Photoisomerization quantum yield of azobenzene-modified DNA depends on local sequence

Article abstract of DOI:10.1021/ja403249u

Photoswitch-modified DNA is being studied for applications including light-harvesting molecular motors, photocontrolled drug delivery, gene regulation, and optically mediated assembly of plasmonic metal nanoparticles in DNA-hybridization assays. We study

Full text of DOI:10.1021/ja403249u

Article
pubs.acs.org/JACS
Photoisomerization Quantum Yield of Azobenzene-Modified DNA
Depends on Local Sequence
Yunqi Yan,‡ Xin Wang,‡ Jennifer I. L. Chen,† and David S. Ginger*,‡
‡Department of Chemistry, University of Washington, Seattle, Washington 98195, United States
†Department of Chemistry, York University, Toronto, Ontario M3J 1P3, Canada
S
* Supporting Information
ABSTRACT: Photoswitch-modified DNA is being studied for
applications including light-harvesting molecular motors,
photocontrolled drug delivery, gene regulation, and optically
mediated assembly of plasmonic metal nanoparticles in DNA-
hybridization assays. We study the sequence and hybridization
dependence of the photoisomerization quantum yield of
azobenzene attached to DNA via the popular d-threoninol
linkage. Compared to free azobenzene we find that the quantum yield for photoisomerization from trans to cis form is decreased
3-fold (from 0.094 0.004 to 0.036 0.002) when the azobenzene is incorporated into ssDNA, and is further reduced 15-fold
(to 0.0056 0.0008) for azobenzene incorporated into dsDNA. In addition, we find that the quantum yield is sensitive to the
local sequence including both specific mismatches and the overall sequence-dependent melting temperature (Tm). These results
serve as design rules for efficient photoswitchable DNA sequences tailored for sensing, drug delivery, and energy-harvesting
applications, while also providing a foundation for understanding phenomena such as photonically controlled hybridization
stringency.
allowed the creation of DNA-hybridization assays capable of
differentiating single-base mismatches using a photonic
stringency wash.5
INTRODUCTION
■
Molecular photoswitches such as azobenzene have a long
history of application in fields ranging from material science1,2
to biology.3 Recently, the modification of DNA with these
molecules has also allowed the addition of stimulus-response
functionality to a wide range of DNA-based technologies.4−8
Notably, Asanuma and co-workers have pioneered the
development of an azobenzene-modified phosphoramidite,9,10
which allows virtually any DNA sequences amenable to solid-
phase synthesis to be readily functionalized with multiple
azobenzene photoswitches. In this approach, the azobenzenes
are incorporated by tethering them to additional sugar/
phosphate linkages along the DNA backbone via a d-threoninol
group (Figure 1). Despite some structural distortion of the
double helix resulting from the volume of the extra phosphate
and azobenzene moieties, the incorporation of a trans form
azobenzene stabilizes a DNA duplex by intercalation between
the neighboring bases.11 This stabilization raises the melting
temperature of the azobenzene-modified DNA above that of an
otherwise identical native sequence.12 Absorption of UV light
(320−370 nm) excites the S0 − S2 transition of the azobenzene
groups, promoting trans-to-cis photoisomerization. In the cis
form the azobenzenes destabilize the DNA duplex, significantly
lowering the melting temperature of the DNA. Blue light
(∼450 nm) converts the cis form back to trans, thereby
permitting reversible optical control of DNA hybridization.
These developments have enabled applications ranging from
photocontrolled gene regulation,13,14 and drug delivery,15 to
optically powered molecular motors.16,17 They have motivated
proposals for DNA-based optical energy harvesting,18 and even
Virtually all of these applications are sensitive to the quantum
yield for azobenzene photoisomerization: the total amount of
optical energy required to achieve DNA denaturation and the
efficiency of optically powered molecular motors will depend
on the quantum yield of photoisomerization. Surprisingly, while
the quantum yield for azobenzene photoisomerization is known
to depend on the local environment,19−21 there is very little
data on how the quantum yield for photoisomerization of
azobenzene is affected by incorporation into different DNA
sequences.
Here, we address this question by studying the quantum
yield of trans-to-cis photoisomerization of azobenzenes inserted
into DNA sequences via the popular Asanuma chemistry.10 We
show that the quantum yield for photoisomerization decreases
upon incorporation into single-stranded DNA (ssDNA), and
decreases further upon incorporation into double-stranded
DNA (dsDNA). Importantly, we show that the quantum yield
in dsDNA is sensitive to the melting temperature of the
sequence and very sensitive to the presence of local
mismatches.
RESULTS AND DISCUSSION
■
To assess how incorporation into various DNA sequences alters
the trans-to-cis photoisomerization of azobenzene we measured
Received: April 1, 2013
Published: May 10, 2013
© 2013 American Chemical Society
8382
dx.doi.org/10.1021/ja403249u | J. Am. Chem. Soc. 2013, 135, 8382−8387

Journal of the American Chemical Society
Article
Figure 1. (a) Absorption spectra of free trans-azobenzene in isooctane (green) and azobenzene-modified ssDNA (red) in buffer. The structure of
free azobenzene in the trans form is shown in the inset. (b) Structure of azobenzene-modified DNA with the d-threoninol linkage highlighted in red,
and adenine and cytosine labeled. (c) DNA nucleotide structures of the abasic site, thymine, and guanine used in the sequences and described in the
text.
the quantum yields in various modified DNA molecules by
quantifying the fraction of cis-azobenzene as a function of UV
(330 nm) irradiation time (Figure S2 in Supporting
Information [SI]) using UV−vis absorption spectroscopy.19,22
Figure 1a compares the absorption spectra of free trans-
azobenzene dissolved in isooctane and trans-azobenzene
incorporated in ssDNA in phosphate buffer. The free
azobenzene exhibits the typical π-to-π* absorption at 315 nm,
whereas the azobenzene incorporated into ssDNA has its
absorption band shifted to about 340 nm, as previously
reported.23
Figure 2a compares the fraction of cis-azobenzene as a
function of the integrated photokinetic factor19 (excitation time
converted to photon dose to account for the changing
absorption of the solution over time) at 28 °C for free
azobenzene, azobenzene attached via d-threoninol linkages10 to
a ssDNA sequence, and the same DNA sequence hybridized to
its (azobenzene-free) complement. For free azobenzene, the
fraction of cis-azobenzene at the photostationary state is 0.93,
which decreases to 0.51 for azobenzene incorporated into
ssDNA, and 0.15 for azobenzene incorporated into dsDNA. We
extracted the photoisomerization quantum yield from the
curves in Figure 2a by fitting the data to eq 1:19,22
y = (y − y∞) exp(−Ax) + y∞
(1)
0
where x is the integrated photokinetic factor19 defined via eq 2,
y is the fraction of cis-azobenzene (y0 and y∞ are the fractions of
cis-azobenzene before photoisomerization and at the photosta-
tionary state, respectively), and A is a prefactor related to the
trans-to-cis quantum yield (see SI for details). The integrated
photokinetic factor is given by:
Figure 2. (a) Representative plots of the measured fraction of cis-
azobenzene vs the integrated photokinetic factor (eq 2) used to obtain
quantum yield. Solid lines are fits of eq 1 to the data shown, and are
labeled with the average quantum yield values measured from at least
three separate experiments. Traces are for azobenzene in isooctane
(green circles), azobenzene incorporated in ssDNA (blue squares) and
dsDNA (red triangles). (b) Similar plots for azobenzene incorporated
in different dsDNA sequences including Seq-mm-abasic (yellow
circles), Seq-mm1T (blue squares), Seq-mm2 (green triangles), and
Seq-complement (red diamonds) (see Table 1 for sequences).
1 − 10−abs(t)
t
x(t) =
dt
∫
0
abs(t)
(2)
Figure 2a shows fits of eq 1 to the data as solid lines. From
these fits we obtained a quantum yield for free azobenzene of
0.094 0.004 when excited at 330 nm, which is consistent with
literature values.19,22
As can be seen from the decrease in cis-azobenzene fraction
in the photostationary state at a given illumination intensity,
incorporation into the ssDNA sequence decreases the photo-
isomerization quantum yield by a factor of ∼3 to 0.036 0.002.
Hybridization of the ssDNA to its complement (Seq-comple-
8383
dx.doi.org/10.1021/ja403249u | J. Am. Chem. Soc. 2013, 135, 8382−8387

Journal of the American Chemical Society
Article
Table 1. Quantum Yields and Melting Temperatures (Tm) of the Azobenzene-Modified DNA
ment) results in an even more dramatic decrease of the
quantum yield to 0.0056 0.0008. Since the DNA solutions do
not absorb at the UV excitation wavelength used here, we
attribute these differences entirely to the attachment of the
azobenzene to the DNA sequences.
While the structuressDNA or dsDNAhas a strong
influence on the azobenzene photoisomerization quantum
yield, we also find that the sequence of the dsDNA
particularly a mismatch in the sequence near the azobenzene
siteplays a role. To explore sequence effects, we used
azobenzene-modified dsDNA with one sequence having
azobenzene incorporated in the center, and the other sequence
bearing a single-base mismatch at varying distances from the
azobenzene site (see Table 1 for sequence data). Figure 2b
compares the fraction of cis-azobenzene as a function of the
integrated photokinetic factor (eq 2) for several different
dsDNA sequences at 28 °C. By fitting eq 1 to the data in Figure
2b, we extracted the photoisomerization quantum yields of the
azobenzene in these different dsDNA sequences. The quantum
yields for all studied sequences are summarized in Table 1.
Figure 2b and the accompanying fits show that the varying
fractions of cis-azobenzene achieved at the photostationary state
are due to sequence-dependent variations in the azobenzene
photoisomerization quantum yield.
Figure 3. Trans-to-cis isomerization quantum yield plotted as a
function of melting temperature (Tm) of azobenzene-modified
dsDNA. Quantum yield decreases as Tm rises. The quantum yield is
very sensitive to the single-base mismatch at the nearest neighbor
position of the azobenzene (red circles), but is less sensitive for
dsDNAs having the mismatched base multiple bases away from the
azobenzene (blue triangles).
yield tends to decrease with increasing Tm of the embedding
DNA sequence. We measure the highest azobenzene quantum
yield (0.020 0.001) for the dsDNA with the lowest Tm (46.7
°C) (Seq-mm-abasic). Likewise, we measure the lowest
photoisomerization quantum yield (0.0056 0.0008) for the
dsDNA with the highest Tm (60.0 °C) (Seq-complement). Our
data are consistent with previous work showing that photo-
isomerization can be used to modulate Tm,12,24 but importantly,
our work shows that the converse is also true: the Tm of the
dsDNA sequence can in turn affect trans-to-cis isomerization
efficiency.
Looking at these end points, one might be tempted to
conclude that Tm is the primary controlling factor in
determining the azobenzene quantum yield. However, closer
inspection of the data reveals that more complicated effects are
at work. For instance, Figure 3 shows that three sequences with
very similar Tm have dramatically different quantum yields for
azobenzene photoisomerization. Interestingly, all three of these
sequences have a single-base mismatch immediately next to the
position of the azobenzene (Seq-mm1T, Seq-mm1C, Seq-
mm1A). The photoisomerization quantum yield is less sensitive
to mismatches that are two or more bases away from the
Of the sequences used in Figure 2b, the lowest quantum
yield of 0.0056
dsDNA with no mismatches (Seq-complement). The quantum
yield increases to 0.011
0.0008 comes from azobenzene-modified
0.001 for dsDNA having an A•C
mismatch two bases away from the azobenzene position (Seq-
mm2). In contrast, a C•T mismatch immediately next to the
incorporated azobenzene (Seq-mm1T) increases the quantum
yield by an additional 45% to 0.016
measure the largest quantum yield (0.020
0.001. Finally, we
0.001) using
modified dsDNA having an abasic site (with no purine or
pyrimidine, Figure 1c, Seq-mm-abasic) as the nearest neighbor
to the azobenzene in the unmodified sequence.
To examine the effects of dsDNA stability and DNA
sequence on quantum yield, Figure 3 plots the quantum yield
as a function of the measured melting temperature (Tm)
(Figure S2 in SI) of the modified dsDNA sequences listed in
Table 1. Broadly, the data show that the azobenzene quantum
8384
dx.doi.org/10.1021/ja403249u | J. Am. Chem. Soc. 2013, 135, 8382−8387

Journal of the American Chemical Society
Article
position of the azobenzene (e.g., Seq-mm2, Seq-mm3, Seq-
mm4).
oxidation potentials of the bases. On the other hand, it is
known that poly(dT) is more flexible than poly(dA).29 This
result also fits well with the free volume hypothesis: azobenzene
in more structurally flexible poly(dT) has a higher quantum
yield than azobenzene in more rigid poly(dA). Likewise, if
energy transfer were dominant, one might expect a correlation
between the positions of the UV−vis spectra of the nucleotides
and the measured quantum yields; however, we were unable to
observe such a correlation (Figure S3 in SI). Thus, we propose
that free volume is likely to be more important, while
acknowledging that it will be difficult to completely separate
electronic and structural control over variation in the
azobenzene photoisomerization quantum yield because an
increase in local free volume around the azobenzene should
also tend to weaken intermolecular electronic interactions.
Finally, we can use the observed variation of azobenzene
quantum yield with the DNA sequence to explain the
properties of azobenzene-modified DNA that facilitate its use
in novel DNA-hybridization assays.5 Previously we have shown
that using only optical inputs, DNA sequences containing
single-base mismatches can be resolved in hybridization
experiments involving gold nanoparticles that are heavily
functionalized with azobenzene-modified DNA.5 The resulting
photonic hybridization stringency wash works because the
denaturation of azobenzene-modified DNA strands occurs at
lower photon doses for sequences with less complementarity.
The results presented herein provide a fundamental mecha-
nistic understanding of this process: partially mismatched
sequences denature at lower photon doses because the
azobenzenes in those sequences photoisomerize more
efficiently than azobenzenes in perfectly complementary
sequences.
We propose that the variations we observe in photo-
isomerization quantum yield when the azobenzene is
incorporated into DNA are largely due to differences in the
local free volume available to the azobenzene in the different
sequences. Generally, the azobenzene quantum yield is known
to depend on the free volume surrounding the azobenzene
site,20,21 and this explanation would be consistent with our
results that the quantum yield decreases on going from an
azobenzene free in solution to that being incorporated in
ssDNA to being incorporated in dsDNA. Furthermore, this
hypothesis could account for the large differences we observe
between single-base mismatches that are next to the
azobenzene site, and those that are further removed: distortions
in the double helix are known to recover over a decay length of
only a few bases.25
However, it is also possible that electronic interactions,
including both energy and charge transfer between the
azobenzene and the nucleic acid bases11,26 are modulating the
quantum yield by changing the energy relaxation pathways
available to the azobenzene on an ultrafast time scale.27 In this
case, one would explain the sequence-dependent differences in
quantum yield as arising from different electronic interactions
between the azobenzene and the different bases. To test this
alternative hypothesis, Figure 4 plots the photoisomerization
To test this hypothesis, we measured the trans-to-cis
isomerization quantum yield of azobenzenes incorporated in
dsDNAs in a classic three-strand capture assay using a linker
strand as shown in Figure 5a. The assay consists of the capture
DNA modified with multiple azobenzenes (Sequence-1), the
unmodified probe DNA (Sequence-2), and the unmodified
target/linker (Sequence-3) that cross-hybridizes with the
capture and probe sequences. We varied the target sequence
by introducing a single base mismatch in the center where it
will form a mismatched base pair next to the azobenzene (see
sequences in Figure 5a). We measure the lowest quantum yield
for the perfectly complementary sequence (Seq-perfect), and
observe an increase in quantum yield from Seq-mmA, to Seq-
mmC and to Seq-mmTall having the single-base mismatch
next to one of the azobenzenes. In order to compare the trend
in quantum yield with that of optical hybridization stringency,
we functionalized gold nanoparticles with the same capture and
probe DNAs and measured the disaggregation rate of
nanoparticle conjugates cross-linked by the same target DNA,
which we have previously shown can exhibit photon-dose-
controlled hybridization stringency.5 We quantify the disag-
gregation rate by monitoring the localized surface plasmon
resonance (LSPR) peak shift of nanoparticle conjugates, which
is expected to blue-shift as conjugates undergo photoinduced
dissociation. Indeed, the inset in Figure 5b shows that the
photon-dose dependence of the nanoparticle disaggregation
process follows the exact same sequence-dependent order as
the measured azobenzene quantum yields, providing strong
evidence that the two trends are linked.
Figure 4. Quantum yield (bars) for the azobenzene incorporated into
the 18-base ssDNA of poly(dC), poly(dA), and poly(dT). The
oxidation potential for the individual DNA nucleotides (C, A, T) vs
saturated calomel electrode (SCE) is also plotted (dots). There does
not appear to be a correlation between the oxidation potential for the
nucleotides and the quantum yield of the azobenzene incorporated
into the ssDNA sequences as might be expected if charge transfer were
the dominant factor governing azobenzene quantum yield differences
between the sequences.
quantum yield (bars) for the azobenzene incorporated in the
middle of 18-base ssDNAs comprising polydeoxyadenosine
(poly(dA)), polydeoxycytidine (poly(dC)), and polydeoxythy-
midine (poly(dT)). In comparison, Figure 4 also plots the
oxidation potential of DNA nucleotides (dots) in buffer (pH =
7) as measured by Fukuzumi et al.28 We note that, if electronic
coupling were dominant, one might expect to see some
correlation between oxidation potential of the neighboring
bases and the quantum yield. However, Figure 4 shows that the
quantum yield of azobenzene contained in poly(dA), poly(dC),
and poly(dT) sequences appears uncorrelated with the
8385
dx.doi.org/10.1021/ja403249u | J. Am. Chem. Soc. 2013, 135, 8382−8387

Journal of the American Chemical Society
Article
Figure 5. Trans-to-cis isomerization quantum yield explains photon-dose-controlled DNA hybridization stringency wash. (a) DNA sequences used
to incorporate azobenzenes and also to cross-link gold nanoparticles. (b) The quantum yield of the azobenzene is shown against each embedding
dsDNA. (Inset) Plot of the gold nanoparticle localized surface plasmon resonance (LSPR) peak position as a function of the photon dose. The
azobenzene photoisomerization quantum yield increases from Seq-perfect, to Seq-mmT, to Seq-mmA and to Seq-mmC. Nanoparticle conjugates
linked by these dsDNA show the same increasing order of photoinduced disaggregation rate.
Sensor Electronic Technology, Inc.), a homemade aluminum stage, a
quartz cuvette with 1-cm optical path length, a stir plate, and a
temperature controller. The temperature of the DNA solution as a
CONCLUSION
■
We have shown that the trans-to-cis photoisomerization
quantum yield for azobenzene decreases upon incorporation
into DNA and is sensitive to both the local DNA sequence and
DNA hybridization state. In general, the photoisomerization
quantum yield tends to increase as the Tm of the attached
dsDNA decreases. However, the biggest variations in quantum
yield are associated with dsDNAs bearing a single-base
mismatch immediately next to the azobenzene site. We
propose that these variations arise due to the structural
fluctuations caused by the adjacent mismatched base inducing
an increase in the local free volume. Not only do these results
provide a mechanism to explain optically controlled DNA
hybridization stringency, but we believe they will also prove
useful for designing systems to more effectively photomodulate
gene regulation and drug delivery or transduce optical energy
into work using DNA nanomachines.
function of temperature controller set point (measured in the
aluminum stage) was calibrated using a thermometer. The temperature
of the DNA solution was kept at 28 °C for all quantum yield
measurements using the calibrated temperature controller. The UV
LEDs were warmed up for 1 h, and then the illumination intensity was
monitored using a silicon photodiode positioned at the other end of
the aluminum stage. The UV intensity was typically 0.37−0.42 mW/
cm2. Azobenzene-modified DNA solutions were added to the quartz
cuvette and were thermally equilibrated at 28 °C for 1 h in the dark
before UV irradiation. During UV irradiation, UV−vis absorption
spectra were recorded by an Agilent 8453 UV−vis spectrometer every
1 min for 15 min and then every 5 min for the remaining 45 min. The
DNA solution was stirred during the measurement. We verified that
the low intensity white light source of the spectrometer had no
significant effect on the photoisomerization process (Figure S4 in SI).
Melting Temperature Measurements. An Agilent 8453 UV−vis
spectrometer operating in the thermal denaturation mode was used to
measure the melting temperature. The temperature ramp started at 5
°C and ended at 95 °C with 2 °C step intervals and a 5-min hold time.
Melting temperature was determined as the temperature at which the
first derivative of the absorbance vs temperature plot was maximum.
See selected data in Figure S2 in SI.
EXPERIMENTAL METHODS
■
Materials. Unmodified DNAs and azobenzene-modified DNAs
were purchased from Integrated DNA Technology (IDT Inc., IA). All
sequences used are shown in Table 1 and Figure 5a. Water was
deionized to 18.0 MΩ with the Millipore filtration system.
Preparation of dsDNA Solution. Aliquots of lyophilized DNA
were dissolved in water and a desired amount of DNA aqueous
solution was then brought to 0.01 M phosphate buffer (pH = 6.5), 0.1
M NaCl and 0.02% sodium azide. Equal moles of complementary
DNAs were mixed at room temperature and annealed at 95 °C for 5
min. The concentration8 of azobenzene modified dsDNA for quantum
yield tests was prepared to be 10 μM; and the concentration for Tm
test was 2 μM. The dsDNA solution was kept at 4 °C before use.
Preparation of the dsDNA That Has Three-Sequence
Structure. The sequences for the three-strand capture assay are
shown in Figure 5a. Sequence-1 and Sequence-2 are both
complementary to part of Sequence-3. The annealing process
consisted of two steps. First, Sequence-2 and Sequence-3 were
combined and annealed at 95 °C for 5 min, followed by gradual
cooling to 55 °C and holding at 55 °C for 30 min. Then, Sequence-1
was added to the solution and annealed for an additional 10 min at 55
°C. The temperature of the DNA solution was lowered to room
temperature and kept at 4 °C before use.
ASSOCIATED CONTENT
* Supporting Information
■
S
Plots of fraction of cis-azobenzene as a function of UV
excitation time, melting curve of dsDNA, UV absorption
coefficient spectra of DNA nucleotides, control experimental
results showing UV−vis spectrometer having no interference
on photoisomerization. This material is available free of charge
via the Internet at http://pubs.acs.org.
AUTHOR INFORMATION
Corresponding Author
ginger@chem.washington.edu
■
Notes
The authors declare no competing financial interest.
ACKNOWLEDGMENTS
This paper is based on work supported by the AFOSR FA9550-
10-1-0474.
■
Quantum Yield Measurements. The UV irradiation setup for
quantum yield measurement consisted of an LED light source
centered at 330 nm with fwhm less than 10 nm (UVTOP325HS
8386
dx.doi.org/10.1021/ja403249u | J. Am. Chem. Soc. 2013, 135, 8382−8387

Journal of the American Chemical Society
Article
REFERENCES
■
(1) Klajn, R. Pure Appl. Chem. 2010, 82, 2247−2279.
(2) Mahimwalla, Z.; Yager, K. G.; Mamiya, J-i.; Shishido, A.; Priimagi,
A.; Barrett, C. J. Polym. Bull. 2012, 69, 967−1006.
(3) Beharry, A. A.; Woolley, G. A. Chem. Soc. Rev. 2011, 40, 4422−
4437.
(4) Tanaka, F.; Mochizuki, T.; Liang, X.; Asanuma, H.; Tanaka, S.;
Suzuki, K.; Kitamura, S-i.; Nishikawa, A.; Ui-Tei, K.; Hagiya, M. Nano
Lett. 2010, 10, 3560−3565.
(5) Yan, Y. Q.; Chen, J. I. L.; Ginger, D. S. Nano Lett. 2012, 12,
2530−2536.
(6) Peng, L.; You, M.; Yuan, Q.; Wu, C.; Han, D.; Chen, Y.; Zhong,
Z.; Xue, J.; Tan, W. J. Am. Chem. Soc. 2012, 134, 12302−12307.
(7) Yang, Y.; Endo, M.; Hidaka, K.; Sugiyama, H. J. Am. Chem. Soc.
2012, 134, 20645−20653.
(8) Nishioka, H.; Liang, X.; Kato, T.; Asanuma, H. Angew. Chem., Int.
Ed. 2012, 51, 1165−1168.
(9) Asanuma, H.; Matsunaga, D.; Liu, M.; Liang, X.; Jhao, J. Nucleic
Acids Res. Suppl. 2003, 117−118.
(10) Asanuma, H.; Liang, X.; Nishioka, H.; Matsunaga, D.; Liu, M.;
Komiyama, M. Nat. Protoc. 2007, 2, 203−212.
(11) Liang, X. G.; Asanuma, H.; Kashida, H.; Takasu, A.; Sakamoto,
T.; Kawai, G.; Komiyama, M. J. Am. Chem. Soc. 2003, 125, 16408−
16415.
(12) Liang, X.; Mochizuki, T.; Asanuma, H. Small 2009, 5, 1761−
1768.
(13) Yamazawa, A.; Liang, X. G.; Asanuma, H.; Komiyama, M.
Angew. Chem., Int. Ed. 2000, 39, 2356−2357.
(14) Matsunaga, D.; Asanuma, H.; Komiyama, M. J. Am. Chem. Soc.
2004, 126, 11452−11453.
(15) Kang, H.; Liu, H.; Zhang, X.; Yan, J.; Zhu, Z.; Peng, L.; Yang,
H.; Kim, Y.; Tan, W. Langmuir 2011, 27, 399−408.
(16) Liang, X. G.; Tekenaka, N.; Nishioka, H.; Asanuma, H. Nucleic
Acids Symp. Ser. 2008, 52, 697−698.
(17) Kang, H.; Liu, H.; Phillips, J. A.; Cao, Z.; Kim, Y.; Chen, Y.;
Yang, Z.; Li, J.; Tan, W. Nano Lett. 2009, 9, 2690−2696.
(18) McCullagh, M.; Franco, I.; Ratner, M. A.; Schatz, G. C. J. Am.
Chem. Soc. 2011, 133, 3452−3459.
(19) Rau, H. In Photoreactive Organic Thin Films, 1st ed.; Sekkat, Z.,
Kroll, W., Eds.; Elsevier: Amsterdam, 2002; pp 3−47.
(20) Mita, I.; Horie, K.; Hirao, K. Macromolecules 1989, 22, 558−563.
(21) Victor, J. G.; Torkelson, J. M. Macromolecules 1987, 20, 2241−
2250.
(22) Zimmerman, G.; Chow, L. Y.; Paik, U. J. J. Am. Chem. Soc. 1958,
80, 3528−3531.
(23) Asanuma, H.; Ito, T.; Komiyama, M. Tetrahedron Lett. 1998, 39,
9015−9018.
(24) Kamei, T.; Akiyama, H.; Morii, H.; Tamaoki, N.; Uyeda, T. Q.
P. Nucleosides, Nucleotides Nucleic Acids 2011, 28, 12−28.
(25) McCullagh, M.; Franco, I.; Ratner, M. A.; Schatz, G. C. J. Phys.
Chem. Lett. 2012, 3, 689−693.
(26) Lewis, F. D.; Wu, Y. S.; Zhang, L. G.; Zuo, X. B.; Hayes, R. T.;
Wasielewski, M. R. J. Am. Chem. Soc. 2004, 126, 8206−8215.
(27) Chen, T.; Igarashi, K.; Nakagawa, N.; Yamane, K.; Fujii, T.;
Asanuma, H.; Yamashita, M. J. Photochem. Photobiol., A 2011, 223,
119−123.
(28) Fukuzumi, S.; Miyao, H.; Ohkubo, K.; Suenobu, T. J. Phys.
Chem. A 2005, 109, 3285−3294.
(29) Goddard, N. L.; Bonnet, G.; Krichevsky, O.; Libchaber, A. Phys.
Rev. Lett. 2000, 85, 2400−2403.
8387
dx.doi.org/10.1021/ja403249u | J. Am. Chem. Soc. 2013, 135, 8382−8387