Novel coumarin- and quinolinone-based polycycles as cell division cycle 25-A and -C phosphatases inhibitors induce proliferation arrest and apoptosis in cancer cells

Article abstract of DOI:10.1016/j.ejmech.2017.04.012

Cell division cycle phosphatases CDC25 A, B and C are involved in modulating cell cycle processes and are found overexpressed in a large panel of cancer typology. Here, we describe the development of two novel quinone-polycycle series of CDC25A and C inhi

Full text of DOI:10.1016/j.ejmech.2017.04.012

Accepted Manuscript
Novel coumarin- and quinolinone-based polycycles as cell division cycle 25-A and -C
phosphatases inhibitors induce proliferation arrest and apoptosis in cancer cells
Clemens Zwergel, Brigitte Czepukojc, Emilie Evain-Bana, Zhanjie Xu, Giulia Stazi,
Mattia Mori, Alexandros Patsilinakos, Antonello Mai, Bruno Botta, Rino Ragno,
Denise Bagrel, Gilbert Kirsch, Peter Meiser, Claus Jacob, Mathias Montenarh, Sergio
Valente
PII:
S0223-5234(17)30267-2
10.1016/j.ejmech.2017.04.012
EJMECH 9359
DOI:
Reference:
To appear in: European Journal of Medicinal Chemistry
Received Date: 19 December 2016
Revised Date: 3 April 2017
Accepted Date: 6 April 2017
Please cite this article as: C. Zwergel, B. Czepukojc, E. Evain-Bana, Z. Xu, G. Stazi, M. Mori, A.
Patsilinakos, A. Mai, B. Botta, R. Ragno, D. Bagrel, G. Kirsch, P. Meiser, C. Jacob, M. Montenarh,
S. Valente, Novel coumarin- and quinolinone-based polycycles as cell division cycle 25-A and -C
phosphatases inhibitors induce proliferation arrest and apoptosis in cancer cells, European Journal of
Medicinal Chemistry (2017), doi: 10.1016/j.ejmech.2017.04.012.
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
O
O
O
O
Sub-G1 arrest, Ser10-histone H3
phosphorylation, PARP and
R1
R2
Cl
Caspase 3 cleavage induction in
A431 cells, growth arrest in A549,
MCF7, PC3, U373n, Hs683,
SKMEL28 cells (IC50 range: 0.1-1.2
µM)
X
O
N
O
H
R3
CDC25-A and -C inhibition
CDC25-C and its splice variant
protein down-regulation in
A431 cells

ACCEPTED MANUSCRIPT
Novel Coumarin- and Quinolinone-Based Polycycles as Cell Division
Cycle 25-A and -C Phosphatases Inhibitors Induce Proliferation
Arrest and Apoptosis in Cancer Cells
Clemens Zwergel,†,#,γ,¶ Brigitte Czepukojc,‡,#,¶ Emilie Evain-Bana,%,¶
Zhanjie Xu,γ,#, Giulia Stazi,† Mattia Mori,€ Alexandros Patsilinakos,†,¥,§
Ω
Antonello Mai,†,χ Bruno Botta,† Rino Ragno,†,¥,§ Denise Bagrel,% Gilbert
Kirsch,γ,* Peter Meiser,Ω Claus Jacob,# Mathias Montenarh,‡,* Sergio
Valente†,*
†Department of Drug Chemistry and Technologies, Sapienza University of Rome, P. le A. Moro 5,
00185 Rome, Italy
χIstituto Pasteur- Fondazione Cenci Bolognetti, Sapienza University of Rome, Via Regina Elena
291, 00161 Rome, Italy
#Division of Bioorganic Chemistry, School of Pharmacy, Saarland State University, Campus, Bld.
B2.1, 66123, Saarbrücken, Germany
%Université de Lorraine, UMR CNRS 7565, Structure et Réactivité des Systèmes Moléculaires
Complexes, Equipe 5 (MIC), Campus Bridoux, rue du Général Delestraint, 57070 Metz Cedex,
France.
‡Medical Biochemistry and Molecular Biology, Saarland University, 66421 Homburg/Saar,
Germany.
γUniversité de Lorraine, UMR CNRS 7565, Structure et Réactivité des Systèmes Moléculaires
Complexes, Equipe 3 (HECRIN), 1 Boulevard Arago, 57078 Metz Technopôle, France
€
Italian Institute of Technology, Center for Life Nano Science@Sapienza, Via Regina Elena 291,
00161 Rome, Italy
¥Rome Center for Molecular Design, Department of Drug Chemistry and Technologies, Sapienza
University of Rome, P. le A. Moro 5, 00185 Rome, Italy
§Alchemical Dynamics srl, P. le A. Moro 5, 00185 Rome, Italy
ΩUrsapharm Arzneimittel GmbH, Industriestraße 35, 66129 Saarbrücken Germany
1

ACCEPTED MANUSCRIPT
*Corresponding authors:
e-mail: gilbert.kirsch@univ-lorraine.fr (GK); Mathias.Montenarh@uniklinikum-saarland.de (MM);
Sergio.valente@uniroma1.it (SV)
¶These authors contributed equally to this work
2

ACCEPTED MANUSCRIPT
ABSTRACT
Cell division cycle phosphatases CDC25 A, B and C are involved in modulating cell cycle
processes and are found overexpressed in a large panel of cancer typology. Here, we describe the
development of two novel quinone-polycycle series of CDC25A and C inhibitors on the one hand
1a-k, coumarin-based, and on the other 2a-g, quinolinone-based, which inhibit either enzymes up to
a sub-micro molar level and at single-digit micro molar concentrations, respectively. When tested in
six different cancer cell lines, compound 2c displayed the highest efficacy to arrest cell viability,
showing in almost all cell lines sub-micro molar IC50 values, a profile even better than the reference
compound NCS95397. To investigate the putative binding mode of the inhibitors and to develop
quantitative structure-activity relationships, molecular docking and 3-D QSAR studies were also
carried out. Four selected inhibitors, 1a, 1d, 2a and 2c have been also tested in A431 cancer cells;
among them, compound 2c was the most potent one leading to cell proliferation arrest and
decreased CDC25C protein levels together with its splicing variant. Compound 2c displayed
increased phosphorylation levels of histone H3, induction of PARP and caspase 3 cleavage,
highlighting its contribution to cell death through pro-apoptotic effects.
3

ACCEPTED MANUSCRIPT
1. INTRODUCTION
Protein phosphorylation is an important post-translational modification for a number of cellular
processes such as cell growth, differentiation, migration, survival, and apoptosis.[1, 2] The balanced
and coordinated activity of protein kinases (PK) and protein phosphatases (PP) has thus been
regarded as crucial for cell functioning in a sort of Yin-Yang relationship.[2] The dysregulation of
PK-PP equilibrium, resulting in aberrant protein phosphorylation, can alter several cellular
pathways and cause several human diseases from cancer to diabetes to neurological or autoimmune
diseases.[3, 4] The protein tyrosine phosphatases (PTPs) constitute a large family of enzymes
including 107 members divided into 4 groups based on their protein sequences and functions.[2, 5]
CDC25 phosphatases, as class III PTP, possess dual specificity and are found in all eukaryotic
organisms except plants.[6] In mammalian cells three different CDC25 isoforms, CDC25A,
CDC25B and CDC25C have been identified, encoded by genes on three different chromosomes and
each with different possible splice variants.[7-10] The N-terminal regulatory regions are highly
divergent in sequence and are subject of alternative splicing generating two variants for CDC25A
and five each for CDC25B and CDC25C. Moreover, several phosphorylation and ubiquitination
sites have been identified at the N-termini, regulating phosphatase activity, protein levels and
protein-protein interactions. Finally, in this region are also located the nuclear import-export
sequences controlling the localization of the phosphatases. Conversely, the C-terminal region
containing the catalytic site is highly conserved among the different CDC25s. In this region is
located the HCX5R signature motif containing the catalytic cysteine in the active site of all the
PTPs.[10] In eukaryotes, CDC25s dephosphorylate the phospho-threonine 14 (p-Thr14) and
phospho-tyrosine 15 (p-Tyr15) residues of cyclin dependent kinases (p-CDK), thus activating them.
They play a key role in cell cycle transition, as well as in mitosis.[11] In particular, CDC25A
interacts with CDK2-cyclin E and CDK2-cyclin A complexes, which control G1/S transition and
4

ACCEPTED MANUSCRIPT
progression through S phase. This isoform is also involved in G2/M transition control interacting
with the CDK1-cyclin B.[12-14] The main target of CDC25B is the CDK1-cyclin B, located in
centrosomes implicated in prophase microtubule dynamics. CDC25B is also responsible for the
CDK2-cyclin A and CDK2-cyclin E activation.[15, 16] CDC25C is present in each cell cycle phase
and it also regulates G2/M transition through CDK1-cyclin B de-phosphorylation.[17, 18] The
apparent functional redundancy between CDC25A, -B and -C remains still an enigma. On the one
hand, Ferguson et al. reported that CDC25B and -C double knock-out mice developed normally,
suggesting that CDC25A can substitute the other two phosphatases.[19] On the other hand
CDC25B was essential for meiotic resumption in female mice, suggesting that the redundancy of
CDC25 isoforms’ are only limited to certain functions.[20]
Activity of CDC25s can also be modulated by miR-483-3p, able to arrest cell cycle progression,
and displaying its effect by direct targeting of the CDC25A phosphatase, thus preventing CDK4/6
activation.[21] A tight regulation of CDC25s activity is required to ensure the maintenance of a
correct level of CDK-cyclin function during normal cell division and in response of checkpoint
activation. This evidence points out how the dysregulation of CDC25s activity could cause genomic
instability, accumulation of genetic mutations and contribute to tumorigenesis. Accordingly,
CDC25s have been found to be overexpressed in numerous cancers including thyroid, laryngeal,
esophageal, hepatocellular, gastric, breast, ovarian, endometrial, prostate and colorectal cancer,
neuroblastoma and glioma, as well as in non-Hodgkin lymphomas.[11, 22] Moreover,
overexpression of CDC25s is often associated with their over-activation and correlation studies
demonstrated that it associates with aggressiveness, higher grade tumors and poor prognosis.[23,
24] An interesting example highlighting a specific role of CDC25s in a certain subtype of cancer is
given by the data recently reported by Sengupta et al.. They explored the role of CDC25s in breast
cancer progression demonstrating that in metastatic breast cancer TGF-β-Smad2 dependent
activation of CDC25A lead to NFAT activation, thus tumor progression. This pathway seems also
to be implicated in maintaining the balance between epithelial-mesenchymal transition (EMT) and
5

ACCEPTED MANUSCRIPT
cell proliferation in the late stages of the disease.[25] All these recent findings underlined the
importance of CDC25 as promising targets for cancer therapy.[11] Several strategies have been
proposed up to now to target CDC25, including the development of CDC25 inhibitors, that can be
divided in various groups such as para-quinonoids, phosphate surrogates and electrophilic entities.
Two CDC25s inhibitors are already under evaluation in clinical trials. Flavopiridol (Alvocidib,
Figure 1) has been studied in numerous phase I and II clinical trials on patients with various types
of relapsed or refractory tumors, either alone or in association with currently used chemotherapeutic
agents.[26] Roscovitine (Figure 1) is currently being investigated in different phase I and II clinical
trials on patients with solid tumors.[27, 28] Para-quinonoids class derives from vitamin K, reported
in 1993 as potent CDC25 inhibitors,[29, 30] and this class of molecules represent the most
numerous and potent one probably acting by irreversible oxidation of the cysteine residue in the
catalytic domain of CDC25s.[31]NSC95397 (Figure 1)[32] and NSC663284 (Figure 1)[33, 34] are
quinone-containing, potent and unspecific CDC25 inhibitors able to arrest the cell cycle in the G1
and G2/M phase and to inhibit the proliferation of human and murine carcinoma cells and human
breast cancer cells. Interestingly, three quinone and non-quinone based compounds (BN82685,
IRC083864 and BN82002, Figure 1) have been proven to be active on miaPaCa2, LNCap and HeLa
cell lines, but also in xenografted mice.[35-37] Furthermore, low doses of BN82685 in association
with paclitaxel arrested HT29 human colon cancer cells proliferation, suggesting an implication of
CDC25 in regulating microtubule dynamics and a possible combination therapy between CDC25
inhibitors and microtubule-targeting agents.[38] NSC119915 (Figure 1) was identified through a
structure-based high-throughput virtual screening. This compound, with a nanomolar Ki, was found
to inhibit human MCF-7, PC-3, and K562 cancer cell proliferation and significantly to impair the
cell cycle progression.[39] Among the non-quinone derivatives, in 2009 a class of
thiazolopyrimidines (Figure 1) was reported to possess CDC25B inhibiting properties and cytotoxic
activity against HeLa and LNCaP cells.[40, 41] In 2015 Lazo and Wipf described a new class of
non-quinoid reversible CDC25B inhibitors containing the 3-aminoisoquinolin-1(2H)-one
6

ACCEPTED MANUSCRIPT
pharmacophore. In this series, compound 13 (Figure 1) was reported as a competitive inhibitor (Ki =
1.9 µM).[42] CHEQ-2 (Figure1), a novel imidazopyridine-based derivative was recently described
as a CDC25A/B inhibitor able to induce S-phase cell cycle arrest in MCF-7, HepG2 and HT-29
cells, and to enhance mitochondrial dysfunction and apoptosis. Finally, CHEQ-2 (Figure 1)
inhibited xenografted human liver tumor growth in nude mice after oral administration (10 mg/kg),
while displaying a low toxicity (LD50 > 2000 mg/kg).[43] In 2015 Lund et al. opened a new
interesting way to target CDC25s, through disruption of the protein-protein interaction (PPI)
between the CDC25B phosphatase and the CDK2/cyclin A substrate. Among the new 2-fluoro-4-
hydroxybenzonitrile based derivatives as PPI disruptor, compound 7 (Figure 1) displayed the
highest potency.[44]
Figure 1. Known CDC25 phosphatases inhibitors.
7

ACCEPTED MANUSCRIPT
In the last years, various series of coumarin-based CDC25 inhibitors were reported by us
(compounds SV8, SV25, SV28, SV29 and SV37, Figure 2),[45, 46] endowed with IC50 values in
the low micro molar range. These compounds led to growth inhibition, apoptosis and impaired the
balance between Bcl2/Bax in HCT116 colon cancer cells.[47] Among them SV37 (Figure 2), able
to inhibit all three CDC25 isoforms through a reversible mode of action, arrested proliferation of
MDA-MB-231 breast cancer cells along with p-CDK accumulation, activation of apoptosis and
DNA cleavage.[46] In addition, in 2013 we disclosed, via a [4+2] thermal Diels–Alder
cycloaddition reaction, a novel quinone-based polycycle 1a endowed with CDC25s inhibition
activity. This compound displayed single digit micro molar range IC50 against CDC25A (5.8 µM)
and CDC25C (2.3 µM), while it was less potent against the CDC25B isoform (14.4 µM).[48] These
previous results prompted us to design and synthesize a novel series of 1a analogues (1b-k) bearing
different substituents at C8 and C9 position of the benzene moiety of the coumarin scaffold.
Furthermore, the coumarin oxygen atom was isosterically replaced by a nitrogen atom to design and
prepare derivatives 2a-g (Figure 2).
O
S
S
S
S
S
H3CO
O
O
H3CO
O
O
O
O
OCH3
O
O
OCH3
S
SV25
SV28
HO
O
O
SV8
S
S
S
O
O
S
O
OCH3
H3CO
O
O
O
SV29
H3CO
O
O
SV37
O
O
R1
R2
O
O
O
For 1b-k, X= O
For 2a-g, X= NH, NCH3
X
O
O
R3
1a
8

ACCEPTED MANUSCRIPT
Figure 2. Rational design of novel compounds 1b-k and 2a-g as CDC25A and CDC25C inhibitors.
All new synthesized compounds (1b-k and 2a-g) and the hit compound 1a were tested against
CDC25A and CDC25C isoforms (Table 1), and screened in six different cancer cell types (A549,
MCF7, PC3, U373n, Hs683, SKMEL28) in order to evaluate the capability to reduce cell viability
by MTT assays. From these assays the most effective compounds were selected for investigating
their anticancer properties together with their cell death mechanism, by evaluating cell viability,
CDC25C expression, phospho-CDK1 accumulation, PARP cleavage and caspases induction in
A431 cells, expressing high CDC25C protein levels.
2. RESULTS
2.1 Chemistry. The known 4-hydroxy-2H-chromen-2-ones 3a-h [48-52] were prepared from the
appropriate commercially available 2-hydroxy acetophenones following the method of Jones et
al.[53]
7-Diethylamino-4-hydroxy-2H-chromen-2-one 3i was synthesized according to a known literature
procedure.[50] The already described 4-hydroxyquinolin-2(1H)-ones 6a-g [54-58] were prepared
according to a known procedure.[59]
The bromination with tetra-n-butylammonium bromide and phosphoruspentoxide of the 4-hydroxy-
2H-chromen-2-ones (3a-i), and the 4-hydroxy-1-methylquinolin-2(1H)-one (6b) led to the 4-bromo-
2H-chromen-2-ones (4a-i), and 4-bromo-1-methylquinolin-2(1H)-one (8b) using the method
recently described by us [48] (Scheme 1).
The bis-bromo 4-bromoquinolin-2(1H)-ones 7a-f were refluxed in an acetic acid/water mixture
using a slightly modified method by Osborne et al., thus obtaining the desired 4-bromoquinolin-
2(1H)-ones 8a, 8c-g [56].
9

ACCEPTED MANUSCRIPT
Scheme 1a
O
O
OBu
Br
OH
O
R1
R2
R1
iii)
R1
R2
R1
R2
iv)
i)
O
O
R2
O
O
O
O
O
R3
R3
R3
R3
1a-g, 1j-k
5a-i
4a-i
3a-i
when 1g
R1= -H,
R2= -OMe.
R3= -H
v)
O
O
vi)
O
O
O
HO
O
O
O
O
1h
1i
O
O
OBu
O
Br
OH
Br
N
R1
R2
R1
R2
R1
R2
R1
R2
iv)
R1
iii)
ii)
for R=H
i)
for R=H
N
R
O
N
R
N
R
O
N
R
O
R2
Br
R3
R3
R3
R3
R3
2a-g
9a-g
7a-f
8a-g
6a-g
i)
for 6b, R= Me
aReagents and conditions: i) P2O5, TBAB, Toluene, 100 °C, 1 h; ii) CH3COOH, H2O, 100°C, 24 h,
iii) butyl vinyl ether, DIPEA, 1,1’-bis(diphenylphosphino)ferrocene (DPPF), Pd2dba3, dioxane,
80°C, 12 h, iv) 1,4-benzoquinone, toluene, 110°C, 8 h, v) BBr3, CH2Cl2, -78°C, 0.5 h then rt, 8 h;
vi) BnBr, NaH 60%, DMF, rt, 10 h.
Table 1. Substitution patterns of final compounds 1a-j, 1j-k and 2a-g
R1
R2
R3
R
R1
R2
R3
H
Cl
F
H
H
H
H
H
H
H
H
H
Cl
F
H
H
H
H
H
H
H
H
H
Me
H
H
H
H
H
Cl
F
H
H
H
H
H
H
H
Cl
F
H
H
H
H
H
H
1a
1b
1c
1d
1e
1f
1g
1j
1k
2a
2b
2c
2d
2e
2f
Me
OMe
N(Et)2
H
H
R2-R3 = Ph
2g
R2-R3 = Ph
The (E)-4-(2-butoxyvinyl)-2H-chromen-2-ones (5a-i), and (E)-4-(2-butoxyvinyl)-quinolin-2(1H)-
10

ACCEPTED MANUSCRIPT
ones (9a-g) were prepared through a very high β-regioselective Heck cross-coupling reaction,
previously described by our research group using 4-bromo-2H-chromen-2-ones (4a-i), and 4-
bromo-quinolin-2(1H)-ones (8a-g) as substrates, with Pd2dba3/DPPF as catalytic system[48]
(Scheme 1).
The 1H-naphtho[1,2-c]chromene-1,4,5-triones (1a-g, 1j-k), and benzo[i]phenanthridine-
1,4,5(6H)-triones (2a-g) were obtained by [4+2] thermal Diels–Alder cycloaddition reactions
between the dienes (E)-4-(2-butoxyvinyl)-2H-chromen-2-ones (5a-i), or (E)-4-(2-butoxyvinyl)-
quinolin-2(1H)-ones (9a-g) and the dienophile 1,4-benzoquinone (Scheme 1) as described
previously.[48]
Demethylation by BBr3 of compound 1g yielded compound 1h (8-hydroxy-1H-naphtho[1,2-
c]chromene-1,4,5-trione). 8-(benzyloxy)-1H-naphtho[1,2-c]chromene-1,4,5-trione 1i was prepared
by alkylation of compound 1h with benzyl bromide using an adapted method of Beauhaire et al.
(Scheme 1).[60]
2.2 Biochemical activities against CDC25A and CDC25C isoforms.
All new synthesized compounds were assayed against isolated human recombinant CDC25
phosphatases CDC25A and CDC25C isoforms in 10-dose IC50 mode with 3-fold serial dilution
starting from 100 µM concentrations (Table 2). From results in Table 2 the following structure-
activity relationship (SAR) can be highlighted: in the coumarin-based series (1a-k) all the
derivatives, including the reference one (1a), displayed a certain, even if not high, selectivity for
CDC25C over CDC25A, with the sole exception of compound 1i. The introduction of a chlorine
atom at the C8 position of polycycle (1d) increased 3.6-fold the inhibition activity against
CDC25A (IC50: 1.61 µM) and 3-fold against CDC25C (IC50: 0.77 µM), when compared to the hit
compound 1a. The same substitution with fluorine atom gave similar results for CDC25C,
although compound 1e was more potent than 1a, but slightly less potent (IC50: 2.19 µM) than the
chlorine analog 1d when evaluated against the CDC25A isoform. Shifting of both halogens from
C8 to C9 position of coumarin moiety resulted in weakly less potent derivatives (1b and 1c), but
11

ACCEPTED MANUSCRIPT
still more potent than the reference compound. Fluorine atom at C9 position (1c) improved the
potency against CDC25C (IC50: 0.4 µM) and the selectivity for this isozyme (ratio
IC50,CDC25A/IC50,CDC25C: 8.75). The addition of a methyl group at coumarin C8 position (1f)
displayed similar behavior to the analogues 1d and 1e, while different electron releasing groups
such as methoxy (1g), hydroxy (1h) and N,N-diethylamino (1j) maintained the inhibition potency
of 1f against both CDC25, decreased the potency for CDC25A (IC50: 6.6 µM), and increased the
inhibition efficiency for both CDC25 isoforms (IC50,CDC25A: 1.2 µM; IC50,CDC25C: 0.3 µM),
respectively. Alkylation of the hydroxy group (1h) with a benzyl substituent (1i) reduced the
inhibition for CDC25C, being 8.9-fold less potent (IC50,CDC25C: 7.12 µM). The addition of a further
benzene ring to 1a to give the pentacyclic derivative 1k led to an increase of both potency and
CDC25C selective inhibition, in particular the lowest IC50 (0.1 µM) and the highest selectivity
ratio was observed (IC50,CDC25A/IC50,CDC25C: 15), whereas the inhibition potency towards CDC25A
(IC50: 1.5 µM) was similar to chlorine-substituted (1d) or N,N-diethylamino (1j) derivatives.
Differently, in the quinolinone-based series (2a-g) the replacement of oxygen atom with its NH
isoster group in all the derivatives markedly reduced the inhibition potency and selectivity of the
derivatives against the CDC25C isoform, with the unique exception of the unsubstituted one (2a),
which was able to maintain 2-fold selectivity for that phosphatase. The nitrogen alkylation with a
methyl group (2b) furnished an almost 2-fold (CDC25A) and 4-fold (CDC25C) less potent inhibitor
than 2a, being the least potent compound among the two series. All the other analogues, chloro- and
fluoro-substituted (2c-2f) (at both C8 and C9 position), as well as the pentacyclic compound 2g
displayed an inhibition activity against CDC25A comparable (IC50 ranging from 2.85 µM (2f) to
4.33 µM (2c)) to the coumarin-based ones, whereas IC50 values towards CDC25C increased from
4.46 µM (2f) to 5.72 µM (2c). Similarly, as in the coumarin series (with the sole exception of 1c),
even in the compounds 2 series the most effective substitution in terms of potency against both
enzymes was in all cases at quinolinone C9 position.
12

ACCEPTED MANUSCRIPT
Table 2. Inhibition values against CDC25A and CDC25C (IC50, µM)a and viability data (IC50,
µM)b in a panel of cancer cell lines for compounds 1a-k and 2a-g.
Inhibition data
(IC50, µM)
Viability (MTT assay, IC50, µM)
Cpd
Structure
CDC25A
CDC25C
A549
26
MCF7
PC3
U373n
Hs683
SKMEL28
29
5.8 ± 0.2
2.57 ± 0.2
3.5 ± 0.2
1.61± 0.4
2.3 ± 0.1
16
32
18
20
1a
1.44 ± 0.2
0.4 ± 0.1
0.77 ± 0.1
25
17
13
3
19
43
12
36
6
33
40
1b
1c
1d
19
1.2
3.2
2.8
4.4
3.7
7.8
2.19 ± 0.2
2.9 ± 1.2
0.69 ± 0.02
1.0 ± 0.2
1
11
82
36
7
4
18
1e
1f
>100
>100
98
>100
>100
2.88 ± 0.1
6.6 ± 1.4
0.98 ± 0.2
0.8 ± 0.1
30
19
9
9
31
27
16
9
24
11
31
29
1g
1h
3.11 ± 0.4
1.2 ± 0.2
7.12 ± 0.3
0.3 ± 0.1
>100
>100
>100
>100
>100
>100
1i
1j
98
58
45
59
98
72
13

ACCEPTED MANUSCRIPT
1.5 ± 0.4
0.1 ± 0.01
70
7
75
55
4
>100
1k
7.73 ± 0.9
13.0 ± 2.2
3.64 ± 0.1
14.9 ± 2.2
1.8
68
0.9
34
0.9
32
2.7
41
2.8
67
1.5
68
2a
2b
4.33± 0.4
3.37 ± 0.1
4.04 ± 0.2
2.85 ± 0.3
5.72 ± 0.5
4.67 ± 0.5
5.5 ± 0.3
0.3
4
1.2
4
0.3
3
0.1
2
0.2
3
0.3
3
2c
2d
2e
2f
0.13
18
2
4
1
1
5
4.46 ± 0.4
27
4
3
9
8
3.23 ± 0.4
4.37 ± 0.9
45
30
42
42
30
72
2g
0.055 ±
0.006
0.078 ±
0.002
NCS
95397
1.45
0.96
3.4
1.85
2.8
2.3
aEach assay was performed in 10-dose IC50 series with 3-fold serial dilution (for concentrations used see
Material and Methods section, SD is given.
bConcentration of tested compound required for 50% inhibition of cell viability. Determined by MTT
assays over 72 h performed in sextuplicate; values are shown are the mean (SD < 10% of the mean).
2.3 Computational Studies
14

ACCEPTED MANUSCRIPT
To investigate the binding mode of the inhibitors and to develop quantitative structure-activity
relationships, molecular docking and 3-D QSAR studies were carried out.
2.3.1 Molecular Docking and 3-D QSAR. Molecular docking and 3-D QSAR studies were
performed to rationalized both structure-based (SB) and ligand-based (LB) relationships. Models
were derived by means of the 3-D QSAutogrid/R procedure python implementation. The models
clearly showed good correlation coefficients and their graphical inspection helped in rationalized
the structure-activity relationships. Furthermore, together with the docking studies the models will
be a valuable tool to design new compounds. Details are reported in Supporting Information.
2.4 Biology
2.4.1 Analysis of cell viability after treating different cancer cell lines with compounds 1a-k
and 2a-g.
In order to assess their capability to target cell viability, all the newly synthesized derivatives 1a-k
and 2a-g were tested in a panel of six typologically different cancer cell lines, such as human lung
adenocarcinoma A549, human breast adenocarcinoma MCF7, human prostate cancer PC3, human
glioma U373n and Hs683, human malignant melanoma SKMEL28. All inhibitors were used at
concentrations ranging from 0.01 µM to 100 µM and cells were treated for 72 h, then IC50 (µM)
was determined by means of the MTT colorimetric assay (Table 2).
The results revealed that in the coumarin-based compounds 1a-k, the hit compound 1a reduced cell
viability at IC50 values from 16 µM (MCF7) to 32 µM (PC3). When we shifted the halogens to the
C8 position, the compounds 1d and 1e were, in almost all cases, more effective than their
corresponding C9 analogues (1b, 1c), showing IC50 at a single digit µM concentration for all cell
lines tested, except for 1e. Compound 1d displayed a low micro molar activity in reducing cell
viability, having an IC50 ranging from 1.2 µM (Hs683) to 7.8 µM (SKMEL28), whereas the fluoro
analogue 1e displayed the lowest IC50 value in A549 cells (1 µM). The introduction of a methyl
group (1f) completely abolished the in-cell activity, this effect is probably due to cell permeability
issue. A similar behavior was shown for compound 1i bearing a benzyloxy group at the same
15

ACCEPTED MANUSCRIPT
position (C8). Among the derivatives having an electron releasing group, compounds 1g (C8-
methoxy) and 1h (C8-hydroxy) were as potent as the hit 1a, whereas compound 1j, in contrast with
its enzymatic IC50, was less efficient than 1a, giving IC50 value ranging from 45 µM to 98 µM,
probably due to poor cell permeability. The pentacyclic compound 1k was deprived of activity in
A549, PC3, U373n and SKMEL28 cell lines, whereas it displayed single digit µM values when
tested in MCF7 (IC50: 7 µM) and Hs683 (IC50: 4 µM).
When the second series of compounds (2a-g) was screened in the same cell lines, an interesting
profile emerged by these assays: compound 2a, without any substituent on the quinolinone moiety,
was able to reach sub-micro molar activities in the viability analysis in MCF7 (IC50: 0.9 µM) and
PC3 (IC50: 0.9 µM) cells. In the latter cell line it was more than 30-fold more potent than the
corresponding oxygen analogue 1a, while it displayed low single digit µM value against all the
other cell lines. However, compound 2a showed a less strong viability effect when compared to the
inhibitor NSC95397. Methylation at the nitrogen atom of 2a, in agreement with its enzymatic
activity, provided a compound (2b) with drastically reduced effect on viability, indeed the IC50
values raised up to 68 µM (45-fold less potent in SKMEL28 when compared to 2a). The halogen-
substituted derivatives 2c-f displayed the best IC50 profile in these assays, but also an opposite
activity when compared to their coumarin-based counterpart 1b-e. Indeed the C9 position was
preferred for in-cell activity, mainly for the chloro-substituted inhibitor 2c, that provided IC50
values ranging from 0.1 µM (U373n) to 1.2 µM (MCF7 cells), and it showed IC50 below 0.3 µM in
all the remaining cell lines. Although not well correlated with its enzyme inhibition activity, the
inhibitor 2c displayed the highest potency in all cell lines tested, even better (up to 18.5-fold in
U373n) than the reference NSC95397. The fluoro-analogue 2d was around 10-fold less potent than
corresponding chloro-substituted one (2c), while the polycycle derivative 2g displayed an activity
comparable to the N-methyl derivative 2b in all cell lines.
2.4.2 Effects of 1a, 1d, 2a and 2c in human epidermoid carcinoma A431 cells
16

ACCEPTED MANUSCRIPT
Based on the results above discussed, we selected the most efficient chloro-containing derivatives
from both coumarin- and quinolinone-based chemical series 1d and 2c (Figure 5), together with the
related unsubstituted hit compounds 1a and 2a (Figure 5) to be evaluated in the human epidermoid
carcinoma A431 cell line.
O
O
O
O
O
O
Cl
O
O
1a
1d
O
O
Cl
O
O
N
H
O
N
H
O
2a
2c
Figure 5. Selected compounds for evaluation in A431 cells
Since in preliminary experiments we have noticed that this cell type expressed considerable
amounts of CDC25C and its splice variant, we have investigated in this cancer cells not only their
influence on cell viability but also their effects on cell cycle arrest, phosphorylation level of CDK1
protein and histone H3, expression of CDC25C protein including a splice variant of this isoform,
PARP and caspase 3 cleavage (Figures 6-10).
To assess whether these compounds might have an influence on the viability of A431, cells were
treated with 1a, 1d, 2a, 2c for 6, 24 and 48 h and then cell viability was determined by MTT assay .
At 10 µM concentration all four compounds showed a reduction in cell viability, although to a
different value (Supporting Information, Figure S1). Compound 2c impaired cell viability up to
more than 60% already after 24 h. The reduction in cell viability was more pronounced after 48 h,
where only around 10% of the cells were viable. Compound 2a displayed a very similar behavior
although at 10 µM the reduction of viability was less remarkable than for compound 2c.
Compounds 1a and 1d worked efficiently only at the highest concentration used in these
experiments (50 µM). Thus, from these results it is evident that the quinolinone-containing
compounds were more potent than the coumarin-counterpart in this assay.
17

ACCEPTED MANUSCRIPT
The aforementioned experiments were replicated three times by using concentrations ranging from
1 to 50 µM and the IC50 values for the four compounds were determined. From data shown in
Figure 6 it turned out that 2c showed the lowest IC50 value with 3.13 µM after 24 h followed by 2a
with an IC50 value of 8.20 µM.
Figure 6. Dose-response curves for cell viability (by MTT assay) of A431 cells (A) and their
corresponding IC50 values (standard deviation are given) (B). Cells were treated with 1a, 1d, 2a or
2c at concentrations 1, 5, 10, 25 and 50 µM for 24 h. Compounds were tested in three independent
experiments.
For supporting these results by a different method, cells were treated with these four selected
compounds and counted after 24, 48 and 72 h. Data of three independent experiments are shown in
Figure 7. The results revealed that 2c was unique to exhibit a strong and time-dependent reduction
in cell growth at 10 µM concentration. These results confirmed the initial observation that 2c
efficiently inhibited cell viability and proliferation already at lower concentrations such as 10 µM.
18

ACCEPTED MANUSCRIPT
12
10
8
not treated
DMSO
1a 10µM
1a 50µM
2c 10µM
2c 50µM
1d 50µM
2a 50µM
6
treatment with compounds
4
2
0
0 h
24 h
48 h
72 h
Figure 7. A431 cell growth curves during 72 h of incubation time. At time point 24 h, the cells
were treated with 10 or 50 µM of 1a, 1d, 2a or 2c. Each point represents the number of cells against
time. The cell viability was assessed using the trypan blue (cell counting dye method). Data are
presented as means of three experiments ± SD. (All measure were performed by triplicate).
Moreover, the effects on cell cycle and/or apoptosis by compounds 1a, 1d, 2a and 2c was
investigated. Therefore, cells were treated with 10 or 50 µM of all four inhibitors and then analyzed
by cytofluorimetry. In Table 4 and Figure 8 the percentage of cells in the different cell cycle phases
and the related histogram of cytofluorimetric analysis are reported, respectively. 2c treated cells
showed a reduction in the percentage of cells in the G1-phase of the cell cycle compared to cells
treated with the DMSO alone control. Moreover, there was a dose-dependent decrease in the
percentage of cells with increasing concentrations of 2c. In addition, cells treated with either 2c or
1a showed a considerable increase in cell number in the subG1 phase, which might indicate
apoptosis. Hence, the reduction in cell proliferation seems to be due to a subG1 arrest of the treated
cells, which was particularly evident for 2c.
Table 4. Influence of 1a, 1d, 2a and 2c on cell cycle progression. Percentage of A431 in cell cycle
phases after treatment with 1a, 1d, 2a and 2c at 50 µM and for 24 h.
19

ACCEPTED MANUSCRIPT
Concentration
[µM]
subG1
(%)
compound
DMSO
G1 (%)
S (%)
G2/M (%)
0.2%
10
53.04
45.05
42.38
48.91
53.78
57.01
53.93
44.50
31.53
17.33
18.71
11.39
18.57
16.24
11.90
6.98
26.23
29.39
29.72
28.92
25.91
24.83
26.28
26.65
27.78
3.40
6.85
1a
1d
2a
2c
50
16.51
3.60
10
50
4.07
10
6.26
50
12.81
15.46
33.63
10
13.39
7.07
50
aA431 cells were treated for 24 h with DMSO (solvent control) or 10 and 50 µM of 1a,
1d, 2a and 2c and then analyzed by FACS
Figure 8. Histograms of cytofluorimetric analysis of cells treated with 1a, 2a and 2c at 50 µM.
20

ACCEPTED MANUSCRIPT
Next, as CDC25C was the enzyme isoform most responsive to the inhibitors, in particular by
coumarin-based series 1, and also prompted by a recent study showing that a compound analogue of
NSC119915 arrested melanoma cells in G2/M, causing a reduction of the protein levels of CDC25A
and CDC25C [61], we have also investigated the effects of selected inhibitors (used at 10 or 50 µM)
on the CDC25C protein expression in A431cells.
Cells treated with 1a, 1d, 2a and 2c were extracted and the cell extracts were analyzed on a SDS-
polyacrylamide gel followed by a Western blot with a CDC25C specific antibody. Figure 9A
highlights a reduction of CDC25C protein level when cells were treated with 50 µM for 1a, 2a and
2c, whereas no significant influence on the level of CDC25C expression was observed after 50 µM
treatment with 1d. With the same antibody we also detected one of the five different splice variants
of such protein, as previously described.[62] Also the amount of this particular splice variant
decreased after treatment with 1a, 2a and 2c (well detectable at 50 µM), whereas no reduction was
observed for 1d. In addition to the phosphatase inhibition, the treatment of A431 cells also led to
CDC25C protein downregulation (Figure 9A).
Since CDC25C dephosphorylates CDK1 at threonine 14 and tyrosine 15, the effect of the selected
compounds was evaluated on the phospho-CDK1 (p-CDK1) level by using phospho-specific
antibodies. After treatment of A431 cells with 10 and 50 µM of the 1a, 2a and 2c compounds, a
weak increase of p-CDK1 level was observed for compound 1a at 50 µM, whereas 2c showed a
strong reduction of p-CDCK1 at the same dosage (Figure 9B).
Next, as a marker for cells in early mitosis[63] we have analyzed the phosphorylation of histone 3
(p-H3) at serine 10 in treated cells, with a phosho-H3 specific antibody. Increasing levels of p-H3
compared to the DMSO control indicated a cell cycle arrest in early mitosis after treatment with the
1a, 2a or 2c. One representative of three Western blots performed is shown in Figure 9C,
highlighting a high level of p-H3 after 2c (10 µM) treatment of A431 cells, whereas no appreciable
21

ACCEPTED MANUSCRIPT
phosphorylation of H3 was observed at 50 µM for 2c. A less remarkable effect was shown by both
1a and 2a.
Figure 9. (A) Effects of 1a, 2a and 2c of the protein level of the phosphatase CDC25C and one of
its splice variants. A431 cells were treated with either DMSO or 10 µM and 50 µM of 1a, 1d, 2a
and 2c for 24 h, and protein expressions were studied by Western blotting, GAPDH was used as
loading control; (B) Western blotting showing the effect of 1a, 2a and 2c on p-CDK1 at 24 h
treatment, tubulin was used as loading control; (C) One representative of three Western blots
highlighting phospho H3 levels after treatment of A431 cells with these compounds for 24 h.
22

ACCEPTED MANUSCRIPT
According to the cytofluorimetry analysis, giving a first indication of apoptosis in 2c, 1a and 2a
treated A431 cells, their effect on poly (ADP-ribose) polymerase (PARP) cleavage was investigated
as a late event in the apoptosis pathway. Cells were treated with 1a, 2c or 2a at 10 or 50 µM then
after 24 h cells were extracted and cell extracts analyzed on a SDS polyacrylamide gel followed by
Western Blot with a PARP specific antibody. The Western blot (Figure 10A) shows only a very
faint band for the PARP cleavage product in cells treated with 50 µM of 1d. Conversely, a
detectable protein band for the PARP cleavage product was observed by using 10 µM 2c, further
supporting the idea that 2c is the most potent compound in killing A431 cells through apoptosis.
Moreover, to assess and to validate their pro-apoptotic mechanism, caspase 3 activation by 1a, 1d,
2a and 2c was analyzed. The Western Blot with caspase 3 specific antibodies, as shown in Figure
10B, demonstrated that treatment with 50 µM of 1a led to the appearance of caspase 3 cleavage
products, which were already observed for 2c at 10 µM. No cleavage products occurred in cells
treated with 1d and 2a at the same concentration. Hence, these data further reinforce our previous
results showing 2c as the most effective compound leading to a reduced cell viability and cell
proliferation, induced G1-arrest and finally to apoptosis, while both 1a and 2a were less effective
even at 50 µM.
23

ACCEPTED MANUSCRIPT
Figure 10. 1a, 2a and 2c compounds induce apoptotic cell death in A431 cells by activation of
PARP cleavage (A) and caspase 3 (B). A431 cells were treated with either DMSO or 10 µM and 50
µM of 1a, 1d, 2a and 2c for 24 h, and protein expressions were studied by Western blotting.
Proteins (50 µg for A and 100 µg for B) were separated on a 10% SDS-polyamide gel for A and on
a 15% SDS-polyamide gel for B, blotted on a PVDF membrane. PARP was visualized with the
appropriate antibody. Full-length caspase 3 and its cleavage products were detected with a caspase
3-specific antibody. GAPDH or ß-actin were used as loading control. One representative of at least
three Western blots is shown.
3. Discussion and Conclusions
Based on our recent experience in developing inhibitors of CDC25 phosphatases[45-48],
particularly from the disclosure of the novel coumarin-based quinone polycycle 1a endowed with
major inhibition potency against CDC25A and CDC25C[48], we describe here the design and
24

ACCEPTED MANUSCRIPT
synthesis of a novel series of 1a analogues (1b-k) bearing different substituents at C8 and C9
position of the coumarin scaffold, and a second series wherein the coumarin moiety was replaced by
a quinolinone, thus preparing the derivatives 2a-g.
The synthesized compounds were analyzed for the effect on CDC25A and CDC25C isoforms to
determine IC50 values and the results revealed that in the coumarin-based series (1a-k) all the
derivatives, displayed some degree of selectivity for CDC25C over CDC25A, with the sole
exception of compound 1i. The substitution with a chlorine atom at the C8 position of the polycycle
(1d) generated the most powerful inhibitor with an 3.6-fold increase in the inhibition of the
phosphatase activity of CDC25A (IC50: 1.61 µM) and 3-fold inhibition of CDC25C (IC50: 0.77
µM), when compared to the lead compound 1a. Interestingly, by adding a further benzene ring to
compound 1a we obtained a pentacyclic derivative (1k) with an improved inhibition potency for the
phosphatase activity of CDC25C, achieving the best IC50 value (0.1 µM) and the highest selectivity
(ratio IC50,CDC25A/IC50,CDC25C: 15) towards this isoform.
In the quinolinone-based series of compounds (2a-g) the replacement of oxygen atom with its NH
isoster group reduced the inhibition potency and selectively against the CDC25C isoform, with the
unique exception of the un-substituted 2a, still maintaining 2-fold selectivity for that phosphatase.
The nitrogen alkylation with a methyl group (2b) furnished an almost 2-fold (CDC25A) and 4-fold
(CDC25C) less potent inhibitor than 2a. All the other analogues 2c-g displayed an inhibition
activity against CDC25A comparable (IC50 ranging from 2.85 µM (2f) to 4.33 µM (2c)) to the
coumarin-based ones, nevertheless the IC50 values towards CDC25C raised from 4.46 µM (2f) to
5.72 µM (2c).
Computational studies, molecular docking and 3-D QSAR, gave some hints on the putative mode of
action of titled compounds. In particular, 3-D QSAR models derived with structure-based aligned
conformation represent a first model to aid the design of further derivatives potentially endowed
with higher inhibitory activities.
25

ACCEPTED MANUSCRIPT
Next, the new derivatives 1a-k and 2a-g were tested in a panel of six different cancer cell lines
(A549, MCF7, PC3, U373n, Hs683, SKMEL28) in order to evaluate their influence on cell viability
by IC50 values.
In the coumarin-based compounds 1a-k the halogen addition to the coumarin moiety at the C8
position gave the best inhibition of cell viability. The best reduction in cell viability was observed
for chloro-substituted compound 1d (IC50 ranging from 1.2 µM (Hs683) to 7.8 µM (SKMEL28)),
which was more potent than the reference compound 1a and than its corresponding C9 analogue 1b,
in this case the effect on cell viability perfectly correlated with the IC50 values against both CDC25
A and C. When the second series of compound (2a-g) was assayed in the same cell lines, compound
2a already displayed an activity in MCF7 in sub-micro molar concentrations (IC50: 0.9 µM) and
PC3 (IC50: 0.9 µM). In the latter cell line 2a was 30-fold more efficient than the corresponding
oxygenated analogue 1a, however displaying low single digit µM values against all other cell lines.
Noteworthy, compound 2a showed a reduced viability profile comparable to the reference inhibitor
NSC95397. The halogen-substituted derivatives 2c-f displayed the best profile in these assays, but
differently from their coumarin-based counterparts 1b-e, the C9 position was the preferred one for
inhibition of cell viability. This result was perfectly exemplified by the chloro-substituted inhibitor
2c, giving IC50 values ranging from 0.1 µM in U373n to 1.2 µM in MCF7 cells. Although not well
correlating with its enzyme inhibition potency, the inhibitor 2c displayed the best reduction in cell
viability in all cell lines tested, even better (up to 18.5-fold more potent in U373n) than the
reference NCS95397.
An explanation for the observed not-correlation between CDC25 inhibition potency and cell
viability effect provided by some compounds such as 2c could come out from additional off-target
mechanism. Due to their planar structure, it could be speculated that these molecules could also
display (in some case) a) DNA intercalating activity or b) the quinone moiety could be involved in
26

ACCEPTED MANUSCRIPT
the production of reactive oxygen species (ROS), thus inducing intrinsic apoptotic pathway or c)
protein kinase inhibition. These points could be further investigated and clarified in a future work.
Based on these results, compounds 1a, 1d, 2a and 2c were selected for additional investigations of
cell cycle arrest, phosphorylation level of the CDK1 protein and histone H3, CDC25C protein
expression including a splice variant of this isoform, PARP cleavage and caspase 3 cleavage. We
decided to perform such experiments in A431 cell line, which express well detectable amounts of
CDC25C and one of its splice variants. Quinolinone-based 2c displayed good reduction of cell
viability and proliferation providing the lowest in A431 IC50 value with 3.13 µM after 24 h of
treatment. Again, 2c showed the highest dose-dependent reduction of cells in the G1-phase.
Moreover, 2c induced cell arrest in the subG1 phase highlighting that the reduction in cell
proliferation was, at least in part, due to apoptosis. Compounds 1a, 2a and 2c, but not 1d reduced
the expression of CDC25C and also its splice variant. Thus, in addition these compounds not only
reduced the phosphatase activity of CDC25C in vitro, but they also led to the reduction in the
expression level of CDC25C in A431 cells. Then the effect of the selected compounds was also
investigated on the phospho-CDK1 level by Western blot and, even if not expected, no remarkable
alteration in the level of p-CDK1 for all selected inhibitors was observed. Conversely, when the
phosphorylation level of histone 3 was investigated, an increasing level of p-H3 was detected,
already at 10 µM, again for compound 2c, thus showing a cell cycle arrest in early mitosis.
Next, to confirm that apoptosis was occurring in A431, the effect of the four inhibitors on PARP
cleavage was studied, and the results highlighted a detectable protein band for the PARP cleavage
product by using 10 µM of 2c, confirming that 2c was the most active compound in A431 cells to
induce apoptosis. To further assess a pro-apoptotic mechanism, caspase 3 activation by 1a, 1d, 2a
and 2c was detected. A Western blot analysis confirmed that treatment with 50 µM 1a or 2a
provided caspase 3 cleavage products, and most remarkable this results was already observed at 10
µM of 2c.
27

ACCEPTED MANUSCRIPT
These data further support previous results obtained in the first panel of cancer cell lines, showing
the 2c was the most efficient antiproliferative compound, hence this inhibitor is an excellent
candidate for a next structural optimization aiming to increase its enzymatic and cellular potency.
4. Experimental Section.
4.1 Chemistry
All solvents were of reagent grade and were purified and dried by standard methods when
necessary. All chemicals were purchased from Acros (Geel, Belgium), Alfa Aesar (Karlsruhe,
Germany) or Sigma Aldrich (Saint-Quentin-Fallavier, France). All reactions were routinely checked
by TLC analysis on an Alugram SIL G/UV254 (Macherey-Nagel, Düren, Germany) with spots
visualized by UV light. The concentration of solutions after reactions and extractions involved the
use of a rotary evaporator operating at reduced pressure. Organic solutions were dried over
anhydrous magnesium sulfate, and column chromatography was performed using silica gel 60 (50-
200 nm diameter, Merck, Darmstadt, Germany). Melting points were determined in open capillaries
on a Stuart SMP3 melting point apparatus and are uncorrected. The 1H and 13C NMR spectra were
measured on an AC Bruker 400 MHz spectrometer in CDCl3 or DMSO-d6; chemical shifts are
reported in parts per million (ppm). All coupling constants (J) are given in Hz. MS spectra were
recorded on an Agilent Technologies GC-MS instrument equipped with a 7683 injector, 6890 N gas
chromatograph and a 5973 mass selective detector. Elemental analysis has been performed on
Thermo Fisher FlashSmart CNHS/O to determine purity of the described compounds, that is ≥95%
(specific data available in supporting info).
The already described 4-hydroxy-2H-chromen-2-ones 3a-3h[48-52] were prepared according to a
known literature procedure from the appropriate commercially available 2-hydroxy
acetophenones.[53]
7-diethylamino-4-hydroxy-2H-chromen-2-one 3i was prepared according to the literature.[50]
28