Interactions of antiviral indolo[2,3-b]quinoxaline derivatives with DNA

Article abstract of DOI:10.1021/jm800787b

Here, we present the synthesis of five novel indoloquinoxaline derivatives and investigate the DNA binding properties of these monomeric as well as dimeric compounds using absorption, fluorescence, and linear dichroism. Several of the mono- and dicationic derivatives presented have previously demonstrated an excellent antiviral effect that is higher than already acknowledged agents against human cytomegalovirus (CMV), herpes simplex virus type 1 (HSV-1), and varicella-zoster virus (VZV). We find that the DNA binding constants of the monomeric and dimeric derivatives are high (～10⁶) and very high (～10⁹), respectively. Results from the spectroscopic measurements show that the planar aromatic indoloquinoxaline moieties upon interaction with DNA intercalate between the nucleobases. Furthermore, we use poly(dAdT) ₂ and calf thymus DNA in a competitive binding experiment to show that all our derivatives have an AT-region preference. The findings are important in the understanding of the antiviral effect of these derivatives and give invaluable information for the future optimization of the DNA binding properties of this kind of drugs.

Full text of DOI:10.1021/jm800787b

7744
J. Med. Chem. 2008, 51, 7744–7750
Interactions of Antiviral Indolo[2,3-b]quinoxaline Derivatives with DNA
L. Marcus Wilhelmsson,† Ngarita Kingi,‡,§ and Jan Bergman*,‡
Department of Chemical and Biological Engineering/Physical Chemistry, Chalmers UniVersity of Technology, SE-41296 Gothenburg, Sweden,
Unit for Organic Chemistry, Department of Biosciences and Nutrition, Karolinska Institute, SE-14157, Huddinge, Sweden, and Drug
DeVelopment, VironoVa AB, Smedjegatan 6, SE-13134 Nacka, Sweden
ReceiVed June 30, 2008
Here, we present the synthesis of five novel indoloquinoxaline derivatives and investigate the DNA binding
properties of these monomeric as well as dimeric compounds using absorption, fluorescence, and linear
dichroism. Several of the mono- and dicationic derivatives presented have previously demonstrated an
excellent antiviral effect that is higher than already acknowledged agents against human cytomegalovirus
(CMV), herpes simplex virus type 1 (HSV-1), and varicella-zoster virus (VZV). We find that the DNA
binding constants of the monomeric and dimeric derivatives are high (∼106) and very high (∼109),
respectively. Results from the spectroscopic measurements show that the planar aromatic indoloquinoxaline
moieties upon interaction with DNA intercalate between the nucleobases. Furthermore, we use poly(dA-
dT)2 and calf thymus DNA in a competitive binding experiment to show that all our derivatives have an
AT-region preference. The findings are important in the understanding of the antiviral effect of these
derivatives and give invaluable information for the future optimization of the DNA binding properties of
this kind of drugs.
Introduction
The indole alkaloid ellipticine (1 in Figure 1), first isolated
in 1959 by Goodwin et al.1 from Ochrosia elliptica, appeared
to display anticancer activity, and within the next two decades,
intense research in both the synthetic and biological areas of
ellipticine and its analogues produced commercial cancer
chemotherapy agents such as celiptium.2 Unfortunately, prob-
lems with the toxicity of these agents have more or less limited
the success of these drugs in the pharmaceutical market.
Figure 1. The alkaloid ellipticine, 1, the indolo[2,3-b]quinoxaline
system, 2, and the parent compound of the indoloquinoxaline derivatives
used in this study: 2,3-dimethyl-6(2-dimethylaminoethyl)6H-indolo-
(2,3-b)quinoxaline (B-220), 3.
Nevertheless research continued with the indolo[2,3-b]qui-
noxaline system (i.e., 2 in Figure 1) as analogues to the cytotoxic
agent ellipticine. The replacement of the methyl groups in
positions 5 and 11 by nitrogen atoms gave little or no
cytotoxicity against most cancer models except for Burkitt’s
lymphoma, a cancer type that is closely related to viral activity.3
This information resulted in the development of 2,3-dimethyl-
6(2-dimethylaminoethyl)6H-indolo-(2,3-b)quinoxaline (B-220a)4
(3 in Figure 1), and further studies revealed that it exhibited
good antiviral activity against herpes simplex virus type 1 (HSV-
1), cytomegalovirus (CMV), and varicella-zoster virus (VZV).
It has been reported that the mechanism of the antiviral action
against HSV-1 appears to involve binding by intercalation into
the DNA helix and, thus, disturbing the processes that are vital
for viral uncoating.5,6 Further development of a new series of
indoloquinoxaline derivatives related to B-220 have shown to
have a significant effect against autoimmune inflammatory
diseases such as multiple sclerosis (MS), rheumatoid arthritis
(RA), and other diseases.7
To improve DNA-binding of the family of ellipticine
analogues and to optimize their antiviral effect, we have
synthesized a series of novel indoloquinoxaline derivatives (4-8
in Figure 2). In this optimization of the drugs, we have varied
parameters that from previous studies are well known to be
important for DNA binding, e.g., the number of positive
charges,8-11 the number of interacting units (comparison
between monomeric or dimeric compounds),12-32 and the
hydrophobicity of the interacting unit (see for example Karlsson
et al.33). The quaternary mono- or dicationic agents (6-8 in
Figure 2) developed using the structurally related 1-3 as parent
compounds have indeed been shown to have excellent inhibitory
effects against human CMV (HCMV), HSV-1, and VZV.
Modified inhibitory tests of antiviral activity against HCMV
were compared with already acknowledged antiviral agents
ganciclovir, phosphonomethanoic acid, and the B-220 com-
pound. This comparison indicated 100% inhibition by the novel
compounds compared to 30, 50, and 20% inhibition of the
acknowledged antiviral agents, respectively. These preliminary
results are not conclusive as to what the underlying antiviral
* To whom correspondence should be addressed. Tel: +46 (0)8 608 92
04. Fax: +46 (0)8 608 15 01. E-mail: jan.bergman@biosci.ki.se.
†
Chalmers University of Technology.
Karolinska Institute.
Vironova AB.
‡
§
a Abbreviations: 11,11′-bidppz, 11,11′-bis(dipyrido[3,2-a:2′,3′-c]phena-
zinyl); B-220, 2,3 dimethyl-6(2-dimethylaminoethyl)6H-indolo-(2,3-b)qui-
noxaline (by Bergman- the 220th synthesis in this project); C4(cpdppz)2,
N,N′ bis(cpdppz)-1,4-diaminobutane; CMV, cytomegalovirus; cpdppz, 12-
cyano-12,13-dihydro-11H-8-cyclopenta[b]dipyrido[3,2-h:2′,3′-j]phenazine-
12-carbonyl; ct-DNA, calf thymus DNA; HCMV, human cytomegalovirus;
HSV-1, herpes simplex type 1; LD, linear dichroism; LDr, reduced linear
dichroism; MQ-water, MilliQ water; MS, multiple sclerosis; phen, 1,10-
phenanthroline; poly(dA-dT)2, duplex of two alternating adenine-thymine
polynucleotides; RA, rheumatoid arthritis; SAR, structure-activity relation-
ship; SDS, sodium dodecyl sulfate; VZV, varicella-zoster virus; YO, oxazole
yellow; YOYO, homodimeric derivative of oxazole yellow (YO).
10.1021/jm800787b CCC: $40.75
2008 American Chemical Society
Published on Web 11/21/2008

AntiViral Indolo[2,3-b]quinoxaline DeriVatiVes
Journal of Medicinal Chemistry, 2008, Vol. 51, No. 24 7745
Table 1. Extinction Coefficients of the Indoloquinoxaline Derivatives
(4-8) Measured in MQ-Water at Room Temperature
drug
λmax (nm)
ε (M-1 cm-1
)
4
5
6
7
8
363
374
363
345
345
17500
15000
19000
21500
21000
21500 M-1 cm-1. The low energy absorption maxima range
from 345 to 374 nm with the monomers at slightly longer
wavelengths (363-374 nm) than the dimers (345 nm). It is
worth noting that the monomers containing exactly the same
chromophoric units, 4 and 6, as expected have very similar
wavelength maxima and extinction coefficients. The slightly
lower extinction coefficient and absorption at longer wavelengths
for monomer 5 compared to those of 4 and 6 can be explained
by the addition of the amino-group in the 9-position of the
indoloquinoxaline system. Also, it is worth noting that the
extinction coefficients and wavelength maxima for 4-6 show
Figure 2. Novel monomeric and dimeric indoloquinoxaline derivatives
synthesized and used in this study: compounds 4, 2,3-dimethyl-6-(2-
dimethylaminoethyl)-6H-indolo[2,3-b]quinoxaline hydrochloride; 5,
9-amino-2,3-dimethyl-6-(2-dimethylaminoethyl)-6H-indolo[2,3-b]qui-
noxaline hydrochloride; 6, 2,3-dimethyl-6-[2-(trimethylamino)ethyl]-
6H-indolo[2,3-b]quinoxaline iodide; 7, bis-indolo-(2,3-b)quinoxaline-
6H-(3,3;10,10-tetramethyl)-3,10-di-(3,10)-azadodecane bromide; and 8,
bis-indolo-(2,3-b)quinoxaline-6H-(3,3;7,7-tetramethyl)-3,10-di-(3,7)-
azaheptane bromide.
mechanism is, although it appears to involve the impairment
of the tegument protein that binds to the viral capsid. The novel
compounds did not show any toxicity as assayed by propidium
iodide staining of cell cultures of infected and uninfected human
lung fibroblasts.34 Obviously, further tests were considered
necessary to comprehend the surprisingly different viral inhibi-
tion results especially exhibited between the novel compounds
and B-220.
Here, we perform extensive spectroscopic studies to inves-
tigate the interaction between the novel indoloquinoxaline
derivatives (4-8) and DNA by determining the binding constant,
the binding geometry, and the AT-specificity. The variation in
the structure, such as in the linker length and its position, the
permanent/nonpermanent quaternary nitrogen salt and the in-
troduction of new substituent groups on the indoloquinoxaline
moiety, are shown to have an immediate significant influence
on the interaction between the ligands and DNA. These
structure-activity relationships (SAR) give important under-
standing in the optimization of the antiviral effects of these
promising drugs.
Results and Discussion
Table 1 shows the wavelength of maximum absorption of
the low energy band and the corresponding extinction coefficient
of the indoloquinoxaline derivatives (4-8) measured in MQ-
water. Furthermore, the appearance of the lowest energy part
of the absorption spectra of the indoloquinoxaline derivatives
can be seen as the solid lines in Figure 3. The extinction
coefficients are all reasonably similar, ranging from 15000 to
Figure 3. Absorption spectra of (a) 4, (b) 5, (c) 6, (d) 7, and (e) 8
in the presence (dashed lines) and absence (solid lines) of calf
thymus DNA. Concentration of calf thymus DNA is 160 µM ([base])
and those of the monomeric and dimeric indoloquinoxaline deriva-
tives 10 and 5 µM, respectively. Measurements performed at room
temperature in a 10 mM phosphate buffer, pH 7.5 (total Na+-
concentration 100 mM).

7746 Journal of Medicinal Chemistry, 2008, Vol. 51, No. 24
Wilhelmsson et al.
Table 2. Equilibrium Constants of the Indoloquinoxaline Derivatives
(4-8) Binding to poly(dA-dT)2 and Ratio in Binding Constants between
poly(dA-dT)2 and Calf Thymus DNA in a Competitive Binding
Experimenta
drug
KAT
KAT/Kct-DNA
4
5
6
7
8
105-106
106-107b
106-107
108-109
108-109c
1.8
1.7
2.4
3.0
1.9
a Measurements were performed at 22°C in a 10 mM phosphate buffer,
pH 7.5 (total Na+-concentration 100 mM). b Because of the lack of
fluorescence of 5, its equilibrium constant of DNA binding was estimated
in a competitive binding experiment with 6. c The equilibrium constant of
DNA binding of 8 was calculated using a binding site size of 3 base pairs
(for the other dimer, 7, a site size of 4 was used).
great resemblance to previously reported values for molecules
built from the same kind of chromophore, even though those
measurements were performed in ethyl acetate.35 As expected,
also dimers 7 and 8 containing the same chromophoric unit have
very similar wavelength maxima and extinction coefficients. The
fact that 7 and 8, which have two chromophoric units that are
essentially identical to monomers 4 and 6, do not exhibit a 2-fold
increase in the extinction coefficient can be explained by
stacking of the two chromophores within the dimers. This
stacking results in an excitonic interaction between the two
chromophores of the dimeric compounds and hence a consider-
able change in spectral properties. Also the change in absorption
maxima from 363 nm in the monomers to 345 nm in the dimers
comes as a result of this excitonic interaction.
Table 2 shows the estimated equilibrium constants of binding
of 4-8 to poly(dA-dT)2 as well as the ratio between the binding
constant when the same drugs interacts with poly(dA-dT)2 and
ct-DNA. The binding constants of the monomeric indoloqui-
noxaline derivatives (4-6) to poly(dA-dT)2 are in the order of
106. Compound 6, which has exactly the same aromatic ring
system as 4, has a slightly higher equilibrium constant.
Consequently, the difference in the binding constant has to be
due to the higher hydrophobicity of 6 as an effect of the
additional methyl on the attached aminoethyl group. In the case
of compound 5, the slight increase in DNA binding constant
compared to that of compound 4 has to be ascribed to the
addition of the amino-group on the aromatic ring system. As
can be seen in Table 2, the binding constants of compounds
(7-8) to poly(dA-dT)2 are considerably higher (108-109). This
is also expected since these dimeric compounds have both
double the amount of functional groups and positive charges
compared to those of 4-6.
A general conclusion about the drugs that can be drawn from
the measurements is that they all prefer AT-DNA. As can be
seen in Table 2, the preference for poly(dA-dT)2 compared to
ct-DNA range from 1.8 to 3.0. Since ct-DNA consists of a
mixture of sites with different base composition and, thus,
different binding constants, and since the number of accessible
drug binding sites in ct-DNA compared to in poly(dA-dT)2 is
difficult to estimate, the actual values of AT-preference should
not be overinterpreted. However, as a consequence of the general
AT-preference among the five drugs and the magnitude of the
KAT/Kct-DNA, there is no doubt that all of the indoloquinoxaline
derivatives, like their parent compounds,36-38 prefer to bind to
AT-DNA (AT-regions).
Figure 4. Emission spectra of (a) 4, (b) 6, (c) 7, and (d) 8 in the
presence (dashed lines) and absence (solid lines) of poly(dA-dT)2.
Compound 5 is virtually nonfluorescent both in the presence and
absence of DNA and therefore not shown. Concentration of poly(dA-
dT)2 is 100 µM ([base]) and those of the monomeric and dimeric
indoloquinoxaline derivatives 10 and 5 µM, respectively. Measurements
performed at room temperature in a 10 mM phosphate buffer, pH 7.5
(total Na+-concentration 100 mM).
when drugs bind to DNA by either intercalation or groove
binding. The large hypochromic effect (14-34%) and red-shift
(6-19 nm) displayed by these drugs indicate intercalation rather
than groove binding. (We will show more distinctive evidence
of intercalation below.) The larger red-shift observed for the
dimers (Figure 3d-e) can be explained by the decrease in
excitonic effects as the dimers unstack upon binding to the DNA.
In Figure 4, the fluorescence spectra of 4, 6, 7, and 8 both in
the presence and absence of poly(dA-dT)2 are presented. As an
effect of the lack of emission both with and without DNA, the
result from the fluorescence measurements of compound 5 is
not shown in the figure. Since the only difference in the
chromophoric unit in 5 compared to that in 4 and 6 is the amino
group in the 9-position, this group has to be the reason for the
quenching of the emission in 5. As can be seen in Figure 4, the
four fluorescent drugs all have an intrinsic emission centered
at 515 nm in buffered solution (solid lines). Furthermore, upon
binding to DNA they all exhibit a considerable increase in
emission (dashed lines) as well as a blue-shift of more than 30
nm. Increase in emission upon binding to DNA is often
considered to be a good indication of intercalation of the drug
between the base pairs of the DNA, and thus, these measure-
ments further support an intercalative mode of binding.
Figure 5 shows the flow linear dichroism (LD) spectra of
the five drugs, 4-8, in the presence of ct-DNA. As can be seen
from the figure, the LD of the bound drugs is purely negative.
In the region where DNA has no influence on the signal (>300
nm), the LD spectra are virtually perfect mirror images of the
corresponding absorption spectra of the DNA-bound drugs (for
comparison see Figure 3). A negative LD proves that the angle
between the DNA helix axis and the planar chromophore unit(s)
of the drugs is larger than the magic angle (54.7°). When
calculating the LDr of the band originating from the five drugs
and subsequently the angles between the DNA helix axis and
the planar chromophore unit(s), they are all in the range between
70 and 90°. This is indeed good evidence for an intercalative
mode of binding for all five drugs. After determination of the
binding mode by LD, SDS was added to all six DNA-drug
solutions. The LD-signal of the drugs disappeared immediately,
leaving only a pure DNA-LD centered around 260 nm. This
Figure 3 shows the absorption spectra of 4-8 in the presence
and absence of ct-DNA. As can be seen in the figure, the general
trend is that the molecules show a considerable hypochromic
effect and a significant red-shift of the absorption maxima upon
addition of ct-DNA (dashed lines). These are common effects

AntiViral Indolo[2,3-b]quinoxaline DeriVatiVes
Journal of Medicinal Chemistry, 2008, Vol. 51, No. 24 7747
data obtained in this investigation it is not possible to directly
determine this number. Nevertheless, from the length of the
linkers and the overall structures of the dimers we estimate the
binding site size to be four and three basepairs for 7 and 8,
respectively. The smaller binding site size of 8 is further
supported by the fact that using this binding site size in the
calculation of the equilibrium constant (see Table 2) results in
a DNA binding constant similar to that of 7, which we expect
because of the high structural resemblance of the two dimers.
Finally, it should be noted that the high similarity in the results
for all derivatives presented here demonstrates that the structural
changes that we introduce to the indoloquinoxaline planar
system only have minor effects on the mode of binding of these
drugs.
The DNA-binding constant is a very important property for
an antiviral drug. In the results obtained here we find that the
monomeric indoloquinoxaline derivates (4-6) have DNA-
binding constants (AT sequence) of approximately 106. DNA
binding constants in the order of 106, at comparable or lower
ionic strengths and similar temperatures, have previously been
reported for familiar intercalators such as ethidium,43 YO,16
Figure 5. Flow linear dichroism spectra of 4 (black), 5 (red), 6 (green),
7 (blue), and 8 (cyan) in the presence of calf thymus DNA. The
concentration of calf thymus DNA is 160 µM ([base]) and those of the
monomeric and dimeric indoloquinoxaline derivatives 10 and 5 µM,
respectively. Measurements performed at room temperature in a 10
mM phosphate buffer, pH 7.5 (total Na+-concentration 100 mM).
actinomycin D,44 daunomycin,43 and [Ru(phen)2dppz]2+26
.
These common intercalators are successfully used or have been
proposed to be used, for example, as DNA stains or as antibiotics
or anticancer therapeutics, and consequently, the indoloqui-
noxaline monomers (4-6) should constitute very promising
DNA-binding drugs. For the dimeric indoloquinoxaline derivates
(7-8), we determined DNA binding constants of 108-109 (a
binding constant of 8 would be even higher if the same binding
site size as that for 7 was used; see Table 2). Previously, binding
constant estimates of other dimeric DNA-ligands such as YOYO
(K ≈1010-1012),16 [µ-C4(cpdppz)2(phen)4Ru2]4+ (K ≈109),28
and [µ-(11,11′-bidppz)(phen)4Ru2]4+ (K ≈1012)29 as well as for
larger DNA-ligands such as distamycin (K ≈109)43 and netropsin
(K ≈109)43 have been performed at comparable or lower ionic
strengths and similar temperatures. When comparing these
values to the ones obtained here for 7 and 8, one has to
remember that for the previously reported dimers the charge is
two times higher and that for one of them, [µ-(11,11′-
Figure 6. Left: 6H-indolo[2,3-b]quinoxaline, 2, and 2,3-dimethyl-6H-
indolo[2,3-b]quinoxaline, 2a. Right: 6-(2-dimethylaminoethyl)-6H-
indolo[2,3-b]quinoxaline, 3, and 2,3-dimethyl-6-(2-dimethylaminoethyl)-
6H-indolo[2,3-b]quinoxaline, 3a.
confirms the noncovalent nature of the DNA interaction of these
drugs and indicates a fast DNA dissociation process.
The very promising antiviral effect exhibited by some of the
indoloquinoxaline derivatives presented here makes them
interesting to investigate in terms of their DNA binding
properties. Moreover, it is important to understand the
structure-activity relationship (SAR) within this set of indolo-
quinoxalines not only for the current derivatives and possible
use of them as drugs but also for their future structural
modification and further optimization.
bidppz)(phen)4Ru2]4+ 29
, the measurements are performed at only
10 mM Na+. Taking this into consideration, both 7 and 8 have
a remarkably high binding constant to DNA, which could
explain their high antiviral activity compared to the already
acknowledged ganciclovir, phosphonomethanoic acid, and the
B-220 compound.
In conclusion, the novel indoloquinoxaline derivatives (4-8)
investigated here comprise a group of very promising antiviral
agents that bind strongly but noncovalently to DNA in an
intercalative mode. They are found to have equally high binding
constants as already established DNA drugs and dyes and show
a preference for AT-regions. The AT-specificity, a property
shared with some of the DNA drugs and dyes mentioned above,
is potentially useful for targeting viral genomes that are
especially AT-rich.
The results obtained from absorption, fluorescence, and LD
measurements all give evidence of a mode of binding where
the planar aromatic parts of the indoloquinoxaline derivates are
noncovalently intercalated between the basepairs of the DNA.
This mode of binding has previously been reported for the parent
compounds (9-OH-)B-22039,40 and is very common for posi-
tively charged planar aromatic molecules such as oxazole yellow
(YO)15-17 and ethidium.41 For the dimers (7-8), this means
that they are bis-intercalators with the interconnecting linker
ending up in one of the grooves of the DNA. From the present
measurements, we cannot conclude whether the linkers are
situated in the minor or major groove, or if they are equally
distributed in the two grooves. For the structurally related bis-
intercalator YOYO as well as for the naturally occurring bis-
intercalator echinomycin, however, the linker interconnecting
the monomeric units has been found to reside in the minor
groove,19,42 suggesting that the linkers of our indoloquinoxaline
dimers may have a minor groove preference. Also, earlier NMR
studies on related compounds suggest that the side chain resides
in the minor groove.36 For the dimers (7-8), it is also important
to identify the size of the DNA binding site. However, with the
Experimental Section
Chemistry. Indolo[2,3-b]quinoxaline (2, Figure 6) and its
derivative (2a, Figure 6) were easily prepared by condensation of
isatin and the appropriate 1,2-phenylenediamine in glacial acetic
acid according to Schunck’s and Marchlewski’s method.45,46
Alkylation of these compounds was achieved when added to a
solution containing 1-chloro-2-dimethylaminoethane hydrochloride
and K2CO3 in acetone and refluxed for 24 h. Compound 4 was
obtained by dissolving the free base (3a, Figure 6) in acetone, and
precipitation was induced by introduction of concentrated HCl. The

7748 Journal of Medicinal Chemistry, 2008, Vol. 51, No. 24
Wilhelmsson et al.
amino compound 5 was prepared via a two-step synthesis starting
with addition of one equivalent of potassium nitrate in sulfuric acid
to compound 3a (Figure 6) to give 2,3-dimethyl-6-(2-dimethylami-
noethyl-9-nitro-6H-indolo[2,3-b]quinoxaline. Reduction of the nitro
group was achieved by catalytic hydrogenation of a suspension in
N,N-dimethylacetamide using Pd/C under hydrogen pressure (2.7
atm) for 24 h. Finally, hydrochloric acid was added to a warm
solution of 9-amino-2,3-dimethyl-6-(2-dimethylaminoethyl)-6H-
indolo[2,3-b]quinoxaline in water, until it reached a pH of 5.
Compound 6 was prepared by refluxing compound 3a (Figure 6)
in acetonitrile with the addition of methyl iodide. The general
procedure for the synthesis of compounds 7 and 8 comprised the
reflux of two equivalents of compound 3 (Figure 6) and one
equivalent of a suitable R,ω-dibromoalkane in acetonitrile.
Melting points were determined on a Leica Kofler hot stage and
are uncorrected. NMR data were recorded on a Bruker DPX at
300.1 MHz for 1H and 75.5 MHz for 13C. IR spectra were acquired
on a Perkin-Elmer FT-IR 1600 spectrophotometer. Elemental
analyses were performed by Mikroanalytisches Laboratorium
KOLBE, Germany and were within ( 0.4% of the calculated values.
General Procedure for Bis-quaternary Bromides.47 6-[2-
(Dimethylamino)ethyl]-6H-indolo[2,3-b]quinoxaline was prepared
according to the general procedure described for compound 3. The
free base (2 eq.) was added to a solution of the appropriate R,ω-
dibromohexane (1 eq.) in acetonitrile and refluxed for 24 h. The
solid thus formed was isolated by filtration.
Extinction Coefficient Determination. The extinction coefficient
of the indoloquinoxaline derivatives were determined by measuring
the absorption of samples of known concentration. Samples were
prepared by weighing out small amounts of the indoloquinoxaline
derivatives, typically 2 mg, and dissolving them in known volumes
of MQ water. Several absorption spectra of each sample were
recorded on a Varian Cary 4000 spectrophotometer at room
temperature and an average of the measurements was calculated.
The extinction coefficients were determined as an average of
between three and five independent sample preparations.
Fluorescence. Emission spectra were recorded at room temper-
ature on a Xenon lamp equipped SPEX fluorolog 3 spectrofluo-
rimeter (JY Horiba) between 365 and 750 nm using an excitation
wavelength of 360 nm. Concentrations of the indoloquinoxaline
derivatives were kept low in order to avoid inner filter effects.
Flow Linear Dichroism. Linear dichroism (LD)48 is defined as
the difference in absorbance of linearly polarized light parallel and
perpendicular to a macroscopic orientation axis (here the flow
direction):
LD(λ) ) AII(λ) - A (λ)
(1)
The reduced linear dichroism LDr is calculated as
LDr(λ) ) LD(λ)/Aiso
(2)
where Aiso represents the absorbance of the same isotropic sample.
The LDr of a single electronic transition, i,
Bis-indolo-(2,3-b)quinoxaline-6H-(3,3;10,10-tetramethyl)-3,10-
di-(3,10)-azadodecane Bromide (7). Yield 72%; mp 260-262 °C;
1H NMR (DMSO-d6) δ 8.42 (d, 1H, J ) 7.6 Hz), 8.28 (d, 1H, J )
8.3 Hz), 8.09 (d, 1H, J ) 8.4 Hz), 7.96 (d, 1H, J ) 8.1 Hz),
7.86-7.83 (m, 2H), 7.76-7.74 (m, 1H), 7.49 (t, 1H, J ) 7.5 Hz),
4.98 (br t, 2H, J ) 6.8 Hz), 3.85 (br t, 2H, J ) 6.8 Hz), 3.53-3.47
(m, 2H), 3.25 (s, 6H), 1.78 (br s, 2H), 1.30 (br s, 2H); 13C NMR
(DMSO-d6) δ 144.8 (s), 143.6 (s), 139.8 (s), 139.6 (s), 139.0 (s),
131.6 (d), 129.3 (d), 129.2 (d), 127.4 (d), 126.6 (d), 122.5 (d), 121.8
(d), 119.1 (s), 110.7 (d), 63.2 (t), 59.2 (t), 50.8 (q), 35.0 (t), 25.4
(t), 21.7 (t). IR (neat) υmax: 3406, 3004, 2933, 1610, 1469, 1413,
1201, 1120, 769, 752 cm-1. Anal. Calcd for (C42H48Br2N8 ·1/4H2O)
C, H, N.
3
2
LDir ) S(3cos2 Ri - 1)
(3)
is related to the angle Ri between the ith transition moment direction
and the molecular orientation axis, in this case the DNA helix axis.
Samples with indoloquinoxaline derivatives and DNA were oriented
in a Couette flow cell with an outer rotating cylinder at a shear
gradient of approximately 3000 s-1. LD spectra were measured
on a Jasco J-720 CD spectropolarimeter equipped with an Oxley
prism to obtain linearly polarized light. All spectra were recorded
between 200 and 650 nm and baseline-corrected by subtracting the
spectrum recorded for the nonoriented sample.
Equilibrium Constant Determination. Buffered solutions
containing 0, 25, 100, and 250 µM poly(dA-dT)2 as well as 10 or
5 µM of the fluorescent monomeric or dimeric indoloquinoxaline
derivatives 4, 6, 7, and 8, respectively, were prepared. Both
absorption (measured on a Varian Cary 4000 spectrophotometer
between 200 and 650 nm) and fluorescence spectra (for details see
above) of the solutions were recorded at 22 °C. Hereafter, the
solutions were diluted 10 and 100 times, and their absorption and
fluorescence spectra once again recorded. The fraction of bound
molecules was estimated from the measured spectra and subse-
quently used to calculate the equilibrium constants of the DNA
binding. When calculating the equilibrium binding constants, the
sizes of the DNA binding sites were assumed to be 2, 2, 4, and 3
base pairs for 4, 6, 7, and 8, respectively. The DNA binding site
size for dimer 8 is assumed to be smaller due to the shorter linker
between the two planar aromatic ring systems compared to that of
dimer 7.
For the virtually nonfluorescent compound 5, competitive DNA
binding with compound 6 was used to estimate the binding constant.
Two buffered solutions containing 6.25 µM poly(dA-dT)2 as well
as 2.5 or 3.75 µM of 6 were prepared. Also, two solutions
containing the same amount of poly(dA-dT)2 and half the amount
of 6 compared to the two above as well as 1.25 or 1.875 µM of 5,
respectively, were prepared. The fluorescence of DNA-bound 6
from the four samples was measured. The changes in emission as
an effect of compound 5 competing with compound 6 about the
DNA binding sites in combination with the previously determined
DNA binding constant of 6 were then used to estimate the
equilibrium binding constant of the nonfluorescent 5 when binding
to DNA.
Bis-indolo-(2,3-b)quinoxaline-6H-(3,3;7,7-tetramethyl)-3,10-
di-(3,7)-azaheptane Bromide (8). Yield 68%; mp 258-260 °C;
1H NMR(DMSO-d6) δ 8.38 (d, 1H, J ) 7.5 Hz), 8.25-8.24 (m,
1H), 8.08-8.06 (m, 1H), 7.98 (d, 1H, J ) 8.2 Hz), 7.81-7.73 (m,
2H), 7.46 (t, 1H, J ) 7.4 Hz), 5.03 (br t, 2H, J ) 7.0 Hz), 3.96 (br
t, 2H, J ) 7.1 Hz), 3.72 (br t, 2H, J ) 7.5 Hz), 3.38 (s, 6H), 2.45
(br s, 2H); 13C NMR (DMSO-d6) δ 144.7 (s), 143.5 (s), 139.7 (s),
139.5 (s), 138.9 (s), 131.5 (d), 129.2 (d), 127.4 (d), 126.5 (d), 122.4
(d), 121.7 (d), 119.0 (s), 110.7 (d), 60.5 (t), 59.9 (t), 50.9 (q), 35.0
(t), 16.9 (t). IR (neat) υmax: 3349, 3004, 1607, 1467, 1415, 1116,
938, 769, 752 cm-1. Anal. Calcd for (C39H42Br2N8 ·1/4H2O) C, H,
N.
Chemicals for DNA-Binding Studies. Calf thymus (ct) DNA,
obtained from Sigma, was dissolved in buffer and filtered twice
through a 0.8 µm Millipore filter before use. The polynucleotide,
poly(dA-dT)2, was purchased from Amersham Biosciences and used
as obtained. All experiments were performed in aqueous buffer (10
mM phosphate buffer at pH 7.5 and NaCl added to a total Na+-
concentration of 100 mM). Sodium dodecyl sulfate (SDS) was
purchased from Sigma-Aldrich. The indoloquinoxaline derivatives
were synthesized as described above.
Sample Preparation. Samples were prepared by mixing equal
volumes of indoloquinoxaline derivatives and DNA dissolved in
buffer. Unless otherwise stated the sample concentrations were 10
µM and 5 µM of the monomeric and dimeric indoloquinoxaline
derivatives, respectively, and 160 µM ([base]) of DNA. Concentra-
tions were determined on a Varian Cary 4B spectrophotometer.
The extinction coefficients used for the DNAs were ε260nm ) 6600
M-1cm-1 for ct-DNA and ε262nm ) 6600 M-1cm-1 for poly(dA-
dT)2. Extinction coefficient used for the indoloquinoxaline deriva-
tives were determined as described below.

AntiViral Indolo[2,3-b]quinoxaline DeriVatiVes
Journal of Medicinal Chemistry, 2008, Vol. 51, No. 24 7749
complexes with 2-color laser-excited confocal fluorescence gel scanner.
Methods Enzymol. 1993, 217, 414–431.
(19) Johansen, F.; Jacobsen, J. P. H-1 NMR studies of the bis-intercalation
of a homodimeric oxazole yellow dye in DNA oligonucleotides.
J. Biomol. Struct. Dyn. 1998, 16, 205–222.
(20) Spielmann, H. P.; Wemmer, D. E.; Jacobsen, J. P. Solution structure
of a DNA complex with the fluorescent bis-intercalator toto determined
by NMR-spectroscopy. Biochemistry 1995, 34, 8542–8553.
(21) Hu, G. G.; Shui, X. Q.; Leng, F. F.; Priebe, W.; Chaires, J. B.;
Williams, L. D. Structure of a DNA-bisdaunomycin complex. Bio-
chemistry 1997, 36, 5940–5946.
(22) Robinson, H.; Priebe, W.; Chaires, J. B.; Wang, A. H. J. Binding of
two novel bisdaunorubicins to DNA studied by NMR spectroscopy.
Biochemistry 1997, 36, 8663–8670.
(23) Leng, F. F.; Priebe, W.; Chaires, J. B. Ultratight DNA binding of a
new bisintercalating anthracycline antibiotic. Biochemistry 1998, 37,
1743–1753.
(24) Friedman, A. E.; Kumar, C. V.; Turro, N. J.; Barton, J. K.
Luminescence of ruthenium(II) polypyridyls: evidence for intercalative
binding to Z-DNA. Nucleic Acids Res. 1991, 19, 2595–2602.
(25) Lincoln, P.; Broo, A.; Norde´n, B. Diastereomeric DNA-binding
geometries of intercalated ruthenium(II) trischelates probed by linear
dichroism: [Ru(phen)dppz]2+ and [Ru(phen)2bdppz]2+. J. Am. Chem.
Soc. 1996, 118, 2644–2653.
The difference in the equilibrium constant of binding of the
various indoloquinoxaline derivatives to the two forms of DNA,
ct-DNA and poly(dA-dT)2, was measured using flow LD. The LD
of each indoloquinoxaline derivative (5 and 10 µM of the dimeric
and monomeric ones, respectively) bound to ct-DNA (160 µM)
was recorded and compared to the LD of the same molecule when
mixed into a solution of a combination of ct-DNA (80 µM) and
poly(dA-dT)2 (80 µM). As a result of the short polynucleotide
strands (699 base pairs), the measured LD of the indoloquinoxaline
derivatives bound to poly(dA-dT)2 (160 µM) has virtually no
intensity.
Supporting Information Available: Details of the preparation
1
of compounds 3, 3a, and 4-6 as well as H NMR and 13C NMR
spectra of compounds 3, 3a, and 4-8; elemental analysis results
of target compounds 4-8. This material is available free of charge
via the Internet at http://pubs.acs.org.
References
(1) Goodwin, S.; Smith, A. F.; Horning, E. C. Alkaloids of Ochrosia
elliptica Labill. J. Am. Chem. Soc. 1959, 81, 1903–1908.
(2) Juret, P.; Tanguy, A.; Girard, A.; Letalaer, J. Y.; Abbatucci, J. S.;
Datxuong, N.; LePecq, J. B.; Paoletti, C. Preliminary trial of
9-hydroxy-2-methyl ellipticinium (Nsc 264-137) in advanced human
cancers. Eur. J. Cancer. 1978, 14, 205–206.
(26) Haq, I.; Lincoln, P.; Suh, D. C.; Norde´n, B.; Chowdhry, B. Z.; Chaires,
J. B. Interaction of ∆-[Ru(phen)2dppz]2+ and λ-[Ru(phen)2dppz]2+
with DNA: a calorimetric and equilibrium binding study. J. Am. Chem.
Soc. 1995, 117, 4788–4796.
¨
(27) Onfelt, B.; Lincoln, P.; Norde´n, B. A molecular staple for DNA:
(3) Bergman, J.; Åkerfeldt, S. G. Substituted Indoloquinoxalines. Patent
WO 87/04436, 1987.
threading bis-intercalating [Ru(phen)2dppz]2+ Dimer. J. Am. Chem.
Soc. 1999, 121, 10846–10847.
(4) Wamberg, M. C.; Hassan, A. A.; Bond, A. D.; Pedersen, E. B.
Intercalating nucleic acids (INAs) containing insertions of 6H-
indolo[2,3-b]quinoxaline. Tetrahedron 2006, 62, 11187–11199.
(5) Harmenberg, J.; Wahren, B.; Bergman, J.; Åkerfeldt, S.; Lundblad,
L. Antiherpesvirus activity and mechanism of action of indolo-(2,3-
b)quinoxaline and analogs. Antimicrob. Agents Chemother. 1988, 32,
1720–1724.
¨
(28) Onfelt, B.; Lincoln, P.; Norde´n, B. Enantioselective DNA threading
dynamics by phenazine-linked [Ru(phen) 2dppz]2+ dimers. J. Am.
Chem. Soc. 2001, 123, 3630–3637.
(29) Lincoln, P.; Norde´n, B. Binuclear ruthenium(II)phenanthroline com-
pounds with extreme binding affinity for DNA. Chem. Communication.
1996, 214, 5–2146.
(6) Harmenberg, J.; Åkesson-Johansson, A.; Gra¨slund, A.; Malmfors, T.;
Bergman, J.; Wahren, B.; Åkerfeldt, S.; Lundblad, L.; Cox, S. The
mechanism of action of the anti-herpes virus compound 2,3-dimethyl-
6(2-dimethylaminoethyl)-6H-indolo-(2,3-b)quinoxaline. AntiViral Res.
1991, 15, 193–204.
(7) Bjo¨rklund, U.; Engqvist, R.; Kihlstro¨m, I.; Gerdin, B.; Bergman, J.
Alkyl Substituted Indoloquinoxalines. Patent WO 2005/123741, 2005.
(8) Crenshaw, J. M.; Graves, D. E.; Denny, W. A. Interactions of acridine
antitumor agents with DNA: binding-energies and groove preferences.
Biochemistry. 1995, 34, 13682–13687.
(9) Petty, J. T.; Bordelon, J. A.; Robertson, M. E. 2000). Thermodynamic
characterization of the association of cyanine dyes with DNA. J. Phys.
Chem. B 2000, 104, 7221–7227.
(10) Frodl, A.; Herebian, D.; Sheldrick, W. S. Coligand tuning of the DNA
binding properties of bioorganometallic (eta(6)-arene)ruthenium(II)
complexes of the type [(η6-arene) Ru(amino acid)(dppz)]n+ (dppz )
dipyrido[3,2-a: 2′,3′-c] phenazine), n)1-3. J. Chem. Soc., Dalton
Trans. 2002, 3664–3673.
(11) White, E. W.; Tanious, F.; Ismail, M. A.; Reszka, A. P.; Neidle, S.;
Boykin, D. W.; Wilson, W. D. Structure-specific recognition of
quadruplex DNA by organic cations: Influence of shape, substituents
and charge. Biophys. Chem. 2007, 126, 140–153.
(12) Capelle, N.; Barbet, J.; Dessen, P.; Blanquet, S.; Roques, B. P.; Lepecq,
J. B. Deoxyribonucleic-acid bifunctional intercalators: kinetic inves-
tigation of the binding of several acridine dimers to deoxyribonucleic-
acid. Biochemistry. 1979, 18, 3354–3362.
(13) Gallego, J.; Reid, B. R. Solution structure and dynamics of a complex
between DNA and the antitumor bisnaphthalimide LU-79553: Inter-
calated ring flipping on the millisecond time scale. Biochemistry. 1999,
38, 15104–15115.
(14) Carlsson, C.; Larsson, A.; Jonsson, M.; Albinsson, B.; Norde´n, B.
Optical and photophysical properties of the oxazole yellow DNA
probes Yo and YoYo. J. Phys. Chem. 1994, 98, 10313–10321.
(15) Larsson, A.; Carlsson, C.; Jonsson, M.; Albinsson, N. Characterization
of the binding of the fluorescent dyes Yo and YoYo to DNA by polarized-
light spectroscopy. J. Am. Chem. Soc. 1994, 116, 8459–8465.
(16) Glazer, A. N.; Rye, H. S. Stable dye-DNA intercalation complexes as
reagents for high-sensitivity fluorescence detection. Nature. 1992, 359,
859–861.
(30) Wilhelmsson, L. M.; Westerlund, F.; Lincoln, P.; Norde´n, B. DNA-
binding of semirigid binuclear ruthenium complex ∆,∆-[µ-(11,11′-
bidppz)(phen) 4Ru-2]4+: Extremely slow intercalation kinetics. J. Am.
Chem. Soc. 2002, 124, 12092–12093.
(31) Wilhelmsson, L. M.; Esbjo¨rner, E. K.; Westerlund, F.; Norde´n, B.;
Lincoln, P. Meso stereoisomer as a probe of enantioselective threading
intercalation of semirigid ruthenium complex µ-(11,11′-bidppz)
(phen) 4Ru-24+. J. Phys. Chem. B 2003, 107, 11784–11793.
(32) Nordell, P.; Westerlund, F.; Wilhelmsson, L. M.; Norde´n, B.; Lincoln,
P. Kinetic recognition of AT-rich DNA by ruthenium complexes.
Angew. Chem., Int. Ed. 2007, 46, 2203–2206.
(33) Karlsson, H. J.; Bergqvist, M. H.; Lincoln, P.; Westman, G. Syntheses
and DNA-binding studies of a series of unsymmetrical cyanine dyes:
structural influence on the degree of minor groove binding to natural
DNA. Bioorg. Med. Chem. 2004, 12, 2369–2384.
(34) Homman, M.; Engqvist, R.; So¨derberg-Naucler, C.; Bergman, J. Novel
Compounds and Use Thereof. Patent WO 2007/084073 A1, 2007.
(35) Przyjazna, B.; Kucybala, Z.; Paczkowski, J. P. Development of new
dyeing photoinitiators based on 6H-indolo[2,3-b]quinoxaline skeleton.
Polymer 2004, 45, 2559–2566.
(36) Behravan, G.; Leijon, M.; Sehlstedt, U.; Norde´n, B.; Vallberg, H.;
Bergman, J.; Gra¨slund, A. The Interaction of ellipticine derivatives with
nucleic-acids studied by optical and H-1-NMR spectroscopy: effect of
size of the heterocyclic ring. Biopolymers 1994, 34, 599–609.
(37) Zegar, I.; Gra¨slund, A.; Bergman, J.; Eriksson, M.; Norde´n, B.
Interaction of ellipticine and an indolo[2,3b]-quinoxaline derivative
with DNA and synthetic polynucleotides. Chem.-Biol. Interact. 1989,
72, 277–293.
(38) Patel, N.; Bergman, J.; Gra¨slund, A. H-1-NMR studies of the
interaction between a self-complementary deoxyoligonucleotide duplex
and indolo[2,3-b]quinoxaline derivatives active against herpes-virus.
Eur. J. Biochem. 1991, 197, 597–604.
(39) Sehlstedt, U.; Aich, P.; Bergman, J.; Vallberg, H.; Norde´n, B.;
Gra¨slund, A. Interactions of the antiviral quinoxaline derivative 9-OH-
B220 {2,3-dimethyl-6-(dimethylaminoethyl)-9-hydroxy-6H-indolo-
[2,3-b]quinoxaline} with duplex and triplex forms of synthetic DNA
and RNA. J. Mol. Biol. 1998, 278, 31–56.
(40) Sandstro¨m, K.; Wa¨rmla¨nder, S.; Bergman, J.; Engqvist, R.; Leijon,
M.; Gra¨slund, A. The influence of intercalator binding on DNA triplex
stability: correlation with effects on A-tract duplex structure. J. Mol.
Recognit. 2004, 17, 277–285.
(17) Rye, H. S.; Yue, S.; Wemmer, D. E.; Quesada, M. A.; Haugland, R. P.;
Mathies, R. A.; Glazer, A. N. Stable fluorescent complexes of double-
stranded DNA with bis- intercalating asymmetric cyanine dyes:
properties and applications. Nucleic Acids Res. 1992, 20, 2803–2812.
(18) Rye, H. S.; Yue, S.; Quesada, M. A.; Haugland, R. P.; Mathies, R. A.;
Glazer, A. N. Picogram detection of stable dye DNA intercalation
(41) Fuller, W.; Waring, M. J. A molecular model for the interaction of
ethidium bromide with deoxyribonucleic acid. Ber. Bunsen-Ges. Phys.
Chem. 1964, 68, 805–808.

7750 Journal of Medicinal Chemistry, 2008, Vol. 51, No. 24
Wilhelmsson et al.
(42) Gilbert, D. E.; Feigon, J. (1991). The DNA-sequence at echinomycin
binding-sites determines the structural-changes induced by drug-
binding: NMR-studies of echinomycin binding to [d(ACG-
TACGT)]2 and [d(TCGATCGA)]2. Biochemistry. 1991, 30, 2483-
2494.
(45) Schunk, E.; Marchlewski, L. Zur Kenntniss der rothen Isomeren des
Indigotins und u¨ber einige Derivate des Isatins. Chem. Ber. 1895, 28,
2525–2531.
(46) Buraczewski, J.; Marchlewski, L. Zur Kenntniss des Isatins. Chem.
Ber. 1901, 34, 4008–4015.
(43) Breslauer, K. J.; Remeta, D. P.; Chou, W. Y.; Ferrante, R.; Curry, J.;
Zaunczkowski, D.; Snyder, J. G.; Marky, L. A. Enthalpy entropy
compensations in drug DNA-binding studies. Proc. Natl. Acad. Sci.
U.S.A. 1987, 84, 8922–8926.
(44) Marky, L. A.; Snyder, J. G.; Remeta, D. P.; Breslauer, K. J.
Thermodynamics of drug-DNA interactions. J. Biomol. Struct. Dyn.
1983, 1, 487–507.
(47) Bergman, J. Formation of N-N bonds by thermolysis of 5-(2-
dimethylaminoethyl)-5H-indolo[2,3-b]quinoxaline. Tetrahedron Lett.
1989, 30, 1837–1840.
(48) Norde´n, B.; Kubista, M.; Kurucsev, T. Linear dichroism spectroscopy
of nucleic-acids. Q. ReV. Biophys. 1992, 25, 51–170.
JM800787B