High-contrast reversible fluorescence photoswitching of dye-crosslinked dendritic nanoclusters in living vertebrates

Article abstract of DOI:10.1002/anie.201107086

Dendrimers crosslinked with a photochromic diarylethene (blue in picture) were labeled with the fluorescent dye Cy3 (red) for reversible fluorescence photoswitching. If the obtained nanoclusters are irradiated with UV light, the Cy3 fluorescence is quenched by the diarylethene in its ring-closed form, whereas upon irradiation with visible light, the nanoclusters show fluorescence. High-contrast fluorescence imaging was achieved inside a living zebrafish. Copyright

Full text of DOI:10.1002/anie.201107086

.
Angewandte
Communications
DOI: 10.1002/anie.201107086
Photoswitchable Compounds
High-Contrast Reversible Fluorescence Photoswitching of Dye-
Crosslinked Dendritic Nanoclusters in Living Vertebrates**
Yoonkyung Kim,* Hye-youn Jung, Yun Hui Choe, Chaewoon Lee, Sung-Kyun Ko, Soonil Koun,
Yohan Choi, Bong Hyun Chung, Byoung Chul Park, Tae-Lin Huh, Injae Shin, and
Eunkyoung Kim
Recent advances in fluorescence imaging have made this
technique indispensable not only for in vitro investigations of
cells and proteins, but also for in vivo imaging.[1] The
sensitivity of in vivo fluorescence imaging, which is generally
limited by autofluorescence, absorbance, light scattering, and
penetration depth, has been improved significantly, for
example, by employing switchable probes.[2] The utility of
such probes mostly for diagnostic purposes in vivo, however,
can be restricted by their low on–off contrast, irreversibility,
poor delivery, and toxicity. In an effort to make fluorescence-
imaging agents for in vivo applications, herein we prepared
biocompatible dendritic nanoclusters that can repeatedly
exhibit high on–off contrast inside the living zebrafish without
any apparent toxicity. Specifically, a diarylethene[3] derivative,
which interconverts between two isomeric forms by alternate
irradiation with UV and visible light,[4] was used to crosslink
dendrimers and to reversibly quench the fluorescence of
a neighboring fluorophore through fluorescence resonance
energy transfer (FRET).[5]
between the activated and quenched states while avoiding (or
minimizing) self-quenching[1,2] of fluorescence that may occur
when the fluorophores with highly extended p conjugation
(e.g., cyanine 3 (Cy3)) are positioned in close proximity to
each other. In fact, the active form of diarylethene (FRET
acceptor) that can truly quench the fluorescence of Cy3
(FRET donor) is limited to the ring-closed isomer[3,5] (derived
from the antiparallel conformation of the ring-open isomer),
which constituted about 50 mol% of all diarylethene avail-
able after irradiating with UV light when monitored by
1H NMR spectroscopy (see Figure S1 in the Supporting
Information). Accordingly, sparse attachment of Cy3 at the
fifth generation (G5) polyamidoamine (PAMAM) dendri-
mer[6] (ca. 5 nm[7] in diameter) was intended, while controlling
the stoichiometry so that its molar content did not exceed
50% of the diarylethene attached to the dendrimer. Second,
instead of using a dendrimer as the platform, dendrimers were
oligomerized[8] to form a nanocluster (3) using the diaryl-
ethene 2[9] as a crosslinker. By forming an oligomer, both
intra- and inter-dendritic FRET within the same nanocluster
can be achieved, and at the same time the fluorescence signal
of Cy3 from a single molecular entity can be amplified
(nanocluster vs. dendrimer). Moreover, rather than attaching
diarylethene moieties to the surface of a preformed nano-
particle of similar dimensions, the formation of such a porous
nanocluster was sought to better distribute the diarylethene
and Cy3 throughout the nanostructure (both interior and
exterior), leading to more efficient FRET. Third, to obtain
a narrow size distribution of nanoclusters, and thus to exert
more homogeneous biological effects, our dendritic nano-
clusters were 1) synthesized according to a procedure to
prepare size-controlled dendritic nanoclusters[8a] (similar
reaction concentration and feed ratio) and 2) partially
fractionated by size using preparative size-exclusion chroma-
tography (SEC). The analysis of NMR integration and
MALDI both indicated that approximately 1.9 diarylethene
units were attached per dendrimer (Figures S5 and S9,
Table S1 in the Supporting Information). Finally, the residual
surface amino groups of dendritic nanocluster 5 were
converted into a more biocompatible functionality in the
final synthetic step by treating them with either 1) succinic
anhydride to make an anionic surface with carboxylate groups
(for 6) or 2) the N-hydroxysuccinimide (NHS) ester of
tetra(ethylene glycol) methyl ether (mTEG) to make a neutral
surface (for 7). Both surface modifications, as confirmed by
measuring the zeta potential (Table S2 in the Supporting
Information), yielded dendritic nanoclusters (ca. 20 nm) with
The synthesis of dendritic nanoclusters is illustrated in
Scheme 1. Our primary concern in designing a reversible
photoswitch was to achieve the maximal fluorescence contrast
[*] Dr. Y. Kim, H.-y. Jung, Y. H. Choe, C. Lee, Y. Choi, Dr. B. H. Chung,
Dr. B. C. Park
Korea Research Institute of Bioscience and Biotechnology
Daejeon, 305-806 (Korea)
E-mail: ykim@kribb.re.kr
S.-K. Ko, Prof. Dr. I. Shin
Department of Chemistry, Yonsei University, Seoul, 120-749 (Korea)
S. Koun, Prof. Dr. T.-L. Huh
School of Life Sciences and Biotechnology
Kyungpook National University, Daegu, 702-701 (Korea)
Prof. Dr. E. Kim
Department of Chemical and Biomolecular Engineering
Yonsei University, Seoul, 120-749 (Korea)
[**] We are grateful to Dr. Haijun Yao at the Mass Spectrometry
Laboratory of the University of Illinois for collecting MALDI spectra
of our dendrimer samples. We thank Drs. Kang Taek Lee and Sang
Hwan Nam for the measurement of the photobleaching rate, Dr.
Nyeon-Sik Eum, Chijung Yun, and Dr. Nam Woong Song for the
assistance with the installation of optical equipments, and Prof. Dr.
Joon Kim for allowing us to use the DeltaVision imaging system.
This work was supported by the National Research Foundation of
Korea (2010-0008056, 2010-0002205) funded by the Ministry of
Education, Science and Technology and the KRIBB Research
Initiative Program.
Supporting information for this article is available on the WWW
under http://dx.doi.org/10.1002/anie.201107086.
2878
ꢀ 2012 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Angew. Chem. Int. Ed. 2012, 51, 2878 –2882

Angewandte
Chemie
Scheme 1. a) Synthesis of dye-crosslinked dendritic nanoclusters. The subscripts “o” and “c” in the compound numbers of 6 and 7 denote the
ring-open and ring-closed isomers of diarylethene derivatives, respectively. Red circles represent the Cy3 moieties as covalently attached to the
surface of dendrimers. For solution samples, hn=510 nm and hn’=566 nm. PyBOP=(benzotriazol-1-yl-oxy)tripyrrolidinophosphonium hexa-
fluorophosphate, DIEA=diisopropylethylamine, DMSO=dimethylsulfoxide. b) Switching between the ring-open and ring-closed isomers of diaryl-
ethene derivatives. c) UV/Vis absorption spectra of the ring-open (A) and ring-closed (B) isomer of 2 and of Cy3 (C). Fluorescence emission
spectrum of Cy3 when excited at 510 nm (D).
excellent water-solubility and relatively narrow size distribu-
tion as confirmed by transmission electron microscopy and
dynamic light scattering (see Figure S10 and Table S2 in the
Supporting Information).
Angew. Chem. Int. Ed. 2012, 51, 2878 –2882
ꢀ 2012 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.angewandte.org 2879

.
Angewandte
Communications
The photoswitching experiment was performed using
aqueous solutions in cuvettes monitoring with UV/Vis and
fluorescence spectroscopy (Figure 1 and Figures S13–S20 in
the Supporting Information). In fact, deionized water is
inappropriate as medium for biological systems (cells and
animals), thus we prepared sample solutions in phosphate-
buffered saline (PBS; higher in ionic strength) as well as in
deionized water. A 590 nm light, which coincided with the
absorption maximum of the ring-closed isomer of 2 and did
not overlap with the absorption band of Cy3 (i.e., no
photobleaching), was used for the visible-light source
(Scheme 1c). As a control, the aqueous solution of diaryl-
ethene 2 or Cy3 was first exposed to alternate irradiation with
UV (365 nm, 2 min) and visible (590 nm, 30 min) light.
Interestingly, the concentration of the ring-closed isomer of
2 after irradiation with UV light was slightly higher in PBS (by
ca. 16%) than in deionized water (Figure S13 in the
Supporting Information). In fact, the cyclization quantum
yield of a diarylethene derivative was reported to be
dependent on the ratio of two conformers, antiparallel and
parallel, in the ring-open isomer.[10] Therefore, this suggests
that for the ring-open isomer of 2, the amount of the
antiparallel conformation relative to that of the parallel
conformation is higher in PBS than in deionized water.
Essentially no fluorescence was detected from diarylethene 2
under the photoswitching conditions in either solution. For
Cy3, no change in the absorption and emission spectra was
found over three cycles of photoswitching (Figure S14 in the
Supporting Information). Next, photoswitching experiments
were conducted for up to five
Figure 1. Reversible photoswitching experiments using dendritic nano-
clusters 6 (a,c) and 7 (b,d) in aqueous solutions (100 mgmLÀ1):
a,b) UV/Vis spectra in PBS (pH 7.4); c,d) fluorescence intensity mea-
sured at 566 nm (irradiation at 510 nm). Photoswitching was repeated
for up to five cycles, and the samples were irradiated alternately with
UV (365 nm, 2 min, 2.60 mWcmÀ2) and visible (590 nm, 30 min, 250–
255 mW) light, beginning with UV light. For UV/Vis spectra, the
average intensity values at each data point of the first five cycles were
plotted.
cycles using dendritic nanoclus-
ter 6 or 7 (Figure 1 and Figur-
es S15 and S16). A small red-
shift in the emission maximum
of Cy3 (562.5 nm as a free dye)
was noticed when conjugated to
the nanocluster (566.0 nm). All
sample solutions of nanoclusters
showed reversible fluorescence
quenching, with 7 prepared in
PBS being the most efficient
(47% reduction in fluorescence
intensity; Table S3 in the Sup-
porting
Information).
The
degree of fluorescence quench-
ing was roughly correlated to the
amount of available ring-closed
isomer of diarylethene relative
to that of Cy3 (A589,afterUV
/
ACy3max,afterVis). Taken together,
the relative availability of the
ring-closed isomer was higher
for the neutral nanocluster 7
compared with the anionic
nanocluster 6 in either solution,
and the fluorescence quenching
was more effective if photo-
switching was performed in
PBS, particularly for 7 (Fig-
Figure 2. Photoswitching experiments performed on living HeLa cells with internalized dendritic nano-
clusters 6 (a,c) or 7 (b,d). Cells were irradiated alternately with UV (365 nm, 2 min) and visible
(590 nm, 30 min) light for up to 10 cycles. a,b) Fluorescence microscopy images of living HeLa cells.
c,d) The average fluorescence intensity values (mean Æ standard deviation (SD)) measured at the
region of interest (ROI; dotted circles, 49 data points each) after each light application are plotted.
Sequential irradiation with UV (loss of fluorescence) and visible (fluorescence recovery) light constitutes
one full cycle of photoswitching (integers on the x-axis). RD-TR-PE: red fluorescence. Scale bars, 20 mm.
2880 www.angewandte.org
ꢀ 2012 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Angew. Chem. Int. Ed. 2012, 51, 2878 –2882

Angewandte
Chemie
ure 1d), than if the photoswitching was performed in
deionized water.
treated with 7 (Figure 3a,c) without any noticeable morpho-
logical defects.[15b] Indeed, the photoswitching of fluorescence
was enabled by the internalized dendritic nanocluster as
confirmed by cryo-sectioning (Figures S26 and S27 in the
Supporting Information).
Next, the photoswitching experiment was conducted using
living HeLa (human epithelial cervical cancer) cell cultures[11]
with internalized nanoclusters under the same conditions as
for the solution samples (Figure 2 and Figure S24 and
Table S3 in the Supporting Information). The optimal con-
centration of nanoclusters (10 mgmLÀ1) and incubation time
(30 min at 378C) were determined
Herein, the internalization of 7 into zebrafish by perme-
ation, however, did not lead to fluorescence inside the blood
vessels. The use of reversibly photoswitchable dendritic
based on the cytotoxicity profile
(Figure S21 in the Supporting
Information) and kinetics of cellu-
lar uptake (Figures S22 and S23 in
the Supporting Information).[12]
Overall, nanocluster
7 showed
higher cell viability and superior
uptake efficiency[13] compared with
6. Interestingly, both nanoclusters
taken up by living cells exhibited
higher on–off contrast when pho-
toswitched (14.2 for 6 and 19.1 for
7; Figure 2c,d) compared with that
measured in a cuvette (1.3–1.9;
Figure 1c,d and Table S3). This is
probably due to the distinct phys-
icochemical properties (pH value,
temperature, salts, etc.) in the
intracellular microenvironment of
living cells compared with those of
the bulk solutions. Moreover, no
apparent change in the cellular
morphology (i.e., no damage) was
noticed during the photoswitching
experiments that were repeated for
up to 10 cycles.
A recent report described pho-
toswitching inside a living worm by
the internalized diarylethene that
regulated its movement.[14] Herein,
a zebrafish (Danio rerio)[15] was
chosen as an in vivo model to
evaluate the feasibility of reversi-
ble fluorescence switching by the
internalized dendritic nanoclus-
ters. Despite the wide utility of
zebrafish in biomedical research,
only a few reported studies have
employed
fluorescent
probes
based on nanomaterials for zebra-
fish imaging.[15d] Dendritic nano-
cluster 7 was used for its uptake,
presumably by
a
cutaneous
route,[15c] by a three-day-old zebra-
fish. Remarkably strong Cy3 fluo-
rescence was observed from the
zebrafish incubated with 7. Fur-
thermore, a successful reversible
photoswitching for up to 10 cycles
Figure 3. Photoswitching experiments performed on living zebrafish with dendritic nanoclusters 7
internalized by a,c) permeation or b,d) microinjection. The zebrafish was irradiated alternately with UV
(365 nm, 2 min) and visible (590 nm, 30 min) light for up to 10 cycles. a,b) Fluorescence microscopy
images of living zebrafish. c,d) The average fluorescence intensity values (mean ÆSD) measured at the
ROI (dotted circles, 49 data points each) after each light application are plotted. Sequential irradiation
with UV (loss of fluorescence) and visible (fluorescence recovery) light constitutes one full cycle of
photoswitching (integers on the x-axis). DIC: differential interference contrast; RD-TR-PE: red fluores-
was achieved from living zebrafish cence. Scale bars, 400 mm.
Angew. Chem. Int. Ed. 2012, 51, 2878 –2882
ꢀ 2012 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.angewandte.org
2881

.
Angewandte
Communications
[2] a) H. Kobayashi, P. L. Choyke, Acc. Chem. Res. 2011, 44, 83 – 90;
b) R. Weissleder, C.-H. Tung, U. Mahmood, A. Bogdanov, Jr.,
Nat. Biotechnol. 1999, 17, 375 – 378.
[3] a) M. Irie, Chem. Rev. 2000, 100, 1685 – 1716; b) C. Yun, J. You, J.
Kim, J. Huh, E. Kim, J. Photochem. Photobiol. C 2009, 10, 111 –
129.
nanoclusters for fluorescence imaging of the intravascular
region may aid in the facile detection of many crucial diseases
such as inflammation, arthritis, cancer, and cardiovascular
disease. Accordingly, microinjection[16] of 7 into a two-day-old
zebrafish was carried out to test the feasibility of photo-
switching inside the blood vessels.[17] The Cy3 fluorescence
was clearly observed from the blood vessels of a living
zebrafish. Subsequently, the photoswitching of nanocluster 7
inside the vascular system of a zebrafish was performed five
days post fertilization (dpf), when the formation of blood
vessels in zebrafish was nearly complete. To determine the
photoswitching efficiency, the lateral region of a zebrafish,
where the tissue layers are shallow enough to visualize the
individual blood vessels, was selected for analysis. Although
slightly less on–off contrast was achieved compared with the
incubation method (6.0 by microinjection vs. 14.8 by perme-
ation; Table S3 in the Supporting Information), the reversible
photoswitching of 7 inside the blood vessels of a living
vertebrate was fulfilled successfully (Figure 3b,d) without
any apparent harmful effects (Figure S28 in the Supporting
Information).
[4] T. Sakata, Y. Yan, G. Marriott, Proc. Natl. Acad. Sci. USA 2005,
102, 4759 – 4764.
[5] a) M. Irie, T. Fukaminato, T. Sasaki, N. Tamai, T. Kawai, Nature
2002, 420, 759 – 760; b) L. Giordano, T. M. Jovin, M. Irie, E. A.
Jares-Erijman, J. Am. Chem. Soc. 2002, 124, 7481 – 7489; c) M.
Sauer, Proc. Natl. Acad. Sci. USA 2005, 102, 9433 – 9434; d) N.
Soh, K. Yoshida, H. Nakajima, K. Nakano, T. Imato, T.
Fukaminato, M. Irie, Chem. Commun. 2007, 5206 – 5208; e) J.
Cusido, E. Deniz, F. M. Raymo, Eur. J. Org. Chem. 2009, 2031 –
2045; f) S. A. Dꢀaz, G. O. Menꢁndez, M. H. Etchehon, L.
Giordano, T. M. Jovin, E. A. Jares-Erijman, ACS Nano 2011, 5,
2795 – 2805.
[6] R. Esfand, D. A. Tomalia, Drug Discovery Today 2001, 6, 427 –
436.
˘
[7] a) P. K. Maiti, T. C¸ agin, G. Wang, W. A. Goddard III, Macro-
molecules 2004, 37, 6236 – 6254; b) H. Lee, J. R. Baker, Jr., R. G.
Larson, J. Phys. Chem. B 2006, 110, 4014 – 4019; c) C. L. Jackson,
H. D. Chanzy, F. P. Booy, B. J. Drake, D. A. Tomalia, B. J. Bauer,
E. J. Amis, Macromolecules 1998, 31, 6259 – 6265.
In summary, reversibly photoswitchable dendritic nano-
clusters for fluorescence imaging in vivo were developed
using the diarylethene derivative 2 as a crosslinker, which
showed high on–off contrast (6.0–19.1) in living systems.
Moreover, to our knowledge, this is the first biocompatible
fluorescent nanostructure that has internalized effectively
into a living zebrafish via two independent routes—perme-
ation and microinjection—for the imaging of interior organs
and blood vessels, respectively. Unlike the recently reported
photoswitchable GFP-like proteins,[18] our biocompatible
dendritic nanoclusters for high-contrast reversible photo-
switching do not require complex biological manipulations
such as genetic encoding and protein expression and can be
simply treated to or injected into living cells and organisms
like small-molecule fluorophores without any noticeable
toxicity or immunogenic concerns. Additional functional
units such as targeting groups can be attached to the surface
of these nanoclusters to accomplish multiple goals simulta-
neously. Furthermore, with proper structural modification of
the diarylethene component, the spectroscopic properties can
be tuned for photoswitching at a desired wavelength. We
envision that our water-soluble and relatively nontoxic
dendritic nanoclusters may facilitate fluorescence imaging
in vivo by enhancing its resolution through the reversibly
controlled exhibition of high on–off contrast.
[8] a) Z. Cheng, D. L. J. Thorek, A. Tsourkas, Angew. Chem. 2010,
122, 356 – 360; Angew. Chem. Int. Ed. 2010, 49, 346 – 350; b) Y.
Kim, M. F. Mayer, S. C. Zimmerman, Angew. Chem. 2003, 115,
1153 – 1158; Angew. Chem. Int. Ed. 2003, 42, 1121 – 1126.
[9] T. Hirose, K. Matsuda, M. Irie, J. Org. Chem. 2006, 71, 7499 –
7508.
[10] S. Kobatake, T. Yamada, K. Uchida, N. Kato, M. Irie, J. Am.
Chem. Soc. 1999, 121, 2380 – 2386.
[11] Y. Zou, T. Yi, S. Xiao, F. Li, C. Li, X. Gao, J. Wu, M. Yu, C.
Huang, J. Am. Chem. Soc. 2008, 130, 15750 – 15751.
[12] O. P. Perumal, R. Inapagolla, S. Kannan, R. M. Kannan,
Biomaterials 2008, 29, 3469 – 3476.
[13] Our preliminary investigation suggested that several cellular
uptake pathways are simultaneously involved in the internal-
ization of both nanoclusters, with macropinocytosis being the
most preferred route. Macropinocytosis is known to allow for the
increased uptake and facile release of macromolecules into the
cytosol while avoiding lysosomal degradation (see Figure S25 in
the Supporting Information): I. A. Khalil, K. Kogure, H. Akita,
H. Harashima, Pharmacol. Rev. 2006, 58, 32 – 45.
[14] U. Al-Atar, R. Fernandes, B. Johnsen, D. Baillie, N. R. Branda, J.
Am. Chem. Soc. 2009, 131, 15966 – 15967.
[15] a) S.-K. Ko, X. Chen, J. Yoon, I. Shin, Chem. Soc. Rev. 2011, 40,
2120 – 2130; b) V. E. Fako, D. Y. Furgeson, Adv. Drug Delivery
Rev. 2009, 61, 478 – 486; c) S. Harper, C. Usenko, J. E. Hutchison,
B. L. S. Maddux, R. L. Tanguay, J. Exp. Nanosci. 2008, 3, 195 –
206; d) S. Rieger, R. P. Kulkarni, D. Darcy, S. E. Fraser, R. W.
Kçster, Dev. Dyn. 2005, 234, 670 – 681; e) C. Nꢂsslein-Volhard,
R. Dahm, Zebrafish, Oxford Univ. Press, Oxford, 2002.
[16] B. M. Weinstein, D. L. Stemple, W. Driever, M. C. Fishman, Nat.
Med. 1995, 1, 1143 – 1147.
[17] A. A. Beharry, L. Wong, V. Trepepe, G. A. Woolley, Angew.
Chem. 2011, 123, 1361 – 1363; Angew. Chem. Int. Ed. 2011, 50,
1325 – 1327.
[18] a) T. Grotjohann, I. Testa, M. Leutenegger, H. Bock, N. T.
Urban, F. Lavoie-Cardinal, K. I. Willig, C. Eggeling, S. Jakobs,
S. W. Hell, Nature 2011, 478, 204 – 208; b) T. Brakemann, A. C.
Stiel, G. Weber, M. Andersen, I. Testa, T. Grotjohann, M.
Leutenegger, U. Plessmann, H. Urlaub, C. Eggeling, M. C. Wahl,
S. W. Hell, S. Jakobs, Nat. Biotechnol. 2011, 29, 942 – 947.
Received: October 6, 2011
Revised: December 12, 2011
Published online: February 3, 2012
Keywords: dendrimers · fluorescence · FRET · photoswitch ·
.
zebrafish
[1] a) H. Kobayashi, M. Ogawa, R. Alford, P. L. Choyke, Y. Urano,
Chem. Rev. 2010, 110, 2620 – 2640; b) V. Ntziachristos, Annu.
Rev. Biomed. Eng. 2006, 8, 1 – 33.
2882 www.angewandte.org
ꢀ 2012 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Angew. Chem. Int. Ed. 2012, 51, 2878 –2882