Functionalization of 9-thioxanthone at the 1-position: From arylamino derivatives to [1]benzo(thio)pyrano[4,3,2-de]benzothieno[2,3-b]quinolines of biological interest

Article abstract of DOI:10.1016/j.bioorg.2019.103347

Original 1-amino substituted thioxanthone derivatives were easily prepared from the bare heterocycle by a deprotometalation-iodolysis-copper-catalyzed C–N bond formation sequence. This last reaction delivered mono- or/and diarylated products depending on the aniline involved. 1-Amino-9-thioxanthone was also prepared and reacted with 2-iodoheterocycles. Interestingly, while 1-(arylamino)-9-thioxanthones could be isolated, their subsequent cyclization was found to deliver original hexacyclic derivatives of helicoidal nature. Evaluation of their photophysical properties revealed high fluorescence in polar media, indicating potential applications for biological imaging. These compounds being able to inhibit PIM1 kinase, their putative binding mode was examined through molecular modeling experiments. Altogether, these results tend to suggest the discovery of a new family of fluorescent PIM inhibitors and pave the way for their future rational optimization.

Full text of DOI:10.1016/j.bioorg.2019.103347

Journal Pre-proofs
Functionalization of 9-thioxanthone at the 1-position: from arylamino deriva-
tives to [1]benzo(thio)pyrano[4,3,2-de]benzothieno[2,3-b]quinolines of bio-
logical interest
Nahida Mokhtari Brikci-Nigassa, Lionel Nauton, Pascale Moreau, Olivier
Mongin, Raphaël Duval, Laurent Picot, Valérie Thiéry, Mohamed Souab,
Sandrine Ruchaud, Stéphane Bach, Rémy Le Guevel, Ghenia Bentabed-
Ababsa, William Erb, Thierry Roisnel, Vincent Dorcet, Florence Mongin
PII:
S0045-2068(19)31210-6
DOI:
https://doi.org/10.1016/j.bioorg.2019.103347
YBIOO 103347
Reference:
To appear in:
Bioorganic Chemistry
Received Date:
Revised Date:
Accepted Date:
27 July 2019
3 October 2019
4 October 2019
Please cite this article as: N. Mokhtari Brikci-Nigassa, L. Nauton, P. Moreau, O. Mongin, R. Duval, L. Picot, V.
Thiéry, M. Souab, S. Ruchaud, S. Bach, R. Le Guevel, G. Bentabed-Ababsa, W. Erb, T. Roisnel, V. Dorcet, F.
Mongin, Functionalization of 9-thioxanthone at the 1-position: from arylamino derivatives to
[
1]benzo(thio)pyrano[4,3,2-de]benzothieno[2,3-b]quinolines of biological interest, Bioorganic Chemistry (2019),
doi: https://doi.org/10.1016/j.bioorg.2019.103347
This is a PDF file of an article that has undergone enhancements after acceptance, such as the addition of a cover
page and metadata, and formatting for readability, but it is not yet the definitive version of record. This version
will undergo additional copyediting, typesetting and review before it is published in its final form, but we are
providing this version to give early visibility of the article. 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.
©
2019 Published by Elsevier Inc.

Graphical Abstract
S
S
O
NH2
I
N
cat. Cu
K2CO3
Bu2O, reflux
S
S
(38% yield)
_
______________________________________________
Functionalization of 9-thioxanthone at the 1-position: from arylamino derivatives to
1]benzo(thio)pyrano[4,3,2-de]benzothieno[2,3-b]quinolines of biological interest
[
a,b
c
c,
a
Nahida Mokhtari Brikci-Nigassa, Lionel Nauton, Pascale Moreau, * Olivier Mongin,
d
e
e
f
f,g
Raphaël Duval, Laurent Picot, Valérie Thiéry, Mohamed Souab, Sandrine Ruchaud,
f,g,
h
b,
a
Stéphane Bach, * Rémy Le Guevel, Ghenia Bentabed-Ababsa, * William Erb,
a
a
a,
Thierry Roisnel, Vincent Dorcet, Florence Mongin *
a
Univ Rennes, CNRS, ISCR (Institut des Sciences Chimiques de Rennes) - UMR 6226,
F-35000 Rennes, France
b
Laboratoire de Synthèse Organique Appliquée, Faculté des Sciences Exactes et Appliquées,
Université Oran1 Ahmed Ben Bella, BP 1524 El M’Naouer, 31000 Oran, Algeria
c
Université Clermont Auvergne, CNRS, SIGMA Clermont, ICCF, F-63000 Clermont-Ferrand, France
d
Université de Lorraine, CNRS, L2CM, F-54506 Nancy, France
e
Laboratoire Littoral Environnement et Sociétés, UMRi CNRS 7266,
La Rochelle Université, 17042 La Rochelle, France
f
Sorbonne Université, CNRS, FR2424, Plateforme de criblage KISSf (Kinase Inhibitor Specialized
Screening facility), Station Biologique de Roscoff, Place Georges Teissier, 29682 Roscoff, France
g
Sorbonne Université, CNRS, UMR 8227, Integrative Biology of Marine Models,
Station Biologique de Roscoff, CS 90074, 29688 Roscoff Cedex, France
h
Univ Rennes, Plateforme ImPACcell, SFR BIOSIT - UMS CNRS 3480, UMS INSERM 018,
F-35000 Rennes, France
*
Corresponding authors.
E-mail addresses: pascale.moreau@uca.fr (P. Moreau), bach@sb-roscoff.fr (S. Bach),
badri_sofi@yahoo.fr (G. Bentabed-Ababsa), florence.mongin@univ-rennes1.fr (F. Mongin).
Keywords: thioxanthone; N-arylation; copper; kinase inhibition; melanoma cells; fluorescence
1

Abstract:
Original 1-amino substituted thioxanthone derivatives were easily prepared from the bare heterocycle
by a deprotometalation-iodolysis-copper-catalyzed C-N bond formation sequence. This last reaction
delivered mono- or/and diarylated products depending on the aniline involved. 1-Amino-9-
thioxanthone was also prepared and reacted with 2-iodoheterocycles. Interestingly, while 1-
(arylamino)-9-thioxanthones could be isolated, their subsequent cyclization was found to deliver
original hexacyclic derivatives of helicoidal nature. Evaluation of their photophysical properties
revealed high fluorescence in polar media, indicating potential applications for biological imaging.
These compounds being able to inhibit PIM1 kinase, their putative binding mode was examined
through molecular modeling experiments. Altogether, these results tend to suggest the discovery of a
new family of fluorescent PIM inhibitors and pave the way for their future rational optimization.
1
.
Introduction
Thioxanthones are part of the aromatic heterocycles found in bioactive compounds (Figure 1).
Lucanthone is for example a DNA intercalating derivative that inhibits the synthesis of macromolecules
by interfering with the activity of topoisomerase I and II during replication and transcription. It
specifically inhibits the DNA repair enzyme apurinic/apyrimidinic endodeoxyribonuclease 1 in a way
that results in unrepaired DNA strand breaks, possibly inducing apoptosis and reducing tumor cell
resistance to radio- and chemotherapy. It also disrupts lysosomal function, inhibiting autophagy. Both
lucanthone and its metabolite hycanthone show antischistosomal and antineoplastic activities. The
analogues SR233377 and SR271425 are also cytotoxic DNA-interacting agents that exhibit a broad
antitumor activity. However, the side effects of all these compounds ˗ mutagenicity (e.g. in the case of
lucanthone and hycanthone, due to the methylene moiety linked to their 4-position) or cardiotoxicity
(e.g. for SR271425) ˗ made clinical use impossible [1]. Investigations are still ongoing in order to
identify more selective antitumor agents [2, 3].
2

Et N
2
Et N
2
Et N
2
Et N
2
O
S
HN
O
HN
O
S
HN
O
HN
H CO
3
S
S
CH3
CH OH
CH NHSO CH
CH NHCHO
2
2
2
2
3
lucanthone
hycanthone
SR233377
SR271425
Figure 1. Examples of thioxanthone derivatives possessing biological properties.
In the course of a previous work, we identified the easily accessible thioxanthone derivative HM107-
g as a potent antibacterial and antifungal agent without significant toxicity on human red blood cells
(Figure 2). In addition, upon evaluation against a short panel of serine/threonine protein kinases, this
compound inhibited human proto-oncogene PIM1 with an IC value of 610 nM [4]. In continuation
5
0
with this preliminary work, we report here the 2- or 3-step syntheses of 9-thioxanthone derivatives, as
well as their photophysical and biological evaluation.
NH2
NH2
NH
N
N
O
I
O
S
8
1
4
H N
2
S
S
HM107-g
1
9-thioxanthone
Figure 2. Retrosynthetic pathway of the PIM1 kinase inhibitor HM107-g [4].
2
.
Results and Discussion
2
.1. Synthesis
The traditional approach to access 1-substituted 9-thioxanthones relies on intramolecular
electrophilic cyclizations from suitably substituted benzoic acids or related compounds. Although able
to deliver various derivatives, it generally requires harsh reaction conditions [1]. A few direct methods
to functionalize the 1-position of 9-thioxanthone are documented, such as some ruthenium-catalyzed
activations [5, 6], the rhodium-mediated reaction with maleimides [7] and the iridium-catalyzed
oxidative heteroarylation with thiophene [8].
3

Both in order to avoid the use of an expensive transition metal and to reduce the competitive 1,8-
difunctionalization sometimes noticed in the previous methodologies, we decided to employ cheaper
reagents, and selected deprotometalation-trapping to access 1-halogeno 9-thioxanthones. Because 9-
thioxanthone is a substrate prone to nucleophilic attack, it was deprotonated by using hindered lithium
amide LiTMP (TMP = 2,2,6,6-tetramethylpiperidino) and ZnCl ·TMEDA (TMEDA = N,N,N’,N’-
2
tetramethylethylenediamine) as in situ trap [9], as reported previously [4, 10]. In THF (THF =
tetrahydrofuran) at -30 °C, the deprotolithiation-transmetalation took place efficiently, as demonstrated
by subsequent iodolysis to give the expected iodide in 84% yield (Scheme 1). Although the competitive
formation of the diiodide 1’ could not be totally avoided (10% isolated yield), this side product was
easily discarded by column chromatography over silica gel. It is worth noting that the corresponding
bromide and chloride should be similarly synthesized [11]; however, the iodide appeared to be more
reactive in the forthcoming copper-catalyzed C-N bond formation reactions.
O
S
1) ZnCl ·TMEDA (1 equiv)
2
O
S
I
8
1
2) LiTMP (1.5 equiv)
THF, -30 °C, 15 min
2) I2, THF
: 84% yielda
1
Scheme 1. Synthesis of 1-iodo-9-thioxanthone (1) and ORTEP diagram (50% probability) of the
a
compound 1’. 1,8-Diiodo-9-thioxanthone (1’) was also isolated in 10% yield.
In order to benefit from the presence of both ketone function and adjacent iodine, we first considered
reacting 1 with 1,2-phenylenediamine in order to access the benzodiazepine derivative 2a (Figure 3).
According to our previous work [12], amidation was performed from the iodo derivative 1 with
catalytic copper(I) iodide and potassium carbonate in dimethylsulfoxide. However, only traces of the
expected product 2a were detected (identified only by X-ray diffraction; see Figure 3). The major
product isolated was unambiguously identified as 1-(2-aminophenylamino)-9-thioxanthone (2a’),
indicating a sluggish cyclization under these conditions. Using 2,3-diaminopyridine similarly led to 2b’
4

(Scheme 2; Figure 3). The yields are moderate, due to competitive deiodination of 1 under the coupling
reaction conditions.
H N
2
O
I
O
S
HN
X
H N
2
CuI (0.2 equiv)
K CO (1 equiv)
2
3
+
H N
2
X
DMSO
S
(2 equiv)
120 °C, overnight
1
2a' (X = CH): 58% yield
b' (X = N): 35% yield
2
Scheme 2. Unsuccessful attempts to reach benzodiazepine-containing derivatives.
N
S
NH
2
a
2
a’
2b’
Figure 3. ORTEP diagrams (50% probability) of the compounds 2a, 2a’ and 2b’.
These results show that the formation of benzodiazepine-containing derivatives from 1 might be
possible, but would require a comprehensive study. Therefore, we opted for the evaluation of simpler
anilines. After a short optimization of the amounts of aniline, base and copper(I) iodide, and the
reaction temperature and time, the conditions shown in Table 1 were applied to the N-arylation of 4-
(trifluoromethylsulfonyl)-, 4-fluoro- and 4-methylaniline by 1. The corresponding coupled products 3a-
5

c were isolated in moderate yields due to the difficulty to separate them from deiodinated 9-
thioxanthone or/and the bis-N-arylated product (entries 1-3; Figure 4). In the case of 4-methylaniline,
this bis-N-arylated product 3c’ could be isolated and identified unambiguously (entry 3).
Table 1. N-arylation of different anilines using 1-iodo-9-thioxanthone (1).
S
Ar
O
I
O
HN
CuI (0.15 equiv)
K CO (2 equiv)
O
O
N
Ar
2
3
Ar
+
H N
+
2
DMSO
20 °C, 24 h
S
S
(1 equiv)
1
1
3
S
3'
3, Yield (%)a
3’, Yield (%)a
Entry
1
H N-Ar
2
-b
3a, 58
3b, 31
3c, 50c
3d, 63
H2N
H2N
H2N
SO2CF3
F
2
3
4
-b
3c’, 6
-d
H N
2
5e
3e, <10c
3e’, 41
H2N
OMe
a
b
After purification (see experimental part). Not isolated (yield not estimated).
c
d
e
Estimated yield due to purification issue. Not found. 0.5 equiv of 4-anisidine was used.
We noticed that the use of 1-naphthylamine did not lead to the bis-N-arylated compound (entry 4;
Figure 4). This could be rationalized by a more important steric hindrance since it became the major
product by reducing the amount of 4-methoxyaniline to 0.5 equivalent (entry 5). These results also
suggest that using this ligandless catalytic system might favor bis-arylation. Indeed, competitive
formation of triarylamines in the course of copper-catalyzed N-arylation of anilines has already been
observed, but when substituted by electron-withdrawing groups [13, 14]; in the present case,
triarylamines were even formed from electron-rich anilines. The bis-coupled products 3c’ and 3e’
6

showed original crystal structures, with both heterocycles being quasi-stacked and arranged head-to-
foot (Figure 4).
Inspired by the N-arylation of 1-amino-9-xanthone with 3-iodoanisole reported by Fujiwara and
Kitagawa in 2000 [15], we attempted to use a mixture of activated copper and copper(I) iodide, and
replaced dimethylsulfoxide by dimethylformamide. Under these conditions, 4-aminopyridine was
reacted with 1 to afford the product 3f, but still in a moderate yield (Scheme 3).
3
a
3c
3
d
3e
3
c’
3e’
Figure 4. ORTEP diagrams (50% probability) of the compounds 3a, 3c, 3d, 3e, 3c’ and 3e’.
7

N
O
I
Cu (0.15 equiv)
CuI (70 mequiv)
K CO (2 equiv)
O
S
HN
N
2
3
+
H N
2
DMF
S
(1 equiv)
140 °C, 6 h
1
3f: 49% yield
Scheme 3. N-arylation of 4-aminopyridine by using 1-iodo-9-thioxanthone (1).
We recently showed that, after N-arylation with 2-iodobenzofuran and 2-iodobenzothiophene, 2-
aminophenones were instantly converted into the corresponding tetracycles, as depicted in Figure 5
[
16]. As these derivatives showed promising antiproliferative activity in melanoma cells, we decided to
prepare 1-amino-9-thioxanthone and attempt the corresponding reactions.
X
O
NH2
X
O
HN
X
N
I
R
R
R
X = O, S
tetracycles
Figure 5. Synthetic pathway of tetracycles of biological interest [16].
A direct access to 1-amino-9-thioxanthone (4) was considered by amination of 1 according to various
conditions reported in the literature for the synthesis of anilines from aryl iodides. First, inspired by
Kim and Chang’s protocol [17], we employed excess aqueous ammonia in the presence of copper(I)
iodide (0.2 equiv), L-proline (0.4 equiv) and potassium carbonate (3 equiv) in dimethylsulfoxide; after
5
days at 80 °C, the expected amino derivative 4 was isolated in 44% yield (the rest was starting 1). We
next adapted a protocol from the work of Ji, Atherton and Page [18], and treated 1 by copper(I) chloride
0.1 equiv), sodium ascorbate (0.1 equiv) and excess ammonia in methanol; after 14 hours at 100 °C,
(
the product 4 was isolated in 20% yield (the rest was 9-thioxanthone, formed by deiodination of 1).
Finally, we applied the protocol using copper(I) iodide (0.1 equiv), DMEDA (DMEDA = N,N’-
dimethylethylenediamine; 0.15 equiv), excess aqueous ammonia and dimethylsulfoxide at 130 °C
(sealed tube) reported by Jeon and co-workers [19]; after 16 hours at this temperature, the expected
8

amine 4 was isolated in 50% yield (Scheme 4). Under these conditions, although no degradation of 1
took place, the competitive formation of a fluorescent unidentified compound was noticed (possessing a
1
singlet at ~9.1 ppm in its H NMR spectra). Replacing DMEDA by PEG300 also gave 4, but in a lower
3
3% yield.
O
I
CuI (0.10 equiv)
DMEDA (0.15 equiv)
O
S
NH2
2
5% NH OH (excess)
4
DMSO
S
130 °C, 16 h
1
4: 50% yield
Scheme 4. Conversion of 1-iodo-9-thioxanthone (1) into 1-amino-9-thioxanthone (4).
We used the amine 4 with various substrates such as 2-iodothiophene, 2-iodobenzothiophene and 2-
iodobenzofuran. In order to favor the formation of polycyclic molecules, we used conditions similar to
those that afforded the tetracycles shown in Figure 5. Thus, the amine 4 was treated by the iodide in the
presence of activated copper (0.2 equiv) and potassium carbonate (2 equiv) at the reflux temperature of
dibutyl ether (Table 2).
Table 2. N-arylation of 1-amino-9-thioxanthone (4) by using heteroaryl iodides.
X
activated Cu
N
O
NH2
(0.2 equiv)
K CO
O
HN
X
2
3
(2 equiv)
+
+
X
Bu O
reflux, 24 h
S
S
I
2
S
4
(1.5 equiv)
5
6
5, Yield (%)a
5a, 40
6, Yield (%)a
Entry
1
Heteroaryl iodide
-b
S
I
2
3
5b, 50
6b, 38
6c, 21
S
I
-b
O
I
a
b
After purification (see experimental part). Not searched.
9

After 24 hours, although hardly separable mixtures were obtained, careful chromatography afforded
compounds 5 or/and 6. In the case of 2-iodothiophene, the product 5a resulting from a mono-N-
arylation of 4 was isolated in 40% yield (entry 1). With 2-iodobenzothiophene, both 5b and the cyclized
product 6b could be isolated (entry 2, respective yields of 50 and 38%; Figure 6). From 2-
iodobenzofuran, the hexacycle 6c was the only isolated product, obtained in a moderate 21% yield
(entry 3; Figure 6). Unfortunately, our attempts to convert the N-arylated products 5 into the fluorescent
products 6 by acid-mediated cyclization [20] only led to degradation. It was however interesting to
study the luminescence properties of the compounds 6b and 6c.
5
b
6b
6c
Figure 6. ORTEP diagrams (50% probability) of the compounds 5b, 6b and 6c.
2
.2. Physicochemical properties
During the purification of the hexacyclic compounds 6b and 6c, we noticed their strong fluorescent
behavior. Therefore, in view of potential applications as fluorescent probes in biological media, we
initiated the early evaluation of their photophysical properties. Their absorption and emission properties
were first investigated in toluene (Figure 7), and the results are gathered in Table 3.
Both compounds absorb in the blue-violet part of the visible region and emit in the green one. The
replacement of the oxygen atom of 6c with a sulfur atom (6b) leads to 12 nm and 20 nm bathochromic
shift in absorption and emission, respectively. Their luminescence quantum yields are quite high and
almost identical. Their luminescence lifetimes are in the range of 15 nanoseconds, which confirms that
1
0

this emission is indeed a fluorescence phenomenon. The radiative decay rate of 6c is slightly higher
than that of 6b (in agreement with a concomitant increase of the molar extinction coefficient), but its
nonradiative decay rate decreases almost similarly, which leads to the maintaining of the fluorescence
quantum yield.
Figure 7. Absorption (solid line) and emission (dotted line) of compounds 6b and 6c in toluene.
Table 3. Absorption and emission properties of 6b and 6c in toluene at 25 °C.
Compound abs a (nm) max b (M cm ) 
-1
-1
c
F d
0.50
0.49
e
f
-1
f
-1
(nm)
 (ns)
k
r
(s )
knr (s )
em
7
7
7
7
6
6
b
c
449
437
6800
8300
532
512
16.5
13.6
3.03 × 10
3.60 × 10
3.03 × 10
3.75 × 10
a
Absorption maximum. b Molar extinction coefficient at  . Emission maximum. Fluorescence quantum
c
d
abs
e
f
yield using as a standard quinine bisulfate in 0.5 M H SO , upon excitation at  . Fluorescence lifetime.
Radiative (k ) and nonradiative (k ) decay rates derived from fluorescence quantum yield and lifetime values
2
4
abs
r
nr
(
k =  /; k = (1- )/).
r
F
nr
F
The influence of the solvent polarity on the absorption and emission properties of the compounds 6b
Figure 8 and Table 4) and 6c (Figure 9 and Table 5) was then studied. Almost no solvatochromism
(
was observed in absorption, whereas a positive (in agreement with the π-π* character of the lowest
energy transition) but weak solvatochromic behavior was observed in emission for both compounds.
More interestingly, their fluorescence quantum yield is also not very sensitive to the increase of the
1
1

solvent polarity, maintaining values higher than 40% in a polar solvent such as acetonitrile. This means
that such compounds should remain fluorescent enough to monitor them in biological media.
Figure 8. Absorption (solid line) and emission (dotted line) of compound 6b in different solvents.
Table 4. Solvatochromic data of 6b at 25 °C.
abs a (nm)
449
em b (nm)
532
F c
0.50
0.56
0.42
Stokes shift (cm-1) d
Solvent
toluene
CH Cl
3470
3820
4210
449
446
542
549
2
2
CH CN
3
a
Absorption maximum. b Emission maximum. Fluorescence quantum yield using as a standard quinine
c
d
bisulfate in 0.5 M H SO , upon excitation at  . Stokes shift = (1/ - 1/ ).
2
4
abs
abs
em
Figure 9. Absorption (solid line) and emission (dotted line) of compound 6c in different solvents.
1
2

Table 5. Solvatochromic data of 6c at 25 °C.
abs a (nm)
437
em b (nm)
512
F c
0.49
0.56
0.42
Stokes shift (cm-1) d
Solvent
toluene
CH Cl
3350
3580
3990
437
434
518
525
2
2
CH CN
3
a
Absorption maximum. b Emission maximum. Fluorescence quantum yield using as a standard quinine
c
d
bisulfate in 0.5 M H SO , upon excitation at  . Stokes shift = (1/ - 1/ ).
2
4
abs
abs
em
2
.3. Evaluation on kinases and molecular modeling experiments
Protein kinases which are often deregulated in diseases such as cancers and neurodegenerative
disorders have become a major class of drug targets since the end of the nineties [21]. Today, 50 FDA-
approved kinase inhibitors are on the market and many drug candidates are undergoing clinical
evaluation [22, 23]. As part of our ongoing research on novel protein kinase inhibitors, some of the
synthesized compounds were also evaluated [16] against a short panel of disease-related protein
kinases: cyclin-dependent kinases 2 (CDK2/Cyclin A) and 9 (CDK9/Cyclin T), proto-oncogene kinase
PIM1, CDC2-like kinase 1 (CLK1), dual specificity tyrosine phosphorylation regulated kinase 1A
(DYRK1A), glycogen-synthase kinase-3 (GSK-3 isoforms α/β), casein kinase 1 (CK1 isoforms δ/ε) and
mitotic kinase Haspin (Table 6).
Table 6. Inhibitory activities of synthesized compounds against a short panel of disease-related protein
kinases. The table displays the remaining kinase activities detected after treatment with 10 µM of the
tested compounds. The values obtained after treatment with 1 µM are given in brackets. Results are
expressed in % of maximal activity, i.e. measured in the absence of inhibitor but with an equivalent
dose of DMSO (solvent of the tested compounds). ATP concentration used in the kinase assays was 10
µmol/L (values are means, n = 2). Kinases are from human origin unless specified: Mm, Mus musculus;
Rn, Rattus norvegicus; Ssc, Sus scrofa domesticus.
Compound CDK2/CyclinA CDK9/CyclinT PIM1
MmCLK1 RnDYRK1A SscGSK3α/β SscCK1δ/ε Haspin
2
3
3
3
3
5
6
6
a’
b
c
c’
e
b
b
c
106 (99)
98 (109)
95 (98)
103 (91)
93 (82)
95 (92)
87 (84)
78 (87)
79 (96)
71 (80)
100 (94)
56 (61)
77 (84)
81 (103)
57 (87)
70 (82)
104 (101) 105 (104) 141 (170)
107 (107) 97 (104) 103 (146)
110 (100) 81 (101) 68 (71)
103 (105)
98 (109)
86 (89)
101 (75)
100 (96)
91 (104)
90 (83)
85 (99)
103 (124) 96 (104)
98 (99)
84 (102)
112 (98)
92 (97)
95 (106)
68 (82)
94 (83)
87 (98)
114 (91)
77 (100)
89 (89)
76 (82)
56 (91)
105 (97)
98 (88)
106 (99)
2 (32)
106 (97) 124 (109)
95 (101) 107 (94)
108 (107) 74 (136)
45 (84) 58 (90)
52 (85) 60 (66)
13 (46)
85 (92)
1
3

While a weak activity was noticed for most of the new thioxanthone derivatives, both compounds 6b
and 6c were found surprisingly active against PIM1 at 10 μM and 1 μM. In order to clearly establish the
potency of the polycyclic compounds 6b and 6c, their IC values of inhibition against both isoforms
5
0
PIM1 and PIM2 were determined (Table 7). As noticed above, compound 6b (0.37 M) is a better
inhibitor of PIM1 than 6c (1.40 M). While the values for 6c are quite similar to those of HM107-g,
showing a higher affinity for PIM2, those of 6b do not show any preference for one of the isoforms.
The presence of a second sulfur atom in 6b seems to contribute to the inhibition of PIM1.
a
Table 7. IC values (M) for the most promising PIM inhibitors (ADP-Glo method at 10 µM ATP).
5
0
Compound
PIM1
PIM2
6
6
b
c
0.37
1.40
1.68 (0.61)
0.09 (0.01)
0.46
0.65
0.42
0.17 (0.63)
b
c
HM107-g
SGI-1776
b
a
b
A known inhibitor (SGI-1776) was also measured as control. Value obtained previously (radioactive method
c
at 15 µM ATP) [4]. Value obtained from the literature (radioactive method) [24].
In order to have a better idea of the selectivity of the inhibitors 6b and 6c, they were tested on an
additional panel of disease-related protein kinases: cyclin-dependent kinase 5 (CDK5/p25), proto-
oncogene 1, non-receptor tyrosine kinase ABL1, tyrosine-protein kinase JAK3, serine/threonine-protein
kinases Aurora B and Nek 6, glycogen-synthase kinase-3 (GSK3; isoforms α and β) and casein kinase 1
(CK1; isoform ε) (Table 8). Since competitive inhibition was only observed for JAK3 and GSK3β at 10
µM, the selectivity is rather good on this panel of 16 kinases including CLK, DYRK, GSK, CK and
Haspin which can be co-inhibited by PIM1 inhibitors [25, 26] (Tables 6 and 8).
Table 8. Inhibitory activities of synthesized compounds against another panel of disease-related protein
kinases. See details in Table 6. Kinases are from human origin.
Compound
CDK5/p25
ABL1
JAK3
AuroraB
Nek6
GSK3α
GSK3β
CK1ε
6
6
b
c
93 (88)
79 (≥100)
63 (88)
20 (89)
80 (≥100) ≥100 (≥100) 60 (98)
41 (86)
62 (≥100)
78 (≥100) 18 (≥100) 93 (≥100) ≥100 (≥100) 55 (≥100) 42 (≥100) 69 (≥100)
1
4

To help rationalize the good inhibitory activity recorded for the hexacyclic compounds, we
investigated the putative binding mode of the best active compound 6b within PIM1 ATP-binding
pocket using molecular modeling experiments. However, as already observed in the corresponding
ORTEP diagram (Figure 6), compound 6b is not fully planar and tends to adopt a helicoidal chiral
structure. As it is known that chiral helicenes might discriminate chiral biological targets [27, 28], and
although we work in racemic series, the possible helix interconversion deserves to be looked at.
The study of the interconversion mechanism between M and P helixes [29] using Gaussian 09 [30]
demonstrated a similar pathway and an energy barrier of the same order of magnitude than those
observed for the [5]helicenes [29, 31, 32], showing that the interconversion of the two forms is possible
(Figure 10).
Figure 10. Isomerization pathway between M and P helixes of 6b.
The docking studies were performed, for the two possible structures, with a PIM1 model generated
from 3JPV structure available in the Protein Data Bank [33] using AutoDock Vina [34, 35]. As 6b M
and P helixes can be interconverted, only the best docking result is described. The binding mode
1
5

presented in Figure 11A depicts the interaction between PIM1 and 6b M helix that, due to a better
adaptation of its curvature to the ATP site, penetrates deeper inside the pocket compared with the 6b P
helix. After molecular dynamic analysis, the observed binding mode showed that hydrogen-bonding
was established between compound 6b and the targeted kinase via water molecules, Wat1 and Wat2.
Both water molecules are commonly observed in PIM1 X-ray crystallographic structures such as 3JPV,
2
C3I [36], 6NO9 [37] and 5V82 [38]. However, the hexacyclic scaffold is highly stabilized in the ATP-
binding pocket via several hydrophobic interactions involving Ile104, Leu120, Val126, Leu174 and
Ile185 residues as well as Leu44 and Val52 (not indicated in Figure 11A to allow a better visualization
of the image).
A
B
Figure 11. A) Binding mode of compound 6b (M helix; A) and HM107-g (B)
within PIM1 ATP-binding site studied by molecular modeling experiments.
The images were produced using UCSF Chimera [39].
In previous studies, compound HM107-g, a related tetracyclic heteroaromatic was identified as well
as a sub-micromolar PIM1 inhibitor [4]. Therefore, in order to compare with 6b, the same method of
molecular modeling docking experiments was used to evidence the molecular interactions established
between HM107-g and PIM1 (Figure 11B). Compound HM107-g is fully planar, inserted into the
1
6

ATP-binding pocket where it is stabilized via hydrophobic interactions with the same residues as 6b.
Moreover, the amino group is H-bonded with backbone carbonyl group of Glu121 hinge region residue.
The results here reported constitute the first information related to the putative binding mode. They
give us valuable information useful to optimize the biological profile of this new series of
heteroaromatic compounds. For example, introduction of substituents on 6b could allow for the
formation of direct H-bonds with PIM1 binding site.
2
.4. Biological evaluation on cancer cell lines
As most of our original thioxanthone derivatives were lacking biological effect on kinases, we finally
evaluated their activity on cancer cell lines. Thus, the 1-(arylamino)-9-thioxanthones 2a’, 3a, 3b, 3c
and 3d were evaluated against K562 lymphoma cells, but none of them exhibited an antiproliferative
activity. The compound 6b as well as the PIM inhibitor HM107-g [4] were tested against HuH7 (liver),
CaCo-2 and HCT116 (colon), MCF7 (breast), MDA-MB-231 and MDA-MB-468 (triple neg. breast),
PC3 (prostate) cancer cell lines. In spite of the expression of PIM kinases by more than half of the
selected cells (e.g. HCT166 [40], MCF7 [41], MDA-MB-231 [42] and PC3 [43]), no clear
antiproliferative activity was detected after 48 h.
The compounds 6b and HM107-g [4] were also evaluated against melanoma cells. Indeed, due to
important mortality rate, the identification of other pharmacological targets and the development of new
drugs to treat melanoma are required [44]. In addition, while PIM kinases are expressed in tumor tissue
of melanoma patients, inhibitors such as SGI-1776 limit the invasion, proliferation and viability of
melanoma cells on in vitro models [45]. Because of their relevant mutations, the A2058 human
melanoma cells are appropriate to study the antiproliferative activity of both compounds. Growth
-5
inhibitions of 38.5% ± 4.2% and 48.1% ± 2.9% were respectively induced at 10 M after 72 h by the
compounds 6b and HM107-g. These anti-melanoma activities are promising [46], and the
rationalization of this effect will soon be studied.
1
7

3
.
Conclusion
Here, we have developed a short methodology to access 1-amino substituted thioxanthone
derivatives, highlighting an interesting N-arylation-cyclization process able to deliver hexacyclic
structures that would otherwise require multi-step syntheses. No biological activity was identified from
the synthesized 1-(arylamino)thioxanthones. In contrast, even if these results will need to be
complemented with more in-depth studies, the fluorescent [1]benzo(thio)pyrano[4,3,2-
de]benzothieno[2,3-b]quinoline derivatives 6b and 6c already proved to inhibit the kinases PIM1 and
PIM2 in a rather selective fashion; in addition, they showed promising results against melanoma cells.
Therefore, in view of the modeling studies related to their putative binding mode, these results lay the
ground for the rational design of a new generation of PIM-targeting fluorescent compounds.
4
.
Experimental
4
.1. General
All reactions were performed under an argon atmosphere. THF was freshly distilled over
sodium/benzophenone. The other solvents did not require any pre-treatment. Column chromatography
separations were achieved on silica gel (40-63 μm). Melting points were measured on a Kofler
1
13
apparatus. IR spectra were taken on a Perkin-Elmer Spectrum 100 spectrometer. H and C Nuclear
Magnetic Resonance (NMR) spectra were recorded either on a Bruker Avance III spectrometer at 300
MHz and 75 MHz respectively, on a Bruker Avance III spectrometer at 400 MHz and 100 MHz
1
respectively, or on a Bruker Avance III HD spectrometer at 500 MHz and 126 MHz respectively. H
1
3
chemical shifts (δ) are given in ppm relative to the solvent residual peak and C chemical shifts are
relative to the central peak of the solvent signal [47]. ZnCl ·TMEDA [48], activated Cu [49], 2-
2
iodobenzothiophene [50, 51] and 2-iodobenzofuran [50, 51] were prepared as described previously. All
reagents not listed in the publication were obtained from commercial sources.
1
8

Crystallography. CCDC 1942631 (1’), 1942632 (2a), 1942633 (2a’), 1942634 (2b’), 1942635 (3a),
942636 (3c), 1942637 (3c’), 1942638 (3d), 1942639 (3e), 1942640 (3e’), 1942641 (5b), 1942642 (6b)
1
and 1942643 (6c) contain the crystallographic data for this paper. These data can be obtained free of
charge from The Cambridge Crystallographic Data Centre.
The X-ray diffraction data were collected by using a D8 VENTURE Bruker AXS diffractometer
equipped with a (CMOS) PHOTON 100 detector at the temperature given in the crystal data. The
samples were studied with monochromatized Mo-K radiation ( = 0.71073 Å, multilayer
monochromator). The structure was solved by dual-space algorithm using the SHELXT program [52],
2
and then refined with full-matrix least-square methods based on F (SHELXL-2014) [53]. All non-
hydrogen atoms were refined with anisotropic atomic displacement parameters. Except the nitrogen-
linked hydrogen atoms that were introduced in the structural model through Fourier difference maps
analysis in the case of 2a, 2b’, 3a, 3c, 3d, 3e and 5b, H atoms were finally included in their calculated
positions (and treated as riding on their parent atom with constrained thermal parameters in the case of
2
a, 2a’, 2b’, 3d, 3e, 3e’ and 6b). The molecular diagrams were generated by ORTEP-3 (version 2.02)
[
54].
4
.2. 1-Iodo-9-thioxanthone (1). To a stirred mixture of 9-thioxanthone (0.20 g, 1.0 mmol) and
ZnCl ·TMEDA (0.26 g, 1.0 mmol) in THF (3 mL) at -30 °C was added dropwise a solution of LiTMP
2
(
prepared by adding BuLi (about 1.6 M hexanes solution, 1.5 mmol) to a stirred, cooled (0 °C) solution
of 2,2,6,6-tetramethylpiperidine (0.25 mL, 1.5 mmol) in THF (3 mL) and stirring for 5 min) cooled at -
0 °C. After 15 min at -30 °C, a solution of I (0.38 g, 1.5 mmol) in THF (5 mL) was introduced, and
3
2
the mixture was stirred overnight before addition of an aqueous saturated solution of Na S O (5 mL)
2
2
3
and extraction with AcOEt (3 x 20 mL). The combined organic layers were dried over MgSO , filtered
4
and concentrated under reduced pressure. The crude product was purified by chromatography over
silica gel (eluent: heptane-AcOEt 80:20). Compound 1 was isolated in 84% yield as a yellow powder:
1
9

1
mp 152 °C; H NMR (CDCl )  7.10 (t, 1H, J = 7.8 Hz), 7.41-7.47 (m, 2H), 7.52-7.60 (m, 2H), 8.14
3
1
3
(
dd, 1H, J = 7.5 and 1.2 Hz), 8.47 (dm, 1H, J = 8.1 Hz); C NMR (CDCl )  94.8 (C), 125.5 (CH),
3
1
26.7 (CH), 126.8 (CH), 127.7 (C), 129.6 (C), 130.2 (CH), 131.8 (CH), 132.3 (CH), 135.0 (C), 138.7
(C), 141.8 (CH), 179.5 (C). The analyses are as described previously [12]. 1,8-Diiodo-9-thioxanthone
(1’) was also isolated in 10% yield as a white powder: mp 162 °C; IR (ATR): 674, 725, 760, 776, 791,
-1 1
8
94, 919, 1060, 1080, 1162, 1205, 1278, 1423, 1545, 1567, 1654, 2852, 2922, 3048 cm ; H NMR
(
CDCl )  7.11 (t, 2H, J = 7.9 Hz), 7.51 (dd, 2H, J = 8.0 and 1.1 Hz), 8.05 (dd, 2H, J = 7.7 and 1.1 Hz);
3
1
3
C NMR (CDCl )  94.5 (2C), 125.9 (2CH), 131.5 (2CH), 132.0 (2C), 136.3 (2C), 140.7 (2CH), 183.7
3
(
C). Crystal data for 1’. C H I OS, M = 464.04, T = 150(2) K, triclinic, P -1, a = 7.5979(14), b =
1
3
6 2
3
1
2
1.307(2), c = 16.538(3) Å, α = 75.648(7), β = 73.865(6), γ = 72.133(6) °, V = 1278.3(4) Å , Z = 4, d =
-3
-1
2
.411 g.cm , μ = 5.062 mm . A final refinement on F with 5588 unique intensities and 308
2
parameters converged at ωR(F ) = 0.1077 (R(F) = 0.0509) for 4445 observed reflections with I > 2σ(I).
CCDC 1942631.
4
4
.3. N-Arylation of aromatic diamines by 1-iodo-9-thioxanthone (1)
.3.1. General procedure 1: A degassed mixture of the required aromatic diamine (2.0 mmol), 1-
iodo-9-thioxanthone (1; 0.34 g, 1.0 mmol), CuI (37 mg, 0.2 mmol), K CO (0.14 g, 1.0 mmol) and
2
3
DMSO (1 mL) was heated at 120 °C overnight. After cooling to room temperature, aqueous 25%
NH OH (5 mL) was added. Extraction using AcOEt (3x20 mL), removal of the solvent and purification
4
by chromatography on silica gel (the eluent is given in the product description) led to the expected
compound.
4
.3.2. 1-(2-Aminophenylamino)-8-thioxanthone (2a’). The general procedure 1 was applied to 1,2-
phenylenediamine (0.22 g) to afford the compound 2a’ (eluent: heptane-AcOEt 90:10) in 58% yield as
an orange powder: mp 184-185 °C; IR (ATR): 670, 718, 735, 769, 1081, 1155, 1183, 1227, 1267, 1307,
-1 1
1
381, 1435, 1463, 1491, 1513, 1556, 1568, 1590, 3360, 3443 cm ; H NMR (CDCl )  3.85 (br s, 2H),
3
2
0

6
7
1
8
.54 (dd, 1H, J = 8.5 and 1.1 Hz), 6.80 (td, 2H, J = 7.9 and 1.2 Hz), 6.85 (dd, 1H, J = 8.0 and 1.4 Hz),
.12 (td, 1H, J = 7.7 and 1.5 Hz), 7.20 (dd, 1H, J = 7.8 and 1.5 Hz), 7.25 (t, 1H, J = 8.0 Hz), 7.43 (ddd,
H, J = 8.2, 7.0 and 1.3 Hz), 7.48 (dd, 1H, J = 8.0 and 1.2 Hz), 7.56 (ddd, 1H, J = 8.3, 7.0 and 1.5 Hz),
1
3
.55 (dd, 1H, J = 8.2 and 1.4 Hz), 11.35 (br s, 1H); C NMR (CDCl )  110.1 (CH), 113.2 (CH), 113.5
3
(
C), 116.2 (CH), 119.0 (CH), 125.3 (CH), 125.6 (C), 126.1 (CH), 127.6 (CH), 128.1 (CH), 129.6 (CH),
30.5 (C), 132.1 (CH), 133.6 (CH), 136.7 (C), 139.5 (C), 143.3 (C), 152.6 (C), 183.8 (C). Crystal data
for 2a’. C H N OS, M = 318.38, T = 150(2) K, triclinic, P -1, a = 3.9485(10), b = 13.778(3), c =
1
1
9
14
2
7.364(7) Å, α = 78.249(8), β = 90.887(10), γ = 81.795(10) °, V = 1441.0(6) Å , Z = 4, d = 1.468 g.cm-
3
2
3,
μ = 0.231 mm . A final refinement on F with 6530 unique intensities and 331 parameters converged
-1
2
2
at ωR(F ) = 0.4732 (R(F) = 0.1994) for 4965 observed reflections with I > 2σ(I). CCDC 1942633.
Traces of 2a also formed in the same reaction, as identified by X-ray diffraction. Crystal data for 2a.
C H N S, M = 300.37, T = 150(2) K, orthorhombic, P b c a, a = 16.0429(8), b = 8.0742(4), c =
1
9
12
2
3
-3
-1
2
2
3
1.7145(9) Å, V = 2812.8(2) Å , Z = 8, d = 1.419 g.cm , μ = 0.227 mm . A final refinement on F with
2
229 unique intensities and 202 parameters converged at ωR(F ) = 0.0921 (R(F) = 0.0373) for 2761
observed reflections with I > 2σ(I). CCDC 1942632.
.3.3. 1-(3-Amino-2-pyridylamino)-9-thioxanthone (2b’). The general procedure 1 was applied to
,3-diaminopyridine (0.22 g) to afford the compound 2b’ (eluent: heptane-AcOEt 70:30) in 35% yield
as an orange powder: mp 196 °C; IR (ATR): 668, 715, 743, 768, 922, 1083, 1156, 1185, 1231, 1271,
4
2
-1 1
1
306, 1431, 1516, 1574, 1630, 2925 cm ; H NMR (CDCl )  3.91 (br s, 2H), 6.85 (dd, 1H, J = 7.7
3
and 4.8 Hz), 7.02 (dd, 1H, J = 7.9 and 1.1 Hz), 7.08 (dd, 1H, J = 7.8 and 1.6 Hz), 7.44 (ddd, 1H, J =
.2, 6.9 and 1.2 Hz), 7.47-7.51 (m, 2H), 7.58 (ddd, 1H, J = 8.2, 7.0 and 1.5 Hz), 7.90 (dd, 1H, J = 4.8
8
1
3
and 1.6 Hz), 8.50 (dd, 1H, J = 8.4 and 1.45 Hz), 8.59 (dd, 1H, J = 8.1 and 1.4 Hz), 12.66 (br s, 1H); C
NMR (CDCl )  113.9 (CH), 114.8 (C), 115.8 (CH), 118.4 (CH), 122.7 (CH), 125.3 (CH), 126.2 (CH),
3
1
29.9 (CH), 130.4 (C), 132.2 (CH), 133.0 (C), 133.7 (CH), 136.9 (C), 137.8 (CH), 139.3 (C), 143.7
(
C), 148.4 (C), 184.1 (C). Crystal data for 2b’. C H N OS, M = 319.37, T = 150(2) K, monoclinic,
1
8
13
3
2
1

3
P 2 /c, a = 13.8914(13), b = 5.2102(4), c = 19.8060(18) Å, β = 90.209(4) °, V = 1433.5(2) Å , Z = 4, d
1
-3
-1
2
=
1.480 g.cm , μ = 0.234 mm . A final refinement on F with 3276 unique intensities and 217
2
parameters converged at ωR(F ) = 0.0955 (R(F) = 0.0490) for 2256 observed reflections with I > 2σ(I).
CCDC 1942634.
4
4
.4. N-Arylation of anilines by 1-iodo-9-thioxanthone (1)
.4.1. General procedure 2: A degassed mixture of the required aniline (1.0 mmol), 1-iodo-9-
thioxanthone (1; 0.34 g, 1.0 mmol), CuI (28 mg, 0.15 mmol), K CO (0.28 g, 2.0 mmol) and DMSO
2
3
(
0.5 mL) was heated at 120 °C for 24 h. The purification procedure is given in the product description.
.4.2. 1-(4-(Trifluoromethylsulfonyl)phenylamino)-9-thioxanthone (3a). The general procedure 2
was applied to 4-[(trifluoromethyl)sulfonyl]aniline (0.23 g). After cooling to room temperature, water
5 mL) was added and the crude product was filtrated. Purification of the solid by chromatography on
4
(
silica gel (eluent: heptane-AcOEt 80:20) led to the expected compound 3a in 58% yield as a yellow
powder: mp 175 °C; IR (ATR): 666, 676, 704, 748, 756, 776, 1071, 1138, 1178, 1201, 1289, 1357,
-1 1
1
436, 1511, 1556, 1568, 3060 cm ; H NMR (CDCl )  7.15 (dd, 1H, J = 6.6 and 2.4 Hz), 7.45-7.55
3
(m, 6H), 7.63 (ddd, 1H, J = 8.3, 6.9 and 1.5 Hz), 7.92-7.95 (m, 2H), 8.52 (ddd, 1H, J = 8.1, 1.5 and 0.6
1
3
Hz), 12.27 (br s, 1H); C NMR (CDCl )  113.1 (CH), 116.5 (C), 118.2 (CH), 119.1 (2CH), 121.9 (C),
3
1
1
4
=
25.4 (CH), 126.7 (CH), 129.8 (CH), 130.2 (C), 132.7 (CH), 133.0 (2CH), 133.1 (CH), 136.6 (C),
40.5 (C), 146.6 (C), 149.4 (C), 184.0 (C), CF not seen. Crystal data for 3a. C H F NO S , M =
3
20 12
3
3 2
35.43, T = 150(2) K, triclinic, P -1, a = 7.2523(12), b = 7.9427(13), c = 16.411(3) Å, α = 98.518(6), β
3
-3
-1
100.915(6), γ = 101.305(6) °, V = 893.2(3) Å , Z = 2, d = 1.619 g.cm , μ = 0.352 mm . A final
2
2
refinement on F with 4054 unique intensities and 140 parameters converged at ωR(F ) = 0.3297 (R(F)
0.1093) for 3779 observed reflections with I > 2σ(I). CCDC 1942635.
.4.3. 1-(4-Fluorophenylamino)-9-thioxanthone (3b). The general procedure 2 was applied to 4-
=
4
fluoroaniline (95 μL). After cooling to room temperature, an aqueous saturated solution of NH Cl (5
4
mL) was added and the product was extracted by using AcOEt (3x20 mL). Drying over Na SO ,
2
4
2
2

removal of the solvent and purification by chromatography on silica gel (eluent: heptane-AcOEt 75:25)
led to the expected compound 3b in 31% yield as a yellow powder: mp 114 °C; IR (ATR): 669, 744,
-
1 1
7
1
59, 826, 1151, 1180, 1210, 1267, 1437, 1512, 1560, 1578, 3060 cm ; H NMR (CDCl )  6.85 (dd,
3
H, J = 7.7 and 1.1 Hz), 6.94 (dd, 1H, J = 8.5 and 1.1 Hz), 7.05-7.13 (m, 2H), 7.24-7.32 (m, 3H), 7.45
(ddd, 1H, J = 8.3, 6.9 and 1.4 Hz), 7.46-7.49 (m, 1H), 7.59 (ddd, 1H, J = 8.3, 6.9 and 1.5 Hz), 8.55
1
3
(
ddd, 1H, J = 8.1, 1.5 and 0.6 Hz), 11.75 (br s, 1H); C NMR (CDCl )  109.6 (CH), 113.5 (CH),
3
1
13.6 (C), 116.4 (d, 2CH, J = 22.4 Hz), 125.2 (CH), 126.2 (CH), 126.4 (d, 2CH, J = 8.1 Hz), 129.6
(CH), 130.4 (C), 132.1 (CH), 133.4 (CH), 136.3 (C), 136.7 (C), 139.7 (C), 152.0 (C), 160.1 (d, C, J =
2
7
44 Hz), 183.7 (C). Anal. Calcd for C H FNOS (321.37): C, 71.01; H, 3.76; N, 4.36. Found: C,
19 12
1.14; H, 3.95; N, 4.29.
.4.4. 1-(4-Tolylamino)-9-thioxanthone (3c). The general procedure 2 was applied to 4-toluidine
0.11 g). After cooling to room temperature, an aqueous saturated solution of NH Cl (5 mL) was added
4
(
4
and the product was extracted by using AcOEt (3x20 mL). Drying over Na SO , removal of the solvent
2
4
and purification by chromatography on silica gel (eluent: heptane-AcOEt-Et N 80:15:5) led to a
3
mixture from which the expected compound 3c (estimated yield from a mixture with 3c’: 50% yield)
1
was identified by NMR and crystal data. H NMR (CDCl , selected data)  1.54 (s, 3H), 6.80 (dd, 1H, J
3
=
7.7 and 1.1 Hz), 7.02 (dd, 1H, J = 8.5 and 1.1 Hz), 8.52 (dd, 1H, J = 8.1 and 1.2 Hz), 11.75 (br s,
1
H). Crystal data for 3c. C H NOS, M = 317.39, T = 150(2) K, monoclinic, P 2 /c, a = 12.8049(11),
2
0
15
1
3
-3
b = 12.0640(10), c = 10.7591(8) Å, β = 114.737(3) °, V = 1509.5(2) Å , Z = 4, d = 1.397 g.cm , μ =
-1
2
0
.218 mm . A final refinement on F with 3443 unique intensities and 212 parameters converged at
2
ωR(F ) = 0.0944 (R(F) = 0.0360) for 2849 observed reflections with I > 2σ(I). CCDC 1942636. The
compound 3c’ was also isolated in 6% yield as a red powder: mp 221 °C; IR (ATR): 670, 717, 741,
-1 1
7
54, 767, 1184, 1229, 1268, 1436, 1497, 1513, 1558, 1568, 1608, 3051 cm ; H NMR (CDCl )  2.30
3
(s, 3H), 7.05-7.19 (m, 12H), 7.28 (ddd, 2H, J = 8.9, 7.3 and 1.6 Hz), 7.36 (ddd, 2H, J = 7.8, 1.6, 0.6
1
3
Hz), 7.45 (t, 2H, J = 8.0 Hz); C NMR (CDCl )  21.1 (CH ), 119.9 (2CH), 121.8 (2C), 124.5 (2CH),
3
3
2
3

1
24.8 (2CH), 126.1 (2CH), 126.6 (2CH), 128.1 (2CH), 130.0 (2CH), 130.9 (2CH), 132.4 (2CH), 132.6
2C), 133.8 (C), 134.9 (2C), 138.4 (2C), 145.8 (C), 150.0 (2C), 181.9 (2C). Crystal data for 3c’.
C H NO S , M = 527.63, T = 150(2) K, monoclinic, P 2 /c, a = 18.500(3), b = 9.1189(18), c =
(
3
3
21
2
2
1
3
-3
-1
1
5.199(2) Å, β = 95.212(6) °, V = 2553.4(8) Å , Z = 4, d = 1.373 g.cm , μ = 0.241 mm . A final
2
2
refinement on F with 5824 unique intensities and 344 parameters converged at ωR(F ) = 0.0952 (R(F)
0.0406) for 4756 observed reflections with I > 2σ(I). CCDC 1942637.
.4.5. 1-(α-Naphthylamino)-9-thioxanthone (3d). The general procedure 2 was applied to α-
=
4
naphthylamine (0.14 g). After cooling to room temperature, an aqueous saturated solution of NH Cl (5
4
mL) was added and the product was extracted by using AcOEt (3x20 mL). Drying over Na SO ,
2
4
removal of the solvent and purification by chromatography on silica gel (eluent: heptane-AcOEt 85:15)
led to the expected compound 3d in 63% yield as an orange powder: mp 198 °C; IR (ATR): 670, 717,
7
42, 755, 767, 1082, 1154, 1184, 1229, 1268, 1376, 1436, 1497, 1514, 1558, 1569, 1592, 1608, 3050
-1 1
cm ; H NMR (CDCl )  6.76 (dd, 1H, J = 8.5 and 1.1 Hz), 6.82 (dd, 1H, J = 7.7 and 1.1 Hz), 7.20 (t,
3
1
7
1
1
1
H, J = 8.1 Hz), 7.45 (ddd, 1H, J = 8.2, 7.0 and 1.3 Hz), 7.42-7.61 (m, 6H), 7.76 (d, 1H, J = 8.2 Hz),
.88-7.93 (m, 1H), 8.15 (dt, 1H, J = 8.1 and 1.0 Hz), 8.59 (ddd, 1H, J = 8.2, 1.5 and 0.6 Hz), 12.2 (br s,
1
3
H); C NMR (CDCl )  110.5 (CH), 113.3 (CH), 113.6 (C), 122.3 (CH), 123.2 (CH), 125.2 (CH),
3
26.0 (CH), 126.0 (CH), 126.1 (CH), 126.5 (CH), 126.5 (CH), 128.5 (CH), 129.7 (CH), 130.1 (C),
30.5 (C), 132.1 (CH), 133.4 (CH), 135.0 (C), 136.5 (C), 136.8 (C), 139.5 (C), 152.9 (C), 183.9 (C).
Crystal data for 3d. C H NOS, M = 353.42, T = 150(2) K, monoclinic, P 2 /c, a = 25.757(6), b =
2
3
15
1
3
-3
-
3
.8988(10), c = 15.909(4) Å, β = 90.430(7) °, V = 1597.5(7) Å , Z = 4, d = 1.469 g.cm , μ = 0.215 mm
1.
A final refinement on F with 3616 unique intensities and 238 parameters converged at ωR(F ) =
2
2
0
.1196 (R(F) = 0.0590) for 2517 observed reflections with I > 2σ(I). CCDC 1942638.
.4.6. 1-(4-Methoxyphenylamino)-9-thioxanthone (3e). The general procedure 2, but with only 0.5
4
mmol of the required aniline, was applied to 4-anisidine (62 mg). After cooling to room temperature, an
aqueous saturated solution of NH Cl (5 mL) was added and the products were extracted by using
4
2
4

AcOEt (3x20 mL). Drying over Na SO , removal of the solvent and purification by chromatography on
2
4
silica gel (eluent: heptane-CH Cl 80:20) led to a mixture from which the compound 3e (estimated
2
2
1
yield < 10% yield) was identified by NMR and crystal data. H NMR (CDCl )  3.84 (s, 3H), 6.77 (dd,
3
1
H, J = 7.7 and 1.1 Hz), 6.86 (dd, 1H, J = 8.5 and 1.1 Hz), 6.92-6.97 (m, 2H), 7.20-7.27 (m, 3H), 7.42
(
ddd, 1H, J = 8.3, 6.9 and 1.4 Hz), 7.44-7.48 (m, 1H), 7.55 (ddd, 1H, J = 8.3, 6.9 and 1.5 Hz), 8.52
ddd, 1H, J = 8.1, 1.5 and 0.6 Hz), 11.65 (br s, 1H). Crystal data for 3e. C H NO S, M = 333.39, T =
(
2
0
15
2
1
50(2) K, triclinic, P -1, a = 7.7122(7), b = 8.5030(7), c = 12.1561(10) Å, α = 80.111(3), β = 79.888(4),
3
-3
-1
2
γ = 82.854(3) °, V = 769.46(11) Å , Z = 2, d = 1.439 g.cm , μ = 0.222 mm . A final refinement on F
2
with 3499 unique intensities and 221 parameters converged at ωR(F ) = 0.1269 (R(F) = 0.0491) for
2
897 observed reflections with I > 2σ(I). CCDC 1942639. The compound 3e’ was isolated in 41% yield
as a red powder: mp 194 °C; IR (ATR): 682, 727, 749, 770, 783, 832, 915, 1026, 1176, 1242, 1261,
-
1 1
1
2
277, 1435, 1504, 1570, 1642, 2965 cm ; H NMR (CDCl )  3.79 (s, 3H), 6.80-6.85 (m, 2H), 7.06 (d,
3
H, J = 7.9 Hz), 7.09-7.17 (m, 8H), 7.16-7.21 (m, 2H), 7.33 (d, 2H, J = 9.0 Hz), 7.44 (t, 2H, J = 8.0
1
3
Hz); C NMR (CDCl )  55.7 (CH ), 114.9 (CH), 119.5 (CH), 121.1 (C), 124.4 (CH), 126.1 (CH),
3
3
1
26.1 (CH), 127.0 (CH), 128.0 (CH), 130.9 (CH), 132.4 (CH), 132.6 (C), 134.8 (C), 138.4 (C), 141.3
(
C), 150.3 (C), 156.9 (C), 182.0 (C). Crystal data for 3e’. C H NO S , M = 543.63, T = 150(2) K,
3
3
21
3 2
monoclinic, P 2 /n, a = 15.0958(5), b = 9.4474(4), c = 18.1199(6) Å, β = 103.5890(10) °, V =
1
3
-3
-1
2
2
511.85(16) Å , Z = 4, d = 1.438 g.cm , μ = 0.251 mm . A final refinement on F with 5710 unique
2
intensities and 353 parameters converged at ωR(F ) = 0.0905 (R(F) = 0.0408) for 4390 observed
reflections with I > 2σ(I). CCDC 1942640.
4
.4.7. 1-(4-Pyridylamino)-9-thioxanthone. A mixture of 4-aminopyridine (94 mg, 1.0 mmol), 1-
iodo-9-thioxanthone (1; 0.40 g, 1.2 mmol), Cu (11 mg, 0.16 mmol), CuI (11 mg, 0.06 mmol), K CO
2
3
(0.28 g, 2.0 mmol) and DMF (0.5 mL) was heated at 140 °C for 6 h. After cooling to room temperature,
water (5 mL) was added and the crude product was filtrated. Purification of the solid by
chromatography on silica gel (eluent: CH Cl -heptane 40:60 to 100:00) led to 3f in 49% yield as a
2
2
2
5

yellow powder: mp 150 °C; IR (ATR): 715, 747, 772, 814, 909, 961, 1079, 1159, 1182, 1277, 1310,
-
1 1
1
1
2
444, 1515, 1557, 1576, 1734, 2855, 2924, 2955 cm ; H NMR (CDCl )  7.09 (dd, 1H, J = 7.1 and
3
.9 Hz), 7.21-7.23 (m, 2H), 7.42-7.53 (m, 4H), 7.60 (ddd, 1H, J = 8.3, 6.9 and 1.5 Hz), 8.45-8.47 (m,
1
3
H), 8.52 (ddd, 1H, J = 8.1, 1.5 and 0.6 Hz), 12.03 (br s, 1H); C NMR (CDCl )  113.0 (CH), 113.8
3
(
(
(
CH), 116.2 (C), 117.5 (CH), 125.4 (2CH), 126.6 (CH), 129.8 (CH), 130.3 (C), 132.6 (CH), 133.1
CH), 136.6 (C), 140.3 (C), 147.0 (C), 148.9 (C), 150.1 (2CH), 183.9 (C). Anal. Calcd for C H N OS
1
8
12
2
304.37): C, 71.03; H, 3.97; N, 9.20. Found: C, 71.37; H, 4.29; N, 8.94.
4
.5. 1-Amino-9-thioxanthone (4) [55] was prepared by slightly modifying a literature procedure
[
19]. A degassed (argon) mixture of 1-iodo-9-thioxanthone (1; 0.51 g, 1.5 mmol), CuI (29 mg, 0.15
mmol), DMEDA (24 μL, 0.22 mmol) and DMSO (0.75 mL) was treated by 25% NH OH (2.3 mL) and
4
the mixture was heated in a sealed tube at 130 °C for 16 h. After cooling to room temperature, brine (10
mL) was added before extraction using AcOEt (3x20 mL), removal of the solvent and purification by
chromatography on silica gel (eluent: heptane-AcOEt 85:15). Compound 4 was isolated in 50% yield as
1
a yellow powder: mp 145 °C; H NMR (CDCl )  6.56 (dd, 1H, J = 8.2 and 1.1 Hz), 6.76 (dd, 1H, J =
3
7
7
.7 and 1.1 Hz), 6.99 (br s, 2H), 7.25 (t, 1H, J = 8.0 Hz), 7.40 (ddd, 1H, J = 8.3, 6.9 and 1.4 Hz), 7.42-
1
3
.46 (m, 1H), 7.53 (ddd, 1H, J = 8.3, 6.9 and 1.5 Hz), 8.50 (ddd, 1H, J = 8.0, 1.5 and 0.6 Hz); C NMR
(
CDCl )  113.0 (CH), 113.2 (CH), 113.4 (C), 125.3 (CH), 126.0 (CH), 129.6 (CH), 130.5 (C), 132.0
3
(CH), 133.3 (CH), 136.7 (C), 139.1 (C), 153.4 (C), 183.4 (C).
4
4
.6. N-arylation of 1-amino-9-thioxanthone (4)
.6.1. General procedure 3: The N-arylated substrates were prepared by slightly modifying a
literature procedure [56]. To 1-amino-9-thioxanthone (4; 0.34 g, 1.5 mmol) and the required iodide (2.2
mmol) in Bu O (2 mL) were successively added activated Cu (19 mg, 0.30 mmol) and K CO (0.43 g,
2
2
3
3
.0 mmol). The mixture was degassed and refluxed under argon for 24 h. After cooling to room
2
6

temperature, the mixture was concentrated. Purification by chromatography on silica gel (the eluent is
given in the product description) led to the expected compound.
4
.6.2. 1-(2-Thienylamino)-9-thioxanthone (5a). The general procedure 3 (reaction time: 24 h) using
-iodothiophene (0.24 mL) gave 5a (eluent: heptane-AcOEt 80:20) in 40% yield as an orange powder:
mp 160 °C; IR (ATR): 747, 1160, 1180, 1194, 1230, 1248, 1284, 1440, 1459, 1478, 1592, 1729, 2925,
2
-
1 1
3
1
057, 3462 cm ; H NMR (CDCl )  6.86-6.90 (m, 2H), 6.97 (dd, 1H, J = 5.7 and 3.7 Hz), 7.01 (dd,
3
H, J = 8.5 and 1.1 Hz), 7.09 (dd, 1H, J = 5.5 and 1.4 Hz), 7.32 (ddd, 1H, J = 8.3, 7.7 and 0.5 Hz), 7.44
(ddd, 1H, J = 8.3, 6.9 and 1.4 Hz), 7.46-7.50 (m, 1H), 7.57 (ddd, 1H, J = 8.3, 6.9 and 1.5 Hz), 8.54
1
3
(
ddd, 1H, J = 8.1, 1.5, 0.6 Hz), 11.75 (s, 1H); C NMR (CDCl )  110.3 (CH), 113.7 (C), 114.2 (CH),
3
1
1
21.4 (CH), 122.0 (CH), 125.3 (CH), 126.2 (CH), 126.2 (CH), 129.7 (CH), 130.3 (C), 132.2 (CH),
33.6 (CH), 136.8 (C), 139.5 (C), 143.3 (C), 152.9 (C), 183.8 (C). Anal. Calcd for C H NOS
1
7
11
2
(
309.40): C, 65.99; H, 3.58; N, 4.53. Found: C, 66.14; H, 3.51; N, 4.36.
.6.3. 1-(2-Benzothienylamino)-9-thioxanthone (5b). The general procedure 3 using 2-
iodobenzothiophene (0.57 g) gave 5b (eluent: heptane) in 50% yield as an orange powder: mp 148 °C;
4
-1 1
IR (ATR): 670, 715, 745, 774, 1084, 1143, 1268, 1436, 1477, 1538, 1558, 1589, 1604, 3054 cm ; H
NMR (CDCl )  6.95 (dd, 1H, J = 7.5 and 1.4 Hz), 7.11 (s, 1H), 7.28-7.38 (m, 3H), 7.38 (t, 1H, J = 7.4
3
Hz), 7.46 (ddd, 1H, J = 8.3, 6.9 and 1.4 Hz), 7.48-7.52 (m, 1H), 7.59 (ddd, 1H, J = 8.3, 6.9 and 1.5 Hz),
1
3
7
.66-7.69 (m, 1H), 7.74-7.76 (m, 1H), 8.56 (ddd, 1H, J = 8.1, 1.5 and 0.6 Hz), 12.16 (s, 1H); C NMR
(
CDCl )  110.8 (CH), 114.0 (C), 115.0 (CH), 115.6 (CH), 122.3 (CH), 122.9 (CH), 123.9 (CH), 124.6
3
(CH), 125.3 (CH), 126.3 (CH), 129.7 (CH), 130.2 (C), 132.3 (CH), 133.5 (CH), 136.4 (C), 136.7 (C),
1
39.0 (C), 139.7 (C), 143.5 (C), 151.2 (C), 183.7 (C). Crystal data for 5b. C H NOS , M = 359.44,
21 13 2
T = 150(2) K, monoclinic, P 2 /c, a = 14.1321(8), b = 15.6356(8), c = 7.3904(4) Å, β = 90.285(2) °, V =
1
3
-3
-1
2
1
632.99(15) Å , Z = 4, d = 1.462 g.cm , μ = 0.335 mm . A final refinement on F with 3686 unique
2
intensities and 230 parameters converged at ωR(F ) = 0.1252 (R(F) = 0.0513) for 2883 observed
reflections with I > 2σ(I). CCDC 1942641.
2
7

4
.6.4. [1]Benzothiopyrano[4,3,2-de]benzothieno[2,3-b]quinoline (6b). The general procedure 3
using 2-iodobenzothiophene (0.57 g) gave 6b (eluent: heptane-AcOEt 80:20) in 38% yield as a yellow
powder: mp 176 °C; IR (ATR): 690, 728, 752, 784, 812, 1064, 1111, 1147, 1160, 1199, 1309, 1354,
-
1 1
1
1
8
1
473, 1537 cm ; H NMR (CDCl )  7.22-7.25 (m, 2H), 7.40-7.46 (m, 3H), 7.54 (dd, 1H, J = 7.9 and
3
.3 Hz), 7.60 (dd, 1H, J = 8.4 and 7.3 Hz), 7.82 (ddd, 1H, J = 7.9, 1.2 and 0.6 Hz), 7.85 (dd, 1H, J =
1
3
.5, 1.1 Hz), 8.19 (dd, 1H, J = 8.0 and 1.3 Hz), 8.23 (dd, 1H, J = 8.2 and 1.0 Hz); C NMR (CDCl ) 
3
20.2 (CH), 123.3 (CH), 123.4 (C), 123.7 (C), 124.0 (CH), 124.3 (CH), 125.4 (CH), 126.3 (CH), 128.0
(
CH), 128.2 (CH), 129.0 (C), 129.1 (CH), 130.0 (C), 130.5 (CH), 131.0 (CH), 133.3 (C), 135.8 (C),
37.6 (C), 138.6 (C), 147.9 (C), 165.9 (C). Crystal data for 6b. C H NS , M = 341.43, T = 150(2) K,
1
2
1
11
2
3
orthorhombic, P b c a, a = 18.151(3), b = 7.8809(10), c = 21.117(2) Å, V = 3020.8(7) Å , Z = 8, d =
-3
-1
2
1
.501 g.cm , μ = 0.353 mm . A final refinement on F with 3461 unique intensities and 217
2
parameters converged at ωR(F ) = 0.1366 (R(F) = 0.0610) for 2353 observed reflections with I > 2σ(I).
CCDC 1942642.
4
.6.5. [1]Benzothiopyrano[4,3,2-de]benzofuro[2,3-b]quinoline (6c). The general procedure 3 using
-iodobenzofuran (0.54 g) gave 6c after removal of the solvent and purification by chromatography on
silica gel (eluent: heptane-CH Cl 95:5) in 21% yield as an orange powder: mp 184 °C; IR (ATR): 671,
2
2
2
7
2
17, 730, 747, 894, 977, 1033, 1079, 1160, 1194, 1228, 1282, 1308, 1325, 1436, 1458, 1562, 1591,
-1 1
924, 3054, 3306, 3429 cm ; H NMR (CDCl )  7.31 (ddd, 1H, J = 8.3, 7.4 and 1.2 Hz), 7.36-7.61
3
(m, 6H), 7.64 (d, 1H, J = 7.8 Hz), 7.85 (dd, 1H, J = 8.4 and 1.2 Hz), 8.22 (d, 1H, J = 7.9 Hz), 8.43 (dd,
1
3
1
H, J = 7.4 and 1.6 Hz); C NMR (CDCl )  112.0 (C), 112.2 (CH), 119.7 (CH), 122.8 (C), 123.0
3
(CH), 123.6 (C), 123.7 (CH), 124.8 (CH), 126.4 (CH), 127.3 (CH), 128.8 (C), 129.2 (CH), 129.3 (CH),
1
29.8 (CH), 130.3 (C), 130.7 (CH), 135.7 (C), 138.3 (C), 147.3 (C), 156.0 (C), 164.3 (C). Crystal data
for 6c. C H NOS, M = 325.37, T = 150(2) K, orthorhombic, P 2 2 2 , a = 7.5722(4), b =
2
1
11
1
1
1
3
-3
-1
1
8.9340(13), c = 20.5024(14) Å, V = 2939.5(3) Å , Z = 8, d = 1.470 g.cm , μ = 0.227 mm . A final
2
8

2
2
refinement on F with 6761 unique intensities and 434 parameters converged at ωR(F ) = 0.1035 (R(F)
=
0.0513) for 5265 observed reflections with I > 2σ(I). CCDC 1942643.
4
.7. Physicochemical measurements
Measurements have been performed on freshly-prepared air-equilibrated solutions at room
temperature (25 °C). UV-Vis absorption spectra were recorded on a Jasco V-570 spectrophotometer.
Steady-state and time-resolved fluorescence measurements were performed in dilute solutions
contained in quartz cells of 1 cm pathlength using an Edinburgh Instrument (FLS920) fluorimeter
equipped with a 450 W Xenon lamp and a Peltier-cooled Hamamatsu R928P photomultiplier tube in
photon-counting mode. Fully corrected emission spectra were obtained, for each compound, after
excitation at the wavelength of the absorption maximum, with Aλex < 0.1 to minimize internal
absorption. Quinine bisulfate in 0.5 M H SO ( = 0.546 at  = 346 nm) was used as a standard.
2
4
ex
Fluorescence lifetimes were measured by time correlated single-photon counting (TCSPC) by using the
same FLS 920 fluorimeter. Excitation was achieved by a hydrogen-filled nanosecond flashlamp
(repetition rate 40 kHz). The instrument response (FWHM ca. 1 ns) was determined by measuring the
light scattered by a Ludox suspension. The TCSPC traces were analyzed by standard iterative
reconvolution methods implemented in the software of the fluorimeter. All compounds displayed
strictly monoexponential fluorescence decays.
4
.8. Evaluation on kinases and molecular modeling experiments
The enzymatic activities of CDK2/Cyclin A, CDK5/p25, CDK9/Cyclin T, PIM1, CLK1, DYRK1A,
GSK-3 isoforms α and/or β, CK1 isoforms δ/ε or ε, Haspin, ABL1, JAK3, Aurora B and Nek 6 kinases
were assayed using the ADP-GloTM bioluminescent kinase assay kit (Promega, Madison, WI) as
previously described [16].
Geometric optimization of 6b structures was obtained with Gaussian 09 [30] at the DFT level of
theory using B3LYP functional and 6-31g basis set. The two helixes differ only in the values of two
2
9