Tuning thiol addition to squaraines by ortho-substitution and the use of serum albumin

Article abstract of DOI:10.1016/j.dyepig.2016.11.060

Tuning the reactivity of squaraine dyes toward nucleophilic addition of thiols was investigated. A series of water soluble, aniline-derived squaraines were synthesized with various ortho substitutions to the squaraine ring. As hypothesized, we found that placing moderately electron donating groups in the ortho position conveyed intermediate reactivity to thiols between the essentially non-reactive hydroxyl-substituted squaraines and very reactive non-substituted squaraines. Furthermore, serum albumin was tested for its influence on the addition of thiols to the squaraines. The dyes bind in the hydrophobic cavities of the protein, and thus we expected serum albumin to affect the squaraines' reactivity. Rather than a protective effect by the protein, we found a cooperative effect for thiol addition.

Full text of DOI:10.1016/j.dyepig.2016.11.060

Dyes and Pigments 141 (2017) 316e324
Contents lists available at ScienceDirect
Dyes and Pigments
journal homepage: www.elsevier.com/locate/dyepig
Tuning thiol addition to squaraines by ortho-substitution and the use
of serum albumin
*
Katharine L. Diehl, J. Logan Bachman, Eric V. Anslyn
Department of Chemistry, University of Texas at Austin, 1 University Station A1590, Austin, TX 78712, United States
a r t i c l e i n f o
a b s t r a c t
Article history:
Received 23 August 2016
Received in revised form
Tuning the reactivity of squaraine dyes toward nucleophilic addition of thiols was investigated. A series
of water soluble, aniline-derived squaraines were synthesized with various ortho substitutions to the
squaraine ring. As hypothesized, we found that placing moderately electron donating groups in the ortho
position conveyed intermediate reactivity to thiols between the essentially non-reactive hydroxyl-
substituted squaraines and very reactive non-substituted squaraines. Furthermore, serum albumin was
tested for its influence on the addition of thiols to the squaraines. The dyes bind in the hydrophobic
cavities of the protein, and thus we expected serum albumin to affect the squaraines' reactivity. Rather
than a protective effect by the protein, we found a cooperative effect for thiol addition.
3
November 2016
Accepted 5 November 2016
Available online 20 February 2017
Keywords:
Chemosensors
Dyes
©
2017 Elsevier Ltd. All rights reserved.
Michael reaction
Substituent effect
1
. Introduction
addition of a nucleophile to a squaraine (Scheme 1) breaks up the
extended conjugation of the molecule, which switches off the
distinctive absorbance and fluorescence of the molecule. Squar-
aines have been shown to be selective probes for thiols at physio-
logical pH [17e28]. While the addition of thiols to squaraines is a
well-known reaction, not all squaraines are susceptible to nucleo-
philic attack. Typically, the squaraines employed for addition are of
the N,N-dialkylaniline type. Squaraines derived from indoles, ben-
zothiazoles, and other heterocyclic compounds do not undergo
nucleophilic attack and are stable in solution. Phenol-derived
squaraines can be electrophilic if the phenol ring is substituted
with electron withdrawing groups such as halogens [29]. For the
aniline-derived squaraines, electron-donating hydroxyl groups
ortho to the four-membered ring of the squaraine render the dye
essentially non-electrophilic. [24,30]
Besides addition to thiols, squaraines bind to serum albumins
(SAs). This has been investigated primarily for protein detection
and non-covalent labeling [31e41], but also for imaging [42e44]. In
these examples, when these squaraines bind to serum albumin,
large increases in emission (10- to 200-fold) were observed and
were accompanied by bathochromic shifts. The effect of serum al-
bumin binding on the absorption spectra of squaraines is more
variable. In the case of the 3-dicyanomethylene-substituted indo-
lenine dye synthesized by Yarmoluk and coworkers, the absorption
of the squaraine increased about three-fold in the presence of BSA
Squaraines are fluorescent molecules prepared by the reaction
of squaric acid with electron rich aryls, such as N,N-dialkylanilines,
benzothiazoles, phenols, and pyrroles, with azeotropic removal of
water [1]. The structure of a squaraine is a resonance-stabilized
zwitterion. The central four-membered ring is electron deficient,
while the oxygen atoms and the two aniline groups are electron
donating, leading to a donor-acceptor-donor (D-A-D) structure.
From theoretical calculations, Bigelow and Freund showed that
0 1
during the S / S transition there is a charge transfer that is
largely from the oxygen atoms to the four-membered ring with
some minor donation from the aniline groups [2]. This intra-
molecular charge transfer and the extended conjugation of the
molecule lead to absorption in the visible to near-IR region with
ꢀ
1
narrow bands (half bandwidth ~ 750 cm ) [1] and large extinction
5
ꢀ1
ꢀ1
coefficients (ε ꢁ 10 cm
M
) [3]. Thus, squaraines have found
application in imaging [1,4,5], photovoltaics [6e8], ion sensing
[9e13], and other areas [14,15].
Martinez-Manez and coworkers demonstrated in 2002 that the
electron deficient four-membered ring in a squaraine is susceptible
to nucleophilic attack [16]. They used a squaraine as a cyanide
probe in acetonitrile/water solutions buffered at pH 9.5. The
[
32]. However, Ramaiah and coworkers observed more complex
*
Corresponding author.
E-mail address: anslyn@austin.utexas.edu (E.V. Anslyn).
absorbance behavior with their phloroglucinol-derived squaraine
http://dx.doi.org/10.1016/j.dyepig.2016.11.060
143-7208/© 2017 Elsevier Ltd. All rights reserved.
0

K.L. Diehl et al. / Dyes and Pigments 141 (2017) 316e324
317
Scheme 1. Thiol addition to a squaraine.
[34]. The authors attributed this phenomenon to site selectivity of
the binding. Similarly, Belfield and coworkers observed an initial
decrease in absorption of their benzoindolenine-based squaraine
upon addition of BSA, followed by an increase at higher equivalents
of protein [42].
Scheme 2. Synthesis of SQ4, 5, 7, 8, and 9.
In this work, we synthesized water-soluble squaraines and used
spectroscopic methods to study their thiol addition properties. We
tested a series of aniline-derived, ortho-substituted squaraines to
determine how different electron-donating groups affected the
reactivity of the squaraines toward thiols. Although we expected to
find that the reactivity would correlate directly with electron-
donating ability of the ortho substituent, we instead found that
the effect of the group is more complex than electronic factors
alone. Furthermore, we investigated the effect of serum albumin on
the reactivity of the squaraines toward thiols. Ramaiah and co-
workers identified a protective effect on phloroglucinol-derived
aniline derivative and squaric acid in benzene/n-butanol via azeo-
tropic removal of water. The procedure was based on a protocol
described by Ghazarossian and coworkers to synthesize SQ4 [46].
We followed that procedure for the synthesis of SQ4 but modified it
for the synthesis of the other squaraines. In that protocol, the sul-
fonated anilinium was used directly in the squaraine formation
reaction along with sodium bicarbonate in situ to obtain the free
base. The sodium bicarbonate interfered with the squaraine for-
mation by deprotonating squaric acid (pK 1.5, 3.4) [47]. Thus, we
a
squaraines by
b-cyclodextrin, although the cyclodextrin did not
obtained the free base of the sulfonated aniliniums by carefully
protect aniline-derived squaraines from thiol addition [29]. To this
end, we first titrated the squaraines with serum albumin to
investigate their binding behavior in the hydrophobic pockets of
the protein. Then, we tested whether the presence of serum albu-
min would enhance or impede the thiol addition to the squaraines.
We found that the presence of serum albumin generally enhanced
thiol addition.
adjusting an aqueous solution of the compound to pH 7 and then
removing the water to isolate the sodium salt. The salt was used to
form the squaraine. Furthermore, while Ghazarossian et al. suc-
cessfully used ethylene glycol and DMSO to ameliorate the poor
solubility of the sulfonated anilinium in benzene/butanol, we found
that the presence of these solvents hindered the formation of
squaraine rather than improving it. While their protocol yielded a
small amount of SQ4, we could not make SQ5 in this manner.
2
. Results and discussion
2.2. Spectroscopic properties of the aniline-based squaraines
2.1. Synthesis
The absorbance spectra for SQ4, SQ5, SQ7, SQ8, and SQ9 are
shown in Fig. 2, and the optical properties are summarized in
Table S1. These compounds exhibit typical squaraine absorbance
properties: bands between 500 and 700 nm and extinction co-
The squaraines synthesized were all water soluble or able to be
dissolved in water with a small amount (<10%) of acetonitrile,
methanol, or dimethyl sulfoxide (DMSO). SQ1 [45], SQ2 [42], SQ3
34,35], and SQ6 [20] (Fig. 1) were synthesized according to liter-
ature procedures.
We synthesized a series of squaraines with different sub-
stituents on the phenyl rings ortho to the cyclobutene ring (Scheme
). To do so, the appropriate aniline derivative was monoalkylated
and then sulfonated. Next, the squaraine was formed from the
5
ꢀ1
ꢀ1
efficients near 10
M
cm . The squaraines are fluorescent in
[
aqueous solution, but the intensity is relatively low. From the
absorbance spectra, we can conclude that these squaraines exist as
both monomers and dimers in solution [30,48,49]. The dimer peak
is around 595 nm, while the monomer peak is around 650 nm. Only
the monomer is emissive. These dimers are H-type aggregates,
2
Fig. 2. Absorbance spectra of SQ4, 5, 7, 8, and 9. The squaraines (10
in 10 mM phosphate buffer, H O, pH 7.00, 0.02% NaN
. 4 ¼ OH, 5 ¼ H, 7 ¼ CH
8 ¼ OCH , and 9 ¼ NH(CO)CH
mM) were dissolved
2
3
3
,
Fig. 1. Structures of squaraines (SQ1-SQ9).
3
3
.

318
K.L. Diehl et al. / Dyes and Pigments 141 (2017) 316e324
which are common for squaraines [3]. SQ5, SQ7, and SQ8 exist
primarily as the monomer in aqueous solution. SQ4 is a mixture of
the two forms, and SQ9 is mostly dimeric.
Table 1
Fitting of binding curves of NAC added to SQ to 1:1 regime.
Squaraine Binding constant Relative standard
Covariance RMS of
ꢀ
1
(M
)
error (%)
of fit
residuals
5
5
4
3
2
2
.3. Thiol addition to squaraines
SQ7
SQ5
SQ8
SQ9
SQ4
8.57 ꢂ 10
4.00 ꢂ 10
1.92 ꢂ 10
2.62 ꢂ 10
6.44 ꢂ 10
9
3
11
11
12
0.017
0.005
0.003
0.009
0.038
0.032
0.012
0.006
0.015
0.019
Evaluation of the addition of thiols to the fluorescent squaraines,
was performed using N-acetylcysteine (NAC). Absorbance and
fluorescence measurements were utilized to monitor thiol addi-
tion. The data for thiol addition to SQ2 (Fig. S1), SQ3 (Figs. S2 and
S3), and SQ6 (Fig. S4) can be found in the SI.
should be more donating than an amide group, it is possible that
the monomer-dimer equilibrium for SQ9 can explain its extra sta-
bility compared to SQ8. SQ9 shows a strong tendency to dimerize
compared to the other squaraines. A dimer would be less suscep-
tible to nucleophilic attack statistically (i.e. two sides per two
squaraines rather than two sides per one squaraine). Additionally, if
the SQ9 dimer were more stable than or similar in stability to the
SQ9-thiol complex, then the equilibrium would favor the SQ9
2
.3.1. SQ4, SQ5, SQ7, SQ8, and SQ9
We set out to investigate if we could tune the reactivity toward
thiols through substitution at the ortho position on the benzene
ring (X in Scheme 2). It has been demonstrated that squaraines with
no ortho substitution (X ¼ H, SQ5) will react readily with thiols. On
the other hand, squaraines with a hydroxyl group in the ortho po-
sition (X ¼ OH, SQ4) are essentially non-reactive to thiols, pre-
sumably because the hydroxyl group donates electron density to
the squaraine core, decreasing electrophilicity. We hypothesized
that placing other donating groups at the ortho position would
result in intermediate reactivity between a hydrogen and a hy-
dimer. These additional factors could make
stituents inapplicable in this case.
s values for sub-
2
.3.2. 1H NMR of thiol addition to squaraine
The desymmetrization of the squaraine upon thiol addition was
droxyl group. We synthesized squaraines with X ¼ CH
3 3
, OCH , and
1
1
observable by H NMR (Figs. S10eS12). The H NMR evidence is
indicative of the reversibility of thiol addition to squaraines. After
one equivalent of NAC was added, the solution decolored
NH(CO)CH (SQ7, SQ8, and SQ9) in order to test this hypothesis.
3
NAC was titrated into each squaraine while monitoring the absor-
bance (Fig. 3 and Fig. S5-S9). From the isotherms, binding constants
were calculated (Table 1).
From Fig. 3 and Table 1, it is clear that SQ5 and SQ7 react with
thiols readily, while SQ8 and SQ9 do not react as readily. SQ4 is
essentially not reactive, as expected. The reactivity of SQ7 appears
to be slightly greater than for SQ5, which is unexpected because a
methyl group is electron-donating. Furthermore, SQ8 is expected to
be less reactive than SQ9 because a methoxy group is a better donor
than an amide, but we observe the opposite order of reactivity. The
1
completely, and the H NMR spectrum indicated complete con-
version of SQ5 to SQ5-thiol. The SQ5-thiol peaks then diminished
over time because the thiols were constantly leaving and re-adding
to SQ5, and at the same time the free thiols in solution were also
being oxidized to disulfides in a separate reaction. The thermody-
namic sink for the system is the state in which the thiols are
oxidized to disulfides and the squaraines are hydrolyzed.
2.4. Reactivity of squaraines to amines
observed trends do not agree well with Hammett
s values
þ
(Table S2) [50], although
s
values appear to be more appropriate
The selectivity of squaraines for thiols over amines at pH 6 has
been demonstrated by Ros-Lis et al. [17] However, we wanted to be
sure that this selectivity was retained at pH 7. A large excess of Boc-
than
s
para, which implies that resonance effects are more signifi-
cant than inductive effects from the substituent. A methyl group
does not provide resonance stabilization, which is consistent with
the similar reactivity of SQ5 and SQ7. Furthermore, perhaps SQ7 is
more reactive than SQ5 due to relief of steric strain between the
methyl group and the oxygen when planarity is broken. The higher
reactivity of SQ5 and SQ7 with thiols compared with SQ8 and SQ9
likely has a basis in a steric bulk effect. The larger methoxy and
amide groups effectively block the site of thiol addition as
compared with the minor steric hindrance of a methyl group or
hydrogen, leading to stronger binding. While a methoxy group
a-lysine (175 equivalents of amine) was added to a solution of SQ7,
the most reactive squaraine in our series (Fig. S13). The solution did
not decolor, even after 1 h, indicating that at pH 7 amine nucleo-
philes are not reactive enough to add into our squaraines.
2.5. Squaraine binding to serum albumin (BSA)
We hypothesized that the squaraines would bind to serum al-
bumin and that this binding would provide a positive or negative
effect with respect to thiol addition to the dyes. In order to test the
first part of this hypothesis, BSA was titrated into each solution of
squaraine, while monitoring the absorbance and fluorescence.
Furthermore, we used indicator displacement studies with dansyl
amide (DNSA) and dansyl proline (DP) to probe the site-specificity
of the binding of the squaraines. SQ2 [42] and SQ3 [34] have already
been reported in the literature to bind to BSA., and the data for BSA
binding to SQ1 (Fig. S14) and SQ6 (Fig. S15) can be found in the SI.
2.5.1. SQ4 binding to BSA
SQ4 dimerizes to a significant extent in aqueous solution (Fig. 2),
and this monomer-dimer equilibrium leads to complex absorbance
and fluorescence behavior of the squaraine upon addition of BSA
(
5
Fig. 4A and B). In the absorbance titration, the dimer peak at
85 nm consistently decreased, while a new peak (605 nm) grew in,
which represents the bound dimer. The monomer peak first
Fig. 3. Isotherms for thiol addition to SQ4, SQ5, SQ7, SQ8, and SQ9 overlaid. The
decrease in absorbance for the monomer peak is described as a percentage of the
initial absorbance of each squaraine. See Figs. S5eS9 for experimental details.

K.L. Diehl et al. / Dyes and Pigments 141 (2017) 316e324
319
Fig. 4. A) Absorbance titration of BSA into SQ4: BSA (0-129
0-50 M) to SQ4 (10 M) in 10 mM phosphate buffer, H
m
M) to SQ4 (20
m
M) in 10 mM phosphate buffer, H
ex ¼ 640 nm.
2
O, pH 7.00, 0.02% NaN
3
. B) Fluorescence titration of BSA into SQ4: BSA
(
m
m
2
O, pH 7.00, 0.02% NaN , l
3
increased up to about 0.2 equivalents of BSA, then decreased from
.2 to 0.7 equivalents of BSA, and finally increased up to at least 6
to reaction with the one free cysteine in BSA (Cys-34) [36]. To test
this hypothesis in our system, we prepared a sample of BSA in
which that cysteine had been blocked with iodoacetamide (c-BSA).
The c-BSA was titrated into SQ5, and the binding isotherm was
compared with that of BSA (Fig. S17). The two isotherms are nearly
identical, which indicates that blocking the thiol does not signifi-
cantly change the emission behavior of SQ5 in the presence of
serum albumin. Of course, SQ5 could be reacting with other side
chains. We found with SQ5 that nearly 200 equivalents of lysine did
not decolorize the solution (Fig. S13), which indicates that amines
0
equivalents of protein. The fluorescence titration also showed
complex behavior: the emission decreased up to 0.5 equivalents of
BSA and then increased. These results are indicative of a binding
stoichiometry greater than 1:1, probably because both the mono-
mer and dimer bind to serum albumin. Due to the complexity of the
binding, we were unable to fit the data to obtain an association
constant/s. As a comparison, we performed the same titration with
Human serum albumin (HSA) (Fig. S16). Similar absorbance
behavior to the addition of BSA was observed; however, small dif-
ferences indicate that the binding of SQ4 is different to HSA than it
is to BSA.
will not react. However, the lysines in BSA have an average pK
9.2 [51], and some individual lysines in the protein, particularly in
the binding pocket, could have even lower pK s.
a
of
a
To determine the binding stoichiometry, we used the method of
continuous variation (Fig. S18). The Job plot indicates a 1:1 stoi-
chiometry, which is consistent with the fluorescence binding
2
.5.2. SQ5 binding to BSA
As BSA was added, the absorbance showed a decrease, while the
5
ꢀ1
(±18%)
isotherm. Thus, an association constant of 1.61 ꢂ 10
M
fluorescence increased (Fig. 5A and B). We were puzzled by the
observation of these apparently opposing phenomena. The absor-
bance decrease implies that the squaraine is becoming desymme-
trized by reaction with a nucleophile, or the charge transfer band is
otherwise being disrupted by steric or electronic factors upon
binding. At the same time, the emission of the squaraine increased,
which implies that binding improves the quantum yield of fluo-
rescence. However, the squaraine was absorbing less light at the
excitation wavelength, so the quantum yield increase must be quite
large. For example, at three equivalents of BSA, the emission had
increased by about five-fold, while the absorbance at the excitation
wavelength had decreased by about five-fold. It is likely that we are
observing the net effect of multiple factors. The data could be
explained by some fraction of the SQ5 reacting with nucleophilic
amino acids in the protein (e.g. lysine, cysteine), while at the same
time some fraction of squaraine binds to the protein and experi-
ences a large increase in its emission due to binding in the hydro-
phobic pocket/s.
was obtained from the data (Fig. 5A and B). To investigate which
site SQ5 binds in, solutions of BSA with dansyl amide (DNSA, Site I-
specific ligand) and with dansyl proline (DP, Site II-specific ligand)
were titrated with SQ5 while measuring the emission of DNSA and
DP (Fig. S19) [51]. Both probes experienced a decrease in their
emission, which is indicative of their displacement from BSA. DP
experienced a larger decrease than DNSA, and thus we concluded
that SQ5 is primarily binding in Sudlow Site II, which is consistent
with the results for SQ2 reported by Belfield et al. [42] Site II
generally prefers aromatic molecules with peripherally-located
negative charges [51], so it makes sense that these sulfonate-
functionalized squaraines would bind in Site II.
Furthermore, a solution of BSA/SQ5 was titrated with DNSA and
DP separately (Figs. S20eS23). We postulated that, if the absor-
bance behavior of SQ5 can be attributed to a non-covalent effect of
being bound in the hydrophobic pocket, then displacing SQ5 would
restore its absorbance and diminish its emission (Scenario 1).
Conversely, if the absorbance behavior is due to SQ5 reacting with
nucleophilic side chains of the protein, we would expect the probes
Ajayaghosh and coworkers attribute the decrease in absorbance
at 600-700 nm of their squaraine upon addition of serum albumin

320
K.L. Diehl et al. / Dyes and Pigments 141 (2017) 316e324
Fig. 5. A) Absorbance titration of BSA into SQ5: BSA (0-50
0-50 M) to SQ5 (10 M) in 10 mM phosphate buffer, H
m
M) to SQ5 (10
m
M) in 10 mM phosphate buffer, H
2 3
O, pH 7.00, 0.02% NaN . B) Fluorescence titration of BSA into SQ5: BSA
(
m
m
2
O, pH 7.00, 0.02% NaN , l
3
ex ¼ 640 nm.
to have no effect on this reaction, so the absorbance would not
change (Scenario 2). While these were the scenarios we hypothe-
sized, we did not observe either of these outcomes. The addition of
DNSA slightly increased the emission of SQ5, which implies that
DNSA has a small cooperative effect on SQ5 binding. From the other
indicator displacement experiments (Fig. S19), we determined that
SQ5 binds in Site II, so perhaps the binding of DNSA in Site I induces
allosteric changes in BSA that render the Site II more hospitable to
SQ5. The absorbance data echoed a cooperative effect by DNSA. The
decrease in absorbance of SQ5 upon addition of DNSA suggests that
the decrease observed upon SQ5 binding to BSA is not due to re-
action with lysines, since DNSA would not affect that reaction. From
both the absorbance and fluorescence data (Figs. S22 and S23), we
observed that DP displaced SQ5 starting at about two equivalents,
which is consistent with our Site II binding model and our Scenario
that were consistent with SQ5 (Fig. 7A and B). Fitting the data to a
1:1 binding regime yielded an association constant of
4
ꢀ1
4.71 ꢂ 10
M
(±69%) from the absorbance and fluorescence ti-
trations. Nevertheless, at < 0.2 equivalents of BSA, a small inflection
is present in both isotherms. This phenomenon is due to a higher
stoichiometry of binding occurring when very little BSA is present
(e.g. dimer binding). Clearly, this effect was much smaller for SQ8
than for SQ7 or SQ4.
2.5.5. SQ9 binding to BSA
From SQ9's absorbance spectrum we found that it exists pri-
marily as the dimer in solution (Fig. 8A and B). Perhaps the amide
groups stabilize the dimer by hydrogen bonding, making it the
dominant form for SQ9, but not for SQ5, SQ7, or SQ8. As BSA was
added to SQ9, the dimer peak diminished and the monomer peak
increased. This titration implies that the BSA only binds the
monomer and so breaks up the dimer. Because SQ9 did not exhibit a
decrease in its monomer peak upon BSA addition as did SQ5, SQ7,
and SQ8, we used the same tests as we performed for SQ5 to probe
the effect of displacement by DNSA and DP to compare to SQ5
1
for indicator displacement. The increase in emission from 0 to 2
equivalents of DP matches with the decrease in the dimer peak in
the absorbance spectrum. The presence of the DP appears to break
up the dimer into monomer, which leads to the initial increase in
emission upon addition of DP. These experiments provided insight
into how SQ5 interacts with BSA; however, it is still unclear exactly
to what phenomenon the decoloration of SQ5 in the presence of
BSA can be attributed.
(Figs. S24e27) and found a Site II binding model also applies to SQ9.
2.6. Thiol addition to squaraines in the presence of serum albumin
2
.5.3. SQ7 binding to BSA
The evaluation of SQ7 again showed that absorbance decreased
Finally, we measured whether the serum albumin exerts a
protective or cooperative effect on thiol addition to squaraines.
Solutions of BSA/SQ were titrated with thiol while monitoring the
absorbance. In order to compare the effects of BSA on thiol addition
to squaraine, the decrease in absorbance of the BSA/SQ solution was
standardized by expressing it as the percent decrease of the initial
absorbance (Fig. 9A and B). The original spectra and titration details
for the experiments can be found in the SI (Figs. S28eS32) as well as
plots comparing thiol addition to SQ and SQ/BSA for each dye
as BSA was added, while the emission increased to about 0.5
equivalents of BSA and then decreased (Fig. 6A and B). The shape is
similar to the binding isotherm for SQ4, but with SQ7 the emission
initially increased rather than decreasing as with SQ4. Thus,
perhaps either the dimer or two separate monomers can bind to
serum albumin. Due to the complexity of the binding, we were
unable to fit the data to obtain an association constant.
(Figs. S33eS37). Table 2 shows the association constants calculated
2
.5.4. SQ8 binding to BSA
The fluorescence and absorbance data for SQ8 showed features
from the data. The order of reactivity from Fig. 3 was largely
retained in the presence of serum albumin.

K.L. Diehl et al. / Dyes and Pigments 141 (2017) 316e324
321
Fig. 6. A) Absorbance titration of BSA into SQ7: BSA (0-129
0-100 M) to SQ7 (20 M) in 10 mM phosphate buffer, H
m
M) to SQ7 (20
m
M) in 10 mM phosphate buffer, H
2
O, pH 7.00, 0.02% NaN
3
. B) Fluorescence titration of BSA into SQ7: BSA
(
m
m
2
O, pH 7.00, 0.02% NaN
3
,
l
ex ¼ 640 nm, em ¼ 690 nm.
l
Fig. 7. A) Absorbance titration of BSA into SQ8: BSA (0-129
m
M) to SQ8 (20
mM) in 10 mM phosphate buffer, H
2 3
O, pH 7.00, 0.02% NaN . B) Fluorescence titration of BSA into SQ8: BSA
(
0-100 M) to SQ8 (20 M) in 10 mM phosphate buffer, H
m
m
2
O, pH 7.00, 0.02% NaN
3
,
l
ex ¼ 630 nm, em ¼ 675 nm.
l
It was observed that BSA exerts a cooperative effect on thiol
thereby increasing the relative concentration of squaraine mono-
mer, resulting in a larger amount of thiol addition. Another alter-
native is that the effect could be due to the squaraines reacting with
nucleophilic side chains of the protein, which decreases the con-
centration of free squaraine in solution, showing the same resulting
loss of fluoresence as would thiol addition. Thus, it would experi-
mentally appear that less thiol is needed to bind the squaraine.
However, the control studies described above indicate that the
squaraine does not react with the SAs. Regardless of the cause, our
results are consistent with those of Ramaiah and coworkers [29], in
addition to the squaraines, and there are several possible mecha-
nisms for this effect. The most clear-cut way for this to occur is that
the SQ-thiol complex is bound more tightly by the protein than the
squaraine alone, and this thermodynamically drives the thiol to add
to the squaraine. However, because the equilibria involved in
squarane dimerization, thiol addition, and SA binding are all
interrelated (Scheme 3), we have considered two other possibil-
ities. One alternative is that SA could have a cooperative effect by
breaking up squaraine dimers that are resistent to thiol addition,

322
K.L. Diehl et al. / Dyes and Pigments 141 (2017) 316e324
Fig. 8. A) Absorbance titration of BSA into SQ9: BSA (0-129
0-100 M) to SQ9 (20 M) in 10 mM phosphate buffer, H
m
M) to SQ9 (20
m
M) in 10 mM phosphate buffer, H
2 3
O, pH 7.00, 0.02% NaN . B) Fluorescence titration of BSA into SQ9: BSA
(
m
m
2
O, pH 7.00, 0.02% NaN
3
,
l
ex ¼ 650 nm, em ¼ 695 nm.
l
Scheme 3. Equilibria in a squaraine (SQ), thiol, and serum albumin (SA) system.
b-cyclodextrin does not influence the thiol addition reaction, unlike
our study with BSA.
To elucidate among the possible mechanisms of the coopera-
tivity with BSA, we performed indicator displacement experiments
with DNSA and DP. A solution of BSA and DP (or DNSA) was titrated
with SQ5 and an SQ5/NAC complex (Figs. S38 and S39). The emis-
sion of the DNSA or DP was monitored to measure displacement of
these probes by the titrants. DP was displaced somewhat less by
SQ5/NAC compared to SQ5 alone, while DNSA was displaced the
same amount by all three titrants. These titrations indicate that the
SQ5-thiol complex binds less well to serum albumin than SQ5
alone. Thus, we believe the mechanism involving perturbation of
the squaraine dimerization by SA is the best explanation for the
observed cooperativity in thiol addition. The thiol adds reversibly
to the monomeric squaraine better than dimeric squaraine, and the
squaraine is reversibly bound to BSA, an equilibrium which tends to
prefer binding of the monomer. These two equilibria are func-
tioning in parallel, and the increased fraction of monomeric
squaraine leads to an increased amount of thiol addition. This
mechanism is further supported by the finding that thiol addition
to SQ8, which exists almost exclusively as the monomer in solution
(Fig. 2), is essentially unaffected by the presence of SA, while the
other squaraines exhibit 2- to 4-fold increases in their thiol affinity
in the presence of SA (Table 2).
Fig. 9. A) Isotherms for N-acetylcysteine (NAC) addition to all of the SQ/BSA combi-
nations overlaid (monomer). The decrease in absorbance for the monomer peak is
described as
a percentage of the initial absorbance of each squaraine. See
Figs. S28eS32 for experimental details and original spectra. B) Isotherms for thiol
addition to all of the SQ/BSA combinations overlaid (dimer). The decrease in absor-
bance for the dimer peak is described as a percentage of the initial absorbance of each
squaraine. See Figures S28eS32 for experimental details and original spectra.
3
. Conclusions
that they did not detect a protective effect of
aniline-derived squaraine, despite detecting binding. They did not
find a cooperative effect by -cyclodextrin either, which means that
b-cyclodextrin on their
Squaraines are organic chromophores with absorption and
b
emission in the red to near-IR region of the spectrum, making them

K.L. Diehl et al. / Dyes and Pigments 141 (2017) 316e324
323
Table 2
Fitting of binding curves of NAC added to BSA/SQ to 1:1 regime.
4.2. SQ4 [46]
Squaraine Binding constant Relative standard Covariance RMS of
To an oven-dried flask was added squaric acid (103 mg,
0.9 mmol, 1 eq), 3-(((3-hydroxyphenyl)(ethyl)ammonio)propane-
1-sulfonate) (490 mg, 1.89 mmol, 2.1 eq), sodium bicarbonate
(159 mg, 1.89 mmol, 2.1 eq), 10 mL butanol, 5 mL benzene, and 3 mL
DMSO. The reaction was refluxed for 3 h. Ether was added to pre-
cipitate the product. The product was very wet, so it was taken up
in methanol. The methanol was filtered to remove solid, and then
the methanol was removed under reduced pressure to give a green
solid. The crude product was purified by RP-HPLC in water/aceto-
nitrile with 0.1% TFA. The product was a green iridescent solid
ꢀ
1
(
M
)
error (%)
of fit
residuals
6
6
4
3
3
SQ7
SQ5
SQ8
SQ9
SQ4
2.64 ꢂ 10
1.60 ꢂ 10
2.15 ꢂ 10
7.18 ꢂ 10
1.35 ꢂ 10
14
22
11
11
11
0.017
0.043
0.002
0.003
0.007
0.010
0.019
0.006
0.009
0.015
useful dyes for a variety of applications. We synthesized a series of
water-soluble squaraines and investigated tuning their reactivity
toward thiols using variable ortho substitution and the addition of
serum albumin. Aniline-derived squaraines with no ortho substi-
tution are very reactive with thiols and other nucleophiles to the
point that they are unstable in aqueous solution. On the other hand,
squaraines with an ortho hydroxyl group are inert to nucleophiles.
We hypothesized that we could tune access to intermediate reac-
tivity with other donating groups in the ortho position. We found
that this substitution did change the reactivity of the squaraine, and
we achieved intermediate reactivity between the H-substituted
squaraine and the OH-substituted squaraine, although the specific
(7.4 mg, 1.4% yield). MS (ESI): 596.2, 297.1 (M-1). HRMS (ESI): ex-
pected m/z 297.07, found m/z 297.07 ((M-1)/2); expected m/z 617.12,
found m/z 617.13 (M-1).
4.3. SQ5
To an oven-dried flask was added squaric acid (64 mg, 0.6 mmol,
eq), sodium 3-(ethyl(phenyl)amino)propane-1-sulfonate [52]
1
(303 mg, 1.2 mmol, 2.1 eq), 16 mL benzene, and 8 mL n-butanol.
A Dean Stark trap was attached, and the reaction was refluxed at
ꢃ
1
00 C for 30 h. The reaction was filtered, and the blue solid was
order of reactivity did not conform to
were able to rationalize these inconsistencies based on additional
factors for the squaraines that the values would not take into
s values for these groups. We
washed with isopropanol. The crude product was purified on a C-18
RediSep column with a CombiFlash instrument eluting with water/
s
1
acetonitrile. The product was a purple solid (5.51 mg, 1.6% yield). H
account. Next, we investigated the binding of these squaraines to
serum albumin as a mechanism by which the protein could
encourage or inhibit thiol addition to the dyes. The sulfonated
squaraines exhibit an increase in their emission upon binding to
BSA. Some of the squaraines bind with greater than 1:1 stoichi-
ometry to BSA, which could be due to the dimer binding in one site
or multiple monomers in different sites. The dominant binding site
for these squaraines seems to be Sudlow Site II, as evidenced by
indicator displacement titrations with DNSA and DP. We detected a
cooperative effect of the serum albumin on the thiol addition to the
squaraines, which we attribute to perturbation of the squaraine
monomer-dimer equilibrium by serum albumin.
NMR (400 MHz, d-DMSO, ppm):
CH ), 2.45 (m, 4H, CH ), 3.53 (m, 4H, CH
H, Ar-H), 8.06 (d, 4H, Ar-H). LRMS (ESI): 281.2 ((M-1)/2). HRMS
ESI): expected m/z 281.07, found m/z 281.07 ((M-1)/2); expected
m/z 585.13, found m/z 585.14 (M-1).
d
1.13 (t, 6H, CH
3
), 1.85 (m, 4H,
2
2
2
), 3.60 (t, 4H CH
2
), 6.98 (d,
4
(
4.4. SQ7
To an oven-dried flask was added squaric acid (100 mg, 0.9 mmol,
eq), sodium 3-(ethyl(m-tolyl)amino)propane-1-sulfonate [52]
1
(489 mg, 1.8 mmol, 2 eq), 16 mL benzene, and 8 mL n-butanol. A
Dean-Stark trap was attached, and the reaction was refluxed at
ꢃ
1
00 C for 24 h. The solvent was removed under reduced pressure.
The crude product was purified on a C-18 RediSep column with a
CombiFlash instrument eluting with water/acetonitrile. The product
was a purple solid (6.8 mg, 1.2% yield). H NMR (400 MHz, d-DMSO,
4
. Experimental section
1
ppm):
d
1.15 (t, 6H, CH
), 3.53 (m, 4H, CH
3
),1.88 (m, 4H, CH
2
), 2.45 (m, 4H, CH
2
), 2.76 (s,
General: Unless otherwise indicated, chemicals and reagents
6H, CH
3
2
), 3.58 (t, 4H CH
2
), 6.77 (d, 2H, Ar-H), 7.09
were obtained from Sigma Aldrich and used without further puri-
fication. Dansyl amide and dansyl proline piperidinium salt were
obtained from TCI America. The fluorescence experiments were
performed with a PTI fluorimeter using an 814 photomultiplier
detection system and a 75W xenon short arc lamp. The absorbance
experiments were performed with a Beckman Coulter DU 800
Spectrophotometer and an Agilent Cary 100 UV-VIS. The binding
constants were calculated using Open Data Fit by the Centre for Bio-
Nano Science (CBNS) at http://supramolecular.org/.
(
(
m, 2H, Ar-H), 8.06 (d, 2H, Ar-H). LRMS (ESI): 295.1 ((M-1)/2), 591.1
M-1). HRMS (ESI): expected m/z 295.09, found m/z 295.09 ((M-1)/
2); expected m/z 613.17, found m/z 613.17 ((M-1)þNa).
4.5. SQ8
To an oven-dried flask was added squaric acid (70 mg,
0.61 mmol, 1 eq), sodium 3-(ethyl(3-methoxyphenyl)amino)pro-
pane-1-sulfonate [52] (381 mg, 1.29 mmol, 2.1 eq), 10 mL benzene,
and 5 mL n-butanol. A Dean-Stark trap was attached, and the re-
action was refluxed at 100 C for 48 h. The reaction mixture was
ꢃ
4
.1. Synthesis
filtered to collect the product, washing with ethyl acetate and
allowing to dry thoroughly. The crude product was purified on a C-
18 RediSep column with a CombiFlash instrument eluting with
The syntheses for SQ1, SQ2, SQ3, and SQ6 were already reported
in the literature as cited above. For the squaraines that had not
previously been synthesized, H NMR and high-resolution MS data
water/acetonitrile. The product was a purple solid (26.5 mg, 6%
1
1
yield). H NMR (400 MHz, d-DMSO, ppm):
d
1.12 (t, 6H, CH
), 3.50 (m, 4H, CH ), 3.65 (t, 4H CH
), 6.35 (s, 2H, Ar-H), 6.51 (d, 2H, Ar-H), 8.53 (d, 2H,
3
), 1.82
were obtained. These squaraines have relatively low solubility in
any solvent (<1 mM) and are unstable in their best solvent (water),
decomposing over the course of several hours. Thus, it was not
(m, 4H, CH
2
), 2.45 (m, 4H, CH
2
2
2
),
3.83 (s, 6H, CH
3
Ar-H). LRMS (ESI): 311.2 ((M-1)/2), 623.2 (M-1). HRMS (ESI): ex-
pected m/z 311.08, found m/z 311.08 ((M-1)/2); expected m/z
645.16, found m/z 645.15 ((M-1)þNa).
13
possible to obtain C NMR data due to the high concentrations and
long scan times necessary.

324
K.L. Diehl et al. / Dyes and Pigments 141 (2017) 316e324
4
.6. Sodium 3-((3-acetamidophenyl)(ethyl)amino)propane-1-
[7] Xiao X, Wei G, Wang S, Zimmerman JD, Renshaw CK, Thompson ME, et al. Adv
Mater 2012;24:1956e60.
sulfonate
[8] Wei G, Wang S, Renshaw K, Thompson ME, Forrest SR. ACS Nano 2010;4:
1
927e34.
[9] Yagi S, Hyodo Y, Hirose M, Nakazumi H, Sakurai Y, Ajayaghosh A. Org Lett
007;9:1999e2002.
10] Arunkumar E, Chithra P, Ajayaghosh A. J Am Chem Soc 2004;126:6590e8.
11] Collins CG, Peck EM, Kramer PJ, Smith BD. Chem Sci 2013;4:2557.
To a flask was added N-(3-(ethylamino)phenyl)acetamide [53]
0.5 g, 2.8 mmol, 1 eq), 1,3-propanesultone (0.3 mL, 3.4 mmol, 1.2
2
(
[
[
eq), and 10 mL acetonitrile. The reaction was refluxed for 18 h. The
reaction mixture was concentrated. Water was added to the res-
idue, and the solution was neutralized with 1M NaOH (aq). The
solution was washed with DCM and then lyophilized. The product
[12] Basheer MC, Alex S, George Thomas K, Suresh CH, Das S. Tetrahedron
006;62:605e10.
2
[
[
13] Ajayaghosh A. Acc Chem Res 2005;38:449e59.
14] Ramaiah D, Eckert I, Arun KT, Weidenfeller L, Epe B. Photochem Photobiol
2004;79:99e104.
15] Dirk CW, Herndon WC, Cervantes-Lee F, Selnau H, Martinez S, Kalamegham P,
et al. J Am Chem Soc 1995;117:2214e25.
16] Ros-Lis JV, Martínez-M ꢀa n~ ez R, Soto J. Chem Commun 2002:2248e9.
1
was a white solid (620 mg, 68% yield). H NMR (400 MHz, d-DMSO,
[
[
ppm):
d
1.02 (t, 3H, CH
), 3.23 (m, 4H, CH
.97 (t, 1H, Ar-H), 9.65 (s, 1H, NH). C NMR (400 MHz, d-DMSO,
11.09 (CH ), 22.55 (CH ), 22.68 (CH ), 27.82 (CH ), 48.67
), 60.41 (CH ), 104.06 (Ar-C), 107.75 (Ar-C), 108.57 (Ar-C),
29.07 (Ar-C), 139.37 (Ar-C), 148.02 (Ar-C), 170.18 (C]O). LRMS
ESI): 299.2 (M-1). HRMS (ESI): expected m/z 299.11, found m/z
3
),1.74 (m, 2H, CH
2 3
),1.96 (s, 3H, CH ), 2.35 (m,
2
H, CH
2
2
), 6.33 (d, 1H, Ar-H), 6.84 (m, 2H, Ar-H),
13
6
[17] Ros-Lis JV, García B, Jim eꢀ nez D, Martínez-M aꢀ n~ ez R, Sancen oꢀ n F, Soto J, et al.
J Am Chem Soc 2004;126:4064e5.
ppm):
CH
d
3
3
2
2
[
18] Ros-Lis JV, Martínez-M ꢀa n~ ez R, Soto J, Villaescusa LA, Rurack K. J Mater. Chem
011;21:5004.
(
2
2
2
1
[19] Hewage HS, Anslyn EV. J Am Chem Soc 2009;131:13099e106.
[20] Yan Z, Guang S, Xu H, Liu X. Analyst 2011;136:1916e21.
21] Fan J, Wang Z, Zhu H, Fu N. Sensors Actuators B Chem 2013;188:886e93.
22] Sreejith S, Divya KP, Ajayaghosh A. Angew Chem Int Ed Engl 2008;47:7883e7.
23] Luo C, Zhou Q, Zhang B, Wang X. New J Chem 2011;35:45.
(
[
[
[
2
99.11 (M-1).
4
.7. SQ9
[24] Liu X-D, Sun R, Ge J-F, Xu Y-J, Xu Y, Lu J-M. Org Biomol Chem 2013;11:
258e64.
25] Liao S, Han W, Ding H, Xie D, Tan H, Yang S, et al. Anal Chem 2013;85:
968e73.
[26] Thomas J, Sherman DB, Amiss TJ, Andaluz SA, Pitner JB. Bioconjug Chem
007;18:1841e6.
27] Luo C, Zhou Q, Lei W, Wang J, Zhang B, Wang X. Supramol Chem 2011;23:
57e62.
[28] Fan J, Chen C, Lin Q, Fu N. Sensors Actuators B Chem 2012;173:874e81.
4
[
To an oven-dried flask was added squaric acid (70 mg,
.61 mmol, 1 eq), sodium 3-((3-acetamidophenyl)(ethyl)amino)
4
0
2
propane-1-sulfonate (416 mg, 1.3 mmol, 2.1 eq), 5 mL benzene, and
5
was refluxed at 100 C for 48 h. The reaction mixture was filtered to
collect the product, washing with ethyl acetate and isopropanol
and allowing to dry thoroughly. The product was a green solid and
[
mL n-butanol. A Dean-Stark trap was attached, and the reaction
6
ꢃ
[
[
29] Arun KT, Jayaram DT, Avirah RR, Ramaiah D. J Phys Chem B 2011;115:7122e8.
30] Chen H, Farahat MS, Law K-Y, Whitten DG. J Am Chem Soc 1996;118:
2584e94.
1
required no further purification (212 mg, 49% yield). H NMR
[31] Terpetsching E, Szmacinski H, Lakowicz JR. Anal Chim Acta 1993;282:633e41.
[
[
32] Volkova KD, Kovalska VB, Tatarets AL, Patsenker LD, Kryvorotenko DV,
Yarmoluk SM. Dye Pigment 2007;72:285e92.
33] Volkova KD, Kovalska VB, Losytskyy MY, Bento A, Reis LV, Santos PF, et al.
J Fluoresc 2008;18:877e82.
(
(
6
400 MHz, d-DMSO, ppm):
s, 3H, CH ), 2.39 (m, 4H, CH
.76 (d, 2H, Ar-H), 8.06 (s, 2H, Ar-H), 8.28 (d, 2H, Ar-H) 11.94 (s, 1H,
d
1.24 (t, 6H, CH
3
), 1.86 (m, 4H, CH
2
), 2.17
3
2
), 3.36 (m, 4H, CH
2
), 3.55 (m, 4H CH
2
),
NH). LRMS (ESI): 338.2 ((M-1)/2). HRMS (ESI): expected m/z 338.09,
found m/z 338.09 ((M-1)/2); expected m/z 699.18, found m/z 699.18
[34] Jisha VS, Arun KT, Hariharan M, Ramaiah D. J Am Chem Soc 2006;128:6024e5.
[35] Jisha VS, Arun KT, Hariharan M, Ramaiah D. J Phys Chem B 2010;114:5912e9.
[36] Anees P, Sreejith S, Ajayaghosh A. J Am Chem Soc 2014;136:13233e9.
[37] Fan X, He Q, Sun S, Li H, Pei Y, Xu Y. Chem Commun (Camb) 2016;52:
(
(M-1)þNa).
1
178e81.
38] Xu Y, Liu Q, Li X, Wesdemiotis C, Pang Y. Chem Commun (Camb) 2012;48:
1313e5.
39] Suzuki Y, Yokoyama K. Angew Chem Int Ed 2007;46:4097e9.
[
[
Acknowledgements
1
The authors are grateful to the National Science Foundation for
support of this research through grant NSF (CHE-1212971) and the
Welch Regents Chair (F-0046). The authors would also like to thank
Dr. Pall Thordarson for his help in using the program for fitting the
data to binding constants.
[40] Xu Y, Malkovskiy A, Pang Y. Chem Commun 2011;47:6662.
[
[
[
[
41] Nakazumi H, Ohta T, Etoh H, Uno T, Colyer CL, Hyodo Y, et al. Synth Met
005;153:33e6.
42] Zhang Y, Yue X, Kim B, Yao S, Bondar MV, Belfield KD. ACS Appl Mater. In-
terfaces 2013;5:8710e7.
43] An F-F, Deng Z-J, Ye J, Zhang J-F, Yang Y-L, Li C-H, et al. ACS Appl Mater. In-
terfaces 2014;6:17985e92.
44] Gao FP, Lin YX, Li LL, Liu Y, Mayerh o€ ffer U, Spenst P, et al. Biomaterials
2
Appendix A. Supplementary data
2014;35:1004e14.
[
[
45] Isgor YG, Akkaya EU. Tetrahedron Lett 1997;38:7417e20.
46] Ghazarossian V, Pease JS, Hu MW, Laney M, Tarnowski TL. Fluorescent Dyes.
Supplementary data related to this article can be found at http://
dx.doi.org/10.1016/j.dyepig.2016.11.060.
1989. EP89308412.
[47] West R, Powell DL. J Am Chem Soc 1963;85:2577e9.
[
[
[
[
48] Das S, Kamat PV, De la Barre B, Thomas KG, Ajayaghosh A, George MV. J Phys
Chem 1992;96:10327e30.
49] Das S, Thanulingam TL, Thomas KG, Kamat PV, George MV. J Phys Chem
References
1993;97:13620e4.
[
[
[
[
[
1] Law KY. Chem Rev 1993;93:449e86.
50] Ritchie CD, Sager WF. Progress in physical organic chemistry. Hoboken, NJ,
USA: John Wiley & Sons, Inc.; 1964.
51] Peters T. All about albumin: biochemistry, genetics, and medical applications.
San Diego, CA: Academic Press; 1996.
2] Bigelow RW, Freund H-J. Chem Phys 1986;107:159e74.
3] Sreejith S, Carol P, Chithra P, Ajayaghosh A. J Mater. Chem 2008;18:264.
4] Law K-Y, Bailey FC. Dye Pigment 1988;9:85e107.
5] Cole EL, Arunkumar E, Xiao S, Smith BA, Smith BD. Org Biomol Chem 2012;10:
[
[
52] Tamaoku K, Murao Y, Akiura K, Ohkura Y. Anal Chim Acta 1982;136:121e7.
53] Sadao A. Production of 3-(N-alkylamino)-acylanilide. 1996. JPH08143523.
5769e73.
[6] Silvestri F, Irwin MD, Beverina L, Facchetti A, Pagani GA, Marks TJ. J Am Chem
Soc 2008;130:17640e1.