Synthesis and effect of heterocycle modification on the spectroscopic properties of a series of unsymmetrical trimethine cyanine dyes

Article abstract of DOI:10.1016/j.dyepig.2014.02.009

Carbocyanine dyes are a class of organic compounds that contain two heterocycles that act as electron donors and acceptors connected by a conjugated methine bridge. Herein the synthesis of a series of 16 novel unsymmetrical trimethine cyanine dyes is reported. Their structures were characterized by various spectroscopic techniques, and their optical properties were measured. Absorption maxima of the dyes were calculated using the time-dependent density-functional theory method and the computational absorption maxima are consistent with the experimental data. The addition of electron withdrawing substituents such as halogens on the heterocycle gave more favorable optical properties such as higher quantum yield and molar absorptivity. The aggregation of these cyanine dyes was studied and compared to a similar series of symmetric cyanine dyes. It was determined that the heterocycle has more effect on aggregation than the side chain and a dye with two different heterocycles will aggregate less than a dye with the same heterocycle. The dyes were also investigated for Lipinski Rule violations as their use is becoming more prevalent for in vivo applications.

Full text of DOI:10.1016/j.dyepig.2014.02.009

Dyes and Pigments 105 (2014) 238e249
Contents lists available at ScienceDirect
Dyes and Pigments
journal homepage: www.elsevier.com/locate/dyepig
Synthesis and effect of heterocycle modification on the spectroscopic
properties of a series of unsymmetrical trimethine cyanine dyes
,
,b,
Andrew Levitz a, Safieh Tork Ladani a, Donald Hamelberg a b, Maged Henary a
*
a Department of Chemistry, Center for Biotechnology and Drug Design, Georgia State University, Atlanta, GA, USA
b Center for Diagnostics and Therapeutics, 100 Piedmont Ave SE, Atlanta, GA, USA
a r t i c l e i n f o
a b s t r a c t
Article history:
Carbocyanine dyes are a class of organic compounds that contain two heterocycles that act as electron
donors and acceptors connected by a conjugated methine bridge. Herein the synthesis of a series of 16
novel unsymmetrical trimethine cyanine dyes is reported. Their structures were characterized by various
spectroscopic techniques, and their optical properties were measured. Absorption maxima of the dyes
were calculated using the time-dependent density-functional theory method and the computational
absorption maxima are consistent with the experimental data. The addition of electron withdrawing
substituents such as halogens on the heterocycle gave more favorable optical properties such as higher
quantum yield and molar absorptivity. The aggregation of these cyanine dyes was studied and compared
to a similar series of symmetric cyanine dyes. It was determined that the heterocycle has more effect on
aggregation than the side chain and a dye with two different heterocycles will aggregate less than a dye
with the same heterocycle. The dyes were also investigated for Lipinski Rule violations as their use is
becoming more prevalent for in vivo applications.
Received 19 December 2013
Received in revised form
8 February 2014
Accepted 9 February 2014
Available online 18 February 2014
Keywords:
Cyanine dye
Unsymmetrical
Trimethine
Synthesis
Optical properties
DFT calculation
Ó 2014 Elsevier Ltd. All rights reserved.
1. Introduction
different, the absorption would not be at the midpoint [16]. If the
basicity of the two nuclei is not equal the absorption should be
Cyanine dyes possess two nitrogen containing heterocyclic
groups that are connected by a conjugated methine bridge as
shown in Fig. 1. The delocalization of electrons across this chain
causes them to be highly fluorescent and exhibit long wavelength
absorption that span from the visible to the near infrared regions
[1e4]. These unique properties as well as their high extinction
coefficients have allowed cyanine dyes to be employed in various
applications. Beginning in the 1800s, cyanine dyes were used in
photographic emulsions and chemotherapy [5], but more recently
cyanine dyes have been used in laser printing [6], pH sensors [7],
fluorescence in vivo imaging [8e11], data storage [12], and as labels
for nucleic acid detection [13e15].
Until the mid-1900s, it was widely accepted that an unsym-
metrical cyanine dye would absorb halfway between the absorp-
tion of the parent symmetrical dyes. Brooker et al. showed that if
the basicities of the nitrogen containing heterocycles are not
identical, or if the relative stabilities of the two forms are not
found at a lower wavelength than the intermediate position. As the
difference in basicity is continually increased, the absorption
maximum gets further from the midpoint, and this deviation in-
creases with an increasing length of the polymethine chain [16].
Although many unsymmetrical cyanine dyes have been made [16],
there is a lack in the literature of cyanine dyes with indolenine
moieties and those with substitutions on the benzene ring of the
indolenine are sporadic.
When designing new dyes for specific applications, it is
important to understand how altering the parent structure, such as
the substituents on the heterocyclic moieties, affects the optical
properties of the dyes. These changes can be made to improve
solubility or increase binding and permeability, but structural
modifications also cause changes in optical properties. Herein we
report the synthesis of a series of unsymmetrical trimethine
cyanine dyes prepared with various heterocyclic modifications to
determine how these substitutions affect the optical properties of
the dyes including molar absorptivity, fluorescence quantum yield,
and stokes shift. Absorption maxima are compared to those of the
symmetrical dyes previously reported [17].
* Corresponding author. Department of Chemistry, Center for Biotechnology and
Drug Design, Georgia State University, Atlanta, GA, USA. Tel.: þ1 404 413 5566;
fax: þ1 404 413 5505.
A popular application of cyanine dyes is for imaging and in vivo
labeling. Lipinski’s Rule of Five is used to evaluate if a chemical
E-mail address: mhenary1@gsu.edu (M. Henary).
http://dx.doi.org/10.1016/j.dyepig.2014.02.009
0143-7208/Ó 2014 Elsevier Ltd. All rights reserved.

A. Levitz et al. / Dyes and Pigments 105 (2014) 238e249
239
2.2.4. 2-Methyl-1-(3-(trimethylammonio)propyl)-3H-indol-1-ium
bromide (5a) and 2-methyl-1-(3-(trimethylammonio)propyl)-3H-
indol-1-ium-5-sulfonate bromide (5h)
2-Methyl-1-(3-(trimethylammonio)propyl)-3H-indol-1-ium
bromide (5a) and 2-Methyl-1-(3-(trimethylammonio)propyl)-3H-
indol-1-ium-5-sulfonate bromide (5h) were previously synthe-
sized by our group [8].
Fig. 1. General structure of cyanine dyes.
compound has properties that would make it orally active in
humans [18]. These properties include molecular weight, hydrogen
bond donors and acceptors, a partition coffecient, and polar surface
area [19]. Partition coefficients, such as log P, are ratios of concen-
trations of a compound in a mixture of two immiscible phases.
Log D is much like log P, but it also takes into account positions that
can be fully protonated or deprotonated such as sulfonate or
carboxylate groups. Based on log D, it can be speculated as to where
a compound will distribute in biological tissues. The Lipinski Rule of
Five properties of each dye were investigated and the log D value of
each dye was predicted to give a better idea of how each substit-
uent affects partitioning of the dye and where it will distribute in
biological tissues.
Previously, a study comparing the optical properties of six
symmetrical and unsymmetrical tricarbocyanine dyes containing
benzothiazole and benzoxazole heterocycles was presented [20].
The authors concluded that the replacement of one sulfur atom
with one oxygen atom blue shifted the absorption wavelength
about 20 nm. Our recent work presented in this paper is similar but
takes into account dyes with various alkyl chains as well as sub-
stitutions on a number of different heterocycles including benz[c,d]
indole, benzoxazole and benzothiazole.
2.2.5. 3-(2,3,3-Trimethyl-3H-indol-1-ium-1-yl)propane-1-
sulfonate (5b)
3-(2,3,3-Trimethyl-3H-indol-1-ium-1-yl)propane-1-sulfonate
(5b) was synthesized following the procedure published by Tang
et al. in 82% yield; mp 217e221 ꢀC (lit mp 165e166 ꢀC [23]).
2.2.6. 1-Ethyl-2,3,3-trimethyl-3H-indol-1ium iodide (5c), 1-butyl-
2,3,3-trimethyl-3H-indol-1ium iodide (5d), 3-butyl-2-methyl-1H-
benzo[e]indol-3-ium iodide (5i) and 1-butyl-2-methylbenzo[cd]
indol-1-ium iodide (7a)
1-Ethyl-2,3,3-trimethyl-3H-indol-1ium iodide (5c), 1-Butyl-
2,3,3-trimethyl-3H-indol-1ium iodide (5d), 3-Butyl-2-methyl-
1H-benzo[e]indol-3-ium iodide (5i) and 1-Butyl-2-methylbenzo
[cd]indol-1-ium iodide (7a) were previously synthesized by our
group [24].
2.2.7. 2,3,3-Trimethyl-1-(3-phenylpropyl)-3H-indol-1-ium
bromide, (5e)
A mixture of 2,3,3-trimethyl-3H-indole and 3-(bromopropyl)
benzene were refluxed for 72 h. The reaction was cooled to room
temperature, poured into cold ether, and the precipitate was
filtered; mp 156e158 ꢀC, 86% yield; 1H NMR (400 MHz, DMSO-d6):
d
1.53 (s, 6 H), 2.13e2.20 (m, 2 H), 2.80e2.83 (m, 5 H), 4.51 (t,
J ¼ 8.0 Hz, 2 H), 7.19e7.32 (m, 5 H), 7.61e7.63 (m, 2 H), 7.84 (t,
2. Experimental
J ¼ 5.2 Hz, 1 H), 7.98 (t, J ¼ 5.2 Hz, 1 H); 13C NMR (100 MHz, DMSO-
d6)
d 14.3, 22.0, 28.8, 31.7, 47.4, 54.1, 115.4, 123.5, 126.0, 128.2, 128.3,
2.1. General information
128.8, 129.3, 140.6,141.0, 141.8, 196.6 HRMS: Calcd for C20H24Nþ m/z
278.1904, obsd m/z 278.1915.
Most reagents were purchased from SigmaeAldrich and were
used without purification. Absorption spectra were recorded on a
Cary 3G UVeVisible Spectrophotometer (Santa Clara, CA) in DMSO
for dyes and 0.1 M NaOH/water solution for fluorescein standard
using VWR disposable two-sided polystyrene cuvettes with a path-
length of 1 cm. Fluorescence (LIF) emission analyses were performed
using a K2 Multifrequency Phase Spectrofluorometer (ISS Inc.,
Champaign, Il) in DMSO for dyes and 0.1 M NaOH/water solution for
the fluorescein standard using SigmaeAldrich disposable poly-
styrene fluorimeter cuvettes with a pathlength of 1 cm. Excitation
was achieved with a 300 W Excelitas Short Arc Xenon Lamp (Fre-
mont, CA) at 645 nm. Slit widths were set to 2 mm and integration
time of 3 s. Microsoft Excel 2010 was used for all calculations.
2.2.8. 5-Bromo-1-butyl-2,3,3-trimethyl-3H-indol-1-ium iodide,
(5f)
Compound 5f was synthesized by a similar procedure to 5e;
mp 176e178 ꢀC, 78% yield; 1H NMR (400 MHz, DMSO-d6):
d 0.93
(t, J ¼ 7.2 Hz, 3 H), 1.39e1.47 (m, 2 H), 1.55 (s, 6 H), 1.76e1.83 (m,
2 H), 2.85 (s, 3 H), 4.45 (t, J ¼ 7.2 H, 2 H), 4.43 (t, J ¼ 7.6 Hz, 2 H),
7.83 (d, J ¼ 8.4 Hz, 1 H), 7.97 (d, J ¼ 8.4 Hz, 1 H), 8.19 (s, 1 H); 13C
NMR (100 MHz, DMSO-d6):
d 14.08, 14.83, 19.79, 22.31, 29.68,
48.23, 54.92, 117.99, 123.28, 127.39, 132.32, 140,89, 144.65,
197.43; HRMS: Calcd for C15H21NBrþ m/z 294.0852, obsd m/z
294.0853.
2.2.9. 1-Butyl-5-methoxy-2,3,3-trimethyl-3H-indol-1-ium iodide,
(5g)
2.2. Synthesis
Compound 5g was synthesized by a similar procedure to 5e; mp
2.2.1. 5-Methoxy-2,3,3-trimethyl-indole, (4b)
145e147 ꢀC, 56% yield; 1H NMR (400 MHz, DMSO-d6):
d 0.93 (t,
5-Methoxy-2,3,3-trimethyl-indole (4b) was synthesized
following the procedure published by Zimmerman and Hennig in
92% yield (lit 95% yield [21]).
J ¼ 7.2 Hz, 3 H), 1.37e1.46 (m, 2 H), 1.53 (s, 6 H), 1.76e1.84 (m, 2 H),
2.80 (s, 3 H), 3.86 (s, 3 H), 4.43 (t, J ¼ 7.6 Hz, 2 H), 7.13 (dd,
J ¼ 10.8 Hz, 1 H), 7.50 (s, 1 H), 7.89 (d, J ¼ 9.2 Hz, 1 H); 13C NMR
(100 MHz, MeOD-d4):
d 12.44, 19.14, 21.61, 29.09, 47.93, 53.81,
2.2.2. Sodium 2,3,3-trimethyl-3H-indole-5-sulfonate (4c)
Sodium 2,3,3-trimethyl-3H-indole-5-sulfonate (4c) was previ-
ously synthesized by our group [8].
55.78, 108.43, 114.50, 116.04, 133.49, 143.54, 160.71, 192.31; HRMS:
Calcd for C16H24NOþ m/z 246.1852, obsd m/z 246.1857.
2.2.10. 5-Bromo-2-methyl-1-(3-(trimethylammonio)propyl)-3H-
indol-1-ium bromide (5j)
5-Bromo-2-methyl-1-(3-(trimethylammonio)propyl)-3H-indol-
1-ium bromide (5j) was previously synthesized by our group [22].
2.2.3. 5-Bromo-2,3,3-trimethyl-3H-indole (4d)
5-Bromo-2,3,3-trimethyl-3H-indole (4d) was previously syn-
thesized by our group [22].

240
A. Levitz et al. / Dyes and Pigments 105 (2014) 238e249
2.2.11. 2,3-Dimethylbenzo[d]thiazol-3-ium iodide (6a)
2,3-Dimethylbenzo[d]thiazol-3-ium iodide (6a) was synthe-
sized following the procedure published by Xiang et al. in 84% yield,
mp 219 ꢀC (lit mp 218e220 ꢀC, 88% yield [25]).
4.19 (t, J ¼ 7.6 Hz, 2 H), 6.60 (d, J ¼ 7.2 Hz, 2 H), 7.31 (t, J ¼ 7.2 Hz, 2
H), 7.37 (d, J ¼ 7.6, 2 H), 7.46 (t, J ¼ 7.6 Hz, 2 H), 7.56 (dd, J ¼ 6.8 Hz, 2
H), 8.55 (t, J ¼ 6.8 Hz, 1 H); 13C NMR (100 MHz, MeOD-d4)
d 12.95,
19.86, 26.91, 27.03, 29.33, 30.87, 43.90, 49.15, 49.22, 102.56, 102.70,
110.90, 111.11, 122.04, 122.19, 125.31, 128.56, 128.61, 140.71, 140.83,
142.01,142.70,150.69,174.50,175.13. HRMS: Calcd for C28H35Nþ2 m/z
399.2795, obsd m/z 399.2810.
2.2.12. 3-Butyl-2-methylbenzo[d]thiazol-3-ium iodide (6c)
3-Butyl-2-methylbenzo[d]thiazol-3-ium iodide (6c) was syn-
thesized following the procedure published by Nigrey and Garito in
94% yield, mp 115e116 ꢀC (lit mp 114e115 ꢀC, 67% yield [26]).
2.2.17. 2-((1E,3E)-3-(3,3-Dimethyl-1-(3-phenylpropyl)indolin-2-
ylidene)prop-1-en-1-yl)-1,3,3-trimethyl-3H-indol-1-ium bromide
(8e)
2.2.13. 2-((1E,3E)-3-(3,3-Dimethyl-1-(3-(trimethylammonio)
propyl)indolin-2-ylidene)prop-1-en-1-yl)-1,3,3-trimethyl-3H-indol-
1-ium bromide (8a)
Compound 8e was synthesized by a similar procedure to 8a. The
dye was purified by column chromatography (4% methanol/
dichloromethane). This compound was obtained in 60.2% yield; mp
A mixture of 3 and 5a were heated at 100 ꢀC for one hour in
acetic anhydride. The solution was allowed to cool to room tem-
perature and then poured into cold diethyl ether. The liquid was
decanted and more ether and the solid was collected by vacuum
filtration. The dye was purified by reverse phase column chroma-
tography (95% methanol/water). This compound was obtained in
134e135 ꢀC; 1H NMR (400 MHz, MeOD-d4):
d 1.75 (s, 6 H), 1.77 (s, 6
H), 2.13e2.21 (m, 2 H), 2.84 (t, J ¼ 7.6 Hz, 2 H), 3.71 (s, 3 H), 4.16 (t,
J ¼ 7.6 Hz, 2 H), 6.21 (d, J ¼ 13.6 Hz, 1 H), 6.29 (d, J ¼ 13.6 Hz, 1 H),
7.22 (d, J ¼ 8.0 Hz,1 H), 7.29e7.31 (m, 4 H), 7.33e7.35 (m, 3 H), 7.37e
7.42 (m, 3 H), 7.45e7.49 (m, 2 H), 7.55 (t, J ¼ 6.8 Hz, 2 H), 8.51 (t,
54% yield; mp 188e190 ꢀC; 1H NMR (400 MHz, MeOD-d4):
d
1.78 (s,
J ¼ 13.6 Hz, 1 H); 13C NMR (100 MHz, MeOD-d4):
d 22.82, 26.75,
6 H), 1.81 (s, 6 H), 2.36 (t, J ¼ 7.8 Hz, 2 H), 3.28 (s, 9 H), 3.78 (s, 3 H),
3.86 (t, J ¼ 8, 2 H), 4.28 (t, J ¼ 8 Hz, 2 H), 6.89e6.93 (m, 2 H), 7.29e
7.37 (m, 2 H), 7.42 (d, J ¼ 13.6 Hz, 1 H), 7.45e7.47 (m, 2 H), 7.48e7.59
(m, 3 H), 8.59 (t, J ¼ 13.6, 1 H); 13C NMR (100 MHz, MeOD-d4):
26.89, 28.53, 30.49, 32.26, 43.02, 49.22, 102.19, 102.42, 110.87,
110.90, 121.99, 122.13, 125.32, 125.41, 126.05, 128.25, 128.31, 128.56,
140.71, 140.76, 141.87, 142.66, 150.61, 174.41, 175.29. HRMS: Calcd
for C33H37Nþ2 m/z 461.2952, obsd m/z 461.2939.
d
20.93, 26.71, 27.13, 31.06, 40.40, 49.01, 49.41, 52.67, 63.02, 102.38,
104.02, 110.63, 111.25, 122.01, 122.20, 125.11, 125.75, 128.61, 128.67,
140.57,140.96,141.73,142.60,151.16,173.91,176.06. HRMS: Calcd for
C
2.2.18. 2-((1E,3E)-3-(5-Bromo-1-butyl-3,3-dimethylindolin-2-
ylidene)prop-1-en-1-yl)-1,3,3-trimethyl-3H-indol-1-ium iodide (8f)
Compound 8f was synthesized by a similar procedure to 8a. No
purification was needed. This compound was obtained in 86% yield;
30H41N23þ m/z 221.6645, obsd m/z 221.6626.
2.2.14. 3-((E)-3,3-Dimethyl-2-((E)-3-(1,3,3-trimethyl-3H-indol-1-
ium-2-yl)allylidene)indolin-1-yl)propane-1-sulfonate (8b)
mp 141e143 ꢀC; 1H NMR (400 MHz, DMSO-d6):
d
0.94 (t, J ¼ 7.2 Hz,
3 H), 1.36e1.46 (m, 2 H), 1.68e1.70 (m, 14 H), 3.68 (s, 3 H), 4.07 (t,
J ¼ 7.2 Hz, 2 H), 6.47 (d, J ¼ 13.6 Hz, 1 H), 6.55 (d, J ¼ 13.6 Hz, 1 H),
7.33 (t, J ¼ 7.2 Hz,1 H), 7.37 (d, J ¼ 8.4 Hz,1 H), 7.50 (m, 2 H), 7.61 (dd,
J ¼ 6.4 Hz, 1 H), 7.65 (d, J ¼ 7.2 Hz, 1 H), 7.91 (s, 1 H), 8.32 (t,
Compound 8b was synthesized by a similar procedure to 8a. The
dye was purified by reverse phase column chromatography (0.5%
HCl/methanol). This compound was obtained in 38% yield; decomp
168 ꢀC; 1H NMR (400 MHz, MeOD-d4):
d
1.78 (s, 12 H), 2.27 (t,
J ¼ 13.6 Hz, 1 H); 13C NMR (100 MHz, DMSO-d6):
d 14.28, 20.00,
J ¼ 7.2 Hz, 2 H), 3.30 (t, J ¼ 6.8 Hz, 2 H), 3.71 (s, 3 H), 4.382 (t,
J ¼ 7.2 Hz, 2 H), 6.50 (d, J ¼ 6.8 Hz, 1 H), 6.60 (d, J ¼ 6.8 Hz, 1 H),
7.29e7.38 (m, 3 H), 7.48 (t, 6.8 Hz, 3 H), 7.55 (d, J ¼ 6.8 Hz, 2 H), 8.59
27.63, 27.83, 29.48, 32.11, 44.16, 49.30, 49.56, 102.64, 104.26, 112.30,
113.60, 117.61, 122.91, 126.06, 126.21, 129.09, 131.81, 141.24, 141.88,
143.03, 143.25, 150.16, 173.24, 176.61. HRMS: Calcd for C28H34BrN2þ
m/z 479.1900, obsd m/z 479.1881.
(t, J ¼ 7.6 Hz, 1 H); 13C NMR (100 MHz, MeOD-d4):
d 22.75, 26.77,
26.95, 30.39, 42.57, 49.19, 49.22, 102.47, 102.78, 110.90, 111.04,
121.93, 122.04, 125.25, 125.36, 128.53, 128.64, 140.76, 141.86, 142.71,
150.91, 174.57, 175.42. HRMS: Calcd for C27H32N2O3S m/z 465.2167,
obsd m/z 465.2231.
2.2.19. 2-((1E,3E)-3-(1-Butyl-5-methoxy-3,3-dimethylindolin-2-
ylidene)prop-1-en-1-yl)-1,3,3-trimethyl-3H-indol-1-ium iodide
(8g)
Compound 8g was synthesized by a similar procedure to 8a. No
purification was needed. This compound was obtained in 77% yield;
2.2.15. 2-((1E,3E)-3-(1-Ethyl-3,3-dimethylindolin-2-ylidene)prop-
1-en-1-yl)-1,3,3-trimethyl-3H-indol-1-ium iodide (8c)
Compound 8c was synthesized by a similar procedure to 8a. The
dye was purified by chromatotron (1% methanol/dichloromethane).
This compound was obtained in 25.6% yield; mp 139e141 ꢀC; 1H
mp 136e138 ꢀC; 1H NMR (400 MHz, DMSO-d6):
d
0. 94 (t, J ¼ 7.2 Hz,
3 H),1.37e1.46 (m, 2 H),1.68 (s, 6 H),1.69 (s, 6 H),1.69e1.70 (m, 2 H),
3.61 (s, 3 H), 3.82 (s, 3 H), 4.12 (t, J ¼ 7.8 Hz, 2 H), 6.38 (d, J ¼ 13.6 Hz,
1 H), 6.47 (d, J ¼ 13.6 Hz, 1 H), 7.00 (d, J ¼ 8.4 Hz, 1 H), 7.26 (t,
J ¼ 7.2 Hz, 1 H), 7.32 (s, 1 H), 7.40e7.45 (m, 3 H), 7.61 (d, J ¼ 7.2 Hz, 1
NMR (400 MHz, MeOD-d4):
d
1.44 (t, J ¼ 7.2 Hz, 3 H), 1.78 (s, 12 H),
3.70 (s, 3 H), 4.23 (q, J ¼ 7.2 Hz, 2 H), 6.47e6.53 (m, 2 H), 7.33 (t,
J ¼ 7.6 Hz, 2 H), 7.37 (d, J ¼ 7.6 Hz, 2 H), 7.46 (t, J ¼ 7.2 Hz, 2 H), 7.55e
7.57 (m, 2 H), 8.57 (t, J ¼ 13.2 Hz, 1 H); 13C NMR (100 MHz, MeOD-
H), 8.29 (t, J ¼ 13.6 Hz, 2 H); 13C NMR (100 MHz, DMSO-d6):
d 14.27,
20.01, 27.81, 27.89, 29.75, 31.62, 44.43, 48.97, 49.74, 56.37, 102.34,
103.21, 109.69, 111.51, 113.05, 114.16, 122.81, 125.19, 128.98, 135.68,
140.76, 143.04, 143.26, 149.15, 158.57, 173.71, 173.98. HRMS: Calcd
for C29H37N2Oþ m/z 429.2901, obsd m/z 429.2913.
d4):
d 11.26, 26.79, 26.82, 30.49, 38.91, 49.18, 49.26, 102.02, 102.37,
110.80, 110.85, 121.98, 122.18, 125.32, 125.35, 128.54, 128.62, 140.68,
140.92, 141.53, 142.70, 150.82, 174.23, 175.23. HRMS: Calcd for
C
26H31Nþ2 m/z 371.2487, obsd m/z 371.2473.
2.2.20. (E)-3,3-Dimethyl-2-((E)-3-(1,3,3-trimethyl-3H-indol-1-
ium-2-yl)allylidene)-1-(3-(trimethylammonio)propyl)indoline-5-
sulfonate bromide (8h)
Compound 8h was synthesized by a similar procedure to 5a. The
dye was purified by recrystallization in acetonitrile. This compound
was obtained in 22.5% yield; decomp 168 ꢀC; 1H NMR (400 MHz,
2.2.16. 2-((1E,3E)-3-(1-Butyl-3,3-dimethylindolin-2-ylidene)prop-
1-en-1-yl)-1,3,3-trimethyl-3H-indol-1-ium iodide (8d)
Compound 8d was synthesized by a similar procedure to 8a. The
dye was purified by column chromatography (95% dichloro-
methane/methanol). This compound was obtained in 20.5% yield;
MeOD-d4):
d 1.78 (s, 6 H), 1.80 (s, 6 H), 2.36e2.40 (m, 2 H), 3.24 (s,
mp 115e117 ꢀC; 1H NMR (400 MHz, MEOD-d4):
H), 1.51e1.59 (m, 2 H), 1.77 (s, 12 H) 1.81e1.87 (m, 2 H), 3.72 (s, 3 H),
d
1.03 (t, J ¼ 7.6 Hz, 3
9H), 3.72 (t, J ¼ 7.6 Hz, 2 H), 3.79 (s, 3 H), 4.24 (t, J ¼ 7.6 Hz, 2 H), 6.67
(d, J ¼ 12.8 Hz,1 H), 6.75 (d, J ¼ 13.6 Hz,1 H), 7.37e7.52 (m, 4 H), 7.74

A. Levitz et al. / Dyes and Pigments 105 (2014) 238e249
241
(t, J ¼ 7.2 Hz, 1 H), 7.78 (d, J ¼ 8.4 Hz, 1 H), 7.91 (s, 1 H), 8.59 (t,
44.58, 48.73, 91.17, 99.24, 110.92, 111.92, 112.55, 122.67, 124.51,
126.38, 126.79, 128.83, 131.56, 140.61, 143.39, 147.17, 148.13, 162.40,
172.75; HRMS: Calcd for C25H29N2Oþ m/z 373.2280, obsd m/z
373.2281.
J ¼ 13.2 Hz, 1 H); 13C NMR (100 MHz, MeOD-d4):
d 20.83, 26.54,
27.09, 31.01, 40.94, 48.80, 49.75, 52.49, 63.06, 102.29, 104.66,
109.98, 111.62, 120.05, 122.10, 126.21, 126.83, 128.69, 140.36, 141.15,
142.12, 142.45, 143.25, 151.23, 173.66, 176.85. HRMS: Calcd for
C
30H40N3O3Sþ m/z 522.2785, obsd m/z 522.2783.
2.2.25. 2-((1E,3E)-3-(1-Butylbenzo[cd]indol-2(1H)-ylidene)prop-1-
en-1-yl)-1,3,3-trimethyl-3H-indol-1-ium iodide (10a)
Compound 10a was synthesized by a similar procedure to 8a.
The dye was purified by column chromatography (1% methanol/
dichloromethane). This compound was obtained in 79% yield; mp
2.2.21. 2-((1E,3E)-3-(3-Butyl-1,1-dimethyl-1H-benzo[e]indol-
2(3H)-ylidene)prop-1-en-1-yl)-1,3,3-trimethyl-3H-indol-1-ium
iodide (8i)
Compound 8i was synthesized by a similar procedure to 8a. No
purification was needed. This compound was obtained in 44% yield;
128e130 ꢀC; 1H NMR (400 MHz, MeOD-d4):
d
1.04 (t, J ¼ 7.6 Hz, 3
H), 1.50e1.60 (m, 2 H), 1.85 (s, 6 H), 1.87e1.95 (m, 2 H), 3.84 (s, 3
H), 4.32 (t, J ¼ 7.6 Hz, 2 H), 6.82 (d, J ¼ 13.6 Hz, 1 H), 6.87 (d,
J ¼ 13.6 Hz, 1 H), 7.41 (t, J ¼ 7.2 Hz, 1 H), 7.39e7.54 (m, 3 H), 7.64
(d, J ¼ 7.6 Hz, 1 H), 7.68 (d, J ¼ 7.6 Hz, 2 H), 7.77 (d, J ¼ 8.4 Hz, 1
H), 7.95 (t, J ¼ 7.6 Hz, 1 H), 8.23 (d, J ¼ 8.4 Hz, 1 H), 8.38 (d,
J ¼ 7.2 Hz, 1 H), 8.91 (t, J ¼ 13.6 Hz, 1 H); 13C NMR (100 MHz,
mp 133 ꢀC; 1H NMR (400 MHz, DMSO-d6):
d
0. 96 (t, J ¼ 7.2 Hz, 3 H),
1.42e1.52 (m, 2 H), 1.74 (s, 6 H), 1.76e1.82 (m, 2 H), 1.98 (s, 6 H), 3.67
(s, 3 H), 4.26 (t, J ¼ 7.2 Hz, 2 H), 6.50 (d, J ¼ 13.6 Hz, 1 H), 6.56 (d,
J ¼ 13.6 Hz,1 H), 7.29e7.32 (m, 1 H), 7.45 (d, 2 H), 7.54 (t, J ¼ 7.6 Hz,1
H), 7.65e7.69 (m, 2 H), 7.79 (d, J ¼ 8.4 Hz, 1 H), 8.07e8.12 (m, 2 H),
8.30 (d, J ¼ 8.4 Hz, 1 H), 8.47 (t, J ¼ 13.6 Hz, 1 H); 13C NMR (100 MHz,
MeOD-d4):
d 12.83, 14.03, 19.88, 27.30, 29.27, 31.06, 43.56, 65.49,
DMSO-d6):
d 0.58, 14.31, 20.02, 27.54, 27.80, 29.95, 31.82, 44.37,
105.97, 106.89, 109.28, 111.55, 122,11, 122.52, 125.09, 126.08,
126.61, 128.67, 129.30, 129.54, 130.29, 131.16, 141.31, 141.44,
142.58, 149.23, 155.41, 175.92. HRMS: Calcd for C29H31Nþ2 m/z
407.2482, obsd m/z 407.2499.
49.26, 51.09, 102.51, 103.17, 111.86, 112.24, 122.70, 122.91, 125.56,
127.86, 128.32, 129.05, 130.38, 130.93, 131.98, 133.76, 139.96, 140.98,
143.18, 149.49, 174.56, 175.89. HRMS: Calcd for C32H37Nþ2 m/z
449.2952, obsd m/z 449.2973.
2.2.26. 1,3,3-Trimethyl-2-((1E,3E)-3-(1-(3-(trimethylammonio)
propyl)benzo[cd]indol-2(1H)-ylidene)prop-1-en-1-yl)-3H-indol-1-
ium bromide (10b)
2.2.22. 2-((1E,3E)-3-(5-Bromo-3,3-dimethyl-1-(3-
(trimethylammonio)propyl)indolin-2-ylidene)prop-1-en-1-yl)-
1,3,3-trimethyl-3H-indol-1-ium bromide (8j)
Compound 10b was synthesized by a similar procedure to 8a.
No purification was needed. This compound was obtained in 75%
Compound 8j was synthesized by a similar procedure to 8a.
No purification was needed. This compound was obtained in 77%
yield; mp 134 ꢀC; 1H NMR (400 MHz, MeOD-d4):
d 1.77 (s, 6 H),
yield; mp 226 ꢀC; 1H NMR (400 MHz, DMSO-d6):
d 1.69 (s, 6 H),
2.23e2.30 (m, 2 H), 3.10 (s, 9 H), 3.62e3.66 (m, 2 H), 4.33 (t,
J ¼ 6.4 Hz, 2 H), 7.05 (d, J ¼ 13.6 Hz, 1 H), 7.12 (d, J ¼ 12.8 Hz, 1 H),
7.42 (t, J ¼ 7.6 Hz, 1 H), 7.54 (t, J ¼ 7.6 Hz, 1 H), 7.58 (d, 7.2 Hz, 1
H), 7.64 (d, J ¼ 8.0 Hz, 1 H), 7.72 (t, J ¼ 6.8 Hz, 1 H), 7.74e7.79 (m,
2 H), 8.00 (t, J ¼ 7.6 Hz, 1 H), 8.28 (d, J ¼ 8.4 Hz, 1 H), 8.36 (d,
J ¼ 7.2 Hz, 1 H), 8.73 (t, J ¼ 13.6 Hz, 1 H); 13C NMR (100 MHz,
1.71 (s, 6 H), 2.11e2.15 (m, 2 H), 3.10 (s, 9 H), 3.59 (t, J ¼ 6.8 Hz, 2
H), 3.71 (s, 3 H), 4.08 (t, J ¼ 7.2 Hz, 2 H), 6.68 (t, J ¼ 12.4 Hz, 2 H),
7.36 (t, J ¼ 7.6 Hz, 1 H), 7.45e7.50 (m, 2 H), 7.55 (d, J ¼ 8.0 Hz, 1
H), 7.67 (t, J ¼ 7.6 Hz, 2 H), 7.93 (s, 1 H), 8.34 (t, J ¼ 13.2 Hz, 1 H);
13C NMR (100 MHz, DMSO-d6):
d 21.09, 27.54, 27.92, 32.37, 41.20,
49.16, 49.76, 53.05, 62.81, 102.58, 104.89, 112.60, 113.32, 117.50,
122.96, 126.26, 126.37, 129.14, 131.74, 141.38, 141.70, 142.98,
143.10, 150.36, 172.76, 176.08. HRMS: Calcd for C22H23N2Sþ m/z
26.6198, obsd m/z 260.62.
MeOD-d4):
d 22.57, 28.09, 32.93, 40.93, 50.30, 52.99, 63.01,
107.73, 108.47, 109.39, 113,17, 122.44, 123.06, 125.23, 126.93,
127.19, 129.22, 129.80, 130.09, 130.23, 130.64, 131.41, 141.40,
142.09, 142.96, 148.87, 154.21, 176.40. HRMS: Calcd for C31H37N32þ
m/z 225.6488, obsd m/z 225.6170.
2.2.23. 1,3,3-Trimethyl-2-((1E,3Z)-3-(3-methylbenzo[d]thiazol-
2(3H)-ylidene)prop-1-en-1-yl)-3H-indol-1-ium iodide (9a)
Compound 9a was synthesized by a similar procedure to 8a. No
purification was needed. This compound was obtained in 67% yield;
2.2.27. 2-((1E,3E)-3-(1-Butylbenzo[cd]indol-2(1H)-ylidene)prop-1-
en-1-yl)-1-methylbenzo[cd]indol-1-ium bromide (13)
Compound 13 was synthesized by a similar procedure to 8a. The
dye was purified by precipitation from methanol with ether. This
compound was obtained in 41% yield; mp >260 ꢀC; 1H NMR
mp 223e225 ꢀC; 1H NMR (400 MHz, DMSO-d6):
d 1.65 (s, 6 H), 3.55
(s, 3 H), 3.94 (s, 3 H), 6.24 (d, J ¼ 12.8 Hz, 1 H), 6.78 (d, J ¼ 13.6 Hz, 1
H), 7.21 (t, J ¼ 7.6 Hz, 1 H), 7.34e7.42 (m, 2 H), 7.51 (t, J ¼ 7.6 Hz, 1 H),
7.58 (d, J ¼ 7.6 Hz, 1 H), 7.65 (t, J ¼ 7.6 Hz, 1 H), 7.87 (d, J ¼ 8.4 Hz, 1
H), 8.09 (d, J ¼ 8.0 Hz, 1 H), 8.40 (t, J ¼ 12.8 Hz, 1 H); 13C NMR
(400 MHz, DMSO-d6):
d
0.96 (t, J ¼ 7.2 Hz, 3 H), 1.44-1.49 (m, 2 H),
1.81e1.85 (m, 2 H), 3.92 (s, 3 H), 4.36 (t, J ¼ 7.2 Hz, 2 H), 7.19 (d,
J ¼ 13.2 Hz, 2 H), 7.70e7.73 (m, 4 H), 7.86 (d, J ¼ 7.2 Hz, 2 H), 8.04e
8.07 (m, 2 H), 8.34 (d, J ¼ 7.6 Hz, 2 H), 8.65 (d, J ¼ 7.2 Hz, 2 H), 9.16 (t,
(100 MHz, DMSO-d6):
d 28.71, 34.51, 48.73, 99.92, 102.82, 110.96,
114.92,122.72,123.69,124.54,125.80,126.46,128.78,128.87,140.65,
142.36, 143.40, 147.64, 168.17, 172.12. HRMS: Calcd for C22H23N2Sþ
m/z 347.1577, obsd m/z 347.1588.
J ¼ 13.2 Hz,1 H); 13C NMR (100 MHz, DMSO-d6):
d 0.59, 14.23, 20.14,
31.28, 31.85, 44.30, 55.40, 111.26, 111.40, 111.50, 111.67, 124.13,
124.19, 124.66, 129.48, 129.55, 129.80, 129.89, 130.11, 130.87, 130.92,
132.38, 141.22, 141.77, 148.51, 155.42, 156.17. HRMS: Calcd for
2.2.24. 2-((1E,3Z)-3-(3-Butylbenzo[d]oxazol-2(3H)-ylidene)prop-
1-en-1-yl)-1,3,3-trimethyl-3H-indol-1-ium iodide (9b)
C
30H27Nþ2 m/z 415.2169, obsd m/z 415.2172.
Compound 9b was synthesized by a similar procedure to 8a. The
dye was purified by chromatotron (2% methanol/dichloromethane).
This compound was obtained in 38% yield; mp 134 ꢀC; 1H NMR
2.3. Stock solutions
(400 MHz, DMSO-d6):
d
0.94 (t, J ¼ 7.2 Hz, 3 H), 1.37e1.44 (m, 2 H),
Stock solutions of the dyes and standard were prepared by
weighing the solid on a 5-digit analytical balance in an amber vial
and adding DMSO or 0.1 M NaOH solution via a class A volumetric
pipette. The vials were vortexed for 20 s and then sonicated for
5 min to ensure complete dissolution. The stock solutions were
stored in a dark freezer at 4 ꢀC when not in use.
1.66 (s, 6 H), 1.75e1.82 (m, 2 H), 3.54 (s, 3 H), 4.29 (t, J ¼ 7.6 Hz, 2 H),
6.23 (d, J ¼ 13.6 Hz,1 H), 6.41 (d, J ¼ 13.2 Hz,1 H), 7.21 (t, J ¼ 7.2 Hz,1
H), 7.34 (d, J ¼ 7.6 Hz, 2 H), 7.39 (t, J ¼ 7.6 Hz, 1 H), 7.47e7.58 (m, 3
H), 7.81 (d, J ¼ 8.0 Hz, 1 H), 7.87 (t, J ¼ 8.0 Hz, 1 H), 8.34 (t, J ¼ 13.2, 1
H); 13C NMR (100 MHz, DMSO-d6):
d 14.08, 19.83, 28.42, 30.21,

242
A. Levitz et al. / Dyes and Pigments 105 (2014) 238e249
2.4. Method of determining molar absorptivity and fluorescence
2.7. Computational methods
quantum yield
All calculations were carried out using Gaussian 03 (Wall-
ington, CT). The structure of each compound was first optimized
using the TD-DFT method with the hybrid exchange-correlation
functional PBE1PEB and 6-31G* basis set. All optimized
structures were obtained using a tight threshold which is based
on a root mean square residual force less than 10ꢁ5 au for the
optimal geometry. To account for solvent effects, energy
calculations were carried out with DMSO as solvent using
PBE1PBE and B3LYP methods. Because experiments have shown
the anion would not affect the spectra, anions were neglected.
Stock solutions were used to prepare six dilutions of dyes and
the standard with concentrations ranging from 2 mM to 6 mM using
a class A volumetric pipette in order to maintain absorption be-
tween 0.1 and 1.0. The dye solutions were diluted ten-fold for
fluorescence in order to minimize inner filter effect. For the stan-
dard, fluorescence concentrations were 0.02e0.06 mM and the area
under the curve was then multiplied by 10. The absorbance spec-
trum of each sample was measured in triplicate using the Cary 3G
UVeVisible Spectrophotometer from 400 to 800 nm. The spectra
were then normalized at 400 nm. The emission spectrum of each
sample was measured in triplicate using the ISS K2 spectrofluo-
rometer with a 465 nm excitation wavelength (lExC).
For molar absorptivity, the absorbance at the wavelength of
maximum absorbance (lmax) was determined and the absorbance
of each sample at lmax was plotted as a function of dye concen-
tration (C). The linear regression equation was computed using
Microsoft Excel.
Only the theoretical excitation energies l(cal) with the highest
oscillator strengths were considered. The electrostatic potential
maps were investigated for the optimized structures at HF/6-
31G* level.
3. Results and discussion
3.1. Synthesis
The fluorescence quantum yields were determined relative to
the fluorescein standard utilizing the gradient from the plot of in-
tegrated fluorescence intensity vs absorbance (Grad), refractive
As shown in Scheme 1, the synthesis of quaternary indolium salts
5 began with a Fischer indole synthesis of 4 by reaction of 4-
index (h) and the published quantum yield of the standard (4 ), as
substituted phenylhydrazines
1 with 3-methyl-2-butanone in
S
per Equation (1). In this equation, the subscripts S and D refer to the
standards and the dyes, respectively.
boiling glacial acetic acid [29e31] as previously published by our
group [8,22]. 2-methylbenzothiazole, 2-methylbenzoxazole, 2,3,3-
trimethylindolenine, 1,1,2-trimethylbenz[e]indole and 1,2,3,3-
tetramethylindolenine were purchased from SigmaeAldrich and
used without purification. The heterocycles were quaternized
through SN2 reaction with alkyl halides in boiling acetonitrile to
form salts 5 and 6. Benz[c,d]indole quaternary ammonium salts 7
were synthesized as previously described [32]. The key step in the
formation of unsymmetrical trimethine cyanine dyes is the forma-
tion of the aldehyde 3 through a Vilsmeier formylation reaction to
form a dimethylamino intermediate [33]. The intermediate was
then heated under reflux in a chloroform/aqueous sodium hydrox-
ide solution to convert the diamine to an aldehyde 3; mp 104e
106 ꢀC, 72% yield (lit mp 116 ꢀC, 60% yield [34]). The aldehyde is then
reacted in acetic anhydride with each quaternary ammonium salt 5,
6 or 7 for one hour at 100 ꢀC toyield 16 novel unsymmetrical dyes 8e
10 and 13 as shown in Scheme 1 and Equation (2). Dyes 8f,g,i,j,9a,
and 10b were produced as pure after crystallization and washing
with ether. Dyes needing purification were purified by column
chromatography using dichloromethane/methanol for regular
phase silica and water/methanol for reverse phase silica. Dyes with
multiple charges 8a,b,j tended to have lower yields due to the use of
reverse phase column chromatography for purification.
Equation 1. Quantum yield equation.
2.5. Hydrophobicity studies
Hydrophobic characteristics of the dyes were assessed by
acquiring absorbance spectra in varying ratios of methanolebuffer
mixture (in the range of 2e100% methanol) using phosphate buffer
pH 7.4. The concentration of methanol was slowly increased until
there were no apparent changes at which point the spectra
overlapped.
2.6. Method of predicting Lipinski rule violations
Molecular weight was calculated using ChemBioDraw (Cam-
bridgeSoft, Cambridge, MA) and do not include the counterion.
Hydrogen bond donors and acceptors, log D, and polar surface area
(PSA) were predicted using MarvinSketch 5.9 (ChemAxon, Buda-
pest, Hungary). Predicted log D values were derived from the cal-
culations by Viswanadhan [27], Klopman [28], and the PHYSPROP
database at pH 7.4 and the three were averaged.
3.2. Solvent determination
After synthesis and full characterization of the dyes, the
spectroscopic properties of the compounds were studied. To
Equation 2. Synthesis of Benz[c,d]indole trimethine cyanine dye 13.

A. Levitz et al. / Dyes and Pigments 105 (2014) 238e249
243
Scheme 1. Synthesis of unsymmetrical trimethine cyanines.
investigate which solvent would be most suitable for studying
the optical properties of these dyes, 3 uM solutions of dye 8d
were dissolved in various solvents that all of the dyes were sol-
uble in. The effect of differing the solvent on absorption was
investigated using acetonitrile, DMSO, ethanol and methanol and
can be seen in Fig. 2. The dyes had the highest absorption in
ethanol and absorbed at the longest wavelength in DMSO.
Methanol and acetonitrile showed no difference from each other.
The dyes were best soluble in DMSO.
absorption shifts caused by the different heterocycles in com-
pounds 8a,i,g, 9b, and 10a can be seen in Fig. 3. The lmax varies from
516 nm to 672 nm in dyes containing a benzoxazole heterocycle
and a benz[c,d]indole heterocycle, respectively. The lmax of trime-
thine cyanine dyes containing two indolenine bases is usually
around 552 nm in DMSO, and the Emax is 565 nm. The alkyl side
chains have little effect on the absorption and fluorescence of the
dyes. The most notable shifts in unsymmetrical dyes with differing
heterocycles come from those containing heterocycles that have
more interaction with the conjugated system. For instance, the
oxygen in benzoxazole heterocycles has more influence on the
conjugated methine chain than the sulfur atom in benzothiazole
heterocycles does due to its electronegativity. This can be seen in
NMR experiments based on the shift of the meso-proton. As shown
3.3. Optical properties
Because the dyes as a whole were best soluble in DMSO it was
chosen as the most suitable solvent for spectroscopic studies. The
Fig. 2. Solvent effects on the absorption of 8d A) full spectrum B) zoomed in to lmax
.

244
A. Levitz et al. / Dyes and Pigments 105 (2014) 238e249
Table 1
Meso-proton shifts in relation to lmax
.
Dye
Meso-proton shift (ppm)
lmax (nm)
10a
8d
9a
8.91
8.55
8.40
8.34
672
552
549
516
9b
Also shown in Figs. 4 and 5 are MEP maps of compounds
8a,b,d,h, 9a,b, and 10a. The negative (red and yellow) and the
positive (blue) regions represent the distribution of electrostatic
potential in the molecule. Analyzing these compounds, the most
negativity is seen around the oxygen atom in compound 9b causing
the meso-carbon to almost be in the surrounding positive blue re-
gion. The proton in this carbon corresponds to a 1H NMR shift of
d
8.34 ppm and a lmax of 516 nm. The next most electronegative is
the sulfur atom in 9a before the dimethyl groups in 8d. Lastly the
benz[c,d]indole heterocycle of 10a has more conjugation in the
heterocycle further delocalizing the electrons in the methine chain
Fig. 3. Absorption spectra of various dyes at 3 mM in DMSO.
and thereby the proton shifts downfield the furthest to
d 8.91 ppm
in Table 1, the observed chemical shift of the meso-proton is most
likely due to electron density from the surrounding atoms. Oxygen
is more electronegative than sulfur and therefore attracts more
electronegativity from the methine chain corresponding to a lmax
shift from 516 nm (9b) to 549 nm (9a). Sulfur is only slightly more
electronegative than carbon causing a smaller meso-proton shift
and corresponding to a lmax to 552 nm (8aee). The benz[c,d]indole
heterocycle shows the largest shift in both the meso-proton and
lmax both due to the lack of an electronegative element around the
methine chain and increased conjugation in the heterocycle. Fig. 4
shows the electrostatic maps of these dyes next to their proton
NMR. Taking into account that the overall charges and the length of
the conjugated system are similar in all of these compounds, the
molecular electrostatic potential (MEP) maps do not present sharp
contrasting features; however, the subtle differences in molecular
structure of the compounds are clearly reflected. The shift of this
meso-proton is even further in the dyes containing the benz[c,d]
indole heterocycle.
and the most red shifted lmax is observed. In addition, shown in
Fig. 5, the maximum negative (red) region is introduced by the
addition of a sulfonate functional group and the addition of a sec-
ond positively charged quaternary nitrogen introduces a strong,
broad positive region, quite distinctive from the rest of the com-
pounds. The MEP maps clearly represent the change in electronic
structure of compounds with different substituents as shown in
Fig. 5(c) where the electrostatic interactions of the sulfonate group
with the additional positive quaternary nitrogen impede the color
shifts that were observed from the addition of one of these charged
functional groups.
Optical properties of the dyes can be seen in Table 2. The dyes
containing bromine and methoxy substituted indolenine 8g,j,f,
benz[e]indole 8i, and benzoxazole 9b show lmax directly between
where symmetrical dyes of the two heterocycles would absorb.
This suggests that the heterocycles are equivalent in basicity to
one another. The dye containing one indolenine and one benz[c,d]
indole 10a heterocycle had lmax that was not in the center of
Fig. 4. Shift of meso-proton due to changing heterocycles.

A. Levitz et al. / Dyes and Pigments 105 (2014) 238e249
245
Fig. 5. Electrostatic maps of compounds A) 8b B) 8h C) 8a.
the symmetrical dyes due to differing basicity in the two hetero-
cycles. Dyes 8aec all have two unsubstituted indolenine hetero-
cycles and show molar absorptivity in the range of 120,000e
127,000 Mꢁ1 cmꢁ1 reinforcing that the side chains have little effect
on the molar absorptivity. The dye with a methoxy substituted
The theoretical data is plotted along with the experimental data in
Fig. 6 for comparison. TD-DFT has been shown to work well for
large conjugated molecules. In DFT methods the orbitals are ob-
tained by solving the KohneSham equation, involving exchange
and correlation (XC) terms [39]. Although a discrepancy gap of
approximately 75 nm maximum is observed between experimental
and theoretical results, the observed trends in absorbance wave-
length are almost the same as seen in experimental data as shown
in Fig. 6. The underestimation of lmax has been shown to be due to
an overestimation of the excitation energies [40]. This finding may
be suggestive of a systematic error in the treatment of the studied
compounds by the mentioned methods that can be statistically
treated for partial correction of these experimental/theoretical
discrepancies.
The majority of the dyes have a quantum yield around 10%. Dye
8f with the bromine substituent on the heterocycle was slightly
higher at 12.5%. This could be due to the electron withdrawing
nature of the bromine substituent or the increased molar absorp-
tivity. The benzoxazole containing dye 9b was also close to 10%, but
the benzothiazole dye was lower around 7%. Benzoxazole dyes tend
to have higher quantum yield due to the lone pair of the oxygen
stabilizing the excited state, while the sulfur atom in benzothiazole
dyes causes the opposite effect possibly due to a lower singlet-
triplet energy gap raising the opportunity for non-radiative relax-
ation [20].
indolenine 8g showed lower molar absorptivity (94,067 Mꢁ1 cmꢁ1
)
and the lowest quantum [35] yield due to the electron donating
nature of the methoxy group introducing electron density back into
the system. Sulfonate groups which are often added to increase
solubility or balance charges increase the molar absorptivity. The
molar absorptivity of dyes containing benzothiazole and benzox-
azole heterocycles is also slightly lower than those with two
indolenine heterocycles. The addition of a benz[c,d]indole hetero-
cycle causes a shift over 100 nm, the greatest effective length [36]
among nitrogen-containing heterocyclic dyes and makes the dye
blue in color, but also causes a decrease in molar absorptivity.
Adding two benz[c,d]indole heterocycles shifts absorption to the
near-infrared region showing a lmax of 758 nm. These shifts are of
importance to in vivo imaging because it allows the fluorescence of
trimethine cyanine dyes to be closer to the NIR range where there is
less autofluorescence from biological tissue but comes at the price
of lower quantum efficiency. Lee et al. have shown that the benz
[c,d]indole moiety causes a large decrease in fluorescence lifetime
[37]. This is probably due to the planar nature of these heterocycles
stacking. The dye containing a benzoxazole heterocycle 9b had a
blue shift further from the NIR region, but also exhibited an
increased Stokes shift of over 20 nm. As mentioned before, this is
likely due to the electronegativity of the oxygen atom. The most
interesting find from quantum yield studies was that the dye with
the bromine substitution actually had an increased quantum yield.
This is in contrast to typical heavy atom effect findings where the
excited state goes through intersystem crossing to the triplet state
and could be due to an increased fluorescence lifetime as described
by Soper et al. [38].
3.4. Hydrophobicity studies
During our spectroscopic studies we noticed that the unsym-
metrical dyes did not show the same aggregation as symmetrical
dyes that we have previously synthesized. Cyanine dyes are known
to aggregate, but in our studies of unsymmetrical trimethine
Theoretical time-dependent density-functional theory (TD-DFT)
evaluations were carried out to compare our absorbance de-
terminations with those expected based on computational studies.
Table 2
Summary of optical properties of unsymmetrical cyanine dyes in DMSO.
Dye
lmax
Emax
Stokes shift
F
3
(Mꢁ1 cmꢁ1 ꢂ103)
8a
8b
8c
8d
8e
8f
8g
8h
8i
552
553
552
552
553
556
566
556
572
556
549
516
672
565
565
565
565
567
569
585
570
586
568
569
537
e
13
12
13
13
14
13
19
14
14
12
20
21
e
0.08
0.10
0.07
0.10
0.08
0.12
0.03
0.10
0.10
0.11
0.07
0.10
e
120
124
127
137
115
128
94
144
107
101
112
111
75
8j
9a
9b
10a
Fig. 6. Experimental and Calculated lmax values.

246
A. Levitz et al. / Dyes and Pigments 105 (2014) 238e249
Fig. 7. Absorption spectra of cyanine dyes as a function of increasing solvent hydrophobicity (% v/v methanol to buffer) at constant dye concentration at constant dye concentration
of 10 M (A) 8d (B) 8i (C) 9a.
m
cyanines, as the concentration of each sample was increased the
ratio of the two absorption peaks intensity remained constant
suggesting that the spectral representation of these compounds
does not show aggregation in DMSO. For example, as can be seen in
indolenine to benze[e]indolenine to benz[c,d]indolenine increases
the amount of aggregation seen in higher polarity solvent.
To show that the vast majority of hydrophobic effects are due to
the heterocycle, an unsymmetrical dye 13 containing two benz[c,d]
indole heterocycles with one methyl and one butyl side chain was
synthesized as shown in Equation (2). The hydrophobic effects of
unsymmetrical dye 13 were compared to symmetrical dyes syn-
thesized by our group [17] and as shown in Fig. 8, there is negligible
change observed from decreasing one butyl group to a methyl.
Changes in the hydrophobic nature of these dyes can make them
less biocompatible with regard to uptake and in vivo transportation.
Table 2 shows that altering the heterocycle from indolenine to
benze[e]indolenine to benz[c,d]indolenine causes a red shift in
absorption but also a decrease in quantum efficiency. A possible
way to counter this is through the use of unsymmetrical dyes. By
changing only one heterocycle biocompatibility may be retained
while only losing a small portion of the dye’s quantum efficiency.
The improved solubility of these less hydrophobic dyes also in-
creases their attractiveness as biological probes. These increases in
hydrophobicity can be predicted through log D values as shown in
Table 3. Compounds with higher log D values are more hydrophobic
and show greater aggregation.
Fig. 7, dye 8i at 6
539 nm is 0.434, a ratio of 1.48. If the concentration is increased to
M the absorption at 572 nm is 0.967 and at 539 nm is 0.653,
mM shows absorption at 572 nm is 0.645 and at
9
m
again a ratio of 1.48. In addition the second absorption maxima at
the lower wavelength are not present in the dyes with a benzo-
thiazole or benzoxazole moiety as can be seen in Fig. 3. The second
peak has been attributed to the more pronounced H-band associ-
ated with dyes containing two dimethyl-indolenine heterocycles.
Hydrophobicity studies were carried out in methanolebuffer
mixtures. As the polarity of the solvent decreased with increasing
methanol concentration from 2% to 100% a general amplification of
the monomeric band was observed as seen in Figs. 7 and 8. As the
polarity decreases there is also an increase of the ratio of monomer
band to the aggregate band up to 30% methanol when the ratio
becomes steady. The increased absorbance is related to the de-
fragmentation of aggregate to form monomers. As explained by our
group this aggregation is predominantly due to a plane-to-plane
arrangement [17]. Although this aggregation is still seen in un-
symmetrical dyes at higher polarities, the aggregation is not as
obvious. This is likely due to the different heterocycles not stacking
as easily as if the heterocycles are the same, i.e. two benzothiazole
heterocycles stacking together better than a benzothiazole and an
indolenine heterocycle. In addition, these dyes all include one
methyl side chain, so the effect of increasing the length and sub-
sequent hydrophobicity of only one side would be expected to have
a smaller effect on aggregation. Examples of this can be seen in
Fig. 7. Increasing the hydrophobicity of the dye by changing
3.5. Predicted electrostatic potentials
Energy transitions in cyanine dyes have been shown to be a
dominant
pep* transition. Molecular orbitals involved in these
transitions can be seen in Fig. 9 along with their energy gaps. The
energy gap between the highest occupied molecular orbital
(HOMO) and lowest unoccupied molecular orbital (LUMO) of
compound 10a is the lowest among all of the synthesized
Fig. 8. Absorption spectra of compound 13 compared to a symmetrical dye synthesized by our lab [17] as a function of solvent hydrophobicity (% v/v methanol to buffer) at constant
dye concentration of 10 M.
m

A. Levitz et al. / Dyes and Pigments 105 (2014) 238e249
247
Table 3
but the benzoxazole heterocycle has slightly more energy than the
indole heterocycle. The next highest energy gaps arise from the
bromine substituted derivatives 8f and 8j again due to the elec-
tronegative nature of the bromine atom. Compounds with the same
heterocycle and different side chains 8a,c,e,h tend to have close
energy gaps in the range of 0.285e0.289 Ha showing that the
heterocycles have a larger effect on energy than do the side chains.
For these compounds both the HOMO energy is spread evenly
throughout the molecule while the LUMO energy is condensed
mostly on the methine chain. The exception here is compound 8b
with a sulfonate on one side chain. The sulfonate causes the energy
gap to decrease 0.268 Ha. In this compound HOMO energy is almost
entirely in the sulfonate group whereas the LUMO energy is spread
throughout the rest of the molecule, still mostly in the methine
chain, but there is no LUMO energy in the sulfonate. This drastic
change in energy is balanced out in compound 8h when the sul-
fonate is a substituent on the heterocycle and the electrons are
shared throughout the conjugated system.
Summary of Lipinski rule properties. Molecular weight was calculated using
ChemBioDraw (CambridgeSoft, Cambridge, MA) and do not include the counterion.
Hydrogen bond donors and acceptors, log D, and PSA were predicted using Mar-
vinSketch 5.9 (ChemAxon, Budapest, Hungary). Predicted log D values were derived
from the calculations by Viswanadhan [27], Klopman [28], and the PHYSPROP
database at pH 7.4 and the three were averaged.
Dye
Molecular
weight
H bond
donors
H bond
acceptors
Polar surface
area (A )
log D
Pass?
2
ꢀ
8a
8b
8c
8d
8e
8f
8g
8h
8i
443.66
464.61
371.53
399.59
461.65
478.48
429.61
522.72
449.64
522.72
347.49
373.51
407.56
0
0
0
0
0
0
0
0
0
0
0
0
0
1
4
1
1
1
1
2
4
1
1
1
2
1
6.25
63.45
6.25
6.25
6.25
15.48
63.45
6.25
6.25
6.25
ꢁ2.15
3.03
3.38
4.35
5.49
5.12
4.19
ꢁ0.94
5.34
ꢁ0.28
1.19
1.72
4.16
☑
☑
☑
☑
,
,
☑
,
,
☑
8j
9a
9b
10a
6.25
15.48
6.25
☑
☑
☑
3.6. Lipinski rule properties
To show that the compounds have potential applications in
medicine and imaging, the Lipinski Rule properties were predicted
using MarvinSketch 5.9. The Lipinski Rule properties for each dye
can be seen in Table 3. The molecular weight (MW) was calculated
without the counterion as it would be lost in vivo. The only dyes
that didn’t meet the Lipinski Rule for molecular weight were 8h and
8j due to the addition of heavy substituents, such as halogens, to
the heterocycle already containing the trimethylammonium side
compounds. This is likely due to the increased conjugation around
the benz[c,d]indole heterocycle which led to further delocalizing of
the electrons making the compound more stable. Compound 9b
with the benzoxazole heterocycle shows the highest energy gap
likely due to the lone pair of electrons around the oxygen atom. In
this compound the HOMO energy is spread evenly throughout the
compound, while the LUMO energy is mostly in the methine chain
Fig. 9. HOMO/LUMO diagrams.

248
A. Levitz et al. / Dyes and Pigments 105 (2014) 238e249
[9] Choi HS, Gibbs SL, Lee JH, Kim SH, Ashitate Y, Liu FB, et al. Targeted zwitter-
ionic near-infrared fluorophores for improved optical imaging. Nat Biotechnol
2013;31(2):148e53.
[10] Licha K, Riefke B, Ntziachristos V, Becker A, Chance B, Semmler W. Hydrophilic
cyanine dyes as contrast agents for near-infrared tumor imaging: synthesis,
photophysical properties and spectroscopic in vivo characterization. Photo-
chem Photobiol 2000;72(3):392e8.
[11] Achilefu S, Dorshow RB, Bugaj JE, Rajagopalan R. Novel receptor-targeted
fluorescent contrast agents for in vivo tumor imaging. Invest Radiol
2000;35(8):479e85.
[12] Nakazumi H. Organic colorants for laser disc optical data storage. J Soc Dye
Colour 2008;104(3):121e5.
[13] Warner IM, Soper SA, McGown LB. Molecular fluorescence, phosphorescence,
and chemiluminescence spectrometry. Anal Chem 1996;68(12):R73e91.
[14] Haugland RP, Spence MTZ, Johnson ID. Handbook of fluorescent probes and
research chemicals. 6th ed. Eugene, OR, USA (4849 Pitchford Ave., Eugene
97402): Molecular Probes; 1996.
[15] Deligeorgiev TG, Zaneva DA, Kim SH, Sabnis RW. Preparation of monomethine
cyanine dyes for nucleic acid detection. Dyes Pigments 1998;37(3):205e11.
[16] Brooker LGS, Keyes GH, Williams WW. Color and constitution. V.1 the ab-
sorption of unsymmetrical cyanines. Resonance as a basis for a classification
of dyes. J Am Chem Soc 1942;64(2):199e210.
chain. None of the dyes contain any hydrogen bond donors. All of
the dyes meet both rules for hydrogen bond donors and acceptors,
less than 5 and less than 10, respectively. They also all meet the
ꢀ
polar surface area (PSA) requirement of less than 140 A. Three dyes
8e,f,i, had log D above the Lipinski Rule value, therefore 8 of the 13
dyes tested meet all the Lipinski rules. One common use of trime-
thine dyes is for cell staining. According to Burton et al. an incre-
mental increase in cell permeability is seen with an increase in
distribution coefficient [18], so it would be expected that the dyes
with the highest log D shown in Table 3 would have the best cell
permeability. The dyes with a butyl substituents 8d,f,g,i and pro-
pylphenyl 8e substituents on them have high log D values, espe-
cially the butyl substituent on the benz[e]indole moiety. The dyes
with the lowest log D values are those with more than one charge
8a,b,h,j and the benzoxazole moiety 9b. It would be expected that
these dyes would have a lower probability of crossing the cell
membrane, but may be more suitable for other applications.
[17] Beckford G, Owens E, Henary M, Patonay G. The solvatochromic effects of side
chain substitution on the binding interaction of novel tricarbocyanine dyes
with human serum albumin. Talanta 2012;92:45e52.
4. Conclusion
[18] Goodwin JT, Conradi RA, Ho NFH, Burton PS. Physicochemical determinants of
passive membrane permeability: role of solute hydrogen-bonding potential
and volume. J Med Chem 2001;44(22):3721e9.
[19] Ghose AK, Viswanadhan VN, Wendoloski JJ. A knowledge-based approach in
designing combinatorial or medicinal chemistry libraries for drug discovery.
1. A qualitative and quantitative characterization of known drug databases.
J Comb Chem 1999;1(1):55e68.
[20] Shi QQ, Sun R, Ge JF, Xu QF, Li NJ, Lu JM. A comparative study of symmetrical
and unsymmetrical trimethine cyanine dyes bearing benzoxazolyl and ben-
zothiazolyl groups. Dyes Pigments 2012;93(1e3):1506e11.
[21] Zimmermann T, Hennig L. Ring transformations of heterocyclic com-
pounds. XXII. Pyrido[1,2-a]indolium salts from 2-methyl-3H-indoles by
A series of 16 novel unsymmetrical trimethine cyanine dyes was
synthesized in very good yield, purified, and characterized by 1H
NMR, 13C NMR, liquid chromatography, and mass spectrometry.
Their optical properties including absorption, fluorescence, Stokes
shift, molar absorptivity, and quantum yield were studied. Solvent
effects were evaluated and DMSO was determined to be the best
solvent for studying the optical properties. Dyes with electron
withdrawing substitutions on the heterocycle show increased
quantum yield and molar absorptivity while those with electron
donating groups show lower molar absorptivity and quantum yield.
Unsymmetrical dyes can be used to fine tune absorption wave-
lengths without the loss of other favorable properties such as
quantum yield. Unsymmetrical dyes were shown to not aggregate
as much as symmetrical dyes most likely due to the inability of the
different heterocycles to stack as well. Trimethine cyanine dyes
could be good oral drugs as the majority of them meet Lipinski’s
Rule of Five.
pyrylium mediated three carbon annelation.
J
Heterocycl Chem
2002;39(2):263e9.
[22] Nanjunda R, Owens EA, Mickelson L, Alyabyev S, Kilpatrick N, Wang SM, et al.
Halogenated pentamethine cyanine dyes exhibiting high fidelity for G-
quadruplex DNA. Bioorg Med Chem 2012;20(24):7002e11.
[23] Tang B, Zhang L, Xu KH. FIA-near-infrared spectrofluorimetric trace deter-
mination of hydrogen peroxide using tricarchlorobocyanine dye (Cy.7.Cl) and
horseradish peroxidase (HRP). Talanta 2006;68(3):876e82.
[24] Sinha SH, Owens EA, Feng Y, Yang Y, Xie Y, Tu Y, et al. Synthesis and evalu-
ation of carbocyanine dyes as PRMT inhibitors and imaging agents. Eur J Med
Chem 2012;54:647e59.
[25] Xiang JF, Liu YX, Sun D, Zhang SJ, Fu YL, Zhang XH, et al. Synthesis, spectral
properties of rhodanine complex merocyanine dyes as well as their effect on
K562 leukemia cells. Dyes Pigments 2012;93(1e3):1481e7.
Acknowledgments
[26] Nigrey PJ, Garito AF. Synthesis and characterization of 3-alkylbenzothiazolium
salts. J Chem Eng Data 1977;22(4):451e2.
This work was supported by the Georgia Research Alliance grant
to MH. MH also appreciates the Georgia State University Center for
Diagnostics and Therapeutics for their support. NSF and the Georgia
Research Alliance fund the research in the group of DH.
[27] Viswanadhan VN, Ghose AK, Revankar GR, Robins RK. Atomic physicochem-
ical parameters for three dimensional structure directed quantitative
structure-activity relationships. 4. Additional parameters for hydrophobic and
dispersive interactions and their application for an automated superposition
of certain naturally occurring nucleoside antibiotics. J Chem Inf Comput Sci
1989;29(3):163e72.
[28] Klopman G, Li J-Y, Wang S, Dimayuga M. Computer automated log P calcu-
Appendix A. Supplementary data
lations based on an extended group contribution approach.
J Chem Inf
Comput Sci 1994;34(4):752e81.
Supplementary data related to this article can be found at http://
dx.doi.org/10.1016/j.dyepig.2014.02.009.
[29] Illy H, Funderburk L. Fischer indole synthesis. Direction of cyclization of iso-
propylmethyl ketone phenylhydrazone. J Org Chem 1968;33(11):4283e5.
[30] Yuan L, Lin W, Song J. Ratiometric fluorescent detection of intracellular hy-
droxyl radicals based on a hybrid coumarin-cyanine platform. Chem Commun
(Camb) 2010;46(42):7930e2.
References
[31] Evans FJ, Lyle GG, Watkins J, Lyle RE. Molecular rearrangements. VIII. The
equilibration of 3,3-dimethyl-2-phenyl- and 2,3-dimethyl-3-phenyl-3H-in-
doles1. J Org Chem 1962;27(5):1553e7.
[32] Le Romancer M, Treilleux I, Bouchekioua-Bouzaghou K, Sentis S, Corbo L.
Methylation, a key step for nongenomic estrogen signaling in breast tumors.
Steroids 2010;75(8e9):560e4.
[1] Wyler H. Die Betalaine. Chemie unserer Zeit 1969;3(5):146e51.
[2] Wyler H. Das Experiment: Papierelektrophorese. Chemie unserer Zeit
1969;3(4):111e5.
[3] Musso H. The pigments of fly agaric, Amanita muscaria. Tetrahedron
1979;35(24):2843e53.
[33] Lee H, Berezin MY, Henary M, Strekowski L, Achilefu S. Fluorescence lifetime
properties of near-infrared cyanine dyes in relation to their structures.
J Photochem Photobiol A Chem 2008;200(2e3):438e44.
[34] Wang J, Cao WF, Su JH, Tian H, Huang YH, Sun ZR. Syntheses and nonlinear
absorption of novel unsymmetrical cyanines. Dyes Pigments 2003;57(2):
171e9.
[35] Murphy S, Yang XQ, Schuster GB. Cyanine borate salts that form penetrated
ion-pairs in benzene solution e synthesis, properties, and structure. J Org
Chem 1995;60(8):2411e22.
[36] Mikhailenko FA, Vasilenko NP, Kachkovskii AD, Rozhinskii RI. Effect of polar
substituents and polymethine chain length on the color of cyanine dyes of the
benz[c,d]indole series. Zh Org Khimþ 1982;18:435e41.
[4] Reichardt C. Chiral polymethine dyes: a remarkable but forgotten conjugated
pi system. J Phys Org Chem 1995;8(12):761e73.
[5] Hamer FM. The cyanine dyes and related compounds. New York: Interscience
Publishers; 1964.
[6] Dähne S, Resch-Genger U, Wolfbeis OS, North Atlantic Treaty Organization.
Scientific Affairs Division. Near-infrared dyes for high technology applications.
Dordrecht; Boston: Kluwer; 1998.
[7] Xu YF, Liu Y, Qian XH. Novel cyanine dyes as fluorescent pH sensors: PET, ICT
mechanism or resonance effect? J Photochem Photobiol A 2007;190(1):1e8.
[8] Choi HS, Nasr K, Alyabyev S, Feith D, Lee JH, Kim SH, et al. Synthesis and
in vivo fate of zwitterionic near-infrared fluorophores. Angew Chem Int Ed
2011;50(28):6258e63.

A. Levitz et al. / Dyes and Pigments 105 (2014) 238e249
249
[37] Branscombe TL, Frankel A, Lee JH, Cook JR, Yang Z, Pestka S, et al. PRMT5
(Janus kinase-binding protein 1) catalyzes the formation of symmetric
dimethylarginine residues in proteins. J Biol Chem 2001;276(35):32971e6.
[38] Flanagan JH, Owens CV, Romero SE, Waddell E, Kahn SH, Hammer RP, et al.
Near-infrared heavy-atom-modified fluorescent dyes for base-calling in DNA-
sequencing applications using temporal discrimination. Anal Chem
1998;70(13):2676e84.
[39] Bickelhaupt FM, Baerends EJ. Kohn-sham density functional theory: predict-
ing and understanding chemistry. Reviews in computational chemistry. John
Wiley & Sons, Inc.; 2007. pp. 1e86.
[40] Champagne B, Guillaume M, Zutterman F. TDDFT investigation of the optical
properties of cyanine dyes. Chem Phys Lett 2006;425(1e3):105e9.