Syntheses and DNA-binding studies of a series of unsymmetrical cyanine dyes: Structural influence on the degree of minor groove binding to natural DNA

Article abstract of DOI:10.1016/j.bmc.2004.02.006

Twelve crescent-shaped unsymmetrical dyes have been synthesized and their interactions with DNA have been investigated by spectroscopic methods. A new facile synthetic route to this type of cyanine dyes has been developed, involving the preparation of 6-substituted 2-thiomethyl-benzothiazoles in good yields. The new dyes are analogues to the minor groove binding unsymmetrical cyanine dye, BEBO, recently reported by us. In this dye, the structure of the known intercalating cyanine dye BO was extended with a 6-methylbenzothiazole substituent. Herein we further investigate the role of the extending benzazole heterocycle, as well as of the pyridine or quinoline moiety of the cyanine chromophore, for the binding mode of these crescent-shaped dyes to calf thymus DNA. Flow LD and CD studies of the 12 dyes show that the extent of minor groove binding to mixed sequence DNA varies significantly between the dyes. We find that hydrophobicity and size are the crucial parameters for recognition of the minor groove. The relatively high fluorescence quantum yield of many of these cyanines bound to DNA, combined with their absorption at long wavelengths, may render them useful in biological applications. In particular, two of the benzoxazole containing dyes BOXTO and 2-BOXTO show a high degree of minor groove binding and quantum yields of 0.52 and 0.32, respectively, when bound to DNA.

Full text of DOI:10.1016/j.bmc.2004.02.006

Bioorganic & Medicinal Chemistry 12 (2004) 2369–2384
Syntheses and DNA-binding studies of a series of unsymmetrical
cyanine dyes: structural influence on the degree of minor groove
binding to natural DNA
H. Jonas Karlsson,* Mattias H. Bergqvist, Per Lincoln and Gunnar Westman*
Department of Chemistry and Bioscience, Chalmers University of Technology, S-412 96 G€oteborg, Sweden
Received 3 September 2003; accepted 3 February 2004
Abstract—Twelve crescent-shaped unsymmetrical dyes have been synthesized and their interactions with DNA have been investi-
gated by spectroscopic methods. A new facile synthetic route to this type of cyanine dyes has been developed, involving the
preparation of 6-substituted 2-thiomethyl-benzothiazoles in good yields. The new dyes are analogues to the minor groove binding
unsymmetrical cyanine dye, BEBO, recently reported by us. In this dye, the structure of the known intercalating cyanine dye BO was
extended with a 6-methylbenzothiazole substituent. Herein we further investigate the role of the extending benzazole heterocycle, as
well as of the pyridine or quinoline moiety of the cyanine chromophore, for the binding mode of these crescent-shaped dyes to calf
thymus DNA. Flow LD and CD studies of the 12 dyes show that the extent of minor groove binding to mixed sequence DNA varies
significantly between the dyes. We find that hydrophobicity and size are the crucial parameters for recognition of the minor groove.
The relatively high fluorescence quantum yield of many of these cyanines bound to DNA, combined with their absorption at long
wavelengths, may render them useful in biological applications. In particular, two of the benzoxazole containing dyes BOXTO and
2-BOXTO show a high degree of minor groove binding and quantum yields of 0.52 and 0.32, respectively, when bound to DNA.
Ó 2004 Elsevier Ltd. All rights reserved.
1. Introduction
Recently, we have developed a series of unsymmetrical
cyanine dyes that bind in the minor groove of DNA
instead of by intercalation (Fig. 1).13;14 These dyes have
the advantage of large fluorescence quantum yields and
absorption at long wavelengths, typical of intercalating
cyanine dyes. In addition, their binding in the minor
groove is not expected to significantly lengthen the
DNA helix, similarly to classical groove-binding DNA
stains such as Hoechst 33258 (Hoechst)15;16 and
DAPI.1;17 Indeed, electrophoresis unwinding assays
show that these cyanines do not unwind natural DNA.14
In agreement with other minor groove binders they have
a preference for AT sequences. Whereas Hoechst and
DAPI absorb in the near-UV region,17;18 where back-
ground from scattering can be a problem, the absor-
bance maxima of the groove-bound cyanine dyes are
between 465 and 515 nm. A potential disadvantage of
the intercalating cyanine dyes is that the insertion of the
dye molecules between the base pairs leads to an
extension of the DNA helix.19–21 This might be unde-
sirable in for example studies of hydrodynamic and
flexibility properties of DNA in solution, where the
polymer length is the crucial variable.22 Moreover, it
has been reported that the minor groove binder DAPI
only exhibits a small effect on the DNA conformations
Dyes that bind to DNA have received considerable
attention due to their potential as anti-tumor drugs and
as tools to study and visualize DNA.1 Intercalating
unsymmetrical cyanine dyes, for example, Thiazole
Orange (TO) (Fig. 1) and its homodimeric counterpart
TOTO, are a widely used class of compounds of the
latter category.2 Upon binding to nucleic acids these
dyes exhibits a dramatic enhancement in fluorescence
intensity. The increase in fluorescence is believed to arise
when the rotation or torsion about the methine bridge
between the two heterocyclic moieties is restricted.3 Due
to their excellent staining properties, cyanine dyes have
been utilized in applications such as flow cytometry,4
DNA fragment sizing,5 as reporter groups in hybrid-
ization probes,6–8 single DNA molecule fluorescence
microscopy,9 gel staining,2 and in real time PCR.10–12
* Corresponding authors. Tel.: +46-31-7723072; fax: +46-31-7723657
(G.W.); tel.: +46-31-7728213 (H.J.K.); e-mail addresses: karjo@
oc.chalmers.se; westman@oc.chalmers.se
0968-0896/$ - see front matter Ó 2004 Elsevier Ltd. All rights reserved.
doi:10.1016/j.bmc.2004.02.006

2370
H. J. Karlsson et al. / Bioorg. Med. Chem. 12 (2004) 2369–2384
Figure 1. Intercalating (BO and TO), groove binding (BEBO, BETO and BOXTO) and the presently investigated unsymmetrical cyanine dyes
(bottom).
but produce less satisfactorily fluorescence images.20
Whereas many of the intercalating cyanine dyes do not
behave very well as reporter dyes in real-time PCR,
BEBO has been shown to be a good complement to
SYBR Green I as a nonspecific probe in real-time
PCR.10 In addition, minor groove binders are generally
more selective toward double than single stranded DNA
compared to intercalators.
dominantly binds as monomers in the minor groove.
Hence, subtle changes in the structure of the dyes affect
the binding to DNA substantially. This motivated us to
further investigate how the nature of the different het-
erocycles in this type of dyes affects their DNA inter-
actions. Herein we present a study of the binding mode
to mixed sequence DNA of a number of cyanine dyes,
focusing on the effect induced by both changing the
heterocycle that is part of the cyanine chromophore as
well as the extending heterocyclic moiety (Fig. 1). Three
different extending heterocyclic moieties were chosen:
benzoxazole, benzothiazole, and 6-methylbenzothiazole.
Since BOXTO, having an extending benzoxazole group,
exhibits the so far most promising staining properties of
these new dyes,14 further studies of dyes having this
extending moiety was motivated. To investigate the
significance of the 6-methyl group for the DNA binding
mode of the dyes, both benzothiazole and 6-methyl
benzothiazole was chosen as extending heterocycles. In
addition to the binding studies, a new facile synthetic
route to these crescent-shaped cyanine dyes is described.
The majority of groove binders interact with DNA by
forming hydrogen bonds to the base pairs,23;24 which
greatly stabilize the complex. However, symmetrical
cyanine dyes absent of hydrogen bond donors have been
found to bind in the minor groove either as monomers25
or as dimers.26 Furthermore, Petty and co-workers
recently proposed that the unsymmetrical trimethine
cyanine dye TO-PRO-3 has a substantial contribution of
cooperative minor groove binding to DNA.27 The dyes
prepared by us also lack the capacity of strong hydrogen
bonding to poly(dA-dT)2, which only have hydrogen
bond acceptors, and studies of these and mentioned
cyanine dyes may provide insight to other important
parameters for minor groove recognition such as
hydrophobicity of the ligands for example.
2. Results and discussion
2.1. Synthesis
In BEBO, the structure of the known intercalating dye
BO was extended with a 6-methyl-benzothiazole sub-
stituent (Fig. 1). The resulting crescent shape of the
molecule is similar to that of other minor groove binders
such as Hoechst. This structural modification induced a
change in binding mode from intercalation toward
minor groove binding.13 The dyes BETO and BOXTO
have the same cyanine chromophore as TO but have an
extending benzothiazole and benzoxazole group,
respectively. BOXTO clearly shows the strongest pref-
erence for minor groove binding of the dyes and also the
largest quantum yield when bound to DNA (0.52).14
Both BEBO and BETO show a more heterogeneous
binding to mixed sequence DNA than BOXTO, which
The recently reported synthetic strategy toward BEBO
and analogous dyes had two problems associated with
it.13 First, a substantial amount of brominated by-
product is produced in addition to the desired 2-amino
benzothiazole by treatment of the 4-benzothiazolyl
aniline 1 with potassium thiocyanate and bromine.
Second, the final condensation of 2-imino-3-methyl-
6-(6-methyl-benzothiazol-2-yl)-benzothiazoline worked
only with 4-methyl pyridinium salts, in modest yields.
Here we instead use the classical cyanine dye conden-
sation method, where a thiomethyl group acts as leaving
group. The required 2-thiomethyl-benzothiazolium salts

H. J. Karlsson et al. / Bioorg. Med. Chem. 12 (2004) 2369–2384
2371
N
S
N
N
ii
95%
N
i
NH2
NH2
S
S
SH
S
95%
Br
3
1
2
N
N
N
iv
OTs
SMe
N
iii
S
N
S
S
40%
96%
S
5
SMe
S
v
4
N
S
S
OTs
6
Scheme 1. Reagents and conditions: (i) KBr, NaBO3Æ4H2O, AcOH, rt, 22 h; (ii) potassium O-ethyl dithiocarbonate, DMF, 140 °C, 4 h; (iii) K2CO3,
iodomethane, DMF, rt, 2 h; (iv) MeOTs (3 equiv), 100 °C, 16 h; (v) MeOTs (large excess), 120 °C, 16 h.
were obtained by a cyclization reaction of the previously
undesired 4-aryl-2-bromo-anilines using potassium O-
ethyl dithiocarbonate yielding 2-thio-benzothiazoles,28
and subsequent dimethylation.
amino thiophenol and 4-nitro-benzoyl chloride
(Scheme 2).
8
2-Substituted benzoxazoles are usually synthesized by
either the condensation of carboxylic acids with 2-am-
inophenols catalyzed by strong acid34;37 or by the oxi-
dative cyclization of phenolic Schiff bases using various
oxidants.39–44 The former method requires activation of
carboxylic acids under strongly acidic conditions at high
temperature. The latter method using DDQ as oxidant
was attempted here,44 but in our hands the condensation
of 2-aminophenol and 4-nitro-benzoyl chloride in NMP
to produce the benzoxazole 9b was more convenient.
This reaction required longer reaction time than the
similar benzothiazole condensation and gave the diacyl-
The commercially available aniline 1 was brominated in
95% yield using a recently reported mild oxidative bro-
mination procedure of anilines, where sodium perborate
is used as oxidant and potassium bromide as the bro-
mine source (Scheme 1).29 The thiol 3 was produced in
95% yield by cyclization of the aniline 2 using potassium
O-ethyl dithiocarbonate in DMF.28 Treatment of the
thiol with methyl iodide in DMF under basic conditions
gave the 2-thiomethyl-benzothiazole
4
in 96%.30
Employing standard procedure, 4 was heated together
with excess methyl tosylate at 120 °C to give a mono-
valent tosylate salt, which, although the NMR spectra
(not shown) were consistent with the desired structure 5,
proved to be unreactive toward quaternary salts using
the cyanine dye condensation protocol. Thioalkyl
groups in the 2-position of benzothiazoles have previ-
ously been reported to be quite labile and susceptible to
thione formation,31 and presumably had the thione 6
been formed using the relatively harsh conditions.
However, milder conditions (3 equiv neat methyl tosyl-
ate at 100 °C) allowed 5 to be isolated in 40% yield.
ated (2-hydroxyphenyl)-(4-nitrophenyl)-amide as
a
by-product. However, the amide was easily removed by
trituration with methanol, affording the benzoxazole in
65% yield.
Both nitro compounds 9a,b were reduced using stan-
nous chloride in refluxing ethanol furnishing the anilines
10a and 10b in 89% and 93% yield, respectively.38 The
monobrominated anilines 11a,b were attained in over
90% yield using potassium bromide and sodium perbo-
rate in acetic acid.29 Cyclization of the bromo-anilines
11a,b and subsequent dimethylation using the same
conditions as in the synthesis of 5 afforded the quat-
ernized salts 14a and 14b in a total yield from the
starting anilines 7a and 7b of 24% and 18%, respectively
(for individual yields of the reactions, see Scheme 2).
Several methods yielding 2-aryl-benzothiazoles have
been described in the literature. These include conden-
sation of an ortho-amino thiophenol with an aromatic
aldehyde,32;33 carboxylic acid,34 acid chloride,35 or
nitrile.36 Some of these methods are quite harsh, for
example, the condensation with carboxylic acids in
polyphosphoric acid at high temperatures.34;37 There are
however milder methods to produce benzothiazoles such
as Jacobson oxidative cyclization of thiobenzanilides
using alkaline potassium ferricyanide.38 In a reasonably
mild methodology reported by Brembilla et al., ortho-
amino thiophenol is condensed with various aromatic
acid chlorides in N-methyl-pyrrolidone (NMP), which
acts as a scavenger for the HCl produced in the reac-
tion.35 Employing this procedure the benzothiazole 9a
was prepared in 93% yield by the condensation of ortho-
The quaternary salts 5, 14a and 14b were condensed
with quinolinium salts and pyridinium salts with the
methyl group in ortho or para position, thus producing
12 cyanine dyes in total (Scheme 3). The reactions pro-
ceeded smoothly in dichloromethane at room tempera-
ture using triethyl amine as base.6 The yields and names
of the dyes are summarized in Table 1. In most cases, the
ortho-methyl salts were more reactive in the condensa-
tions than the corresponding salts with para-substituted
methyl groups. Presumably this is due to the stronger
acidity of the methyl groups in 2-position providing a

2372
H. J. Karlsson et al. / Bioorg. Med. Chem. 12 (2004) 2369–2384
NO2
N
X
NH2
XH
i: X=S
iii
N
NH2
NO2
ii: X=O
X
10a: X = S (89%)
10b: X = O (93%)
9a: X = S (93%)
9b: X = O (64%)
7a: X = S
8
Cl
O
7b: X = O
N
N
N
X
vi
v
iv
NH2
SH
X
S
Br
12a: X = S (92%)
12b: X = O (89%)
11a: X = S (91%)
11b: X = O (93%)
N
X
N
N
vii
N
vi
SMe
OTs
SMe
X
S
S
14a: X = S (37%)
14b: X = O (43%)
13a: X = S (94%)
13b: X = O (86%)
Scheme 2. Reagents and conditions: (i) NMP, 100 °C, 1 h; (ii) NMP, 100 °C, 17 h; (iii) SnCl2Æ2H2O, EtOH, 12 h, reflux; (iv) KBr, NaBO3Æ4H2O,
AcOH, rt, 22 h; (v) potassium O-ethyl dithiocarbonate, DMF, 140 °C, 4 h; (vi) K2CO3, iodomethane, DMF, rt, 2 h; (vii) MeOTs (3 equiv), 100 °C,
16 h.
N
N
S
X
N
N
N
i
R
OTs
N
X
N
SMe
OTs
S
R
N
S
N
5: R = Me, X = S
X
N
14a: R = H, X = S
14b: R = H, X = O
OTs
R
Scheme 3. Reagents and conditions: (i) NEt3, CH2Cl2, rt, 20 h.
higher concentration of reactive methylene anions in the
reaction mixture.
der. By dividing the LD spectrum with the absorption
spectrum of the randomly oriented sample (isotropic
absorption) the reduced LD (LDr) is obtained. From
this value the angle between the transition moments of
the ligands and the DNA helix axis can be calculated
using Eq. 1. Minor groove binders such as DAPI46 and
Hoechst47 binds with an angle of around 45° compared
to the DNA axis while intercalators binds with an angle
close to 90°.48
2.2. DNA binding studies
2.2.1. Flow linear dichroism. Linear dichroism (LD) is
the difference in absorption of polarized light perpen-
dicular and parallel to an orientation axis. By applying a
flow gradient, DNA of sufficient length can become
oriented yielding a macromolecular orientation axis.
Subsequently, small molecules that bind to DNA in a
specific way will also be oriented and their electronic
transitions will generate LD signals according to their
interaction with DNA.45 Since the electronic transitions
of the DNA bases are perpendicular to the DNA helix
axis they will give rise to a negative LD. Thus, interca-
lating ligands will also have a negative LD whereas
minor groove binding compounds have a positive signal.
The orientation of the DNA in these studies was
achieved using a Couette cell with outer rotating cylin-
3
LDr ¼ S ꢀ O ¼ ꢀ S ꢀ ð3 cos2 a ꢁ 1Þ
ð1Þ
2
Flow-LD spectra were measured for the dyes in presence
of calf thymus DNA (ctDNA) at a mixing ratio, dye:-
bases, of 0.025. To illustrate differences in binding mode
induced by the extending heterocycle, spectra of dyes
having the same cyanine chromophore are jointly pre-
sented in the figures. The names of the dyes reflect their
basic structure and in order to clarify the following
discussions, an explanation of the names of the dyes is
motivated. Dyes having a benzoxazole or benzothiazole

H. J. Karlsson et al. / Bioorg. Med. Chem. 12 (2004) 2369–2384
Table 1. Yields and extinction coefficients of the dyes synthesized
2373
370 nm there are also substantial LD signals. These give
rise to a higher LDr and a corresponding angle of 47°.
Interestingly, in this case the methyl group on 2-BEBO
does not impose any significant differences in binding
compared to that of H-2-BEBO. 2-BOXBO has a very
low LD indicating a heterogeneous binding with both
intercalation and minor groove binding affording a low
LD. However, it could also be due to a substantial
amount of randomly oriented dye. However, this is
contradicted by results from CD, absorption, and fluo-
rescence measurements.
Dye
R
X
HetAr
Yield (%) ea (Mꢁ1 cmꢁ1
)
2-BEBO
BEBO
Me
Me
S
2-Pyridine
4-Pyridine
43
46
33,500
32,500
34,000
41,500
41,000
36,500
36,000
38,000
37,000
32,500
37,500
38,000
S
Me-2-BETO Me
S
2-Quinoline 68
4-Quinoline 61
Me-BETO
2-BOXBO
BOXBO
2-BOXTO
BOXTO
H-2-BEBO
H-BEBO
2-BETO
Me
H
H
H
H
H
H
H
H
S
O
O
O
O
S
2-Pyridine
4-Pyridine
81
40
2-Quinoline 79
4-Quinoline 53
2-Pyridine
4-Pyridine
79
47
S
S
2-Quinoline 89
4-Quinoline 79
BETO
S
2.2.1.3. 4-Quinolinium dyes. In contrast to that of
BOXBO having a relatively large negative signal,
BOXTO has a clear positive LD providing a strong
indication of minor groove binding (Fig. 2c). We have
studied BOXTO extensively and found it to have the
highest preference for monomeric minor groove binding
of all unsymmetrical cyanine dyes studied so far.14 The
fact that BOXBO is only one fused benzene ring smaller
than BOXTO nicely illustrate the large effect relatively
small variations in structure have upon the DNA rec-
ognition of these dyes. The other dyes with a 4-quino-
linium moiety, BETO and Me-BETO, also have some
interesting differences in their LD spectra. BETO has
been shown to bind by a combination of monomeric
groove binding and as a dimer or a higher oligomer with
random orientation. Thus, the second peak in the iso-
tropic absorption spectrum, not present in the LD
spectrum, is due to these aggregated nonordered dye
molecules. The methyl-substituted dye Me-BETO on the
other hand has only a weak LD signal. Symmetrical
cyanine dyes, studied by Armitage and co-workers, have
been shown to dimerize in the minor groove forming
helical aggregates on the DNA template.26 The degree of
dimerization was proposed to be dependent on the
hydrophobicity of the dye.49 Given the higher hydro-
phobicity of Me-BETO compared to BETO and the
insignificant LD signal, randomly ordered dimers might
be the dominant interaction of Me-BETO. However, it
cannot be ruled out that these dimers are bound in one
of the grooves with an angle close to 54°, which would
not generate a LD signal either (Eq. 1). Usually, minor
groove binding ligands bind with an angle close to 45°
though.
a Extinction coefficient of free dyes in 5 mM sodium phosphate buffer
(pH 7.0).
group have the prefix BOX or BE, respectively.ꢀ The
suffix of the dyes is dependent on the type of cyanine
chromophore and subsequently the parent intercalating
dye (TO or BO). Thus, dyes with a pyridinium hetero-
cycle have the suffix –BO and those with a quinolinium
group –TO. Finally, dyes having the quaternized nitro-
gen in the 2-position (2-pyridinium or 2-quinolinium)
have the number 2 indicated in the name (see Table 1).
To further clarify the presentation of the results, dyes
with the same cyanine chromophore will be referred to
as 2-pyridinium, 4-pyridinium, 2-quinolinium and 4-
quinolinium dyes.
2.2.1.1. 4-Pyridinium dyes. In Figure 2a normalized
LD and absorption spectra of dyes with the 4-pyrid-
inium heterocycle (BEBO, H-BEBO, and BOXBO) are
shown.ꢁ We have previously showed that BEBO have a
mixed binding to ctDNA but a significant contribution
of minor groove binding.13 However, the dyes H-BEBO
and BOXBO have a more negative LD signal indicating
a higher contribution of intercalation. The LDr spec-
trum (data not shown) of BOXBO even suggests that
intercalation is the major binding mode. It seems that
for these dyes the methyl substituent on the benzo-
thiazole moiety in BEBO is important affording a higher
degree of minor groove binding.
2.2.1.2. 2-Pyridinium dyes. Figure 2b shows the
spectra of DNA complexes of dyes having a 2-pyrid-
inium heterocycle (2-BEBO, H-2-BEBO, and 2-BOX-
BO). The benzothiazole dyes have a positive LD of a
similar shape as that of their isotropic absorption
spectra, indicating groove interactions. The LDr is fairly
constant throughout the long wavelength transition for
both dyes and, using Eq. 1, the average binding angle is
calculated to be 49°. At the absorption band at around
2.2.1.4. 2-Quinolinium dyes. In Figure 2d the LD and
absorption spectra of the dyes with a 2-quinolinium
moiety are shown (Me-2-BETO, 2-BETO, 2-BOXTO).
All of these dyes have a strong positive LD with a
similar shape to their absorption spectra, indicating
minor groove binding. From the LDr values, the average
binding angle at the absorption maxima for the two
peaks were calculated using Eq. 1 to be between 43° and
46°. This is similar to the angle for known minor groove
binders, for example, DAPI.46 The peak at shorter
wavelength might at the first look be attributed to a
band arising from dimers or aggregates, as in the 4-
quinolinium case, but is here more probable due to a
vibronic progression, as observed for other monome-
thine cyanine dyes such as YO.3 The latter explanation
ꢀ
Some of the dyes also have a Me- or H-prefix in order to distinguish
them from the already published dyes BEBO, BETO and BOXTO.
ꢁ
Throughout the entire paper, spectra corresponding to dyes with the
benzoxazole, benzothiazole, and 6-methyl-benzothiazole as the
extending heterocycle appears in red, blue, and black, respectively.

2374
H. J. Karlsson et al. / Bioorg. Med. Chem. 12 (2004) 2369–2384
Figure 2. Normalized flow LD and absorption spectra of the cyanine dyes in presence of ctDNA at a mixing ratio, dye:bases, of 1:40, normalized at
the DNA base transition. (a) 4-Pyridinium dyes: BEBO (black), H-BEBO (blue), and BOXBO (red); (b) 2-pyridinium dyes: 2-BEBO (black), H-2-
BEBO (blue), and 2-BOXBO (red); (c) 4-quinolinium dyes:Me-BETO (black), BETO (blue), and BOXTO (red); (d) 2-quinolinium dyes: Me-2-BETO
(black), 2-BETO (blue), and 2-BOXTO (red). The absorption spectra are in all cases the above spectra of the same color. [dye] ¼ 10 lM in all spectra.
is supported by the similar LD and absorption profiles
for the benzoxazole and unsubstituted benzothiazole
compared to the 6-methyl derivative. For this com-
pound, the most hydrophobic dye Me-2-BETO, the
peak at 465 nm is more pronounced than for the other
dyes, indicating a higher abundance of dye–dye inter-
actions. Interestingly, the LDr value for the first band is
lower than for the second indicating the presence of
intercalating or nonordered monomer dye molecules.
which has the most negative LD of all the dyes studied
in this series, also have a relatively low CD signal further
supporting that the dominant interaction is by intercal-
ation.
2.2.2.2. 2-Pyridinium dyes. Figure 3b shows the CD
spectra of DNA complexes of dyes having a 2-pyrid-
inium heterocycle (2-BEBO, H-2-BEBO, and 2-BOX-
BO). As in the LD measurements, the CD spectra of the
two dyes with a benzothiazole moiety are similar with
a relatively large ICD. Although 2-BOXBO only had a
very weak LD signal, the ICD indicates that there is
a relatively high abundance of groove bound dye mole-
cules. In comparison to the 4-pyridinium dyes, the ICD
at the absorption band at around 370 nm is much larger
for these dyes. The reason for this is not clear, although
it might be due to that the extending heterocycle is
rotated slightly out of co-planarity by the groove
interactions, affording a chiral conformation of the bis-
benzazole moiety. However, it could be noted that also
rigid heterocycles can obtain very strong ICD when
bound in the minor groove.52 Interestingly, the large
ICD at 370 nm of the 2-pyridinium dyes is comparable
to that at 350 nm of the bis-benzimidazole dye Hoe-
chst,53 indicating that the bis-benzazole and bis-benz-
imidazole chromophores may be bound in a similar way.
2.2.2. Circular dichroism. Upon binding of achiral mol-
ecules to the chiral nucleic acid helix structure, they
acquire an induced CD (ICD), which is characteristic of
their interaction with DNA. Compared to the intercal-
ation site between the base pairs, the minor groove of
DNA provides a more chiral environment affording a
larger ICD of minor groove binders.50;51 In addition,
nonrigid ligands binding in the minor groove, such as
DAPI, obtain a CD signal induced by a chiral defor-
mation of the molecule when interacting with the DNA
helix.
2.2.2.1. 4-Pyridinium dyes. CD spectra were measured
on the same samples as those in the LD measurements
(mixing ratio 0.025). In Figure 3a the CD spectra of the
dyes with a 4-pyridinium moiety are shown (BEBO,
H-BEBO, BOXBO). In agreement with LD results
obtained, BEBO has the strongest ICD indicating a
higher abundance of minor groove binding. BOXBO,
2.2.2.3. 4-Quinolinium dyes. In Figure 3c the CD
spectra of the 4-quinolinium dyes in presence of ctDNA

H. J. Karlsson et al. / Bioorg. Med. Chem. 12 (2004) 2369–2384
2375
Figure 3. CD spectra of the cyanine dyes in presence of ctDNA at a mixing ratio, dye:bases, of 1:40, normalized at the DNA base transition. (a)
4-Pyridinium dyes: BEBO (black), H-BEBO (blue), and BOXBO (red); (b) 2-pyridinium dyes: 2-BEBO (black), H-2-BEBO (blue), and 2-BOXBO
(red); (c) 4-quinolinium dyes: Me-BETO (black), BETO (blue), and BOXTO (red); (d) 2-quinolinium dyes: Me-2-BETO (black), 2-BETO (blue), and
2-BOXTO (red), [dye] ¼ 10 lM in all spectra.
are shown. BOXTO has an ICD of large amplitude
further supporting groove binding. The stronger pref-
erence for monomeric groove binding by BOXTO is
illustrated by roughly a twice as large ICD of BOXTO
than to that of BETO. The spectrum of BETO more
clearly exhibits a small contribution of exciton coupling
interactions between closely bound chromophores, with
the cross-over of the CD couplet at around 490 nm. The
higher hydrophobicity of Me-BETO relative to BETO
might imply that the driving force for dimerization of
Me-BETO upon interaction with DNA is larger. Indeed,
the CD spectrum shows an exciton-coupled signal
around 490 nm as the major signal, indicating randomly
bound dimers as the dominant interaction.
couplet should be found. However, the red-shifted band
of the CD couplet is possibly cancelled out due to the
positive sign of the ICD from bound monomer dye
molecules at the same wavelength.
Similarly to the dyes with the 2-pyridinium moiety, the
2-quinolinium series of dyes have a large ICD at the
short wavelength transition band around 370 nm (see
above).
2.2.3. Absorption measurements. A common feature of
cyanine dyes is their tendency to aggregate in water
solutions.54 Dimerization and further aggregation of
cyanine dyes upon binding to DNA has also been
observed in a number of studies.26;49;55;56 In Figures 5–8
are shown comparisons between the dyes absorption
spectra in presence of ctDNA at a mixing ratio 1:40 (a)
and in methanol (b), where the dye molecules are pre-
sumably present as monomers.13;54;57
2.2.2.4. 2-Quinolinium dyes. In Figure 3d the CD
spectra of the dyes with a 2-quinolinium moiety are
shown. For both 2-BOXTO and 2-BETO the ICD at the
long wavelength transition is relatively large supporting
minor groove binding. In contrast, the ICD of Me-2-
BETO is very low at this wavelength. This is in agree-
ment with the LD results (Fig. 2d), which shows that the
contribution of groove-bound monomers of Me-2-
BETO is smaller than that of the other dyes with the
same cyanine chromophore. According to the absorp-
tion and LD measurements Me-2-BETO has a high
abundance of dimer interaction with ctDNA. The dimer
absorption peak of Me-2-BETO has its maximum at
around 465 nm, where the cross-over of the exciton
2.2.3.1. 2- and 4-Pyridinium dyes. In Figure 4a, b
spectra of dyes having a 4-pyridinium moiety are shown.
The shape of the spectra in methanol and in presence of
DNA is quite similar indicating that most of the dyes are
bound as monomers. In the absorption spectra of the
dyes with a 2-pyridinium moiety in presence of ctDNA,
there is a small shoulder at shorter wavelengths, which

2376
H. J. Karlsson et al. / Bioorg. Med. Chem. 12 (2004) 2369–2384
Figure 6. Normalized absorption spectra of Me-BETO (black), BETO
(blue), and BOXTO (red) in presence of ctDNA at a mixing ratio,
dye:bases, of 1:40 (a) and free in methanol solution (b).
Figure 4. Normalized absorption spectra of BEBO (black), H-BEBO
(blue), and BOXBO (red) in presence of ctDNA at a mixing ratio,
dye:bases, of 1:40 (a) and free in methanol solution (b).
Figure 5. Normalized absorption spectra of 2-BEBO (black), H-2-
BEBO (blue), and 2-BOXBO (red) in presence of ctDNA at a mixing
ratio, dye:bases, of 1:40 (a) and free in methanol solution (b).
Figure 7. Normalized absorption spectra of Me-2-BETO (black), 2-
BETO (blue), and 2-BOXTO (red) in presence of ctDNA at a mixing
ratio, dye:bases, of 1:40 (a) and free in methanol solution (b).
might be due to the presence of dimers (Fig. 5). How-
ever, the same shoulder is also observed in the LD
spectra and since the LDr (data not shown) is reason-
ably constant throughout this transition it is likely that

H. J. Karlsson et al. / Bioorg. Med. Chem. 12 (2004) 2369–2384
2377
Figure 8. Energy minimized structures of BOXTO (left) and 2-BOXTO (right) docked into poly(dA-dT)2.
it originates from the same electronic transition and that
the shoulder correspond to different vibronic transitions
to the same electronic excited state.3
4-quinolinium series, the most hydrophobic dye with a
6-methyl-benzothiazole moiety (Me-2-BETO) has a
dominant dimer peak at around 465 nm. In case of 2-
BETO and 2-BOXTO, there is also another peak at
shorter wavelengths, which is similar to that of the 2-
pyridinium dyes are also present in the LD spectra.
Since there is a shoulder observed in the methanol
spectra at similar wavelengths, which again probably
originate from other vibronic transitions within the
same electronic excited state,3 this suggests that the
short wavelength peak in the presence of DNA more
strongly correspond to different vibronic transitions
rather than bound dimers. In addition, these peaks are
slightly red-shifted in comparison with the presumed
dimer peak of Me-2-BETO. Upon binding to DNA the
dyes will become more rigid, which might explain the
more distinct vibronic peaks in the DNA spectra, com-
pared to the spectra of free dye in methanol.
2.2.3.2. 4-Quinolinium dyes. Figure 6 shows the
absorption spectra of the 4-quinolinium containing
dyes. Interestingly, the most hydrophobic dye,
Me-BETO, shows the highest degree of dimer interac-
tion with DNA, illustrated by the dominant dimer peak
at 490 nm. The dimeric interaction is also supported by
CD results, showing an exciton-coupled signal at 490 nm
(Fig. 3c). BETO clearly exhibits a substantial dimeric
binding, whilst the least hydrophobic dye, BOXTO,
binds predominantly as a monomer.
2.2.3.3. 2-Quinolinium dyes. In Figure 7 the similar
absorption spectra as above are shown for dyes with a
2-quinolinium heterocycle. Analogous to that of the
Table 2. Absorption, fluorescence, and binding properties of the investigated dyes bound to ctDNAa
c
d
Dye
Ext.
HetArb
kabs: max:
(nm)
kems: max:
(nm)
/
/
bound=/free
Fbound=Ffree
LDrdye
/
Main binding modef /main
bound formg
F
e
LDrDNA
BEBO
2-BEBO
Me-Bt.
Me-Bt.
Me-Bt.
Me-Bt.
Bt.
467
455
493
465
467
454
516
500
463
452
515
495
492
502
570
550
491
495
561
542
487
488
552
535
0.18
16
7
300
60
)0.15
)0.29
+0.002
)0.48
+0.20
)0.24
)0.24
)0.59
+0.61
+0.017
)0.60
)0.58
Intermediate/monomer
(Minor groove)h/monomer
Intermediate/dimer
0.062
0.086
0.079
0.19
Me-BETO
Me-2-BETO
H-BEBO
H-2-BEBO
BETO
6
40
9
100
300
200
130
180
340
100
260
370
Minor groove/dimer
Intermediate/monomer
(Minor groove)h/monomer
6
Bt.
Bt.
0.081
0.21
0.22
15
14
24
32
30
50
72
h;i
(Minor groove)h/(dimer)
2-BETO
Bt.
Minor groove/monomer
Intercalation/monomer
Intermediate/monomer
Minor groove/monomer
Minor groove/monomer
BOXBO
2-BOXBO
BOXTO
2-BOXTO
Box.
Box.
Box.
Box.
0.41
0.082
0.52
0.32
a Measured at 25 °C in 5 mM sodium phosphate buffer (pH 7.0) at a mixing ratio, dye:bases, of 1:100.
b Extending heterocycle on the cyanine chromophore. Me-Bt., Bt., and Box. correspond to methyl-benzothiazole, benzothiazole, and benzoxazole,
respectively.
c Fluorescence quantum yields, /F were determined relative to fluorescein in 0.1 M NaOH, assuming a /F of 0.93.
d Fbound=Ffree is the largest increase in fluorescence intensity (dependent on wavelength) upon binding to DNA.
e The ratio between the reduced LD at kabs: max : and at the DNA transition (260 nm). A ratio of )0.50 corresponds to an average binding angle of 45°
(minor groove binding) and a ratio of +1.0 an angle of 90° (intercalation) (from Eq. 1). Dye:bases ratio 1:40.
f The main binding mode as deduced from LD and CD results. Dye:bases ratio 1:40.
g The main bound form of the dyes to ctDNA, roughly estimated from absorption and CD results.
h The brackets note a slight majority of the designated binding interaction.
i The minor groove interaction of BETO is dominated by monomers.14

2378
H. J. Karlsson et al. / Bioorg. Med. Chem. 12 (2004) 2369–2384
2.2.4. Fluorescence measurements. The fluorescence
properties of the dyes in presence of ctDNA at a mixing
ratio, dye:bases, of 1:100 are shown in Table 2. The
relatively high quantum yield of many of these cyanines
bound to DNA combined with their absorption at long
wavelengths may render them useful in biological
applications. Although some of these dyes bind by a
combination of groove binding and intercalation they
may still be useful as nonselective stains for DNA due to
their large quantum yield when bound. For example,
BOXBO has a quantum yield of 0.41, which can be
compared to 0.083 for the parent intercalating dye BO.58
As a further comparison, the monovalent intercalating
cyanine dye TO has a quantum yield of 0.11,55 and the
tetravalent cyanine dyes YOYO and TOTO of around
0.4 when bound to DNA.59 Both the 2-pyridinium and
the 2-quinolinium dyes have a slightly lower absorption
and emission maximum than the corresponding para-
substituted cyanines, affording a possibly useful exten-
sion of the spectral characteristics of this type of cyanine
dyes.
an angle of 45°.13;14 In addition, many of these dyes bind
with an angle close to 45° to mixed sequence ctDNA.
The 2-pyridinium dyes have a higher proportion of
groove binding than the dyes with the 4-pyridinium
dyes. Similarly, the 2-quinolinium dyes are principally
bound in the groove and have a larger preference for
minor groove binding than the 4-quinolinium dyes, with
the exception of BOXTO. The thermodynamically most
stable conformations of these dyes would have the
methylated nitrogen directed out from the groove (Figs.
1 and 8). This might impose less steric hindrance and
stronger van der Waals interactions with the walls of the
groove. In addition, the 2-quinolinium dyes will have the
fused benzene ring in a different direction compared to
the 4-quinolinium dyes, affording a longer crescent-
shaped dye with a slight difference in overall curvature.
The requirement of isohelicity of minor groove binding
ligands have been highlighted in several studies60–62 and
the 2-quinolinium dyes may be more isohelical with the
minor groove than the 4-quinolinium dyes. Indeed,
energy minimized structures of BOXTO and 2-BOXTO
docked into an alternating AT oligomer indicate that
the fused benzene ring of 2-BOXTO follows the groove,
whereas in BOXTO this ring is pointing outwards (Fig.
8). However, both dyes fit very well in the groove,
covering roughly 4 AT base pairs. The delocalization of
charge will also vary between the different types of
cyanine chromophores. One could only speculate to
what extent each of these parameters will contribute to
the affinity for the minor groove and ultimately crystal
or NMR structures of the dye–DNA complexes is
warranted.
For all new dyes, a blue-shift in emission maxima upon
binding to DNA of between 30 and 70 nm is observed
(data not shown). This blue-shift can be of advantage
since the fluorescence intensity for the free dye at
wavelengths near the emission maxima for the bound
dye is low, resulting in a larger enhancement in fluo-
rescence intensity when measured at shorter wavelengths
compared to the total increase in quantum yield. The
increase in quantum yield of the dyes upon binding to
DNA varies between 6- and 72-fold, whereas the maxi-
mum increase in fluorescence intensity is between 40 and
370. The most prominent monomeric minor groove
binder of the presently investigated dyes, BOXTO, also
has the largest quantum yield when bound to ctDNA
(0.52). The benzoxazole containing dye 2-BOXTO also
has a high quantum yield, 0.32, and the largest increase
in quantum yield upon binding.
The variation of the extending benzazole heterocycle
also provides some curious differences in binding mode.
The significance of the methyl group on the benzo-
thiazole ring is more apparent in the case of the 4-py-
ridinium than the 2-pyridinium dyes, illustrated by the
higher degree of groove binding of BEBO than H-BEBO
lacking the methyl group and the negligible difference
between 2-BEBO and H-2-BEBO. Possibly, steric hin-
drance toward the minor groove induced by the methyl
group in BEBO forces the nitrogen atom on the ben-
zothiazole moiety to be directed into the groove. This
might stabilize the complex to a higher degree than
when free rotation of the benzothiazole group allows the
sulphur to be directed inwards. Alternatively, it might be
a hydrophobic effect. In a study by Mikheikin et al.,
trimethine symmetrical cyanine dyes were found to bind
in the minor groove as monomers.25 They proposed that
the DNA-binding ligands should be sufficiently hydro-
phobic to provide a high affinity for the minor groove.
For the 4-pyridinium dyes this seems to fit reasonably
well: the more hydrophobic BEBO has the strongest
affinity for the minor groove and BOXBO being the
least hydrophobic binds with a high degree of interca-
lation. However in the case of 2-BEBO and H-2-BEBO,
which interact with DNA in a very similar fashion, the
methyl group is nonsignificant. In the 4-quinolinium
dyes the methyl group induce an increase of the abun-
dance of dimer interaction. This is most evident in the
case of Me-BETO, where the dimer interactions clearly
2.3. Implications of structural influence on binding mode
In order to facilitate discussions of the relative degree of
groove binding of these dyes, the ratios between the LDr
value of the cyanine transitions and that of the DNA
base transitions are shown in Table 2 (also known as the
dichroism ratio parameter).47 A value of )0.5 corre-
spond to an angle of 45° relative to the DNA helix axis
(Eq. 1), which is characteristic of minor groove binding
ligands. Purely intercalating compounds would have a
ratio close to +1.0, which corresponds to the transition
moment of the ligand being oriented completely parallel
to the DNA bases. In the following discussion it is as-
sumed that lower LD signals, affording a less negative or
a positive ratio, for some of the dyes are due to a con-
tribution of either intercalated or nonoriented externally
bound dye. Alternatively, oriented dye binding in the
minor groove with an angle of 54° would have no LD
and thus also lead to a decrease in the total LD signal.
This is less likely however, since BEBO, BOXTO, and
BETO bind in the minor groove of poly(dA-dT)2 with

H. J. Karlsson et al. / Bioorg. Med. Chem. 12 (2004) 2369–2384
2379
dominate (Table 2). The benzothiazole dye BETO has a
4. Experimental
4.1. Spectroscopic measurements
faint majority of dimer interactions, whereas BOXTO
binds mainly as a monomer. It seems that sufficient
hydrophobicity and size is crucial for monomeric rec-
ognition of the minor groove of these dyes. However,
slight increases in hydrophobicity cause the dyes to bind
in a dimeric fashion.
UV–vis spectra were measured on a Varian Cary4
spectrophotometer. Fluorescence spectra were recorded
using a SPEX fluorolog s2 spectrofluorimeter. The flow-
LD and CD spectra were recorded on a JASCO-720
spectropolarimeter. The orientation of the DNA com-
plexes was achieved using a flow Couette cell with outer
rotating cylinder. All spectroscopic measurements were
performed at 25 °C in 5 mM sodium phosphate buffer
(pH 7.0). Aqueous solutions of the dyes used were typ-
ically obtained from 1 to 3 mM stock solutions in
DMSO. The DNA concentration per nucleotide was
determined by absorption spectroscopy using extinction
coefficient 6600 Mꢁ1 cmꢁ1 at 260 nm. The cyanine dye
concentrations were estimated using the extinction
coefficients found in Table 1. Calf thymus DNA was
purchased from Fluka.
3. Conclusions
A new facile synthetic route to a number of crescent-
shaped unsymmetrical dyes has been developed. The
yields of the reactions toward the bis-benzazoles 4 and
13a,b are good to excellent. The quaternizations of the
benzothiazole nitrogen proceeds in relatively modest
yields, whereas the cyanine dyes condensations proceeds
in satisfactorily yields. Flow LD and CD studies of the
12 dyes show that the extent of minor groove binding to
mixed sequence ctDNA varies between the dyes. The 2-
quinolinium dyes (Me-2-BETO, 2-BETO, and 2-BOX-
TO) bind principally in the minor groove both as
monomers and as dimers (Me-2-BETO). In the case of
the 4-quinolinium dyes (Me-BETO, BETO, and BOX-
TO), nonordered dimeric interactions increase with the
hydrophobicity of the dyes. BOXTO exhibits the most
pronounced monomeric minor groove binding of the
dyes studied as yet. The 2-pyridinium dyes (BEBO, H-
BEBO, and BOXBO) have a higher contribution of
intercalation going from higher to lower hydrophobic-
ity. However, this increase in the degree of intercalation
can be a pure structural effect and also be dependent on
the size of the dyes. In BEBO the methyl group on the
extending benzothiazole moiety might impose a more
favorable conformation for minor groove binding.
Relative to the 4-pyridinium dyes, the 2-pyridinium dyes
(2-BEBO, H-2-BEBO, and 2-BOXBO) have a higher
contribution of minor groove binding.
4.2. Mathematical modeling
Structures of 2-BOXTO and BOXTO were optimized
with PM3, docked into the minor groove of an alter-
nating AT oligomer in canonical B-DNA conformation,
and further energy minimized in the MM+ force field of
the HYPERCHEM software package (HyperCube, Inc.).
4.3. Synthesis
Flash and column chromatography was performed using
aluminium oxide (activated, neutral, approx. 150 mesh),
deactivated by addition of water to Brockman grade III,
and silica gel (Merck, grade 60, 70–230 mesh). 1H
(400 MHz) and 13C (100.6 MHz) NMR spectra were
recorded at rt using a Varian UNITY-400 NMR spec-
trometer. Chemical shifts are in ppm, relative to solvent
1
The relatively high fluorescence quantum yield of many
of these cyanines bound to DNA combined with their
absorption at long wavelengths may render them useful
in biological applications. In particular, two of the
benzoxazole containing dyes BOXTO and 2-BOXTO
show a high degree of minor groove binding and
quantum yields of 0.52 and 0.32, respectively, when
bound to DNA. Upon binding to ctDNA the dyes
exhibit varyingly large enhancements in quantum yield.
The lowest increase is observed for Me-BETO and the
largest for 2-BOXTO, with a 6-fold and a 72-fold
increase, respectively. In most cases, dyes with a larger
contribution of dimer interaction have a lower quantum
yield when bound to DNA. Due to a blue-shift in
emission maxima upon binding to DNA, the increase in
fluorescence intensity are larger than the quantum yield
enhancements when measured at slightly shorter wave-
lengths than the emission maxima and varies between
40- and 370-fold. Finally, the introduction of cyanine
chromophores having the quinolinium or pyridinium
moiety in 2-position provide a possibly useful extension
of the spectral characteristics of this type of cyanine
dyes.
peaks for DMSO-d6 (d 2.50 for H and dC 39.51 for 13C
NMR), CDCl3 (dTMS 0 for 1H dC 77.0 for 13C NMR) and
methanol-d4 (d 3.31 for 1H and dC 49.0 for 13C NMR). J
values are given in Hertz. High resolution mass spectra
were performed with a VG 7070E magnetic sector
instrument (VG Analytical/Micromass, Manchester,
UK). Conditions for FAB (fast atom bomardment): Xe
gun at 8 kV, matrix glycerol or 3-nitrobenzylalcohol with
PEG600 as mass reference. A signal from a coil in the
magnet field was used for mass calibration. Commercial
reagents were purchased from Sigma-Aldrich and used
without further purification. 2-(4-Aminophenyl)benzo-
thiazole 10a was prepared from 2-aminothiophenol 7a in
two steps as previously described.35;38. 2-(4-Nitrophen-
yl)benzoxazole 9b was reduced according to a previously
reported procedure to afford 2-(4-aminophenyl)benz-
oxazole 10b in 93% yield.63 Potassium O-ethyl dithio-
carbonate64 and 4-nitrobenzoyl chloride65 were prepared
as described previously.
4.3.1. 2-(4-Nitrophenyl)benzoxazole 9b. 4-Nitrobenzoyl-
chloride (2.25 g, 15 mmol) was added in portions to a

2380
H. J. Karlsson et al. / Bioorg. Med. Chem. 12 (2004) 2369–2384
solution of 2-aminophenol (1.63 g, 15 mmol) in dry
NMP (10 mL) (caution, exothermic reaction). The mix-
ture was stirred at 100 °C for 17 h. After cooling, the
mixture was poured into water and the pH was adjusted
to 8–9 by the addition of aqueous ammonia. The pre-
cipitate formed was collected by filtration, washed with
methanol, and air dried to afford 9b as a white solid
(2.31 g, 64%).1H NMR (CDCl3): d 7.40–7.48 (2H, m,
ArH), 7.65 (1H, d, J ¼ 8, ArH), 7.84 (1H, d, J ¼ 8,
ArH) 8.36–8.50 (4H, m, ArH). These data are consistent
with those reported for the previously prepared com-
pound.66
(6.8 g, 42 mmol) in portions. The mixture was heated to
140 °C for 4 h. After cooling the mixture was diluted
with water (30 mL) and acidified by 6 N HCl. The pre-
cipitate formed was collected by filtration, washed with
water and dried to obtain the thiol as a faint yellow
solid.
4.3.3.1. 6-(6-Methyl-benzothiazol-2-yl)-2-mercapto-
benzothiazole 3. Obtained from the bromo-aniline 2
(6.05 g, 19 mmol) using the above-described procedure
1
(5.68 g, 95%). H NMR (DMSO-d6): d 2.45 (3H, s, Ar-
CH3), 7.35 (1H, d, J ¼ 8, ArH), 7.41 (1H, d, J ¼ 8,
ArH), 7.90 (2H, m, ArH), 8.08 (1H, d, J ¼ 8, ArH), 8.42
(1H, s, ArH); 13C NMR (DMSO-d6): d 21.15, 112.9,
120.6, 121.9, 122.3, 126.4, 128.2, 129.0, 130.6, 134.6,
135.4, 143.3, 151.7, 165.3, 191.0; HR-FAB-MS m=z
found: 315.001 C15H11N2S3 (M+Hþ): requires M,
315.0084.
4.3.2. General procedure for bromination to afford 2 and
11a,b. NaBO3Æ4H2O (3.38 g, 22 mmol) was added in one
portion to a suspension of the appropriate aniline
(20 mmol) and KBr (2.86 g, 24 mmol) in concentrated
acetic acid (15 mL). After stirring at rt for 22 h, the
reaction was quenched by the addition of ice water. The
precipitate formed was collected by filtration, washed
with water, and air dried to afford the ortho-brominated
aniline.
4.3.3.2. 6-(Benzothiazol-2-yl)-2-mercaptobenzothiazole
12a. Obtained from the bromo-aniline 11a (5.80 g,
19 mmol) using the above-described procedure (5.25 g,
4.3.2.1. 2-(4-Amino-3-bromophenyl)-6-methyl-benzo-
thiazole 2. Obtained from the aniline 1 (4.81 g,
20 mmol) as a white solid (6.07 g, 95%) using the above-
1
92%). H NMR (DMSO-d6): d 7.46 (2H, m, ArH), 7.55
(1H, d, J ¼ 8, ArH) 8.03 (1H, d, J ¼ 8, ArH), 8.14 (2H,
m, ArH) 8.48 (1H, s, ArH), 14.02 (1H, s, SH); HR-FAB-
MS m=z found: 301.422 C14H9N2S3 (M+Hþ): requires
M, 301.430.
1
described procedure. H NMR (DMSO-d6): d 2.43 (3H,
s, Ar-CH3), 6.07 (2H, s, NH2), 6.89 (1H, d, J ¼ 8, ArH),
7.30 (1H, d, J ¼ 8, ArH), 7.74 (1H, dd, J1 ¼ 8, J2 ¼ 2,
ArH), 7.81 (1H, d, J ¼ 8, ArH), 7.84 (1H, s, ArH), 8.02
(1H, d, J ¼ 2, ArH). These data are consistent with
those reported for the previously prepared compound.38
4.3.3.3. 6-(Benzoxazol-2-yl)-2-mercaptobenzothiazole
12b. Obtained from bromo-aniline 11b (5.49 g,
19 mmol) using the above-described procedure (4.81 g,
1
89%). H NMR (DMSO-d6): d 4.32 (1H, s, SH), 7.43
4.3.2.2. 2-(4-Amino-3-bromophenyl)benzothiazole 11a.
Obtained from the aniline 1 (4.53 g, 20 mmol) as a white
solid (5.55 g, 91%) using the above-described procedure.
1H NMR (CDCl3): d 4.45 (2H, s, NH2), 6.83 (1H, d,
J ¼ 8, ArH), 7.35 (1H, d, J ¼ 8, ArH), 7.47 (1H, d,
J ¼ 8, ArH), 7.82 (1H, d, J ¼ 8, ArH), 7.87 (1H, d,
J ¼ 8, ArH), 8.0 (1H, d, J ¼ 8, ArH), 8.21 (1H, s, ArH).
These data are consistent with those reported for the
previously prepared compound.38
(2H, m, ArH), 7.48 (1H, d, J ¼ 8, ArH) 7.79 (1H, dd,
J1 ¼ 8, J2 ¼ 2, ArH), 8.22 (1H, d, J ¼ 8, ArH) 8.57 (1H,
s, ArH); 13C NMR (DMSO): d 110.8, 113.0, 119.8,
120.9, 122.2, 124.9, 125.5, 126.6, 130.5, 141.5, 143.8,
150.2, 161.8, 191.2; HR-FAB-MS m=z found: 284.001
(Mþ): C14H8N2OS2 requires M, 284.008.
4.3.4. General procedure for preparation of 2-thiomethyl
benzothiazoles 4 and 13a,b. To the appropriate 2-mer-
capto benzothiazole (16 mmol) and potassium carbonate
(3.3 g, 24 mmol) in anhydrous DMF (50 mL) was added
iodomethane (2.71 g, 19 mmol). After stirring at rt for
2 h water (50 mL) was added and the precipitate formed
was collected by filtration, washed with water, and air
dried. The analytical samples were recrystallized from
ethyl acetate to give the benzothiazoles as a faint yellow
solids.
4.3.2.3. 2-(4-Amino-3-bromophenyl)benzoxazole 11b.
Obtained from the aniline 9b (4.20 g, 20 mmol) as a
white powder (5.38 g, 93%) using the above-described
1
procedure. H NMR (CDCl3): d 4.51 (2H, s, NH2) 6.86
(1H, d, J ¼ 8, ArH), 7.33 (2H, m, ArH), 7.54 (1H, d,
J ¼ 8, ArH), 7.71 (1H, d, J ¼ 8, ArH), 8.00 (1H, d,
J ¼ 8, ArH), 8.35 (1H, s, ArH); 13C NMR (DMSO-d6):
d 106.7, 110.5, 114.5, 115.0, 119.1, 124.6, 127.8, 127.9,
131.4, 141.8, 149.3, 149.9, 162.0; HR-FAB-MS m=z
found 289.001, C13H9BrN2O (M+Hþ): requires M,
289.006.
4.3.4.1. 6-(6-Methyl-benzothiazol-2-yl)-2-thiomethyl-
benzothiazole 4. Obtained from the mercaptobenzo-
thiazole 3 (5.0 g, 16 mmol) using the above-described
1
procedure (5.04 g, 96%). %). H NMR (DMSO-d6): d
4.3.3. General procedure for preparation of 2-mercapto-
benzothiazoles 3 and 12a,b. To a solution of the appro-
priate bromoaniline (19 mmol) in anhydrous DMF
(30 mL) was added potassium O-ethyl dithiocarbonate
2.46 (3H, s, Ar-CH3), 2.83 (3H, s, S-CH3), 7.37 (1H, d,
J ¼ 8, ArH), 7.95 (3H, m, ArH), 8.14 (1H, d, J ¼ 8,
ArH), 8.79 (1H, s, ArH); HR-FAB-MS m=z found:
329.032 C16H13N2S3 (M+Hþ): requires M, 329.0241.

H. J. Karlsson et al. / Bioorg. Med. Chem. 12 (2004) 2369–2384
2381
4.3.4.2. 6-(Benzothiazol-2-yl)-2-thiomethyl-benzothiaz-
ole 13a. Obtained from the mercaptobenzothiazole 12a
(4.81 g, 16 mmol) using the above-described procedure
(4.73 g, 94%). 1H NMR (DMSO-d6): d 2.84 (3H, s,
S-CH3), 7.48 (1H, t, J ¼ 8, ArH), 7.56 (1H, t, J ¼ 8,
ArH), 7.99 (1H, d, J ¼ 8, ArH), 8.07 (1H, d, J ¼ 8,
ArH), 8.18 (2H, m, ArH), 8.84 (1H, s, ArH); HR-FAB-
MS m=z found: 315.013 C15H11N2S3 (M+Hþ): requires
M, 315.0084.
(1H, d, J ¼ 8, ArH), 9.29 (1H, s, ArH); HR-FAB-MS
m=z found: 313.0459 C16H13N2OS2 (Mþ): requires M,
313.0469.
4.3.6. General procedure for the cyanine dye condensa-
tions. The 2-thiomethyl benzothiazolium salt 5, 14a or
14b (0.05 mmol) and the appropriate picolinium or
lepidinium salt (0.05 mmol) were stirred in dichlorome-
thane (2 mL). Triethyl amine (0.2 mmol, 20 mg, 28 lL)
was added to form a deeply colored mixture, which was
stirred at rt for 20 h. The mixture was concentrated and
the residue purified by flash chromatography on neutral
Al2O3 with methanol–dichloromethane (2–4% MeOH)
as eluant. During chromatography, the tosylate coun-
terion was in all cases exchanged for chloride, as judged
by absence of tosylate protons in the NMR spectra and
positive silver halide tests. The chloride anions most
likely originate from traces of HCl formed by decom-
position of the chlorinated solvent. Chemical shifts in
deuterated methanol were found to be quite sensitive to
concentration, probably due to aggregation. The data
4.3.4.3. 6-(Benzoxazol-2-yl)-2-thiomethyl-benzothiaz-
ole 13b. Obtained from the mercaptobenzothiazole 12b
(4.55 g, 16 mmol) using the above-described procedure
1
(4.1 g, 86%). H NMR (CDCl3): d 2.85 (3H, s, S-CH3),
7.38 (1H, d, J1 ¼ 8, J2 ¼ 2, ArH) 7.60 (1H, t, J ¼ 8,
ArH), 7.78 (1H, t, J ¼ 8, ArH), 7.98 (1H, d, J ¼ 8,
ArH), 8.33 (1H, d, J ¼ 8, ArH), 8.69 (1H, s, ArH); 13C
NMR (CDCl3): d 16.22, 110.8, 120.2, 120.7, 121.8,
123.1, 124.9, 125.3, 125.8, 136.0, 142.4, 151.0, 155.5,
162.8, 171.6; HR-FAB-MS m=z found: 299.032
C15H11N2OS2 (M+Hþ): requires M, 299.0313.
obtained here are from solutions of roughly 3 mg mLꢁ1
.
4.3.5. General procedure for the preparation of the
benzothiazolium salts 5 and 14a,b. The appropriate
benzothiazole (1 mmol) was heated with methyl tosylate
(0.56 g, 3 mmol) at 100 °C for 16 h. The glassy solid
formed was dissolved in DMF (2 mL) and water (15 mL)
was added. The mixture was refrigerated over night and
the crystals formed were collected by filtration.
4.3.6.1. 4-[(3-Methyl-6-(6-methyl-benzothiazol-2-yl)-
2,3,-dihydro-(benzo-1,3-thiazole)-2-methylidene)]-1-meth-
yl-pyridinium chloride (BEBO). Obtained from the ben-
zothiazolium salt 5 (26 mg, 0.05 mmol) and 4-picolinium
tosylate (14 mg, 0.05 mmol) using the above-described
procedure as a yellow solid (10 mg, 46%). 1H NMR
(DMSO-d6): d 2.47 (3H, s, Ar-CH3), 3.76 (3H, s, N-
CH3), 4.02 (3H, s, N-CH3), 6.34 (1H, s, @CHA), 7.38
(1H, d, J ¼ 8, ArH), 7.47 (2H, d, J ¼ 7, PyH), 7.70 (1H,
d, J ¼ 8, ArH) 7.93 (1H, d, J ¼ 8, ArH), 7.95 (1H, s,
ArH), 8.18 (1H, d, J ¼ 8, ArH), 8.39 (2H, d, J ¼ 7,
PyH), 8.65 (1H, s, ArH). These data were consistent
with those of the previously prepared compound.13
4.3.5.1. 6-(6-methyl-benzothiazol-2-yl)-3-methyl-2-thio
methyl-benzothiazolium tosylate 5. Obtained from the
benzothiazole 4 (0.33 g, 1 mmol) using the above-
described procedure (0.21 g, 40%). 1H NMR (methanol-
d4): d 2.35 (3H, s, Ar-CH3), 2.53 (3H, s, Ar-CH3), 3.18
(3H, s, S-CH3), 4.18 (3H, s, NCH3), 7.21 (1H, d, J ¼ 8,
ArH), 7.45 (2H, d, J ¼ 8, ArH), 7.69 (1H, d, J ¼ 8,
ArH), 7.78 (2H, d, J ¼ 8, ArH), 7.87 (1H, s, ArH), 7.96
(1H, d, J ¼ 8, ArH), 8.20 (1H, d, J ¼ 8, ArH), 8.92 (1H,
s, ArH); HR-FAB-MS m=z found: 343.050 C17H15N2S3
(Mþ): requires M, 343.0397.
4.3.6.2. 2-[(3-Methyl-6-(6-methyl-benzothiazol-2-yl)-
2,3,-dihydro-(benzo-1,3-thiazole)-2-methylidene)]-1-methyl-
pyridinium chloride (2-BEBO). Obtained from the ben-
zothiazolium salt 5 (26 mg, 0.05 mmol) and 2-picolinium
tosylate (14 mg, 0.05 mmol) using the above-described
procedure as a yellow solid (9.5 mg, 43%). 1H NMR
(methanol-d4): d 2.46 (3H, s, Ar-CH3), 3.70 (3H, s, N-
CH3), 3.91 (3H, s, N-CH3), 5.56 (1H, s, @CHA), 7.10
(1H, t, J ¼ 8, ArH), 7.30 (1H, d, J ¼ 8, ArH), 7.46 (1H,
d, J ¼ 8, ArH) 7.65 (1H, s, J ¼ 8, ArH), 7.78 (2H, m,
ArH), 8.03 (1H, t, J ¼ 8, ArH), 8.07 (1H, d, J ¼ 8,
ArH), 8.23 (1H, d, J ¼ 7, ArH) 8.26 (1H, s, ArH); HR-
FAB-MS m=z found: 402.104 C23H20N3S2 (Mþ): re-
quires M, 402.1099; UV–vis (sodium phosphate buffer
5 mM, pH 7.0): kmax ðeÞ ¼ 437 (33,500), 374 (11,500),
340 (10,400), 273 nm (11,000 Mꢁ1 cmꢁ1).
4.3.5.2. 6-(Benzothiazol-2-yl)-3-methyl-2-thiomethyl-
benzothiazolium tosylate 14a. Obtained from the
benzothiazole 13a (0.31 g, 1 mmol) using the above-
1
described procedure (0.19 g, 37%). H NMR (DMSO-
d6): d 2.28 (3H, s, Ar-CH3), 3.17 (3H, s, S-CH3), 4.15
(3H, s, NCH3), 7.10 (2H, d, J ¼ 8, ArH), 7.46 (2H, d,
J ¼ 8, ArH), 7.54 (1H, t, J ¼ 8, ArH), 7.62 (1H, t,
J ¼ 8, ArH), 8.13 (1H, d, J ¼ 8, ArH), 8.24 (1H, d, J ¼
8, ArH), 8.35 (1H, d, J ¼ 8, ArH), 8.52 (1H, d, J ¼ 8,
ArH), 9.20 (1H, s, ArH); HR-FAB-MS m=z found:
329.032 C16H13N2S3 (Mþ): requires M, 329.0241.
4.3.6.3. 4-[(3-Methyl-6-(6-methyl-benzothiazol-2-yl)-
2,3,-dihydro-(benzo-1,3-thiazole)-2-methylidene)]-1-meth-
yl-quinolinium chloride (Me-BETO). Obtained from the
benzothiazolium salt 5 (26 mg, 0.05 mmol) and 4-lepid-
inium tosylate (16 mg, 0.05 mmol) using the above-
described procedure as a red solid (15 mg, 61%). 1H
NMR (methanol-d4): d 2.50 (3H, s, Ar-CH3), 4.03 (3H,
s, N-CH3), 4.25 (3H, s, N-CH3), 6.97 (1H, s, @CHA),
4.3.5.3.
6-(Benzoxazol-2-yl)-3-methyl-2-thiomethyl-
benzothiazolium tosylate 14b. Obtained from the
benzothiazole 13b (0.30 g, 1 mmol) using the above-
described procedure (0.21 g, 43%). H NMR (DMSO-
d6): d 2.28 (3H, s, Ar-CH3), 3.18 (3H, s, S-CH3), 4.15
(3H, s, NCH3), 7.10 (2H, d, J ¼ 8, ArH), 7.47 (4H, m,
ArH), 7.88 (2H, m, ArH), 8.41 (1H, d, J ¼ 8, ArH), 8.61
1

2382
H. J. Karlsson et al. / Bioorg. Med. Chem. 12 (2004) 2369–2384
7.36 (1H, d, J ¼ 8, ArH), 7.56 (1H, d, J ¼ 7, ArH),
7.75–7.89 (4H, m, ArH) 8.0–8.10 (2H, m, ArH), 8.27
(1H, d, J ¼ 7, ArH), 8.51 (1H, d, J ¼ 7, ArH), 8.58 (1H,
s, ArH), 8.70 (1H, d, J ¼ 8, ArH); HR-FAB-MS m=z
found: 452.130 C27H22N3S2 (Mþ): requires M, 452.1255;
UV–vis (sodium phosphate buffer 5 mM, pH 7.0):
kmax ðeÞ ¼ 489 (41,500), 356 (11,500), 292 nm
(15,000 Mꢁ1 cmꢁ1).
benzothiazolium salt 14a (25 mg, 0.05 mmol) and 4-le-
pidinium tosylate (16 mg, 0.05 mmol) using the above-
described procedure as a red solid (19 mg, 79%). 1H
NMR (methanol-d4): 4.04 (3H, s, N-CH3), 4.25 (3H, s,
N-CH3), 6.98 (1H, s, @CHA), 7.44 (1H, t, J ¼ 8, ArH),
7.52–7.58 (2H, m, ArH), 7.77 (1H, d, J ¼ 8, ArH), 7.83
(1H, t, J ¼ 8, ArH), 7.99–8.10 (4H, m, ArH), 8.30 (1H,
dd, J1 ¼ 8, J2 ¼ 2, ArH), 8.51 (1H, d, J ¼ 7, ArH), 8.62
(1H, s, ArH), 8.71 (1H, d, J ¼ 8, ArH). These data were
consistent with those of the previously prepared com-
pound.14 However, the shifts are somewhat different,
which is explained by a lower dye concentration in the
present NMR sample compared to the previously re-
ported data (vide supra).
4.3.6.4. 2-[(3-Methyl-6-(6-methyl-benzothiazol-2-yl)-
2,3,-dihydro-(benzo-1,3-thiazole)-2-methylidene)]-1-meth-
yl-quinolinium chloride (Me-2-BETO). Obtained from
the benzothiazolium salt 5 (26 mg, 0.05 mmol) and 2-
lepidinium tosylate (16 mg, 0.05 mmol) using the above-
described procedure as a red-brown solid (17 mg, 68%).
1H NMR (DMSO-d6): d 2.495 (3H, s, Ar-CH3), 3.96
(3H, s, N-CH3), 4.18 (3H, s, N-CH3), 6.24 (1H, s,
@CHA), 7.39 (1H, d, J ¼ 8, ArH), 7.67 (1H, t, J ¼ 8,
ArH), 7.88 (1H, d, J ¼ 8, ArH) 7.96 (3H, m, ArH), 8.08
(1H, d, J ¼ 8, ArH), 8.13 (1H, d, J ¼ 9, ArH), 8.21 (1H,
d, J ¼ 9, ArH), 8.26 (1H, d, J ¼ 9, ArH), 8.57 (1H, d,
J ¼ 9, ArH), 8.76 (1H, s, ArH); HR-FAB-MS m=z
found: 452.137 C27H22N3S2 (Mþ): requires M, 452.1255;
UV–vis (sodium phosphate buffer 5 mM, pH 7.0):
kmax ðeÞ ¼ 459 (34,000), 349 nm (12,500 Mꢁ1 cmꢁ1).
4.3.6.8.
2-[6-(Benzothiazol-2-yl)-(3-methyl-2,3-dihy-
dro-(benzo-1,3-thiazole)-2-methylidene)]-1-methyl-quino-
linium chloride (2-BETO). Obtained from the
benzothiazolium salt 14a (25 mg, 0.05 mmol) and 2-lep-
idinium tosylate (16 mg, 0.05 mmol) using the above-
described procedure as a red-brown solid (21 mg, 89%).
1H NMR (methanol-d4): 3.98 (3H, s, N-CH3), 4.23 (3H,
s, N-CH3), 6.22 (1H, s, @CHA), 7.45 (1H, t, J ¼ 8,
ArH), 7.55 (1H, t, J ¼ 8, ArH), 7.67 (1H, t, J ¼ 8,
ArH), 7.78 (1H, d, J ¼ 8, ArH), 7.96 (1H, t, J ¼ 7,
ArH), 7.99-8.05 (3H, m, ArH), 8.13 (1H, d, J ¼ 9, ArH),
8.22 (1H, d, J ¼ 9, ArH), 8.31 (1H, dd, J1 ¼ 8, J2 ¼ 2,
ArH), 8.43 (1H, d, J ¼ 9, ArH), 8.62 (1H, s, ArH); HR-
FAB-MS m=z found: 438.115 C26H20N3S2 (Mþ): re-
quires M, 438.1099; UV–vis (sodium phosphate buffer
5 mM, pH 7.0): kmax ðeÞ ¼ 468 (37,500), 355 (13,000),
279 nm (13,000 Mꢁ1 cmꢁ1).
4.3.6.5.
4-[6-(Benzothiazol-2-yl)-(3-methyl-2,3-dihy-
dro-(benzo-1,3-thiazole)-2-methylidene)]-1-methyl-pyridi-
nium chloride (H-BEBO). Obtained from the
benzothiazolium salt 14a (25 mg, 0.05 mmol) and
4-picolinium tosylate (14 mg, 0.05 mmol) using the
above-described procedure as a yellow solid (10 mg,
1
47%). H NMR (DMSO-d6): 3.76 (3H, s, N-CH3), 4.03
(3H, s, N-CH3), 6.35 (1H, s, @CHA), 7.48 (3H, m,
ArH), 7.57 (1H, t, J ¼ 8, ArH), 7.71 (1H, d, J ¼ 8,
ArH), 8.05 (1H, d, J ¼ 8, ArH), 8.17 (1H, d, J ¼ 8,
ArH), 8.21 (1H, d, J ¼ 8, ArH), 8.40 (2H, d, J ¼ 7,
ArH), 8.69 (1H, s, ArH); HR-FAB-MS m=z found:
388.095 C22H18N3S2 (Mþ): requires M, 388.0942; UV–
vis (sodium phosphate buffer 5 mM, pH 7.0): kmax ðeÞ ¼
444 (32,500), 339 (9500), 277 nm (6000 Mꢁ1 cmꢁ1).
4.3.6.9. 4-[6-(Benzoxazol-2-yl-(3-methyl-)-2,3-dihydro-
(benzo-1,3-thiazole)-2-methylidene)]-1-methyl-pyridinium
chloride (BOXBO). Obtained from the benzothiazolium
salt 14b (24 mg, 0.05 mmol) and 4-picolinium tosylate
(14 mg, 0.05 mmol) using the above-described procedure
1
as a yellow solid (8 mg, 40%). H NMR (methanol-d4):
3.75 (3H, s, N-CH3), 4.03 (3H, s, N-CH3), 6.17 (1H, s,
@CHA), 7.40–7.43 (5H, m, ArH), 7.59 (1H, d, J ¼ 8,
ArH), 7.64–7.71 (2H, m, ArH), 8.15 (1H, d, J ¼ 7,
ArH), 8.31 (1H, d, J ¼ 8, ArH), 8.54 (1H, s, ArH); HR-
FAB-MS m=z found: 372.112 C22H18N3OS (Mþ):
requires M, 372.1171; UV–vis (sodium phosphate buffer
5 mM, pH 7.0): kmax ðeÞ ¼ 440 (36,500), 329 (6400),
271 nm (11,000 Mꢁ1 cmꢁ1).
4.3.6.6.
2-[6-(Benzothiazol-2-yl)-(3-methyl-2,3-dihy-
dro-(benzo-1,3-thiazole)-2-methylidene)]-1-methyl-pyridi-
nium chloride (H-2-BEBO). Obtained from the
benzothiazolium salt 14a (25 mg, 0.05 mmol) and 2-pic-
olinium tosylate (14 mg, 0.05 mmol) using the above-
described procedure as a yellow solid (16.5 mg, 79%).
1
%). H NMR (methanol-d4): 3.83 (3H, s, N-CH3), 4.11
(3H, s, N-CH3), 5.89 (1H, s, @CHA), 7.24 (1H, t, J ¼ 7,
ArH), 7.45 (1H, t, J ¼ 7, ArH), 7.55 (1H, d, J ¼ 8,
ArH), 7.63 (1H, d, J ¼ 8, ArH), 8.0–8.06 (3H, m, ArH),
8.15 (1H, t, J ¼ 7, ArH), 8.23 (1H, d, J ¼ 8, ArH), 8.41
(1H, d, J ¼ 7 ArH), 8.50 (1H, s, ArH); HR-FAB-MS
m=z found: 388.094 C22H18N3S2 (Mþ): requires M,
388.0942; UV–vis (sodium phosphate buffer 5 mM,
pH 7.0): kmax ðeÞ ¼ 437 (37,000), 371 (10,500), 275 nm
(12,500 Mꢁ1 cmꢁ1).
4.3.6.10. 2-[6-(Benzoxazol-2-yl-(3-methyl-)-2,3-dihy-
dro-(benzo-1,3-thiazole)-2-methylidene)]-1-methyl-pyridi-
nium chloride (2-BOXBO). Obtained from the
benzothiazolium salt 14b (24 mg, 0.05 mmol) and 2-pic-
olinium tosylate (14 mg, 0.05 mmol) using the above-
described procedure as a yellow solid (16.5 mg, 81%). 1H
NMR (DMSO-d6): 3.82 (3H, s, N-CH3), 4.12 (3H, s, N-
CH3), 5.94 (1H, s, @CHA), 7.37 (1H, t, J ¼ 8, ArH),
7.44 (2H, m, ArH), 7.76–7.83 (3H, m, ArH), 7.97 (1H, d,
J ¼ 8, ArH), 8.24–8.32 (2H, m, ArH), 8.62 (1H, d,
J ¼ 7, ArH), 8.74 (1H, s, ArH); HR-FAB-MS m=z
found: 372.107 C22H18N3OS (Mþ): requires M,
372.1171; UV–vis (sodium phosphate buffer 5 mM,
4.3.6.7. 4-[6-(Benzothiazol-2-yl)-(3-methyl-2,3-dihy-
dro-(benzo-1,3-thiazole)-2-methylidene)]-1-methyl-quino-
linium chloride (BETO). Obtained from the

H. J. Karlsson et al. / Bioorg. Med. Chem. 12 (2004) 2369–2384
2383
7. Ishiguro, T.; Saitoh, J.; Yawata, H.; Otsuka, M.; Inoue,
T.; Sugiura, Y. Nucleic Acids Res. 1996, 24, 4992–4997.
8. Seitz, O.; Bergmann, F.; Heindl, D. Angew. Chem., Int.
Ed. 1999, 38, 2203–2206.
9. Gurrieri, S.; Wells, K. S.; Johnson, I. D.; Bustamante, C.
Anal. Biochem. 1997, 249, 44–53.
pH 7.0): kmax ðeÞ ¼ 436 (41,000), 362 (12,000), 323
(11,000), 270 nm (16,000 Mꢁ1 cmꢁ1).
4.3.6.11. 4-[6-(Benzoxazol-2-yl-(3-methyl-)-2,3-dihy-
dro-(benzo-1,3-thiazole)-2-methylidene)]-1-methyl-quino-
linium chloride (BOXTO). Obtained from the
benzothiazolium salt 14b (24 mg, 0.05 mmol) and 4-
lepidinium tosylate (16 mg, 0.05 mmol) using the above-
described procedure as a red solid (12 mg, 53%). 1H
NMR (methanol-d4): 4.05 (3H, s, N-CH3), 4.28 (3H, s,
N-CH3), 7.02 (1H, s, @CHA), 7.40–7.44 (2H, m, ArH),
7.61 (1H, d, J ¼ 6, ArH), 7.66–7.77 (2H, m, ArH), 7.81–
7.87 (2H, m, ArH), 8.05 (1H, t, J ¼ 8, ArH), 8.12 (1H,
d, J ¼ 8, ArH), 8.45 (1H, dd, J1 ¼ 8, J2 ¼ 2, ArH), 8.55
(1H, d, J ¼ 8, ArH), 8.73 (1H, s, ArH), 8.74 (1H, d,
J ¼ 8, ArH). These data were consistent with those
of the previously prepared compound.14 However,
the shifts are somewhat different, which is explained
by a lower dye concentration in the present NMR
sample compared to the previously reported data (vide
supra).
10. Bengtsson, M.; Karlsson, H. J.; Westman, G.; Kubista, M.
Nucleic Acids Res. 2003, 31, e45.
11. Yin, J. L.; Shackel, N. A.; Zekry, A.; McGuinness, P. H.;
Richards, C.; Putten, K. V.; McCaughan, G. W.; Eris, J.
M.; Bishop, G. A. Immunol. Cell. Biol. 2001, 79, 213–221.
12. Svanvik, N.; Stahlberg, A.; Sehlstedt, U.; Sjoback, R.;
Kubista, M. Anal. Biochem. 2000, 287, 179–182.
13. Karlsson, H. J.; Lincoln, P.; Westman, G. Bioorg. Med.
Chem. 2003, 11, 1035–1040.
ꢁ
14. Karlsson, H. J.; Eriksson, M.; Perzon, E.; Akerman, B.;
Lincoln, P.; Westman, G. Nucleic Acids Res. 2003, 31,
6227–6234.
15. Frau, S.; Bernadou, J.; Meunier, B. Bull. Soc. Chim. Fr.
1996, 133, 1053–1070.
16. Pjura, P. E.; Grzeskowiak, K.; Dickerson, R. E. J. Mol.
Biol. 1987, 197, 257–271.
17. Kapuscinski, J. Biotech. Histochem. 1995, 70, 220–233.
18. Cosa, G.; Focsaneanu, K. S.; McLean, J. R. N.; McNa-
mee, J. P.; Scaiano, J. C. Photochem. Photobiol. 2001, 73,
585–599.
4.3.6.12. 2-[6-(Benzoxazol-2-yl-(3-methyl-)-2,3-dihy-
dro-(benzo-1,3-thiazole)-2-methylidene)]-1-methyl-quino-
linium chloride (2-BOXTO). Obtained from the
benzothiazolium salt 14b (24 mg, 0.05 mmol) and 2-
lepidinium tosylate (16 mg, 0.05 mmol) using the above-
described procedure as a red-brown solid (17 mg, 76%).
1H NMR (methanol-d4): 3.99 (3H, s, N-CH3), 4.24 (3H,
s, N-CH3), 6.24 (1H, s, @CHA), 7.40–7.47 (2H, m,
ArH), 7.66–7.79 (3H, m, ArH), 7.82 (1H, d, J ¼ 8,
ArH), 7.97 (1H, t, J ¼ 8, ArH), 8.03 (1H, d, J ¼ 8,
ArH), 8.15 (1H, d, J ¼ 9, ArH), 8.24 (1H, d, J ¼ 9,
ArH), 8.41-8.49 (2H, m, ArH), 8.70 (1H, s, ArH); HR-
FAB-MS m=z found: 422.127 C26H20N3OS (Mþ):
requires M, 422.1327; UV–vis (sodium phosphate buffer
5 mM, pH 7.0): kmax ðeÞ ¼ 463 (37,500), 344 (13,500),
274 nm (15,500 Mꢁ1 cmꢁ1).
19. Bordelon, J. A.; Feierabend, K. J.; Siddiqui, S. A.; Wright,
L. L.; Petty, J. T. J. Phys. Chem. B 2002, 106, 4838–4843.
20. Matsuzawa, Y.; Yoshikawa, K. Nucleos. Nucleot. 1994,
13, 1415–1423.
21. Wang, A. H. J.; Ughetto, G.; Quigley, G. J.; Rich, A.
Biochemistry 1987, 26, 1152–1163.
22. Perkins, T. T.; Smith, D. E.; Chu, S. In Flexible Polymer
Chains in Elongational Flow; Nguyen, T. Q., Kausch,
H.-H., Eds.; Springer, 1999; pp 283–334.
23. Neidle, S. Biopolymers 1997, 44, 105–121.
24. Dervan, P. B. Bioorg. Med. Chem. 2001, 9, 2215–2235.
25. Mikheikin, A. L.; Zhuze, A. L.; Zasedatelev, A. S.
J. Biomol. Struct. Dyn. 2000, 18, 59–72.
26. Seifert, J. L.; Connor, R. E.; Kushon, S. A.; Wang, M.;
Armitage, B. A. J. Am. Chem. Soc. 1999, 121, 2987–2995.
27. Sovenyhazy, K. M.; Bordelon, J. A.; Petty, J. T. Nucleic
Acids Res. 2003, 31, 2561–2569.
28. Chaudhuri, N. C. Synth. Commun. 1996, 26, 3783–
3790.
29. Roche, D.; Prasad, K.; Repic, O.; Blacklock, T. J.
Tetrahedron Lett. 2000, 41, 2083–2085.
Acknowledgements
We are grateful to Dr. Gunnar Stenhagen at Stenhagen
analys AB for high resolution mass measurements, and
Kristin Samuelson for help concerning the quaterniza-
tion of benzothiazoles.
30. Khurana, J. M.; Sahoo, P. K. Synth. Commun. 1992, 22,
1691–1702.
31. Kendall, J. D.; Suggate, H. G. J. Chem. Soc. 1949, 1503–
1509.
32. Lee, C. L.; Lam, Y. L.; Lee, S. Y. Tetrahedron Lett. 2001,
42, 109–111.
References and notes
33. Deligeorgiev, T. G. Dyes Pigments 1990, 12, 243–248.
34. Hein, D. W.; Alheim, R. J.; Leavitt, J. J. J. Am. Chem.
Soc. 1957, 79, 427–429.
35. Brembilla, A.; Roizard, D.; Lochon, P. Synth. Commun.
1990, 20, 3379–3384.
36. Tale, R. H. Org. Lett. 2002, 4, 1641–1642.
37. Haugwitz, R. D.; Angel, R. G.; Jacobs, G. A.; Maurer,
B. V.; Narayanan, V. L.; Cruthers, L. R.; Szanto, J.
J. Med. Chem. 1982, 25, 969–974.
38. Shi, D. F.; Bradshaw, T. D.; Wrigley, S.; McCall, C. J.;
Lelieveld, P.; Fichtner, I.; Stevens, M. F. G. J. Med. Chem.
1996, 39, 3375–3384.
1. Reddy, B. S. P.; Sondhi, S. M.; Lown, J. W. Pharmacol.
Therapeut. 1999, 84, 1–111.
2. Rye, H. S.; Yue, S.; Wemmer, D. E.; Quesada, M. A.;
Haugland, R. P.; Mathies, R. A.; Glazer, A. N. Nucleic
Acids Res. 1992, 11, 2803–2812.
3. Carlsson, C.; Larsson, A.; Jonsson, M.; Albinsson, B.;
ꢀ
Norden, B. J. Phys. Chem. 1994, 98, 10313–10321.
4. Hirons, G. T.; Fawcett, J. J.; Crissman, H. A. Cytometry
1994, 15, 129–140.
5. Yan, X.; Grace, W.; Yoshida, T.; Habbersett, R.; Velap-
pan, N.; Jett, J.; Keller, R.; Marrone, B. Anal. Chem. 1999,
71, 5470–5480.
39. Srivastava, R. G.; Venkataramani, P. S. Synth. Commun.
1988, 18, 1537–1544.
6. Svanvik, N.; Westman, G.; Wang, D.; Kubista, M. Anal.
Biochem. 2000, 281, 26–35.
40. Park, K. H.; Jun, K.; Shin, S. R.; Oh, S. W. Tetrahedron
Lett. 1996, 37, 8869–8870.

2384
H. J. Karlsson et al. / Bioorg. Med. Chem. 12 (2004) 2369–2384
41. Varma, R. S.; Saini, R. K.; Prakash, O. Tetrahedron Lett.
1997, 38, 2621–2622.
42. Varma, R. S.; Kumar, D. J. Heterocycl. Chem. 1998, 35,
1539–1540.
43. Cho, C. S.; Kim, D. T.; Zhang, J. Q.; Ho, S. L.; Kim, T. J.;
Shim, S. C. J. Heterocycl. Chem. 2002, 39, 421–423.
44. Chang, J. B.; Zhao, K.; Pan, S. F. Tetrahedron Lett. 2002,
43, 951–954.
54. West, W.; Pearce, S. J. Phys. Chem. 1965, 69, 1894–1903.
55. Nygren, J.; Svanvik, N.; Kubista, M. Biopolymers 1998,
46, 39–51.
56. Ogulꢂchansky, T. Y.; Losytskyy, M. Y.; Kovalska, V. B.;
Yashchuk, V. M.; Yarmoluk, S. M. Spectrochim. Acta A
2001, 57, 1525–1532.
57. Khairutdinov, R. F.; Serpone, N. J. Phys. Chem. B 1997,
101, 2602–2610.
ꢀ
45. Norden, B.; Kubista, M.; Kurucsev, T. Quart. Rev.
Biophys. 1992, 25, 51–170.
58. Isacsson, J.; Westman, G. Tetrahedron Lett. 2001, 42,
3207–3210.
59. Larsson, A.; Carlsson, C.; Jonsson, M. Biopolymers 1995,
36, 153–167.
60. Goodsell, D.; Dickerson, R. E. J. Med. Chem. 1986, 29,
727–733.
ꢁ
46. Kubista, M.; Akerman, B.; Norden, B. Biochemistry 1987,
ꢀ
26, 4545–4553.
47. Bailly, C.; Colson, P.; Henichart, J. P.; Houssier, C.
Nucleic Acids Res. 1993, 21, 3705–3709.
48. Larsson, A.; Carlsson, C.; Jonsson, M.; Albinsson, B.
J. Am. Chem. Soc. 1994, 116, 8459–8465.
61. Goodsell, D. S.; Dickerson, R. E. Nucleic Acids Res. 1994,
22, 5497–5503.
49. Garoff, R. A.; Litzinger, E. A.; Connor, R. E.; Fishman,
I.; Armitage, B. A. Langmuir 2002, 18, 6330–6337.
ꢀ
50. Lyng, R.; Rodger, A.; Norden, B. Biopolymers 1992, 32,
1201–1214.
51. Eriksson, M.; Norden, B. Linear and Circular Dichroism
of Drug–Nucleic Acid Complexes. In Drug–Nucleic Acid
Interactions; 2001; Vol. 340, pp 68–98.
52. Tanious, F. A.; Ding, D. Y.; Patrick, D. A.; Tidwell, R.
R.; Wilson, W. D. Biochemistry 1997, 36, 15315–15325.
53. Jorgenson, K. F.; Varshney, U.; van de Sande, J. H.
J. Biomol. Struct. Dyn. 1988, 5, 1005–1023.
62. Zasedatelev, A. S. FEBS Lett. 1991, 281, 209–211.
63. Stevens, M. F. G.; Shi, D. F.; Castro, A. J. Chem. Soc.
Perkin Trans. 1 1996, 83–93.
64. Vogel, A. I. Textbook of Practical Organic Chemistry,
5th ed.; Longman Scientific
p 793.
&
Technical: 1989;
65. Moriyama, H.; Hiramatsu, Y.; Kiyoi, T.; Achiha, T.;
Inoue, Y.; Kondo, H. Bioorg. Med. Chem. 2001, 9, 1479–
1491.
66. Zhang, J. Z.; Zhu, Q.; Huang, X. Synth. Commun. 2002,
32, 2175–2179.