Synthesis and hybridization properties of oligonucleotide-perylene conjugates: Influence of the conjugation parameters on triplex and duplex stabilities

Article abstract of DOI:10.1039/b410695d

We report here the synthesis of oligo-2′-deoxyribonucleotides (ODNs) conjugated with perylene. Introduction of perylene, coupled either directly or via a propyl linker to the anomeric position of a 2′-deoxyribose residue, induces the formation of two anom

Full text of DOI:10.1039/b410695d

View Online / Journal Homepage / Table of Contents for this issue
A R T I C L E
Synthesis and hybridization properties of oligonucleotide–perylene
conjugates: influence of the conjugation parameters on triplex and
duplex stabilities†
Yves Aubert and Ulysse Asseline*
Centre de Biophysique Mole´culaire CNRS, UPR 4301, affiliated with the University of Orle´ans
and with INSERM. Rue Charles Sadron, Cedex 02, Orle´ans, France 45071.
E-mail: asseline@cnrs-orleans.fr; Fax: +33-2-38-63-15-17
Received 15th July 2004, Accepted 4th October 2004
First published as an Advance Article on the web 3rd November 2004
We report here the synthesis of oligo-2ꢀ-deoxyribonucleotides (ODNs) conjugated with perylene. Introduction of
perylene, coupled either directly or via a propyl linker to the anomeric position of a 2ꢀ-deoxyribose residue, induces
the formation of two anomers. Single incorporations of each pure anomer of these sugar–perylene units have been
performed at either the 5ꢀ-end or an internal position of a pyrimidic pentadecamer. The binding properties of these
modified ODNs with their single- and double-stranded DNA targets were studied by absorption spectroscopy.
Double incorporations of the sugar–perylene unit most efficient at stabilizing the triplex and duplex structures (the
b-anomer involving the propyl linker) have been performed at both the 5ꢀ-end and at an internal position (or both the
5ꢀ- and 3ꢀ-ends) of the ODN chain. Comparison has been made with ODN–perylene conjugates involving either one
or two perylenes attached via a longer polymethylene chain to either the 5ꢀ- or 3ꢀ- (or both the 5ꢀ-and 3ꢀ-) terminal
phosphate groups. The ODNs involving two perylenes are more efficient at stabilizing the triplex and the duplex
structures than the ODNs involving only one perylene and, among these, the ODN–perylene conjugate involving two
sugar–perylene units attached at both termini is the most efficient. The results of the fluorescence studies have shown
an important increase in the intensity of the fluorescent signal upon hybridization of the ODNs involving two
perylenes with either the single- or the double-stranded targets. This increase in the intensity of the fluorescent signal
could be used as proof of the hybridization.
ODN–perylene conjugates by covalently linking perylene to the
Introduction
anomeric position of a 2ꢀ-deoxyribose, that could be incorpo-
Synthetic oligonucleotides (ODNs), which are widely used in a
rated at various positions inside the sequence. In a preliminary
great variety of areas such as biotechnology and diagnostics, of-
report we showed that when linked to the anomeric position
fer interesting prospects for the regulation of gene expression.1–5
Progress made in the genomic sequencing of an increasing
of a 2ꢀ-deoxyribose via a propynyl linker, the perylene stabilizes
both the duplex and the triplex structures.25 In order to further
number of organisms, including the human genome, opens a new
explore the possibility of increasing the duplex and triplex
application for modified ODNs as tools for functional genomics
stabilities, we chose to incorporate new sugar–perylene units,
involving the perylene attached either directly or via a propyl
and for validating targets.6,7 To be efficient in regulating gene
expression, antisense and anti-gene ODNs must efficiently cross
linker to the anomeric position of the sugar residue, into a
cell membranes, be resistant to nucleases and specifically and
pyrimidic pentadecamer and to compare the binding properties
strongly hybridize to their targets on mRNA or double-stranded
DNA. Numerous ODN analogs involving a modified backbone
of these new ODNs with those of the corresponding ODNs
involving the perylene attached via a longer polymethylene
linker to the terminal phosphodiester groups. The influence
or nucleic acid bases conjugated, or not, to various ligands
have been engineered in order to fulfill these requirements.8–12
of the incorporation of two perylene units per ODN on the
Despite these efforts, the stability of triplexes has needed further
duplex and triplex stabilities was also studied. We report here
improvement in many cases. One way to stabilize the triplexes
the preparation of these new ODN–perylene conjugates, as
stems from the covalent attachment of intercalators.13–21
A
well as their binding properties with complementary single-
and double-stranded DNA target sequences which were studied
using absorption and fluorescence spectroscopies.
perylene derivative has been used to bridge the two pyrimidine
strands of DNA in a pyrimidine–purine–pyrimidine triplex.22
The planar linker involving seven rings provides interactions
with all three base residues. To test the ability of the perylene
derivative involving five fused six-membered rings to stabilize
duplex and triplex structures, we previously linked it to the
5ꢀ- and the 3ꢀ-ends of a pyrimidine decamer via the terminal
phosphates.23 The results showed strong stabilization of the
triplex when the perylene was attached to the 5ꢀ-end of the
sequence. Since a survey of the literature together with the results
of our previous work clearly showed that the properties of ODN–
intercalator conjugates were dependent on the parameters of the
linkage between both entities,16–18,21,24 we chose to prepare other
Results and discussion
Perylene was covalently linked to a 15-mer oligopyrimidic
sequence d-5ꢀ TTCTTTTTCTTCTCT3ꢀ via the anomeric position
of a 2ꢀ-deoxyribose unit without a linker, or by the intermediate
of a trimethylene linker. Parameters such as the anomeric
configuration and the incorporation position of the sugar–
perylene units inside the sequence were investigated. To study the
influence of the parameters of the linkage between the ODN and
the perylene, the latter was also linked to the terminal phosphates
of the ODN via polymethylene linkers. Double incorporations of
the perylene were also performed. We report here the synthesis of
these ODN–perylene conjugates and the results of their binding
properties with both their double- and single-stranded DNA
† Electronic supplementary information (ESI) available: 1H NMR
data for fully deprotected compounds 28a, 28b, 33a and 33b; and 1H
NMR data for H-phosphonate derivatives 30a, 30b, 35a and 35b. See
http://www.rsc.org/suppdata/ob/b4/b410695d/
3 4 9 6
O r g . B i o m o l . C h e m . , 2 0 0 4 , 2 , 3 4 9 6 – 3 5 0 3
T h i s j o u r n a l i s
T h e R o y a l S o c i e t y o f C h e m i s t r y 2 0 0 4
©

View Online
targets, evaluated by absorption and fluorescence spectroscopy.
The structures of the modified ODNs 4–18 and the ODNs used
as targets 1 and 2 and reference 3 are depicted in Fig. 1.
hydroxyhexyl)perylene 23 (Scheme 1). The latter was obtained
following a procedure adapted from that of Schlichting et al.27
The azido-derivative 24 was obtained by treatment of 23
with LiN3, CBr4 and triphenylphosphine, after which the 3-(6-
aminohexyl)perylene 25 was obtained by catalytic hydrogena-
tion of compound 24.
Sugar–perylene units. The protected sugar–perylene deriva-
tives 27a and 27b were obtained by glycosylation of
perylene 19 with 1-o-methyl-2-deoxy-3,5-di(p-toluoyl)-erythro-
pentafuranose 2628 in the presence of SnCl4 (Scheme 2). A
mixture of anomers was obtained and separated by HPLC
using a silica column with 43% yield and a b : a, 73 : 27 ratio.
The anomeric configuration was assigned by 2D NMR (NOE)
experiments performed on the fully deprotected compounds
(vide infra). The protected sugar–perylene derivatives 32a and 32b
involving the propyl linker were obtained by glycosylation be-
tween 3-(3-hydroxypropyl)-perylene 22 and 1-a-chloro-2-deoxy-
3,5-di(p-toluoyl)-erythro-pentafuranose 3129, in the presence of
˚
ZnBr2, NaHCO3 and 4 A molecular sieves (Scheme 2). A mixture
of anomers was obtained and separated by semi-preparative
chromatography on silica plates with 78% yield and a b : a, 62
: 38 ratio (vide infra). The 5ꢀ- and 3ꢀ-hydroxyl functions of the
fully protected sugar–perylene units 27a, 27b, 32a and 32b were
deprotected by treatment with a sodium methylate solution to
give the unprotected sugar–perylene units 28a, 28b, 33a and 33b.
The latter were then dimethoxytritylated on their 5ꢀ-hydroxyl
function, using a classical procedure, to give compounds 29a, 29b,
34a and 34b, which were transformed into their 3ꢀ-phosphonate
derivatives 30a, 30b, 35a and 35b by treatment with 2-chloro-4 H-
1,3,2-benzodioxaphosphorin-4-one30 using classical conditions.
Structural assignments. Structural assignments of the
anomeric configuration for sugar–perylene units 28a, 28b, 33a
and 33b were made by 1H NOE studies. The method used is
based on the separate irradiation of the H-2ꢀ proton resonances
and the observation of enhancements at vicinal 1ꢀ and 3ꢀ protons.
Examination of the structures of aand b sugar–perylene units
shows that for a-anomers the 2ꢀ-b proton is in close proximity to
both the 1ꢀ and the 3ꢀ protons, while the 2ꢀ-a proton is not near
either of these protons. In the b-anomers, the 2ꢀ-a is only near
the 1ꢀ proton while the 2ꢀ-b proton is near only the 3ꢀ proton.
Thus, in an a-anomer separate irradiation of each of the 2ꢀ
protons should lead to two and zero enhancements at the vicinal
protons, while in a b-anomer these two irradiations would lead
to one significant enhancement for each irradiation. The results,
indicated in Table 1, confirm that for one anomer (assigned to be
in the aconfiguration) of each pair only one irradiation gives two
strong enhancements of the vicinal protons, while for the other
anomer (assigned to be in the b configuration) one significant
enhancement is observed for each of the two irradiations.
Fig. 1 Chemical structures of the ODNs 1–18.
Synthesis
Perylene linkers derivatives. The perylene linker deriva-
tive 22, involving a propyl linker, was obtained following
a three-step procedure (Scheme 1). A Sonogashira coupling
reaction between 3-bromoperylene 20, obtained from pery-
lene 19 following a reported procedure,26,27 and propargyl
alcohol led to the 3-(3-hydroxy-1-propynyl)perylene 21. The
3-(3-hydroxypropyl)perylene 22 was then obtained by cat-
alytic hydrogenation of 21. The perylene linker 25 was ob-
tained following a two-step procedure starting from the 3-(6-
ODN–perylene conjugates. In a series of syntheses, sugar–
perylene units, involving the perylene attached via its 3ꢀ-position
with or without a propyl linker to the anomeric position of
the sugar, were incorporated either between the fifth and the
sixth base of the sequence (starting from the 3ꢀ-end) or at
the 5ꢀ-end of the sequence. The internal position was chosen
to allow intercalation of the perylene unit between the TxAT
Scheme 1 Synthesis of the perylene derivatives 22 and 25. Reagents and conditions: i: ref. 26, ii: propargyl alcohol, THF–piperidine, CuI,
tetrakis(triphenylphosphine) palladium(0); iii: H2, Pd/C, THF; iv: ref. 27; v: PPh3, LiN3, CBr4, DMF; vi: H2, Pd/C, THF.
O r g . B i o m o l . C h e m . , 2 0 0 4 , 2 , 3 4 9 6 – 3 5 0 3
3 4 9 7

View Online
Scheme 2 Synthesis of the sugar–perylene units 27–30 and 32–35. Reagents and conditions: i: C2H4Cl2, SnCl4; ii: MeONa, CH2Cl2–MeOH (50 : 50,
˚
v/v); iii: DMTrCl, Py; iv: 2-chloro-4H-1,3,2-benzodioxaphosphorine-4-one, CH2Cl2–pyridine (4 : 6, v/v); v: ZnBr2, NaHCO3, 4 A molecular sieves,
THF.
Table 1 Proton NOE data for fully deprotected sugar–perylene units
28a, 28b, 33a and 33b
ODNs 5, 7, 9 and 11, involving the perylene at the b-anomeric
position of the sugar, were a little higher than those of the
corresponding ODNs 4, 6, 8, and 10, involving the perylene at
the a-anomeric position of the sugar. Retention times of ODNs
involving two perylene derivatives 12, 15 and 18 were higher
than those of the ODNs involving only one perylene (ODNs
4–11, 13, 14, 16 and 17). These results indicate the formation
of more lipophilic compounds. The UV-visible spectra of the
ODN–perylene ODNs were recorded between k = 230 nm and
k = 500 nm. The spectra contains two main absorption bands
in the visible region between k = 350 nm and k = 500 nm
(where only the perylene absorbs light) with kmax vis ≈ 448–
450 nm and kmax vis ≈ 420–423 nm, with a shoulder at k ≈
395 nm, for all the ODNs. The other absorption band in the
UV range corresponded to the absorbance of the ODN and
the perylene with a kmax UV ≈ 259–267 nm for all the ODNs.
The comparison of the two series of ODNs involving a single
sugar–perylene unit incorporation showed that in each series
the positions of the main absorption bands were identical;
independent of the position of incorporation or the anomeric
configuration of the sugar–perylene unit. Only the UV : visible
absorbance ratio was slightly different inside each series (UV :
visible absorbance ratio ≈ 4.25–4.50 for ODNs 4–7 and ≈ 4.40–
4.65 for ODNs 8–11). The UV-visible spectra of ODNs 13 and
14, involving a b-sugar-perylene unit with a propyl linker at the
3ꢀ-end of the sequence, and those of the ODNs 16 and 17, involv-
ing a polymethylene linker to connect the perylene to the ODN
via a phosphodiester linkage, were nearly identical to that of the
ODN 11, involving a b-sugar–perylene unit with a propyl linker
at the 5ꢀ-end of the sequence. In the case of the ODNs 12, 15
and 18, involving two perylenes, the kmax vis ≈ 449 and ≈ 423 nm
were unchanged, while a slight blue-shift (≈ 7–8 nm) of the
kmax UV was observed as compared to the kmax UV (265–267 nm)
of the corresponding ODNs involving only one perylene. The
main change concerns the UV : visible absorbance ratio which
is lower in the case of the ODNs involving two perylenes 12,
15 and 18 (3 to 3.4), as compared to those observed for the
corresponding ODNs involving only one perylene (vide supra).
The UV-visible spectra of the ODN–perylene ODNs 11 and 15
are shown in Fig. 2. All the modified ODNs were characterized
by electrospray mass spectrometry (Table 2).
Irradiation at:
Compounds
NOE observed
H2ꢀ-b (%)
H2ꢀ-a (%)
28a
28b
33a
33b
H1ꢀ
H3ꢀ
H1ꢀ
H3ꢀ
H1ꢀ
H3ꢀ
H1ꢀ
H3ꢀ
100.0
67.6
10.2
103.3
103.4
100.0
20.0
13.3
18.1
100.0
34.1
35.1
9.2
100.0
−2.4
28.9
base triplets. Coupling at the 5ꢀ-position of the sequence was
also chosen because studies carried out with other intercalator–
ODN conjugates showed that in the case of triple-stranded DNA
structures the junction between the triplex and the overhanging
double-stranded DNA target was a preferential binding site for
the intercalators. The preparation of these ODNs was performed
by manual incorporation of the sugar–perylene units via their
H-phosphonate derivatives during the ODN chain elongation.
After a standard deprotection step, purification was carried
out by reversed-phase chromatography to give ODNs 4–12.
For the incorporation of the sugar–perylene unit at the 3ꢀ-
end of the ODNs 14 and 15 we chose a support described
for the preparation of 3ꢀ-aminooligonucleotides.31 This choice
resulted in higher yields than by using a previously described 3ꢀ-
phosphate support32 initially chosen for the incorporation of the
sugar–perylene unit at the 3ꢀ-end (ODN 13). Deprotection and
purification steps were performed as for ODNs 4–12. The ODN
16, bearing the perylene via the 5ꢀ-terminal phosphate, was ob-
tained using our previously reported perylene phosphoramidite
derivative.23 ODN 17 was obtained by reacting the perylene
derivative 25 with an H-phosphonate group, obtained after
coupling of the H-phosphonate derivative of the 3ꢀ-terminal
nucleoside, as previously reported,23 followed by the full length
sequence assembly.
The perylene derivative 25, with a hexamethylene linker, was
used in place of our previously reported perylene derivative
involving a linker containing a secondary amino function.23 This
change was made to prevent acetylation of the amino function
during the ODN chain assembly that led to a weak stabilization
of the complexes involving the acetylated ODNs owing to
steric hindrance.23 ODN 18 was obtained as described for the
preparation of ODN 17 and by coupling the second perylene to
the 5ꢀ-end of the sequence via the phosphoramidite derivative of
perylene.23 Purification of the ODNs was performed by reversed-
phase chromatography. The retention times (Table 2) for the
Thermal denaturation studies and fluorescence measurements
Thermal denaturation studies. Experiments were followed
by absorption spectroscopy. One transition was observed in
the melting profile of each duplex while two transitions were
observed in the melting of each triplex (data not shown). The
double-stranded target has previously been circularised through
the use of two hexaethylene linkers in order to increase its
3 4 9 8
O r g . B i o m o l . C h e m . , 2 0 0 4 , 2 , 3 4 9 6 – 3 5 0 3

View Online
Table 2 Characterizations for ODNs 1, 3–18. Tms and fluorescence data. Column 1: Retention times obtained by reversed-phase HPLC analyses
for the ODNs 1 and 3–18 performed on a Lichrospher RP 18 (5 lm) column (125 mm × 4 mm) from Merck using a linear gradient of CH3CN (12.5%
to 42.5% over 40 min) in 0.1 M aqueous TEAA, pH 7, with a flow rate of 1 cm3 min−1. Column 2: Mass analysis data for ODNs 1 and 3–18. Column
3: Tm values for triplexes formed between ODNs 3–18 and the double-stranded target 1 in a 10 mM sodium cacodylate, pH 7, buffer containing
140 mM KCl and 5 mM MgCl2. ODN concentrations were 1 lM of the circularized duplex target 1 and 1.5 lM of the third strand. Column 4:
Tm values for duplexes formed between ODNs 3–18 and the single-stranded DNA target 2 in a 10 mM sodium cacodylate, pH 7, buffer containing
100 mM NaCl. ODN concentrations were 1 lM (each strand). Column 5: Relative fluorescence intensity between duplexes and ODNs 4–18, at 16 ◦C,
in the same buffer as used for the melting studies. Column 6: Relative fluorescence intensity between triplexes and ODNs 9, 11, 12, 15 and 18, at 4 ◦C,
in the same buffer as used for the melting studies
Mass analysis
Triplexes
Duplexes
Reversed-phase
Rt/min
Fduplex
FODNs
/
Ftriplex
FODNs
/
ODNs
Calculated
Found
Tm/◦C
DTm
Tm/◦C
DTm
1
3
—
C517H668N182O318P52 = 16130.66
16134.41
4440.6
—
—
—
—
—
—
—
—
—
—
—
—
—
—
—
0.50
—
1.06
1.46
—
—
2.12
—
6 min 20 sec
9 min 46 sec
10 min 23 sec
15 min 11 sec
18 min 47 sec
15 min 05 sec
15 min 31 sec
24 min 20 sec
26 min 53 sec
30 min 05 sec
21 min 45 sec
20 min 18 sec
34 min 08 sec
15 min 25 sec
28 min 32 sec
33 min 37 sec
C
146H192N34O99P14 = 4440.93
171H211N34O104P15 = 4871.33
17.0
18.5
16.0
19.0
25.5
25.0
25.0
27.0
30.0
36.0
25.0
25.0
39.0
30.0
25.0
37.0
38.0
41.0
40.0
44.0
45.0
43.0
40.0
45.0
45.0
44.0
44.0
44.5
51.0
44.0
44.0
46.0
4
C
4871.06
4870.97
4871.01
4871.33
4929.38
4929.10
4929.22
4929.30
5416.81
5008.52
5107.93
5595.84
4883.97
4850.95
5292.14
+1.5
−1.0
+2.0
+8.5
+8.0
+8.0
+10.0
+13.0
+19.0
+8.0
+8.0
+22.0
+13.0
+8.0
+20.0
+3.0
+2.0
+6.0
+7.0
+5.0
+2.0
+7.0
+7.0
+6.0
+6.0
+6.5
+13.0
+6.0
+6.0
+8.0
1.20
0.91
1.03
0.97
1.63
1.73
1.22
1.19
3.71
1.10
1.05
3.0
5
C171H211N34O104P15 = 4871.33
6
C
C
171H211N34O104P15 = 4871.33
171H211N34O104P15 = 4871. 33
7
8
C174H217N34O105P15 = 4929.41
9
C
174H217N34O105P15 = 4929.41
10
11
12
13
14
15
16
17
18
C174H217N34O105P15 = 4929.41
C
C
174H217N34O105P15 = 4929.41
202H242N34O111P16 = 5417.89
C174H218N34O108P16 = 5009.39
C180H231N35O108P16 = 5108.57
C
208H256N35O114P17 = 5597.04
C173H218N35O102P15 = 4884.42
0.97
1.05
2.90
C
C
172H216N35O101P15 = 4854.39
199H242N36O104P16 = 5297.84
—
2.76
significant stabilization. In this case, an equivalent stabilization
(D Tm = +8 ◦C) was observed with both anomers (ODNs 8 and
9). The absence of stabilization induced by the incorporation of
the sugar–perylene unit, without the linker, at internal position
of the ODN can be attributed to a problem of steric hindrance.
When the incorporation was performed at the 5ꢀ-end of the
ODN (ODNs 6, 7, 10 and 11) a significant stabilization was
obtained (D Tm ≥ +8.5 ◦C), except in the case of ODN 6
involving the a-anomer of the sugar–perylene unit without the
linker. However, the strongest stabilization (D Tm = +13 ◦C)
was obtained by the incorporation of the b-anomer of the sugar–
perylene unit involving the propyl linker (ODN 11). The stronger
stabilization provided by the incorporation of the b-anomer
(ODNs 7 and 11) compared to that of the a-anomer (ODNs
6 and 10) at the 5ꢀ-end of the third strand can be explained by
the fact that in the case of the b-anomer the stacking of the
perylene with the base triplet at the triplex–duplex junction is
favoured. The stabilization observed in the case of the ODN 11
is equivalent to that obtained when the perylene was attached via
a polymethylene linker to the 5ꢀ-terminal phosphate of the ODN
(ODN 16). A lower, but similar, stabilization (D Tm = +8 ◦C)
was obtained with the three ODNs 13, 14 and 17 involving
the perylene linked to the 3ꢀ-end. In the case of ODN 14, the
presence of an alkyl amino linker at the 3ꢀ-end did not increase
the stability of the triplex as compared to that of the triplex
involving the ODN 13, ending with a phosphate group at the
3ꢀ-end. This can be due to unfavourable interactions between the
linker and the 3ꢀ-end of the triplex structure.
Fig. 2 Absorption spectra of the ODNs 11 (full line) and 15 (broken
line) recorded between 230 and 500 nm in a 10 mM sodium cacodylate,
pH 7, buffer containing 100 mM NaCl.
stability, allowing the dissociation of the third strand to be
obtained before the melting of the duplex target.33 The transition
with the higher Tms corresponds to the melting of the target
◦
duplex (around 77 C for all complexes, data not shown) and
the transitions with lower Tms to the dissociation of the third
strand. The Tm values, listed in Table 2, indicate that the triplex
and duplex structures formed between the modified ODNs 4–
18 and the duplex 1 or single-stranded 2 targets were stabilized
in most cases (except in the case of the hybridization of ODN
5 with the double-stranded DNA target), as compared to the
corresponding complexes formed with the unmodified ODN 3
which was used as a reference.
Duplex stabilization. In the case of the duplex structure, the
difference in the stability increase observed was less dependent
on the position of the sugar–perylene unit than in the case of the
triplex structure. Except in the case of the incorporation of the
sugar–perylene at the internal position of the sequence (ODNs
4, 5, 8 and 9), leading to a weaker stabilization, the maximum
stabilization observed was +7 ◦C. These results indicate that
the perylene is more efficient in stabilizing the triplex than the
duplex structures as previously observed.23,25
ODNs involving one perylene. Triplex stabilization. The
following observations can be made. Among the sugar–
perylene units, those involving a linker (ODNs 8–11) are
the most efficient at stabilizing the triplex structures. When
the incorporation was performed at internal positions of the
sequence (ODNs 4, 5, 8 and 9) the presence of a linker, between
the perylene and the 2ꢀ-deoxyribose, was necessary to obtain a
ODNs involving two perylenes. Triplex stabilization. In all
cases, the presence of a second perylene (ODNs 12, 15, 18)
O r g . B i o m o l . C h e m . , 2 0 0 4 , 2 , 3 4 9 6 – 3 5 0 3
3 4 9 9

View Online
induced a stabilization increase of the triplex as compared
to that obtained with the ODNs involving only one perylene
(Fig. 3). The strongest stabilization was observed with ODN
15, involving two sugar–perylene units at both the 5ꢀ- and the
3ꢀ-ends of the ODN (D Tm = +22 ◦C). This D Tm increase was
nearly equal to the sum of the stability increase observed with
ODNs 11 (D Tm = +13 ◦C) and 14 (D Tm = +8 ◦C) involving
the sugar–perylene unit at the 5ꢀ- and the 3ꢀ-ends, respectively.
When the perylenes were linked via an hexamethylene linker
decrease in the fluorescence intensity was observed for the ODN
9, while no change was observed for the ODN 11.
ODNs involving two perylenes. In the case of the modified
ODNs involving two perylene units 12, 15 and 18, a greater
increase of the fluorescence intensity signal was observed upon
hybridization with the single-stranded target 2 (3.71, 3.0 and 2.90
fold, respectively). Using the same conditions, no change in the
fluorescence intensity signal was observed upon hybridization
of ODN 15 with another single-stranded target sequence [d-
5ꢀ CCCCTTTTCTTTTAAAAAGTGGC3ꢀ ], indicating that the
fluorescent signal modification was specific to the hybridization.
The fluorescence intensities of the duplexes involving two
perylenes differed little from those of the duplex involving only
one perylene. But the fluorescence intensities of the modified
ODNs involving two perylenes were lower than those of the
modified ODNs involving only one perylene. This corresponds
to a quenching of the fluorescence of the modified ODNs 12,
15 and 18 with two perylenes. The most efficient quenching
was observed with ODN 12, involving the perylenes linked
at both the 5ꢀ-position (via a sugar residue) and the internal
position (after ten nucleotides). When the two perylenes are
linked to the ends of the sequence, the distance is greater, but
the intramolecular folding brings the two perylenes into closer
proximity allowing the quenching.
◦
to the terminal phosphates (ODN 18) a D Tm = +20 C was
observed. The incorporation of sugar–perylene units at both
the 5ꢀ-end and the internal position (ODN 12) also led to an
important stabilization of the triplex (D Tm = +19 ◦C). This
marked increase in the stability of the triplex structures provided
by the presence of a second intercalator was also previously
reported in the case of ODN–naphtalene conjugates.20
Studies carried out with ODNs 12, 15 and 18 and the double-
stranded target 1 also indicated an increase in the fluorescence
intensity signal (1.46, 2.12 and 2.76 fold, respectively).
Conclusions
Fig. 3 UV melting profiles (recorded at k = 260 nm) corresponding to
the third strand dissociation of the triplex formed by ODNs 11 (–) and 15
(D) and by the unmodified ODN 3 (♦) in the presence of the circularised
duplex 1. Concentrations were 1 lM of the target and 1.5 lM of the
third strands in a 10 mM sodium cacodylate buffer, pH 7, containing
140 mM KCl and 5 mM MgCl2.
We have reported the synthesis and the binding properties with
double- and single-stranded targets, studied by absorption and
fluorescence spectroscopy, of a series of pyrimidine sequences
covalently linked to perylene. The linkage of the latter to the
ODN sequence was performed via the anomeric position of a
2ꢀ-deoxyribose or via polymethylene linkers attached to the 5ꢀ-
and 3ꢀ-terminal phosphates. In the case of the incorporation of
the perylene by the intermediate of the sugar–perylene units, the
polycyclic aromatic molecule was either directly connected to
the anomeric position of the sugar residue or via a propyl linker.
In each case, the sugar–perylene units were obtained as pure
a- and b-anomers and incorporated separately into the ODNs.
First, single incorporations were performed at either the 5ꢀ-end
or an internal position of the sequence. The results of the binding
studies indicated that the presence of a linker between the sugar
residue and the perylene is necessary to obtain a significant
stabilization of both the triplex and duplex structures. In the
absence of the linker it is probable that, due to a problem of
steric hindrance, the perylene cannot adopt a position favourable
to afford a good stabilization. When a linker was used to
connect the perylene and the sugar, in all cases stabilization was
provided by the presence of the perylene and, in most examples,
the perylene was more efficient at stabilizing the triplex than
the duplex structures. However, the strongest stabilization was
observed when the b-anomer was incorporated at the 5ꢀ-end
of the sequence. In this case, the stabilization provided by the
use of a sugar–perylene unit, involving the propyl linker, was
equivalent to that obtained when the perylene was attached
to the terminal phosphate of the ODN via a polymethylene
linker. After the most efficient sugar–perylene unit was selected,
it was incorporated twice at both the 5ꢀ-and the 3ꢀ-ends (or
both the 5ꢀ-end and an internal position) of the ODN. Two
perylenes were also attached to both the 5ꢀ- and the 3ꢀ- terminal
phosphates of the ODNs via flexible linkers. In the case of triplex
formation, the covalent attachment of two perylene units led to
a higher stability increase than in the case of the linking of only
one perylene unit to the same sequence (the D Tm increase is
the sum of the D Tm increase observed for ODNs involving
one perylene at either position). The best result was obtained
Duplex stabilization. In the case of the duplex, the stabilization
provided by the presence of two perylene units was almost
equivalent to that observed with ODNs involving only one
perylene, except in the case of ODN 15 involving a sugar–
perylene unit at each end of the sequence (D Tm = +13 ◦C).
This weak stability increase provided by the attachment of two
intercalators (via terminal phosphates) to an ODN as compared
to the stability of the duplexes formed with the same ODN
linked to only one intercalator previously been observed with
ODN–acridine conjugates.24
Fluorescence studies
Fluorescence emission spectra of ODNs 4–18 were compared
with those of the duplexes and triplexes. The kmax em were almost
equivalent for all the ODNs (451 nm for ODNs 4–7; 453 nm for
ODNs 8–11, 13 and 14; 454 nm for ODNs 12, 15, 16 and 18;
and 455 nm for ODN 17). Upon hybridization with either the
double-stranded 1 or the single-stranded target 2, the kmax em did
not change as compared to those observed for the ODNs. The
main changes concerned the intensity of the fluorescent emission
(Table 2).
ODNs involving one perylene. In the presence of the single-
stranded target sequence 2, two tendencies were clearly ob-
served. In the case of the ODNs 4–7, 13, 14, 16 and 17, the
intensity of the fluorescence emission was only slightly modified
in the presence of the target. More important changes, (a little
more than 50% of the intensity increase) were observed in the
case of ODNs 8–11, involving the sugar–perylene unit with
the propyl linker. The weak fluorescence increase observed for
these ODNs, upon hybridization with the target sequence, was
previously observed in the case of the linkage of a perylene
derivative to the 3ꢀ-end of an ODN via a pseudonucleoside unit.34
In the presence of the double-stranded target sequence 1, a 50%
3 5 0 0
O r g . B i o m o l . C h e m . , 2 0 0 4 , 2 , 3 4 9 6 – 3 5 0 3

View Online
with the incorporation of two sugar–perylene units at each end
of the ODNs, probably because the entropic factor was more
favourable than with ODNs involving the perylenes attached
via longer flexible linkers. These results confirm the paramount
influence of the linkage parameters between the perylene and
the ODNs. In addition, in the case of the ODNs involving
two perylenes a significant intensity increase of their fluorescent
signal was observed upon their hybridization with both their
single-stranded and double-stranded targets that can be used as
a proof of their binding.
CH2), 3.02 (2H, t, J 7.8, Ar-CH2), 3.46–3.54 (2H, m, CH2O),
4.56 (1H, t, J 4.9, OH), 7.40 (1 H, d, J 7.8, Ar-H), 7.48–7.60
(3H, m, Ar-H), 7.75 (2 H, t, J 8.8, Ar-H), 7.94 (1 H, d, J 7.8,
Ar-H), 8.22–8.38 (4H, m, Ar-H). ESI-MS: m/z, C23H18O calc.
310.4, found 311.2 (M + H+).
3-(6-Aminohexyl) perylene 25 (Scheme 1). The synthesis of
the perylene linker derivative 23 was adapted from Schlichting
et al.27 Compound 23 (183 mg, 0.52 mmol), triphenylphosphine
(207 mg, 0. 79 mmol) and LiN3 (77 mg, 1.57 mmol) were dried in
a dessicator. DMF (5 cm3) was added and then a solution of CBr4
(259 mg, 0.78 mmol) in DMF (2 cm3) was added dropwise. After
2 h of reaction, MeOH (1 cm3) was added and the solvents were
removed by distillation under reduced pressure. The residue was
purified by flash chromatography on silica gel (30% CH2Cl2 in
Experimental
General methods
cyclohexane v/v) to give 24 as a brown oil (192 mg, 98%). Rf24
=
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. The reagents were dried for 15 h in a
desiccator containing KOH and P2O5, under a reduced pressure
(20 mm Hg). During all the synthesis and purification steps pery-
lene derivatives were protected from light. Analytical thin-layer
chromatography (TLC) was performed on precoated alumina
plates (Merck silica gel 60F 254 ref. 5554) and preparative TLC
on glass-backed silica plates (Merck silica gel 60 F254 ref. 5717).
For flash chromatography, Merck silica gel 60 (40–63 lM) (ref.
9385 from Merck) was used. NMR spectroscopy was performed
0.55, Rf23 = 0 [cyclohexane : CH2Cl2 (50 : 50, v/v)]. Compound
24 (188 mg, 0.50 mmol) was solubilized with THF (25 cm3),
10% Pd/charcoal (100mg) was added and hydrogenation was
performed for 28 h as reported above for compound 22. After
Celite filtration, the mixture was concentrated and the residue
purified on preparative silica plates (10% MeOH in CH2Cl2,
v/v). The plates were eluted four times. Brown solid: (62 mg,
35%). Rf25 0.46, [CH2Cl2 : MeOH (80 : 20, v/v) then CH2Cl2 :
MeOH (50 : 50, v/v) then CH2Cl2 : MeOH : NH4OH (50 : 50 :
5, v/v/v)]. dH (DMSO-d6) 1.30–1.47 (6H, m, 3 CH2), 1.62–1.71
(2H, m, CH2), 2.75–2.80 (2H, m, CH2NH2), 2.98 (2H, t, J 8.0,
Ar-CH2), 7.38 (1 H, d, J 6.5, Ar-H), 7.48–7.60 (3H, m, Ar-H),
7.75 (2 H, t, J 9.0, Ar-H), 7.91 (1 H, d, J 8.0, Ar-H), 8.24–8.39
(4H, m, Ar-H). ESI-MS: m/z, C26H25N calc. 351.49, found 352.3
(M + H+).
1
on a Varian Unity 500 spectrometer. H Chemical shifts were
referenced to either residual solvent peak DMSO (2.54 ppm) or
Me4Si. 31P Chemical shifts were referenced to H3PO4 (external
reference). 1H NMR coupling constants are reported in Hz and
refer to apparent multiplicities. Mass analysis was performed on
a Quattro II (Micromas) instrument. ODNs were synthesized
using cyanoethyl phosphoramidite chemistry and an Expedite
Nucleic Acid Synthesis System 8909 from Perseptive Biosystems.
Reversed-phase chromatography analysis and purification were
performed on a 600 E System Controller equipped with a Waters
990 photodiode array detector. UV spectra were recorded on
an Uvikon 860 spectrophotometer. Fluorescence spectra were
recorded on a Fluoromax 2 (ISA-Jobin-Yvon) spectrofluorime-
ter in 0.5 cm path-length Suprasil quartz cuvettes (Hellma) with
slits set at 0.5 mm (band pass = 2 nm).
Sugar–perylene units 27a and 27b (Scheme 2). Perylene 19
(252 mg, 1 mmol) and 1-o-methyl-2-deoxy-3,5-di(p-toluoyl)-
erythro-penta-furanose 26 (384 mg, 1 mmol)28 were dried in
a dessicator and solubilized with 1,2-dichloroethane (100 cm3)
under an argon atmosphere. SnCl4 (60 ll, 0.51 mmol) was
added with a syringe to the mixture, under stirring, at room
temperature. After 3 h of reaction, the mixture was washed
with a 0.5 M aqueous NaHCO3 solution (50 cm3), then H2O
(50 cm3) and with a 1 M aqueous NaCl solution (50 cm3). The
organic phase was dried over Na2SO4, filtered and concentrated
to dryness. The brown residue was purified twice by flash
chromatography (50% cyclohexane in CHCl3, v/v). Separation
of the anomers was then performed by semi-preparative HPLC
using a Hibar Lichrosorb column Si 60, 7 lm, (250 × 25 mm)
from Merck (50% cyclohexane in CHCl3, v/v) at a flow rate
of 4 cm3 min−1. Before loading, the samples were filtered
using Acrodisc Gelman CR, PTFE 0.45 lM, diameter 25 mm.
Detection was performed at 460 nm. Yellow solids [27a 70 mg;
27b 190 mg, 43% (b + a) with a b : a ratio 73 : 27]. Rf27a 0.27,
Rf27b 0.34, Rf19 0.87 and Rf26 0.15 (cyclohexane : EtOAc, 80 :
20, v/v). dH (CDCl3). 27a: 2.32 (3H, s, CH3), 2.40–2.48 (4H, m,
H-2ꢀ, CH3), 3.15–3.24 (1H, m, H-2ꢀꢀ), 4.63–4.72 (2H, m, H-5ꢀ,
H-5ꢀꢀ), 4.86–4.90 (1H, m, H-4ꢀ), 5.66–5.71 (1H, m, H-3ꢀ), 6.00
(1H, pseudo t, J 5.6, J 7.7, H-1ꢀ), 7.09 (2H, d, J 7.7, Ar-H), 7.28
(2 H, d, J 8.1, Ar-H), 7.47–7.55 (3 H, m, Ar-H), 7.65–7.76 (5 H,
m, Ar-H), 7.79 (1H, d, J 7.7, Ar-H), 8.03 (2 H, d, J 8.1, Ar-H),
8.19–8.25 (4 H, m, Ar-H). 27b: 2.31–2.40 (4H, m, H-2ꢀ, CH3),
2.46 (3H, s, CH3), 2.80–2.87 (1H, m, H-2ꢀꢀ), 4.65–4.69 (1H, m,
H-4ꢀ), 4.70–4.80 (2H, m, H-5ꢀ, H-5”), 5.67–5.71 (1H, m, H-3ꢀ),
5.91 (1H, dd, J 5.1, J 10.7, H-1ꢀ), 7.20 (2 H, d, J 8.1, Ar-H), 7.32
(2 H, d, J 8.1, Ar-H), 7.46–7.53 (3 H, m, Ar-H), 7.67–7.71 (2H,
m, Ar-H), 7.77–7.86 (2 H, m, Ar-H), 7.95 (2 H, d, J 8.1, Ar-H),
8.05 (2H, d, J 8.6, Ar-H), 8.13–8.24 (4H, m, Ar-H). ESI-MS:
m/z 27a and 27b, C41H32O5 calc. 604.7, found 27a 627.5 (M +
Na+) and 27b 627 (M + Na+).
Synthesis
Perylene-linker derivatives.
3-(3-Hydroxypropyl) perylene 22 (Scheme 1). 3-Bromopery-
lene 20 obtained as previously reported26 (1.01 g, 3.05 mmol)
was placed in a round-bottomed flask, dried in a dessicator
and solubilized with a THF–piperidine (50 : 50, v/v) mixture
(60 cm3). Propargyl alcohol (342 mg, 6.10 mmol) and then
tetrakis-(triphenylphosphine) palladium(0) (156 mg, 135 lmol)
and copper iodide (32 mg, 168 lmol), previously dried over P2O5
in an oven at 60 ◦C for 4 h under a reduced pressure (20 mm Hg),
were added under an argon atmosphere. The reaction mixture
was refluxed at 70 ◦C for 2 h, poured into an ice–HCl (3 : 1,
v/v) mixture (180 cm3) and extracted with CH2Cl2 (300 cm3 and
then 150 cm3 × 3). The organic phase was dried over Na2SO4,
filtered and concentrated to dryness. The residue was purified
by flash chromatography (CH2Cl2) to give 21 as a yellow solid
(615 mg, 66%). Rf21 0.28, Rf20 0.88 (CH2Cl2). dH (DMSO-d6)
4.5 (2H, s, CH2), 7.50–7.90 (5 H, m, Ar-H), 8.15 (1H, d, J 9.0,
Ar-H), 8.30–8-48 (5 H, m, Ar-H).
Compound 21 (940 mg, 3.07 mmol) was placed in a round-
bottomed flask, solubilized with THF (90 cm3) and 10%
Pd/charcoal (280 mg) was added under argon. The flask was
evacuated and flushed twice with argon, equipped with a balloon
filled with hydrogen, evacuated and flushed twice. After 15 h of
reaction, the mixture was filtered over Celite, concentrated and
the residue was purified by flash chromatography (CH2Cl2) to
give the perylene-linker derivative 22 as a yellow solid (800 mg,
84%). Rf22 0.13 (CH2Cl2). dH (DMSO-d6) 1,76–1.85 (2H, m,
Sugar–perylene units 32a and 32b (Scheme 2). Compounds
22 (664 mg, 2.14 mmol) and 3129 (1.248 g, 3.21 mmol),
ZnBr2 (hygroscopic) (133 mg, 0.58 mmol), NaHCO3 (360 mg,
˚
4.29 mmol) and activated molecular sieves 4 A (3.3 g) were
dried in a dessicator. THF (40 cm3) was added with a syringe
O r g . B i o m o l . C h e m . , 2 0 0 4 , 2 , 3 4 9 6 – 3 5 0 3
3 5 0 1

View Online
under an argon atmosphere, and the reaction mixture was
stirred at room temperature. After a 4 h reaction, the mixture
was filtered and concentrated to dryness. The yellow residue was
solubilized with CH2Cl2 (130 cm3). The organic phase was
washed with H2O (50 cm3), dried over Na2SO4, filtered and
concentrated to dryness. The residue was purified by flash
chromatography (CH2Cl2). Separation of the anomers was
performed on preparative silica plates (30% cyclohexane in
CH2Cl2, v/v). The plates were eluted eight times. Two bands
corresponding to each pure anomers of the sugar–perylene unit
32 were separated and extracted with CH2Cl2 to give yellow
solids corresponding to each isomer [32a 417 mg, 32b 685 mg,
78% (b + a) with a b : a ratio 62 : 38]. Rf32a 0.44, Rf32b 0.51, Rf22
0.13 and Rf31 0.02 (CH2Cl2). dH (CDCl3) 32a: 2.02–2.12 (2H,
m,CH2), 2.28–2.36 (4H, m, H-2ꢀ, CH3), 2.38 (3H, s, CH3), 2.52–
2.60 (1H, m, H-2ꢀꢀ), 3.06–3.22 (2H, m, Ar-CH2), 3.54–3.60 (1H,
m, CH2O), 3.87–3.93 (1H, m, CH2O), 4.52–4.58 (1H, m, H-4ꢀ),
4.60–4.68 (2 H, m, H-5ꢀ, H-5ꢀꢀ), 5.35 (1H, d, J 5.1, H-1ꢀ), 5.44–
5.48 (1H, m, H-3ꢀ), 7.16 (2H, d, J 7.7, Ar-H), 7.20 (2H, d, J 8.1,
Ar-H), 7.33 (1H, d, J 7.7, Ar-H), 7.40–7.50 (3H, m, Ar-H), 7.67
(2H, t, J 8.1, Ar-H), 7.86 (1H, d, J 8.6, Ar-H), 7.91 (2H, d, J 8.1,
Ar-H), 7.97 (2H, d, J 8.1, Ar-H), 8.08 (1 H, d, J 7.7, Ar-H), 8.14
(1 H, d, J 6.8, Ar-H), 8.17–8.21 (2H, m, Ar-H). 32b: 1.96–2.04
(2H, m, CH2), 2.32 (3H, s, CH3), 2.37–2.46 (4H, m, H-2ꢀ, CH3),
2.60–2.67 (1H, m, H-2ꢀꢀ), 2.98–3.10 (2H, m, Ar-CH2), 3.46–3.53
(1H, m, CH2O), 3.84–3.91 (1H, m, CH2O), 4.49–4.53 (1H, m,
H-4ꢀ), 4.54–4.58 (1H, m, H-5ꢀ), 4.59–4.64 (1H, m, H-5ꢀꢀ), 5.38 (1
H, dd, J 2.6, J 5.6, H-1ꢀ), 5.62–5.66 (1H, m, H-3ꢀ), 7.15 (2H,
d, J 8.1, Ar-H), 7.25 (2 H, d, J 8.1, Ar-H), 7.30 (1H, d, J 7.7,
Ar-H), 7.44–7.53, (3H, m, Ar-H), 7.66 (2H, t, J 7.9, Ar-H), 7.85
(1H, d, J 8.1, Ar-H), 7.91–7.98 (4H, m, Ar-H), 8.09 (1H, d, J
7.7, Ar-H), 8.14 (1H, d, J 7.3, Ar-H), 8.17–8.22 (2H, m, Ar-H).
ESI-MS: m/z 32a and 32b, C44H38O6 calc. 662.7, found 32a 663.2
(M + H+) and 32b 663.2 (M + H+).
pyridine mixture (50 : 50, v/v), under an argon atmosphere, to
obtain a 50 mM solution. Then, 1.2 eq. of 2-chloro-4H-1,3,2-
benzodio-xaphosphorin-4-one30 [100 mM solution in a CH2Cl2–
pyridine mixture (1 : 5, v/v)] was added to the reaction mixture
at 0 ◦C. After 1.5 h, water containing TEA (20% by volume)
was added to the reaction mixture and vigorous stirring was
maintained for 30 min. The crude product was extracted with
CH2Cl2, the organic phase dried over Na2SO4 and concentrated
under a reduced pressure to give a brown oil. The residue
was purified by flash chromatography [5% MeOH in CH2Cl2
containing 0.5% NEt3 (v/v/v)] to give 30a (83 mg, 95%), 30b
(260 mg, 60%), 35a (154 mg, 58%), 35b (250 mg, 62%). Rf30a 0.12,
1
Rf30b 0.07, Rf35a 0.09 and Rf35b 0.15, using the same eluent. H-
NMR data are given as supplementary material†. dp(CDCl3) 30a
4.10, 30b 3.69, 35a 4.16 and 35b 3.29. ESI-MS: m/z 30a and 30b:
C46H39O7P calc 734.8, found 30a 735.1 (M + H+) and 30b 735.2
(M + H+). 35a and 35b: calculated mass, C49H45O8P = 792.9;
found 35a 995.7 (M + H+ + 2 TEA); found 35b 995.5 (M +
H+ + 2 TEA).
ODN–perylene conjugates 4–18. The ODNs 4–15 were as-
sembled using classical phosphoramidite chemistry on a CPG
support at a one lmol scale. To introduce the perylene–sugar
residue into the ODNs, a manual coupling step was used
for the H-phosphonate derivatives 30a, 30b, 35a and 35b. The
syntheses were performed as follows. At the position chosen
for the incorporation of the sugar–perylene unit an additional
detritylation step was performed. Then, after drying of the
support, the selected H-phosphonate derivatives 30a, 30b, 35a
or 35b, 15.2 eq. [in a pyridine–CH3CN mixture (50 : 50, v/v)
3
˚
˚
(0.38 cm of a 40 mM solution) dried overnight on 3 A and 4 A
molecular sieves] and pivaloyl chloride 49.4 eq. [in a pyridine–
CH3CN mixture (50 : 50, v/v) (0.38 cm3 of a 130 mM solution),
prepared one hour before use] were added simultaneously to
either the support or to the ODN chain bound to the support.
After 2.5 min of reaction the solution was removed and the
support washed with an anhydrous pyridine–CH3CN mixture
(50 : 50, v/v) (1 cm3 × 6). The coupling yield was monitored by
a trityl cation assay. Then, the H-phosphonate derivative was
transformed into the 2-cyanoethyl phosphotriester group by a
15 min treatment of the ODN bound to the support with a 10%
solution (1 cm3) of 3-hydroxypropionitrile in a CCl4–CH2Cl2–
CH3CN–NEt3–NMI mixture (90 : 10 : 10 : 5 : 5, v/v/v/v/v)
Deprotection of compounds 27a, 27b, 32a and 32b. Compounds
27a, 27b, 32a and 32b were solubilized separately with a CH2Cl2–
MeOH mixture (50 : 50, v/v) and deprotected by an excess of
sodium methylate (5 eq.) in methanol. A precipitate appeared
during the process. The starting materials (Rf27a 0.90, Rf27ß 0.90,
Rf32a 0.92 and Rf32ß 0.94) were transformed into new yellow
products with lower Rf values (Rf28a 0.26, Rf28ß 0.23, Rf32a 0.34
and Rf32ß 0.33) (EtOAc). After a 15 h reaction, the mixtures were
neutralized by addition of a 18 M solution of acetic acid. The
yellow solids were filtered, washed with a MeOH–H2O mixture
(1 : 1, v/v) and dried. 28a (105 mg, 86%). 28b (260 mg, 86%). 33a
(221 mg, 89%). 33b (361 mg, 92%). 1H-NMR data are given as
supplementary material†. ESI-MS: m/z 28a and 28b, C25H20O3
calc. 368.4, found 28a 369.3 (M + H+) and 28b 369.3 (M + H+).
33a and 33b, C28H26O4 calc. 426.5, found 33a 444.4 (M + H2O)+
and 33b 426.3 (M + H+).
˚
˚
dried overnight on 3 A and 4 A molecular sieves following
a procedure adapted from a literature report.35 Finally, the
residual H-phosphonate linkage was oxidized by a 1 h treatment
with 0.1 M iodine solution in pyridine–H2O mixture (98 : 2, v/v).
After removal of the solution, the support was washed with an
anhydrous pyridine–CH3CN mixture (50 : 50, v/v) (3 × 1 cm3)
and then with anhydrous CH3CN (1 cm3). Except in the case
of the 5ꢀ-terminal addition of the H-phosphonate derivative,
the support was treated with a mixture of capping solutions
used on the synthesizer (0.5 cm3 each) for 10 min, washed with
CH3CN (4 × 1 cm3) and dried. The ODN chain assemblies
were completed via phosporamidite chemistry to give the fully
protected ODNs. ODN 16 was obtained by using the previously
reported perylenyl phosphoramidite.23 ODN 17 was obtained
by using the perylene derivative 25 and proceeding as described
in our previous report,23 followed by the full length sequence
assembly. ODN 18 was obtained as described for the preparation
of ODN 17 and by coupling the second perylene to the 5ꢀ-end
of the sequence using our previously reported phosphoramidite
derivative.23
Tritylation of compounds 28a, 28b, 33a and 33b. Compounds
28a, 28b, 33a and 33b were dried separately by co-evaporation
with pyridine, solubilized with pyridine and 4,4ꢀ-dimethoxytrityl
chloride (1.1 eq.) was added. After a 21 h reaction, the
mixture was diluted with CH2Cl2, washed with a 0.5 M
aqueous NaHCO3 solution, dried over Na2SO4, filtered and
concentrated to dryness. The orange-colored oil was purified
by flash chromatography [3% EtOAc in CH2Cl2 containing 1%
NEt3 (v/v/v)] to give the tritylated compounds 29a (71 mg,
37%), 29b (360 mg, 82%), 34a (239 mg, 74%) and 34b (364 mg,
85%). Rf29a 0.48, Rf29b 0.29, Rf34a 0.55 and Rf34b 0.13, using
the same eluent. ESI-MS: m/z 29a and 29b, C46H38O5 calc.
670.8, found 29a 670.2 (M + H+) and 29b 671.3 (M + H+).
34a and 34b, C49H44O6 calc. 728.9, found 34a 729.3 (M + H+)
and 34b 729.3 (M + H+). Compounds 29a, 29b, 34a and 34b were
used in the next synthesis step without additional characteri-
zation.
The deprotection step was performed by a 28% aqueous
ammonia treatment for 18 h at 20 ◦C, either alone or in the
presence of DTT in the case of the preparation of ODNs 13,
17 and 18 involving the use of a modified support containing a
disulfide bridge.32 For ODNs 4–15 and 17 the detritylation step
was performed by a treatment with a 80% acetic acid solution
for 30 min before the purification step. Analyses and purifica-
tions by reversed-phase chromatography were performed on a
H-phosphonate derivatives 30a, 30b, 35a and 35b. Compounds
29a, 29b, 34a and 34b were dried separately by co-evaporation with
pyridine, dried in a dessicator and solubilized with a CH2Cl2–
3 5 0 2
O r g . B i o m o l . C h e m . , 2 0 0 4 , 2 , 3 4 9 6 – 3 5 0 3

View Online
Lichrospher 100 RP18 column (5 lM, 125 mm × 4 mm) from
Merck with a linear gradient of CH3CN (12.5% to 42.5% over
40 min) in 0.1 M aqueous triethylammonium acetate, pH 7,
with a flow rate of 1cm3 min−1 and detection at k = 260 nm.
The mass values of the modified ODNs were confirmed by
electrospray mass spectrometry analysis (Table 2). The overall
yield after the synthesis and purification steps was 11–32%
for ODNs 4–12 and 16, while lower yields were obtained
for ODNs 13 (3%), 14 (7%), 15 (5%), 17 (1%) and 18 (1%),
involving the incorporation of a perylene unit at the 3ꢀ-end of
the ODNs by reaction of the perylene derivatives on modified
supports.
Acknowledgements
We thank H. Meudal for recording NMR spectra and C. Bure´
for running the electrospray mass spectrometer.
References
1 T. Christopoulos, Anal. Chem., 1999, 71, 425R.
2 N. C. Seeman, Nature, 2003, 23, 427.
3 D. D. F. Ma, T. Rede, N. A. Naqvi and P. D. Cook, Biotechnol. Annu.
Rev., 2000, 5, 155.
4 K. R. Fox, Curr. Med. Chem., 2000, 7, 17.
5 A. C. Stephens and R. P. A. Rivers, Curr. Opin. Mol. Ther., 2003, 5,
118.
6 C. F. Bennett and L. M. Coswert, Curr. Opin. Mol. Ther., 1999, 1,
359.
Hybridization studies
7 J. S. Sun and C. He´le`ne, Nucleosides, Nucleotides Nucleic Acids, 2003,
22, 489.
Melting studies. Experiments were carried out by absorption
spectroscopy using the double-stranded ꢀ DNA target d-5ꢀ
TAGTTTCTCTTCTTTTTCTTCTCTT3ꢀ /3 ATCAAAGAGA-
AGAAA-AAGAAGAGAA5ꢀ 1 [circularized using two hexa-
ethylene linkers33 in order to increase its stability] and the
single-stranded complementary DNA target d-5ꢀ CTCAGAGA-
AGAAAAAGAACTC3ꢀ 2. Concentrations of the circularized
double-stranded ODN 1, as well as those of the unmodified
ODNs 2 and 3, were calculated using molar extinction
coefficients at 260 nm determined using the nearest-neighbour
8 S. M. Freier and K. H. Altmann, Nucleic Acids Res., 1997, 25, 4429.
9 M. Manoharan, Antisense Nucleic Acid Drug Dev., 2002, 12, 103.
10 U. Asseline, in Modified Nucleosides, Synthesis and Application, ed.
D. Loakes, Research Signpost, Kerala, India, 2002, p. 31.
11 M. J. Gait, Cell. Mol. Life Sci., 2003, 60, 844.
12 S. Buchini and C. J. Leumann, Curr. Opin. Chem. Biol., 2003, 7, 717.
13 J. S. Sun, C. Giovannangeli, J. C. Francˇois, R. Kurfu¨rst, T. Montenay-
Garestier, U. Asseline, N. T. Thuong and C. He´le`ne, Proc. Natl. Acad.
Sci. USA., 1991, 88, 6023.
14 A. A. Koshkin, K. Y. Kropachev, S. V. Mamaev, N. V. Bulychev, S. G.
Lokhov, V. V. Vlassov and A. V. Lebedev, J. Mol. Recognit., 1994, 3,
177.
model36 (ODN 1: e260 = 497 700 M−1 cm−1, ODN 2: e260
=
222 800 M−1 cm−1, ODN 3: e260 = 118 500 M−1 cm−1). The molar
extinction coefficients (e) for the ODN–perylene ODNs 5, 11
and 15 were determined by titration of the ODN solutions in
a 10 mM sodium cacodylate, pH 7, buffer containing 100 mM
NaCl at 3 ◦C with a solution of single-stranded complementary
15 J. F. Mouscadet, C. Ketterle´, H. Goulaouic, S. Carteau, F. Subra, M.
Le Bret and C. Auclair, Biochemistry, 1994, 33, 4187.
16 A. Garbesi, S. Bonazzi, S. Zanella, M. L. Capobianco, G. Giannini
and F. Arcamone, Nucleic Acids Res., 1997, 25, 2121.
17 G. C. Silver, C. H. Nguyen, C. H. Boutorine, E. Bisagni, T. Garestier
and C. He´le`ne, Bioconjugate Chem., 1997, 8, 15.
18 G. C. Silver, J. S. Sun, C. H. Nguyen, C. H. Boutorine, E. Bisagni
and C. He´le`ne, J. Am. Chem. Soc., 1997, 119, 263.
19 M. D. Keppler, M. C. McKeen, O. Zegrocka, L. Strekowski, T. Brown
and K. R. D. Fox, Biochim. Biophys. Acta, 1999, 1447, 137.
20 D. A. Gianolio, J. M. Segismundo and L. W. McLaughlin, Nucleic
Acids Res., 2000, 28, 2128.
21 S. Vinogradov, V. Roig, Z. Sergueeva, C.-H. Nguyen, P. Arimondo,
T. T. Nguyen, E. Bisagni, J.-S. Sun, C. He´le`ne and U. Asseline,
Bioconjugate Chem., 2003, 14, 120.
22 S. Bevers, S. Schutte and L. W. Mc Laughlin, J. Am. Chem. Soc.,
2000, 122, 5905.
23 U. Asseline and E. Cheng, Tetrahedron Lett., 2001, 42, 9005.
24 U. Asseline, E. Bonfils, D. Dupret and N. T. Thuong, Bioconjugate
Chem., 1996, 7, 369.
25 Y. Aubert and U. Asseline, Nucleosides, Nucleotides Nucleic Acids,
2003, 22, 1223.
26 R. Lapouyade, J. Pereyre and P. Garrigues, C. R. Acad. Sci., Ser. II,
1986, 303(10), 903.
27 P. Schlichting, U. Rhor and K. Mu¨llen, Liebigs Ann. Chem., 1997,
395.
sequence 2 (ODN 5: e260 = 154 400 M−1 cm−1, ODN 11: e260
=
154 100 M−1 cm−1 and ODN 15: e260 = 165 800 M−1 cm−1).
Considering the minor differences observed for the UV-visible
spectra corresponding to each series of ODNs, the e value
obtained for ODN 5 was used for ODNs 4, 6 and 7. The e
value obtained for ODN 11 was used for ODNs 8–10, 13, 14,
16 and 17. The e value obtained for the ODN 15 was used
for ODNs 12 and 18. All concentrations are given on a per
strand basis. In the case of the duplexes, a 1 lM ODNs (each
strand) was used in a 10 mM sodium cacodylate buffer, pH 7,
containing 100 mM NaCl. In the case of the triplexes, the
experiments were performed with a 1 lM concentration in the
circularized double-stranded target and a 1.5 lM concentration
in the third ODN strand in a 10 mM cacodylate buffer, pH 7,
containing 140 mM KCl and 5 mM MgCl2. Duplex and triplex
stabilities were determined by thermal denaturation. Results are
given in Table 2. The uncertainty in the Tm values reported
0.5 ◦C.
28 K. Ness, in Synthetic Procedures in Nucleic Acids Chemistry, eds
W. W. Zorbach and R. S. Tipson, Interscience Publishers, New York,
1968, p. 183.
Fluorescence studies. Fluorescence studies were performed
using the same buffer conditions as used for the binding studies.
The emission spectra of 1 lM concentrations of the ODNs were
first recorded and then a small volume of the target sequences 1
or 2 (1 eq.) was added. The mixtures were allowed to hybridize
in the dark at 16 ◦C for the duplexes (30 min) and at 4 ◦C
for the triplexes (overnight) to ensure complete hybridization.
The emission spectra of the duplex and triplex structures were
recorded between k = 450 and 550 nm using the same kexc as for
the ODN–perylene conjugates (kexc = 446 nm for ODNs 4–7, 13
29 M. Hoffer, Chem. Ber., 1960, 93, 2777.
30 J. E. Marugg, M. Tromp, E. Kuyl-Yeheskiely, G. A. van der Marel
and J. H. Boom, Tetrahedron Lett., 1986, 27, 2661.
31 C. R. Petrie, M. W. Reed, A. D. Adams and R. B. Meyer, Bioconjugate
Chem., 1992, 3, 85.
32 U. Asseline and N. T. Thuong, Tetrahedron Lett., 1989, 30, 2521.
33 S. Rumney and E. T. Kool, J. Am. Chem. Soc., 1995, 117, 5635.
34 K. V. Balakin, V. A. Korshun, D. S. Esipov, I. I. Mikhalev and Y. A.
Berlin, Nucleosides Nucleotides, 1998, 18, 1279.
35 B. C. Froehler, Tetrahedron Lett., 1986, 27, 5575.
36 C. R. Cantor, M. M. Warshaw and H. Shapiro, Biopolymers, 1970,
9, 1059.
and 18; kexc = 448 nm for ODNs 8–11, 16 and 17; and kexc
449 nm for ODNs 12, 14 and 15).
=
O r g . B i o m o l . C h e m . , 2 0 0 4 , 2 , 3 4 9 6 – 3 5 0 3
3 5 0 3