Bio-inspired benzo[k,l]xanthene lignans: Synthesis, DNA-interaction and antiproliferative properties

Article abstract of DOI:10.1039/c3ob42521e

In this work twelve benzo[k,l]xanthene lignans were synthesized by biomimetic, Mn-mediated oxidative coupling of caffeic esters and amides. These compounds, bearing different flexible pendants at position C1/C2 of the aromatic core, interact with DNA in a dual mode, as confirmed by DF-STD NMR analysis and molecular docking: the planar core acts as a base pair intercalant, whereas the flexible pendants act as minor groove binders. Their antiproliferative activity was evaluated on a panel of six tumor cell lines: HT-29, Caco-2, HCT-116 (human colon carcinoma), H226, A549 (human lung carcinoma), and SH-SY5Y (human neuroblastoma). All compounds under study, except 29, resulted in activity against one or more cell lines, and the markedly lipophilic esters 13 and 28 showed the highest activity. Compound 13 was more active than the anticancer drug 5-fluorouracil (5-FU) towards HCT-116 (colon, GI₅₀ = 3.16 μM) and H226 (lung, GI₅₀ = 4.33 μM) cell lines. This journal is the Partner Organisations 2014.

Full text of DOI:10.1039/c3ob42521e

Organic &
Biomolecular Chemistry
View Article Online
View Journal | View Issue
PAPER
Bio-inspired benzo[k,l]xanthene lignans: synthesis,
DNA-interaction and antiproliferative properties†
Carmela Spatafora,a Vincenza Barresi,b Vedamurthy M. Bhusainahalli,‡a
Simone Di Micco,c Nicolò Musso,b Raffaele Riccio,c Giuseppe Bifulco,c
Daniele Condorellib and Corrado Tringali*a
Cite this: Org. Biomol. Chem., 2014,
12, 2686
In this work twelve benzo[k,l]xanthene lignans were synthesized by biomimetic, Mn-mediated oxidative
coupling of caffeic esters and amides. These compounds, bearing different flexible pendants at position
C1/C2 of the aromatic core, interact with DNA in a dual mode, as confirmed by DF-STD NMR analysis
and molecular docking: the planar core acts as a base pair intercalant, whereas the flexible pendants act
as minor groove binders. Their antiproliferative activity was evaluated on a panel of six tumor cell lines:
HT-29, Caco-2, HCT-116 (human colon carcinoma), H226, A549 (human lung carcinoma), and SH-SY5Y
(human neuroblastoma). All compounds under study, except 29, resulted in activity against one or more
cell lines, and the markedly lipophilic esters 13 and 28 showed the highest activity. Compound 13 was
more active than the anticancer drug 5-fluorouracil (5-FU) towards HCT-116 (colon, GI50 = 3.16 µM) and
H226 (lung, GI50 = 4.33 µM) cell lines.
Received 17th December 2013,
Accepted 21st February 2014
DOI: 10.1039/c3ob42521e
www.rsc.org/obc
related compounds, well-known for their large structural
variety and for including examples of highly bioactive com-
Introduction
Renewed attention to natural or nature-derived bioactive com- pounds, the most famous being podophyllotoxin (1, Fig. 1)
pounds as lead compounds in drug discovery has been whose optimization has afforded the anticancer drugs etopo-
observed in recent years.1 This trend is probably connected side (2), etopophos (3) and teniposide (4).5
with the observation that the significant decrease in the
Dimeric lignans are biosynthetically originated by coupling
number of new molecular entities (NMEs) drugs approved by of two phenylpropanoid (C6C3) units, and a number of lignans
the U.S. Food and Drug Administration over the period and analogues have been synthesized by biomimetic oxidative
1990–2010 might be related to the reduced attention to natural coupling;6 among these, benzo[k,l]xanthene lignans (in the fol-
products as sources of new therapeutic leads towards the end lowing, more briefly ‘benzoxanthene lignans’) have recently
of 20th century.2 Moreover, many anticancer drugs available been made available by some of us through a simple bio-
today have been developed from natural leads,3 and a number mimetic methodology,7 in the frame of our studies on nature-
of new anticancer candidates derived from natural products derived compounds as potential antitumor agents.8 The oxi-
are currently in Phase II or Phase III clinical trials.4 Thus, dative coupling of caffeic esters allowed us to obtain in good
many families of natural products have been investigated in an yields both natural and ‘unnatural’ lignans with a planar and
effort to discover or develop new bio-inspired antitumor intensively fluorescent benzo[k,l]xanthene core, very rarely
agents, and hundreds of promising synthetic analogues have found in nature. To the best of our knowledge, to date only
been obtained. Many of these studies refer to lignans and five natural benzoxanthene lignans have been reported in the
literature, namely rufescidride (5) from Cordia rufescens9 and
Taraxacum mongolicum,10 mongolicumin A (6) from T. mongoli-
cum,10 yunnaneic acid H (7) from Salvia yunnanensis,11 chilian-
aDipartimento di Scienze Chimiche, Università degli Studi di Catania, Viale A. Doria
tin D (8) from Rhoiptelea chiliantha12 and dodegranoside C (9)
6, I-95125 Catania, Italy. E-mail: ctringali@unict.it
from Dodecadenia grandiflora.13 By biomimetic dimerization of
bDipartimento di Scienze Bio-Mediche, sezione di Biochimica, Università degli Studi
methyl caffeate (10) and CAPE (Caffeic Acid Phenethyl Ester,
11), the latter being an anti-inflammatory, antioxidant and
antiproliferative component of propolis,14 we have firstly
obtained, respectively, the synthetic benzoxanthenes 12 and
13; also the natural benzoxanthenes 5, 6 were synthesized. We
have proposed a probable mechanism of formation of these
di Catania, Viale A. Doria 6, I-95125 Catania, Italy
cDipartimento di Farmacia, Università di Salerno, via Giovanni Paolo II, 132,
84084 Fisciano (SA), Italy
†Electronic supplementary information (ESI) available. See DOI: 10.1039/
c3ob42521e
‡Present address: Department of Inorganic and Physical Chemistry, Indian Insti-
tute of Science, CV Raman Road, Bangalore, Karnataka 560012, India.
2686 | Org. Biomol. Chem., 2014, 12, 2686–2701
This journal is © The Royal Society of Chemistry 2014

View Article Online
Organic & Biomolecular Chemistry
Paper
Fig. 1 Structures of compounds 1–9 and 12–17.
compounds (corroborated by calculations) by a β–β′ (8–8′) oxi- guanines and with the anomeric oxygen of deoxyribose,
dative coupling of two caffeoyl residues. whereas the C-1/C-2 pendants are collocated along the grooves
The lack of biological data on benzoxanthene lignans, as of the nucleic acid and make contact with the external deoxy-
well as their planar and extensively conjugated nucleus, ribose/backbone. These results were in agreement with biologi-
prompted us to study their DNA-interaction and antiprolifera- cal data: indeed, the benzoxanthene 13 (GI50 = 2.57 µM on
tive properties: to establish the role of the hydroxyl groups in SW480 cells), with the predicted higher affinity for the minor
C-6, C-9 and C-10 and define some structure–activity relation- groove, proved significantly more active than 12 (GI50
=
ships, we prepared the permethyl derivatives 14 and 15 and 4.76 µM on SW480 cells); the permethylated derivatives 14 and
the peracetyl derivatives 16 and 17. These were evaluated, 15 were substantially inactive, indicating that the free hydroxyl
together with 5, 12 and 13, both for their DNA interaction groups are essential for the antiproliferative activity. Interest-
(studied by STD-NMR experiments and molecular docking) ingly, the peracetates 16 and 17 were active, and docking calcu-
and for their antiproliferative activity towards SW480 (colon lations showed that the interactions of the benzoxanthene core
carcinoma) and HepG2 (hepatoblastoma) cancer cells.15 The with the macromolecular counterpart are realized through the
results of this study demonstrated that benzoxanthene lignans change of hydrogen bond donors into acceptors. Rufescidride
are able to interact with DNA in a dual mode: the planar core (5), lacking flexible pendants, also proved inactive and this
acts as a base pair intercalant, whereas the flexible pendants confirmed the importance of the ester pendants for the anti-
in C-1 and C-2 act as minor groove binders.
proliferative activity of these compounds. It is also worth men-
In particular, docking calculations revealed that the tioning here that the CAPE-derived lignan 13 has recently been
hydroxy groups in C-6, C-9 and C-10 are of crucial importance reported by some of us as an inhibitor of both new vessel
for intercalation, being involved in hydrogen bonds both with growth in the angiogenesis bioassay and VEGF secretion.16
This journal is © The Royal Society of Chemistry 2014
Org. Biomol. Chem., 2014, 12, 2686–2701 | 2687

View Article Online
Paper
Organic & Biomolecular Chemistry
Compound 13 also inhibited colon adenocarcinoma (WiDr) focused our study on these pendants and report in the follow-
cells more effectively than its analog 12, showed DNA-damage ing the synthesis of new benzoxanthene lignans, the experi-
and pro-apoptotic properties and induced autophagy in WiDr mental measurement of their interaction with DNA through
cells.17 Very recently, we have also shown that compounds 12 DF-STD, the molecular docking study of these products with
and 13 react efficiently with peroxyl radicals in both polar and DNA, and the evaluation of their antiproliferative activity
nonpolar solvents.18 On the basis of these promising results, towards six tumor cell cultures, in comparison with the ben-
we decided to continue the study of ‘bio-inspired’ benzo- zoxanthenes 12 and 13, previously found active towards colon
xanthene lignans not only for their rarity in nature and conse- (SW480) and hepatic (HepG2) cancer cells.
quent poor biological evaluation, but also for their dual-mode
Synthesis
interaction with DNA. Anticancer agents that bind to DNA
commonly act either by intercalation between base pairs or by
minor groove binding; molecules able to establish a dual
mode interaction with DNA have been less frequently reported
and, in principle, should give stronger interaction with DNA
and a potential inhibitory activity towards DNA processing
enzymes;19 to cite one example, combilexins are a promising
group of DNA hybrid ligands having a minor groove binding
element combined with an intercalating planar core. These
hybrid molecules should be able to interact with DNA to a
greater extent, interfering with DNA–protein interaction and
resulting inhibitors of topoisomerase I and II.20 Thus, we
planned the synthesis of a new set of compounds bearing
different pendants in order to implement the previous SAR
data, in particular on the role of the flexible pendants and
their minor groove binding interactions.
The benzoxanthene lignans discussed here have been syn-
thesized through
a biomimetic methodology previously
reported by some of us7 and based on a Mn-mediated oxidative
coupling. A series of caffeic acid derivatives, namely the esters
10, 11 and 18–23, as well as the amides 24–26 were subjected
to the oxidative coupling reaction in the presence of Mn(OAc)3,
according to the general Scheme 1, and afforded the benzo-
xanthene lignans 12, 13 and 27–36.§
The coupling reactions of both methyl caffeate (10) and
CAPE (11) to obtain the dimers 12 and 13 were carried out in
chloroform as previously reported. Under analogous con-
ditions, ethyl caffeate (18) and butyl caffeate (19), prepared by
esterification of caffeic acid, afforded respectively the benzo-
xanthene diesters 27 and 28 as the main products; these were
purified and characterized by spectral analysis. The signals in
Bearing in mind the observed different antiproliferative
activity of lignans 5, 12 and 13, we planned to evaluate and
compare the effects of the following modifications of the C-1/
C-2 pendants: (a) acyl residues of different dimensions, namely
with a longer acyl chain or an aromatic ring, able to establish
van der Waals interactions with the minor groove; (b) the pres-
ence of polar groups with donor or acceptor properties in the
formation of hydrogen bonds; (c) isosteric substitution of the
ester with amide functions: the latter might prove to be of some
importance, because of the higher metabolic stability of the
amides, if the compounds under study were degraded by intra-
cellular esterase enzymes before reaching the biological target.
Thus, we report here the synthesis of new benzoxanthene
lignans together with the study of their interaction with DNA
through a differential frequency-saturation transfer difference
(DF-STD) protocol,21 based on STD-NMR experiments,22 and by
molecular docking calculations to study the three-dimensional
complex of ligands and DNA; these compounds, in addition to
the previously active ones 12 and 13, were evaluated for their
antiproliferative activity against six cell lines, namely HT-29,
Caco-2, HCT-116 (human colon carcinoma), H226, A549 (human
lung carcinoma), and SH-SY5Y (human neuroblastoma).
Results and discussion
Scheme 1 Synthesis of compounds 12, 13, and 27–36; the first number
Our previous work on the interaction of benzoxanthene
lignans with DNA15 established the importance of C-6, C-9 and
C-10 hydroxy groups in stabilizing the intercalation between
base pairs, and also suggested that the C-1/C-2 pendants have
refers to the monomer, and the second to dimers.
§All data concerning the reported compounds will be included in LIBIOMOL,
the chemical library of natural and synthetic bioactive molecules, accessible at
a significant role in minor groove binding. In this work we the website http://www.libiomol.unina.it
2688 | Org. Biomol. Chem., 2014, 12, 2686–2701
This journal is © The Royal Society of Chemistry 2014

View Article Online
Organic & Biomolecular Chemistry
Paper
Scheme 2 Synthesis of compounds 29–31.
their 1H and 13C NMR spectra were easily assigned by compari-
son with those of the related dimers 12 and 13. In order to
evaluate the effect of pendants with polar groups, caffeic acid
was esterified with ethylene glycol to obtain the hydroxyethyl
caffeate; the oxidative coupling of this substrate required the
addition of ethanol to the reaction mixture, and this resulted in
a lower yield of the dimer 29, in agreement with the solvent
effects reported in our previous study. We therefore turned to a
chemo/regioselective enzymatic methodology, summarized in
Scheme 2, which allowed obtaining not only the dimer 29 with
Fig. 2 Key HMBC correlations for the signal assignments of the C1/C2
higher yields, but also the diacetate 30 and the monoacetate 31, pendants in compound 31.
useful to compare the effect of different groups in the pendant
chain. Indeed, employing lipase from Candida antarctica (CaL)
the phenolic components of olive oil),23 to evaluate the effect
and vinyl acetate, we were able to initially obtain the acetoxy-
ethyl caffeate 21, and this substrate was effectively dimerized
with Mn(OAc)3 in chloroform to give 30 as the main product.
Subsequently, this dimer was submitted to an enzymatic
alcoholysis with n-butanol in the presence of CaL; the enzyme
removed preferentially the acetyl group from the C-2 chain, so
in a first step the monoacetylated benzoxanthene 31 was
obtained. By prolonging the reaction time to 72 h, the acetate
at C-1 was also removed, and a mixture of compounds 29, 30
and 31 was obtained. After prolonged extension of the reaction
time, no further conversion of the starting material was
observed. Compounds 29–31 were purified and a complete
of a p-hydroxy group on the aromatic ring, in comparison with
the dimer derived from CAPE. The tyrosil caffeate 22, obtained
in the presence of DCC and DMAP, was subjected to the coup-
ling reaction with Mn(OAc)3 in a chloroform–ethanol mixture,
and afforded 32 as the main product. As a variation of this
model, we carried out the synthesis of a similar ester employ-
ing, instead of tyrosol, the 2-(4-aminophenyl)ethanol, in order
to have pendants with a p-amino group on the aromatic ring.
The 4-aminophenethyl caffeate 23, prepared analogously to 22,
was dimerized with Mn(OAc)3 in chloroform–ethanol, and the
diester 33 was the main product.
In addition to the above cited benzoxanthene esters, we
carried out the synthesis of some benzoxanthene amides; as a
matter of fact, it is well known that amides have a higher meta-
bolic stability than esters, the latter frequently being substrates
for intracellular esterases. Consequently, amides could be
more active than the corresponding esters because of the
greater probability of reaching the biological target without
chemical modifications. Hence we prepared the primary
amides 24 and 25, isosteric respectively with the esters 19 and
22; a further amide 26 was prepared to test the effect of a more
crowded N,N-disubstituted amide on both minor groove
1
assignment of their H and 13C NMR spectra was carried out;
an unambiguous assignment was achieved, when necessary,
by analysis of two-dimensional NMR spectra, as exemplified in
Fig. 2, reporting selected HMBC correlations for the signal
assignments of the C1/C2 pendants in compound 31.
A further dimer, 32, was designed on the basis of our pre-
vious study, indicating 13 as the best DNA-interacting benzo-
xanthene lignan, and highlighting the strong binding of the
two phenylethyl groups along the DNA minor groove. Thus, we
planned a slight modification of this pendant by esterification
of caffeic acid with the natural phenolic alcohol tyrosol (one of
This journal is © The Royal Society of Chemistry 2014
Org. Biomol. Chem., 2014, 12, 2686–2701 | 2689

View Article Online
Paper
Organic & Biomolecular Chemistry
binding and antiproliferative activity. These amides were syn-
thesized from caffeic acid employing BOP/Et3N; the dimeriza-
tion reactions were carried out in ethanol and afforded
respectively the benzoxanthene diamides 34, 35 and 36 as the
main products, although with lower yields with respect to the
diesters. Also these further benzoxanthenes were subjected to
1
a complete NMR study to confirm the structure and assign H
and 13C NMR spectra.
DF-STD analysis
In order to detect the interaction of compounds 27–36 with
the DNA and to understand their binding mode, the DF-STD
method21 was carried out by using a poly(dG–dC)·poly(dG–dC)
copolymer as a macromolecule. This method is based on the
acquisition of two STD-NMR22 experiments at two different
saturation frequencies,21 and it allows distinguishing the
binding mode of DNA ligands (minor groove binders, inter-
calators, external binders). Analogously to our previous
DF-STD investigation of benzoxanthene lignans with DNA,15
two distinct binding mode indexes (BMI, Table 1) were calcu-
lated for compounds 27–36 (see the Experimental section): the
BMI, relative to the polycyclic aromatic core, and the BMI′,
referred to the chemical appendages in positions C-1 and C-2
of the benzoxanthene core. This strategy was also adopted for
the doxorubicin in our original contribution on DF-STD,21 and
was recently reported by Gomez-Monterrey et al. for the ana-
lysis of DNA-interacting spiro derivatives.24 The BMI is a
numerical parameter that, taking into account the relative
intensities of STD effects at different saturation frequencies,
gives insights on binding contacts of the ligands on the DNA
surface.21 The experimental DF-STD results showed that all
compounds 27–36 interacted with the macromolecule
(Fig. 45S–53S†), as exemplified in Fig. 3 for compound 35.
In particular, the results of the DF-STD analysis confirmed
that the polycyclic aromatic portion intercalates between the
base pairs of the nucleic acid (BMI values, Table 1), whereas
the flexible pendants at C-1 and C-2 are placed in the minor
groove (BMI′ values, Table 1). These data are in agreement
Fig. 3 DF-STD spectra of the 35–DNA complex. (a, b) STD spectra
recorded upon saturation in the aromatic (8.5 ppm) and deoxyribose/
backbone (5.0 ppm) spectral regions, respectively. (c) Reference STD
spectrum with off-resonance irradiation (−16 ppm).
with our previous outcomes obtained from the interactions of
12, 13 and other benzoxanthene lignans with the poly(dG–
dC)·poly(dG–dC) copolymer.15 In Table 1, the BMI and BMI′
values previously obtained for 12 and 13 have been added to
aid structure–activity relationship considerations.
The BMI values of 27–36 show that the planar core esta-
blishes extended van der Waals contacts with the base pair aro-
matic rings and is also involved in interactions with the DNA
backbone. Indeed, the obtained BMI values (ranging from 0.7
to 1.1) are clearly in line with data reported for unusual DNA
intercalators such as ethidium bromide (BMI = 0.9)21 which
interacts with external negatively charged phosphate by its
amino groups. Moreover, these values agree with the BMI of
0.9 obtained by us for rufescidride (5),15 the above cited
natural benzoxanthene anhydride with the same polyaromatic
portion of 27–36 but lacking flexible pendants. The DF-STD
data suggest that the hydroxyl groups in C-6, C-9 and C-10 of
the planar moiety could be involved in hydrogen bonds with
the deoxyribose/backbone of the nucleic acid. The observed
differences in BMI values for compounds sharing the same
polyaromatic core suggest that the substituents at C-1 and
C-2 have an influence on the interaction with the macromole-
cule by determining different contact distances. The BMI′
values obtained for compounds 27–36 show that the 1,2 substi-
tuents are accommodated along the minor groove of the
macromolecule. In particular, the BMI′ values of 27, 28, 30, 32,
33 and 35 (ranging from 0.5 to 0.8) highlight a comparable
saturation diffusion irradiating on the aromatic and aliphatic
DNA resonances. They are consistent not only with our pre-
vious data obtained for benzoxanthenes 12 and 13 but also
with similar values obtained for well-known minor groove
binders, such as distamycin A (BMI = 1.1) and netropsin (BMI
= 0.9).21 The BMI′ are also in agreement with the experimental
data describing the interaction of distamycin A with the minor
Table 1 Binding mode index values for compounds 12, 13 and 27–36
calculated by eqn (1)
Compound
BMIa
BMI′a
12b
13b
27
28
29
30
31
32
33
34
35
36
1.2
1.4
0.7
0.7
0.7
0.9
0.7
1.1
0.9
0.7
0.8
0.9
0.7
0.8
0.6
0.7
0.3
0.7
0.4
0.5
0.8
0.4
0.8
0.4
a Signal-to-noise ratios have been calculated integrating
4
empty
regions in the spectra, finding a maximum error of 8%. b Values for 12
and 13 are reported from our previous work.15
2690 | Org. Biomol. Chem., 2014, 12, 2686–2701
This journal is © The Royal Society of Chemistry 2014

View Article Online
Organic & Biomolecular Chemistry
Paper
groove of a DNA quadruplex (BMI = 0.8).25 The BMI′ values planar core and located the flexible appendages at different
obtained for 29, 31, 34 and 36 suggest a much closer contact points of the DNA surface (Fig. 5), in accord with the findings
with the deoxyribose/backbone than the base pairs of the bio- obtained by the experimental DF-STD data.
logical target. Taken as a whole, the DF-STD data clearly indi-
Concerning the analysis of the interactions with Model A,
cate that the different structures of the various flexible the polycyclic aromatic moiety of 27–36 presents the same
substituents at C-1, C-2 influence the overall interaction orientation between the base pairs of the DNA (Fig. 5a). In par-
contact with the DNA.
ticular, the hydroxyl group in C-10 of 27–36 donates hydrogen
bonds both to the anomeric and 5′ oxygen of C13 (Fig. 5a),
with atom distance ranges respectively of 1.76–2.30 Å and
2.32–2.75 Å. The different lengths and angles of the hydrogen
bonds are dependent on the different chemical structures of
the substituents in C-1 and C-2. These outcomes are in line
with the above reported experimental BMI values (Table 1).
Moreover, compound 30 donates a hydrogen bond by the OH
in C-9 to N-7 of G3. This interaction could potentially be made
also by the other compounds thanks to the rotation of the
hydroxyl group. The flexible pendants are located along the
minor groove of the DNA, in agreement with the experi-
mentally determined BMI′ data (Table 1), and contribute to
the complex stability by hydrogen bonds and hydrophobic
interactions. The ester oxygen in C-1 accepts a H-bond from
the NH2 group of G14 (Fig. 6a). Concerning the oxygen in C-2,
we observed the same interaction with N-7 of G3 by only 29
and 33. Compounds 12, 13, 27, 28, 30–32, 34–36 were not
involved in interaction with N-7 of G3, due to the structural
features of the flexible substituents. The chemical switch of
the ester into amide entails the conversion of the hydrogen
bond acceptor into a donor with loss of interaction with the
macromolecular counterparts. Indeed, 34–36 were not hydro-
gen bonded by their amide groups to the aromatic base pairs
along the minor groove of the DNA (Fig. 6). Behind the
Molecular docking studies
Along with STD NMR experiments, we performed molecular
docking calculations to acquire/produce a 3D model of the
complex between compounds 27–36 and the DNA, and to shed
light on the structure–activity relationships of these benzo-
xanthene lignans, in particular for the possible role of the
flexible pendants at C-1 and C-2. As previously described for
compounds 12 and 13, two binding sites were considered
(Fig. 4) for the investigation of base pair intercalation to under-
stand if the stacking into the DNA and the hydrogen bond for-
mation were base-dependent.
The analysis of docking calculations confirmed that lignans
27–36 intercalate between the base pairs of the DNA by their
Fig. 4 Schematic representation of DNA Models (A and B) used in
docking calculations. The sequence and numbering of two 10mers (A
and B) DNA duplexes are reported. The arrows indicate the intercalation
points.
Fig. 5 Superimposition in the binding site of Models A (a, b purple) and B (c, d pink) of 12, 13 and 27–36. The ligands and DNA are represented by
tube (a–d). In b and d, the macromolecule is also depicted by molecular surface. The ligands are coloured: 12, lime; 13, brown; 27, grey; 28, yellow;
29, light blue; 30, dark blue; 31, blue; 32, green; 33, dark green; 34, orange; 35, light green; 36, light grey. The OH in position 10 of 12, 13, 27–36,
and the anomeric and 5’ oxygen of C13 (a) and G14 (b) are coloured by atom type: polar H, white; O, red. The yellow line indicates hydrogen bonds.
The figure highlights the overlapping binding modes of 12, 13, 27–36 in the binding site of Models A and B.
This journal is © The Royal Society of Chemistry 2014
Org. Biomol. Chem., 2014, 12, 2686–2701 | 2691

View Article Online
Paper
Organic & Biomolecular Chemistry
Fig. 6 Superimposition in the binding site of Models A (a, purple) and B (b, pink) of 32 and 35. The ligands and DNA are represented by tube. The
ligands are coloured: 32, green; 35, light green. The ester and amide groups in positions 1 and 2, and the NH2 of G14 (a) and G5 (b) are coloured by
atom type: polar H, white; O, red; N, sky blue. Yellow lines indicate hydrogen bonds. The figure is an example of the loss of a hydrogen bond by the
amide group (35), compared to the ester functionality (32) along the minor groove of Models A (a) and B (b).
hydrogen bond formation, the substituents in C-1 and C-2 G14, instead of C13 with Model A (Fig. 5). These theoretical
positions contribute to the DNA binding by hydrophobic inter- results are in agreement with the experimental BMI values
actions thanks to their carbon chains. In particular, as pre- reported in Table 1. Moreover, in our proposed 3D model the
viously observed for compound 13, the phenylethyl groups of compounds 29, 30 and 33 donate a hydrogen bond to N-7 of
32, 33 and 35 contribute more efficiently to the complex stabi- G14 by their OH in C-9. The same interaction can potentially
lity, giving wider van der Waals contacts along the minor be made by 27–29, 31–33, 35 and 36. As found for docked
groove of the DNA, compared to those lacking in the phenethyl small molecules in Model A, compounds 29–33 and 35 accept
groups (27–31, 34 and 36).
and donate hydrogen bonds to and from DNA residues by the
Moreover, compounds 29–33 and 35 bear different chemi- chemical appendages in positions C-1 and C-2. As reported
cal groups at the end of the carbon chain, such as OH, NH2, above for Model A, the two hydroxy groups of 29 donate hydro-
COCH3, which are hydrogen bond donors and acceptors. gen bonds to N-7 of G14 and the anomeric oxygen of C6 and
Indeed, these chemical functionalities are involved in inter- C13. Compound 30 accepts a H-bond from NH2 of G3 and G5
actions with macromolecular counterparts, further contribut- by its CO of acetyl groups. The oxygen of the acetate at the end
ing to the binding towards the nucleic acid. In particular, 29 of the C-2 substituent accepts a hydrogen bond from NH2 of
by the two hydroxy groups of its appendages donates hydrogen G14. The binder 31 donates a hydrogen bond to N-7 of G14 by
bonds to N-7 of G3 and G13; 30 interacts with the same nucleo- its hydroxyl group, and accepts a H-bond by its CO of the
tides found for 29, accepting H-bonds by the CO of its acetate acetate from NH2 of G5. Compound 32 donates hydrogen
groups. Compound 31 accepts a H-bond from NH2 of G5 by its bonds to the anomeric and 3′ oxygens of C15 and G12, by the
acetate on the appendage, and donates a H-bond by its OH to hydroxyl group of the substituents in C1 and C-2, respectively.
the oxygen 3′ of G12. The binders 32 and 35 donate H-bonds to By the NH2 of the chemical arm bound to C-2, compound 33
the anomeric and 3′ oxygen of G14; these compounds, by the donates a H-bond to CO in 2 of C4 and accepts a H-bond from
other OH of their appendages, donate a hydrogen bond to the NH2 of G3. The other amine group donates a H-bond to the
oxygen in 3′ of G3. Compound 33 donates hydrogen bonds to anomeric oxygen of C14. By the OH of the C-2 appendage, the
CO in 2 and the oxygen 3′ of C6, by its NH2 of the chemical arm binder 35 donates a H-bond to CO in 2 of C4 and accepts a
bound to C-1. The other NH2 donates hydrogen bonds to the H-bond from NH2 of G3. The other hydroxyl group donates a
anomeric oxygen of G3 and to N-7 of G12.
hydrogen bond to the anomeric oxygen of G7 and accepts a
Comparable findings are obtained by docking calculations hydrogen bond from the NH2 of G16.
on Model B. Indeed, the predicted bioactive conformations are
similar, giving the same number of hydrogen bonds with com-
Lipophilicity measurements
parable distances (Fig. 5b). The common planar moiety forms
π–π interactions with base pairs of Model B (Fig. 5b), and the
two pendants (in C-1 and C-2) are located in the minor groove
establishing van der Waals contacts and hydrogen bonds
(Fig. 5b). It is noteworthy that, besides the π-stacking, the
planar portions of compounds 27–36 have the same orien-
tation between two adjacent base pairs both in Model A and B
(Fig. 5). With Model B, the OH in C-10 of 27–36 donates to
both the anomeric (1.64–2.58 Å) and 5′ oxygen (2.03–3.14 Å) of
To acquire information about the lipophilicity of the com-
pounds under study, as a first attempt we tried to run log P
measurements, employing the classical method based on
n-octanol/water partition.26 However, we realized that the
majority of compounds do not give reliable results, probably
due to their poor solubility both in n-octanol and water. We
therefore turned to an alternative method based on an
RP-HPLC measurement.27 This method has been applied in
several cases to compare the relative lipophilicity of bioactive
2692 | Org. Biomol. Chem., 2014, 12, 2686–2701
This journal is © The Royal Society of Chemistry 2014

View Article Online
Organic & Biomolecular Chemistry
Paper
molecules.28 Hence, we carried out a C-18 reverse-phase HPLC carcinoma), SH-SY5Y (neuroblastoma), by means of the MTT
analysis of the compounds to be evaluated. In Fig. 7 we report bioassay. The results are summarized in Table 2 and presented
a superimposition of their HPLC profiles, which provides an as the concentration inhibiting the 50% cell growth (GI50).
indication of their relative lipophilicity; below we list these 5-Fluorouracil (5-FU), one of the most effective anticancer
compounds in the order of decreasing lipophilicity, indicating drugs, was used as positive reference.
the pertinent capacity factor (k) value27b in brackets: 28 (4.25)
≈ 13 (4.21) > 27 (3.24) > 32 (3.01) > 12 (2.92) > 34 (2.83) ≈ 30 compounds exhibit significant antiproliferative activity against
(2.82) > 36 (2.59) > 31 (2.47) > 35 (2.39) > 33 (2.31) > 29 (2.09). at least one cell line, although highly diverse biological
These data show that, with the exception of 29, the tested
Interestingly, calculated log P values (in brackets) are in the responses were observed for the different cell lines. Interest-
order: 13 (6.65) > 28 (5.75) > 32 (5.34) > 27 (3.71) > 33 (4.28) > ingly, the CAPE-derived lignan 13 is more active than 5-FU
34 (3.66) = 35 (3.66) > 12 (2.69) > 30 (2.52) > 36 (2.32) > 31 against HCT-116 (GI50 = 3.16 µM with respect to 5.81 µM) and
(1.60) > 29 (0.69). Both data concur well in indicating a marked H226 (GI50 = 4.33 µM with respect to 16.41 µM); it is also
lipophilicity of compounds 13 and 28 (and 27 with minor highly active against A549 and SH-SY5Y, with GI50 of 11.07 and
agreement), compared to all other benzoxanthenes.
11.28 µM, respectively. Also compound 28 is more active than
5-FU against HCT-116 cells (GI50 = 4.22 µM). The GI50 values
obtained for 13 are in good agreement with those reported in
Antiproliferative activity
The benzoxanthene lignans 27–36, in addition to the pre- our previous study,15 namely 2.57 µM and 4.76 µM towards
viously reported 12 and 13, were evaluated for their antiproli- SW480 (colon carcinoma) and HepG2 (hepatoblastoma) cells,
ferative activity against six human cancer cell lines, namely respectively, and confirm the promising antiproliferative
HT-29, Caco-2, HCT-116 (colon carcinoma), H226, A549 (lung activity of this compound. It has to be highlighted that
Fig. 7 Stacked plot of RP-HPLC-UV for compounds 12, 13 and 27–36. Note: a minimal shift to the right has been applied progressively from the
lower to the higher profile.
Table 2 Antiproliferative activity of compounds 12, 13, and 27–36
GI50 (µM) SD*
Compound
HT29a
Caco-2b
HCT-116c
H226d
A549e
SH-SY5Yf
12
13
27
28
29
30
31
32
33
34
35
36
5-FU
100.00 6.28
27.38 1.45
80.20 4.61
35.41 2.43
>100
>100
>100
31.12 1.71
>100
>100
>100
>100
106.08 13.47
8.32 0.67
33.27 1.37
14.22 1.49
>100
30.98 1.45
41.11 2.19
46.21 1.91
47.25 4.12
35.27 3.19
>100
31.46 0.62
3.16 0.38
32.41 1.97
4.22 0.15
>100
63.32 2.51
>100
26.72 1.41
>100
>100
21.23 0.15
4.33 0.32
71.43 4.88
23.17 1.27
>100
>100
>100
>100
>100
73.55 4.81
>100
31.51 2.51
16.41 0.91
35.50 2.10
11.07 0.70
68.28 6.31
33.18 2.57
>100
>100
>100
>100
>100
>100
>100
23.18 2.03
11.28 0.93
3.56 0.41
35.22 3.21
>100
>100
>100
>100
>100
82.61 7.57
48.21 3.93
27.41 2.31
1.51 0.11
>100
>100
5.81 0.31
>100
1.11 0.05
38.43 3.49
9.01 0.51
8.09 0.51
*GI50 (concentration causing 50% cell growth inhibition) calculated after 72 h of continuous exposure relative to untreated controls. Values are
the means ( SD) of four experiments. Values lower than 5 µM are given in bold. a HT29: colon. b Caco-2: colon. c HCT-116: colon. d H226: lung.
e A549: lung. f SH-SY5Y neuroblastoma.
This journal is © The Royal Society of Chemistry 2014
Org. Biomol. Chem., 2014, 12, 2686–2701 | 2693

View Article Online
Paper
Organic & Biomolecular Chemistry
Fig. 8 (a, a’) Compound 13; (b, b’) compound 28; (c, c’) compound 32; (d, d’) compound 34; (e, e’) compound 35; (f, f’) no treatment. (a–f): Phase
contrast image; (a’–f’): fluorescence microscope image (excitation 345 nm; emission 455 nm). Cells were incubated with the indicated compounds
for 15 min before taking the microphotographs except for compounds 34 and 35 whose incubation lasted for 1 hour to show the lack of staining
even after prolonged treatment.
compound 27 showed high selectivity and was more active Analogously to the previously evaluated lignans 12 and 13, the
towards SH-SY5Y cells (GI50 = 3.56 µM).
compounds 27–36 now examined were obtained by a bio-
It is noteworthy that by microscopic observation some com- mimetic, Mn-mediated oxidative coupling; caffeic amides were
pounds show a net fluorescence inside the cell, clearly indicat- also used in this study in addition to esters. These com-
ing an effective uptake by the tumor cells. In particular, as pounds, bearing different flexible pendants at position C1/C2
shown in Fig. 8 with reference to the cell line HTC-116, this of the aromatic core, were subjected to a DF-STD NMR analy-
marked fluorescence was observed for the most potent and sis, as well as to molecular docking calculations; these studies
markedly lipophilic benzoxanthenes 13 and 28 (Fig. 8a′–b′), confirmed that all benzoxanthene lignans interact with DNA
whereas a weaker fluorescence was observed for less potent in a dual mode: the planar core acts as a base pair intercalant,
compounds such as 32 (Fig. 8c′) or no significant fluorescence whereas the flexible pendants act as minor groove binders; the
above the background could be observed for inactive com- reported data also showed that the structure of the pendant
pounds such as 35 (Fig. 7e′). Also interesting is the comparison may affect the interaction with DNA. In particular, the di-
between the esters 28 and 32 (Fig. 8b′–c′), showing a discern- amides 34–36 lack a hydrogen bonding interaction with the
ible fluorescence, and the structurally related, although less aromatic base pairs along the minor groove of the DNA,
lipophilic, amides 34 and 35 (Fig. 8e′–d′) where no significant observed for the diesters. A panel of six human cancer cell cul-
fluorescence could be observed.
tures, namely HT-29, Caco-2, HCT-116 (colon carcinoma),
H226, A549 (lung carcinoma), and SH-SY5Y (neuroblastoma),
was used to evaluate the antiproliferative activity of com-
pounds 27–36 with those of 12 and 13, previously showing
antiproliferative activity towards SW480 (colon carcinoma) and
HepG2 (hepatoblastoma) cancer cells. The biological data
show that almost all the compounds tested are active against
one or more cell lines, although significant differences were
observed in the antiproliferative activity: compounds 13 and
Conclusions
In this work, we have extended our previous study on a rare
group of natural and ‘unnatural’ lignans, obtained by bio-
mimetic oxidative coupling of caffeic esters and possessing a
planar and intensively fluorescent benzo[k,l]xanthene core.
2694 | Org. Biomol. Chem., 2014, 12, 2686–2701
This journal is © The Royal Society of Chemistry 2014

View Article Online
Organic & Biomolecular Chemistry
Paper
28, originated respectively by coupling of CAPE (11) and n- to be greater than 95% by HPLC. Elemental analyses were per-
butyl caffeate (19), proved more active than 5-fluorouracil formed on a Perkin-Elmer 240B microanalyzer.
towards some cell lines and are the most lipophilic according
The enzymatic reactions were incubated at 40 °C under
to both experimental and calculated log P values. The less lipo- shaking at 400 rpm; control reactions without enzyme were
philic amides 34–36, though metabolically more stable than carried out under the same conditions. The progress of the
the esters, proved only moderately active. The lignan 29, the chemo-enzymatic reactions was monitored at regular time
only one that is not active toward any of the cell lines evalu- intervals by TLC. Reactions were quenched on completion by
ated, was the least lipophilic. Microscopic observation on cell filtering off the enzyme. Candida antarctica (Chirazyme L-2,
line HT116 showed that, at least for this cell line, the most c.-f. C2) lipase was a gift from Roche.
active compounds 13 and 28 are effectively absorbed by the
Preparation of esters 10, 18–20. Concd H2SO4 (1 mL) was
cells, presumably due to their higher lipophilicity; a significant added to a solution of caffeic acid (500 mg, 2.77 mmol) in
comparison concerns the esters 28 and 32, more lipophilic MeOH, EtOH, BuOH or ethylene glycol (100 mL). The resulting
than the related amides 34 and 35: these are less effectively mixture was stirred for 6 h at reflux. After cooling at room
absorbed and less active towards HCT-116 than the esters. It is temperature, the solution was diluted with ethyl acetate
also noteworthy that compound 27, showing a selective activity (250 mL) and washed with an aqueous solution of NaHCO3;
towards SH-SY5Y cells, although less lipophilic than 13 and the organic phase was washed with saturated brine, dried with
28, was more lipophilic than all the other compounds under anhydrous Na2SO4 and the solvent evaporated under vacuum
study. The microscopic observations coupled to the lipophili- to yield 10, 18, 19 and 20:
(E)-Methyl 3-(3,4-dihydroxyphenyl)acrylate (10). Yield 99%. Rf
(TLC) = 0.52 (8% MeOH–CH2Cl2). 1H NMR data of the product
are in agreement with those previously reported in the
literature.29
city data for the compounds under study suggest that their
ability to penetrate cell membranes might be a structural
feature of greater impact than the minor groove binding pro-
perties of the C1/C2 pendants.
(E)-Ethyl 3-(3,4-dihydroxyphenyl)acrylate (18). Yield 95%. Rf
These results, while confirming that benzoxanthene
lignans are able to interact with DNA in a dual mode, reveal
that their potential optimization as antitumor agents should
take into account the structural features influencing their cel-
lular uptake. In conclusion, this scarcely explored family of
compounds appears worthy of further studies, also in view of
their intense fluorescence, their significant antiradical pro-
perties and their possible interaction with DNA quadruplex
structures, as suggested by the references cited above.
1
(TLC) = 0.66 (8% MeOH–CH2Cl2). H and 13C NMR data of 18
are in agreement with those previously reported in the
literature.30
(E)-Butyl 3-(3,4-dihydroxyphenyl)acrylate (19). Yield 92%. Rf
1
(TLC) = 0.60 (7% MeOH–CH2Cl2). H and 13C NMR data of 19
are in agreement with those previously reported in the
literature.30
(E)-2-Hydroxyethyl 3-(3,4-dihydroxyphenyl)acrylate (20). Yield
90%. Rf (TLC) = 0.44 (12% EtOH–CH2Cl2). 1H and NMR data of
20 are in perfect agreement with those previously reported in
the literature.31
Preparation of (E)-2-acetoxyethyl 3-(3,4-dihydroxyphenyl)-
acrylate (21). Chirazyme (218 mg) was added to a solution of
Experimental
General
All chemicals were of reagent grade and were used as pur- 20 (220 mg, 0.98 mmol) in EtOH (50 mL) containing vinyl
chased from Sigma-Aldrich. ESI-MS spectra were carried out in acetate (0.2 mL) and the suspension was kept at 40 °C under
negative mode on a mass spectrometer Agilent MS G1956A shaking (400 rpm). After 3 h incubation, the reaction was
using a capillary voltage of 3.5 kV. NMR spectra were run on a quenched by filtering off the enzyme and the filtrate was taken
Varian VNMR-S spectrometer operating at 499.86 (1H) and to dryness under reduced pressure. Rf (TLC) = 0.62 (12%
1
125.70 MHz (13C) and equipped with a gradient-enhanced EtOH–CHCl3). H NMR (500 MHz, CDCl3): δ = 2.1 (s, 3H), 4.35
3
3
reverse detection probe. All NMR spectrometry experiments, (m, 2H), 4.41 (m, 2H), 6.28 (d, J = 16.0, 1H), 6.88 (d, J = 8.5,
3
4
3
including two-dimensional spectra, i.e., COSY, HSQC, and 1H), 7.02 (dd, J = 8.5, J = 1.5, 1H), 7.09 (d, J = 1.5, 1H), 7.60
HMBC, were performed using software supplied by the manu- (d, 3J = 16.0, 1H). 13C NMR (125 MHz, CDCl3): δ = 20.98 (C-13),
facturers and acquired at a constant temperature (298 K). 62.26 (C-11), 62.41 (C-10), 114.40 (C-2), 114.97 (C-5), 115.49
CDCl3, CD3OD, and (CD3)2CO were used as solvents; chemical (C-8), 122.60 (C-6), 127.37 (C-1), 143.94 (C-4), 145.82 (C-7),
shifts (δ) were indirectly referred to TMS using solvent signals. 146.55 (C-3), 167.38 (C-12), 171.28 (C-9).
STD-NMR experiments were performed on a Bruker Avance
DRX 600 MHz spectrometer, equipped with a cryoprobe, at (1.45 mmol) and tyrosol (3 mmol) or 2-(4-amminophenyl)-
300 K (see below for details). ethanol (3 mmol) in DMF (3 mL) was prepared and a solution
Preparation of esters 22 and 23. A solution of caffeic acid
LiChroprep Si-60 and LiChroprep DIOL 25–40 (Merck) were of DCC and DMAP in dichloromethane (20 mL) was added; the
used as stationary phases for column chromatography. Thin reaction mixture was stirred mechanically at room tempera-
layer chromatography (TLC) was carried out on Merck 60 F254 ture; N,N-dicyclohexylurea formed was filtered off. The organic
plates using cerium sulphate and phosphomolybdic acid as phase was washed with a NaHCO3 solution and saturated
chromogenic reagents. Purity of all compounds was assessed brine, dried with anhydrous Na2SO4 and the solvent
This journal is © The Royal Society of Chemistry 2014
Org. Biomol. Chem., 2014, 12, 2686–2701 | 2695

View Article Online
Paper
Organic & Biomolecular Chemistry
1
evaporated under vacuum. The residue was purified on a silica EtOH–EtOAc). H and 13C NMR data of 23 are in perfect agree-
gel DIOL column as reported for each compound.
ment with those previously reported in the literature.33
(E)-4-Hydroxyphenethyl 3-(3,4-dihydroxyphenyl)acrylate (22).
(E)-3-(3,4-Dihydroxyphenyl)-N,N-diethylacrylamide (26). The
The residue was purified on a silica gel DIOL column with residue was purified on a silica gel DIOL column with a gradi-
80% of CHCl3–petroleum ether and subsequently with 2% of ent of EtOAc–petroleum ether (from 30 to 50%); 26 was
EtOH–CHCl3; 22 was obtained as white powder, yield 25%. Rf obtained as white powder, yield 50%. Rf (TLC) = 0.41 (10%
1
(TLC)
=
0.55 (10% EtOH–CHCl3); 1H NMR [500 MHz, EtOAc–petroleum ether). H NMR (CD3OD, 500 MHz): δ = 1.17
(CD3)2CO]: δ = 2.06 (t, J = 7.0, 2H), 2.90 (bs, 1H), 4.30 (t, J = (t, 3J = 7.0, 3H), 1.25 (t, 3J = 7.0, 3H), 3.46 (m, 2H), 3.52 (m,
3
3
3
3
3
3
7.0, 2H), 6.27 (d, J = 16.0, 1H), 6.87 (d, J = 8.0, 1H), 6.79 (d, 2H), 4.88 (s, OH), 6.76 (d, J = 15.3, 1H, H-7), 6.78 (d, J = 8.5,
3J = 8.5, 2H), 7.05 (dd, J = 8.0, J = 2.0, 1H), 7.13 (d, J = 8.5, 1H, H-5), 6.96 (dd, J = 8.5, J = 2.0, 1H, H-6), 7.05 (d, J = 2.0,
3
4
3
3
4
3
2H), 7.16 (d, J = 2.0, 1H), 7.54 (d, J = 16.0, 1H), 8.30 (bs, 2H). 1H, H-2), 7.47 (d, 3J = 15.3, 1H, H-8). H NMR data of 26 are in
13C NMR [125 MHz, (CD3)2CO]: δ = 34.1 (C-7′), 64.8 (C-8′), agreement with those previously reported in the literature.34
114.4 (C-2), 114.8 (C-8), 115.0 (C-3′ and C-5′), 115.2 (C-5), 121.6 13C NMR (CD3OD, 125 MHz): δ = 12.1 (C-13), 13.8 (C-11), 41.0
(C-6), 126.8 (C-1); 128.9 (C-1′), 129.9 (C-2′ and C-6′), 144.8 (C-7), (C-12), 42.2 (C-10), 113.7 (C-2), 113.9 (C-8), 115.1 (C-5), 120.8
3
3
1
145.4 (C-3), 147.8 (C-4), 156.0 (C-4′), 166.5 (C-9).
(C-6), 127.2 (C-1), 143.0 (C-7), 145.3 (C-4), 147.5 (C-3), 167.2
(C-9).
Synthesis of compounds 12 and 13. Dimethyl 6,9,10-trihy-
droxybenzo[k,l]xanthene-1,2-dicarboxylate (12) and bis(2-phe-
nylethyl)-6,9,10-trihydroxybenzo[k,l]xanthene-1,2-dicarboxylate
(13) were synthesized as previously described.7
(E)-3-(3,4-Dihydroxyphenyl)-N-(4-hydroxyphenethyl)acrylamide
(23). The residue was purified on a silica gel DIOL column
with a gradient of EtOH–CHCl3 (from 0 to 100%); 23 was
obtained as yellow powder, yield 75%. Rf (TLC) = 0.27 (12%
EtOH–CHCl3). 1H NMR (500 MHz, CD3OD): δ = 2.80 (t, 3J = 7.0,
2H), 3.74 (t, 3J = 7.0, 2H), 7.06 (d, 3J = 2.0, 1H), 6.55 (d, 3J =
Synthesis of compounds 27–36. To a solution of caffeic
ester or caffeic amides (0.70 mmol) dissolved in CHCl3
(50 mL), CHCl3–EtOH (9–1) or in EtOH was added Mn(OAc)3
(substrate : oxidative agent ratio 1 : 4). The reaction mixtures
were stirred for 3–8 h at room temperature; the progress of the
reactions was monitored at regular time intervals by TLC; the
solvent was evaporated in vacuo and the residue was treated
with a saturated solution of ascorbic acid in H2O; subsequently
the mixture was extracted twice with EtOAc; the organic phase,
after washing with a solution of NaHCO3 and saturated brine,
was dried over anhydrous Na2SO4 and the solvent evaporated
under vacuum. The residue was purified on a chromatographic
column as reported for each compound. In the following, the
assignments of NMR signals for the side chains are referred to
the numbering system reported in ESI.†
3
3
4
16.0, 1H), 6.79 (d, J = 8.0, 1H), 6.95 (dd, J = 8.0, J = 2.0, 1H),
3
3
3
7.20 (d, J = 8.5, 2H), 7.53 (d, J = 16.0, 1H), 7.57 (d, J = 8.5,
2H). 13C NMR (125 MHz, CD3OD): δ = 38.3 (C-7′), 62.8 (C-8′),
113.8 (C-2), 115.1 (C-5), 117.4 (C-8), 119.9 (C-3′ and C-5′), 120.9
(C-6), 126.9 (C-1), 128.9 (C-2′ and C-6′), 134.9 (C-1′), 136.9
(C-4′), 141.8 (C-7), 145.4 (C-3), 147.6 (C-4), 166.0 (C-9).
Preparation of amides 24–26. The caffeic acid (300 mg,
1.62 mmol) is dissolved in 8 mL of DMF and 0.23 mL of tri-
ethylamine (1.62 mmol). The solution is cooled in an ice-water
bath and 1.62 mmol of the amine are added followed by a
solution of BOP (733 mg, 1.62 mmol) in 10 mL of CH2Cl2. The
mixture is stirred at 0 °C for 30 min and then at room tempera-
ture for 24 h. CH2Cl2 is removed under reduced pressure and
the solution is diluted with 40 mL of water. The mixture is
extracted with ethyl acetate. The extract is washed successively
Diethyl 6,9,10-trihydroxybenzo[k,l]xanthene-1,2-dicarboxylate
with 1 M HCl, 1 M NaHCO3, and water, dried over Na2SO4, fil- (27). The residue was purified on a silica gel DIOL column
tered and evaporated. The residue is purified on a silica gel with a gradient of CH2Cl2–MeOH (from 0 to 2%); 27 was
DIOL column as reported for each compound.
(E)-N-Butyl-3-(3,4-dihydroxyphenyl)acrylamide
obtained as yellow powder, yield 47%. Rf (TLC) = 0.55 (6%
MeOH–CH2Cl2). 1H NMR [500 MHz, (CD3)2CO]: δ = 1.382 (t,
(24). The
3
3
3J = 7.0, 3H, H-2′″), 1.387 (t, J = 7.0, 3H, H-2″), 4.35 (m, J =
residue was purified on a silica gel DIOL column with an
eluent 60% ethyl acetate–petroleum ether; 24 was obtained as
white powder, yield 58%. Rf (TLC) = 0.50 (8% MeOH–CH2Cl2).
3
7.0, 2H, H-1′″), 4.37 (m, J = 7.0, 2H, H-1″), 6.72 (s, 1H, H-8),
3
3
7.26 (s, 1H, H-11), 7.34 (d, J = 8.5, 1H, H-5), 7.52 (d, J = 8.5,
1H, H-4), 8.15 (s, 1H, H-3) ppm. 13C NMR [125 MHz,
(CD3)2CO]: δ = 13.2 (C-2″), 13.6 (C-2′″), 60.9 (C-1″), 61.2 (C-1′″),
103.8 (C-8), 110.0 (C-11a), 111.5 (C-11), 119.6 (C-5), 121.1 (C-4),
121.2 (C-1), 123.3 (C-11c), 124.1 (C-2), 125.5 (C-11b), 126.7
(C-3a), 128.5 (C-3), 141.7 (C-10), 141.9 (C-6), 146.9 (C-9), 148.2
(C-7a), 165.9 (CO-2′), 169.9 (CO-1′) ppm. ESI-MS m/z 408.8
[M − H]−; elemental analysis C, 64.39 and H, 4.42 calculated
for C22H18O8. Found C, 64.72 and H, 4.38.
3
1H NMR (CD3OD, 500 MHz): δ = 0.94 (t, J = 7.5, 3H), 1.38 (m,
3
2H), 1.53 (m, 2H), 3.28 (t, J = 7.0, 2H), 4.89 (s, OH and NH),
6.38 (d, 3J = 15.3, 1H, H-7), 6.76 (d, 3J = 8.5, 1H, H-5), 6.96 (dd,
4
3
3J = 8.0, J = 1.5, 1H, H-6), 7.01 (d, J = 2.0, 1H, H-2), 7.40 (d,
3J = 15.3, 1H, H-8). H NMR data of 24 are in agreement with
1
those previously reported in the literature.32 13C NMR (CD3OD,
125 MHz) δ = 12.7 (C-14), 19.8 (C-13), 31.2 (C-12), 38.9 (C-11),
113.7 (C-2), 115.1 (C-5), 117.1 (C-8), 120.7 (C-6), 127.0 (C-1),
140.7 (C-7), 145.3 (C-4), 147.3 (C-3), 169.8 (C-9).
Dibutyl 6,9,10-trihydroxybenzo[k,l]xanthene-1,2-dicarboxylate
(E)-3-(3,4-Dihydroxyphenyl)-N-(4-hydroxyphenethyl)acrylamide (28). The residue was purified on a silica gel DIOL column
(25). The residue was purified on a silica gel DIOL column
with a gradient of MeOH–CH2Cl2 (from 0 to 10%); 25 was
obtained as white powder, yield 74%. Rf (TLC) = 0.76 (10%
with a gradient of MeOH–CH2Cl2 (from 0 to 3%); 28 was
obtained as yellow powder, yield 40%. Rf (TLC) = 0.5 (8%
MeOH–CH2Cl2). 1H NMR [500 MHz, (CD3)2CO]: δ = 0.94 (t, 3J =
2696 | Org. Biomol. Chem., 2014, 12, 2686–2701
This journal is © The Royal Society of Chemistry 2014

View Article Online
Organic & Biomolecular Chemistry
Paper
3
3
1
7.0, 3H, H-4″), 0.99 (t, J = 7, 3H, H-4′″), 1.44 (m, J = 7.5, 2H, the unconverted substrate (9.9 mg, 45%). H NMR [500 MHz,
3
3
H-3″), 1.49 (m, J = 7.5, 2H, H-3′″), 1.74 (m, J = 7.5, 4H, H-2″ (CD3)2CO] of 31: δ = 1.98 (s, 3H, H-4′″), 3.90 (bt, 2H, H-2″), 4.39
and H-2′″), 4.40 (t, 3J = 6.5, 2H, H-1′″), 4.30 (t, 3J = 6.5, 2H, (bt, 2H, H-1″), 4.43 (bt, 2H, H-2′″), 4.64 (bt, 2H, H-1′″), 6.73 (s,
H-1″), 6.71 (s, 1H, H-8), 7.25 (s, 1H, H-11), 7.31 (d, 3J = 9.0, 1H, 1H, H-8), 7.25 (s, 1H, H-11), 7.34 (d, J = 8.5, 1H, H-5), 7.52 (d,
3
H-5), 7.45 (d, 3J = 9.0, 1H, H-4), 8.16 (s, 1H, H-3) ppm. 13C NMR 3J = 8.5, 1H, H-4), 8.25 (s, 1H, H-3). 13C NMR [125 MHz,
[125 MHz, (CD3)2CO]: δ = 13.1 (C-4″), 13.1 (C-4′″), 19.0 (C-3″ and (CD3)2CO]: δ = 20.4 (C-4′″), 60.6 (C-2″), 62.6 (C-2′″), 64.3 (C-1′″),
C-3′″), 30.1 (C-2″), 30.6 (C-2′″), 64.7 (C-1″), 65.2 (C-1′″), 103.8 67.7 (C-1″), 104.6 (C-8), 110.7 (C-11a), 112.3 (C-11), 120.5 (C-5),
(C-8), 110.1 (C-11a), 111.5 (C-11), 119.7 (C-5), 121.2 (C-4), 121.3 121.9 (C-4), 122.1 (C-1), 124.0 (C-11c), 125.0 (C-11b), 125.8
(C-1), 123.3 (C-11c), 124.1 (C-2), 125.5 (C-11b), 126.7 (C-3a), (C-2), 127.5 (C-3a), 142.5 (C-10), 147.6 (C-9), 149.0 (C-7a), 166.6
128.5 (C-3), 136.7 (C-6a), 141.71 (C-10), 141.78 (C-6), 146.9 (C-9), (C-2′), 170.7 (C-1′), 170.8 (C-3′″). Selected HMBC correlations: C
148.1 (C-7a), 165.9 (CO-2′), 170.0 (CO-1′) ppm. ESI-MS m/z 464.9 (H), 166.6 (8.25; 4.39), 170.8 (4.43; 1.98), 170.7 (4.64), 60.6
[M − H]−, elemental analysis C, 66.94 and H, 5.62 calculated for (4.39). ESI-MS m/z 482.8 [M − H]−, elemental analysis C, 59.11
C26H26O8. Found C, 66.85 and H, 5.60.
and H, 4.16 calculated for C24H20O11. Found C, 59.48 and
Bis(2-hydroxyethyl)-6,9,10-trihydroxybenzo[k,l]xanthene-1,2-
dicarboxylate (29). The residue was purified on a silica gel
H, 4.17.
Bis(4-hydroxyphenethyl)6,9,10-trihydroxybenzo[k,l]xanthene-1,2-
DIOL column with a gradient of EtOH–CHCl3 (from 3 to 5%); dicarboxylate (32). The residue was purified on a silica gel
for direct dimerization of 20, 29 was obtained as yellow DIOL column with a gradient of EtOH–CH2Cl2 (from 0 to 7%);
powder, yield 12.8%; Rf (TLC) = 0.32 (8% EtOH–CHCl3); 1H 32 was obtained as yellow powder, yield 55%. Rf (TLC) = 0.47
NMR [500 MHz, (CD3)2CO] of 29: δ = 3.88 (bt, 2H, H-2′″), 3.97 (12% EtOH–CHCl3).
3
(bt, 2H, H-2″), 4.36 (bt, 2H, H-1′″), 4.54 (bt, 4H, H-1″), 6.69 (s,
1H NMR (500 MHz, CD3OD): δ = 2.93 (t, J = 8.0, 2H, H-2′″),
3
1H, H-8), 7.33 (d, J = 8.5, 1H, H-5), 7.42 (s, 1H, H-11), 7.52 (d, 2.97 (t, 3J = 7.0 2H, H-2″), 4.43 (t, 3J = 7.0, 2H, H-1″), 4.56 (t, 3J =
3J = 8.5, 1H, H-4), 8.25 (s, 1H, H-3). 13C NMR [125 MHz, 8.0, 2H, H-1′″), 6.69 (d, J = 8.5, 2H, H-3IV and H-5IV), 6.71 (s,
3
(CD3)2CO]: δ = 59.9 (C-2′″), 60.1 (C-2″), 67.0 (C-1′″), 67.3 (C-1″), 1H, H-8), 6.75 (d, 3J = 8.0, 2H, H-3V and H-5V), 7.10 (m, 5H,
104.6 (C-8), 110.8 (C-11a), 112.3 (C-11), 120.5 (C-5), 121.9 (C-1), H-11, H-2IV, H-6IV, H-2V and H-6V), 7.21 (d, J = 8.5, 1H, H-5),
3
122.0 (C-4), 124.1 (C-11c), 124.7 (C-11b), 125.8 (C-2), 127.4 7.32 (d, 3J = 8.5, 1H, H-4), 7.94 (s, 1H, H-3). 13C NMR
(C-3a), 129.7 (C-3), 137.4 (C-6a); 142.4 (C-10), 142.7 (C-6), 147.6 (125 MHz, CD3OD): δ = 33.3 (C-2′″), 33.8 (C-2″), 66.0 (C-1″),
(C-9), 148.9 (C-7a), 166.8 (C-2′), 170.9 (C-1′). ESI-MS m/z 440.8 66.8 (C-1′″), 103.5 (C-8), 109.6 (C-11a), 111.2 (C-11), 114.9
[M − H]−, elemental analysis C, 59.73 and H, 4.10 calculated (C-3IV and C-5IV), 115.0 (C-3V and C-5V), 118.5 (C-1), 119.4
for C22H18O8. Found C, 59.79 and H, 4.12.
(C-5), 120.2 (C-2), 120.9 (C-4), 123.3 (C-11c), 124.5 (C-1IV), 124.7
(C-11b and C-1V), 126.6 (C-3a), 128.6 (C-3), 129.6 (C-2IV and
C-6IV), 129.7 (C-2V and C-6V), 136.9 (C-6a), 141.7 (C-6), 141.8
(C-7a), 147.0 (C-10), 148.3 (C-9), 155.5 (C-4IV), 155.7 (C-4V),
166.4 (C-2′), 171.6 (C-1′). ESI-MS: m/z 592.8 [M − H]−, elemen-
tal analysis C, 68.68 and H, 4.41 calculated for C34H26O10.
Found C, 68.61 and H, 4.39.
Bis(2-acetoxyethyl)6,9,10-trihydroxybenzo[k,l]xanthene-1,2-dicar-
boxylate (30). The residue was purified on a silica gel DIOL
column with a gradient of EtOH–CH2Cl2 (from 0 to 3%); 30
was obtained as yellow-green powder, yield 45%. Rf (TLC) =
0.55 (10% EtOH–CHCl3). 1H NMR [500 MHz, (CD3)2CO]: δ =
1.98 (s, 3H, H-4″), 2.07 (s, 3H, H-4′″), 4.43 (bt, 4H, H-2″ and
H-2′″), 4.54 (bt, 2H, H-1″), 4.64 (bt, 2H, H-1′″), 6.73 (s, 1H,
Bis(4-aminophenethyl)6,9,10-trihydroxybenzo[k,l]xanthene-1,2-
3
3
H-8), 7.25 (s, 1H, H-11), 7.35 (d, J = 8.5, 1H, H-5), 7.53 (d, J = dicarboxylate (33). The residue was purified on a silica gel
8.5, 1H, H-4), 8.21 (s, 1H, H-3). 13C NMR [125 MHz, (CD3)2CO]: DIOL column with a gradient of EtOH–CHCl3 (from 5 to 13%);
δ = 20.4 (C-4″), 20.5 (C-4′″), 62.2 (C-2″), 62.5 (C-2′″), 63.8 (C-1″), 33 was obtained as yellow powder, yield 25%. Rf (TLC) = 0.44
1
64.3 (C-1′″), 104.5 (C-8), 110.6 (C-11a), 112.0 (C-11), 120.5 (C-5), (20% EtOH–CH2Cl2). H NMR (500 MHz, CD3OD): δ = 2.79 (m,
121.2 (C-1), 122.1 (C-4), 124.0 (C-11c), 125.0 (C-11b), 125.4 4H, H-2″ and H-2′″), 3.74 (t, J = 7.0, 4H, H-1″ and H-1′″), 6.61
3
(C-2), 127.4 (C-3a), 129.7 (C-3), 137.4 (C-6a), 142.4 (C-10), 142.7 (s, 1H, H-8), 7.13 (m, 6H, H-4, H-5, H-2IV, H-6IV, H-2V and
(C-6), 147.6 (C-9), 148.9 (C-7a), 166.3 (C-2′), 170.4 (C-1′), 170.8 H-6V), 7.30 (s, 1H, H-11), 7.56 (m, 5H, H-3, H-3IV, H-5IV, H-3V
(C-3″ and C-3′″). ESI-MS m/z 524.8 [M − H]−, elemental analysis and H-5V). 13C NMR (125 MHz, CD3OD): δ = 39.70 (C-2′″)*1,
C, 59.32 and H, 4.21 calculated for C26H22O12. Found C, 59.41 39.73 (C-2″)*1, 64.2 (C-1″ and C-1′″), 104.7 (C-8), 111.6 (C-11a),
and H, 4.17.
113.8 (C-11), 119.9 (C-11b), 120.2 (C-5), 121.3 (C-4), 122.0 (C-3IV
and C-5IV)*4, 122.6 (C-3V and C-5V)*4, 123.8 (C-1), 124.1 (C-11c),
125.8 (C-3), 128.2 (C-3a), 130.0 (C-2V and C-6V)*3, 130.1 (C-2IV
and C-6IV)*3, 133.4 (C-2), 136.3 (C-1IV)*2, 136.7 (C-1V)*2, 137.8
(C-6a), 138.1 (C-4IV)*5, 138.2 (C-4V)*5, 141.7 (C-6), 142.9 (C-7a),
148.4 (C-10), 149.4 (C-9), 169.7 (C-2′), 171.3 (C-1′). *n: inter-
changeable with n. ESI-MS: m/z 590.7 [M − H]−, elemental ana-
lysis C, 68.91, H, 4.76, and N, 4.73 calculated for C34H28N2O8.
Found C, 68.87, H, 4.72 and N, 4.75.
2-(2-Acetoxyethyl)1-(2-hydroxyethyl)6,9,10-trihydroxybenzo[k,l]-
xanthene-1,2-dicarboxylate (31). The substrate 30 (22 mg,
0.04 mmol) was dissolved in EtOAc (12 mL) and n-BuOH
(1 mL). Candida antarctica lipase (24 mg) was added to this
solution and the mixture was maintained under shaking at
40 °C for 72 h. The reaction was quenched by filtering off the
enzyme and the filtrate was evaporated in vacuo. The residue
was purified on a silica gel DIOL column with a gradient of
EtOH–CHCl3 (from 0 to 3%) to obtain the deacylated ana-
N,N-Dibutyl-6,9,10-trihydroxybenzo[k,l]xanthene-1,2-dicarbox-
logues 31 (8.5 mg, 70% yield) and 29 (3.6 mg, 30% yield) and amide (34). The residue was purified on a silica gel DIOL
This journal is © The Royal Society of Chemistry 2014
Org. Biomol. Chem., 2014, 12, 2686–2701 | 2697

View Article Online
Paper
Organic & Biomolecular Chemistry
column with a gradient of MeOH–CH2Cl2 (from 0 to 3%); 34 Measurements of the capacity factor (K)
was obtained as yellow-green powder, yield 15%. Rf (TLC) =
The measurements of the capacity factor were performed
0.53 (8% MeOH–CH2Cl2). 1H NMR [500 MHz, (CD3)2CO]: δ =
0.94 (m, 6H, H-4″ and H-4′″), 1.46 (m, 4H, H-3″ and H-3′″),
using a Luna-C18 column (5 µm; 4.6 × 250 mm; Phenomenex)
using as an eluent a gradient of H2O–CH3CN; the elution gra-
dient had the following profile: t0 min B (5%), t10 min B (100%),
3
1.62 (m, 4H, H-2″ and H-2′″), 3.43 (q, J = 6.0, 2H, H-1″), 3.74
(m, 2H, H-1′″), 6.62 (s, 1H, H-8), 7.20 (d, 3J = 8.0, 1H, H-5), 7.30
(d, 3J = 8.0, 1H, H-4), 7.38 (t, 3J = 6.0, 1H, NH), 7.51 (s, 1H,
H-11), 7.61 (s, 1H, H-3), 7.65 (t, 6.0, 3J = 1H, NH). 13C NMR
[125 MHz, (CD3)2CO]: δ = 14.1 (C-4″ and C-4′″), 20.7 (C-3′″),
20.9 (C-3″), 31.5 (C-2′″), 32.4 (C-2″), 40.1 (C-1′″), 40.5 (C-1″),
104.3 (C-8), 111.8 (C-11a), 113.2 (C-11), 119.0 (C-5), 120.7 (C-4),
123.2 (C-11c), 124.1 (C-11b), 124.7 (C-1), 125.6 (C-3),
127.0 (C-3a), 133.0 (C-2), 137.5 (C-6a), 140.0 (C-6), 142.3 (C-7a),
147.5 (C-10), 148.5 (C-9), 168.7 (C-2′), 171.4 (C-1′). ESI-MS:
m/z 463.3 [M − H]−, elemental analysis C, 67.23, H, 6.08 and
N, 6.03 calculated for C26H28N2O6. Found C, 67.12, H, 6.02
and N, 6.09.
t15
B (100%). The flow rate was 1 mL min−1. All experi-
min
ments were performed using an Agilent 1100 pump, a photo-
diode array detector Agilent Series 1200 G1315C/D; an
autosampler Series 1100 G1313A (Agilent) was used for sample
injections. The capacity factor K was calculated using the
following expression:
K ¼ ðtR ꢀ toÞ=to
where tR is the retention time of the tested substance, and to is
the dead-time determined using ascorbic acid as an analyte.
Calculated log P values were obtained with ACD/labs log P
program version 11.
6,9,10-Trihydroxy-N,N-bis(4-hydroxyphenethyl)benzo[k,l]-
xanthene-1,2-dicarboxamide (35). The residue was purified on a
silica gel DIOL column with a gradient of MeOH–CH2Cl2
(from 0 to 4%); 35 was obtained as yellow-green powder, yield
DF-STD NMR
As the compounds under study were not water soluble, pre-
liminary solubility tests were carried out on compounds 27–36
before the NMR measurements; as a result, different DMSO–
water mixtures containing buffered solution (10 mM of NaCl,
10 mM of NaH2PO4) at pH 7.4 were used to solve the mole-
cules. For 27 and 30–36 the final concentration was 1 mM and
for the DNA, 50 μM (expressed as molarity of phosphate
groups), in 200 μl of the following solvent: [D6]DMSO–D2O
50/150 for 32, 33 and 35 and 10/190 for 27, 30, 31, 34 and 36.
The final concentration of 28 and 29 was 0.5 mM vs. 25 μM of
DNA, by using the solvent mixture [D6]DMSO–D2O in the ratio
10/190; for 12 and 13 see literature data.15
The DNA used as a biological target was a poly(dG–dC)·poly-
(dG–dC) copolymer, presenting an average of 750 base
pairs. The copolymer was dissolved in the above described
buffer aqueous solution and underwent annealing for 5 min at
80 °C. The copolymer was dried and then dissolved in the
deuterated buffer with the percentage of DMSO as described
1
20%. Rf (TLC) = 0.21 (17% MeOH–CHCl3). H NMR (500 MHz,
3
3
CD3OD): δ = 2.82 (t, J = 7.5, 4H, H-2″ and H-2′″), 3.55 (t, J =
7.5, 4H, H-1″ and H-1′″), 6.69 (s, 1H, H-8), 6.70 (d, 3J = 8.5, 2H,
H-3IV and H-5IV), 6.75 (d, J = 8.5, 2H, H-3V and H-5V), 7.13 (d,
3
3J = 8.5, 2H, H-2V and H-6V), 7.19 (d, 3J = 8.5, 1H, H-5), 7.26 (d,
3J = 8.5, 1H, H-4), 7.36 (s, 1H, H-11), 7.42 (s, 1H, H-3). 13C
NMR (125 MHz, CD3OD): δ = 33.2 (C-2′″), 33.5 (C-2″), 41.4
(C-1″), 41.8 (C-1′″), 103.3 (C-8), 110.2 (C-11), 112.2 (C-11a),
114.9 (C-3IV, C-5IV, C-3V and C-5V), 119.0 (C-11c), 119.7 (C-5),
124.2 (C-3), 124.3 (C-4), 122.6 (C-1), 127.0 (C-3a), 129.3 (C-2IV,
C-6IV, C-2V and C-6V), 129.5 (C-1IV and C-1V), 129.9 (C-2), 131.9
(C-11b), 136.9 (C-6a), 140.3 (C-10), 141.6 (C-6), 147.0 (C-9),
148.0 (C-7a), 155.4 (C-4IV), 155.5 (C-4V), 170.2 (C-1′), 171.4
(C-2′). ESI-MS: m/z 591.3 [M − H]−, elemental analysis C, 68.91,
H, 4.76, and N, 4.73 calculated for C34H28N2O8. Found C,
68.75, H, 4.72 and N, 4.78.
N,N,N,N-Tetraethyl-6,9,10-trihydroxybenzo[k,l]xanthene-1,2-
dicarboxamide (36). The residue was purified on a silica gel above.
DIOL column with a gradient of MeOH–CH2Cl2 (from 0 to
Two STD NMR spectra for each ligand were recorded irra-
2%); 36 was obtained as yellow-brown powder, yield 35%. Rf diating on the aromatic and sugar proton resonances (Table 3)
(TLC) = 0.44 (50% EtOAc–petroleum ether). 1H NMR [500 MHz, of DNA by Gaussian train pulses reaching a total saturation
3
3
(CD3)2CO]: δ = 0.92 (t, J = 7.0, 3H, H-4′″), 1.13 (t, J = 7.0, 3H, time of 4 s.
3
3
H-2″), 1.19 (t, J = 7.0, 3H, H-4″), 1.23 (t, J = 7.0, 3H, H-2′″),
For each DF-STD spectrum (saturation time = 4 s) 288 scans
3.01 (bs, 1H, -OH), 3.19 (m, 2H, H-3′″), 3.40 (m, 2H, H-1″), 3.58 were recorded, whereas 144 scans were recorded for the
1
(m, 2H, H-3″), 3.74 (m, 2H, H-1′″), 6.67 (s, 1H, H-8), 7.24 (d, reference STD spectrum. The water H signal was suppressed
3
3J = 9.0, 1H, H-5), 7.34 (d, J = 9.0, 1H, H-4), 7.37 (s, 1H, H-3), using a 3–9–1935 pulse sequence. The STD effects of the indi-
7.44 (s, 1H, H-11), 8.41 (bs, 2H, -OH). 13C NMR [125 MHz, vidual protons were calculated for each compound relative to a
(CD3)2CO]: δ = 12.1 (C-2′″), 12.8 (C-4″), 13.2 (C-4′″), 13.9 (C-2″), reference spectrum with off-resonance saturation at δ =
39.2 (C-1″), 39.3 (C-1′″), 43.0 (C-3″), 43.8 (C-3′″), 104.2 (C-8), −16 ppm.
111.6 (C-11a), 113.0 (C-11), 119.8 (C-5), 120.3 (C-4), 122.2
In our original contribution, based on the analysis of
(C-11c), 122.4 (C-3), 123.1 (C-2 and C-11b), 127.9 (C-3a), 133.7 ligands with a well-known binding mode to DNA, we defined
(C-1), 137.4 (C-6a), 142.3 (C-7a), 140.6 (C-6), 147.3 (C-10), 148.4 three BMI ranges: 0 < BMI < 0.50 for external (nonspecific),
(C-9), 169.8 (C-2′), 170.2 (C-1′). ESI-MS: m/z 463.3 [M − H]−, electrostatic backbone binding; 0.90 < BMI < 1.10 for minor
elemental analysis C, 67.23, N, 6.08 and H, 6.03 calculated for groove binding; and 1.20 (0.90) < BMI < 1.50 for base-pair
C26H28N2O6, found C, 66.98, H, 6.05 and N, 6.00.
intercalation.21
2698 | Org. Biomol. Chem., 2014, 12, 2686–2701
This journal is © The Royal Society of Chemistry 2014

View Article Online
Organic & Biomolecular Chemistry
Paper
All ligand structures were built and their geometries
optimized through the MacroModel 8.5 software36 package
and using the MMFFs force field.39 The MonteCarlo
Multiple Minimum (MCMM) method (10 000 steps) of the
MacroModel package36 was used in order to allow a full
exploration of the conformational space. The so-obtained geo-
metries were optimised using the Polak–Ribier Conjugate
Gradient algorithm (PRCG, maximum derivative less than
0.001 kcal mol−1).
Table 3 Irradiation resonances used in the STD NMR experiments for
12, 13 and 27–36
Saturation frequency (ppm)
Compound
Aromatic
Aliphatic
12a
13a
27
28
29
30
31
32
33
34
35
36
9.0
9.2
8.8
9.2
9.0
9.0
8.7
8.8
8.5
9.0
8.5
8.8
2.0
1.3
5.4
5.4
2.5
3.2
5.8
5.4
5.0
5.1
5.0
5.4
Autodock 4.240 was used for all docking calculations. The
software rapidly takes into account the ligand–macromolecule
interactions by precalculating atomic affinity potentials (grid
maps) for each atom type in the substrate molecule by the grid
method. These maps are calculated by AutoGrid, where the
protein is embedded in a three-dimensional grid and a probe
atom is placed at each grid point. The energy of interaction of
this single atom with the protein is assigned to the grid point.
For all the docking calculations with Models A and B, a grid
box size of 54 × 54 × 52 with a spacing of 0.375 Å between the
grid points was used and centred on the following x, y and z
coordinates: −1.189; 2.505; 23.5 (Model A), and −1.189; 2.505;
20.2 (Model B). The above described grid boxes included the
external deoxyribose/backbone.
a Values for 12 and 13 are reported from our previous work.15
The BMI were calculated using the following equation:21
ꢀ
ꢁ
P
SNaromatic=SNrif
SNaliphatic=SNrif
i
BMI ¼
ð1Þ
ni
In order to achieve a representative conformational space
during the docking calculations on ligands under investi-
gation, ten calculations consisting of 256 runs were performed,
obtaining 2560 ligand conformations (256 × 10). The Lamarck-
ian genetic algorithm was used for dockings. An initial popu-
lation of 700 randomly placed individuals, a maximum
number of 6 × 106 energy evaluations, and a maximum
number of 7 × 106 generations were taken into account. A
mutation rate of 0.02, a crossover rate of 0.8 and a local search
where SNaromatic and SNaliphatic are the signal to noise ratios of
the ligand peaks obtained from the on-resonance STD spectra
irradiating in the aromatic and aliphatic regions, respectively.
SNrif is instead the intensity of the same signal in the off-
resonance spectrum (saturation at δ = −16 ppm), and ni is the
number of ligand signals.
Computational methods
As described in our previous paper,15 Models A and B were frequency of 0.26 were used. All the 3D models were depicted
built using the graphical interface Maestro version 6.0, Schrö- using the Phyton software.
dinger, LLC, New York, NY, 2003. A B-DNA decamer (dG–
dC)·(dG–dC) was built, in order to have a complete helix turn,
using the available base pair models of Maestro (see ESI† for
Human cell cultures
the three-dimensional coordinates of the models). To study The lung carcinoma cell line A549 (ATCC number: CCL-185),
the intercalation between base pairs, the binding site was built the lung squamous carcinoma cell line NCI-H226 (ATCC
taking into account the distances from experimental structures number: CRL-5826), the colorectal adenocarcinoma Caco-2
of complexes between nucleic acids and intercalators and the (ATCC number: HTB-37) and the HT29 (ATCC number:
induced local unwinding. Two intercalation points were built HTB-38) were obtained from the American Type Culture Collec-
(Models A and B), and considered in the calculations. In tion (ATCC, Teddington, UK). The SH-SY5Y (ATCC CRL-2266)
detail, the binding site of Model A was formed by the G3·C13 was kindly provided by Prof. Maria Angela Sortino (University
and C4·G14 base pairs, whereas for Model B the intercalation of Catania, Italy). The A549, NCI-H226 and Caco-2 cell lines
point was delimited by C4·G14 and G5·C15 base pairs. Both were grown in RPMI 1640 medium (cat. 61870-010, GIBCO by
models were optimized using Macromodel 8.5 software36 by Life Technologies, Italy). The HT-29 cells were grown in DMEM
the Polak–Ribiere Conjugate Gradient method using the Force with glutamax (4.5 g L−1) (cat. no. 31965-023, GIBCO by Life
Field Amber37 (threshold 0.005 kJ mol−1 Å−1). The GB/SA (Gen- Technologies, Italy). The SH-SY5Y cells were grown in
eralized Born/surface area)38 solvent treatment was used, DMEM-F12 (cat. no. 21331-046, GIBCO by Life Technologies,
mimicking the presence of H2O, in the calculations for redu- Italy). All media were supplemented with 10% (vol/vol) heat-
cing the artifacts derived from the absence of the solvent. inactivated foetal bovine serum, 2 mM L-alanyl-L-glutamine,
During the optimization of the geometries and energies of penicillin-streptomycin (50 units – 50 mg per ml) and cell cul-
Models A and B, a penalty of 100 kJ Å−2 was applied for the tures were incubated at 37 °C under a humidified atmosphere
distance violations along the hydrogen bonds between nucleo- of 5% CO2, 95% air. The culture media were changed twice
tide aromatic rings of each complementary strand.
a week.
This journal is © The Royal Society of Chemistry 2014
Org. Biomol. Chem., 2014, 12, 2686–2701 | 2699

View Article Online
Paper
Organic & Biomolecular Chemistry
Treatment with antitumor compounds and MTT colorimetric
assay
4 G. M. Cragg, P. G. Grothaus and D. J. Newman, Chem. Rev.,
2009, 109, 3012–3043.
5 M. Gordaliza, M. A. Castro, J. M. M. del Corral and A. San
Feliciano, Curr. Pharm. Des., 2000, 6, 1811–1839.
Human cancer cell lines (2.5–3 × 103 cells per 0.33 cm2) were
plated in 96 well plates “Nunclon TM Microwell TM” (Nunc),
and were incubated at 37 °C. Treatment with antitumor mole-
cules, cytotoxicity and cell proliferation was conducted as pre-
viously described. After 24 h, cells were treated with the
compounds (final concentration 0.01–100 μM). Cells treated
with 0.5–1% of DMSO were used as controls. Microplates were
incubated at 37 °C under a humidified atmosphere of 5% CO2,
95% air for 3 days and then cytotoxicity was measured with the
colorimetric assay based on the use of tetrazolium salt MTT (3
(4,5 dimethylthiazol-2-yl)-2,5-diphenyl tetrazolium bromide).
The results were read on a multiwell scanning spectropho-
tometer (Multiscan reader) using a wavelength of 570 nm.
Each value was the average of 8 wells (standard deviations were
less than 10%). The GI50 value was calculated according to
NCI: thus, GI50 is the concentration of the test compound
where 100 × (T − T0)/(C − T0) = 50 (where T is the optical
density of the test well after a 72 h period of exposure to the
test compound, T0 is the optical density at time zero, and C is
the DMSO control optical density). The cytotoxicity effect was
calculated according to NCI when the optical density of treated
cells was lower than the T0 value using the following formula:
100 × (T − T0)/T0 < 0.
6 C. Spatafora and C. Tringali, Phenolic oxidative coupling in
the biomimetic synthesis of heterocyclic lignans, neo-
lignans and related compounds, in Targets in Heterocyclic
Systems. Chemistry and Properties, ed. D. S. O. Attanasi,
Società Chimica Italiana, Roma, 2007, vol. 11, pp. 284–312.
7 C. Daquino, A. Rescifina, C. Spatafora and C. Tringali,
Eur. J. Org. Chem., 2009, 6289–6300.
8 (a) F. Mazue, D. Colin, J. Gobbo, M. Wegner, A. Rescifina,
C. Spatafora, D. Fasseur, D. Delmas, P. Meunier, C. Tringali
and N. Latruffe, Eur. J. Med. Chem., 2010, 45, 2972–2980;
(b) G. Valdameri, L. P. Rangel, C. Spatafora, J. Guitton,
C. Gauthier, O. Arnaud, A. Ferreira-Pereira, P. Falson,
S. M. B. Winnischofer, M. E. M. Rocha, C. Tringali and
A. Di Pietro, ACS Chem. Biol., 2012, 7, 321–329;
(c) V. M. Bhusainahalli, C. Spatafora, M. Chalal,
D. Vervandier-Fasseur, P. Meunier, N. Latruffe and
C. Tringali, Eur. J. Org. Chem., 2012, 5217–5224.
9 S. A. S. Silva, A. L. Souto, M. F. Agra, E. V. L. Cunha,
J. M. Barbosa-Filho, M. S. Silva and R. Braz, ARKIVOC,
2004, 54–58.
10 S. Shi, Y. Zhang, K. Huang, S. Liu and Y. Zhao, Food Chem.,
2008, 108, 402–406.
11 T. Tanaka, A. Nishimura, I. Kouno, G. Nonaka and
C. R. Yang, Chem. Pharm. Bull., 1997, 45, 1596–1600.
12 Z. H. Jiang, T. Tanaka and I. Kouno, Chem. Pharm. Bull.,
1996, 44, 1669–1675.
13 M. Kumar, P. Rawat, N. Rahuja, A. K. Srivastava and
R. Maurya, Phytochemistry, 2009, 70, 1448–1455.
14 M. A. E. Watanabe, M. K. Amarante, B. J. Conti and
J. M. Sforcin, J. Pharm. Pharmacol., 2011, 63, 1378–1386.
15 S. Di Micco, F. Mazué, C. Daquino, C. Spatafora,
D. Delmas, N. Latruffe, C. Tringali, R. Riccio and
G. Bifulco, Org. Biomol. Chem., 2011, 9, 701–710.
16 G. Basini, L. Baioni, S. Bussolati, F. Grasselli, C. Daquino,
C. Spatafora and C. Tringali, Invest. New Drugs, 2012, 30,
186–190.
Acknowledgements
This research was supported by MIUR, Ministero dell’Univer-
sità e della Ricerca (PRIN 2009, Rome, Italy), and by grants
from the Università degli Studi di Catania (Progetti di Ricerca
di Ateneo, Catania, Italy) and Università di Salerno.
Notes and references
1 (a) J. Clardy and C. Walsh, Nature, 2004, 432, 829–837;
(b) A. D. Kinghorn, L. Pan, J. N. Fletcher and H. Y. Chai, J.
Nat. Prod., 2011, 74, 1539–1555; (c) C. Spatafora and 17 (a) V. Vijayakurup, C. Spatafora, C. Daquino, C. Tringali,
C. Tringali, Natural Polyphenols to Synthetic Antitumor
Agents, in: Bioactive Compounds from Natural Sources,
Natural Products as Lead Compounds in Drug Discovery, II
edn, CRC Press-Taylor & Francis, 2012, pp. 299–338;
P. Srinivas and S. Gopala, Life Sci., 2012, 91, 1336–1344;
(b) V. Vijayakurup, C. Spatafora, C. Tringali,
P. C. Jayakrishnan, P. Srinivas and S. Gopala, Mol. Biol.
Rep., 2014, 41, 85–94.
(d) D. J. Newman and G. M. Cragg, J. Nat. Prod., 2012, 75, 18 C. Spatafora, C. Daquino, C. Tringali and R. Amorati, Org.
311–335. Biomol. Chem., 2013, 11, 4291–4294.
2 T. A. Johnson, J. Sohn, W. D. Inman, S. A. Estee, 19 (a) C. Hotzel, A. Marotto and U. Pindur, Eur. J. Med. Chem.,
S. T. Loveridge, H. C. Vervoort, K. Tenney, J. K. Liu,
K. K. H. Ang, J. Ratnam, W. M. Bray, N. C. Gassner,
Y. Y. Shen, R. S. Lokey, J. H. McKerrow, K. Boundy-Mills,
A. Nukanto, A. Kanti, H. Julistiono, L. B. S. Kardono,
L. F. Bjeldanes and P. Crews, J. Nat. Prod., 2011, 74, 2545–
2555.
2003, 38, 189–197; (b) C. Keuser and U. Pindur, Pharmazie,
2006, 61, 261–268; (c) M. H. David-Cordonnier,
M. P. Hildebrand, B. Baldeyrou, A. Lansiaux, C. Keuser,
K. Benzschawel, T. Lemster and U. Pindur, Eur. J. Med.
Chem., 2007, 42, 752–771.
20 (a) C. Bourdouxhe Housiaux, P. Colson, C. Houssier,
M. J. Waring and C. Bailly, Biochemistry, 1996, 35, 4251–
4264; (b) R. Zhang, X. Wu, L. J. Guziec, F. S. Guziec,
3 G. M. Cragg and D. J. Newman, Phytochem. Rev., 2009, 8,
313–331.
2700 | Org. Biomol. Chem., 2014, 12, 2686–2701
This journal is © The Royal Society of Chemistry 2014

View Article Online
Organic & Biomolecular Chemistry
Paper
G. L. Chee, J. C. Yalowich and B. B. Hasinoff, Bioorg. Med.
Chem., 2010, 18, 3974–3984; (c) U. Pindur, M. Jansen and
T. Lemster, Curr. Med. Chem., 2005, 12, 2805–2847.
E. G. Van Meir and M. M. Goodman, J. Med. Chem., 2012,
55, 6738–6750.
29 M. C. Foti, C. Daquino and C. Geraci, J. Org. Chem., 2004,
69, 2309–2314.
21 S. Di Micco, C. Bassarello, G. Bifulco, R. Riccio and
L. Gomez-Paloma, Angew. Chem., Int. Ed., 2006, 45, 224– 30 B. Etzenhouser, C. Hansch, S. Kapur and C. D. Selassie,
228. Bioorg. Med. Chem., 2001, 9, 199–209.
22 (a) M. Mayer and B. Meyer, Angew. Chem., Int. Ed., 1999, 38, 31 M. DellaGreca, L. Previtera, R. Purcaro and A. Zarrelli,
1784–1788; (b) M. Mayer and B. Meyer, J. Am. Chem. Soc.,
2001, 123, 6108–6117.
23 S. Grasso, L. Siracusa, C. Spatafora, M. Renis and
C. Tringali, Bioorg. Chem., 2007, 35, 137–152.
J. Nat. Prod., 2007, 70, 1664–1667.
32 Y. Naito, M. Sugiura, Y. Yamaura, C. Fukaya, K. Yokoyama,
Y. Nakagawa, T. Ikeda, M. Senda and T. Fujita, Chem.
Pharm. Bull., 1991, 39, 1736–1745.
24 I. Gomez-Monterrey, P. Campiglia, A. Carotenuto, 33 D. K. Kim, J. P. Lim, J. W. Kim, H. W. Park and J. S. Eun,
D. Califano, C. Pisano, L. Vesci, T. Lama, A. Bertamino, Arch. Pharmacal Res., 2005, 28, 39–43.
M. Sala, A. Mazzella di Bosco, P. Grieco and E. Novellino, 34 J. Fu, K. Cheng, Z. M. Zhang, R. Q. Fang and H. L. Zhu,
J. Med. Chem., 2007, 50, 1787–1798. Eur. J. Med. Chem., 2010, 45, 2638–2643.
25 L. Martino, A. Virno, B. Pagano, A. Virgilio, S. Di 35 V. Sklenar, M. Piotto, R. Leppik and V. Saudek, J. Magn.
Micco, A. Galeone, C. Giancola, G. Bifulco, L. Mayol Reson., 1993, 102, 241–245.
and A. Randazzo, J. Am. Chem. Soc., 2007, 129, 16048– 36 MacroModel, version 8.5, Schrodinger LLC, New York, NY,
16056. 2003.
26 A. Leo, C. Hansch and D. Elkins, Chem. Rev., 1971, 71, 525– 37 (a) W. D. Cornell, P. Cieplak, C. I. Bayly, I. R. Gould,
616.
K. M. Merz, D. M. Ferguson, D. C. Spellmeyer, T. Fox,
J. W. Caldwell and P. A. Kollman, J. Am. Chem. Soc., 1995,
117, 5179–5197; (b) S. J. Weiner, P. A. Kollman, D. A. Case,
U. C. Singh, C. Ghio, G. Alagona, S. Profeta, Jr. and
P. K. Weiner, J. Am. Chem. Soc., 1984, 106, 765–784;
(c) P. A. K. S. J. Weiner, D. T. Nguyen and D. A. Case,
J. Comput. Chem., 1986, 7, 230–252; (d) P. K. Weiner and
P. A. Kollman, J. Comput. Chem., 1981, 2, 287–303.
27 (a) Y. Henchoz, D. Guillarme, S. Rudaz, J. L. Veuthey and
P. A. Carrupt, J. Med. Chem., 2008, 51, 396–399;
(b) C. Giaginis and A. Tsantili-Kakoulidou, J. Liq. Chroma-
togr., 2008, 31, 79–96.
28 (a) M. Calvani, L. Critelli, G. Gallo, F. Giorgi,
G. Gramiccioli, M. Santaniello, N. Scafetta, M. O. Tinti and
F. De Angelis, J. Med. Chem., 1998, 41, 2227–2233;
(b) A. Drews, S. Bovens, K. Roebrock, C. Sunderkotter, 38 W. C. Still, A. Tempczyk, R. C. Hawley and T. Hendrickson,
D. Reinhardt, M. Schafers, A. van der Velde, J. Am. Chem. Soc., 1990, 112, 6127–6129.
A. S. Elfringhoff, J. Fabian and M. Lehr, J. Med. Chem., 39 T. A. Halgren, J. Comput. Chem., 1996, 17, 616–641.
2010, 53, 5165–5178; (c) J. Mun, A. A. Jabbar, N. S. Devi, 40 R. Huey, G. M. Morris, A. J. Olson and D. S. Goodsell,
S. M. Yin, Y. Z. Wang, C. Tan, D. Culver, J. P. Snyder,
J. Comput. Chem., 2007, 28, 1145–1152.
This journal is © The Royal Society of Chemistry 2014
Org. Biomol. Chem., 2014, 12, 2686–2701 | 2701