Live-cell-permeant thiophene fluorophores and cell-mediated formation of fluorescent fibrils

Article abstract of DOI:10.1021/ja2065522

In our search for thiophene fluorophores that can overcome the limits of currently available organic dyes in live-cell staining, we synthesized biocompatible dithienothiophene-S,S-dioxide derivatives (DTTOs) that were spontaneously taken up by live mouse embryonic fibroblasts and HeLa cells. Upon treatment with DTTOs, the cells secreted nanostructured fluorescent fibrils, while cell viability remained unaltered. Comparison with the behavior of other cell-permeant, newly synthesized thiophene fluorophores showed that the formation of fluorescent fibrils was peculiar to DTTO dyes. Laser scanning confocal microscopy of the fluorescent fibrils showed that most of them were characterized by helical supramolecular organization. Electrophoretic analysis and theoretical calculations suggested that the DTTOs were selectively recognized by the HyPro component of procollagen polypeptide chains and incorporated through the formation of multiple H-bondings.

Full text of DOI:10.1021/ja2065522

ARTICLE
pubs.acs.org/JACS
Live-Cell-Permeant Thiophene Fluorophores and Cell-Mediated
Formation of Fluorescent Fibrils
Ilaria Palamꢀa,† Francesca Di Maria,§ Ilenia Viola,|| Eduardo Fabiano,† Giuseppe Gigli,†,‡ Cristian Bettini,^ and
Giovanna Barbarella*,§,^
†NNL-CNR Nanoscience Institute and ‡Dip. Ingegneria Innovazione, Universitꢀa del Salento, Via Arnesano, 73100 Lecce, Italy
§Istitute for Organic Synthesis and Photoreactivity and ^Mediteknology srl, National Research Council, Via Gobetti 101,
40129 Bologna, Italy
NNL, Istituto Nanoscienze ꢀ CNR, via Arnesano, 73100 Lecce, and c/o Dipartimento di Fisica, Universitꢀa La Sapienza, Rome, Italy
S Supporting Information
b
ABSTRACT: In our search for thiophene fluorophores that can
overcome the limits of currently available organic dyes in live-cell
staining, we synthesized biocompatible dithienothiophene-S,S-
dioxide derivatives (DTTOs) that were spontaneously taken up
by live mouse embryonic fibroblasts and HeLa cells. Upon
treatment with DTTOs, the cells secreted nanostructured fluor-
escent fibrils, while cell viability remained unaltered. Comparison
with the behavior of other cell-permeant, newly synthesized
thiophene fluorophores showed that the formation of fluorescent
fibrils was peculiar to DTTO dyes. Laser scanning confocal
microscopy of the fluorescent fibrils showed that most of them were characterized by helical supramolecular organization.
Electrophoretic analysis and theoretical calculations suggested that the DTTOs were selectively recognized by the HyPro
component of procollagen polypeptide chains and incorporated through the formation of multiple H-bondings.
’ INTRODUCTION
All fluorophores reported so far have limits and drawbacks.
For example, fluorescent proteins cannot stain non-encodable
cell components such as glycans, lipids, and DNA. Quantum
dots, which have very appealing optical properties, are not
genetically encodable, need mechanical methods such as micro-
injection to get into the cells, and, once inside, are unable to
target a specific protein. Moreover, they have to be coated with
biologically compatible polymers and are scarcely reproducible.14
Small fluorescent molecules, which are nano-objects of well-
defined structure and dimension that can be engineered by
organic synthesis, may also be difficult to load within the cells
and retain for a sufficiently long time. Organic fluorophores with
good retention times within the cells are scarce, as are those that
uniformily and selectively stain the cytoplasmic compartments.15,16
Furthermore, organic fluorophores often photobleach, becoming
dark and consequently untraceable.
All types of fluorophores, to be useful, must be nontoxic to the
cells, must have sufficiently high absorption coefficients and
quantum yields, must have emission frequencies that are easy
to discriminate from background autofluorescence, must not
photobleach, and must have sufficiently large Stokes shifts to
filter out the exciting light. Last, they should be low cost and easy
to handle for extensive application with the instrumentation
In the past few years, interest in fluorescence optical imaging
has grown exponentially in cell biology, biophysics, and biotech-
nology, prompted by advances in instrumentation1 and light-
emitting probes.2 The increasing availability of sophisticated
techniques and newly designed fluorescent probes opens new
routes to cell exploration and expands the boundaries of possible
discoveries and consequent industrial applications. On one hand,
super-resolution methods—achieving nanoscale resolution in
biological systems—offer unprecedented insights into specific
cellular constituents and intracellular dynamics and processes.1
On the other hand, more efficient synthetic methodologies allow
the preparation of a wealth of new fluorophores: small organic
molecules,3 nanoparticles,4,5 biocompatible6 and nonblinking7
quantum dots, polymer dots,8 photoactivable organic fluorophores,9
target-cell-specific activatable probes,10 and linear and branched
conjugated oligo- and polyelectrolytes.11ꢀ13 Meanwhile, geneti-
cally encodable fluorescent proteins evolve toward increasingly
sophisticated applications for monitoring the inner components
of live cells in real time.2 All these dyes—acting with different
modalities with living cells and their components and through
different types of interactions—demonstrate the tremendous
potential of fluorescence techniques for cell imaging and track-
ing, obtaining real-time mechanistic insights into protein aggre-
gation processes and related diseases, and detection of modified
optical and electronic properties at the cellular level.
Received: July 14, 2011
Published: September 27, 2011
r
2011 American Chemical Society
17777
dx.doi.org/10.1021/ja2065522 J. Am. Chem. Soc. 2011, 133, 17777–17785
|

Journal of the American Chemical Society
ARTICLE
Chart 1. Molecular Structure of the Fluorophores Described in This Study
commonly available in chemical, physical, and cell biology
laboratories. All these requirements make the development of
new fluorophores a rather complex task, since no precise criteria
are available so far to predict their properties. Thus, researchers
are more than ever involved in the search for better fluorophores
that can overcome the limitations of those currently available.
In the past few years, we have developed a few families of
thiophene fluorophores that are optically stable and have easily
tunable properties.17 While there are a fair amount of publications
on thiophene oligomers and polymers as organic semiconductors,18
the number of those describing their use as fluorescent markers
for biological applications is very limited. Due to the versatile
thiophene chemistry, the properties of thiophene fluorophores
can be modulated to make them more robust, more brightly
fluorescent, and easier to conjugate to molecules and polymers of
biological interest.17 However, so far, to be able to penetrate the
membrane of living cells, they need to be conjugated to an appro-
priate carrier.19 Consequently, the development of thiophene
fluorophores with high living cell internalization efficiency remains
challenging for application of these dyes to biological problems.
Here, we report a study describing novel biocompatible
thiophene fluorophores capable of spontaneously penetrating
the membrane of living cells and uniformily staining the cyto-
plasm with long-lasting fluorescence and no harm to the cells. In
the course of this study, we discovered that some dithienothio-
phene-S,S-dioxide derivatives (DTTOs) are cell-permeant fluor-
escent dyes that cause the secretion of fluorescent fibrils by live
cells. We describe the synthesis of the DTTOs, the formation of
fluorescent fibrils when the DTTOs are administered to live cells,
and the characterization of the fibrils by electrophoretic analysis,
confocal microscopy, and atomic force microscopy (AFM). A
hypothesis accounting for the formation of the fluorescent fibrils,
based on experimental data and theoretical calculations, is
reported and discussed.
The synthetic pattern for the preparation of compounds 1ꢀ4 is
reported in Scheme 1, while the preparation of compounds 5ꢀ7 is
reported in Schemes S2 and S3 in the Supporting Information.
The synthesis of the fluorophores was carried out starting from
commercial precursors and was based on Suzuki20 and/or Stille
coupling21 in the presence of palladium catalysts. The fluorophores
were prepared taking advantage of ultrasound (US) and microwave
(MW) assistance. US and/or MW irradiation—low-energy, low-
cost enabling technologies—allowed rapid and efficient prepara-
tions, the fluorophores being obtained in very pure form.
Compounds 1, 1b, 1c, and 1d had already been prepared
through different synthetic pathways.22 Here, the synthesis of
the rigid 3,5-dimethyldithieno[2,3-d:30,2-b]thiophene inner core
(1b) from 3-bromo-4-methylthiophene was made more expedi-
ent by employing commercial bis(tri-n-butyltin)sulfide to obtain
the precursor bis(4-methylthiophen-3-yl) sulfide in 90% yield
(see SI for details). Use of US irradiation allowed us to obtain the
monobromo derivative 1e, almost free of the corresponding
dibromo derivative (1d), in 80% yield in 30 min at room
temperature, using NBS and CH2Cl2:CH3COOH (6:4 v/v) as
the solvent. In the absence of US, NBS needed to be added stepwise
during many hours, and the resulting product always contained a
non-negligible amount of 1d. The reactions of the mono- and
dibromo derivatives with the appropriate stannanes were carried
out with MW assistance, which allowed us to obtain the desired
compound rapidly (1 h) and in high yields (80ꢀ95%).
The detailed descrition of all compounds in Schemes 2, S1,
and S2 is reported in the Supporting Information and in Figures
S1ꢀS26.
Optical Properties. Absorption and photoluminescence max-
imum wavelengths (λmax, λPL), molar absorption coefficients (ε),
and quantum yields (j) of fluorophores 1ꢀ7 in DMSO are given
in Table 1. The absorption and photoluminescence spectra are
reported in Figure S27, together with the experimental details.
Table 1 shows that DTTOs 1ꢀ4 display the highest j values in
DMSO, in the range 0.5ꢀ0.9, whereas the nonfused fluoro-
phores 5ꢀ7 display values in the range 0.3ꢀ0.4, as expected for
ter- and quaterthiophene in solution.22,23 For 1, an absolute
fluorescence quantum yield of 0.87 in dichloromethane, measured
using an integrated sphere, has already been reported by us.22
’ RESULTS AND DISCUSSION
Synthesis. The molecular structures of the fluorophores
described in this study are shown in Chart 1.
17778
dx.doi.org/10.1021/ja2065522 |J. Am. Chem. Soc. 2011, 133, 17777–17785

Journal of the American Chemical Society
ARTICLE
Scheme 1. Ultrasound and Microwave-Assisted Synthesis of DTTO Fluorophores 1ꢀ4 from Commercial 3-Bromo-
4-methylthiophenea
a Reagents and conditions: (i) bis(tri-n-butyltin) sulfide, Pd(PPh3)4, toluene, 130 ꢀC, 90%; (ii) n-BuLi, CuCl2, ethyl ether, 0 ꢀC, 50%;
(iii) 3-chloroperbenzoic acid, CH2Cl2, 70%; (iv) 1 or 2 mmol of NBS, CH3COOH/CH2Cl2, US, 80%, 99%; (v) tributyl(phenyl)stannane, 5%
Pd(PPh3)4, toluene, MW, 80 ꢀC, 95%; (vi) tributyl(phenyl)stannane, 5% Pd(PPh3)4, toluene, MW, 80 ꢀC, 95%; (vii) NBS, CH3COOH/CH2Cl2, US 30
min, room temperature, 99%; (viii),(ix),(x) tributyl(5-methylthiophen-2-yl)stannane, tributyl(5-(2-((2-methoxyethoxy)methoxy)ethyl)thiophen-
2-yl)stannane, tert-butyldimethyl((5(tributylstannyl)thiophen-2-yl)methoxy)silane, 5% Pd(PPh3)4, toluene, MW, 1 h, 80 ꢀC, 80%, 80%, 95%;
(xi) Et3N 3HF, THF, 99%.
3
1ꢀ7 were also stable under prolonged irradiation at the max-
Table 1. Absorption (λmax, nm) and Emission (λPL, nm)
Wavelengths, Molar Absorption Coefficients (ε, cmꢀ1 Mꢀ1),
imum absorption wavelength, and neither appreciable photo-
bleaching nor blinking was observed after many hours of
and Quantum Yields (u) of Fluorophores 1ꢀ7 in DMSO
exposure to the irradiation source, as already reported for
compd λmax λPL
ε
j
compd λmax λPL
ε
j
thiophene fluorophores.17
Cell Staining and Fluorescent Fibrils Formation. The
fluorophores displayed a different degree of solubility in water:
poor solubility for DTTO fluorophores 1ꢀ4, good solubility for
the two alkylsulfonate-substituted fluorophores 6 and 7, and
insolubility for the benzothiadiazole-based 5. Nevertheless, all of
them had the right hydrophilicity/hydrophobicity balance and
the stereoelectronic requirements needed to spontaneously cross
the cell membrane.
Live NIH 3T3 cells—mouse embryonic fibroblast cells of
mesodermal origin whose function is to form the structural fibrils
of connective tissues, including collagen—and HeLa cells—
cervical cancer cells—were employed for the staining experi-
ments. No special loading procedures were needed, since all
fluorophores were spontaneously uptaken by the cells without
apparent leakage. The cells were first incubated for 1 h with the
different fluorophores in buffered solution, and then they were
extensively washed with culture medium to remove the unbound
fluorophore (see SI for details). Through a trial-and-error
procedure it was found that the optimum fluorophore concen-
tration in buffered solution was 0.05 mg/mL. After complete
replacement of the culture medium, the cells were continuously
cultured for several days and monitored at fixed times.
All fluorophores were able to stain with bright fluorescence the
cytoplasm of the cells, while the nucleus remained dark. In all
1
2
3
4
408 514 18 030 0.85
430 544 22 393 0.49
429 543 22 397 0.50
428 542 20 079 0.51
5
6
7
479 613 26 142 0.34
346 428 12 645 0.30
409 589 20 873 0.42
The j values of Table 1 are only indicative of a possible trend
within the cellular environment. Indeed, it is known that the
viscous cellular milieu affects the fluorescence in a way that is very
difficult to predict.17 Probably, two opposing factors are operat-
ing within the cells. On one side, the increase in viscosity leads to
the rigidification of the molecular skeleton, causing the increase
of j; on the other, molecular aggregation causes a drop of j with
respect to the value in DMSO. We notice that, in the case of
DTTO 1, the absolute j value drops to 0.24 on going from
solution to microcrystalline powder.22 Anyway, what we observe
experimentally is that—whatever their precise j value in the
cellular environment—the fluorophores remain intensely fluores-
cent within the cells once they have crossed the cellular membrane.
The molar absorption coefficients of Table 1 are in line with
those already reported for thiophene fluorophores in solution, in
the range 12 000ꢀ26 000 cmꢀ1 M
.
ꢀ1 17 All fluorophores display
large Stokes shifts from absorption to emission wavelengths, also
in line with the trend of thiophene fluorophores.17 Compounds
17779
dx.doi.org/10.1021/ja2065522 |J. Am. Chem. Soc. 2011, 133, 17777–17785

Journal of the American Chemical Society
ARTICLE
conditions.17,25 The same reaction cannot occur within the
nucleus, the acidic region of the cell containing nucleic acids,
which consequently is not stained.
Clearly, an alternative mechanism must operate with DTTO
fluorophores 1ꢀ4 lacking the NHS functionality. Based on our
observations, these fluorophores are first distributed uniformily
into the cytoplasm; afterward they migrate to the perinuclear
region of the cells and concentrate in multiple clusters, acting as
nucleation areas for the formation of fibrils via a cell-mediated
process. All this is illustrated in Figures 1 and 2.
Figure 1A compares the laser scanning confocal microscopy
(LSCM) images of NIH 3T3 cells stained with fluorophore 1
with those of NIH 3T3 cells stained with N-hydroxysuccinimidyl
ester 5. In both cases the images were taken 1 and 48 h after
treatment and estensive washing. It is seen that, after 1 h of
treatment, the cytoplasmatic regions are brightly green or red
fluorescent, depending on the fluorophore, with no or undetect-
able staining of the nuclei. The bright red staining of the cells
treated with 5 persisted after 48 h, during which the cells were
normally proliferating (see below for the cytotoxicity tests) and
the fluorescence was transmitted from mother to daughter cells.
By contrast, in the cells treated with 1, after 48 h there was the
appearance of numerous fibrils, randomly oriented and so
intensely fluorescent as to mask the emission of stained cells
on the background. The amount of fluorescent fibrils formed
increased progressively over time while the cells were normally
proliferating (see below for the cytotoxicity tests). Their forma-
tion and progressive increase were reproducible and were verified
several times with fresh cells and fresh fluorophore, always with
the same results.
To confirm the different cell-labeling mechanisms of the two
different types of fluorophores and simultaneously check
whether the formation of fluorescent fibrils can take place also
in the case of multilabeling of the cells, we treated NIH 3T3 cells
with both fluorophore 5 and fluorophore 1.
Figure 1. (A) Laser scanning confocal microscopy images of NIH 3T3
cells stained with fluorophores 1 taken 1 and 48 h after treatment and
estensive washing. After 48 h, numerous fluorescent fibrils are present.
(B) LSCM images of NIH 3T3 cells stained with thiophene-based N-
hydroxysuccinimidyl ester 5 taken 1 and 48 h after treatment and
washing. (C) Overlay of images of NIH 3T3 cells treated first with 5 and
then with 1, showing the formation of green fluorescent fibrils. Scale
bars, 75 μm.
NIH 3T3 cells were first incubated for 1 h with 5 and washed;
afterward they were incubated for 1 h with 1 and washed again.
The results are reported in Figure 1C, showing the overlay of the
corresponding fluorescence images taken 48 and 168 h (7 days)
after treatment with the fluorophores. Immediately after uptake
of both fluorophores, merging of green and red cellular labeling
was visible. After 48 h, the cells were still proliferating and the
fluorescence was transmitted from mother to daughter cells (see
also Figure S29). A few days after treatment, a few green fluo-
rescent fibrillar structures were detected while the red staining of
cells cytoplasm by 5 became predominant. Apparently, the
presence of the second fluorophore causes only some delay in
the secretion of fluorescent fibrils.
Figure 2 shows that intensely green fluorescent fibrils were
also produced by NIT 3T3 cells treated with DTTO fluoro-
phores 2 and 3. With 2 and 3 the formation of fibrils was slower
than with 1. Parts A-c and B-c of Figure 2 show fluorescent fibrils
formed 72 h after treatment. Apparently, the size and morphol-
ogy of the fibrils changed on changing the molecular structure of
the fluorophore, shorter and in tight bundles with 2, longer and
more isolated with 3.
cases, the fluorescence persisted for at least 7 days (the maximum
time we were able to monitor the cells), during which the cells
were normally proliferating, and was transmitted from mother to
daugther cells during the replication process (Figures 1 and 2).
However, there was a distinctive difference between the behaviors
of fluorophores 5ꢀ7 and DTTO fluorophores 1ꢀ4 causing the
production of fluorescent fibrils.
Fluorophores 5ꢀ7 stain the cytoplasm of live NIH 3T3 cells
with bright and persistent fluorescence, as shown in Figure 1 for
the red-emitting fluorophore 5. Since these fluorophores are
functionalized with the succinimidyl ester group (NHS), they are
likely to be retained within the cells through the covalent binding
of NHS to primary amino groups of intracellular proteins. It has
been shown that this is indeed the mechanism leading to
cytoplasm staining of live cells by carboxyfluorescein diacetate
succinimidyl ester, one of the most commonly used dyes to stain
the cytoplasm of living cells.24 We have already demonstrated
that, with thiophene N-succinimidyl esters, the reaction of NHS
with primary amino groups of proteins to form an amidic bond is
rapid and occurs at very low concentrations in mildly basic
Fluorophore 1 was also tested with HeLa cells, and also in this
case the production of fluorescent fibrils was observed. The
experiment carried out with HeLa cells is illustrated in Figure 2C.
The cells were incubated for 1 h with fluorophore 1 at different
times, 18 and 72 h, after they had been seeded on Petri dishes.
The cells were then washed to remove the unbound fluorophore,
17780
dx.doi.org/10.1021/ja2065522 |J. Am. Chem. Soc. 2011, 133, 17777–17785

Journal of the American Chemical Society
ARTICLE
Figure 2. (A) LSCM images of NIH 3T3 cells stained with fluorophore 3 taken 1 h (a) and 48 h (b) after treatment and magnification of some
fluorescent fibrils present 72 h after treatment (c, overlay of images from light transmission and fluorescence microscopies). (B) Same as A with
fluorophore 2. (C) HeLa cells treated with fluorophore 1, 18 (a) and 72 h (b,c) after having been seeded on Petri dishes. Incubation with the fluorophore
lasted for 1 h in both cases, and the images were taken 24 h after treatment.
and LSCM images were taken 24 h later in both cases. It was
found that the amount of fibrils produced by cells treated 18 h
after seeding was much higher than that produced by those
treated 72 h after seeding. Interestingly, the latter cells showed
the presence of green clusters around the perinuclear region,
in particular in the rough endoplasmic reticulum, i.e., in the region
of cells where proteins synthesis takes place (Figure 2Ca). This
experiment suggested that the formation of fibrils was a physio-
logical process and that the fibrils might have a proteic nature. We
envisaged that the difference in the amount of fibrils produced
could be due to the fact that the cells seeded 72 h before
incubation with the fluorophore had already produced all the
typical extracellular matrix proteins and remained in a stationary
protein production phase, while those treated 18 h after seeding
were still in a synthesis phase and were able to incorporate into
the fibrils most of the fluorophore administered to the cells.
In agreement with the assumption that the formation of
fluorescent fibrils was due to a cellular physiological process,
repeated attempts to stain with DTTOs commercial fibrils of
type-I collagen and other main extracellular matrix proteins, such
as laminin and fibronectin, invariably failed. No green fluorescent
fibrils were obtained in these experiments, even using highly con-
centrated fluorophore solutions (see, for example, Figure S28).
The fluorophores were not toxic to the cells. This is illustrated in
Figure3, showing3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyltetrazolium
bromide (MTT) cytotoxicity tests on NIH 3T3 cells treated with
the different fluorophores compared to untreated cells. The tests
indicate that all fluorophores but 4 were not toxic to the cells,
their viability being near 80% or much more, even when the cells
were secreting the fluorescent fibrils. Contrary to the treatment
with the other fluorophores, treatement with 4 caused the
viability of the cells to drop to less than 20%. This compound
had a free terminal hydroxyl group, highly toxic to the cells.
However, when the OꢀH functionality was protected with a
MEM group (ꢀCH2OCH2CH2OCH3), as in 3, the fluorophore
was no more toxic to the cells and contributed to the production
of fluorescent fibrils.
Fibrils Characterization. Most of the fluorescent fibrils secreted
by the cells in the presence of DTTO fluorophores displayed helical
supramolecular organization. Some helical fluorescent fibrils formed by
17781
dx.doi.org/10.1021/ja2065522 |J. Am. Chem. Soc. 2011, 133, 17777–17785

Journal of the American Chemical Society
ARTICLE
Figure 3. MTT cytotoxicity tests for NIH 3T3 cells treated with fluorophores 1ꢀ7 compared to untreated cells (NT).
modulation of the PL intensity related to the helical arrangement.
In general, the fluorescence periodicity seemed to be related to
the complexity of the superhelices (helices of helices) formed.
A wide range of fibrils dimensions was always observed, probably
depending not only on the molecular structure of fluorophore
but also on the stage of development and the number of helices
wrapped up in themselves.
The formation of helical superstructures was also observed by
AFM topography. Figure 6A shows the AFM image of a fixed
HeLa cell having adherent fluorescent fibrils (picked out by
LSCM) obtained upon incubation with fluorophore 1. Figure 6B
shows the AFM image of fluorescent fibrils obtained upon
incubation of NIH 3T3 cells with fluorophore 2 and isolated
from the cells. The helical structures in Figure 6 show the same
structural periodicity measured by confocal microscopy with the
helix step in the range of 2ꢀ5 μm.
Figure 4. Sodium dodecyl sulfateꢀpolyacrylamide gel electrophoresis
of fluorescent fibrils formed by NIH 3T3 cells upon uptake of fluor-
ophores 1ꢀ3 and isolation from the cells.
the cells and picked out by LSCM were isolated from cell medium and
analyzed by sodium dodecyl sulfateꢀpolyacrylamide gel electrophor-
esis (SDS-PAGE), the method commonly used to analyze proteins.
Figure 4 shows the SDS-PAGE carried out with fluorescent
fibrils produced by NIH 3T3 fibroblasts 72 h after treatment with
1ꢀ3 and isolated from the cells (see SI for details). In this
experiment the behavior of the fibrils was compared to that of
commercial fibrils of type-I collagen, which is the predominant
product of fibroblasts activity.26ꢀ28 In Figure 4, lane A shows the
bands of the molecular weight marker employed to monitor the
progress of the electrophoretic run, lane B the bands of type-I
collagen, and lanes CꢀE the bands pertaining to the fluorescent
fibrils formed after treatment of the cells with 1ꢀ3. The
irreversible staining with Coomassie Blue of the electrophoreti-
cally separated bands allowed the visualization of the separated
proteins.
SDS-PAGE unambiguously indicated the proteic nature of the
helical fluorescent fibrils, which were characterized by high
molecular weights, similar to those of type-I collagen. Remark-
ably, the fibrils secreted by the cells upon treatment with 1ꢀ3 in
separate experiments all displayed the same proteic composition.
Figure 5 shows LSCM images, spatially resolved photolumi-
nescence spectra (SR-PL), and 3D LSCM images obtained by PL
z-reconstruction of fibrils formed upon incubation of HeLa cells
with fluorophore 1 and NIH 3T3 cells with fluorophore 3. For
comparison, the supramolecular arrangement of the same fluor-
ophores in cell culture medium (DMEM) but in the absence of
cells is also reported. The figure shows that, in the absence of
cells, the fluorophores form shapeless aggregates, emitting green
light and characterized by a photoluminescence spectrum similar
to that of the dyes in solution. By contrast, the fluorescent fibrils
released in the presence of cells were well structured and
characterized by fluorescence periodicity ranging from 2 to 10 μm.
The spatially resolved PL spectra along the fibrils structure
showed an emission similar to that of the dyes in solution and a
Hypothesis for the Interaction of DTTO Fluorophores with
Protocollagen Polypeptide Chains and Theoretical Calcula-
tions. It is known that fibroblasts and other cell types in culture
are able to produce collagen fibrils that are randomly distributed
in the cell culture.26ꢀ29 Collagen is the most abundant protein in
mammals, synthesized predominantly in the form of type-I
collagen and self-assembled into fibrillar triple helices.30 The
helical morphology of the fluorescent fibrils secreted by our cells
in culture is in agreement with the fact that they contain collagen,
as shown by electrophoretic analysis.
To account for our experimental results, we made the hypoth-
esis that DTTO fluorophores are specifically recognized by some
component of protocollagen polypeptide chains28ꢀ30 and in-
corporated via nonbonding interactions at some stage during the
self-assembly of protocollagen and the subsequent formation of
collagen triple helices. The spontaneous aggregation of triple-
helical collagen molecules into fibrils with embedded DTTOs
renders the fibrils fluorescent and traceable by fluorescence
techniques. The hypothesis appears reasonable in view of the
fact that in vitro studies on collagen growth and self-assembly
from synthetic collagen-mimetic peptides have demonstrated
that the incorporation of exogenous compounds such as metal
complexes can stabilize the triple helix of collagen,31 probably via
stereoelectronic effects.30,32,33
It is known that collagen consists of three left-handed helical
chains coiled around each other to form a right-handed supercoil
stabilized by interchain hydrogen bonds. The polypeptide chains
of procollagen are made of approximately 300 repeats of the
sequence Gly-X-Y, where Gly stands for glycine and X and Y are
predominantly L-proline (Pro) and 4(R)-hydroxy-L-proline
(HyPro).29 For each Gly-X-Y triplet, there is a hydrogen bond
17782
dx.doi.org/10.1021/ja2065522 |J. Am. Chem. Soc. 2011, 133, 17777–17785

Journal of the American Chemical Society
ARTICLE
Figure 5. Laser scanning confocal microscopy (LSCM) images (a), spatially resolved photoluminescence spectra (SR-PL) (b), 3D LSCM images
obtained by PL z-reconstruction (c), and the corresponding optical sections (d) of fibrils formed upon incubation of HeLa cells with fluorophore 1 (A)
and upon incubation of NIH 3T3 cells with fluorophore 3 (B). For comparison, LSCM images (a) and spatially resolved photoluminescence spectra
(SR-PL) (b) of self-assembled fluorophores 1 and 3 in the same medium used for cells culture (DMEM) but in the absence of cells are also reported.
Figure 6. (A) AFM of a frozen HeLa cell treated with 1 and showing adherent helical fibrils. (B) AFM of helical fibrils formed by NIH 3T cells treated
with 2 and isolated from the cells. LSCM showed that, in both cases, the helical fibrils were brightly green fluorescent.
between the amide hydrogen atom of Gly in one chain and the
carbonyl oxygen atom of X in the adjacent chain.
We carried out theoretical calculations on a system consisting
of three sequences of the tripeptide Gly-Pro-HyPro, as model for
17783
dx.doi.org/10.1021/ja2065522 |J. Am. Chem. Soc. 2011, 133, 17777–17785

Journal of the American Chemical Society
ARTICLE
or intermolecular bridges.30 The inner core of DTTO fluoro-
phores, with one oxygen above and one below the molecular
plane, is in principle capable of forming bridges between proto-
collagen chains, contributing to lateral fibrils growth during
the self-assembly process. This matter will be the subject of
further studies aimed at elucidating the precise supramolecular
mechanism by which the fluorescent fibrils are built in the
presence of DTTOs. In this respect, it is worth recalling that
LSCM measurements (Figure 2) suggest that the molecular
structure of the fluorophore might play a role in modulating the
shape and dimension of the fluorescent fibrils and that it is known
that incorporation of metal complexes31 or the presence of
macromolecules such as fibromodulin27,28 can modulate the
shape and the size of self-assembled collagen aggregates.
’ CONCLUSIONS
We have described two classes of biocompatible thiophene
fluorophores capable of spontaneously crossing the membrane
of live cells: the first is capable of staining the cytoplasm
through reaction of the N-succinimidyl ester functionality with
primary amino groups of intracellular proteins; the second,
based on the dithienothiophene-S,S-dioxide moiety (DTTO),
is capable of participating in the physiological formation of
fluorescent fibrils.
We have shown that live cells, upon spontaneous uptake of
DTTO fluorophores, secrete collagen containing fluorescent
fibrils, most of which display helical supramolecular organization.
The process was tested repeatedly and the results were highly
reproducible. Based on electrophoretic analysis, morphology and
theoretical calculations, we interpreted these results on the basis
of molecular recognition between the DTTO and the hydro-
xyproline component of procollagen polypeptide chains, leading
to incorporation of the fluorophore through multiple hydrogen-
bondings.
The results concening DTTO fluorophores are the proof of
principle that cell-permeant, small, and nontoxic fluorescent
molecules can spontaneously transfer their properties to an
intracellular protein in live cells through molecular recognition
and participation in a supramolecular self-assembly process.
The spontaneous uptake of the DTTOs by live cells is not only
a powerful strategy to visualize intracellular processes by fluor-
escence tecniques but also an invaluable tool to introduce
chemical diversity and additional properties into collagen form-
ing within live cells, in a single step. As is the case with many
conjugated molecules, besides fluorescence, DTTOs are also
characterized by additional properties, in particular electro-
activity.22 Thus, DTTO fluorophores are potentially a vehicle
to introduce electroactivity within collagen fibrils. Work is in
progress in this direction with the objective of creating innovative
classes of biocompatible and biologically active materials for
biomedical applications.
Figure 7. Calculated “complex” of fluorophore 1 with three sequences
of the tripeptide Gly-Pro-HyPro (glycine-proline-hydroxyproline) as
model for a collagen strand. The circle indicates the H-bonding formed
in the inner tripeptide between the OꢀH group of HyPro and the
OꢀSꢀO group of the dithienothiophene core of the fluorophore.
procollagen chains,32 in the presence of fluorophore 1, chosen
since it is a symmetric molecule. The structure of isolated
procollagen chain and 1 was preoptimized at the AM1 semi-
empirical level.34 A conformational search was then performed to
probe the possible binding sites for 1 by performing 650
geometry optimizations at the AM1 level, starting from different
relative positions of 1 and the polypeptide chain. The 10 optimized
structures with the lowest energy were further optimized at the
B3LYP/TZVP level35ꢀ38 to refine the results and identify the
binding site of 1 with the lowest total energy. The binding energy
for the finally selected structure was computed at the B3LYP/
TZVP level with a counterpoise procedure to correct for the basis
set superposition error.38 AM1 calculations were performed
using GAUSSIAN03,39 while B3LYP calculations were per-
formed with the TURBOMOLE program package.40 The results
are shown in Figure 7. The calculations predicted an energe-
tically favorable configuration (binding energy 8.2 kcal/mol,
H-bond length 1.95 Å) involving a hydrogen bond between
the HyPro and one of the oxygen atoms of 1. According to the
calculations, the plane of the inner DTTO core was parallel to
the polypeptide axis, while the external phenyl rings were
tilted by 58ꢀ and the oxygen atoms were on opposite sides of
the plane, a geometry corresponding to that determined by
X-ray analysis.22
The calculations support our hypothesis, according to which
DTTO fluorophores are recognized by a component of collagen
polypeptide chains—specifically HyPro—and incorporated via
nonbonding interactions within the polypeptide chain. The
periodicity of DTTO fixation through multiple hydrogen bond-
ings, related to the presence of HyPro at regular distances into
the procollagen polypeptdide chains, creates periodicity in
fluorescence emission at the nanometer scale. The fluorescence
periodicity observed at the micrometer scale in the fibrils
secreted by the cells is probably related to the wrapping up of
several nanometer-sized fluorescent fibrils into themselves.
It is known that hydroxyproline coordinates a network of
water molecules within the triple helix of collagen, forming intra-
’ ASSOCIATED CONTENT
S
Supporting Information. Full experimental details for
b
the synthesis of 1ꢀ7 and their precursors; H and 13C NMR
1
spectra of 1ꢀ7 and their precursors; UVꢀvis and PL spectra of
1ꢀ7; fluorescence microscopy characterization of commercial
synthetic ECM proteins treated with DTTO 1; experimental
details of cell staining; complete ref 39. This material is available
free of charge via the Internet at http://pubs.acs.org.
17784
dx.doi.org/10.1021/ja2065522 |J. Am. Chem. Soc. 2011, 133, 17777–17785

Journal of the American Chemical Society
ARTICLE
’ AUTHOR INFORMATION
(29) Prockop, D. J.; Fertala, A. J. Biol. Chem. 1998, 273, 15598–
15604.
(30) Shoulders, M. D.; Raines, R. T. Annu. Rev. Biochem. 2009,
78, 929–958.
Corresponding Author
barbarella@isof.cnr.it
(31) Marcos, M. P.; Chmielewski, J. J. Am. Chem. Soc. 2009, 131,
2706–2712.
(32) Improta, R.; Mele, F.; Crescenzi, O.; Benzi, C.; Barone, V. J. Am.
Chem. Soc. 2002, 124, 7857–7865.
(33) Hodges, M. D.; Raines, R. T. J. Am. Chem. Soc. 2005,
127, 15923–15932.
(34) Dewar, M. J. S.; Zoebisch, E. G.; Healy, E. F.; Stewart, J. J. P.
J. Am. Chem. Soc. 1985, 107, 3902–3909.
’ ACKNOWLEDGMENT
This research was supported by project FIRB RBPR05JH2P_
ITALNANONET and by ERC Starting Grant FP7 Project
DEDOM (No. 207441).
(35) Becke, A. D. J. Chem. Phys. 1993, 98, 5648–5652.
(36) Lee, C.; Yang, W.; Parr, R. G. Phys. Rev. B 1988, 37, 785–789.
(37) Sch€afer, A.; Huber, C.; Ahlrichs, R. J. Chem. Phys. 1994,
100, 5829–5835.
(38) van Duijneveldt, F. B.; van Duijneveldt-van de Rijdt, J. G. C. M.;
van Lenthe, J. H. Chem. Rev. 1994, 94, 1873–1885.
(39) Frisch, M. J.; et al. Gaussian 03, Revision C.02; Gaussian, Inc.:
Wallingford, CT, 2004.
(40) Ahlrichs, R.; B€ar, M.; H€aser, M.; Horn, H.; K€olmel, C. Chem.
Phys. Lett. 1989, 162, 165–169.
(41) Geng, Y.; Li, H.; Wu, S.; Duan, Y.; Su, Z.; Liao, Y. Theor. Chem.
Acc. 2011, 129, 247–255.
’ REFERENCES
(1) Hell, S. W. Science 2007, 316, 1153–1158.
(2) Newman, R. H.; Fosbrink, M. D.; Zhang, J. Chem. Rev. 2011,
111, 3614–3666.
(3) Krumova, K.; Cosa, J. J. Am. Chem. Soc. 2010, 132, 17560–17569.
(4) Ueno, Y.; Jose, J.; Loudet, A.; Pꢁerez-Bolívar, C.; Anzenbacher, P,
Jr.; Burgess, K. J. Am. Chem. Soc. 2011, 133, 51–55.
(5) Lu, Y.; Dasog, M.; Leontowich, A. F. G.; Scott, R. W. J.; Paige,
M. F. J. Phys. Chem. C 2010, 114, 17446–17454.
(6) Liu, W.; Howarth, M.; Greytak, A. B.; Zheng, Y.; G. Nocera, D.;
Ting, A. Y.; Bawendi, M. G. J. Am. Chem. Soc. 2008, 130, 1274–1284.
(7) Mahler, B.; Spinicelli, P.; Buil, S.; Quelin, X.; Hermier, J. P.;
Dubertret, B. Nat. Mater. 2008, 7, 659–664.
(8) Wu, C.; Schneider, T.; Zeigler, M.; Yu, J.; Schiro, P. G.; Burnham,
D. R.; McNeill, J. D.; Chiu, D. T. J. Am. Chem. Soc. 2010, 132, 15410–15417.
(9) Lee, H. D.; Lord, S. J.; Iwanaga, S.; Zhan, K.; Xie, H.; Williams,
J. C.; Wang, H.; Bowman, G. R.; Goley, E. D.; Shapiro, L.; Twieg, R. J.;
Rao, J.; Moerner, W. E. J. Am. Chem. Soc. 2010, 132, 15099–15101.
(10) Kobayashi, H.; Choyke, P. L. Acc. Chem. Res. 2011, 44, 83–90.
(11) Garner, L. E.; Park, J.; Dyar, S. M.; Chworos, A.; Sumner, J. J.;
Bazan, G. C. J. Am. Chem. Soc. 2010, 132, 10042–10052.
(12) Bj€orka, P.; Nilsson, K. P. R.; Lenner, L.; Kagedal, B.; Persson,
B.; Ingan€as, O.; Jonasson, J. Mol. Cell. Probe 2007, 21, 329–337.
(13) Li, K.; Pu, K. Y.; Cai, L.; Liu, B. Chem. Mater. 2011, 23,
2113–2119.
(14) Resch-Ganger, U.; Grabolle, M.; Cavaliere-Jaricot, S.; Nitschke,
R.; Nann, T. Nat. Methods 2008, 5, 763–775.
(15) Goldys, E. M. Fluorescence Applications in Biotechnology and Life
Sciences; Wiley-Blackwell: Hoboken, New Jersey, 2009.
(16) The Molecular Probes Handbook, A Guide to Fluorescent Probes
and Labeling Technologies, 11th ed.; Johnson, I., Spence, M. T. Z., Eds.;
Life Technologies: Eugene, Oregon, 2010.
(17) Zambianchi, M.; Di Maria, F.; Cazzato, A.; Gigli, G.; Piacenza, M.;
Della Sala, F.; Barbarella, G. J. Am. Chem. Soc. 2009, 131, 10892–10900.
(18) Handbook of Thiophene-Based Materials; Perepichka, I. E.,
Perepichka, D. F., Eds.; John Wiley & Sons: Chichester, 2009.
(19) Duca, M.; Dozza, B.; Lucarelli, E.; Santi, S.; Di Giorgio, A.;
Barbarella, G. Chem. Commun. 2010, 46, 7948–7950.
(20) Suzuki, A. Proc. Jpn. Acad., Ser. B 2004, 80, 359–371.
(21) Espinet, P.; Echavarren, A. M. Angew. Chem., Int. Ed. 2004,
43, 4704–4734.
(22) Barbarella, G.; Favaretto, L.; Sotgiu, G.; Antolini, L.; Gigli, G.;
Cingolani, R.; Bongini, A. Chem. Mater. 2001, 13, 4112–4122.
(23) Becker, R. S.; Seixas de Melo, J.; Macanita, A. L.; Elisei, F.
J. Phys. Chem. 1996, 100, 18683–18695.
(24) Wang, X. Q.; Duan, X. M.; Liu, L. H.; Fang, Y. Q.; Tan, Y. Acta
Biochim. Biophys. Sin. 2005, 37, 379–385.
(25) Barbarella, G.; Zambianchi, M.; Ventola, A.; Fabiano, E.; Della
Sala, F.; Gigli, G.; Anni, M.; Bolognesi, A.; Polito, L.; Naldi, M.;
Capobianco, M. Bioconjugate Chem. 2006, 17, 58–67.
(26) Falini, G.; Fermani, S.; Foresti, E.; Parma, B.; Rubini, K.; Sidoti,
M. C.; Roveri, N. J. Mater. Chem. 2004, 14, 2297–2302.
(27) Canty, E. G.; Kadler, K. E. J. Cell Sci. 2005, 118, 1341–1353.
(28) Hulmes, D. J. S. J. Struct. Biol. 2002, 137, 2–10.
17785
dx.doi.org/10.1021/ja2065522 |J. Am. Chem. Soc. 2011, 133, 17777–17785