Interaction of anthraquinone anti-cancer drugs with DNA:Experimental and computational quantum chemical study

Article abstract of DOI:10.1016/j.molstruc.2016.08.007

Anthraquinones form the basis of several anticancer drugs. Anthraquinones anticancer drugs carry out their cytotoxic activities through their interaction with DNA, and inhibition of topoisomerase II activity. Anthraquinones (AQ4 and AQ4H) were synthesized and studied along with 1,4-DAAQ by computational and experimental tools. The purpose of this study is to shade more light on mechanism of interaction between anthraquinone DNA affinic agents and different types of DNA. This study will lead to gain of information useful for drug design and development. Molecular structures were optimized using DFT B3LYP/6-31?+?G(d). Depending on intramolecular hydrogen bonding interactions two conformers of AQ4 were detected and computed as 25.667?kcal/mol apart. Molecular reactivity of the anthraquinone compounds was explored using global and condensed descriptors (electrophilicity and Fukui functions). Molecular docking studies for the inhibition of CDK2 and DNA binding were carried out to explore the anti cancer potency of these drugs. NMR and UV-VIS electronic absorption spectra of anthraquinones/DNA were investigated at the physiological pH. The interaction of the three anthraquinones (AQ4, AQ4H and 1,4-DAAQ) were studied with three DNA (calf thymus DNA, (Poly[dA].Poly[dT]) and (Poly[dG].Poly[dC]). NMR study shows a qualitative pattern of drug/DNA interaction in terms of band shift and broadening. UV-VIS electronic absorption spectra were employed to measure the affinity constants of drug/DNA binding using Scatchard analysis.

Full text of DOI:10.1016/j.molstruc.2016.08.007

Accepted Manuscript
Interaction of anthraquinone anti-cancer drugs with DNA; experimental and
computational quantum chemical study
Jamelah S. Al-Otaibi, Paul Teesdale Spittle, Tarek M. El Gogary
PII:
S0022-2860(16)30835-3
10.1016/j.molstruc.2016.08.007
MOLSTR 22833
DOI:
Reference:
To appear in: Journal of Molecular Structure
Received Date: 6 June 2016
Revised Date: 3 August 2016
Accepted Date: 5 August 2016
Please cite this article as: J.S. Al-Otaibi, P. Teesdale Spittle, T.M. El Gogary, Interaction of
anthraquinone anti-cancer drugs with DNA; experimental and computational quantum chemical study,
Journal of Molecular Structure (2016), doi: 10.1016/j.molstruc.2016.08.007.
This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to
our customers we are providing this early version of the manuscript. The manuscript will undergo
copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please
note that during the production process errors may be discovered which could affect the content, and all
legal disclaimers that apply to the journal pertain.

ACCEPTED MANUSCRIPT

ACCEPTED MANUSCRIPT
Interaction of anthraquinone anti-cancer drugs with DNA; Experimental and
computational quantum chemical study.
Jamelah S. Al-Otaibia,d, Paul Teesdale Spittleb and Tarek M. EL Gogaryc
a
Department of Chemistry, College of Science, Princess Nourah bint Abdulrahman
University, Riyadh 11951, Saudi Arabia.
b
School of Biological Sciences, Victoria University of Wellington, Wellington, New
Zealand.
c Department of Chemistry, Faculty of Science, Jazan University, Jazan, Saudi Arabia.
dDeanship of Scientific Research, Princess Nora Bint Abdul Rahman University.
Abstract
Anthraquinones form the basis of several anticancer drugs. Anthraquinones anticancer
drugs carry out their cytotoxic activities through their interaction with DNA, and
inhibition of topoisomerase II activity. Anthraquinones (AQ4 and AQ4H) were
synthesized and studied along with 1,4-DAAQ by computational and experimental
tools. The purpose of this study is to shade more light on mechanism of interaction
between anthraquinone DNA affinic agents and different types of DNA. This study
will lead to gain of information useful for drug design and development. Molecular
structures were optimized using DFT B3LYP/6-31+G(d). Depending on
intramolecular hydrogen bonding interactions two conformers of AQ4 were detected
and computed as 25.667 kcal/mol apart. Molecular reactivity of the anthraquinone
compounds was explored using global and condensed descriptors (electrophilicity and
Fukui functions). Molecular docking studies for the inhibition of CDK2 and DNA
binding were carried out to explore the anti cancer potency of these drugs. NMR and
UV-VIS electronic absorption spectra of anthraquinones/DNA were investigated at
the physiological pH. The interaction of the three anthraquinones (AQ4, AQ4H and
1,4-DAAQ) were studied with three DNA (calf thymus DNA, (Poly[dA].Poly[dT]) and
(Poly[dG].Poly[dC]). NMR study shows a qualitative pattern of drug/DNA interaction in
terms of band shift and broadening. UV-VIS electronic absorption spectra were
employed to measure the affinity constants of drug/DNA binding using Scatchard
analysis.
Keywords: Anthraquinone; Anti-cancer; reactivity descriptors; Docking; NMR; Scatchard
analysis
.

ACCEPTED MANUSCRIPT
1. Introduction
The development of novel potent chemotherapeutics has been a topic of interest for
decades [1-4]. DNA is a primary target for drugs in the cell. The mode of action of
many chemotherapeutic agents is thought to be due to their ability to intercalate
between double helix DNA base pairs . Originally, this model was first suggested in
1961 by Lerman in the context of the acridines and DNA interaction. These
chemotherapeutics are generally thought to exert their biological actions through
binding with DNA followed by an interference with DNA replication, transcription,
and inhibition of gene expression. The design of small drug molecules that
selectively target DNA, with high binding affinities, has led to the discovery of many
anticancer, antibiotic, and antiviral drugs [5-14].
Quinonoid compounds are a series of widespread compounds found in the living
organisms. These compounds are known to perform many biochemical and
physiological processes in living organisms. Quinones are organic compounds with
various interesting antitumor, antibacterial, antibiotical, and antifungal properties
[15-16]. Anthraquinones, as a group of natural quinones, are widely used in treatment
of cancer. Anthraquinones anticancer drugs carry out their cytotoxic activities through
their interaction with DNA, preferentially at guanine/cytocine rich sites. This
interaction is believed to cause significant conformational changes in the DNA
leading to inhibition of the DNA replication [17]. This may lead to DNA damage. On
the other hand, they can cause inhibition of topoisomerase II activity, leading to DNA
damage.
Most DNA-targeted molecules start their binding with double helix DNA non-
covalently which subsequently may developed to covalent binding. Non-covalent
binding may include π- stacking, hydrogen bonding, electrostatic, charge transfer,
and hydrophobic interactions [17]. All these interactions may contribute to the
drug/DNA interaction mechanism so that the main objective of this study is to explore
the dominant interaction. This information is crucial for design and development of
new drugs. Molecular docking of anthraquinones with DNA and Cyclin-dependent
kinase-2 (CDK2) will shed more light on the binding modes of these drugs with
biological molecules. CDK2 takes an important part in regulating various events of
living cell cycle. Therefore, CDK2 is considered as a potentially therapeutic target for

ACCEPTED MANUSCRIPT
cancer therapy. In addition to DNA CDK2 was chosen for our study as some
anthraquinone derivatives showed a good inhibitory effect toward CDK2 enzyme. For
this study the following anthraquinone derivatives (Scheme 1), 1,4-bis {[2-
(methylamino)ethyl]amino}-5,8-dihydroxyanthracene-9,10-dione, (AQ4), 1,4-bis {[2-
(methylamino)ethyl]amino}anthracene-9,10-dione,
diaminoanthraquinone were chosen for this study.
(AQ4H)
and
1,4-
O
NH2
N
N
OH
O
HN
O
HN
OH
O
HN
O
HN
N
N
O
NH2
AQ4
AQ4H
1,4-DAAQ
Scheme 1. Studied anthraquinones

ACCEPTED MANUSCRIPT
2. Experimental details
The UV-VIS absorption spectra were measured using a Perkin Elmer Lambda-16 UV-
VIS Spectrophotometer. NMR spectra were recorded in CDCl3 using a Bruker AC250
at 30 °C.
Calf thymus DNA, polydeoxyadenylic acid-polythymidylic acid (Poly[dA]. Poly[dT])
and polydeoxyguanylic acid-polydeoxycytidylic acid (Poly[dG]. Poly[dC]) were
purchased from Sigma Chemical Co and were used without further purification. 1,4-
DAAQ and D2O (99.9% D) were purchased from Aldrich. Trizma base
(Tris[hydroxymethyl] aminomethane) and NaCl were supplied from Sigma and used
for buffer preparation without further purification.
Synthesis of anthraquinone drugs:
1,4-bis {[2-(methylamino)ethyl]amino}-5,8-dihydroxyanthracene-9,10-dione, (AQ4),
1,4-bis {[2-(methylamino)ethyl]amino}anthracene-9,10-dione, (AQ4H) were
synthesized according to the following method [18-20]:
AQ4 Synthesis:
The synthesis of AQ4 was done in three steps;
1- Synthesis of 1,5-diamino-4,8-dihydroxy anthracene-9,10-dione. A stirred solution
of 1,5-DAAQ (6 g, 25 mmol) in concentrated sulphuric acid (97%) was
o
maintained at –5 C. Anhydrous sodium chlorate (7.2 g, 67 mmol) was slowly
added over 30 min. The mixture was allowed to warm to room temperature and
stirred for further 3 h. A precipitate resulted on slowly adding a chilled solution of
1000 cm3 of 1% sodium hydrogen sulfite, which, was removed by filtration and
washed thoroughly with water and freeze-dried to yield (6 g, 88%) of title
compound.
2- Synthesis of leuco-1,4,5,8-tetrahydroxyanthracene-9,10-dione. A stirred mixture
of 1,5-diamino-4,8-dihydroxyanthracene-9,10-dione (5 g, 19 mmol) and sodium
hydroxide (10 g, 25 mmol) in water was heated under reflux for 3 h until ammonia
was no longer evolved and the solution colour turned blue to brown. The solution
was allowed to cool to room temperature and then acidified to pH 3 with 5 M

ACCEPTED MANUSCRIPT
HCl. The resulting precipitate was removed by filtration and washed with water
and dried to afford (4.5 g, 75%) of brown powder.
3- Synthesis of 1,4-bis{[2-(methylamino)ethyl]amino}-5,8-dihydroxyanthracene-9,
10-dione. A partial solution of leuco-1,4,5,8-tetrahydroxyanthracene-9,10-dione
(1 g, 3.6 mmol) and 10 ml of N,N-dimethylethylendiamine in 50 ml of ethanol
was heated under reflux for 1 h under an atmosphere of argon and then allowed to
cool at room temperature. The suspension was stirred at room temperature for 24
h to open air. Aqueous sodium hydroxide solution (5 M, 0.5 cm3) was added and
stirred for 1 h. The solvent was removed under vacuum and gave the resulting
dark blue powder (0.6 g, 40%) after recrystallization from ethanol as the titled
1
compound. Rf (9:1:0.1 CH2Cl2/MeOH/H3N) :0.33. H NMR δ (CDCl3): 7.15 (s,
2H), 7.10 (s, 2H), 3.45 (m, 4H), 2.65 (t, 4H), 2.35 (s, 12H); 13C NMR δ (CDCl3):
185, 156, 147, 125, 126, 116, 109, 58, 46, 41 ; mass spectrum, m/z 413 (m+ + 1).
IR ν (KBr): 3500, 3100, 2800, 1650, 1625, 1475, 1375, 1465, 1225, 850 cm-1.
Anal. Calcd. For C22H28O4N4.025 H2O: C, 63.3; H, 7.0; N, 13.4. Found: C, 63.3;
H, 7.0; N, 13.4.
AQ4H Synthesis:
A partial solution of leucoquinizarine (2 g, 8.3 mmol) and N,N-
dimethylethylendiamine (10 ml ) was heated under reflux for 8 h under an
atmosphere of argon and then allowed to cool at room temperature. The
suspension was stirred at room temperature for 24 h to open air and diluted with
water to precipitate the title compound. The filtered solid was washed with water
and dried. The solid was chromatographed (SiO2, (9:1 CH2Cl2/ MeOH; 9:1:21
CH2Cl2/ MeOH/NH3), then washed with hexane and recrystallized from ethanol
giving the title compound. 1H NMR δ (CD3OD): 8.3 (q,2H), 7.75 (q, 2H), 7.38 (s,
2H), 3.45 (m, 4H), 2.7 (t, 4H), 2.4 (s, 12H); 13C NMR δ (CDCl3): 187, 158, 149,
138, 135, 129, 128, 62, 48, 44 ; mass spectrum, m/z 381 (m+ + 1). Anal. Calcd.
For C22H28O2N4: C, 69.45; H, 7.42; N, 14.73. Found: C, 69.50; H, 7.22; N, 14.81.

ACCEPTED MANUSCRIPT
3. Computations
All computations were done using G09 suit of programs [21]. Molecular geometry of
anthraquinone compounds were optimized in the gas phase at DFT B3LYP/6-31+G(d,p) level
of theory. A frequency job was performed on the optimized geometry to confirm a minimum
energy structure. Fukui functions were calculated using DMol3 module [22,23] employing
B3LYP/DND method implemented in Material studio program [24].

ACCEPTED MANUSCRIPT
4. Results and Discussion
4.1. Computational work
4.1.1. Molecular Geometry:
Since there is no reported experimental molecular geometrical data for AQ4 and
AQ4H, we calculated the geometry of AQ4, AQ4H and 1,4-DAAQ at B3LYP/6-
31+G(d). Two minimum energy conformers, Conf-1 and Conf-2, of AQ4 were
studied. Conf-1 allows two intra-molecular hydrogen bonds while Conf-2 allows four
intra-molecular hydrogen bonds and it is 25.667 kcal/mol more stable than AQ4
Conf-1 in the gas phase. Optimized geometry of AQ4 is shown in Fig. 1. Some
geometrical parameters of AQ4 Conf-1 are represented in Table S1 in the
Supplementary Material. The carbonyls bond lengths C7=O15 and C10=O16 are
calculated as 1.271 and 1.233 Å respectively. The reason for elongation of the former
bond length is due to the engagement of the O15 with two intramolecular hydrogen
bonds. Formation of hydrogen bond results in more polarization of the carbonyl group
and hence decreasing duple bond character which affect the elongation of the bond.
On the other hand, O16 is not engaged with hydrogen bonding formation. Bond
lengths of the two hydrogen bonds, O17
−
H35…O15 and N20 H38…O15 are 1.645 and
−
1.780 Å respectively. The former hydrogen bond is stronger than the latter as can be
inferred from its 0.135 Å shorter bond length. This could be attributed to the higher
electronegativity of the O atom with respect to N atom. Hydrogen bond formation
also affects the length of O17
Table S1 the bond length O17
The elongation of the bond length is a direct reason of hydrogen bond formation. The
angle C5 C4 C3 C7 is
C10 is computed to 121.5o while the corresponding angle C2
calculated is computed 119.6o. The later angle is decreased to permit effective
hydrogen bond interaction. The same argument is true for the angle C7 C8 C14
(119.8°) and the angle C10 C9 C11 (121.0°) where the contraction of the former angle
facilitates the N H…O hydrogen bond formation. The ring system in Conf-1 is not
planar as appeared from the values of the dihedral angles, C7 C8 C9 C10, C11 C9-
C8 C14, C2 C3 C4 C5, C3-C4 C10 O16 and C8 C9 C10 C16 are 13.6, 6.4, 1.9, 158.8 and
−
H35 in comparison with O18
−
H36. As can be seen from
−
H35 ( 0.996 Å) is larger than that of O18
−
H36 (0.970 Å).
−
−
−
−
−
−
−
−
−
−
−
−
−
−
−
−
−
−
−
−
−
−
153.6 o respectively (see Table S1). All these angles deviate from planarity.
Selected geometrical parameters of AQ4 Conf-2 are represented in Table S2 in the
Supplementary Material. Conf-2 is 25.667 kcal/mol more stable than AQ4 Conf-1.

ACCEPTED MANUSCRIPT
This stabilization is due to the extra two hydrogen bonds and the more relaxation of
the side chains (see Fig. 1). As can be seen from Table S2, the carbonyl C5=O10 and
C3 O18 bond lengths are equal and longer (1.273 Å) than the corresponding bonds in
−
Conf-1. This elongation results from decreasing of the double bond character as a
result of hydrogen bond formation. The ring C=C bond lengths are ranging from
1.381-1.460 Å which are characteristic to aromatic C=C bond. The ring system in this
conformer is planar as can be confirmed by the flatness of the ring dihedral angles in
Table S2. This is in contrast to the puckered structure of the Conf-1 as above.
Optimized geometrical parameters of AQ4H and 1,4-DAAQ are displayed in Table
S3 and S4 respectively in the Supplementary Material. From Table S3 the ring system
is planar and the amines are not planar as appeared from the diheadral angle values.
Intra molecular hydrogen bonds in AQ4H and 1,4-DAAQ are formed as can be seen
from Fig. 1 and Tables S3 and S4. In AQ4H, the hydrogen bond N25−H26…O10 is
1.759 Å with an angle O10
−
H26
−
N25 136.7 and hydrogen bond N23−H24…O19 is 1.769
o
Å with an angle O19 H24
−
−
N23 135.7 as shown from Table S3. In 1,4-DAAQ, the
amino group is planar as shown from Table S4. The intra molecular hydrogen bond
N26−H27…O10 is 1.842 Å with an angle O10
−H27
−N26 129.4 and hydrogen bond
o
N23−H24…O19 is also 1.842 Å with an angle O19
−H24
−
N23 129.4 as shown from
Table S4. This similarity of the hydrogen bonds strength and geometry results from
the symmetry of the molecule.

ACCEPTED MANUSCRIPT
AQ4, Conf-1
AQ4, Conf-2
1,4-DAAQ
AQ4H
Fig. 1. Optimized Structure of AQ4, AQ4H and 1,4-DAAQ showing atomic numeration and
H-bonding.

ACCEPTED MANUSCRIPT
4.1.2. Chemical reactivity:
Chemical reactivity descriptors were computed to explore the readability of
anthraquinones to interact with DNA. The density of electrons on an atom is an
important property that contains all the information about the molecular systems.
Descriptors of chemical reactivity are good tools for predicting and understanding
reactivity of compounds. These descriptors initially developed within the density
functional theory framework.
Global quantities as chemical potential (µ), electronegativity (χ), hardness (ƞ),
softness (S), and electrophilicity (ω) of anthraquinones were calculated according to
the equations presented elsewhere [25] and presented in Table 1. From the results in
Table 1, it may be observed that the global hardness decreases in the order 1,4-DAAQ
> AQ4 > AQ4H. Global electrophilicity of AQ4 is slightly smaller than that of AQ4H
while 1,4-DAAQ has the smallest global electrophilicity. This suggests that the
electrophilic behavior of these anthraquinone decreases in the order AQ4H > AQ4 >
1,4-DAAQ.

ACCEPTED MANUSCRIPT
Table 1
Calculated global quantities of anthraquinones: chemical potential (
), hardness ( ), softness (S), electrophilicity ( ) and Calculated charge transfer ∆N
between the studied antraquinones and DNA bases and base pairs.
µ
), electronegativity
(
χ
ƞ
ω
µ
Ƞ
1.207
S
ω
6.324
χ
-3.907
3.907
0.829
AQ4
AQ4H
-3.879
-4.048
3.879
4.048
1.189
1.307
6.329
6.269
0.841
0.765
T
1,4-DDAQ
A
G
C
AT
GC
0.082
0.136
-0.201
-0.429
-0.375
-0.713
0.020
0.071
-0.247
0.666
0.717
0.399
-0.245
-0.194
-0.510
-0.565
-0.519
-0.810
AQ4
AQ4H
1,4-DAAQ
The calculated charge transfer ∆N between two systems A and B (the studied
anthraquinones and DNA bases and DNA base pairs) was computed according to the
equation [25-27]:
µB
−
µA
η B
∆N =
2
(
η A
+
)
∆N is summarized in Table 1. Inspection of the calculated charge, as shown in the
Table shows that the charge transfer takes place from DNA base pairs (AT and GC) to
anthraquinones. The higher transferred charge with GC reflects the extent of
interaction between anthraquinone and DNA base pair which is in agreement with our
previous experimental studies [17]. AQ4 and AQ4H are electron donors to A, C, and
T while they are electron acceptor toward G. 1,4-DAAQ has an electron accepting
nature toward all bases and base pairs except T. Comparing AQ4 and AQ4H with 1,4-
DAAQ confirm the electron donating character of the side chain in the former. AQ4H
is stronger electron donor than the AQ4 which indicates the electron withdrawing
nature of the hydroxyl groups.

ACCEPTED MANUSCRIPT
Fukui [28] introduced a qualitative approach of chemical reactivity in the form of
what we call Frontier Orbital Theory. Later, this theory was demonstrated [29,30] in
the framework of DFT. In a molecular system, the atomic site, which possesses
highest condensed Fukui function, favors the higher reactivity. Tables 5S-7S in the
Supplementary Material show the condensed Fukui functions and local softness as
calculated based on Mullikin and Hirchfeld charges for AQ4, AQ4H and 1,4-DAAQ.
Isosurface maps for Fukui functions of electrophilic, nucleophilic and radical attack
are given in Fig. S1 in the Supplementary Material. Isosurface maps show that AQ4
acts better as nucleophilic and free radical attack than electrophilic attack. Ring
system in AQ4 has no electrophilic centers, these centers are shown on oxygen and
nitrogen atoms. On the other hand, nucleophilic and free radical attack are shown on
the ring system. From the results of Table 5S, the reactivity order for the radical
attack was found on C1, C6, C7, C12, C13, O15, O16, O17, N19 and N20. Among
these atoms the reactivity of radical attack could be ranked as N19 > O16 > N20 =
O15 > C12 > C13 > C6 = C1 > C7 > O17. For nucleophilic attack, the most reactive
sites are O16 and O16. The order of reactivity toward nucleophilic attack could be
ranked as O16 = O15 > C7 > C6 > C13 > C1 > C10 = O17 > N20 > H33 > H32 >
H31 = N19 > H34. On the other hand, for electrophilic attack, the most reactive site is
N19. The reactivity order toward electrophilic attack could be ranked as N19 > N20 >
C9 > C8 > O17 > O16 > C13 > O15 > C14 = C1 > C6 = H37 > C12 > N28 = H33 =
H34. From Table 6S, it is clear that AQ4H is less reactive than AQ4 when comparing
values (Mullikin and Hirshfeld) of the local softness and condensed Fukui functions
for electrophilic and nucleophilic attack. O19 has larger tendency (0.74) for
nucleophilic attack than O10 (0.73) as shown from Table 6S. N23 and N25 are also
sites for nucleophilic attack. 1,4-DAAQ (Table 7S) shows the highest values of local
softness and condensed Fukui functions for electrophilic and nucleophilic attack. O15
and O16 (0.081) are most nucleophilic sites and N17 and N18 (0.091) are the most
electrophilic sites.

ACCEPTED MANUSCRIPT
4.1.3. Molecular docking and binding energy calculations
Anthraquinone anticancer drugs are known to exert their anticancer activity via
inhibition of human CDK2 and DNA intercalation. Accordingly, we docked AQ4,
AQ4H and 1,4-DAAQ into the crystal structures of CDK2 and two DNA employing
both intercalation and minor groove interactions. The coordinates for the protein
structure and DNA was obtained from the RCSB Protein Data Bank (PDB; 2A4L,
(Resolution: 2.40 Å) [31] ,2GVR, (Resolution: 1.65 Å) [32] and 1Z3F, (Resolution:
1.50 Å) [33]). Protein Structures were prepared using Discovery Studio (DS 2.0)
software package [34]. The invalid or missing residues were added and the structures
were aligned using the protein structure alignment module. Hydrogen atoms were
added and the structure was minimized using CHARMm force field to relax the
backbone and to remove the clashes. Anthraquinones were optimized by DFT method
(B3LYB/6-31+G(d)) using G09 software to eliminate bond length and bond angle
biases and saved to be used in docking calculations. C-Docker with its default
parameters was used to dock the anthraquinone compounds into CDK2and two DNA
active sites (classic and minor groove sites) to study all possible interactions with
DNA. Figure 2 shows binding model of 1,4-DAAQ, AQ4 and AQ4H with human
Cyclin-dependent kinases (CDK2) transferase. The results are summarized in Table 2.
Table 2
C-Docker interaction energy and binding energy of anthraquinones with CDK2 enzyme and
DNA.
System
AQ4/CDK2
binding energy
-38.880
interaction energy
-58.077
AQ4H/CDK2
-33.340
-52.914
1,4-DAAQ/CDK2
AQ4/DNA (intercalation)
AQ4H/DNA (intercalation)
-20.632
-53.665
-58.357
-30.751
-54.189
-57.331
1,4-DAAQ/DNA (intercalation) -15.909
-26.080
AQ4/DNA (minor groove)
AQ4H/DNA ( minor groove )
-34.244
-29.496
-52.100
-50.394

ACCEPTED MANUSCRIPT
(a)
(b)
(c)
Fig. 2. Binding model of (a)1,4-DAAQ, (b) AQ4, and (c) AQ4H, with human Cyclin-dependent kinases (CDK2) transferase
hydrogen bonding (green dotted line) and non-bonding interaction (white dotted line).

ACCEPTED MANUSCRIPT
By analysis the data in Table 2 it is obvious that compound AQ4 has the highest C-
Docker binding energy and the highest C-Docker interaction energy. This is because
it fits well with the CDK2 pocket and its side chain has a freedom to interact well
with the residues in active site as shown in Fig. 2. Also one of the hydroxyl groups
has a hydrogen bond with Lys89 and another hydrogen bond with Glu12, which make
its interaction more favorable than AQ4H (Fig.2). While 1,4-DAAQ lacks the side
chains so its interactions with the receptor are missing the most efficient non-bonding
interactions hence its poor binding (Fig.2).
Classic DNA interactions:
Intercalation binding model of 1,4-DAAQ, AQ4 and AQ4H with DNA showing
hydrogen bonding, non-bonding interaction and hydrophobic interactions are
presented in Fig. 3 and computed energies are given in Table 2.
The data in Table 2 showed that compound AQ4H has the best affinity towards DNA
(classical interactions) over compound AQ4. The interpretation of this can be
concluded from the docking pictures, which showed that AQ4H buried in the pocket
very well and its non-bonding interactions was greater than AQ4. It forms two
hydrogen bonds with A:Cyt5, B:Gua2 and forms many hydrophobic and non-bonding
interactions with A:Cyt5, A:Gya6, B:Gua2, B:Cyt1. Compound 1,4-DAAQ is small
and has not too small conformations so its interactions with DNA is small and
consequently its modeling values were too small.
Minor groove interactions:
Minor groove binding model of AQ4H and AQ4 with DNA showing hydrogen
bonding and hydrophobic interactions are shown in Fig. 3. The docking results of
AQ4 and AQ4H in DNA minor groove were shown in Table 2. The results revealed
that AQ4 has favorable interactions over compound AQ4H, which can be attributed to
its two hydroxyl groups that oriented to the solvent exposure area . Finally, the classic
interactions were favorable over the minor groove interactions as the docking results
showed because of the one of anthraquinone rings was oriented to the solvent
exposure area, which is not favorable. This result was also confirmed by the results of
the DNA spectrophotometric titration as shown below (see Table 6). The UV-VIS

ACCEPTED MANUSCRIPT
spectrophotometric DNA titration shows that the intercalation (classic) affinity
constants are much greater than the surface binding affinity constants.
(a)
(b)

ACCEPTED MANUSCRIPT
(c)
(d)
(e)
Fig. 3. Intercalation binding model of (a) 1,4-DAAQ, (b) AQ4, (c) AQ4H, (d) AQ4H (minor
groove binding) and (e) AQ4 (minor groove binding) with DNA showing hydrogen bonding
(green dotted line) and hydrophobic interactions (pink dotted lines).

ACCEPTED MANUSCRIPT
4.2. Experimental Study
In this work the anthraquinone drug/DNA interactions have been discussed. Three
types of DNA were used, for such a study, namely calf thymus DNA,
polydeoxyadenylic
acid-polythymidylic
acid
(Poly[dA].Poly[dT])
and,
polydeoxyguanylic acid-polydeoxycytidylic acid (Poly[dG].Poly[dC]). The drugs
used include 1,4-bis {[2-(methylamino)ethyl]amino}-5,8-dihydroxyanthracene-9,10-
dione, (AQ4), 1,4-bis {[2-(methylamino)ethyl]amino}anthracene-9,10-dione,
(AQ4H), and 1,4-DAAQ. This study has been done using NMR and UV-VIS
spectroscopy.
4.2.1. NMR Studies:
The NMR spectra of drug/DNA mixtures were obtained in D2O buffer at pD=7.2 in
different drug/DNA ratios. No NMR signals for DNA were observed by 250 MHz
machine under the conditions used. Different drug NMR signals in the spectra of
drug/DNA mixtures were compared with that of the corresponding drug in its pure
solution. Generally two observations were obtained, the first is the band shift and the
second is the band broadening.
NMR spectra of AQ4, AQ4H, and 1,4-DAAQ with calf thymus DNA,
(Poly[dA].Poly[dT]) and (Poly[dG].Poly[dC]) were obtained in D2O buffer at pD=7.2
o
in different drug/DNA ratios at 30 C in presence of TMS as an internal standard.
Figs. S2-S4 in the Supplementary Material show the NMR spectra of the AQ4/calf
thymus DNA, AQ4/P(dA).P(dT) and AQ4/P(dG).P(dC) as an example. Tables 3-5
gives the degree of band shift (in Hz, defined as (shift unbound)-(shift bound).) and
the presence of the broadening for different drug/DNA ratios of the studied drugs with
different DNA. Delta shift value would show the change in the electronic
environment compared to unbound.
Different drug NMR signals in the spectra of drug/calf thymus DNA mixtures were
compared with that of the corresponding drug in its pure solution. It was observed that
the side chain protons were always shifted downfield whereas the aromatic protons
show up-field shift, except for 1,4-DAAQ. Methyl protons of AQ4H are much more
shifted (9.47 HZ) than that in AQ4 (0.38 HZ). As shown above in the molecular
modeling section that the hydroxyl groups of the intercalated drug in AQ4 form

ACCEPTED MANUSCRIPT
intermolecular hydrogen bonding that leads to locating of the N(CH3)2 outside the
interaction zone (see Fig. 3b). In case of intercalated AQ4H the N(CH3)2 group is
present within the interaction site as demonstrated in the modeling Fig. 3c. This
explains the high band shift of the methyl groups in AQ4H as presented in Table 3.
Another interesting observation is the band broadening which depends on the
drug/DNA ratio. The degree of broadening increases with increasing of DNA
concentration and at high DNA concentration the drug bands are too broad to be seen
at all. On the other hand, at low DNA concentration there is an equilibrium between
the bound and the free drug, and an average-positioned signal was observed as shown
from Table 3.
NMR spectroscopy is the most efficient experimental method, since it allows not only
the equilibrium analysis, but also the determination of the structures of complexes on
the atom-atom level. The order of shift of the N(CH3)2 of AQ4 is P(dG).P(dC) >
P(dA).P(dT) > calf thymus DNA. The shift in the aromatic protons of 6&7 positions
is larger than that of protons of 2&3. This fact is the same in 1,4-DAAQ in which
protons at positions 5 and 6 are more shifted than protons at positions 2&3. To
interpret this observation, there are two possible ways of intercalation of the 1,4-
disubstituted anthraquinones into the DNA : (1) side chains bind to different
phosphate backbones (the drug molecular axis is prependicular to the axis between the
two backbones) and (2) side chains bind to the same phosphate backbone (the drug
molecular axis is parallel to the axis between the two backbones). The magnitude of
the sift of the protons of AQ4H and 1,4-DAAQ suggests that these drugs are GC
selective, which is inconsistent with the results from the spectrophotometric titration.

ACCEPTED MANUSCRIPT
Table 3
1H NMR chemical shift differences, ∆ (Hz) for different drug protons in their mixtures with
calf thymus DNA:
Drug
:
N(CH3)2
ArHNCH2
Ara
Broadening
Broadening
Broadening
System
DNA
∆
∆
AQ4/calf
thymus
1:0.5 0.08
-1.23
-3.12
-4.1
√
√
√
√
1:1
0.38
9.47
√
√
√
√
-4.39
AQ4H/calf 1:1
thymus
too broad
√
1:2
1:5
7.62
Unseen
-
too broad
unseen
6.51
√
√
-
√
√
-
√
√
√
1,4-
1:0.5
DAAQ/calf
thymus
10.75
7.92
1:1
-
-
-
2.07
√
8.87
4.23
a Aromatic protons in AQ4 exhibit two shifts for protons at positions 2&3 and 6&7 are presented
sequentially. For AQ4H and 1,4-DAAQ the shifts for protons at positions 2&3, 5&8 and 6&7 are
presented sequentially.

ACCEPTED MANUSCRIPT
Table 4
1H NMR chemical shift differences, ∆ (Hz) for different drug protons in their mixtures with
P(dA).P(dT ):
Drug:
DNA
N(CH3)2
ArHNCH2
Ara
Broadening
Broadening
Broadening
System
∆
∆
AQ4/
P(dA).P(dT)
1:0.25 2.93
4.27
0.40
√
√
√
√
√
1:0.5
1:0.5
1.84
7.21
-0.04
-3.26
too broad
-0.6
√
√
√
√
AQ4H/
P(dA).P(dT)
√
√
√
√
√
√
√
√
√
-
-1.0
1:1
6.98
-0.84
-2.93
-0.76
√
-
√
-
1,4-DAAQ/
P(dA).P(dT)
1:0.5
-
4.1
-
-
1:1
-
-
-
-
a
Aromatic protons in AQ4 exhibit two shifts for protons at positions 2&3 and 6&7 are presented
sequentially. For AQ4H and 1,4-DAAQ the shifts for protons at positions 2&3, 5&8 and 6&7 are
presented sequentially.

ACCEPTED MANUSCRIPT
Table 5
1H NMR chemical shift differences, ∆ (Hz) for different drug protons in their mixtures with
P(dG).P(dC):
Drug:
DNA
1:0.5
N(CH3)2
ArHNCH2
Ara
Broadening
Broadening
Broadening
System
∆
∆
2.45
1.06
AQ4/
P(dG).P(dC)
2.24
√
√
√
√
√
√
√
√
√
√
√
√
-
AQ4H/
P(dG).P(dC)
1:0.5
7.21
-1.38
-2.58
unclear
1:0.25 6.88
-1.77
√
√
-3.98
unclear
1,4-DAAQ/
P(dG).P(dC)
1:0.5
-
-
-
8.54
16.62
10.88
-
-
a Aromatic protons in AQ4 exhibit two shifts for protons at positions 2&3 and 6&7 are presented
sequentially. For AQ4H and 1,4-DAAQ the shifts for protons at positions 2&3, 5&8 and 6&7 are
presented sequentially.

ACCEPTED MANUSCRIPT
4.2.2. UV-VIS Studies :
The interactions between three types of DNA (calf thymus DNA, (Poly[dA].Poly[dT])
and, (Poly[dG].Poly[dC])) and anthraquinone drugs (AQ4, AQ4H and 1,4-DAAQ)
were investigated in Tris buffer at pH=7.2 by absorption spectroscopy.
The absorption spectra of AQ4, AQ4H and 1,4-DAAQ with DNA were obtained in
Tris buffer at the physiological pH in different drug/DNA ratios at room temperature.
Fig. 4 shows the absorption spectra of AQ4/calf thymus DNA system as an example.
Similar spectra were obtained for other systems. As can be seen from the spectra the
changes in the uv-vis spectra induced by excess amounts of DNA to the drug solution,
show hypochromicity in the absorbance of the drug. Furthermore, a shift to higher
wavelength was also observed as compared to the free drug. Table S8 in the
Supplementary Material gives the degree of the red shift (in nm) observed in the
spectra of the drugs with DNA. The modifications of these absorption spectra are
common features of drugs upon binding to DNA due to strong interactions between
the electronic states of the DNA bases with the drug [35]. This leads to an electronic
redistribution of the bound drug, leading usually to spectral changes (red-shifts and
hypochromicity) [36,37].
The absorption titration data for these drugs with DNA were employed to calculate
the affinity constants and number of base pairs occupied by a bound drug molecule.
The experiments were done in triplicates to estimate the uncertainity in the measured
parameters. Fig. S5 in the Supplementary Material shows the uncertainity (as a
standard deviation) in the absorbance values at 605 nm against DNA/drug ratio for the
system AQ4/ P(dA).P(dT) . Similar plots for other systems were obtained. The
absorbance titration data for the drugs bound to DNA were analyzed according to the
MC Ghee and Von Hippel model [38] and plotted as a Scatchard representation. Fig.
5 shows the Scatchard plot for the binding of AQ4/ P(dA).P(dT). Similar Scatchard
plots were obtained for other systems.

ACCEPTED MANUSCRIPT
0.8
0.7
0.6
0.5
0.4
0.3
0.2
0.1
0.0
6
5
4
3
1
2
1
6
300
400
500
600
700
nm
Fig. 4. The electronic absorption spectra of AQ4/calf thymus DNA in Tris buffer
solution (pH=7.2). Concentration of AQ4 is 2.314x10-5 M, concentrations of DNA
are: (1) 0.0 (2) 3.636x10-6, (3) 9.695x10-6, (4) 1.818x10-5, (5) 3.636x10-5 and (6)
5.454x10-5 M.
500000
400000
300000
200000
100000
0
0.05
0.10
0.15
0.20
0.25
r
Fig. 5. Scatchard plot of AQ4 bound to P(dA).P(dT). Solid line shows a theoretical
curve based on smoothing to adjacent averaging.

ACCEPTED MANUSCRIPT
The curvature of the Scatchard plot is not suprising since two types of binding modes
of the drug to DNA are expected. The data points in the Scatchard plots can be
devided to two portions and treated separately to determine the binding parameters for
each mode of binding. The least squares method was used to fit the data points
(linearly) of each data point portion of the Scatchard plot. Affinity constants (K) and
number of base pairs occupied by a bound drug molecule (n) were calculated from the
slope and the intercept with the X-axis, respectively, and are displayed in Table 6
where K1 and n1 are for intercalation mode and K2 and n2 for surface binding mode. In
order of decreasing intercalation affinity constant of the drugs to all types of DNA,
the drugs ranked as follows: AQ4 > AQ4H > 1,4-DAAQ. This result is consistent
with the calculated charge transfer ∆N (Table 1). However, in order of decreasing
surface binding affinity constant of the drugs to calf thymus DNA the drugs ranked as
follows: AQ4H > AQ4. AQ4/calf thymus DNA shows the most stable complex
(2.7x107 M-1). AQ4 shows largest tendency and selectivity to AT sites with an affinity
constant (4.9x107 M-1). This is consistent with the NMR investigations which show
highest chemical shift with P(dA).P(dT) as displayed in Table 4. However, this result
is inconsistent with the calculated charge transfer ∆N (Table 1) which shows that
highest charge transfer, ∆N, was found with GC. Affinity constants for surface
binding (≈ 2-8 x 104 M-1) are much smaller than that of the intercalation binding (≈
0.2-5 x 107 M-1). Surface binding affinity constants values are comparable with that
measured elsewhere [1,2]. On the other hand, the measured intercalation affinity
constants suggest strong drug/DNA interaction which are much more than that
measured for flavonoid/DNA intercalation [39,40]. A general value of n is 2.5-3.0
was observed only for the systems AQ4/calf thymus DNA and AQ4/P(dG).P(dC)
which is in the range typically observed for DNA intercalators.

ACCEPTED MANUSCRIPT
Table 6
Affinity constants (K)a and numbers of base pairs bound per drug molecule (n)b for the
binding of AQs drugs with DNA.
Drug/DNA
K1
n1
K2
n2
AQ4/calf
thymus
AQ4H/ calf
thymus
1,4-DAAQ/
calf thymus
AQ4/
P(dA).P(dT)
AQ4H/
P(dA).P(dT)
1,4-DAAQ/
P(dA).P(dT)
AQ4/
P(dG).P(dC)
AQ4H/
P(dG).P(dC)
1,4-DAAQ/
P(dG).P(dC)
2.7x107 ±8.5x105
1.4x107±2.2x106
1.7x106±1.8x105
4.9x107 ±7.6x106
1.8x107±6.28x106
2.3x106±1.5x105
9.1x106 ±1.9x106
5.1x106±1.6x106
2.29x106±126735
2.63±0.027
8.33±0.003
12.8±0.0038
10.50±0.000
12.9±0.002
1.02±0.0028
3.28±0.001
10.7±0.004
0.89±0.003
5.0x104±19940
8.0x104±20045.6
not detected
2.3x104±44641
Not detected
Not detected
not detected
not detected
not detected
1.06±0.22
5.13±0.01
not detected
3.20±0.04
not detected
not detected
not detected
not detected
not detected
a: K1 intercalation binding affinity constant, K2 surface binding affinity constant
b: n1 number of drug molecules intercalated per base pair, n2 number of drug molecules
surface binding per base pair.

ACCEPTED MANUSCRIPT
Conclusion
AQ4 and AQ4H were synthesized and confirmed. Anthraquinone DNA affinic agents
AQ4, AQ4H and 1,4-DAAQ were investigated computationally and experimentally to
explore their potency as anti-cancer drugs. According to the computational studies at
B3LYP/6-31+G(d) level two conformers of AQ4 were identified which show 25.667
kcal/mol difference due to intra molecular hydrogen bonding. Values of computed
Fukui functions and atomic softness for nucleophilic attack prove the higher
susceptibility of these anthraquinones for nucleophilic attack. The calculated charge
shows that the charge transfer takes place from DNA base pairs (AT and GC) to
anthraquinones. The higher transferred charge with GC reflects the extent of
interaction between anthraquinone and DNA base pair which is in agreement with the
molecular docking studies (Table 2) and our previous experimental studies [17].
Molecular docking of the studied anthraquinones were performed to CDK2 and DNA.
AQ4 has the highest binding energy and the highest interaction energy as it fits well with the
CDK2 pocket. In addition to that its side chain has a freedom to interact with the residues in
active site. Also one of the hydroxyl groups has a hydrogen bond interactions Lys89 and
Glu12, which make its interaction more favorable over AQ4H. Molecular docking studies
and UV-VIS spectrophotometric DNA titration studies prove the presence of two
binding modes namely intercalation and surface binding as shown from the molecular
modeling diagram in Fig. 3 and the curvature of the Schatchard plot in Fig. 5. Both
molecular docking and UV-VIS spectrophotometric DNA titration studies confirm the
predominance of the intercalation binding as appeared from the values of the
computed interaction energy and binding energy (Table 2) and measured affinity
constants (Table 6). UV-VIS spectrophotometric DNA titration studies show
hypochromicity and red shifts upon interaction. AQ4/calf thymus DNA shows the
most stable complex (2.7x107 M-1). AQ4 shows largest tendency and selectivity to
AT sites with an affinity constant (4.9x107 M-1). NMR studies qualitatively, conclude
the drug/DNA interaction in terms of band shift and band broadening.

ACCEPTED MANUSCRIPT
Acknowledgement
This work was supported by the Deanship of Scientific Research, Princess Nora Bint Abdul
Rahman University.