An efficient fluorescence resonance energy transfer (FRET) between Pyrene and Perylene assembled in a DNA duplex and its potential for discriminating single-base changes

Article abstract of DOI:10.1002/chem.200902078

To increase the apparent Stokes' shift of perylene, pyrene (donor) and perylene (acceptor) were assembled in a DNA duplex to achieve the efficient fluorescence resonance energy transfer (FRET) from pyrene to perylene. Multiple donors were introduced in the vicinity of acceptors through d-threoninol and natural base pairs were inserted between the dyes. Accordingly, donors and acceptors could be accumulated inside the DNA without forming an undesired excimer/ exciplex. When two pyrene moieties were located in proximity to one perylene with one base pair inserted between them, efficient FRET occurred within the duplex. Thus, strong emission at 460 nm was observed from perylene when excited at 345 nm at which pyrene has its absorption. The apparent Stokes' shift became as large as 115 nm with a high apparent FRET efficiency (Fφ1). However, the introduction of more than two pyrenes did not enhance the fluorescence intensity of perylene, due to the short Fcrster radius (R₀) of the donor pyrene. Next, this FRET system was used to enlarge the Stokes' shift of the DNA probe, which can discriminate a one-base deletion mutant from wild type with a model system by incorporation of multiple donors into DNA. Two perylene moieties were tethered to the DNA on both sides of the intervening base, and two pyrenes were further inserted in the vicinity of the perylenes as an antenna. Hybridization of this FRET probe with a fully matched DNA allowed monomer emission of perylene when the pyrenes were excited. In contrast, excimer emission was generated by hybridization with a one-base deletion mutant. Thus, the apparent Stokes' shift was enhanced without loss of efficiency in the detection of the deletion mutant.

Full text of DOI:10.1002/chem.200902078

FULL PAPER
DOI: 10.1002/chem.200902078
An Efficient Fluorescence Resonance Energy Transfer (FRET) between
Pyrene and Perylene Assembled in a DNA Duplex and Its Potential for
Discriminating Single-Base Changes
Hiromu Kashida,[a] Tomohiko Takatsu,[a] Koji Sekiguchi,[a] and Hiroyuki Asanuma*[a, b]
Abstract: To increase the apparent
Stokesꢀ shift of perylene, pyrene
(donor) and perylene (acceptor) were
assembled in a DNA duplex to achieve
the efficient fluorescence resonance
energy transfer (FRET) from pyrene
to perylene. Multiple donors were in-
troduced in the vicinity of acceptors
through d-threoninol and natural base
pairs were inserted between the dyes.
Accordingly, donors and acceptors
could be accumulated inside the DNA
without forming an undesired excimer/
exciplex. When two pyrene moieties
were located in proximity to one pery-
lene with one base pair inserted be-
tween them, efficient FRET occurred
within the duplex. Thus, strong emis-
sion at 460 nm was observed from per-
ylene when excited at 345 nm at which
pyrene has its absorption. The apparent
Stokesꢀ shift became as large as 115 nm
with a high apparent FRET efficiency
(F>1). However, the introduction of
more than two pyrenes did not en-
hance the fluorescence intensity of per-
ylene, due to the short Fçrster radius
(R0) of the donor pyrene. Next, this
FRET system was used to enlarge the
Stokesꢀ shift of the DNA probe, which
can discriminate a one-base deletion
mutant from wild type with a model
system by incorporation of multiple
donors into DNA. Two perylene moiet-
ies were tethered to the DNA on both
sides of the intervening base, and two
pyrenes were further inserted in the vi-
cinity of the perylenes as an antenna.
Hybridization of this FRET probe with
a fully matched DNA allowed mono-
mer emission of perylene when the pyr-
enes were excited. In contrast, excimer
emission was generated by hybridiza-
tion with a one-base deletion mutant.
Thus, the apparent Stokesꢀ shift was
enhanced without loss of efficiency in
the detection of the deletion mutant.
Keywords: DNA · energy transfer ·
fluorescent probes
pyrene
· perylene ·
Introduction
fects sensitivity because the scattered light of excitation in-
hibits the detection of emission. Although conventional flu-
orescent dyes, such as Cy3 and Cy5, have a high quantum
yield and can be applied to the labeling of ODNs, their
Stokesꢀ shifts are only 15 and 17 nm, respectively.[2,3] Since
such a small Stokesꢀ shift lowers the sensitivity, efforts have
been made to expand the difference between the excitation
and emission wavelengths.[4] One promising strategy for this
purpose is the utilization of fluorescence resonance energy
transfer (FRET).[5] Several studies on the application of
FRET to DNA duplexes have reported an increase in the
“apparent” Stokesꢀ shift.[6–12] For example, pyrene and pery-
lene, which are known as a FRET pair,[13] were introduced
into an ODN to increase the apparent Stokesꢀ shift. The
emission band of pyrene appeared at 370–430 nm, whereas
the absorption band of perylene appeared at 390–470 nm.
Previously, these donor–acceptor pairs were tethered at the
termini of ODNs,[14] or at the 2’-position of a locked nucleic
acid (LNA).[15] Thus, pyrene and perylene were located at
the end of the ODN, or in the minor groove, so that the flu-
Fluorescent dyes are commonly used for the detection of
oligodeoxyribonucleotides (ODN) due to their high sensitiv-
ity.[1] While the quantum yield of the fluorescent dye is the
dominant factor that determines its sensitivity, the Stokesꢀ
shift, that is, the difference between the band maxima of ab-
sorption and emission of identical dyes, also significantly af-
[a] Dr. H. Kashida, T. Takatsu, K. Sekiguchi, Prof. Dr. H. Asanuma
Graduate School of Engineering
Nagoya University, Furo-cho, Chikusa-ku
Nagoya 464-8603 (Japan)
Fax : (+81)52-789-2488
E-mail: asanuma@mol.nagoya-u.ac.jp
[b] Prof. Dr. H. Asanuma
CREST (Japan) Science and Technology Agency
4-1-8 Honcho, Kawaguchi
Saitama 332-0012 (Japan)
Supporting information for this article is available on the WWW
under http://dx.doi.org/10.1002/chem.200902078.
Chem. Eur. J. 2010, 16, 2479 – 2486
ꢁ 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
2479

orophores were not rigidly fixed but could move to some
extent. However, it might be difficult to introduce multiple
donors as an antenna in close proximity to an acceptor dye
by using these molecular designs, because fluorescence will
be quenched upon dimer formation of donors or following
donor–acceptor dimerization.
Here, we describe another FRET system in which donors
are accumulated in close proximity to acceptors within the
duplex. In this study, pyrene and perylene were introduced
into DNA via a d-threoninol linker to facilitate the interca-
lation of this FRET pair between base pairs:[16] pyrene and
perylene are intercalated inside the duplex and intervening
base pairs separate these chromophores as depicted in
Scheme 1. Consequently, fluorophores can be accumulated
quences of the ODNs used in this study. Both pyrenes (P in
Scheme 1) and perylenes (E) were tethered on d-threoninol
in the middle of the ODN and one to three natural nucleo-
tides were inserted between the pyrenes and perylenes to
avoid undesired excimer or exciplex formation. The comple-
mentary strands used were C1 (for ODN1–ODN4) and C2
(for ODN5–ODN11). In these strands, pyrene and perylene
residues were inserted into the ODN as bulges (i.e. not in
place of a natural nucleotide). By this sequence design, nat-
ural base pairs were kept intact in the duplex state. To eval-
uate the effect of distance between donors and acceptors,
ODNs, which have one and three bases between pyrene and
perylene, were synthesized (ODN2 and ODN3). In addition,
ODN with four pyrenes (ODN4) was also synthesized to in-
vestigate the effect of the
number
of
donors.
One
acceptor and two or four
donors were introduced into
the ODNs (ODN2–ODN4) for
the calculation of FRET effi-
ciencies. In addition, ODNs
without a donor (ODN1) or ac-
ceptor (ODN2d–ODN4d) were
also synthesized as controls. As
shown on the right-hand side of
Scheme 1, probe-type sequen-
ces, which tethered two pery-
lenes, were also prepared
Scheme 1. Sequences of the ODNs synthesized in this study.
(ODN6–ODN11). To calculate
in close proximity without forming undesired excimers/exci-
plexes or quenching their fluorescence. Molecular cluster
FRET efficiencies, ODN7d, ODN8d, and ODN10d were
synthesized as controls for ODN7, ODN8, and ODN10, re-
spectively. Most of the spectroscopic measurements were
carried out at 08C in which full duplex hybridization was as-
certained (unless otherwise noted).
motifs reported by us and other research groups have diffi-
[17–22]
culty in suppressing these interactions.
The design pre-
sented herein makes it possible to introduce more than one
donor (pyrene) near the acceptor (perylene), which facili-
tates energy transfer from pyrene to perylene. In the present
study, we investigated the effect of the position and number
of donors on FRET efficiencies. We then applied this FRET
system to enlarge the Stokesꢀ shift of the fluorescent probe
for potential application to the one-base deletion poly-
morphisms. Recently, much attention has been paid to non-
single-nucleotide polymorphism (SNP) variations, such as
small deletions, due to their frequent occurrence.[23] We
have developed ODN probes detecting base-deletion poly-
morphisms.[24,25] Here, we demonstrate that the Stokesꢀ shift
of the fluorescent probe can be extended by locating donors
in proximity to the acceptor. This strategy has a potential to
be used for other fluorescent probes tethering intercala-
tors.[26–30]
Fluorescent behavior of ODNs tethering one perylene: First,
effect of the distance between the donor pyrene and accept-
or perylene on the FRET efficiency was investigated by
changing the number of base pairs between the pyrene and
perylene. Figure 1 shows the fluorescence emission spectra
of ODN1–ODN4 hybridized with their complementary
strand (C1) excited at 345 nm, a wavelength at which pyrene
has its absorption.[31] The corresponding UV/Vis spectra are
depicted in Figure S1 in the Supporting Information. The
emission spectra of ODN1/C1, which has single perylene (E
residue) and no pyrene, also showed a small emission at 460
and 491 nm because perylene still has very weak absorption
at 345 nm (see Figure S1 in Supporting information). In con-
trast, other sequences exhibited much stronger emission at
450–550 nm than did ODN1/C1. In particular, the emission
intensity of ODN2/C1, which has only one base pair be-
tween pyrene and perylene, was much higher than that of
ODN1/C1, which demonstrates efficient energy transfer be-
tween pyrene and perylene. Similarly, insertion of three nat-
ural base pairs between perylene and pyrene (ODN3/C1)
also showed a higher emission at 461 and 491 nm than
ODN1/C1, although its intensity was lower than that of
Results
Design of modified ODNs by using pyrenes and perylenes
as donors and acceptors, respectively, and their fluorescent
behavior: All the modified ODNs were synthesized by the
standard phosphoramidite method. Scheme 1 shows the se-
2480
www.chemeurj.org
ꢁ 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2010, 16, 2479 – 2486

An Efficient Fluorescence Resonance Energy Transfer between Pyrene and Perylene
FULL PAPER
of ODN2/C1 was greater than 1 (1.18), which may have
been caused by a structural disturbance (see the Discussion
section). As expected from the fluorescence emission spec-
tra, FT of ODN3/C1 was lower than that of ODN2/C1.
Moreover, ODN4/C1, containing twice as many pyrenes as
ODN2/C1, showed a similar FT to that of ODN3/C1. Thus,
the FT of ODN4/C1 was lower than that of ODN2/C1, al-
though the emission intensity was almost the same.
The melting temperatures (Tms) of these modified ODNs
are listed in Table 1. The Tm of ODN3/C1 was 54.68C, which
was 9.68C higher than that of ODN1/C1, which indicates
that the introduced pyrene stabilized the duplex by a stack-
ing interaction. However, the Tm of ODN2/C1 was about
208C lower than that of ODN3/C1. Similarly, the Tm of
ODN4/C1 was about 108C lower than that of ODN2/C1.
These low Tm values are attributed to distortion of the
duplex by alternately introduced chromophores and natural
nucleobases. Despite this destabilization, pyrenes were effi-
ciently separated by the intervening base pairs because no
excimer or exciplex emission was observed (see Figure 1).[32]
Figure 1. Fluorescence emission spectra of ODN1/C1–ODN4/C1 at 08C.
Spectra at 370–420 nm are magnified in the inset. The excitation wave-
length was 345 nm. The solution conditions are as follows: [ODNn]=
1.0 mm, [C1]=1.2 mm, [NaCl]=100 mm, pH 7.0 (10 mm phosphate buffer).
ODN2/C1 (compare the solid blue line with the green one
in Figure 1). Concurrent with these effects, the emission
peaks of pyrene at 370–430 nm were increased by the inser-
tion of base pairs (see the inset in Figure 1). Accordingly,
the FRET efficiency of ODN2/C1 is higher than that of
ODN3/C1.
Next, FRET efficiency of the accumulated donor pyrenes
was evaluated by comparing ODN2/C1 with ODN4/C1,
which involves four donor pyrenes. An increase in the
number of donors did not enhance the emission of the ac-
ceptor and the emission intensities of ODN2/C1 and ODN4/
C1 were almost identical. Note that no excimer emission of
pyrene was observed in ODN4/C1 although five chromo-
phores (one perylene and four pyrenes) were introduced
and thus absorption at 345 nm was increased (see Figure S1
in the Supporting Information). This result demonstrates
that each donor and acceptor is separated by an intervening
base pair according to the design. We also calculated the
FRET efficiencies (FT) of these duplexes by using ODN2d/
C1–ODN4d/C1 as controls (see Table 1). The apparent FT
Spectroscopic behavior of probe-type ODNs tethering two
perylenes: We also investigated the spectroscopic behavior
of probe-type sequences (ODN5–ODN11, see the next sec-
tion for details), which tether two perylenes in the middle of
the sequence to evaluate the FRET efficiency of accumulat-
ed donor pyrenes near the perylenes, and effect of the dis-
tance between the fluorophores.[25] Figure 2 shows the fluo-
Table 1. Spectroscopic properties of modified ODNs involving an
acceptor.
Figure 2. Fluorescence emission spectra of ODN5/C2–ODN11/C2 at 08C.
The excitation wavelength was 345 nm. Spectra at 370–420 nm are magni-
fied in the inset. Solution conditions are as follows: [ODNn]=1.0 mm,
[C2]=1.2 mm, [NaCl]=100 mm, pH 7.0 (10 mm phosphate buffer).
[b,d]
Sequence
Tm [8C][a]
Relative
FT
intensity[b,c]
ODN1/C1
ODN2/C1
ODN3/C1
ODN4/C1
45.0
34.8
54.6
24.9
1.0
15.9
7.6
–
>1[e]
0.53
0.61
rescence emission spectra of ODN5/C2–ODN11/C2 excited
at 345 nm. The emission intensity of ODN6/C2, which has
one pyrene, was almost double that of ODN5/C2. Moreover,
ODN7/C2, which has two pyrenes showed much more in-
tense peaks than ODN5/C2. As the number of base pairs
between pyrene and perylene increased, (ODN8/C2 and
ODN9/C2), the emission intensity at 450–550 nm decreased.
In particular, the emission of an ODN9/C2 duplex with five
base pairs between pyrene and perylene was as weak as that
of ODN5/C2, which has no pyrenes (compare the red solid
16.1
[a] [ODNn]=5.0 mm, [C1]=5.0 mm, [NaCl]=100 mm, pH 7.0 (10 mm
phosphate buffer). [b] [ODNn]=1.0 mm, [C1]=1.2 mm, [NaCl]=100 mm,
pH 7.0 (10 mm phosphate buffer), 08C. Excitation wavelength: 345 nm.
[c] Relative intensity at 460 nm relative to that of ODN1/C1. [d] Effi-
ciencies of FRET were calculated according to the method described in
the Experimental Section. [e] Apparent FRET efficiency was calculated
as 1.18.
Chem. Eur. J. 2010, 16, 2479 – 2486
ꢁ 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
2481

H. Asanuma et al.
line with the dotted blue line in Figure 2); this indicates that
the antenna effect of the pyrenes was almost nil. On the
other hand, emission at 370–430 nm showed a slightly com-
plicated behavior; the emission intensity of ODN8/C2 was
higher than ODN9/C2, although that of ODN7/C2 was
lower than that of ODN9/C2. Electron transfer between
pyrene and natural bases might affect emission intensity of
pyrene in these duplexes. Incorporation of four pyrenes
(ODN10/C2 and ODN11/C2) did not enhance the emission
of perylene (compare the green solid line with the green
dotted and purple solid lines in Figure 2). As the distance
between donors and acceptors decreased, more intense
emission of perylene was observed, which shows that energy
transfer from pyrene to perylene occurred.[33] The FRET
efficiencies of ODN7/C2, ODN8/C2 and ODN10/C2 were
calculated by using ODN7d/C2, ODN8d/C2 and ODN10d/
C2, respectively, as listed in Table 2. The FT of ODN7/C2
Figure 3. CD spectra of ODN7/C2 at various temperatures. Solution con-
ditions are as follows: [ODN7]=1.0 mm, [C2]=1.2 mm, [NaCl]=100 mm,
pH 7.0 (10 mm phosphate buffer).
Table 3. Effect of the donor position on polymorphism-discrimination
ability.
[a]
Sequence
I550/I461
C1[c]
Tm [8C][b]
C1[c]
Table 2. Spectroscopic properties of modified ODNs involving two ac-
ceptors.
C2[d]
C2[d]
49.9
45.3
36.3
46.4
54.6
25.7
38.8
[a,c]
Sequence
Relative
FT
ODN5
ODN6
ODN7
ODN8
ODN9
ODN10
ODN11
0.84[e]
0.39[f]
0.61[f]
0.86[f]
0.21[f]
0.47[f]
0.94[f]
0.05[e]
0.06[f]
0.07[f]
0.05[f]
0.05[f]
0.19[f]
0.10[f]
46.7
40.3
33.0
42.7
50.9
21.0
33.9
intensity[a,b]
ODN5/C2
ODN6/C2
ODN7/C2
ODN8/C2
ODN9/C2
ODN10/C2
ODN11/C2
1.0
2.2
3.7
3.3
1.0
3.6
2.9
–
nd[d]
0.36
0.27
nd[d]
0.14
nd[d]
[a] [ODNn]=1.0 mm, [C1]=[C2]=1.2 mm, [NaCl]=100 mm, pH 7.0, 208C
(10 mm phosphate buffer). [b] [ODNn]=5.0 mm, [C1]=[C2]=5.0 mm,
[NaCl]=100 mm, pH 7.0 (10 mm phosphate buffer). [c] One-base deletion
mutant. [d] Full match. [e] Excited at 425 nm. [f] Excited at 345 nm.
[a] [ODNn]=1.0 mm, [C2]=1.2 mm, [NaCl]=100 mm, pH 7.0 (10 mm
phosphate buffer), 08C. Excitation wavelength: 345 nm. [b] Relative in-
tensity at 460 nm relative to that of ODN5/C2. [c] FRET efficiencies
were calculated according to the method described in the Experimental
Section. [d] Not determined.
characteristic of B-form DNA, was observed at around
260 nm, although the couplet was not symmetrical, probably
due to the CD induced by perylene at this region. Although
ODN7/C2 involves two pyrenes, no CD was observed at
around 350 nm at which pyrene has its absorption.[34]
was 0.36, which was much lower than that of ODN2/C1. In-
sertion of one more base pair between pyrene and perylene
decreased FT (ODN8/C2). ODN10/C2, with four pyrenes
exhibited a much lower FT than ODN7/C2 with two pyr-
enes, although their emission intensities were almost the
same.
To investigate the orientation and position of chromo-
phores, circular dichroism (CD) spectra of ODN7/C2 at var-
ious temperatures were measured (Figure 3). At a tempera-
ture higher than the Tm, no obvious CD was observed above
350 nm (The Tm values are listed in Table 3). On the other
hand, as the temperature decreased below 208C, an induced
CD was observed at 400–460 nm in which perylene has its
absorption. There are two positive peaks and two negative
peaks at this region, which indicates that the two perylenes
were oriented in a right-handed manner. A positive couplet,
Potential application to the detection of a one-base deletion
with a FRET probe: We next used ODN5– ODN11 to eval-
uate the discrimination ability of deletion polymorphisms
from full match. The probe design for a one-base deletion is
depicted in Scheme 2. Two perylenes and several donors
were introduced into ODN. When this ODN is hybridized
Scheme 2. Schematic illustration of the design of the FRET probe.
2482
www.chemeurj.org
ꢁ 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2010, 16, 2479 – 2486

An Efficient Fluorescence Resonance Energy Transfer between Pyrene and Perylene
FULL PAPER
with a fully complementary strand (full match), all the chro-
mophores are, most likely, intercalated between base pairs
and monomer emission of perylene should be observed by
excitation at 345 nm at which the donor pyrene has its ab-
sorption due to efficient FRET between perylene and
pyrene. On the other hand, in a one-base-deletion mutant,
two perylenes should come close to each other by forming a
bulge-like structure so that an excimer of two perylenes
should be formed.[25]
that base-deletions of other nucleotides (not only A but C,
G and T) could be detected by this method, although dis-
crimination abilities varied depending on the sequence (see
Tables S1 and S2 in the Supporting Information).[38]
The I550/I461 ratios of other probes are summarized in
Table 3. The ratios of the ODN8 and ODN9 probes were
almost the same as that of ODN7. These results demon-
strate that incorporation of two pyrenes does not affect exci-
mer formation as long as the chromophores are separated
by intervening base pairs. Furthermore, an increase in the
number of pyrenes did not affect the ratio. The I550/I461 of
ODN11 was evaluated as 0.94 for C1, which was nine-times
larger than that for C2 (0.10). On the other hand, the differ-
ence in I550/I461 between ODN10/C1 and ODN10/C2 was
smaller, because of the low melting temperatures of these
duplexes (the Tm values of ODN10/C1 and ODN10/C2 were
21.0 and 25.78C, respectively (see Table 3)). Accordingly, at
08C, which is far below the Tm of ODN10/C1 or ODN10/C2,
I550/I461 of ODN10/C2 decreased to 0.06, whereas that of
ODN10/C1 was 0.38 (data not shown), demonstrating that
incorporation of multiple donors did not lower the poly-
morphism-discriminating ability. Thus, all the results pre-
sented in this section show that the apparent Stokesꢀ shift
was successfully enlarged by the introduction of multiple
donors without disturbing the polymorphism-discriminating
ability.
The fluorescence emission spectra of ODN7/C1 and
ODN7/C2, excited at 345 nm, are shown in Figure 4. When
ODN7 was hybridized with full-matched target (C2), strong
Figure 4. Fluorescence emission spectra of ODN7/C1 (c) and ODN7/
C2 (g) at 208C. The excitation wavelength was 345 nm. Solution con-
ditions are as follows: [ODN7]=1.0 mm, [Cn]=1.2 mm, [NaCl]=100 mm,
pH 7.0 (10 mm phosphate buffer).
Discussion
Efficiency of energy transfer from pyrene to perylene: The
FRET efficiencies (FT) of various duplexes were evaluated
as listed in Table 1. The apparent FT was estimated as >1
(1.18) for ODN2/C1, which indicated that efficient energy
transfer occurred in this duplex. Sufficient proximity of the
donor to the acceptor contributed to this highly efficient
energy transfer. Thus, the separation of dyes by intervening
base pairs is an effective strategy for the accumulation of
chromophores in ODNs. However, a FT value that is larger
than 1 is not a reasonable value. One of the causes of this
discrepancy is a structural disturbance in ODN2/C1. Be-
cause three fluorophores and four natural base pairs were
alternately introduced into the centre of the duplex, the
double-helical structure is expected to be disturbed to some
extent. Thus, quenching by neighbouring bases may be re-
duced in the case of ODN2/C1 as compared with ODN1/C1,
which raised the emission from perylene (IAD) and gave a
large FT. Indeed, the melting temperature of ODN2/C1 was
only 34.88C, which was 10.28C lower than that of ODN1/
C1. These structural differences might affect the fluores-
cence intensity of perylene.[39]
monomer emission of perylene was observed. In contrast,
monomer emission was significantly quenched upon hybridi-
zation with the deletion mutant C1. Concurrently, a new
shoulder peak appeared at 550 nm, which indicated that an
excimer of two perylenes was formed. The ratio of excimer
to monomer emission (I550/I461) at 208C was as high as 0.61
for ODN7/C1 corresponding to the deletion mutant due to
excimer emission and strong quenching of monomer emis-
sion. This value was eight-times larger than that of ODN7/
C2 (0.07) corresponding to full match.[35,36] ODN5, a probe
without pyrene as an antenna,[25] showed similar results: I550
/
I461 was 0.84 for the deletion mutant C1, whereas it was 0.05
for full-match C2 when excited at 425 nm at which perylene
has its absorption maximum (see Table 3 and Figure S6 in
the Supporting Information for the actual fluorescence spec-
tra of ODN5/C1 and ODN5/C2). Note that the Stokesꢀ shift
of ODN5 was as small as 35 nm, whereas that of ODN7,
which has pyrenes as an antenna was as large as 115 nm.
Thus, incorporation of pyrenes into the appropriate posi-
tions, such as those in ODN7, could efficiently enhance the
apparent Stokesꢀ shift without losing the ability to discrimi-
nate between the wild type and the deletion mutant.[37] We
also investigated sequence dependence on the discrimina-
tion ability by changing the intervening base. It was found
In our system, FRET efficiency strongly depended on the
distance between the acceptor and the donors. For example,
the FT value of ODN3/C1 was as low as 0.53, although
there are only three base pairs between pyrene and pery-
lene. The dependency of FT on the distance (r) is expressed
by using Equations (1) and (2):
Chem. Eur. J. 2010, 16, 2479 – 2486
ꢁ 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
2483

H. Asanuma et al.
FT ¼ R60=ðR60 þ r6Þ
donor (pyrene) emission, although the quantum yield of
pyrene was not determined directly.[44,45]
ð1Þ
ð2Þ
R0 ¼ 0:2108½JðlÞk2nꢀ4FDꢁ
1=6
Since the ODNs in the series ODN5–ODN11 have two
acceptors, we expected that FRET efficiencies should de-
crease to one-half of the value of the ODNs of the series
ODN1– ODN4 that have a single acceptor. However, FT of
ODN7/C2 was as small as 0.36, whereas that of ODN2/C1
was greater than 1. The relative intensity of ODN7/C2 with
respect to that of ODN5/C2 was only 3.7. Although the rea-
sons for this result are still not clear, the structural differen-
ces between ODN2/C1 and ODN7/C2 might affect the fluo-
rescence intensity. A comparison of the fluorescence intensi-
ty of ODN6/C2 with that of ODN7/C2 indicates that two ac-
ceptors are required for efficient energy transfer. Since the
R0 of pyrene is rather short, excited energy transfers only to
the nearest perylene and not to the distant perylene.[46] Such
a short R0 also inhibits the enhancement of fluorescence in-
tensity even if the number of donors increases from two to
four. Thus, the intensity of ODN10/C2 was almost the same
as that of ODN7/C2. Similarly, the intensity of ODN11/C2
was even lower than that of ODN8/C2. On the basis of
these results, we concluded that two pyrenes are sufficient
for efficient energy transfer as long as one base pair is in-
serted between the donor (pyrene) and acceptor (perylene)
due to the short R0 of the intercalated pyrene.
in which R0 is the Fçrster radius, J(l) is the integral of spec-
tral overlap between the donor and acceptor, k is the orien-
tational factor, n is the refractive index of the solution and
FD is the quantum yield of the donor.[40] Masuko et al. esti-
mated R0 as 22.3 ꢂ assuming k2 to be 2/3.[14] In the present
study, R0 should be shortened because a three-base separa-
tion strongly decreased the FRET efficiency as observed
with ODN3/C1.[41] Energy-minimized structures of ODN2/
C1 and ODN3/C1, calculated by using InsightII/Discover3,
are shown in Figure 5. In both duplexes, pyrenes and pery-
The ODN series of ODN6–ODN11 successfully discrimi-
nated the deletion mutant (C1) from the full match (C2)
based on the change in I550/I461. By use of a FRET pair, the
apparent Stokesꢀ shift of these new probes was much larger
(115 nm) than that of the previous probe composed of only
perylene (35 nm).[25] As shown in Table 3, the I550/I461 ratios
of ODN5/C2–ODN9/C2 did not show much change, which
demonstrated that the polymorphism-discrimination ability
did not decrease despite the fact that many donors were in-
serted. Even exciplex emission, which should be observed
when a pyrene–perylene heterodimer is formed,[47,48] was not
observed at all. These results demonstrate that one base
pair is enough to separate donor and acceptor or multiple
donors.
Duplex stabilities: The melting temperatures were strongly
influenced by the position of pyrene. As shown in Table 1,
insertion of pyrene in the nearest-neighbour base pairs
largely destabilized the duplex (ODN2/C1). On the other
hand, ODN3/C1, which has three base pairs between pyrene
and perylene, stabilized the duplex. The same tendency was
observed with ODN7–ODN9, in which, as the number of in-
tervening base pairs increased, the Tm value increased. Note
that the Tm of ODN7/C2, which has one base pair between
pyrene and perylene, was much lower than that of ODN8/
C2, which has two base pairs. This phenomenon is consistent
with the nearest-neighbour exclusion principle: free interca-
lators cannot be intercalated between the nearest-neighbour
base pairs.[49,50] The same is true for ODN10 and ODN11,
both of which have four pyrenes. The Tm value of ODN10/
C2 was 25.78C, which was 13.18C lower than that of
ODN11/C2 (38.88C).
Figure 5. Energy-minimized structures of ODN3/C1 (left side) and
ODN2/C1 (right side). The relative orientations of the chromophores are
shown at the bottom. Pyrene and perylene moieties are coloured in blue
and yellow, respectively.
lenes were stacked within the duplex.[42] The relative orien-
tations of the chromophores are also shown at the bottom
of Figure 5. The angles between the transition dipoles of
pyrene and perylene were 13 and 708 in ODN2/C1, but were
2 and 558 in ODN3/C1, which indicated that k2 was not very
low in these duplexes.[43] From these results, the relatively
short R0 may be attributed to the low quantum yield of
2484
www.chemeurj.org
ꢁ 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2010, 16, 2479 – 2486

An Efficient Fluorescence Resonance Energy Transfer between Pyrene and Perylene
FULL PAPER
Measurement of the melting temperature: The melting curve of duplex
DNA was obtained with a Shimadzu UV-1800 by measurement of the
change in absorbance at 260 nm versus temperature. The melting temper-
ature (Tm) was determined from the maximum in the first derivative of
the melting curve. Both the heating and cooling curves were measured,
and the calculated Tm values agreed to within 2.08C. The temperature
ramp was 0.58C minꢀ1. The sample solutions were as follows (unless oth-
erwise noted): [NaCl]=100 mm, pH 7.0 (10 mm phosphate buffer),
[ODNn]=[Cn]=5.0 mm.
Conclusion
An efficient FRET system of pyrene and perylene interca-
lated inside a duplex was constructed. Chromophores can be
accumulated in close proximity by this design because base
pairs completely suppress their interactions. The apparent
Stokesꢀ shift of the FRET probe was successfully increased
by the introduction of multiple donors close to acceptors
without disturbing the excimer formation. In these probes,
chromophores were strictly arranged in a predetermined po-
sition and order in the duplex without undesirable interac-
tions so that energy transfer and excimer formation were
strictly controlled. This strategy is especially useful for the
monitoring of biological samples because autofluorescence
from the sample that often interferes with the detection is
minimized.
Calculation of FRET efficiencies: FRET efficiencies (FT) were calculated
according to Equation (3):[40]
ꢀ
ꢁ
AA IAD
AD IA
FT
¼
ꢀ1
ð3Þ
in which AA is the absorbance at 345 nm of sequences which have only
acceptor(s) (ODN1/C1 or ODN5/C2). AD is the absorbance at 345 nm of
duplexes without acceptors (ODNnd/C1 or C2). IAD and IA are the fluo-
rescence intensities at the acceptor emission wavelength (460 nm) of se-
quences in the presence (ODN2–4/C1 or ODN7, 8 or 10/C2) or absence
(ODN1/C1 or ODN5/C2) of donors, respectively.
Experimental Section
Molecular modeling: The Insight II/Discover 98.0 program package was
used for conformational energy minimization. ODNs modified by pery-
lene and pyrene were built from canonical B-form DNA by a graphical
program. The AMBER95 force field was used for the calculation. All the
structures were energy-minimized to an RMS derivative of
<0.001 kcalꢂꢀ1. Computations were carried out on a Silicon Graphics
O2+ workstation with the operating system IRIX64 Release 6.5.
Materials: All the conventional phosphoramidite monomers, CPG col-
umns, reagents for DNA synthesis and Poly-Pak II cartridges were pur-
chased from Glen Research. Other reagents for the synthesis of phos-
phoramidite monomers were purchased from Tokyo Kasei Co and Al-
drich.
Synthesis of the DNA modified with pyrene and perylene: All the modi-
fied ODNs were synthesized on an automated DNA synthesizer (ABI-
3400 DNA synthesizer, Applied Biosystems) by using phosphoramidite
monomers bearing dye molecules. Synthetic procedures for the perylene
monomer were reported previously.[25] See Scheme S1 in the Supporting
Information for synthetic procedures and NMR spectroscopic assign-
ments of the pyrene monomer. The coupling efficiency of the monomers
corresponding to the modified residues was as high as the conventional
ones as judged from the coloration of the released trityl cation. After the
recommended workup, they were purified by reversed phase (RP)-HPLC
and characterized by MALDI-TOFMS (Autoflex II, BRUKER DAL-
TONICS). Purities of all the modified ODNs were >99% estimated by
RP-HPLC analysis.
Acknowledgements
This work was supported by Core Research for Evolution Science and
Technology (CREST), Japan Science and Technology Agency (JST). Par-
tial support by a Grant-in-Aid for Scientific Research from the Ministry
of Education, Culture, Sports, Science and Technology (Japan) and the
Mitsubishi Foundation (for H.A.) are also acknowledged.
MALDI-TOFMS data for the ODNs: ODN1: calcd for [ODN1+H+]:
4118; found: 4118; ODN2: calcd for [ODN2+H+]: 4992; found: 4992;
ODN3: calcd for [ODN3+H+]: 4992; found: 4992; ODN4: calcd for
[ODN4+H+]: 5866; found: 5866; ODN5: calcd for [ODN5+H+]: 4904;
found: 4904; ODN6: calcd for [ODN6+H+]: 5341; found: 5341; ODN7:
calcd for [ODN7+H+]: 5778; found: 5778; ODN8: calcd for [ODN8+H+
]: 5778; found: 5778; ODN9: calcd for [ODN9+H+]: 5778; found: 5779;
ODN10: calcd for [ODN10+H+]: 6653; found: 6654; ODN11: calcd for
[ODN11+H+]: 6653; found: 6652; ODN2d: calcd for [ODN2d+H+]:
4519; found: 4518; ODN3d: calcd for [ODN3d+H+]: 4519; found: 4518;
ODN4d: calcd for [ODN4d+H+]: 5393; found: 5394; ODN7d: calcd for
[ODN7d+H+]: 4832; found: 4832; ODN8d: calcd for [ODN8d+H+]:
4832; found: 4832; ODN10d: calcd for [ODN10d+H+]: 5706; found:
5705.
[1] M. S. T. GonÅalves, Chem. Rev. 2009, 109, 190–212.
[2] R. B. Mujumdar, L. A. Ernst, S. R. Mujumdar, C. J. Lewis, A. S.
Waggoner, Bioconjugate Chem. 1993, 4, 105–111.
[3] Fluorescence of Cy3 and Cy5 is dependent on the neighboring se-
quence.
[4] X. Peng, F. Song, E. Lu, Y. Wang, W. Zhou, J. Fan, Y. Gao, J. Am.
Chem. Soc. 2005, 127, 4170–4171.
[5] Kim E. Sapsford, L. Berti, I. L. Medintz, Angew. Chem. 2006, 118,
4676–4704; Angew. Chem. Int. Ed. 2006, 45, 4562–4589.
[6] P. R. Selvin, S. Kenneth, Methods Enzymol. 1995, 246, 300–334.
[7] S.-C. Hung, R. A. Mathies, A. N. Glazer, Anal. Biochem. 1998, 255,
32–38.
[8] S. Kawahara, T. Uchimaru, M. Shigeo, Chem. Commun. 1999, 563–
564.
Spectroscopic measurements: CD spectra were measured on a JASCO
model J-820 with a 10 mm quartz cell that was equipped with pro-
grammed temperature controllers. A slight excess of complementary
strand was used to ensure that all the probe ODN formed duplex. The
sample solutions were as follows: [NaCl]=100 mm, pH 7.0 (10 mm phos-
phate buffer), [ODN7]=1.0 mm, [C2]=1.2 mm.
[9] A. K. Tong, S. Jockusch, Z. Li, H.-R. Zhu, D. L. Akins, N. J. Turro, J.
Ju, J. Am. Chem. Soc. 2001, 123, 12923–12924.
[10] Y. Ohya, K. Yabuki, M. Hashimoto, A. Nakajima, T. Ouchi, Biocon-
jugate Chem. 2003, 14, 1057–1066.
[11] F. D. Lewis, L. Zhang, X. Zuo, J. Am. Chem. Soc. 2005, 127, 10002–
10003.
Fluorescence spectra were measured on a JASCO model FP-6500 with a
microcell. The excitation wavelength was 345 nm. For FRET analysis, ex-
periments were conducted at 08C to ensure sufficient hybridization. The
sample solutions were as follows (unless otherwise noted): [NaCl]=
100 mm, pH 7.0 (10 mm phosphate buffer), [ODNn]=1.0 mm, [Cn]=
1.2 mm.
[12] A. Iqbal, S. Arslan, B. Okumus, T. J. Wilson, G. Giraud, D. G.
Norman, T. Ha, D. M. J. Lilley, Proc. Natl. Acad. Sci. USA 2008,
105, 11176–11181.
[13] M. Tomura, E. Ishiguro, N. Mataga, J. Phys. Soc. Jpn. 1967, 22, 1117.
[14] M. Masuko, S. Ohuchi, K. Sode, H. Ohtani, A. Shimadzu, Nucleic
Acids Res. 2000, 28, 34e.
Chem. Eur. J. 2010, 16, 2479 – 2486
ꢁ 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
2485

H. Asanuma et al.
[15] D. Lindegaard, A. S. Madsen, I. V. Astakhova, A. D. Malakhov,
B. R. Babu, V. A. Korshun, J. Wengel, Bioorg. Med. Chem. 2008, 16,
94–99.
[16] X. G. Liang, H. Asanuma, H. Kashida, A. Takasu, T. Sakamoto, G.
Kawai, M. Komiyama, J. Am. Chem. Soc. 2003, 125, 16408–16415.
[17] H. Asanuma, K. Shirasuka, M. Komiyama, Chem. Lett. 2002, 490–
491.
[18] C. Brotschi, C. J. Leumann, Angew. Chem. 2003, 115, 1694–1697;
Angew. Chem. Int. Ed. 2003, 42, 1655–1658.
[19] S. M. Langenegger, R. Hꢃner, Chem. Commun. 2004, 2792–2793.
[20] M. Nakamura, Y. Ohtoshi, K. Yamana, Chem. Commun. 2005,
5163–5165.
[21] E. Mayer-Enthart, H.-A. Wagenknecht, Angew. Chem. 2006, 118,
3451–3453; Angew. Chem. Int. Ed. 2006, 45, 3372–3375.
[22] L.-A. Fendt, I. Bouamaied, S. Thoni, N. Amiot, E. Stulz, J. Am.
Chem. Soc. 2007, 129, 15319–15329.
[23] S. Levy, G. Sutton, P. C. Ng, L. Feuk, A. L. Halpern, B. P. Walenz, N.
Axelrod, J. Huang, E. F. Kirkness, G. Denisov, Y. Lin, J. R. Mac-
Donald, A. W. C. Pang, M. Shago, T. B. Stockwell, A. Tsiamouri, V.
Bafna, V. Bansal, S. A. Kravitz, D. A. Busam, K. Y. Beeson, T. C.
McIntosh, K. A. Remington, J. F. Abril, J. Gill, J. Borman, Y.-H.
Rogers, M. E. Frazier, S. W. Scherer, R. L. Strausberg, J. C. Venter,
PLoS Biol. 2007, 5, e254.
[24] H. Kashida, H. Asanuma, M. Komiyama, Chem. Commun. 2006,
2768–2770.
[25] H. Kashida, T. Takatsu, H. Asanuma, Tetrahedron Lett. 2007, 48,
6759–6762.
[26] U. B. Christensen, E. B. Pedersen, Nucleic Acids Res. 2002, 30,
4918–4925.
[27] U. B. Christensen, E. B. Pedersen, Helv. Chim. Acta 2003, 86, 2090–
2097.
[28] A. D. Malakhov, M. V. Skorobogatyi, I. A. Prokhorenko, S. V. Gon-
tarev, D. T. Kozhich, D. A. Stetsenko, I. A. Stepanova, Z. O. Shen-
karev, Y. A. Berlin, V. A. Korshun, Eur. J. Org. Chem. 2004, 1298–
1307.
[29] O. Kçhler, D. V. Jarikote, O. Seitz, ChemBioChem 2005, 6, 69–77.
[30] D. Baumstark, H.-A. Wagenknecht, Angew. Chem. 2008, 120, 2652–
2654; Angew. Chem. Int. Ed. 2008, 47, 2612–2614.
[31] The wavelength of UV light for excitation (l=345 nm) might be too
short especially for in vivo analysis because it potentially damages
other biomolecules.
[32] Incorporation of a mismatch into ODN4/C1 lowered the melting
temperature although the degree of destabilization was dependent
on the position of the mismatch. See Table S3 in the Supporting In-
formation for melting temperatures.
[36] The difference in I550/I461 was enhanced with increasing ionic
strength. See Figure S5 in the Supporting Information.
[37] The probe presented in this paper also emits fluorescence even in
the single-stranded state. This emission should disturb the detection
of the target especially when the amount of the target is small.
[38] The I550/I461 ratio of ODN7/C7 (5’-GGTATPCEAEGPCAATC-3’/3’-
CCATAGGCGTTAG-5’) containing an A–G mismatch was larger
than that of full-matched ODN7/C2, but smaller than ODN7/C1
containing one-base deletion. The increase in the I550/I461 ratio by in-
troducing mismatch is attributable to the disordering around the in-
tervening bases that facilitated the dimerization of perylene. These
results also support proper base pairing of “full match” even though
multiple perylenes and pyrenes were incorporated in the nearest
neighboring positions. See Table S4 in the Supporting Information.
[39] In other reports, chromophores were mainly introduced into grooves
or outside duplex. Consequently, this structural problem could be
avoided in those reports.
[40] B. Valeur, Molecular Fluorescence, Wiley-VCH, Weinheim, 2002.
[41] The distance between the donor and acceptor was about 13 ꢂ, esti-
mated from the energy-minimized structure shown in Figure 5.
[42] CD spectra indicated that both ODN2/C1 and ODN3/C1 took simi-
lar B-DNA structures. See Figure S7 in the Supporting Information.
[43] When the chromophores are stacked in the parallel plane, the orien-
tation factor, k, is determined from the following equation: k2 =
cos2q, in which q is the angle between the transition moments of the
dyes. Since R0 is proportional to sixth root of k2, the k contributes to
R0 significantly only when q is close to 908. For example, in the case
of q=138, R0 increases only 6% relative to that calculated by as-
suming k2 as 2/3. On the other hand, R0 decreased 25% in the case
of q=708.
[44] In our study, peaks of pyrene were substantially weaker than those
of perylene.
[45] Quenching of pyrene is probably due to electron transfer between
pyrene and neighboring bases. For quenching towards pyrene by
natural bases, see: M. Manoharan, K. L. Tivel, M. Zhao, K. Nafisi,
T. L. Netzel, J. Phys. Chem. 1995, 99, 17461–17472.
[46] As a result, two pyrenes were required for the efficient energy trans-
fer.
[47] J. N. Wilson, J. Gao, E. T. Kool, Tetrahedron 2007, 63, 3427–3433.
[48] J. Chiba, S. Takeshima, K. Mishima, H. Maeda, Y. Nanai, K.
Mizuno, M. Inouye, Chem. Eur. J. 2007, 13, 8124–8130.
[49] D. M. Chothers, Biopolymers 1968, 6, 575–584.
[50] We believe that all the chromophores were intercalated within the
duplex because of the following reasons. No obvious excimer or ex-
ciplex emission was observed even with ODN10/C2, which has six
chromophores in the nearest-neighboring positions. If the chromo-
phores were flipped out from the duplex, they should form an exci-
mer or exciplex, because flipped out hydrophobic chromophores
with flexible ethylene or propylene linkers prefer to be stacked
against each other in the aqueous environment. Similar results were
also reported by Murakami and Yamana (M. Nakamura, Y. Shimo-
mura, Y. Ohtoshi, K. Sasa, H. Hayashi, H. Nakano, K. Yamana,
Org. Biomol. Chem. 2007, 5, 1945–1951, A. Mahara, R. Iwase, S. Sa-
kamoto, K. Yamana, T. Yamaoka, A. Murakami, Angew. Chem.
2002, 114, 3800–3802; Angew. Chem. Int. Ed. 2002, 41, 3648–3650).
They demonstrated that bispyrene flipped out from the duplex was
assembled to form an excimer, whereas intercalated bispyrene
showed only monomer emission as observed in our case. However,
we cannot completely rule out the possibility that some chromo-
phores were flipped out without forming an excimer and exciplex
because of the lack of direct evidence of intercalated structure.
[33] Electron transfer between pyrene and perylene is rather unlikely be-
cause location of pyrene near the perylene did not quench the fluo-
rescence from perylene. Furthermore, quenching of pyrene was not
necessarily facilitated by approximating pyrene to perylene (com-
pare the fluorescence from pyrene (l=370–430 nm) of ODN8/C2
and ODN9/C2 in Figure 2).
[34] When ODN7 was hybridized with full-matched C2, peaks of pyrene
at 330–360 nm showed bathochromicity in absorption spectra, which
indicated that pyrenes were intercalated between base pairs. On the
other hand, peaks of perylene slightly shifted to shorter wavelength.
This shift can be attributed to the change of dimeric perylene in the
single-stranded state to monomeric state in the duplex with C2. See
Figures S2 and S3 in the Supporting Information for UV/Vis and
emission spectra. For the relationship between bathochromicity and
intercalation, see: W. Mꢄller, D. M. Crothers, J. Mol. Biol. 1968, 35,
251–290.
[35] Although the change in I550/I461 was not so large, the difference in
fluorescence can be discriminated by the naked eye. See Figure S4
in the Supporting Information for the photograph.
Received: July 27, 2009
Revised: November 4, 2009
Published online: January 11, 2010
2486
www.chemeurj.org
ꢁ 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2010, 16, 2479 – 2486