display no significant inhibition of COX-1 expression and activity (Fig. 7B and 7A respectively). This
confirmed the previous results and identified compound 5d as a selective inhibitor of COX-2 activity.
2.4 Molecular modeling
To understand the ability of the synthesized compounds to inhibit COX-2 activity and to get some
insights on their binding mode into the COX-2 binding site, the selected compounds were docked into
the binding site of COX-2, defined using the co-crystallized selective COX-2 inhibitor as a reference,
SC-558 (PDB ID: 1CX2). The choice of the PDB structure was based on a previous benchmarking study
[12] in which the selected structure 1CX2 was associated with the highest performance in discriminating
COX-2 inhibitors from decoy compounds [13], i.e. presumed inactive compounds. The predicted
conformations of the ligand in the binding site were generated with the docking software AutoDock
Vina and Surflex-dock that display different conformational search algorithms and different scoring
functions. To validate our docking protocols with AutoDock VINA and Surflex-Dock, a preliminary re-
docking study of the 1CX2 co-crystallized ligand, the selective COX-2 inhibitor SC-558, was performed
in the 1CX2 active site. We were able to reproduce the experimental binding mode of the co-crystallized
SC-558 with a RMSD (root mean square deviation) of 1.405 Å and of 1.067Å with AutoDock VINA
and Surflex-Dock respectively. We then performed the molecular docking of our synthesized
cyclocoumarol analogues associated with the best and the worst PGE2 inhibition potency (respectively
compound 5d and compound 5g) to understand their structure-activity relationships. Both docking
software predicted similar binding mode for compound 5d with a RMSD value computed between
AutoDock VINA and Surflex-Dock poses equal to 0.576 Å. As presented in Fig. 8, the coumarin
scaffold of compound 5d is predicted to be bound in the COX-2 binding site pocket formed by Met113,
12

European Journal of Medicinal Chemistry
ACCEPTED MANUSCRIPT
Val116, Val349, Tyr355, Leu359 and Leu 531 [14] and stabilized through hydrogen bonds with Arg120
and Tyr355.
Fig. 8. Binding mode of compound 5d in the binding site of COX-2.
The methoxybenzene moiety of compound 5d is accommodated in a cavity that could not be observed
in the COX-1 structure [14] and the oxygen atom of the methoxy group is linked by hydrogen bonds to
His90 and Arg513. This ability to occupy the selective COX-2 cavity is a common feature of COX-2
selective inhibitors and could explain the COX-2 selectivity experimentally observed for compound 5d.
We re-scored the compound 5d docking pose with the HYDE scoring function resulting in a predicted
affinity in the µM to nM range (Fig. 9), consistent with experimental results (see Biological section).
13

European Journal of Medicinal Chemistry
ACCEPTED MANUSCRIPT
Fig. 9. Predicted affinity of compound 5d (A) in the nM to µM range and of compound 5g (B) in the
µM to mM range using the HYDE scoring function implemented in SeeSAR.
We then explored the binding mode obtained for compound 5g that displayed no PGE2 inhibition
potency in the biological experiments. The predicted binding mode of compound 5g was very similar to
the predicted binding mode of compound 5d (Fig. 10).
Fig. 10. Binding mode of compound 5g in the binding site of COX-2.
The coumarin scaffold of compound 5g is stabilized in the COX-2 binding site pocket formed by
Met113, Val116, Val349, Tyr355, Leu359 and Leu 531 through hydrogen bonds with Arg120 and
Tyr355 and the nitrobenzene moiety is oriented towards the COX-2 specific cavity [14]. However, nitro
group are known to be weak hydrogen bond acceptors [15], and the additional hydrogen bond observed
between compound 5d and the COX-2 specific cavity could be lost with compound 5g which could
explain the loss of experimental activity. Re-scoring the compound 5g docking pose with HYDE scoring
function led to an estimated affinity in the range of mM to µM activity (Fig. 9), which is in accordance
with the experimental results.
3.
Conclusion
In summary, based on finding selective COX-2 inhibitors for inflammation treatment, a serie of novel
cyclocoumarol derivatives was designed, synthesized, characterized and biologically evaluated. Most of
14

European Journal of Medicinal Chemistry
ACCEPTED MANUSCRIPT
the synthesized compounds exhibited significant inhibitory activity on COX-2 without any significant
inhibition on COX-1 activity. Among these, 2-methoxy-2-methyl-(1-(4-methoxyphenyl))-3,4-
dihydropyrano[3,2-c]chromen-5(2H)-one, (compound 5d), have showed the highest anti-inflammatory
activity with a dose-dependent inhibition profile. The molecular modeling study allowed us to
understand the inhibitory potential of synthesized compounds and suggested that the remarkable binding
affinity of compound 5d to COX-2 enzyme may be attributed to its ability to occupy a selective COX-2
cavity establishing additional hydrogen bonds between the oxygen of the methoxybenzene moiety and
the His90 and Arg513 of the enzyme.
Our results revealed a new type of selective COX-2 inhibitors containing pyranocoumarin skeleton and
proposed an interesting route towards the discovery of drugs with anti-inflammatory activities and may
be a challenge in the research of COX-2 inhibitory therapy. The antiproliferative potential of the best
compounds is currently being investigated and will be reported in due course.
3.
Experimental section
4.1. Chemistry
4.1.1. General
All reagents were obtained from commercial sources unless otherwise noted, and used as received.
Heated experiments were conducted using thermostatically controlled oil baths and were performed
under an atmosphere oxygen-free in oven-dried glassware when necessary. All reactions were
monitored by analytical Thin Layer Chromatography (TLC) and GC-MS analysis. TLC was performed
on aluminium sheets precoated silica gel plates (60 F254, Merck). TLC plates were visualized using
irradiation with light at 254 nm. Flash column chromatography was carried out when necessary using
silica gel 60 (particle size 0.040-0.063 mm, Merck). Separation of diastereomers was performed on an
Automate ISCO CombiFlash System RF75 PSI.
The structure of the products prepared by different methods was checked by comparison of NMR, IR
13
and MS data and by the TLC behavior. 1H and C-NMR spectra were recorded on a Bruker BioSpin
GmbH spectrometer 400 MHz, at room temperature. Chemical shifts are reported in δ units, parts per
15

European Journal of Medicinal Chemistry
ACCEPTED MANUSCRIPT
million (ppm). Coupling constants (J) are measured in hertz (Hz). Splitting patterns are designed as
followed: s, singlet; d, doublet; dd, doublet of doublets; m, multiplet; br, broad; ddd: doublet of doublet
of doublet. Various 2D techniques and DEPT experiments were used to establish the structures and to
assign the signals. GC-MS analysis were performed with an Agilent 6890N instrument equipped with a
dimethyl polysiloxane capillary column (12 m x 0.20 mm) and an Agilent 5973N MS detector-column
temperature gradient 80-300°C (method 80): 80°C (1 min); 80°C to 300°C (12.05°C/min); 300°C (2
min); Infrared spectra were recorded over the 400-4000 cm-1 range with an Agilent Technologies Cary
630 FTIR/ ATR/ ZnSe spectrometer.
Melting points were recorder on a Kofler hot block Heizbank type 7841 and were uncorrected. The
reactions using microwaves as the activating/heating source were accomplished using a microwave oven
(CEM Discover 1TM) under pressure and were monitored by ChemDriver.
X-ray diffraction data for compound 5a, 5a’ and 5d were obtained at room temperature on a Rigaku
XtaLabPro diffractometer equipped with a microfocus source (MicroMax003_Mo) and multilayer
confocal mirrors (Mo Kα radiation, λ = 0.71075 Å). Data were indexed, integrated and scaled using
CrysAlisPro [16] software suite. They were also corrected for polarization, Lorentz and absorption
effects (SCALE3 ABSPACK).
The structure was solved with the ShelXT [17] structure solution program using Direct Methods and
refined with the ShelXL [18] refinement package using Least Squares minimization. All non-hydrogen
atoms were refined with anisotropic displacement parameters and C-bound H atoms have been added
geometrically and treated as riding on their parent atoms. O-bound H atom was located in a difference
Fourier and N-H distance was restrained to 0.87 Å using DFIX command and standard uncertainty (0.02
Å).
High-resolution mass spectra (HRMS) analyses were acquired on a LTQ-Orbitrap XL from Thermo
Scientific (Thermo Fisher Scientific, Courtaboeuf, France) mass spectrometer and operated in positive
ionization mode, with a spray voltage at 3.6 kV. Detection was achieved in the Orbitrap with a
resolution set to 60,000 (at m/z 400) and a m/z range between 110-1200 in profile mode. Spectrum was
analyzed using the acquisition software XCalibur 2.1. The automatic gain control (AGC) allowed
16

European Journal of Medicinal Chemistry
ACCEPTED MANUSCRIPT
accumulation of up to 2.105 ions for FTMS scans, Maximum injection time was set to 300 ms and 1
µscan was acquired. 5µL was injected using a Thermo Finnigan Surveyor HPLC system (Thermo Fisher
Scientific, Courtaboeuf, France) with a continuous infusion of methanol at 100 µL.min-1.
4.1.2 Synthesis and characterization
Synthesis procedures of final compounds 5a-g and 5a’-g’ are described below. All the experimental data
1
and detailed attribution of the different H and 13C signals of compounds 4a-g are available in the
Supplementary material section.
Fig. 11. Convention adopted to assign signals of 1H and 13C-NMR spectra
4.1.2.1. 2-Methoxy-2-methyl-4-phenyl-3,4-dihydropyrano[3,2-c]chromen-5(2H)-one 5a. Hydroxy-3-(3-
oxo-1-phenylbutyl)-2H-chromen-2-one (1 equiv., 100 mg, 0.325 mmol, 4a), was refluxed in HCl 4%
MeOH (10 equiv., 3.25 mmol) for 23 h (monitoring by TLC, cyclohexane, AcOEt :7/3). Then the
mixture was cooled down at room temperature and 1 mL of distilled water was added, followed by
extraction with AcOEt. The diastereomers were separated by flash chromatography (eluent:
cyclohexane/AcOEt : 8/2) and were isolated to afford 60 mg (53%) of the major diastereomer 5a as a
white powder and 10 mg (9%) of the minor diastereomer 5a’ as a white powder too. 5a: H14 in axial
position, Rf = 0.33, cyclohexane/AcOEt : 8/2; Mp: 165-166°C; 1H NMR (CDCl3, 300 MHz) : δ (ppm):
1.70 (s, 3H, H21), 2.01 (dd, 1H, H13ax, Jgem = 14.0 Hz, J13ax-14ax = 11.8 Hz ), 2.52 (dd, 1H, H13eq,
Jgem = 14.0 Hz, Jeq-ax = 6.9 Hz ), 3.36 (s, 3H, H22), 4.16 (dd ,1H, H14ax, Jax-ax = 11.8 Hz, Jax-eq =
6.9 Hz ), 7.24-7.26 (m, 2H, H16, H20), 7.28-7.35 (m, 5H, H6, H8, H17, H18, H19), 7.53-7.58 (ddd, 1H
13
,H7, J = 9.4, 7.4, 1.7 Hz), 7.91 (dd, 1H ,H5, J = 8.0, 1.3 Hz). C NMR (CDCl3, 75 MHz) : δ (ppm):
22.35 (C21), 35.60 (C14), 43.52 (C13 ), 49.94 (C22), 101.44 (C3), 105.63 (C12), 115.80 (C10), 116.72
(C8), 122.46 (C5), 123.84 (C6), 126.54 (C18), 127.16 (C16, C20 or C17, C19), 128.72 (C17, C19 or
17

European Journal of Medicinal Chemistry
ACCEPTED MANUSCRIPT
C16, C20), 131.62 (C7), 143.55 (C15), 153.14 (C9), 158.25 (C4), 161.19 (C2). NMR analysis are in
agreement with literature [10] HRMS (ESI): m/z calcd. for C20H18O4 [M+Na]+ : 345,1097; found :
345,1096. Minor Diastereomer 5a’: H14 in equatorial position Rf = 0.29, cyclohexane/AcOEt : 8/2; 1H
NMR (CDCl3, 300 MHz) : δ (ppm): 1.69 (s, 3H, H21), 2.38 (dd, 1H, H13-ax, Jgem = 14.0 Hz, J13ax-
14eq = 7.4 Hz ), 2.52 (dd, 1H, H13eq, Jgem = 14.1 Hz, J13eq-eq = 3.8 Hz), 3.29 (s, 3H, H22), 4.19 (dd
,1H, H14eq, Jeq-ax = 7.2 Hz, Jeq-eq = 3.7 Hz), 7.21-7.32 (m, 5H, H16, H17, H18, H19, H20), 7.34-
7.43 (m, 2H, H6, H8), 7.60-7.62 (m, 1H, H7), 7.96 (dd, 1H, H5, J = 7.8, 1.2 Hz). 13C NMR (CDCl3, 75
MHz) : δ (ppm): 22.58 (C21), 35.45 (C14), 39.93 (C13 ), 49.41 (C22), 102.58 (C3), 103.30
(C12),115.89 (C10), 116.86 (C8), 122.67 (C5), 123.97(C6), 126.26 (C18), 127.52 (C16, C20 or C17,
C19), 128.21 (C17, C19 or C16, C20), 131.87 (C7), 143.18 (C15), 153.11 (C9), 159.07 (C4), 162.00
(C2).
4.1.2.2. 2-Methoxy-2-methyl-(1-(p-tolyl))-3,4-dihydropyrano[3,2-c]chromen-5(2H)-one 5b. 4-hydroxy-
3-(3-oxo-1-p-tolylbutyl)-2H-chromen-2-one (1 equiv., 100 mg, 0.31 mmol, 4b), was refluxed in HCl
4% MeOH (10 equiv., 3.1 mmol) for 31 h (monitoring by TLC, cyclohexane, AcOEt :7/3). Then the
mixture was cooled down at room temperature and 1 mL of distilled water was added, followed by
extraction with AcOEt. The diastereomers were separated by flash chromatography (eluent:
cyclohexane/AcOEt : 8/2 and were isolated to afford 48 mg (46%) of the major diastereomer 5b as a
yellow powder and 13 mg (13%) of the minor diastereomer 5b’ as a yellow oil. 5b: Rf = 0.42,
1
cyclohexane/AcOEt : 8/2; Mp: 164-165°C; H NMR (CDCl3, 300 MHz) : δ (ppm): 1.82 (s, 3H, H21),
2.13 (t, 1H, H13ax, Jgem = 13.7, J13ax-14ax = 12.0 Hz), 2.46 (s, 3H, H23), 2.62 (dd, 1H, H13eq, Jgem
= 14.0 Hz, J13eq-14ax = 6.9 Hz ), 3.48 (s, 3H, H22), 4.24 (dd, 1H, H14-ax, Jax-ax = 11.7 Hz, Jax-eq =
6.9 Hz ), 7.26 (m, 4H, H16, H17, H19, H20), 7.40-7.47 (m, 2H, H6, H8), 7.67 (t, 1H, H7, J = 7.6 Hz),
13
8.03 (d, 1H, H5, J = 7.7 Hz). C NMR (CDCl3, 75 MHz) : δ (ppm): 21.19 (C23 ), 22.35 (C21), 35.16
(C14), 43.58 (C13 ), 49.92 (C22), 101.43 (C3), 105.78 (C12), 115.82 (C10), 116.70 (C8), 122.44 (C5),
123.80 (C6), 127.01 (C16, C20 or C17, C19), 129.45 (C17, C19 or C16, C20), 131.56 (C7), 135.96
(C18), 140.47 (C15), 153.12 (C9), 158.15 (C4), 161.21 (C2). HRMS (ESI): m/z calcd. for C21H20O4
18

European Journal of Medicinal Chemistry
ACCEPTED MANUSCRIPT
1
[M+Na]+ : 359,1254 ; found : 359,1245 . 5b’: Rf = 0.35, cyclohexane/AcOEt : 8/2; H NMR (CDCl3,
300 MHz) : δ (ppm): 1.61 (s, 3H, H21), 2.48 (s, 3H, H23), 2.53-2.65 (m, 2H, H13ax , H13eq), 3.43 (s,
3H, H22), 4.27-4.30 (m, 4H, H16, H17, H19, H20), 7.48-7.55 (m, 2H, H6, H8), 7.74 (t, 1H, H7, J = 7.4
13
Hz), 8.08 (d, 1H, H5, J = 7.7 Hz). C NMR (CDCl3, 75 MHz) : δ (ppm): 21.21 (C23 ), 22.59 (C21),
35.16 (C14), 40.05 (C13 ), 49.42 (C22), 102.57 (C3), 103.54 (C12),115.79 (C10), 116.85 (C8), 122.65
(C5), 123.94 (C6), 127.38 (C16, C20 or C17, C19), 129.02 (C17, C19 or C16, C20), 131.81 (C7),
135.68 (C18), 140.18 (C15), 153.10 (C9), 158.96 (C4), 161.99 (C2).
4.1.2.3. 2-Methoxy-2-methyl-(1-(4-tert-butylphenyl)-3,4-dihydropyrano[3,2-c]chromen-5(2H)-one 5c.
3-(1-(4-tert-butylphenyl)-3-oxobutyl)-4-hydroxy-2H-chromen-2-one (1 equiv., 100 mg, 0.27 mmol, 4c),
was refluxed in HCl 4% MeOH (10 equiv., 2.7 mmol) for 52 h (monitoring by TLC, cyclohexane,
AcOEt :7/3). Then the mixture was cooled down at room temperature and 1 mL of distilled water was
added, followed by extraction with AcOEt. The diastereomers were separated by flash chromatography
(eluent: cyclohexane/AcOEt : 8/2 ) and were isolated to afford 55 mg (57%) of the major diastereomer
5c as a white powder and 15 mg (16%) of the minor diastereomer 5c’ as a white powder too. 5c : Rf =
1
0.48, cyclohexane/AcOEt : 8/2; Mp: 134-135°C ; H NMR (CDCl3, 300 MHz) : δ (ppm): 1.26 (s, 9H,
H23), 1.64 (s, 3H, H21), 1.96 (t, 1H, H13ax, Jgem = J13ax-14ax = 13.8 Hz), 2.45 (dd, 1H, H13-eq,
Jgem = 14.0 Hz, J13eq-14ax = 7.0 Hz), 3.30 (s, 3H, H22), 4.07 (dd , 1H, H14ax, Jax-ax = 11.6 Hz, Jax-
eq = 7.1 Hz ), 7.10-7.12 (m, 2H, H16, H20), 7.26-7.29 (m, 4H, H6 ,H8, H17, H19), 7.50 (t, 1H, H7, J =
7.7 Hz), 7.85 (d, 1H, H5, J = 8.3 Hz). 13C NMR (CDCl3, 75 MHz) : δ (ppm): 22.36 (C21), 31.51 (C23),
35.07 (C14), 43.59 (C13), 49.92 (C22), 101.46 (C3), 105.87 (C12), 115.86 (C10), 116.70 (C8), 122.46
(C5), 123.80 (C6), 125.62 (C17, C19), 126.75 (C16, C20), 131.56 (C7), 140.32 (C15), 149.06 (C18),
153.12 (C9), 158.12 (C4), 161.23 (C2). HRMS (ESI): m/z calcd. for C24H26O4 [M+Na]+ : 401.1723;
found: 401.1728, 5c’: Rf = 0.39, cyclohexane/AcOEt : 8/2; 1H NMR (CDCl3, 300 MHz) : δ (ppm): 1.28
(s, 9H, H23), 1.63 (s, 3H, H21), 2.29 (dd, 1H, H13-ax, Jgem = 14.1 Hz, J13ax-14ax = 7.4 Hz), 2.45 (dd,
1H, H13eq, Jgem = 14.0 Hz, J13eq-14eq = 4.1 Hz), 3.25 (s, 3H, H22), 4.10-4.12 (m, 1H, H14eq), 7.09-
7.12 (m, 2H, H16, H20), 7.25-7.37 (m, 4H, H6 , H8, H17, H19), 7.56 (tb, 1H, H7, J = 7.3 Hz), 7.90 (d,
19

European Journal of Medicinal Chemistry
ACCEPTED MANUSCRIPT
13
1H, H5, J = 7.8 Hz). C NMR (CDCl3, 75 MHz) : δ (ppm): 22.56 (C21 ), 31.54 (C23), 35.18 (C14),
39.97 (C13 ), 49.46 (C22), 102.68 (C3), 103.61 (C12),115.82 (C10), 116.83 (C8), 122.69 (C5), 123.93
(C6), 125.18 (C17, C19), 127.07 (C16, C20), 131.80 (C7), 139.98 (C15), 148.70 (C18), 153.09 (C9),
158.97 (C4), 162.00 (C2).
4.1.2.4. 2-Methoxy-2-methyl-(1-(4-methoxyphenyl))-3,4-dihydropyrano[3,2-c]chromen-5(2H)-one 5d.
4-hydroxy-3-(1-(4-methoxyphenyl)-3-oxobutyl)-2H-chromen-2-one (1 equiv., 100 mg, 0.296 mmol,
4d), was refluxed in HCl 4% MeOH (10 equiv., 1 g, 2.96 mmol) for 66 h (monitoring by TLC,
cyclohexane, AcOEt :7/3). Then the mixture was cooled down at room temperature and 1 mL of
distilled water was added, followed by extraction with AcOEt. The major diastereomer 5d was purified
by flash chromatography (eluent : cyclohexane/AcOEt : 8/2) and was isolated to afford 66 mg (63 %) as
a white powder. 5d: Rf = 0.39, cyclohexane/AcOEt : 8/2; Mp: 168 °C; 1H NMR (CDCl3, 300 MHz) : δ
(ppm): 1.67 (s, 3H, H21), 1.96 (dd, 1H, H13ax, Jgem = 14.0 Hz, J13ax-14ax = 11.7 Hz), 2.46 (dd, 1H,
H13eq, Jgem = 14.0 Hz, J13eq-14ax = 7.0 Hz), 3.33 (s, 3H, H22), 3.77 (s, 3H, H23), 4.08 (dd, 1H,
H14ax, Jax-ax = 11.9 Hz, Jax-eq = 7.2 Hz ), 6.82-6.85 (m, 2H, H17, H19),7.13-7.16 (m, 2H, H16, H20),
7.26-7.32 (m, 2H, H6, H8), 7.53 (t, 1H, H7, J = 8.6 Hz), 7.89 (d, 1H, H5, J = 6.7 Hz). 13C NMR (CDCl3,
75 MHz) : δ (ppm): 22.36 (C21), 34.75 (C14), 43.56 (C13), 49.92 (C22 ), 55.31 (C23), 101.49 (C3),
105.86 (C12), 114.17 (C17, C19), 115.83 (C10), 116.70 (C8), 122.46 (C5), 123.81 (C6), 128.12 (C16,
C20), 131.57 (C7), 135.46 (C15), 153.12 (C9), 158.08 (C18), 158.21 (C4), 161.20 (C2). HRMS (ESI):
m/z calcd. for C21H20O5 [M+Na]+ 375,1203; found: 375,1203.
4.1.2.5. 2-Methoxy-2-methyl-(1-(4-fluorophenyl))-3,4-dihydropyrano[3,2-c]chromen-5(2H)-one 5e. 3-
(1-(4-fluorophenyl)-3-oxobutyl)-4-hydroxy-2H-chromen-2-one (1 equiv., 100 mg, 0.29 mmol, 4e), was
refluxed in HCl 4% MeOH (10 equiv., 2.90 mmol) for 31 h (monitoring by TLC, cyclohexane, AcOEt
:7/3). Then the mixture was cooled down at room temperature and 1 mL of distilled water was added,
followed by extraction with AcOEt. The diastereomers were separated by flash chromatography (eluent:
cyclohexane/AcOEt : 8/2 ) and were isolated to afford 47 mg (44%) of the major diastereomer 5e as a
white powder and 21 mg (19%), of the minor diastereomer 5e’ as a yellow powder. 5e: Rf = 0.34,
20

European Journal of Medicinal Chemistry
ACCEPTED MANUSCRIPT
cyclohexane/AcOEt : 8/2; Mp: 164-166°C; 1H NMR (CDCl3, 300 MHz) : δ (ppm): 1.98 (s, 3H, H21),
2.24 (dd, 1H, H13ax, Jgem = 14.0 Hz, J13ax-14ax = 12 Hz), 2.75 (dd, 1H, H13eq, Jgem = 14.0 Hz,
J13eq-14ax = 6.9 Hz), 3.62 (s, 3H, H22), 4.41 (dd ,1H, H14ax, Jax-ax = 11.8 Hz, Jax-eq = 6.9 Hz ),
7.23-7.29 (m, 2H, H17, H19), 7.45-7.49 (m, 2H, H16, H20), 7.55-7.62 (m, 2H, H6, H8), 7.82 (m, 1H,
H7), 8.17 (m, 1H, H5,). 13C NMR (CDCl3, 75 MHz) : δ (ppm) : 22.31 (C21 ), 34.92 (C14), 43.50 (C13),
49.96 (C22 ), 101.45 (C3), 105.40 (C12), 115.51 (d, C17, C19, 2JC-F = 21.8 Hz), 115.71 (C10), 116.72
(C8), 122.51 (C5), 123.92 (C6), 128.60 (d, C16, C20, 3JC-F = 8.3 Hz), 131.74 (C7), 139.16 (C15),
153.12 (C9), 159.14 (d, C18, JC-F = 120.8 Hz), 161.17 (C4), 163.17 (C2). HRMS (ESI): m/z calcd. for
1
C20H17FO4 [M+Na]+ 363,1003; found 363,1003. 5e: Rf = 0.27, cyclohexane/AcOEt : 8/2; H NMR
(CDCl3, 300 MHz) : δ (ppm): 1.96 (s, 3H, H21), 2.64 (dd, 1H, H13ax, Jgem = 14.1 Hz, J13ax-14eq =
7.3 Hz), 2.75 (dd, 1H, H13eq, Jgem = 14.1 Hz, J13eq-14eq = 3.0 Hz), 3.54 (s, 3H, H22), 4.45 (m ,1H,
H14eq ), 7.23-7.29 (m, 2H, H17, H19), 7.45-7.49 (m, 2H, H16, H20), 7.63-7.70 (m, 2H, H6, H8), 7.89
(t, 1H, H7, J = 7.3 Hz), 8.23 (d, 1H, H5, J = 7.8 Hz). 13C NMR (CDCl3, 75 MHz) : δ (ppm): 22.63 (C21
), 34.57 (C14), 39.87 (C13), 49.38 (C22 ), 102.42 (C3), 103.12 (C12), 115.00 (d, C17, C19, 2JC-F =
21.8 Hz), 115.65 (C10), 116.88 (C8), 122.68 (C5), 124.07 (C6), 129.0 (d, C16, C20, 3JC-F = 8.3 Hz),
132.00 (C7), 138.87 (C15), 153.09 (C9), 159.42 (C18), 162.02 (C4), 163.01 (C2).
4.1.2.6. 2-Methoxy-2-methyl-(1-(4-trifluoromethyl)phenyl))-3,4-dihydropyrano[3,2-c]chromen-5(2H)-
one 5f. 4-hydroxy-3-(3-oxo-1-(4-(trifluoromethyl)phenyl)butyl)-2H-chromen-2-one (1 equiv., 100 mg,
0.27 mmol, 4f), was refluxed in HCl 4% MeOH (10 equiv, 2.7 mmol) for 23 h (monitoring by TLC,
cyclohexane, AcOEt :7/3). Then the mixture was cooled down at room temperature and 1 mL of
distilled water was added, followed by extraction with AcOEt. The diastereomers were separated by
flash chromatography (eluent: cyclohexane/AcOEt : 8/2 ) and were isolated to afford 42 mg (38%) of
the major diastereomer 5f and 9 mg (8%) of the minor diastereomer 5f’ as a yellow powders. Mp: 180-
182°C; 13C NMR (CDCl3, 75 MHz) : δ (ppm) = 22.32 (C21), 35.61 (C14), 43.26 (C13), 50.05 (C22),
101.40 (C3), 104.80 (C12), 115.66 (C10), 116.85 (C8), 122.56 (C5), 124.04 (C6), 128.8 (q, C18, 2JC-F
= 32.2 Hz), 125.75 (q, C17, C19, 3JC-F = 3.8 Hz), 125.75 (C16, C20), 131.96 (C7), 147.86 (C15),
21

European Journal of Medicinal Chemistry
ACCEPTED MANUSCRIPT
153.19 (C9), 158.72 (C4), 161.23 (C2), C23-F non visible. HRMS (ESI): m/z calcd. for C21H17F3O4
[M+Na]+ : 413,0971; found: 413,0980, 5f’: Rf = 0.27, cyclohexane/AcOEt : 8/2: 1H NMR (CDCl3, 300
MHz) : δ (ppm): 1.65 (s, 3H, H21), 2.35 (dd, 1H, H13ax, Jgem = 14.2 Hz, J13ax-14eq = 7.4 Hz), 2.48
(dd, 1H, H13-eq, Jgem = 14.2 Hz, J13eq-14eq = 2.8 Hz), 3.19 (s, 3H, H22), 4.19 (m, 1H, H14-eq), 7.29
(d, 2H, H16, H20, J= 8.3 Hz ), 7.34-7.39 (m, 2H, H6, H8), 7.50 (d, 2H, H17, H19, J= 8.1 Hz ), 7.57 (td,
13
1H, H7, J = 7.4 Hz, 1.6 Hz), 7.92 (dd, 1H, H5, J = 7.9 Hz, 1.5 Hz). C NMR (CDCl3, 75 MHz) : δ
(ppm): 22.62 (C21), 35.00 (C14), 39.60 (C13), 49.33 (C22 ), 102.30 (C3), 102.47 (C12), 115.57 (C10),
116.99 (C8), 122.71 (C5), 124.18 (C6), 125.15 (q, C17, C19, 3JC-F = 3.8 Hz), 128.00 (C16, C20),
132.19 (C7), 147.46 (C15), 153.17 (C9), 159.31 (C4), 162.07(C2), C18 et C23-F non visible.
4.1.2.7. 2-Methoxy-2-methyl-(1-(4-nitrophenyl))-3,4-dihydropyrano[3,2-c]chromen-5(2H)-one 5g. 4-
hydroxy-3-(1-(4-nitrophenyl)-3-oxobutyl)-2H-chromen-2-one (1 equiv., 100 mg, 0.283 mmol, 8g), was
refluxed in HCl 4% MeOH (10 equiv., 2.83 mmol) for 22 h (monitoring by TLC, cyclohexane, AcOEt
:7/3). Then the mixture was cooled down at room temperature and 1 mL of distilled water was added,
followed by extraction with AcOEt. The diastereomers were separated by flash chromatography (eluent:
cyclohexane/AcOEt : 8/2 ) and were isolated to afford 63 mg (64%) of the major diastereomer 7g as a
white powder and 11 mg (12%) of the minor diastereomer 7g’ as a yellow powder. 7g: Rf = 0.24,
1
cyclohexane/AcOEt : 8/2; Mp: 204-206°C; H NMR (CDCl3, 300 MHz) : δ (ppm): 1.56 (s, 3H, H21),
1.95 (dd, 1H, H13ax, Jgem = 13.9 Hz, J13ax-14ax= 12.1 Hz), 2.48 (dd, 1H, H13eq, Jgem = 13.9 Hz,
J13eq-14ax= 6.8 Hz), 3.35 (s, 3H, H22), 4.24 (dd ,1H, H14-ax, Jax-ax = 12.1 Hz, Jax-eq = 6.8 Hz ),
7.30-7.35 (m, 2H, H6, H8), 7.38 (d, 2H, H16, H20, J = 8.8 Hz), 7.57 (ddd, 1H, H7 , J =9.4 Hz, 7.3 Hz,
1.6 Hz), 7.89 (dd, 1H, H5, J = 9.7 Hz, 1.3 Hz), 8.16 (d, 2H, H17, H19, J = 8.8 Hz ). 13C NMR (CDCl3,
75 MHz) : δ (ppm): 22.31 (C21), 35.72 (C14), 42.91 (C13), 50.11 (C22), 101.36 (C3), 104.29 (C12),
115.54 (C10), 116.90 (C8), 122.60 (C5), 124.15 (C6 ,C17, C19), 128.11 (C16, C20), 132.16 (C7),
146.79 (C18), 151.65 (C15), 153.20 (C9), 158.95 (C4), 161.21(C2). HRMS (ESI): m/z calcd. for
1
C20H17NO6 [M+Na]+: 390,0948; found: 390,0950. 7g’: Rf = 0.16, cyclohexane/AcOEt : 8/2; H NMR
(CDCl3, 300 MHz) : δ (ppm): 1.66 (s, 3H, H21), 2.37 (dd, 1H, H13ax, Jgem = 14.2 Hz, J13ax-14eq =
22

European Journal of Medicinal Chemistry
ACCEPTED MANUSCRIPT
7.4 Hz), 2.50 (dd, 1H, H13eq, Jgem = 14.2 Hz, J13eq-14eq = 2.5 Hz), 3.17 (s, 3H, H22), 4.22 (dd , 1H,
H14eq, Jeq-ax = 7.3 Hz, Jeq-eq = 2.5 Hz ), 7.34 (d, 2H, H16, H20, J = 8.4 Hz ), 7.38-7.41 (m, 2H, H6,
H8), 7.61 (ddd, 1H, H7, J =9.2, 7.3, 1.6 Hz), 7.92 (dd, 1H, H5, J = 7.9 Hz, 1.5 Hz), 8.13 (d, 2H, H17,
H19, J = 8.8 Hz). 13C NMR (CDCl3, 75 MHz) : δ (ppm): 22.61 (C21), 34.92 (C14), 39.45 (C13), 49.32
(C22), 101.94 (C3), 102.18(C12), 116.06 (C10), 117.06 (C8), 122.75 (C5), 123.46 (C17, C19), 124.30
(C6), 128.54 (C16, C20), 132.40 (C7), 146.43 (C18), 151.29 (C15), 153.18 (C9), 159.12 (C4), 162.10
(C2).
4.1.3 X-ray analysis
The crystallographic analysis of 2-methoxy-2-methyl-4-phenyl-3,4-dihydropyrano[3,2-c]chromen-
5(2H)-one 5a, 5a’ and 2-methoxy-2-methyl-(1-(4-methoxyphenyl))-3,4-dihydropyrano[3,2-c]chromen-
5(2H)-one 5d are described below. The complete characterization containing the crystal data and
structure refinement is available in the Supplementary material section.
The relative configuration of 2-methoxy-2-methyl-4-phenyl-3,4-dihydropyrano[3,2-c]chromen-5(2H)-
one 5a was characterized as (S, S) and (R, R). Crystal Data for 5a (M =322.34 g/mol): monoclinic,
space group P21/n (no. 14), a = 5.8669(3) Å, b = 16.7563(8) Å, c = 16.4159(9) Å, β = 94.830(5)°, V =
1608.08(15) Å3, Z = 4, T = 292.7(5) K, µ(MoKα) = 0.092 mm-1, Dcalc = 1.331 g/cm3, 16505 reflections
measured (6.962° ≤ 2Θ ≤ 60.034°), 4107 unique (Rint = 0.0458, Rsigma = 0.0464) which were used in all
calculations. The final R1 was 0.0478 (I > 2σ(I)) and wR2 was 0.1397 (all data) (Fig. 12).
23

European Journal of Medicinal Chemistry
ACCEPTED MANUSCRIPT
Fig.12. ORTEP-3 plot of major diastereomer 5a. Ellipsoids are drawn at the 50% probability level and
H atoms are shown as spheres of arbitrary radius.
The relative configuration of 2-methoxy-2-methyl-4-phenyl-3,4-dihydropyrano[3,2-c]chromen-5(2H)-
one 5a’ was characterized as (S, R) and (R, S). Crystal Data for 5a’ (M =322.34 g/mol): monoclinic,
space group I2/a (no. 15), a = 20.1207(12) Å, b = 8.5036(4) Å, c = 19.9586(14) Å, β = 107.014(7)°, V =
3265.5(4) Å3, Z = 8, T = 292.6(4) K, µ(MoKα) = 0.091 mm-1, Dcalc = 1.311 g/cm3, 18115 reflections
measured (7.124° ≤ 2Θ ≤ 60.164°), 4158 unique (Rint = 0.0587, Rsigma = 0.0579) which were used in all
calculations. The final R1 was 0.0549 (I > 2σ(I)) and wR2 was 0.1328 (all data) (Fig. 13).
Fig. 13. ORTEP-3 plot of 5a’. Ellipsoids are drawn at the 50% probability level and H atoms are shown
as spheres of arbitrary radius.
The relative configuration of the 2-methoxy-2-methyl-(1-(4-methoxyphenyl))-3,4-dihydropyrano[3,2-
c]chromen-5(2H)-one (5d), major diastereomer was characterized as (S, S) and (R, R). Crystal Data for
5d (M =352.37 g/mol): monoclinic, space group C2/c (no. 15), a = 23.694(2) Å, b = 5.6932(4) Å, c =
26.775(3) Å, β = 102.720(9), V = 3523.2(6) Å3, Z = 8, T = 292.5(7) K, µ(MoKα) = 0.095 mm-1, Dcalc =
1.329 g/cm3, 19052 reflections measured (7.372° ≤ 2Θ ≤ 59.93°), 4454 unique (Rint = 0.0552, Rsigma
=
0.0590) which were used in all calculations. The final R1 was 0.0490 (I > 2σ(I)) and wR2 was 0.1362
(all data) (Fig. 14).
24

European Journal of Medicinal Chemistry
ACCEPTED MANUSCRIPT
Fig. 14. ORTEP-3 plot of 5d. Ellipsoids are drawn at the 50% probability level and H atoms are shown
as spheres of arbitrary radius.
Crystallographic data for 2-methoxy-2-methyl-4-phenyl-3,4-dihydropyrano[3,2-c]chromen-5(2H)-one
5a and 5a’ and 2-methoxy-2-methyl-(1-(4-methoxyphenyl))-3,4-dihydropyrano[3,2-c]chromen-5(2H)-
one 5d, have been deposited with the Cambridge Crystallographic Data Centre (deposit no. CCDC
1563254-1563256). Copies of the data can be obtained, free of charge, from The Cambridge
Crystallographic Data Centre via www.ccdc.cam.ac.uk/data_request/cif.
4.1.4 Biology
4.1.4.1. Assay of COX-2 activity
Raw 264.7 murine macrophages (ATCC TIB 71; American Type Culture Collection, Rockville,MD,
USA) were maintained in DMEM supplemented with 10% (v/v) heat-inactivated fetal calf serum, 100
U/mL penicillin and 100 µg/mL streptomycin. The cells were grown in a humidified incubator at 37 °C
and 5% CO2. Raw 264.7 cells were seeded at 2.105 cells/well during 24 h, then 10 µM of each
synthesized cyclocoumarol derivatives was added. After 2 h, LPS (10 ng/mL) was added in the culture
medium during 24 h. The PGE2 levels were quantified in culture media supernatants from treated and
control cells by enzyme immunoassay using an EIA Kit (Cayman Chemical) as previously described
25

European Journal of Medicinal Chemistry
ACCEPTED MANUSCRIPT
[19].The results were expressed by the standard deviation average of ± 8 independent experiments. NS-
398, a well-known COX-2 selective inhibitor was used as reference.
4.1.4.2. Assay of COX-1 activity
The same protocol described before was used except for the absence of LPS and deprivation of SVF
(1%) to measure the COX-1 activity. The PGE2 levels were quantified in culture media supernatants
from treated and control cells by enzyme immunoassay using an EIA Kit (Cayman Chemical). The
results were expressed by the standard deviation average of ± 6 independent experiments.
4.1.4.3 Assay of dose-dependent response for compound 5d
The same protocol described in Assay of COX-2 activity section was used except for the concentration
of the molecules (1-20 µM). The PGE2 levels were quantified in culture media supernatants from treated
and control cells by enzyme immunoassay using an EIA Kit (Cayman Chemical). The results were
expressed by the standard deviation average of ± 6 independent experiments.
4.1.4.4. Protein extraction and Western-Blot analysis
For total protein extraction, Raw 264.7 cells were washed in PBS, then, the total cell pool was
centrifuged at 200 g for 5 min at 4°C and homogenized in RIPA lysis buffer (50 mM HEPES, pH 7.5,
150 mM NaCl, 1% sodium deoxycholate, 1% NP-40, 0.1% SDS, 20 mg/mL of aprotinin) containing
protease inhibitors (CompleteTM Mini, Roche Diagnostics) according to the manufacturer’s
instructions. Proteins (10-100 µg) were separated by electrophoresis on 10% SDS–PAGE gels and
transferred to polyvinylidene fluoride (PVDF) membranes (Amersham Pharmacia Biotech, Saclay,
France) and probed with human primary antibodies. Human COX-1 and COX-2 antibodies were
purchased from Cayman Chemical (Bertin Pharma, Montigny le Bretonneux, France). After incubation
with secondary antibodies (Dako France S.A.S., Trappes, France), blots were developed using the ECL
Plus Western Blotting Detection System (Amersham Pharmacia Biotech) and G: BOX system (Syngene,
26

European Journal of Medicinal Chemistry
ACCEPTED MANUSCRIPT
Ozyme, Saint Quentin en Yvelines, France). Membranes were then reblotted with human anti-β-actin
(Sigma Aldrich, Saint Quentin Fallavier, France) used as a loading control.
4.1.5. Molecular modeling study
Coordinates of Mus musculus COX-2 in complex with a selective COX-2 inhibitor, SC-558, were
retrieved from the X-ray structure available in the Protein Data Bank (PDB) [20] (accession code 1CX2,
resolution 3.0 Å). Only chain A was retained. The ligand bound in the active site was removed and the
protein structure was prepared for docking studies using the DockPrep tool of UCSF CHIMERA [21] to
delete the co-crystallized water molecules, repair truncated side chains, protonate the protein residues
and assign atomic partial charges using the AMBER force field ff14SB [22]. The 3D structures of
compounds 5a-g were generated with Corina online demo [23]. OpenBabel [24] was used to calculate
Gasteiger’s partial atomic charges. Compounds 5a-g were then docked individually using AutoDock
VINA [25] with default parameters. To confirm the predicted docking poses obtained with AutoDock
VINA, we used Surflex-Dock v. 2.5 [26], with the +premin and +remin options. SeeSAR v6.0 [27] was
then used to re-score the docked poses using the HYDE scoring function [28].
4.1.6. Statistical analysis
Data are expressed as the arithmetic means ± standard error of the mean (SEM) of separate experiments.
The statistical significance of results obtained from in vitro studies was evaluated by the two tailed
unpaired Student's t-test, with p<0.05 being considered as significant.
Acknowledgements
The authors acknowledge the Ministry of National Education, Research and Technology (MENRT) for
the graduate fellowship awarded to Anita Rayar. Authors also thank Pr Marie-Claude Viaud-Massuard
and Dr Cécile Croix for their help in the use of the CombiFlash system. The authors would like to thank
Pr Ajay Jain for providing Surflex and BiosolveIT Gmbh for providing SeeSAR.
27

European Journal of Medicinal Chemistry
ACCEPTED MANUSCRIPT
Supplementary data
Supplementary data related to this article can be found http://dx.doi.org/XX.XXXX/j.ejmech.2017
.XX.XXX
References
[1]
[2]
[3]
[4]
F. Celotti, S. Laufer, Anti-inflammatory drugs: new multitarget compounds to face an old problem. The
dual inhibition concept, Pharmacol. Res. 43 (2001) 429-436.
A.L. Blobaum, L.J. Marnett, Structural and Functional Basis of Cyclooxygenase Inhibition, J. Med.
Chem. 50 (2007) 1425-1441.
D.L. Simmons, R.M. Botting, T. Hla, Cyclooxygenase Isozymes: The Biology of Prostaglandin Synthesis
and Inhibition, Pharmacol. Rev. 56 (2004) 387.
S. Harirforoosh, W. Asghar, F. Jamali, Adverce effects of nonsteroidal antiinflammatory drugs: An
update of gastrointestinal, cardiovascular and renal complications, J. Pharm. Pharma. Sci. 16 (2013) 821-
847.
[5]
[6]
A.-M. Rayar, N. Lagarde, C. Ferroud, J.-F. Zagury, M. Montes, M. Sylla-Iyarreta Veitía, Update on
COX-2 Selective Inhibitors: Chemical Classification, Side Effects and their Use in Cancers and Neuronal
Diseases, Curr. Top. Med. Chem. 17 (2017) 1-22.
M. Sylla-Iyarreta Veitía, D. Siverio Mota, V. Lerari, M. Marín, R. M. Giner, L. Vicet Muro, Y. Rivero
Guerra, F. Dumas, C. Ferroud, P. A. M. de Witte, A. D. Crawford, V. J. Arán, Y. Marrero Ponce, Fishing
Anti-Inflammatories from Known Drugs: In Silico Repurposing, Design, Synthesis and Biological
Evaluation of Bisacodyl Analogues, Curr. Top. Med. Chem. 17 (2017). 2866 - 2887.
A. Rayar, M.S.-I. Veitia, C. Ferroud, An efficient and selective microwave-assisted Claisen-Schmidt
reaction for the synthesis of functionalized benzalacetones, SpringerPlus 4 (2015) 221.
M. Ikawa, M.A. Stahmann, K.P. Link, Studies on 4-Hydroxycoumarins. V. The Condensation of α,β-
Unsaturated Ketones with 4-Hydroxycoumarin1, J. Am. Chem. Soc. 66 (1944) 902-906.
W.M. Barker, M.A. Hermodson, K.P. Link, 4-Hydroxycoumarins. Synthesis of the metabolites and some
other derivatives of warfarin, J. Med. Chem. 14 (1971) 167-169.
[7]
[8]
[9]
28