Targeting DNA with light-up pyrimidine triple-helical forming oligonucleotides conjugated to stabilizing fluorophores (LU-TFOs)

Article abstract of DOI:10.1039/b813289e

The synthesis of triple-helix-forming oligonucleotides (TFOs) linked to a series of cyanine monomethines has been performed. Eight cyanines including one thiocyanine, four thiazole orange analogues, and three quinocyanines were attached to the 5′-end of 10-mer pyrimidine TFOs. The binding properties of these modified TFOs with their double-stranded DNA target were studied by absorption and steady-state fluorescence spectroscopy. The stability of the triplex structures depended on the cyanine structure and the linker size used to connect both entities. The most efficient cyanines able to stabilize the triplex structures, when attached at the 5′-end of the TFO, have been incorporated at both ends and provided triplex structures with increased stability. Fluorescence studies have shown that for the TFOs involving one cyanine, an important intensity increase (up to 37-fold) in the fluorescent signal was observed upon their hybridization with the double-stranded target, proving hybridization. The conjugates involving thiazole orange attached by the benzothiazole ring provided the most balanced properties in terms of triplex stabilization, fluorescence intensity and fluorescence enhancement upon hybridization with the double-stranded target. In order to test the influence of different parameters such as the TFO sequence and length, thiazole orange was used to label 17-mer TFOs. Hybridizations of these TFOs with different duplexes, designed to study the influence of mismatches at both internal and terminal positions on the triplex structures, confirmed the possibility of triplex formation without loss of specificity together with a strong fluorescence enhancement (up to 13-fold).

Full text of DOI:10.1039/b813289e

View Article Online / Journal Homepage / Table of Contents for this issue
PAPER
www.rsc.org/obc | Organic & Biomolecular Chemistry
Targeting DNA with “light-up” pyrimidine triple-helical forming
oligonucleotides conjugated to stabilizing fluorophores (LU-TFOs)
Brice-Lo¨ıc Renard, Re´my Lartia† and Ulysse Asseline*
Received 1st August 2008, Accepted 19th August 2008
First published as an Advance Article on the web 16th October 2008
DOI: 10.1039/b813289e
The synthesis of triple-helix-forming oligonucleotides (TFOs) linked to a series of cyanine
monomethines has been performed. Eight cyanines including one thiocyanine, four thiazole orange
analogues, and three quinocyanines were attached to the 5¢-end of 10-mer pyrimidine TFOs. The
binding properties of these modified TFOs with their double-stranded DNA target were studied by
absorption and steady-state fluorescence spectroscopy. The stability of the triplex structures depended
on the cyanine structure and the linker size used to connect both entities. The most efficient cyanines
able to stabilize the triplex structures, when attached at the 5¢-end of the TFO, have been incorporated
at both ends and provided triplex structures with increased stability. Fluorescence studies have shown
that for the TFOs involving one cyanine, an important intensity increase (up to 37-fold) in the
fluorescent signal was observed upon their hybridization with the double-stranded target, proving
hybridization. The conjugates involving thiazole orange attached by the benzothiazole ring provided
the most balanced properties in terms of triplex stabilization, fluorescence intensity and fluorescence
enhancement upon hybridization with the double-stranded target. In order to test the influence of
different parameters such as the TFO sequence and length, thiazole orange was used to label 17-mer
TFOs. Hybridizations of these TFOs with different duplexes, designed to study the influence of
mismatches at both internal and terminal positions on the triplex structures, confirmed the possibility
of triplex formation without loss of specificity together with a strong fluorescence enhancement (up to
13-fold).
fluorogenic probes able to detect specific sequences on DNA
Introduction
targets are peptide nucleic acid (PNA) molecular beacons.11 These
probes are able to dissociate the double strand to hybridize
with their target sequences. During the past decade a great
deal of effort has been devoted to modifying TFOs to improve
binding properties to their cognate sequences in a cellular
environment.12,13 These developments have led to TFOs able to
specifically and strongly hybridize to their double-stranded DNA
targets. These TFOs can be used to detect specific sequences on
DNA duplexes without having to separate both strands. We now
report the synthesis and photophysical properties of pyrimidine
TFOs labelled with monomethine cyanines. The influence of
the labelling on the triplex stability was studied by thermal
denaturation followed by absorption spectroscopy. The changes in
the fluorescence emission, upon triplex formation, were evaluated
by steady-state fluorescence experiments. We have also investigated
the possibility of increasing the difference in the fluorescence
emission between the free and the hybridized labelled TFO by
quenching the fluorescence of the free TFOs by hybridization
with a complementary short probe involving a quencher. The
influence of different parameters such as the TFO sequence
and length and the presence of mismatches at both the internal
and terminal positions on the triplex structures have also been
studied. The possibility of forming a triplex with increased stability
by using a labelled TFO designed to hybridize with a duplex
involving a pyrimidine–purine base-pair inversion has also been
investigated.
The detection of specific nucleic acid sequences by the hybridiza-
tion of fluorescent oligonucleotides is widelyused inbasic research,
diagnostics and biotechnology.1–6 In the past few years a series
of fluorescent oligonucleotides able to produce a fluorescence
emission change in the presence of single-stranded target se-
quences, proving the hybridization, has been reported. However,
only a few examples of emission changes in the presence of their
double-stranded DNA targets have been described.7–11 The first
concerns pairs of TFOs conjugated to labels designed in such a
way as to produce FRET upon hybridization with the double-
stranded DNA target.7 The second involves light-up probes, that
is one TFO labelled with one (or two identical) fluorophores,
whose fluorescence is enhanced upon triplex formation. They were
labelled with either oxazole orange, providing a 3-fold fluorescence
enhancement upon hybridization8 or a ruthenium complex giving
a 10-fold increase.9 Pyrimidine TFOs doubly labelled with perylene
have also shown a moderate 3-fold fluorescence enhancement
in the presence of the target sequence.10 Another example of
Centre de Biophysique Mole´culaire, CNRS UPR 4301, affiliated with the
University of Orle´ans and INSERM, Rue Charles Sadron, 45071 Orle´ans
Cedex 02, France. E-mail: asseline@cnrs-orleans.fr; Fax: + 33-2-38-63-
15-17
† Present address: De´partement de Chimie Mole´culaire, UMR 5250,
ICMG FR 2607, CNRS, Universite´ Joseph Fourier, BP 53, 38041 Grenoble
Cedex 9, France
This journal is
The Royal Society of Chemistry 2008
Org. Biomol. Chem., 2008, 6, 4413–4425 | 4413
©

View Article Online
development of intercalator–TFO conjugates.19,20 The pyrimidine
DNA sequences were reported to hybridize with their duplex
DNA targets in a parallel orientation with respect to the purine
strand.12,21–22 Labelling was first performed at the 5¢- position
because the duplex–triplex junction at the 5¢-end of the triplex
nucleic acid structure is favourable for intercalation.21–23 We chose
linkers with eight methylene groups because it has been shown that
linkers with lengths equal or superior to five or six carbon atoms
are required to enable intercalation.22,24,25
To study the influence of the linker size used to connect the TFO
and the fluorophore, the most efficient cyanine at stabilizing the
triplex structure, 4TO¢, was also linked to the 5¢- end of the TFO
via linkers involving seven, six, five and four methylene groups.26
Bis-conjugations at both the 5¢- and 3¢-ends of the TFO of the
most efficient cyanines at stabilizing the triplex structure when
attached to the 5¢-end of the TFO, 4TO¢ and 2TO, were also
performed using the linker size (eight methylene groups) showing
the strongest stability when 4TO¢ was attached to the 5¢-end. Bis-
labelling was also achieved with quinocyanine 24¢Q giving the
greatest fluorescence increase upon triplex formation. The linker
size used to connect 4TO¢ to the 3¢-end was also increased by three
Results and discussion
Structures and experimental design
These fluorescent TFOs are made of an oligopyrimidine sequence
involving either one monomethine cyanine attached at the 5¢-
end or two (identical ones) linked to both ends. Their structures
are shown in Fig. 1. The eight monomethine cyanines, including
one thiocyanine (Th), four thiazole orange analogues (2TO,
4TO, 2TO¢and 4TO¢), and three quinocyanines (22¢Q, 24¢Q and
42¢Q), used as the labels were previously prepared14 but were
not used to label TFOs. Monomethine cyanines were chosen
because they are able to intercalate into double-stranded nucleic
acid targets. Thus, reduction of the rotation around the methine
bond, provided by intercalation, has been reported to be one of
the parameters responsible for a strong increase in fluorescence
observed on binding of the most frequently used monomethine
cyanine, thiazole orange, with DNA duplexes.15–18 They were,
firstly, covalently linked to the 5¢-end of a 10-mer oligopyrimidine
sequence d-5¢TTTTCTTTTC3¢ via a phosphothiolodiester bond
and an octamethylene linker (Fig. 1). This sequence was chosen
because it had been used previously in our laboratory for the
Fig. 1 Structures of labelled TFOs and oligonucleotides.
4414 | Org. Biomol. Chem., 2008, 6, 4413–4425 This journal is
The Royal Society of Chemistry 2008
©

View Article Online
atoms because it has been demonstrated that the stabilization
of triple helices by an acridine derivative attached to the 3¢-
end of a TFO required the use of a longer linker as compared
with the size of the one used to link this intercalator at the 5¢-
position.27 We report here the synthesis of these TFO-conjugates
and their binding properties with their double-stranded DNA
targets, evaluated by absorption and fluorescence spectroscopy.
Synthesis
The attachment of labels to oligonucleotides can be achieved
following two strategies.28 The first consists in their direct in-
corporation using the phosphoramidites of the labels (or the
phosphoramidite derivatives of modified nucleosides bearing the
labels). The 3¢-modification can also be achieved by using modified
supports involving these labels. However, the direct incorporation
of labels during the oligonucleotide synthesis requires labels or
nucleoside analogues able to withstand the chemical conditions
needed for the deprotection step and that are soluble in organic sol-
vents. Another strategy consists in the post-synthetic attachment
of labels to various positions of the oligonucleotides by specific
reactions between convenient functional groups incorporated at
pre-selected positions on both entities. Since our previous work
has shown that cyanines are not very stable in strong alkali
conditions (unpublished results), we chose the post-synthetic
coupling strategy for the preparation of the labelled TFOs (vide
infra).
Synthesis of modified TFOs bearing the cyanines at their 5¢-
ends. Synthesis of conjugates involving cyanines linked to the
5¢-end of the TFO 44TO¢8 to 442¢Q8 relies on the incorporation of a
thiophosphate group at the 5¢-end of the oligonucleotide bound
to the support 7 (Scheme 1a). The TFO was synthesized by phos-
phoramidite chemistry29 and after an additional detritylation step,
a masked thiophosphate group was added using our previously
reported method.30 After the deprotection step, the crude 5¢-
thiophosphorylated TFO 4 was reacted with the halogenoalkyl
groups of the cyanines, obtained as previously described14,26 in a
mixture of an aqueous bicarbonate buffer and dimethylformamide
(see Fig. 1 for the structures). The conjugates were purified by
reversed-phase chromatography (see Fig. 2 and Table 1) and
characterized by electrospray mass spectrometry (Table 1) and
UV–visible analyses (see Table 1 for lmax absorption in the UV
and visible range). The visible absorption bands of the different
conjugates span from 427 to 561 nm. The UV–visible spectra of
the 5¢-labelled TFOs 424¢Q8, 42TO8 and 44TO¢8 are shown in Fig. 3.
Scheme 1 Synthesis of 5¢-labelled TFOs 44TO¢8 to 442¢Q8 (a) and 5¢-,3¢-bis-
labelled TFOs 54TO¢8, 52TO8 and 524¢Q8 (b).
Synthesis of TFOs bearing cyanines at both the 5¢- and 3¢-
ends. The synthesis of the bis-conjugated TFOs 52TO8, 524¢Q8
and 54TO¢8 involving both cyanines attached via the same linker
length was performed by incorporation of a thiophosphate group
at both ends of the TFO (Scheme 1b). The synthesis was
performed on our previously reported support 8.31 A masked
thiophosphate group was then incorporated at the 5¢-end as
described above. After the deprotection and purification steps,
the bis 5¢-,3¢-thiophosphorylated TFO 5 was reacted with the
selected iodoalkylated cyanines to give TFO conjugates 52TO8, 524¢Q8
and 54TO¢8. The synthesis of the bis-conjugated TFO involving
two cyanines attached with linkers of different sizes 64TO¢8.4 was
performed by incorporation of a thiophosphate group at its 5¢-
end and a thiol masked function at its 3¢-end in order to perform
a sequential coupling at each end (Scheme 2). The incorporation
of the thiol function at the 3¢-end of the TFO was achieved by
using a new modified support leading, after the deprotection step,
to oligonucleotides with a 6,6¢-dithiodihexanol linker.
The new support (Scheme 2) was obtained following a five-
step procedure. First, thiohexanol 9 was dimerized into 6,6¢-
dithiodihexanol. The latter was monodimethoxytritylated to give
10. Then reaction with succinic anhydride followed by coupling
with p-nitrophenol in the presence of N,N¢-carbodiimide led to
the activated ester. The latter was then reacted with aminopropyl
controlled pore glass to give the modified support 11 with a loading
of 36 mmol g-1. The coupling of the bis-modified TFO 6 with
This journal is
The Royal Society of Chemistry 2008
Org. Biomol. Chem., 2008, 6, 4413–4425 | 4415
©

View Article Online
cation of the TFO conjugates was performed by reversed-phase
chromatography. The retention times (Table 1) for the conjugates
involving two cyanines were higher than those involving only one
cyanine. These results indicate the formation of more lipophilic
compounds. The UV–visible spectra of the bis-labelled TFOs 52TO8
524¢Q8 and 54TO¢8 are shown in Fig. 3.
,
Absorption studies
Free TFO conjugates. The UV–visible spectra of the TFO-
labelled conjugates were recorded between l = 230 nm and l =
700 nm. The spectra of the 5¢-conjugated TFOs contained two
main absorption bands. The absorption band in the UV range
corresponded to the absorbance of the TFO and the cyanines with
a lUV ~267 nm for all conjugates. The second absorption band in
the visible region between l = 350 nm and l = 700 nm corresponds
to the cyanine. The general shape involves a main band with a
shoulder at a shorter wavelength. The lmax values in the visible
range and the intensity ratios of the UV–visible bands depend on
the cyanine considered (data not shown). A comparison of the
UV–visible spectra of the mono- and bis-labelled TFOs, involving
2TO, 24¢Q and 4TO¢ as the labels, shows different shapes in the
visible range (Fig. 3).
Fig. 2 Reversed-phase HPLC analysis of the coupling reaction between
TFO 5 and the cyanine-linker derivative 4TO¢-(CH2)8I performed on a
LiChrospher RP 18 (5 mm) column (125 mm ¥ 4 mm) from Merck using
a linear gradient of CH3CN (5 to 38.5% over 45 min) in 0.1 M aqueous
ammonium acetate, pH = 7, with a flow rate of 1 cm3 min-1. Detection
l = 260 nm.
4TO¢-(CH2)8-I followed by release of the thiol function by cleavage
of the disulfide bridge by treatment with a reducing agent,32
and reaction with 4TO¢-(CH2)4-I led to the bis-derivatized 64TO¢8.4
involving two TO¢ attached via linkers of different sizes. Purifi-
Conjugate 54TO¢8 exhibited a main band in the visible range
at l = 478 nm and a shoulder at l = 508 nm with the one at
Table 1 Characterization of conjugates and Tm data
c
Mass analysis b
E/M-1 cm-1
Tm/◦C ( 0.5 ◦C)d DTm/◦C ( 1 ◦C)
a
ODNs Reversed-phase Rt
Calculated
Found
lmax UV
lmax vis
4
11 min 37 sec
28 min 49 sec
26 min 27 sec
24 min 05 sec
22 min 28 sec
22 min 21 sec
28 min 42 sec
28 min 19 sec
28 min 36 sec
28 min 29 sec
28 min 20 sec
28 min 51 sec
28 min 32 sec
11 min 48 sec
32 min 47 sec
34 min 27 sec
32 min 58 sec
11 min 24 sec
32 min 13 sec
25 min 50 sec
27 min 18 sec
27 min 30 sec
28 min 12 sec
24 min 08 sec
23 min 24 sec
C98H128N22O68P10S = 3044.00
C124H158N24O68P10S2 = 3446.60
C123H156N24O68P10S2 = 3432.57
C122H154N24O68P10S2 = 3418.54
C121H152N24O68P10S2 = 3404.52
C120H150N24O68P10S2 = 3390.49
C124H158N24O68P10S2 = 3446.60
C122H156N24O68P10S3 = 3452.62
C124H158N24O68P10S2 = 3446.60
C124H158N24O68P10S2 = 3446.60
C126H160N24O68P10S = 3440.57
C126H160N24O68P10S = 3440.57
C126H160N24O68P10S = 3440.57
C98H128N22O68P10S = 3139.03
C150H188N26O70P11S4 = 3944.24
C150H188N26O70P11S4 = 3944.24
C154H192N26O70P11S2 = 3932.19
C110H153N22O72P11S3 = 3372.43
C152H192N26O71P11S4 = 3988.30
C121H148N54O54P10 = 3530.58
C101H124N44O43P8 = 2889.15
C91H112N39O38P7 = 2575.94
3044.06
3446.37
3431.63
3417.55
3403.64
3389.88
3446.39
3452.23
3446.01
3446.27
3440.05
3440.00
3439.81
3141.54
3942.79
3943.36
3931.59
3374.41
4002.41
3530.58
2888.89
2575.95
2263.10
5532.59
5557.11
80 100 (267)
116 500 (267)
116 500 (267)
116 500 (267)
116 500 (267)
116 500 (267)
120 300 (267)
95 600 (267)
93 500 (267)
89 600 (267)
82 600 (267)
83 600 (267)
82 400 (267)
80 100 (267)
145 000 (267)
154 500 (267)
85 800 (267)
80 100 (267)
145 000 (267)
125 500 (256)
103 700 (256)
91 700 (256)
79 700 (256)
169 800 (267)
172 400 (267)
—
—
—
44TO¢8
44TO¢7
44TO¢6
44TO¢5
44TO¢4
42TO8
4Th8
82 700 (508)
82 700 (508)
82 700 (508)
82 700 (508)
82 700 (508)
62 700 (488)
77 100 (427)
48 400 (486)
65 300 (508)
58 600 (526)
59 200 (561)
42 100 (560)
—
143 600 (478)
143 900 (459)
74 500 (519)
—
111 500 (479)
29 300 (477)
29 300 (477)
29 300 (477)
29 300 (477)
82 700 (508)
82 700 (508)
37.5
35.0
32.5
29.0
26.5
32.5
31.0
28.0
27.5
26.5
25.5
19.5
—
45.0
39.0
30.5
—
43.0
—
+24.5
+22.0
+19.5
+16.0
+13.5
+19.5
+18.0
+15.0
+14.5
+13.5
+12.5
+6.5
—
42TO¢8
44TO8
422¢Q8
424¢Q8
442¢Q8
5
54TO¢8
52TO8
524¢Q8
6
64TO¢8.4
12dab
13dab
14dab
15dab
164TO¢8
174TO¢8
+32.0
+26.0
+17.5
—
+30.0
—
—
—
—
—
—
—
C81H100N34O33P6 = 2262.73
C191H247N41O114P17S2 = 5531.93
See Table 3
See Table 3
C191H246N44O113P17S2 = 5556.95
a Retention times obtained by reversed-phase HPLC analyses for the ODNs performed on a LiChrospher RP 18 (5 mm) column (125 mm ¥ 4 mm)
from Merck using a linear gradient of CH3CN (5% to 38.5% over 45 min) in 0.1 M aqueous ammonium acetate, pH 7, with a flow rate of 1 cm3 min-1.
b Mass analysis data for the modified TFOs and their conjugates and dabcyl conjugates. c Molar absorption coefficients for the modified TFOs and their
conjugates (the values in the brackets indicate the corresponding wavelength). For 5¢-labelled TFOs 4, 16 and 17 the e values at l = 260 nm were the
approximate sum of the e values of the TFOs and labels.14 E values were determined experimentally for labelled TFOs 5 and 6. The same e values were
used for the conjugates involving the same sequence and different linker lengths to connect the label and the TFO. d Tm values for triplexes formed by
the labelled TFOs and the double-stranded target (1 + 2) in a 10 mM sodium phosphate, pH 6, buffer containing 140 mM KCl and 5 mM MgCl2.
Concentrations were 1 mM in the duplex target and TFOs. Under the same experimental conditions, the Tm value for the unmodified triplex (1 + 2 + 3)
was 13 ◦C.
4416 | Org. Biomol. Chem., 2008, 6, 4413–4425
This journal is
The Royal Society of Chemistry 2008
©

View Article Online
Scheme 2 Synthesis of the bis-labelled TFO 64TO¢8.4
addition to the shape difference observed for the visible spectra of
mono- and bis-labelled TFOs, the observation of the UV–visible
absorption ratio indicated that the absorbance intensities in the
visible range for the bis-labelled TFOs were not twice those of
the corresponding mono-labelled TFOs. The shift of the visible
absorption band to the shortest wavelength observed for the bis-
conjugated TFO indicated aggregation of the cyanine dyes. Similar
spectrum changes have been previously reported for the thiazole
orange dimer.16 It is likely that, even at 60 ◦C, due to the flexibility
of the oligonucleotide chain, contacts between the two labels were
possible.
Fig. 3 UV–visible absorption spectra of water solutions of the mono-
labelled TFOs (full line) 424¢Q8 (a) 42TO8 (b) and 44TO¢8 (c); and bis-labelled
TFOs (bold line) 524¢Q8 (a), 52TO8 (b) and 54TO¢8 (c) recorded between l = 230
and l- = 700 nm.
Molar absorption coefficient determination. The molar ab-
sorption coefficients (e260nm values) for the conjugates derived
from TFO 4 were estimated to be the sum of the e values
of the TFO and cyanines deducted from ref. 14 (see Table 1).
The same e260nm values were used for the series of conjugates
44TO¢n involving different linker sizes (n = 4, 5, 6, 7 and 8).
Because the UV–visible absorption ratio observed for the bis-
conjugated TFOs 54TO¢8, 52TO8 and 524¢Q8 indicated intensities not
simply twice those of the corresponding mono-labelled TFOs, their
extinction coefficients at 260 nm were determined by titrations
of the conjugate solutions in a 10 mM KH2PO4 buffer (pH 6),
containing 140 mM KCl and 5 mM MgCl2 performed at 3 ◦C,
with solutions of the single-stranded complementary sequence
5¢GGGAAAAGAAAATTT3¢.19 The molar extinction coefficient
l = 478 nm being the more intense, while for the corresponding
conjugate 44TO¢8 involving only one TO¢ residue, the absorption
at 478 nm appeared only as a shoulder and the main band was
detected at l = 508 nm. Upon a temperature increase, from 20
◦
to 60 C, a blue-shift of 3 nm was observed for the visible band
with a slight hypochromism, 2–3%, for the 5¢-mono-conjugated
TFO 44TO¢8 while only hypochromism (4% at 478 nm) was detected
for the corresponding bis-conjugated TFO 54TO¢8. Similar intensity
changes were observed for TFOs 42TO8 and 52TO8 labelled with the
2TO. The spectrum of the mono-labelled TFO 42TO8 showed a
main peak with l = 488 nm and a shoulder at l = 470 nm,
while that of the bis-labelled TFO 52TO8 exhibited a main peak
at l = 459 nm and a shoulder at l = 500 nm. When 24¢Q was
used to label the TFO the visible band of the mono-labelled TFO
showed a main band at l = 561 nm with a shoulder at l =
528 nm. The visible spectra of the corresponding bis-labelled TFO
exhibited two main bands at l = 519 nm and l = 560 nm. In
at 260 nm used for 64TO¢8.4 was the same as for 54TO¢8
.
Hybridized TFO-conjugates. Mixing of the mono-conjugated
TFO derived from 4 with the double-stranded DNA target 1 + 2 at
3 ◦C induced only weak spectral modifications that were reversed
upon a temperature increase to 65◦. These changes showed a
This journal is
The Royal Society of Chemistry 2008
Org. Biomol. Chem., 2008, 6, 4413–4425 | 4417
©

View Article Online
slight red-shift (~4 nm) of the visible absorption band and a weak
hyperchromism (4%) (data not shown). In the case of conjugate
54TO¢8 involving two cyanines attached through octamethylene
linkers, more important changes were observed resulting in a
decrease (24%) in the intensity of the band at 478 nm and an
increase (34%) at 508 nm. For conjugate 64TO¢8.4 involving two
4TO¢ attached via linkers of different sizes, hybridization with
the double-stranded target 1 + 2 resulted in similar changes except
that the intensities were 18% hyprochromism at 481 nm and 20%
hyperchromism at 511 nm. Similar trends were also observed upon
hybridization of TFO 52TO8, resulting in a slight red shift (458 to
465 nm) with 13% hypochromism, and of TFO 524¢Q8, giving a
smaller red-shift (490 to 494 nm) with 20% hyperchromism. These
changes indicate that in the presence of the duplex, the interactions
between the two cyanine units are suppressed. Since a DNA triplex
is more rigid than a single-stranded sequence, it is likely that upon
hybridization of the conjugates of TFOs 4, 5 and 6 with the double-
stranded target the cyanines are able to interact with only the end
of the triplex and a few bases on the overhanging double-stranded
target adjacent to the triplex on the side of the cyanine attachment.
(DTm = +18 ◦C), 2TO¢8 (DTm = +15 ◦C), 4TO8 (DTm = +14 ◦C)
and the four quinocyanines provided stabilizations ranging from
DTm = +6.5 ◦C to DTm = +13.5 ◦C. Since it has previously
been shown that the attachment of intercalating groups at the 5¢-
end of TFOs via linkers involving five or six methylene groups
corresponds to the optimal size for the greatest stability, we
chose to attach the most efficient cyanine 4TO¢ via linkers of
decreasing lengths involving seven to four methylene groups.
Tm measurements indicated that the stability depended on the
linker size and decreased with the length of the linker. The DTm
was approximately 3 ◦C per methylene group. TFO conjugates
involving one label attached to the 5¢-end via an eight-methylene
linker form the most stable triplex structures.
Bis-labelled TFOs 5 and 6. In all cases the presence of a second
cyanine (conjugates 54TO¢8, 52TO8, 524¢Q8 and 64TO¢8.4) induced an
increase in the stabilization of the triplex as compared with that
obtained with the mono-conjugated TFOs. UV melting profiles
corresponding to the third strand dissociation of the triplex formed
by TFOs 54TO¢8 and 44TO¢8 and by the unmodified TFO 3 are shown in
Fig. 4. Stability depended on both the cyanine and the linker length
used to attach the second◦ cyanine to the 3¢-end. The strongest
stabilization (DTm = +32 C) was observed with 54TO¢8 involving
two thiazole orange labels attached via a linker of the same length
corresponding to the eight methylene groups. An increase in the
linker length to attach the cyanine to the 3¢-end of the TFO 64TO¢8.4
led to a slight decrease of the triplex stability (DTm = +30 ◦C). This
result is different from the one observed with the TFO labelled
with an acridine derivative at both the 5¢- and 3¢-ends.27 The
presence of the second intercalator 4TO8, 2TO8 or 24¢Q8 led
to similar stability increases (DTm = +7.5 ◦C, DTm = +6.5 ◦C,
Thermal denaturation studies. Experiments were carried out
by absorption spectroscopy. Triplex stabilities were determined
by thermal denaturation using a 1 : 1 mixture of the conjugates
and double-stranded target 1 + 2. The corresponding unmodified
TFO 3 was used as a reference. Two transitions were observed
in the melting of each triplex (Fig. 4). The transition with the
higher Tm corresponds to the melting of the target duplex (around
62 ◦C for all complexes) and the transitions with lower Tm to the
dissociation of the third strand. The Tm values, listed in Table 1,
indicate that the triplex structures formed between the conjugates
derived from TFOs 4–6 and the duplex 1 + 2 target were stabilized
in all cases as compared to the corresponding complexes formed
with the unmodified TFO 3 used as a reference.
◦
and DTm = +5 C respectively). This increase in the stability of
the triple-stranded structures provided by the presence of a second
intercalator was also previously reported for TFO–naphthalene33
and TFO–perylene conjugates.10
Fluorescence studies
Fluorescence emission spectra of the conjugates were compared
with those of the triplexes. Experiments were performed under the
same conditions (concentration, buffer) as for the Tm measure-
ments (see Experimental Section and Table 2). The lexc and lem
depended on the cyanine (Table 2). Upon hybridization with the
double-stranded target 1 + 2, the lem did not change as compared
with those observed for the free conjugates. The main changes
concerned the intensity of the fluorescence emission (Table 2 and
Fig. 5).
Free labelled TFOs. As can be seen from Table 2, the lem for
the mono-conjugated TFOs 4 range from 467 nm for 4Th8 to 576 nm
for 442¢Q8. Comparison of the fluorescence intensities observed for
the mono-conjugated TFOs 4 indicated that they depend on the
cyanine composition. For the same concentrations of conjugates,
the fluorescence intensity was higher for the thiocyanine derivative,
than for the thiazole orange isomers. The quinocyanines gave the
lowest signals. The presence of a second label at the 3¢-end of
the TFOs resulted only in fluorescence intensity changes. For the
conjugates 54TO¢8, 524¢Q8 and 64TO¢8.4, fluorescence intensities were
not very different from those of the corresponding mono-labelled
TFOs. These results can be explained by intra-molecular inter-
actions between the cyanines resulting in significant fluorescence
Fig. 4 UV melting profiles (recorded at l = 260 nm) of the triplex
structures formed by conjugates 44TO¢8 (A) and 54TO¢8 (B) and the unmodified
TFO 3 (C) in the presence of the duplex target (1 + 2). Concentrations
were 1 mM in the duplex and in the TFO, in a 10 mM sodium phosphate,
pH 6, buffer containing 140 mM KCl and 5 mM MgCl2.
TFOs involving one label at the 5¢-end. Different stabilizations
were observed depending on the cyanine structures and the linker
sizes. A comparison of the stabilization afforded by the cyanines
attached via the same linker involving eight methylene groups
indicated that the most efficient stabilizing group was 4TO¢8
(DTm = +24.5 ◦C) followed by 2TO8 (DTm = +19.5 ◦C), Th8
4418 | Org. Biomol. Chem., 2008, 6, 4413–4425
This journal is
The Royal Society of Chemistry 2008
©

View Article Online
Table 2 Fluorescence data a
b
Free conjugate
Triplex
em/nm 4 ◦C
F
Triplex/FTFO
TFOs
lex/nm 4 ◦
C
l
em/nm 4 ◦
C
Stokes shift
FTFO 4 ◦
C
l
FTriplex 4 ◦
C
4 ◦
C
20 ◦
C
44TO¢8
44TO¢7
44TO¢6
44TO¢5
44TO¢4
42TO8
4Th8
42TO¢8
44TO8
422¢Q8
424¢Q8
442¢Q8
54TO¢8
52TO8
524¢Q8
64TO¢8.4
511
511
513
512
511
491
429
490
510
527
561
560
512
459
519
513
527
528
527
528
526
544
467
549
526
574
572
573
528
579
583
528
16
17
14
16
15
53
38
59
16
47
11
13
16
120
64
16
4.81
4.71
7.70
4.14
3.09
1.42
26.26
0.90
3.44
0.22
0.70
0.33
5.86
4.24
0.58
3.00
527
528
528
527
525
531
453
550
527
540
570
576
529
529
570
528
58.32
62.78
56.92
40.12
53.12
17.88
89.26
10.07
24.09
5.14
22.75
5.69
54.77
28.93
7.77
12.12 (12.13)
13.33 (13.33)
7.39 (7.43)
24.47 (24.60)
27.63 (27.62)
14.96 (14.97)
15.36 (15.40)
32.20 (32.90)
18.29 (18.60)
5.31 (5.90)
15.57 (15.84)
12.73 (12.98)
24.07 (29.84)
37.84 (39.51)
16.11 (19.52)
18.32 (18.54)
7.70 (10.02)
7.97 (15.96)
6.63 (6.84)
9.69 (9.71)
17.19 (17.20)
12.59 (13.13)
3.40 (3.70)
11.19 (11.31)
7.00 (7.06)
23.36 (27.12)
32.50 (32.83)
17.24 (17.61)
9.35 (9.39)
6.82 (12.92)
13.40 (19.15)
5.42 (5.45)
16.26
a Emission spectra were recorded in a 10 mM sodium phosphate, pH 6, buffer containing 140 mM KCl and 5 mM MgCl2. Concentrations were 1 mM in
TFOs and the duplex target. b The values indicated are calculated from the intensities measured at the lem of the triplexes and the lem of the free labelled
TFOs. The values in the brackets correspond to the ratio of intensities measured at the lem of the triplexes.
Triplex structures. Upon formation of the triplex structures
by hybridization of the conjugates derived from TFOs 4, 5 and
6 with the double-stranded target, there were either no changes
or a very slight shift of the lem [except for the conjugate 52TO8
(50 nm)] together with a great increase in the fluorescent signal
(Table 2). Measurements were performed at 4 and 20 ◦C. The
intensity increases depended on both the cyanine and linker
size.
Mono-labelled TFOs. The most important intensity increases
upon triplex formation were observed at 20 ◦C. Among the
changes, a 37-fold increase was observed for 442¢Q8, 32-fold for
44TO¢4, 27-fold for 44TO¢7, and 24-fold for 44TO¢8. Emission spectra
of the free and hybridized TFOs 44TO¢4 and 44TO¢7 are shown in
Fig. 5. Under the same experimental conditions, the hybridization
of conjugate 44TO¢8 with its complementary single-stranded target
5¢GGGAAAAGAAAATTT3¢ resulted in only a 30% fluorescence
increase, clearly showing the great potential of our new probes
for the detection of double-stranded nucleic acid structures (data
not shown). It is important to note that when 4TO¢ was used as a
label, the most important fluorescence increase did not correspond
to the more stable triplex structure. In the case of the strongest
stabilization, it is likely that its intercalation allowed efficient
quenching of the label by the neighboring nucleobases. Conversely,
when attached by a shorter linker corresponding to four methylene
groups, it is possible that the label could not be fully intercalated.
In this case, although the structure formed could block the methine
bond rotation, it is possible that it prevented efficient quenching
by the nucleobases at the duplex–triplex junction. These results
agree with our previous studies obtained with 4TO¢-conjugated
oligonucleotides able to detect terminal mismatches on DNA
duplexes, with the mismatched duplexes being more fluorescent
than the perfect ones.26
Fig. 5 Fluorescence emission spectra of the free (a) and hybridized (b)
5¢-labelled TFOs 44TO¢8 (full line) and 44TO¢4 (dotted line) in a 10 mM sodium
phosphate buffer, pH 6, containing 140 mM KCl, and 5 mM MgCl2
recorded at 20 ◦C (bold line) and 4 ◦C (plain line). (Same concentrations
and buffer conditions as for Fig 3. lexc = 465 nm.) Hybridizations were
performed overnight.
quenching rates. Comparison of the fluorescence of the conjugates
54TO¢8 and 64TO¢8.4 indicates that the presence of the longer linker at
the 3¢-end results in greater quenching. It is possible that in this
case the interactions between both 4TO¢ units are favoured. These
results are in accordance with literature results concerning the bis-
labelling of hairpin oligonucleotides with Cy 5.5.34 For conjugate
52TO8, the presence of a second label resulted in an important red
shift of the lem (from 544 to 579 nm) and an increase in the
fluorescence intensity. A comparison of fluorescence intensities
for mono-conjugated TFOs 4 and bis-conjugated TFOs 5 and 6
Bis-labelled TFOs. For modified TFOs involving two cyanines
54TO¢8, 52TO8, 524¢Q8, and 64TO¢8.4, an increase in the fluorescence
◦
at 4 and 20 C indicated greater values at the lower temperature
(data not shown).
intensity signal was also observed upon hybridization with the
This journal is
The Royal Society of Chemistry 2008
Org. Biomol. Chem., 2008, 6, 4413–4425 | 4419
©

View Article Online
double-stranded target (18, 10, 15 and 6-fold, respectively). These
values are lower than those observed with the corresponding
mono-labelled TFOs.
Increasing the fluorescence intensity changes between the free
labelled TFO and the triplex structure. In order to increase the
difference in the fluorescence emission intensity between the triplex
structure and the free labelled TFOs, the possibility of using the
strand transfer strategy (Fig. 6) previously reported with single-
stranded targets was studied.35 This method is based on the
hybridization of the labelled oligonucleotidic probe with a shorter
sequence complementary to the 5¢-end, bearing a quencher at its
3¢-end. In the presence of the target sequence, strand displacement
occurs allowing the hybridization with the latter. In our study,
three labelled TFOs 44TO¢8, 44TO¢4 and 54TO¢8 were used. As quenching
sequences we used the 10-mer with the complementary sequence
to the TFO, 12dab [5¢GAAAAGAAAAp-(CH2)6NH-dabcyl3¢] as
well as shorter sequences 13dab [5¢AAAGAAAAp-(CH2)6NH-
dabcyl3¢], 14dab [5¢AAGAAAAp-(CH2)6NH-dabcyl3¢] and 15dab
[5¢AGAAAAp-(CH2)6NH-dabcyl3¢] complementary to the 5¢- end
of the TFO, bearing a dabcyl residue at their 3¢-ends (See Table 1
for characterizations). The quencher was attached by the reaction
of its succinimidyl ester with an aminohexyl linker incorporated
at the 3¢-ends of the oligonucleotides. Each of the labelled TFOs
was first hybridized with the series of oligonucleotide–dabcyl
conjugates. Fluorescence measurements indicated quenching (up
to at least 50%). Then, one equivalent of the double-stranded
target was added. Hybridizations were performed in the conditions
previously used for triplex formation.
Fig. 7 Fluorescence intensities for the free conjugates (hatched), the
quenched duplex (labelled TFO + 15dab) (black) and the quenched duplex in
the presence of the DNA target (1 + 2) (white). Concentrations and buffer
conditions were the same as for Fig. 3. Experiments were performed at
20 ◦C.
◦
◦
44TO¢8 + 15dab and 44TO¢4 + 15dab (23.5 C and 18 C, respectively),
and those of the corresponding triplexes (37.5 ◦C and 26.6 ◦C,
respectively).
Other sequence contexts
In order to test the ability of our labelling method to detect
triplex structures in different contexts, we have designed and
synthesized other TFOs labelled at their 5¢-end with TO¢ via
an eight methylene linker (Table 3). We chose, first, a 17-mer
pyrimidine TFO involving a cytosine at the 5¢-end 16. The
synthesis was performed as reported for the synthesis of the
conjugates of 4 (see Table 1 for characterizations). This TFO was
designed to hybridize with 17-contiguous purines on a double-
stranded DNA of 26 base-pairs (triplex I). Other double-stranded
targets were also chosen in order to allow mismatched triplex
formation. Mutations were incorporated on either the purine or
the pyrimidine strands at different positions to give either terminal
or internal mismatched triplexes (triplexes II to V). Another 5¢-
TO¢ labelled 17-mer pyrimidine TFO 17 was designed to bind to
a duplex target involving an AT base-pair inversion at an internal
position (triplex VI). On the basis of literature indicating that the
mismatched GTA triplet was among the least destabilizing,36–38
we chose to incorporate a guanine in the TFO in the position
facing the AT base-pair inversion. The stability provided by the
TO¢ labelling in these different contexts was evaluated by thermal
denaturation studies followed by absorption spectroscopy using
the corresponding unlabelled triplexes as references. Fluorescence
intensity changes observed upon hydridization were determined
by steady-state experiments.
Fig. 6 Schematic representation of the strand transfer from the duplex
(44TO¢8 + 15dab) (a) to the triplex structure (b). The concentrations were 1 mM
in each duplex and the buffer was the same as described in Fig. 3.
A large increase in the fluorescence emission was observed
when the target was added to the quenched duplex involving
the 6-mer-dabcyl conjugate 15dab. The fluorescence intensity of
the mixture reached 80% of that observed for the corresponding
triplex formed in the absence of 15dab after 4 h of incubation,
indicating the efficiency of the strand transfer. The experiments
were performed at 4 and 20 ◦C. When performed at 20 ◦C, the
strand transfer strategy resulted in a 47-fold fluorescence increase
with 44TO¢8 (Fig. 7). In such a strategy, the use of the 4TO¢ labelled
TFO involving the eight methylene linker forming the more stable
triplex structure was required for efficient strand transfer. This
result is in accordance with the Tm values observed for the duplexes
Absorption studies. The longer length of the duplex target al-
lowed us to distinguish between the triplex-to-duplex and the
duplex-to-single strand transitions. The Tm value for the fully
matched triplex I is 41.5 ◦C (Tm value for the fully matched
duplex is 69.5 ◦C). The presence of mismatches in the triplex
structures induced different effects that depended on both their
position on the triplexes (terminal versus internal position) and
the strand where mutations were incorporated (pyrimidine versus
purine strand). In agreement with the literature,36–38 our results
indicated that the incorporation of a purine base in the pyrimidine
strand, which affects only the Watson and Crick (WC) base-pair
4420 | Org. Biomol. Chem., 2008, 6, 4413–4425
This journal is
The Royal Society of Chemistry 2008
©

View Article Online
Table 3 Tm and fluorescence data for the 17-mer triplexes I–VI
formation, had either no (triplex II) or only a moderate effect
on the triplex stabilities (triplex III) while the incorporation of a
pyrimidine base in the purine strand had a greater destabilizing
effect (triplexes IV and V). In addition to its effect on the Watson
and Crick (WC) base-pair formation, this kind of modification also
altered the Hoogsteen base-pairing. In this case, the stability of the
triplex ◦involving a mismatch at the end (triplex IV) is decreased
by 2.5 C while no triplex was detected, under the experimental
conditions used, when a pyrimidine was incorporated at an
internal position. The hybridization of the 17-mer TFO 17,
containing a guanine, with the duplex involving an AT base-pair
inversion also resulted in the formation of a triplex (triplex VI) but
with decreased stability (by 8.5 ◦C) as compared with that of the
perfectly matched one (triplex I). In all cases, labelling the TFOs
with the intercalator thiazole orange TO¢ induced the formation of
triplexes with increased stabilities as compared with the unlabelled
ones. However, the stability increase due to labelling of the TFOs
was reduced when the size of the triplex was increased from 10
to 17 base-triplets. Similar trends were also previously observed
with other intercalators such as acridine and perylene.10,19,25
A
comparison of the stabilities of the labelled triplexes with those
of the corresponding unlabelled ones clearly showed that the
binding specificity was not reduced due to the presence of the sta-
◦
bilizing label. A stability increase of +6 C was observed for the
fully matched triplex (I) and +5.5 ◦C for the triplex involving the
AT base-pair inversion (VI). Similar stability increases were also
observed for triplexes II and III involving the terminal mismatches
and the triplex with the internal mutation on the pyrimidic strand
(triplex IV). When the mutation was incorporated in the purine
strand (triplex V) the presence of the label induced the formation of
a triplex structure but with low stability (Tm = 26 ◦C) as compared
with those of the other triplexes.
Fluorescence studies. A comparison of the fluorescence inten-
sities emitted by the free labelled TFOs indicated that they depend
on both their length and base composition. The fluorescence of
the 17-mer pyrimidine TFO 164TO¢8 was 1.4-fold higher than that
of the 10-mer TFO 4TO¢8. The replacement of a thymine by a
This journal is
The Royal Society of Chemistry 2008
Org. Biomol. Chem., 2008, 6, 4413–4425 | 4421
©

View Article Online
guanine in the TFO 174TO¢8 led to a further increase (about 2-fold)
in the fluorescence emission. This can be explained by the back-
binding effect previously reported for thiazole orange labelled
peptide nucleic acids and oligonucleotides.39,14 The fluorescence
emissions of the purine rich oligonucleotides labelled with thiazole
orange were also higher than those in the pyrimidine series
due to stronger stacking interactions between the purine rings
and the thiazole orange.39–41 Furthermore, in the case of TFO
174TO¢8, the concomitant presence of the intercalating molecule
and the possibility of an internal GC base-pair formation was an
additional factor in favor of the back-binding structure. Upon an
increase in temperature, a decrease in the fluorescence emission
was observed (Table 3 and data not shown). Upon hybridization
of the labelled TFOs with the different duplexes, fluorescence
emission increases were observed◦in all cases. A comparison of the
were greater at a lower temperature (4 ◦C versus 25 ◦C). However,
the following observations can be made. At 4 ◦C, the fluorescence
intensities of triplexes I, III and VI were nearly identical while
that of triplex V with weak stability, was weaker. The weakest
fluorescence intensities were observed for the triplexes II and
IV involving mismatched triplets in the position adjacent to the
intercalation site at the duplex to triplex junction. When the
experiments were performed at 25 ◦C, the fluorescence intensities
were decreased by a third for all the triplexes except for triplex V for
which a more important reduction was observed (more than two
thirds). The resulting fluorescence intensity ratio at 25 ◦C, between
the triplex structures and the labelled TFOs, was ~13.7 fold for
fully matched triplex I and triplex III involving a mutation at an
internal position on the pyrimidine strand. Due to the presence of
the terminal mismatches, the ratio was slightly weaker for triplexes
II and IV (9.9-fold and 8.5-fold, respectively, Fig. 8) while it was the
weakest for triplex V with the mutation at an internal position of
the purine strand. For triplex VI with the AT base-pair inversion,
the ratio was only 6.5-fold due to the high fluorescence intensity of
the free labelled TFO 174TO¢8. These results confirm the possibility
of detection of double-stranded nucleic acid structures by triple-
helix formation with thiazole orange 5¢-labelled TFOs.
Conclusion
We have reported the possibility of detecting double-stranded
nucleic acid sequences by hybridization of “light-up” TFOs
involving fluorescent labels with intercalating properties. These
labels are monomethine cyanines. One thiocyanine, four thiazole
orange isomers and three quinocyanines were linked to the 5¢-
end of a 10-mer pyrimidine TFO. The binding properties of these
labelled TFOs with the double-stranded nucleic acid target, stud-
ied by absorption and steady-state fluorescence spectroscopies,
indicated that strong stability can be obtained by choosing a
convenient linker size. Stability also depends on the label structure.
The presence of one benzothiazole ring is required to obtain
good stability. The linkage of a second label at the 3¢-end of
the TFOs provides additional stability. Steady-state fluorescence
experiments showed an important intensity increase upon triplex
formation. While the most important fluorescence increase was
observed with a quinocyanine (37-fold), the conjugates involving
thiazole orange attached by the benzothiazole ring provided the
most balanced properties in terms of triplex stabilization, fluores-
cence intensity and fluorescence enhancement upon hybridization
with the double-stranded target. It is also possible to increase
the fluorescence emission upon triplex formation by quenching
the fluorescence of the labelled TFO by its hybridization with
a shorter sequence complementary to the 5¢-end of the TFO
and bearing a quencher at its 3¢-end. Due to the stabilizing
properties of the labels, strand displacement occurs resulting in
the formation of a triplex structure. Formation of the triplex is
accompanied by a strong fluorescence enhancement (up to about
47-fold). The formation of longer fully matched triplex structures
(17 triplets) also induced an important fluorescence increase (13-
fold), while the presence of a mismatched base-pair on the duplex,
at the duplex–triplex junction, resulted in a reduced fluorescence
increase. A 17-mer TO¢-labelled TFO involving a guanine facing
an AT inversion on the duplex target also formed a triplex structure
detectable by a large fluorescence enhancement (6.5-fold). These
new “light up” TFOs (LU-TFOs) will be useful tools for the
detection of specific sequences on double-stranded nucleic acids.
◦
results obtained at 4 C and 25 C indicated that the intensities
Experimental
General methods
All solvents used were of the highest purity and did not contain
more than 10 ppm H2O. All chemicals were used as obtained
unless otherwise stated. During all the syntheses and purification
steps cyanine-linker derivatives and TFO–cyanine conjugates were
protected from light. Analytical thin-layer chromatography (TLC)
was performed on precoated alumina plates (Merck silica gel 60F
254 ref. 5554). For flash chromatography, Merck silica gel 60 (40–
63 mM) (ref. 9385 from Merck) was used. NMR spectroscopy
was performed on a Varian Unity 500 spectrometer. 1H chemical
shifts were referenced to Me4Si. 1H NMR coupling constants are
reported in Hz and refer to apparent multiplicities. Mass analysis
was performed on a Quattro II (Micromass) instrument. TFOs
were synthesized using cyanoethyl phosphoramidite chemistry
Fig. 8 Fluorescence emission spectra of the free 5¢-labelled TFO 16 (a);
fully matched triplex I (b) and end-mismatched triplexes II (c) and IV (d)
in a 10 mM sodium phosphate buffer, pH 6, containing 140 mM KCl, and
5 mM MgCl2 recorded at 25 ◦C (same concentrations and buffer conditions
as for Fig 3. lexc = 465 nm.) Hybridizations were performed overnight.
4422 | Org. Biomol. Chem., 2008, 6, 4413–4425
This journal is
The Royal Society of Chemistry 2008
©

View Article Online
and an Expedite nucleic acid synthesis system 8909 from Perseptive
Biosystems. Reversed-phase chromatography and purification
were performed on a 600 E (System Controller) equipped with
a photodiode array detector (Waters 990) using a LiChrospher
100 RP 18 (5 mm) column (125 mm ¥ 4 mm) from Merck.
Analyses and purifications by ion-exchange chromatography were
carried out on a Pharmacia FPLC with a MonoQ (8 mm, 10 ¥
100 mm, Pharmacia). UV spectra were recorded on an Uvikon
860 spectrophotometer. Fluorescence spectra were recorded on
a Fluoromax 2 (ISA-Jobin-Yvon) spectrofluorimeter in 0.5 cm
path-length Suprasil quartz cuvettes (Hellma) with slits set at
0.5 mm (band pass = 2 nm). Oligonucleotides 1, 2, 3 and the 26-
mer oligonucleotides were from Eurogentec. Their concentrations
were calculated using molar extinction coefficients at 260 nm,
determined by the nearest-neighbor model.42
A
solution of 4,4¢-O-dimethoxytritylchloride (550 mg,
1.65 mmol, 0.27 eq.) in pyridine (20 cm3) was then added to
6,6¢-dithiodihexanol (1.65 g, 1 eq., 6.2 mmol) previously dried
by coevaporation with anhydrous pyridine. After 5 h of stirring
at rt, 4,4¢-O-dimethoxytritylchloride (0.1 eq., 51 mg, 0.15 mmol)
was added and the stirring was continued overnight. The pyridine
was azeotroped with toluene and the residue purified on a silica
gel column using a MeOH gradient (0 to 3%) in CH2Cl2 to give
10 as yellow oil (950 mg, 1.6 mmol, 95%). Rf14 = 0.55 (CH2Cl2–
1
MeOH, 90 : 10, v/v). H NMR (500 MHz, CDCl3, TMS) d =
7.45 (2H, d, 4JH-H = 7.3 Hz, -ODMT), 7.33 (4H, d, 4JH-H = 8.9 Hz,
3
-ODMT), 7.29 (2H, m, -ODMT), 7.21 (1H, t, JH-H = 7.1 Hz,
4
-ODMT), 6.83 (4H, d, JH-H = 8.9 Hz, -ODMT), 3.80 (6H, s,
-ODMT), 3.64 (2H, t, 3JH-H = 6.5 Hz, -CH2ODMT), 3.05 (2H, t,
3JH-H. = 6.5 Hz, -CH2OH), 2.68 (4H, dd, 3JH-H = 7.5 Hz, 3JH-H
=
9 Hz, -CH2-SS-CH2-), 1.68 (4H, m, -CH2CH2S-SCH2CH2-),
1.60 (4H, m, -CH2CH2OH + -CH2CH2ODMT), 1.40 [8H, m,
-(CH2)n-]. ESI-MS: m/z, C33H44O4S2, calc. 568.84, found 591.73
(M + Na+). 1-O-Dimethoxytrityl-6,6¢-dithiodihexanol 10 (900 mg,
1.6 mmol, 1 eq.), succinic anhydride (159 mg, 1.59 mmol, 0.95 eq.)
and 4-dimethylaminopyridine (183 mg, 1.5 mmol, 0.9 eq.) were
solubilized with anhydrous pyridine (12 cm3) and the mixture
was stirred for 24 h at rt. The pyridine was azeotroped with
toluene. The residue was washed with a 10% aqueous citric acid
solution (10 cm3) and H2O (10 cm3) then extracted with CH2Cl2
(3 ¥ 20 cm3). The organic phase was dried over magnesium
sulfate and concentrated. Brown oil. (1.04 g, 1.52 mmol) Yield:
95%. Rf = 0.35 (CH2Cl2–MeOH, 90 : 10, v/v). ESI-MS: m/z,
C37H48O7S2, calc. 668.90, found 667.49 (M-). Functionalization
of the support was then performed as previously reported31 by
replacing the 1-O-dimethoxytrityloxyethyl-1¢-succinyl-ethyl-2,2¢-
disulfide with 1-O-dimethoxytrityl-6,6¢-dithiodihexanol 10 and
Synthesis
Synthesis of the cyanine-linker derivatives. The syntheses of
cyanine-linker derivatives 4TO¢-(CH2)n-I (n = 4, 5, 6, 7 and
8), 2TO-(CH2)8-I, 2Th-(CH2)8-I, 2TO¢-(CH2)8-I, 4TO-(CH2)8-I,
22¢Q-(CH2)8-I, 24¢Q-(CH2)8-I, and 42¢Q-(CH2)8-I were performed
as previously described.14,26
Synthesis of the 5¢-thiophosphorylated TFO 4 (Scheme 1a). The
TFO was assembled at the 1 mmol scale. At the end of the chain
assembly an additional detritylation step was performed followed
by the incorporation of the thiophosphate group according to our
previously reported procedure.30 TFO 4 was purified by reversed-
phase chromatography using a linear gradient of CH3CN (5 to
50% over 60 min) in 0.1 M aqueous ammonium acetate, pH 7,
with a flow rate of 1 cm3 min-1.
Synthesis of the bis 3¢-,5¢-bis-thiophosphorylated TFO
5
˚
aminopropyl fractosil with LCCA-CPG 500 A (Chemgenes).
Loading 36 mmol g-1.
(Scheme 1b). The synthesis was performed at the 1 mmol scale
on our previously reported modified support 8 leaving, after the
deprotection step, a thiophosphate group at the 3¢-end.31 At the
end of the chain assembly an additional detritylation step and
the incorporation of the thiophosphate group were performed
following our previously reported procedure.30 TFO 5 was purified
by ion-exchange chromatography using a linear gradient of NaCl
in a 25 mM Tris–HCl buffer, pH 7, containing 10% MeOH.
Yield (30%).
Labelling of 5¢- and bis 3¢-,5¢-thiophosphorylated TFOs 4, 5, 16
and 17. DMF solutions of the selected cyanine-linker derivatives
(1.5 mg in 0.3 cm3, 10 eq.) were added to vortexed solutions
of 5¢-thiophosphorylated TFOs 4, 5, 16 and 17 (10 OD each)
in a 2.5% aqueous bicarbonate buffer, pH = 9, (0.2 cm3). The
mixture was stirred for 4 to 10 h at rt. The coupling efficiency was
monitored by reversed-phase analysis using the same conditions
as described above. TFO–cyanine conjugates were obtained with
increased retention times as compared to those of the starting
TFOs 4 or 5. During the entire work-up, conjugate solutions were
protected from light. Mixtures were evaporated to dryness under
reduced pressure. The residue was dissolved with water (2 cm3)
and extracted several times with CH2Cl2. The aqueous phase was
passed over a size exclusion Sephadex G 10 column, using water as
the eluant. Fast eluting colored fractions were purified by reversed-
phase chromatography with a linear gradient of CH3CN (5% to
50% over 60 min for mono-labelled conjugates and bis-labelled
conjugates) in 0.1 M aqueous ammonium acetate, pH 7, with a
flow rate of 1 cm3 min-1. (See Table 1 for retention times and
mass analysis data.) Yields were: 4Th8 (37%), 42TO¢8 (32%), 44TO¢8
(30%), 42TO8 (27%), 44TO8 (35%), 422¢Q8 (48%), 424¢Q8 (28%), 442¢Q8
(40%), 54TO¢8 (27%), 52TO8 (40%), 524¢Q8 (25%), 164TO¢8 (25%) and
174TO¢8 (20%). Electrospray mass analysis confirmed the mass of
all the conjugates.
Synthesis of the 3¢-thiolated-,5¢-thiophosphorylated TFO 6
(Scheme 2). The synthesis was performed at the 1 mmol scale
on modified support 11 described below. At the end of the chain
assembly an additional detritylation step and the incorporation of
the thiophosphate group at the 5¢-end were performed as described
above. TFO 6 was purified by ion-exchange chromatography using
the conditions reported above. Yield (31%).
The synthesis of the modified support involving a masked thiol
function 11 was obtained as follows. A solution of thiohexanol
9 (5 g, 37.25 mmol) in EtOH–NH4OH (1 : 2, v/v), (30 cm3) was
stirred with air bubbling for ten days with addition of NH4OH
(2 cm3) every two days to give the 6,6¢-dithiodihexanol. The
mixture was concentrated to dryness and then purified on a silica
gel column using a MeOH gradient (0 to 3%) in CH2Cl2 to give a
white powder (1.7 g, 6.4 mmol, 35%). Rf = 0.30 (CH2Cl2–MeOH,
90 : 10, v/v). ESI-MS: m/z, C12H26O2S2, calc. 266.46, found
265.40 (M-).
This journal is
The Royal Society of Chemistry 2008
Org. Biomol. Chem., 2008, 6, 4413–4425 | 4423
©

View Article Online
Bis-labelling of the 3¢-thiolated,5¢-thiophosphorylated TFO 6.
5¢-Labelling was performed as described for the labelling of TFOs
4, and 5 using 20 OD of TFO 6 and the 4TO¢-(CH2)8-I cyanine-
linker derivative. After purification of the 5¢-labelled TFO by
reversed-phase chromatography, the organic solvent was removed
by evaporation and the buffer by lyophilization. The solution of
5¢-labelled TFO 6 in a 2.5% aqueous bicarbonate buffer, pH =
9, (0.2 cm3) was degassed for 15 min by a stream of argon.
Then, a degassed TCEP32 solution (5 mg in 0.75 cm3) in a 2.5%
aqueous bicarbonate buffer, pH = 9, (0.02 cm3) was added. After
10 min, a degassed DMF solution of 4TO¢-(CH2)4-I (1.5 mg in
0.3 cm3) was added and the mixture vigorously vortexed. After 2 h
of reaction, the coupling efficiency was monitored by reversed-
phase analysis using the same conditions as described above.
During the entire work-up, the conjugate solutions were protected
from light. Mixtures were evaporated to dryness using a toluene–
DMF azeotrope. The residue was dissolved with a 1 M NaCl
solution in H2O–MeOH (80 : 20, v/v) (2 mL) and extracted
several times with CH2Cl2. The aqueous phase was passed over
a size exclusion Sephadex G10 column, using water as the eluant.
Fast eluting colored fractions were purified by reversed-phase
chromatography using the conditions described above. (See Table 1
for retention times and mass analysis data.) The bis-conjugated
64TO¢8.4 was purified by reversed-phase chromatography using the
conditions described vide supra. Yield (5%).
determined by titrating the conjugate solutions in a 10 mM
sodium phosphate, pH 6, buffer containing 100 mM NaCl at 5 ◦C
with a solution of the single-stranded complementary sequence
5¢GGGAAAAGAAAATTT3¢. The molar extinction coefficients
(e) for oligonucleotide–dabcyl conjugates 12dab, 13dab, 14dab and
15dab were determined by titrations with the complementary
sequence 3.
Tm measurements. The Tm values for the triplex formed by the
mono- and bis-labelled TFOs and the double-stranded target 1 + 2
were determined by thermal denaturation followed by absorption
spectroscopy using 1 mM solutions of TFOs and 1 mM of the
double-stranded duplex in a 10 mM KH2PO4 buffer (pH 6),
containing 140 mM KCl and 5 mM MgCl2. The samples were
left to hybridize overnight at 4 ◦C (far below the Tm value) in the
dark. Results are given in Table 1. The uncertainty in the Tm values
reported was 1 ◦C.
Fluorescence studies
Fluorescence excitation and emission spectra were recorded on
a Fluoromax 2 (ISA-Jobin-Yvon) spectrofluorimeter in 0.5 cm
path-length Suprasil quartz cuvettes (Hellma) with slits set at
0.5 mm (band pass = 2 nm). A 1 mM solution of labelled TFOs
was prepared in a 10 mM KH2PO4 buffer (pH 6), containing
140 mM KCl and 5 mM MgCl2 The triple-stranded samples were
prepared by the addition of a small volume (0.005 cm3) of a
concentrated solution of the target (1 eq.) at rt. The sample was left
to hybridize overnight at 4 ◦C, 20 ◦C or 25 ◦C in the dark before the
measurements were performed. The fluorescence emission spectra
of free and hybridized labelled TFOs were recorded between l =
450 and 700 nm. Errors in fluorescence values were estimated to
be 10%.
Synthesis of the ON–dabcyl conjugates. ONs involving an
aminohexyl linker at the 5¢-end were synthesized at the 1 mmol scale
on a previously reported modified support.43 Deprotection and
purification was then performed by ion-exchange chromatography
as reported above. A DMF solution of dabcyl succinimidyl ester
(1.5 mg in 0.3 cm3, 10 eq.) was added to vortexed solutions of 3¢-
aminohexyl-ONs (10 OD each) in a 2.5% aqueous bicarbonate
buffer, pH = 9 (0.2 cm3). The mixture was stirred for 4 h
at rt. The coupling efficiency was monitored by reversed-phase
analysis using the same conditions as described above. ON–dabcyl
conjugates 12dab, 13dab, 14dab and 15dab were obtained with increased
retention times as compared with the starting 3¢-aminohexyl-
oligonucleotide conjugates. During the entire work-up, conjugate
solutions were protected from light. Mixtures were evaporated
to dryness under reduced pressure. The residue was dissolved
with water (2 cm3) and extracted several times with CH2Cl2. The
aqueous phase was passed over a size exclusion Sephadex G 10
column using water as the eluant. Fast eluting colored fractions
were purified by reversed-phase chromatography with a linear
gradient of CH3CN (5% to 50% over 60 min) in 0.1 M aqueous
ammonium acetate, pH 7, with a flow rate of 1 cm3 min-1. (See
Table 1 for retention times and mass analysis data.) Yields were
12dab (60%), 13dab (55%), 14dab (70%) and 15dab (75%). Electrospray
mass analysis confirmed the mass of all conjugates.
Acknowledgements
We thankthe RegionCentre for adoctoralfellowshiptoBrice-Lo¨ıc
Renard. We thank the “Plateforme de Spectrometrie de masse” of
the “Centre de Biophysique Mole´culaire” for analysis facilities,
C. Bure´ and G. Gaband for running the mass spectrometers and
H. Meudal for recording the NMR spectra.
References
1 A. Okamoto, Y. Saito and I. Saito, Photochem. Photobiol., C, 2005, 6,
108.
2 R. Ranasinghe and T. Brown, Chem. Commun., 2005, 5487.
3 U. Asseline, Curr. Org. Chem., 2006, 10, 491.
4 J. N. Wilson and E. T. Kool, Org. Biomol. Chem., 2006, 4, 4265.
5 K. E. Sapsford, L. Berti and I. L. Medintz, Angew. Chem., Int. Ed.,
2006, 45, 4562.
6 A. A. Marti, S. Jockusch, N. Stevens, J. Ju and N. J. Turro, Acc. Chem.
Res., 2007, 40, 402.
7 J. L. Mergny, T. Garestier, M. Rouge´e, A. V. Lebedev, M. Chassignol,
N. T. Thuong and C. He´le`ne, Biochemistry, 1994, 33, 15321.
8 T. Inoue, Y. Sugiura, J. Saitoh, T. Ishiguro and M. Otsuka, Bioorg. Med.
Chem., 1999, 7, 1207.
9 G. N. Grimm, A. S. Boutorine, P. Lincoln, B. Norden and C. He´le`ne,
ChemBioChem, 2002, 3, 324.
10 Y. Aubert and U. Asseline, Org. Biomol. Chem., 2004, 2, 3496.
11 H. Kuhn, Antisense Nucleic Acid Drug Dev., 2001, 11, 265.
12 J. S. Sun and C. He´le`ne, Nucleosides, Nucleotides Nucleic Acids, 2003,
22, 489.
Absorption studies
Molar absorption coefficient determination. The molar absorp-
tion coefficients (e260nm values) for the conjugates derived from
TFOs 4, 16 and 17 were estimated to be the sum of the e values
of the TFO and cyanines deducted from ref. 14 (see Table 1).
The same e260nm values were used for the series of conjugates
44TOn involving different linker sizes (n = 4, 5, 6, 7 and 8).
The molar extinction coefficients (e) for conjugates of 5 were
4424 | Org. Biomol. Chem., 2008, 6, 4413–4425
This journal is
The Royal Society of Chemistry 2008
©

View Article Online
13 E. Brunet, M. Corgnali, L. Perrouault, V. Roig, U. Asseline, M. D.
Sorensen, B. R. Babu, J. Wengel and C. Giovannangeli, Nucleic Acids
Res., 2005, 33, 4223.
14 R. Lartia and U. Asseline, Chem.–Eur. J., 2006, 12, 2270.
15 L. G. Lee, C.-H. Chen and L. A. Liu, Cytometry, 1986, 7, 508.
16 J. Nygren, N. Svanvik and M. Kubista, Biopolymers, 1998, 46, 39.
17 S. Prodhomme, J.-P. Demaret, S. Vinogradov, U. Asseline, L. Morin-
Allory and P. Vigny, J. Photochem. Photobiol., B, 1999, 53, 60.
18 A. Larsson, C. Carlsson, M. Jonsson and B. Albinsson, J. Am. Chem.
Soc., 1994, 117, 8459.
28 N. T. Thuong and U. Asseline, in Current Protocols in Nucleic Acid
Chemistry. ed. S. Beaucage, D. E. Bergstrom, G. D. Glick and
R. A. Jones. John Wiley & Sons, New York, 2000, pp. 4.2.1–4.2.34.
29 S. L. Beaucage and M. H. Caruthers, Tetrahedron Lett., 1981, 22,
1859.
30 R. Lartia and U. Asseline, Tetrahedron Lett., 2004, 45, 949.
31 U. Asseline, E. Bonfils, R. Kurfu¨rst, M. Chassignol, V. Roig and N. T.
Thuong, Tetrahedron, 1992, 48, 1233.
32 J. A. Burns, J. C. Butler, J. Moran and G. M. Whitesides, J. Org. Chem.,
1991, 56, 2648.
19 U. Asseline and E. Cheng, Tetrahedron Lett., 2001, 42, 9005.
20 S. Vinogradov, V. Roig, Z. Sergueeva, C.-H. Nguyen, P. Arimondo, J.-S.
Sun, N. T. Thuong, E. Bisagni, C. He´le`ne and U. Asseline, Bioconjugate
Chem., 2003, 76(14), 120.
21 D. A. Collier, J.-L. Mergny, N. T. Thuong and C. He´le`ne, Nucleic Acids
Res., 1991, 19, 4219.
33 D. A. Gianolio, J. M. Segismundo and L. McLaughlin, Nucleic Acids
Res., 2000, 28, 2128.
34 V. Metelev, R. Weissleder and A. Bogdanov Jr, Bioconjugate Chem.,
2004, 15, 1481.
35 Q. Li, G. Luan, Q. Guo and J. Liang, Nucleic Acids Res., 2002,
30, e5.
22 U. Asseline, N. T. Thuong and C. Helene, New J. Chem., 1997, 21, 5.
23 J. S. Sun, C. Giovannangeli, J. C. Franc¸ois, R. Kurfu¨rst, T. Montenay-
Garestier, U. Asseline, N. T. Thuong and C. He´le`ne, Proc. Natl. Acad.
Sci. U. S. A., 1991, 88, 6023.
24 U. Asseline, E. Bonfils, D. Dupret and N. T. Thuong, Bioconjugate
Chem., 1996, 7, 369.
25 U. Asseline, M. Delarue, G. Lancelot, F. Toulme, N. T. Thuong, T.
Montenay-Garestier and C. Helene, Proc. Natl. Acad. Sci. U. S. A.,
1984, 81, 3297.
36 J.-S. Sun, J.-L. Mergny, R. Lavery, T. Montenay-Garestier and C.
He´le`ne, J. Biomol. Struct. Dyn., 1991, 9, 411.
37 J.-L. Mergny, J.-S. Sun, M. Rouge´e, T. Montenay-Garestier, F. Barcelo,
J. Chomilier and C. He´le`ne, Biochemistry, 1991, 30, 9791.
38 D. A. Rusling, T. Brown and K. R. Fox, Biophys. J., 2006, 123, 134.
39 N. Svanvik, J. Nygren, G. Westman and M. Kubista, J. Am. Chem. Soc.,
2001, 123, 803.
40 E. Privat, T. Melvin, U. Asseline and P. Vigny, Photochem. Photobiol.,
2001, 74, 532.
26 U. Asseline, M. Chassignol, Y. Aubert and V. Roig, Org. Biomol. Chem.,
2006, 4, 1949.
27 T. Montenay-Garestier, J. S. Sun, J. Chomilier, J. L. Mergny, M.
Takasugi, U. Asseline, N. T. Thuong, M. Rougee and C. Helene,
in “Molecular basis of specificity in nucleic acid-drug interactions”
ed. B. Pullman and J. Jortner, Kluwer Academic Publishers, The
Netherlands, 1990, p. 275.
41 E. Privat, T. Melvin, F. Merola, G. Schweizer, S. Prodhomme, U.
Asseline and P. Vigny, Photochem. Photobiol., 2002, 75, 201.
42 C. R. Cantor, M. M. Warshaw and H. Shapiro, Biopolymers, 1970, 9,
1059.
43 Y. Aubert, S. Bourgerie, L. Meunier, R. Mayer, A.-C. Roche, M.
Monsigny, N. T. Thuong and U. Asseline, Nucleic Acids Res., 2000,
28, 818.
This journal is
The Royal Society of Chemistry 2008
Org. Biomol. Chem., 2008, 6, 4413–4425 | 4425
©