Design of artificial nucleobases for the recognition of the AT inversion by triple-helix forming oligonucleotides: A structure-stability relationship study and neighbour bases effect

Article abstract of DOI:10.1016/S0968-0896(03)00229-3

We report herein on the synthesis, the incorporation into triplex forming oligonucleotides (TFO) and the recognition properties of a series of synthetic nucleosides designed for the specific recognition of an inverted A·T base pair in a pyrimidine triple

Full text of DOI:10.1016/S0968-0896(03)00229-3

Bioorganic & Medicinal Chemistry 11 (2003) 2751–2759
Design of Artificial Nucleobases for the Recognition of the AT
Inversion by Triple-Helix Forming Oligonucleotides: A Structure–
Stability Relationship Study and Neighbour Bases Effect
Dominique Guianvarc’h,a Jean-Louis Fourrey,b Rosalie Maurisse,a
Jian-Sheng Suna and Rachid Benhidab,c,
*
a
´
Laboratoire de Biophysique, UR 565 INSERM, UMR 8646 CNRS, Museum National d’Histoire Naturelle,
´
43 rue Cuvier 75231 Paris Cedex 05, France
bInstitut de Chimie des Substances Naturelles, CNRS, 1 avenue de la Terrasse, 91198 Gif-sur-Yvette, France
´ ´
Laboratoire de Chimie Bioorganique, UMR 6001 CNRS, Universite de Nice-Sophia Antipolis, Parc Valrose, 06108 Nice Cedex 2, France
c
Received 4February 2003; accepted 1 April 2003
Abstract—We report herein on the synthesis, the incorporation into triplex forming oligonucleotides (TFO) and the recognition
properties of a series of synthetic nucleosides designed for the specific recognition of an inverted Aꢀ T base pair in a pyrimidine triple
helix motif. These analogues were designed on the basis of the results obtained with our previously reported compounds S and Bt,
in order to define a structure–stability relationship. We report also on the chemical nature effect of the bases flanking S in the case
of S-containing TFOs, in order to get further informations about the recognition process within the Aꢀ TxS triplet. This study
establishes guidelines for the conception of more potent analogues for the recognition of both Aꢀ T and Gꢀ C inverted base pairs.
# 2003 Elsevier Science Ltd. All rights reserved.
Introduction
been proposed to overcome this sequence limitation.5
One of these strategies is the universal base approach
which consists into conjugating intercalating agents to
the 50- or 30-end or to internal positions of the TFO in
order to stabilise the triplex containing base-pair inter-
ruptions in the purine motif.6 An alternative approach
is the specific base strategy which calls for the synthesis
of new modified bases able to form hydrogen bonds
with one or both partners of the Aꢀ T or the Gꢀ C Wat-
son–Crick inverted base pairs in the major groove.7 This
specific base approach seems to be a promising way,
although in most cases the proposed analogues appear
to interact by intercalation rather than by forming spe-
cific H-bonds. Moreover, whereas a large number of
extended nucleobases have been engineered for the
recognition of the Gꢀ C base pair,7 compounds designed
for targeting the Aꢀ T base pair are still lacking. In this
respect, it is noteworthy that the presence of the
5-methyl group of thymine constitutes a supplementary
obstacle for the recognition of Aꢀ T inversions due to
steric hindrance.
Pyrimidine oligonucleotides can interact in a parallel
orientation with the oligopurine strand of an oligopyr-
imidineꢀ oligopurine DNA duplex through the forma-
tion of Tꢀ AxT and Cꢀ GxC+ triplets.1 This recognition
process occurs in the major groove of the double stran-
ded DNA (ds-DNA) and involves the formation of
specific Hoogsteen hydrogen bonds between the pyri-
midines of the triplex forming oligonucleotide (TFO)
and the purine Watson–Crick bases.2 Accordingly,
TFOs have attracted a considerable interest because of
their potential applications in gene expression modula-
tion and in gene targeting technologies.3 Unfortunately,
triplex formation by TFOs is limited to oligopyr-
imidineꢀ oligopurine ds-DNA targets and one observes
that a single Aꢀ T or Gꢀ C base pair interruption reduces
strongly the stability of the triplex.4 To date, the recog-
nition of mixed purine/pyrimidine sequences remains a
challenge. In the last few years, two main approaches
which involve a number of chemical modifications have
We have recently reported on the synthesis and the
recognition properties of two non-natural C-nucleoside
*Corresponding author. Tel.: +33-492-076153; fax: +33-492-076151;
e-mail: benhida@unice.fr
0968-0896/03/$ - see front matter # 2003 Elsevier Science Ltd. All rights reserved.
doi:10.1016/S0968-0896(03)00229-3

2752
D. Guianvarc’h et al. / Bioorg. Med. Chem. 11 (2003) 2751–2759
analogues S and Bt (see structure in Fig. 1) when incor-
porated into pyrimidine-motif TFOs.8 When the thi-
In order to have further insight into the recognition
process within the Aꢀ TxS triplet, we report also on the
chemical nature effect of the bases flanking S in the case
of S-containing TFOs.
azolyl-aniline derivative
S was incorporated into
oligonucleotides, the resulting TFO was found to selec-
tively recognise an Aꢀ T base pair in ds-DNA with an
affinity similar to those obtained in the case of canonical
triplets (Tꢀ AxT and Cꢀ GxC+). On the other hand, we
found quite surprisingly the triplex constructed with a
TFO containing the thiazolyl-benzimidazole derivative
Bt to be less stable than the analogue incorporating S.
Indeed, although structurally related to S, Bt is con-
formationally less flexible than S owing to extended
aromaticity of the benzimidazole ring. Accordingly,
because of favourable entropic factors, we expected Bt
to exhibit more efficient triplex stabilising properties
than S. In addition, Bt induced only a moderate dis-
crimination between the Aꢀ T and Gꢀ C base pairs, sug-
gesting a non sequence-specific interaction with the
probe, probably due to base intercalation.
Results andDiscussion
Syntheses
The preparations of the phosphoramidites of S, S0, Bt,
and B have been described previously.8 The phosphor-
amidite derivative of S1 (i.e., 8) was obtained by cou-
pling the protected 2-deoxyribose derivative 2 with the
3-iodo-aniline 3 activated as its Weinreb’s amide9
(Scheme 1). Compound 2 was prepared in 85% yield
from ester 1 according to literature.10 The protection of
the 3-hydroxyl position being performed with
TBDMSCl in DMF. Then, 3-iodo-aniline 3, activated
as its dimethylaluminium derivative, was coupled with 2
to furnish 4 in 90% yield. Compound 6 was prepared
using a previously described procedure11 by application
of a palladium catalysed Stille12 cross-coupling reaction
between the 5-(tributyltin)-thiazolyl derivative 513 and
the iodo compound 4 in 87% yield. Successive TBDMS
cleavage (Bu4NF, THF) and phosphitylation under
standard conditions of the resulting compound 7 pro-
vided phosphoramidite 8 in 53% overall yield (two
steps).
However, the presence of the amino-thiazole moiety in
these S and Bt C-nucleobases was determinant for their
recognition properties. This observation led us to fur-
ther investigate the recognition process leading to the
formation of the Aꢀ TxS and Aꢀ TxBt triplets. Accord-
ingly, we have designed a series of analogues of the
nucleobases S and Bt whose modifications were chosen
to delineate the effects of: (i) the nature of the hetero-
atom X (X=N or S; Fig. 1) facing the 6-amino group of
A within the Aꢀ T base pair, (ii) the length of the linker
(n=0 or 1) between the 2-deoxyribose unit and the
nucleobase, and (iii) the flexibility/aromaticity of the
nucleobase (i.e., aniline with Z=O or benzimidazole
with Z=N).
The synthesis of the carboxamides S1, Sa, Sa0 and Sa1
was performed starting from the protected ribonic acid
9 (Scheme 2). Thus, 9 was obtained in two steps from
3,5-di-O-toluoyl-a-1-chloro-2-deoxy-d-ribofuranose14
following nucleophilic substitution (TMSCN) and sub-
sequent hydrolysis of the 1-cyano intermediate (HCl/
dioxane).15 The amide coupling was accomplished in
high yield by condensation of 9 with either aniline 10,
compound 118a or 3 using the Mukaiyama’s reagent
(2-chloro-1-methylpyridinium iodide).16 Compound 15
was obtained following Stille-type coupling of the iodo
compound 14 with the tin derivative 5. Selective sapo-
nification of the resulting toluoyl esters 12, 13 and 15
was accomplished using sodium methanolate in metha-
nol, producing 16, 17 and 18, respectively, in good
yields. Successive 50-dimethoxytritylation and 30-phos-
phitylation of compounds 16, 17 and 18 under standard
conditions provided the corresponding phosporamidites
22, 23 and 24, respectively. Phosphoramidites 8, 22, 23
and 24 were then incorporated into the 18-mer pyri-
midine-motif TFOs (III) listed in Table 1, at the internal
position Z using automated oligonucleotide synthesis.
We report here on the synthesis of these analogues, their
incorporation into TFOs, together with their recogni-
tion properties of an Aꢀ T base-pair interruption.
Structure–stability relationship study
The triple helix binding properties of the TFOs III were
examined by means of thermal denaturation experi-
ments with the oligopyrimidineꢀ oligopurine 26-mer
duplex (Iꢀ II) containing an Aꢀ T interruption
(Xꢀ Y=Aꢀ T) (Table 1).17 All the new synthetic thiazolyl
S1, Sa and Sa1 nucleobases, showed a lower capacity to
recognise an Aꢀ T inverted base pair in ds-DNA than S,
as shown by the lower Tm values measured for their
Figure 1. Structure of the artificial nucleobases and proposed model
for their specific recognition of an inverted Aꢀ T base pair.

D. Guianvarc’h et al. / Bioorg. Med. Chem. 11 (2003) 2751–2759
2753
Scheme 1. Synthesis of phosphoramidite 8. Reagents and reaction conditions: (a) TBDMSCl, imidazole, DMF (85%); (b) AlMe3, molecular sieves
4 A, toluene, (90%); (c) 5, Pd(Ph3)4, DMF, 60 ꢂC (87%); (d) Bu4NF, THF (87%); (e) ClP(OCH2CH2CN)(NiPr)2, N,N-diisopropylethylamine,
CH2Cl2 (61%).
Scheme 2. Synthesis of phosphoramidites 22, 23 and 24. Reagents and reaction conditions: (a) chloromethylpyridinium iodide, Et3N, CH2Cl2, 60 ꢂC
(90–95%); (b) 5, Pd(Ph3)4, DMF, 60 ꢂC (92%); (c) NaOMe, MeOH (91–97%); (d) DMTCl, Et3N, DMF (80%); (e) ClP(OCH2CH2CN)(NiPr)2, N,N-
diisopropylethylamine, CH2Cl2 (63–80%).
corresponding triplexes (entries 6, 8 and 9 as compared
with entry 5). Interestingly, some observations concern-
ing the structure–stability relationship were obtained
from this study and are given below.
opposite result was obtained when going from Sa0 to Sa1
(ÁTm=ꢁ8 ꢂC). Moreover, the Tm value of Sa1-TFO is
identical to the one observed with a truncated 12-mer
TFO matching until the Z position as presented in Table 1
(Tm=28 ꢂC, entry 11). Hence, in the triplex with Sa1-TFO,
the Hoogsteen pairing is likely stopped at the Sa1 site, and
all the Sa1TTCTT bases are not associated. Furthermore,
the Tm value obtained with Sa0-TFO (Tm=36 ꢂC, entry
10) suggests that the Hoogsteen pairing of the five 30-end
bases of this TFO is partially destroyed.
Aminothiazole effect. In our preliminary work, we
showed that the presence of the aminothiazole moiety of
S or Bt was essential since incorporation of S0 or B into
the TFO led to an important loss of stability compared
with the S or Bt-based TFOs, respectively (ÁTm=ꢁ18
or ꢁ9 ꢂC, respectively).8 In line with these results, we
also found that when going from S0 to S1 or from Sa0 to
Thiazole endo-heteroatom effect. The effect of the het-
eroatom nature (X=N or S; Fig. 1) facing the 6-amino
group of A within the Aꢀ T base pair is shown by com-
Sa
a
significant increase of Tm was observed
(ÁTm=+12 or +7 ꢂC, respectively). However, the

2754
D. Guianvarc’h et al. / Bioorg. Med. Chem. 11 (2003) 2751–2759
Table 1. Sequence of the Iꢀ IIxIII triplexes and melting temperature
values (Tm ꢃ1 ꢂC)
to
a
similar observation, the decrease being
ÁTm=ꢁ16 ꢂC.
30-CGTA-TTTTCTTCTCTTXTTCTT-AGTG-50
50-GCAT-AAAAGAAGAGAAYAAGAA-TCAC-30
50-TTTTCTTCTCTTZTTCTT-30
I
II
III
Flexibility/aromaticity of nucleobase. One can observe
that Bt possesses the same ability to form hydrogen
bond contacts as Sa1. However, the Sa1-TFO forms a
less stable triplex with ds-DNA than the Bt-TFO (entries
3 and 9, ÁTm=ꢁ15 ꢂC). This low stability obtained with
Sa1-TFO corresponds to the one observed in the case of
the truncated triplex (entry 11, Tm=28 ꢂC). As there is a
great structural analogy between the aglycone part of the
50-TTTTCTTCTCTT-30
Truncated TFO
Entry
Xꢀ Y
(ds-DNA Iꢀ II)
Z
(TFO III)
Structure of
the nucleobase
Tm
(ꢂC)
1
2
Tꢀ A
Aꢀ T
T
G
—
—
51
4
two heterocyclic systems of B and S , this result should
t
5
a1
be likely ascribed to conformational factors. Indeed, in
the case of compound Sa1, the ring oxygen of deoxyr-
ibose and of the carbonyl amide function can adopt a
trans or cis relationship by rotation around the glycosidic
C10–CO bond (Fig. 2).
3
4
5
6
7
8
9
Aꢀ T
Aꢀ T
Aꢀ T
Aꢀ T
Aꢀ T
Aꢀ T
Aꢀ T
Bt
B
43
34
50
44
32
43
28
The cis conformation is expected to be defavourable
because of the oxygen–oxygen anomeric effect. Unfor-
tunately, the thermodynamically stable trans conforma-
tion induced a base orientation change which destroyed
the base pairing at the interrupted site, resulting in a
triplex destabilisation. In the case of Bt, the highly
delocalised electronic charge on the benzimidazole ring
and the hydrogen prototropy disfavoured the anomeric
effect and hence, favoured the free rotation around the
C–C glycosidic bond. In this case, both conformations
could exist without any energetic penalty. In contrast,
in the case of Sa-TFO, the moderate Tm value
(Tm=43 ꢂC, entry 8) might reflect favourable interac-
tions with the Aꢀ T base pair preventing this moiety to
adopt a trans conformation. Indeed, the best amino-
thiazole system of Sa could compensate the anomeric
effect keeping the nucleobase in the best conformation
for adenine recognition.
S
S1
S0
Sa
Sa1
Sa0
10
11
Aꢀ T
Iꢀ II
36
28
A truncated 12 mer TFO
ꢀ
Nearest-neighbour effects aroundthe A TxS non-natural
base triplet
parison of S/S1 and Sa/Sa1 (Table 1). Thus, S and Sa
differ from S1 and Sa1 respectively by the position of
attachment of the aminothiazole moiety to the phenyl
ring, the thiazole being linked at C-5 in S1 and Sa1
whereas it is connected at C-4in the case of S and Sa.
Both nucleobases possess the required amino group
(NH) for the recognition of the adenine N7 nitrogen.
We found that the stabilisation increased when going
from S1 to S (ÁTm=+6 ꢂC) or from Sa to Sa1
(ÁTm=+15 ꢂC). This result indicates that the best tri-
plex stabilisation could be obtained with aminothiazole
in which the nitrogen endo is involved in the recognition
of the adenine (NH2) Watson–Crick base rather than
sulfur.
In order to get further informations about the stabilising
effect observed within the triple helice containing the
Aꢀ TxS triplet (Table 1, entry 5), we investigated the effect
of the base triplets flanking the Aꢀ TxS one on the triplex
ꢀ
ꢀ
stabilisation process (X1 Y1xZ1 and X2 Y2xZ2 triplets).
Towards this end, S and G (as a control) were incorpo-
rated into the 18-mer TFO (III) at the internal position Z
flanked by C+ and/or T at positions Z1 and Z2. The
triple helix binding properties of these TFOs harbouring
the four possible combinations (Table 2, entries 1 to 4,
Z1ZZ2=TST, TSC+, C+ST, C+SC+) were examined
by means of thermal denaturation experiments with
their respective 26-mer ds-DNA (Iꢀ II) target.
Linker effect. Incorporation of Sa into the TFO led to a
less stable triplex compared to the one containing S
(ÁTm=ꢁ7 ꢂC). This compound only differs by the
length of the linker between the sugar unit and the
aglycone part, S featuring a supplementary methylene.
Accordingly, the length of the linker in the case of S
seems to be well adapted for the recognition of the Aꢀ T
inverted base pair. Comparison of the Tm values of S1
(entry 6, Tm=44 ꢂC) and Sa1 (entry 9, Tm=28 ꢂC) leads
Figure 2. trans and cis conformations of compound Sa1
.

D. Guianvarc’h et al. / Bioorg. Med. Chem. 11 (2003) 2751–2759
2755
Table 2. Nearest-neighbour effects within triplexes containing S (or
G) in the Hoogsteen-paired strand (III)
of more potent analogues for the recognition of both
Aꢀ T and Cꢀ G inverted base pairs. Indeed, in the con-
tinuation of this project, the synthesis of new cyclised
analogues related to the benzimidazole series is in pro-
gress, in order to increase the triplex stability and nat-
ural base pairs discrimination. Furthermore, the
neighbour effect of the base pairs flanking the Aꢀ T
inversion provides some informations about the inter-
actions stabilising the non-natural Aꢀ TxS triplet. Thus,
it was found that in all cases, the triplexes containing
the Aꢀ TxS triplet are more stable than those containing
the A TxG one. This gain of stability is more important
for the triplexes in which S is flanked on each side by T
or C+. Finally, it was found that neighbour base effect,
i.e., destabilisation in presence of C+, is the same on the
50-side as on the 30-side favouring the proposed selective
hydrogen bonds binding mode.
30-CGTA-TTTTCTTCTCTX1AX2TCTT-AGTG-50
50-GCAT-AAAAGAAGAGAY1TY2AGAA-TCAC-30
50-TTTTCTTCTCTZ1ZZ2TCTT-30
I
II
III
Entry
Triplet X1ꢀ Y1xZ1
Triplet X2ꢀ Y2xZ2
Tm (ꢂC)
Z=G
Z=S
1
2
3
4
Tꢀ AxT
Tꢀ AxT
Cꢀ GxC+
Cꢀ GxC+
Tꢀ AxT
Cꢀ GxC+
Tꢀ AxT
45
4
41
344
50
1
43
ꢀ
3
4
Cꢀ GxC+
0
As shown in Table 2, a neighbour bases effect was
observed with the non-natural Aꢀ TxS triplet. Indeed,
the triple helix was most stable when Aꢀ TxS triplet was
flanked by two Tꢀ AxT triplets. In contrast, the stability
of the triplexes decreased as the number of adjacent
Cꢀ GxC+ around Aꢀ TxS 0triplet increased. Although, we
did not find any 50- or 3 -Cꢀ GxC+ side effect (entries 2
and 3, Tm=43 ꢂC). Finally, in all cases, the Aꢀ TxS trip-
let provides a better triplex stabilisation compared to
the best natural base triplet Aꢀ TxG. This stabilisation is
moderate (ÁTm=+2 ꢂC) when the Aꢀ TxS triplet is
flanked by Tꢀ AxT and Cꢀ GxC+ neighbours in either the
50- or the 30-direction (entries 2 and 3, respectively). In
contrast, this stabilisation is higher when the Aꢀ TxS
triplet is flanked by two Tꢀ AxT, (ÁTm=+5 ꢂC) or two
Cꢀ GxC+ triplets (ÁTm=+6 ꢂC).
Experimental
General
Solvents were dried by standard methods. All reagents
were of commercial quality (Acros, Aldrich or Lan-
caster) and were used without further purification.
Melting points were determined with a Kofler apparatus
and are uncorrected. Microanalyses were performed by
the ‘Service de Microanalyses de l’ICSN-CNRS’. 1H,
13C and 31P NMR spectra were conducted on Bruker
AC250, AM300 and AM400 spectrometers. NMR
chemical shifts (d) are given in ppm relative to TMS
(1H) or H3PO4 (31P). Mass spectra were recorded on a
AEI MS-9 (CI), KRATOS MS-80-RF (FAB). Column
chromatography were carried out on Silica gel Merck
Kiesegel 60 (230–400 mesh). Analytic TLC were per-
formed using Merck silica gel 60 F254, the spots were
visualized with UV/vis light, iodine or phosphomolyb-
dic acid spray followed by heating. HPLC analyses were
recorded on Gilson apparatus. All reactions were car-
ried out under a nitrogen atmosphere under anydrous
conditions unless otherwise noted.
Previously, Dervan et al. have examined the effects of
altering the flanking sequence around an Aꢀ TxG triplet
using the affinity cleaving method.18,19 Both results are
in accordance excepting that, in our case, we did not
observe any 30- or 50-Cꢀ GxC+ side effects (entries 2 and
3, Tm=41 ꢂC).
Altogether, these results indicate that the use of com-
pound S for the recognition of an inverted Aꢀ T base
pair is more favourable than G whatever the sequence
composition. In contrast to the reported work,20 the
neighbour bases effect, in our case, is the same on both
the 30- and 50-sides. This might suggest that the mode
of interaction of S probably depends more on favour-
able hydrogen bond contacts rather than on base
intercalation. Although, the base stacking contribu-
tions will play an important role in the overall triplex
stabilisation.
Ethyl-(2-deoxy-3-O-tert-butyldimethylsilyl-5-O-dime-
thoxytrityl-ꢀ-D-ribofuranosyl)-acetate (2). To a solution
of compound 1 (216 mg, 0.42 mmol) in DMF (5 mL)
were successively added imidazole (6 equiv, 163 mg)
and TBDMSCl (3.2 equiv, 181 mg). The reaction
mixture was stirred for 16 h at room temperature,
then evaporated under reduced pressure to give a
crude oil. Silica gel column chromatography purifica-
tion using a gradient of ethyl acetate in heptane with
triethylamine (1%) afforded 2 as an oil (221 mg,
85%). Rf 0.65 (7:3 heptane/ethyl acetate). 1H NMR
(CDCl3) d 0.02 (2s, 6H), 0.83 (s, 9H), 1.21 (t, 3H, J=7.1
Hz) 1.70 (m, 1H), 1.95 (m, 1H), 2.50 (m, 2H), 2.65 (m,
2H), 3.06 (m, 2H), 3.73 (s, 6H), 3.91 (m, 1H), 4.13 (q,
2H, J=7.1 Hz), 4.25 (m, 1H), 4.50 (m, 1H), 6.80 and
7.30 (m, 8H), 7.10–7.50 (m, 5H). 13C NMR (CDCl3) d
14.21, 17.98, 25.78, 40.77, 41.09, 55.12, 60.41, 64.14,
74.25, 74.62, 85.98, 86.58, 113.04, 126.67, 127.72,
128.21, 130.08, 136.18, 145.0, 158.42, 171.10. MS (FAB)
m/z=643 [M+Na]+.
Conclusion
In summary, we synthesised a series of nucleoside ana-
logues featuring highly functionalised nucleobases. To
the best of our knowledge, S is the best nucleobase
analogues for the recognition of an Aꢀ T base pair
inversion since it gives a high increase of triplex stabi-
lity. Moreover, the structure–stability relationship study
of these compounds establishes guidelines for the design

2756
D. Guianvarc’h et al. / Bioorg. Med. Chem. 11 (2003) 2751–2759
C-(30 -O-(tert-Butyldimethylsilyl)-20 -deoxy-50 -O-dime-
thoxytrityl-ꢀ-D-ribofuranosyl-10)-N-(3-iodophenyl) car-
boxamide (4). To a solution of 2-iodo-aniline 3 (3
equiv, 212 mg) in dry toluene (5 mL) in the presence of
molecular sieves (4A ), was added dropwise at 0 ꢂC a
2 M AlMe3 solution in toluene (3 equiv, 0.5 mL) under
argon atmosphere. After stirring for 2 h at room tem-
perature, a solution of compound 2 (200 mg, 0.32
mmol) in toluene was added and the reaction mixture
was stirred for 2 h at room temperature. Then, the
reaction mixture was poured into aq NaHCO3 solution
and extracted with ethyl acetate. The organic layer was
dried over MgSO4, filtred and the solvent removed in
vacuo. The crude residue was purified by silica gel
chromatography using methylene chloride/ethyl acetate
as solvents with triethylamine (1%) to provide 4 as an
oil (230 mg, 90%). Rf 0.50 (7:3 heptane/ethyl acetate).
1H NMR (CDCl3) d 0.02 (2s, 6H), 0.85 (s, 9H), 1.75 (m,
1H), 1.95 (m, 1H), 2.49 (m, 2H), 2.62 (dd, 2H, J=3.0
and 15.0 Hz), 3.2 (m, 2H), 3.75 (s, 6H), 4.05 (m, 1H),
4.25 (m, 1H), 4.55 (m, 1H), 6.80 (m, 4H), 6.9 (t, 1H,
J=8.1 Hz), 7.10–7.45 (m, 11H), 7.85 (s, 1H), 8.55 (s,
1H). 13C NMR (CDCl3) d 4.54, 18.03, 25.86, 41.24,
43.58, 55.29, 64.07, 73.78, 74.94, 86.32, 87.24, 94.18,
113.23, 119.06, 127.93, 128.23, 128.42, 129.25, 130.10,
130.39, 132.97, 135.91, 139.40, 144.82, 158.60, 169.29.
MS (FAB) m/z=816 [M+Na]+.
(m, 1H), 2.67 (m, 1H), 3.25 (m, 2H), 3.56 (s, 3H), 3.73
(s, 6H), 4.07 (m, 1H), 4.37 (m, 1H), 4.57 (m, 1H), 6.80
and 7.30 (m, 8H, H–Ar), 7.10–7.40 (m, 7H), 7.43 (m,
1H), 7.56 (s, 1H), 7.68 (s, 1H), 8.43 (s, 1H). 13C NMR
(CDCl3) d 28.21, 33.94, 40.98, 43.57, 55.12, 64.35, 73.51,
74.91, 86.23, 86.67, 113.13, 117.21, 119.04, 121.83,
126.85, 127.83, 128.10, 129.47, 129.98, 130.02, 132.94,
133.29, 135.81, 138.60, 144.71, 153.10, 158.46, 160.88,
169.30. MS (FAB) m/z=788 [M+Na]+, 303 [DMT]+.
Phosphoramidite (8). A solution of compound 7 (200
mg, 0.26 mmol) and N,N0-diisopropylethylamine (96
mL, 2 equiv) in 5 mL of anhydrous methylene chloride,
under an inert atmosphere, was treated with 2-cyano-
ethyl diisopropylchlorophosphoramidite (68 mL, 1.2
equiv). After stirring for 15 min at room temperature,
the reaction mixture was poured into aq NaHCO3
solution, extracted and whashed twice with water. The
organic layer was separated, dried over MgSO4, filtred
and the solvent removed in vacuo. The crude residue
was purified by silica gel chromatography using hep-
tane/ethyl acetate (9/1) as solvents with triethylamine
(1%) to provide 8 as a white foam (155 mg, 61%, mix-
ture of two diastereoisomers). Rf 0.44 and 0.51 (1:1
heptane/ethyl acetate). 31P NMR (CDCl3, 243 MHz) d
145.02, 145.19.
C-(20-Deoxy-30,50-di-O-toluoyl-ꢀ-D-ribofuranosyl-10)-N-
phenylcarboxamide (12). The compound 9 (600 mg, 1.5
mmol) was dissolved in dry methylene chloride (30 mL)
and then triethylamine (2 equiv, 4.18 mL), chloromethyl
pyridinium iodide (2 equiv, 766 mg) and aniline chlor-
hydrate 10 (1 equiv, 195 mg) were added successively.
After refluxing for 1 h, the solvent was evaporated and
the crude product was extracted with methylene chlo-
ride, dried over MgSO4, filtred and evaporated. The
obtained oil was purified under silica gel chromato-
graphy using heptane/ethyl acetate (9/1) as solvants
giving 675 mg (95%) of 12 as a white solid. Rf 0.38 (7:3
heptane/ethyl acetate). Mp 123–125 ꢂC. 1H NMR
(CDCl3) d 2.34(s, 3H), 2.37 (m, 1H), 2.41 (s, 3H), 2.71
(dd, 1H, J=14.0 and 6.1 Hz), 4.56 (m, 1H), 4.67 (dd,
1H, J=4.6 and 11.9 Hz), 4.82 (m, 2H), 5.53 (d, 1H,
J=4.8 Hz), 7.01 (m, 2H), 7.26 (m, 5H), 7.61 (d, 2H,
J=8.1 Hz), 7.87 (dd, 2H, J=1.3 and 8.1 Hz), 7.96 (dd,
2H, J=1.3 and 8.1 Hz), 8.65 (s, 1H). 13C NMR (CDCl3)
d 21.67, 21.74, 37.25, 64.17, 76.05, 79.14, 84.99, 119.83,
124.40, 126.53, 128.86, 129.30, 129.74, 129.78, 137.34,
144.34, 144.45, 166.16, 166.87, 169.28. MS (IC)
m/z=474 [M+H]+, 353 [M-PhNHCO+H]+, 119
[Tol]+. Anal. calcd for C28H27NO6: C, 71.02; H, 5.75;
N, 2.96. Found: C, 70.79; H, 5.81; N, 2.72.
C-(30 -O-(tert-Butyldimethylsilyl)-20 -deoxy-50 -O-dime-
thoxytrityl-ꢀ-D-ribofuranosyl-10)-N-[3-(2-N-methyl-N-
tert-butoxycarbonyl-thiazole-5-yl)phenyl]acetamide (6).
To a solution of compound 4 (660 mg, 0.83 mmol) in
dry DMF (10 mL) were successively added, under argon
atmosphere, thiazolyltin derivative 5 (1.5 equiv, 800 mg)
and Pd(PPh3)4 (80 mg). The mixture was stirred at 60 ꢂC
for 2 h and then filtred through Celite. The filtrate was
evaporated to give a crude residue which was purified
by flash chromatography on silica gel, using 20% of
ethyl acetate in heptane with triethylamine (1%), to
afford 6 as an oil (635 mg, 87%). Rf 0.30 (7:3 heptane/
1
ethyl acetate). H NMR (CDCl3) d 0.02 (2s, 6H), 0.85
(s, 9H), 1.59 (s, 9H), 1.75 (m, 1H), 1.95 (m, 1H), 2.59
(m, 1H), 2.70 (m, 1H), 3.20 (m, 2H), 3.56 (s, 3H), 3.75
(s, 6H), 4.05 (m, 1H), 4.25 (m, 1H), 4.56 (m, 1H), 6.80
and 7.30 (m, 8H), 7.10 and 7.40 (m, 7H), 7.43 (m, 1H),
7.55 (s, 1H), 7.70 (s, 1H), 8.52 (s, 1H). 13C NMR
(CDCl3) d ꢁ4.62, 18.90, 25.79, 28.27, 33.94, 41.30,
43.69, 55.17, 64.01, 73.79, 74.95, 86.24, 87.25, 113.14,
117.18, 118.97, 121.8, 126.86, 127.84, 128.17, 129.48,
130.01, 132.90, 133.01, 133.40, 135.87, 138.73, 145.10,
153.50, 158.51, 162.30, 169.28. MS (FAB) m/z=902
[M+Na]+.
C-(20-Deoxy-50-O-dimethoxytrityl-ꢀ-D-ribofuranosyl-10)-
N-[3-(2-N-methyl-N-tert-butoxycarbonyl-thiazole-5-yl)-
phenyl]acetamide (7). Compound 6(457 mg, 0.52 mmol)
was dissolved in THF (15 mL) and 1 M TBAF in THF
(1.1 equiv, 0.6 mL) was added. The reaction mixture
was stirred at room temperature for 1 h and then eva-
porated to a syrup. The residue was purified over a silica
gel column, using 20% of ethyl acetate in heptane, to
afford 7 as a white foam (345 mg, 87%). 1H NMR
(CDCl3) d 1.59 (s, 9H), 1.90 (m, 1H), 2.08 (m, 1H), 2.59
C-(20-Deoxy-30,50-di-O-toluoyl-ꢀ-D-ribofuranosyl-10)-N-
[3-(2-acetylaminothiazole-4-yl)phenyl]carboxamide (13).
By the same procedure as described for the synthesis of
12, the acid 9 (400 mg, 1 mmol) and the amine 11 (1.5
equiv, 300 mg) were coupled to give compound 13
(white solid, 582 mg, 95%). Rf 0.44 (1:1 heptane/ethyl
acetate). Mp 204–206 ꢂC. H NMR (CDCl3) d 2.13 (s,
1
3H), 2.28 (s, 3H), 2.36 (m, 1H), 2.42 (s, 3H), 2.70 (dd,
1H, J=6.2 and 13.9 Hz), 4.57 (m, 1H), 4.67 (dd, 1H,
J=12.0 and 4.9 Hz), 4.81 (m, 2H), 5.52 (d, 1H, J=7.6

D. Guianvarc’h et al. / Bioorg. Med. Chem. 11 (2003) 2751–2759
2757
Hz), 7.05 (s, 1H), 7.08 (d, 1H, J=8.1 Hz), 7.28 (m, 4H),
7.51 (d, 1H, J=8.1 Hz), 7.60 (d, 1H, J=8.1 Hz), 7.84
(d, 2H, J=8.2 Hz), 7.93 (d, 2H, J=8.2 Hz), 8.01 (m,
1H), 8.78 (s, 1H). 13C NMR (CDCl3) d 21.71, 21.82,
23.15, 37.33, 64.28, 76.07, 79.25, 85.11, 108.18, 117.69,
119.56, 122.28, 128.98, 129.38, 129.79, 129.87, 130.17,
137.88, 149.17, 167.02, 169.56. MS (ES) m/z=614
[M+H]+. Anal. calcd for C33H31N3O7S: C, 64.59; H,
5.09; N, 6.85. Found: C, 64.28; H, 5.03; N, 6.49.
methylene chloride afforded 16 as a white solid (138 mg,
91%). Rf 0.20 (95:5 CH2Cl2/MeOH). Mp 88–90 ꢂC. H
1
NMR (CDCl3+CD3OD) d 2.23 (m, 1H), 2.34(m, 1H),
3.63 (dd, 1H, J=11.9 and 3.9 Hz), 3.77 (dd, 1H, J=3.3
and 11.9 Hz), 3.95 (m, 1H), 4.24 (m, 1H), 4.63 (t, 1H,
J=7.0 Hz), 7.04(t, 1H, J=7.3 Hz), 7.20 (m, 2H), 7.50
(d, 2H, J=7.6 Hz), 9.40 (s, 1H, NH). 13C NMR
(CDCl3+CD3OD) d 39.19, 61.24, 70.75, 77.69, 86.99,
120.34, 124.86, 128.96, 129.16, 130.02, 137.08, 172.34.
MS (ES) m/z=260 [M+Na]+, 238 [M+H]+. Anal.
calcd for C12H15NO4: C, 60.75; H, 6.37; N, 5.90. Found:
C, 60.70; H, 6.31; N, 5.85.
C-(20-Deoxy-30,50-di-O-toluoyl-ꢀ-D-ribofuranosyl-10)-N-
(3-iodophenyl)carboxamide (14). By the same procedure
as described for the synthesis of 12, the acid 9 (500 mg,
1.25 mmol) and the 2-iodo-aniline 3 (1.1 equiv, 300 mg)
were coupled to give compound 14 (white foam, 674
C-(20-Deoxy-ꢀ-D-ribofuranosyl-10)-N-[3-(2-acetylamino-
thiazole-4-yl)phenyl]carboxamide (17). By the same pro-
cedure as described for the synthesis of 16, 13 (320 mg,
0.52 mmol) was converted into compound 17 (white
solid, 190 mg, 97%). Rf 0.16 (95:5 CH2Cl2/MeOH). Mp
180–182 ꢂC. 1H NMR (CDCl3+DMSO-d6) d 2.25 (s,
3H), 2.30 (m, 2H), 3.69 (m, 1H), 3.81 (m, 1H), 3.97 (m,
1H), 4.34 (m, 1H), 4.64 (t, 1H, J=7.0 Hz), 5.02 (d, 1H,
J=4.3 Hz), 5.32 (t, 1H, J=4.8 Hz), 7.22 (s, 1H), 7.33 (t,
1H, J=7.9 Hz), 7.55 (t, 2H, J=8.3 Hz), 8.24(s, 1H),
9.85 (s, 1H), 10.92 (s, 1H). 13C NMR (CDCl3+DMSO-
d6) d 21.88, 38.58, 59.77, 69.06, 77.04, 86.33, 106.66,
1
mg, 90%). Rf 0.5 (7:3 heptane/ethyl acetate). H NMR
(CDCl3) d 2.30 (m, 1H), 2.38 (s, 3H), 2.43 (s, 3H), 2.74
(m, 1H), 4.56 (m, 1H), 4.64 (dd, 1H, J=11.7 and 4.8
Hz), 4.76 (dd, 1H, J=6.1 and 10.6 Hz), 4.85 (dd, 1H,
J=2.1 and 11.7 Hz), 5.53 (d, 1H, J=5.7 Hz), 7.00 (t,
1H, J=7.9 Hz), 7.16 (d, 2H, J=8.0 Hz), 7.28 (d, 4H,
J=7.7 Hz), 7.42 (d, 1H, J=7.8 Hz), 7.64(dd, 1H,
J=1.3 and 8.1 Hz), 7.86 (d, 2H, J=8.3 Hz), 7.99 (d, 2H,
J=8.3 Hz), 8.00 (d, 1H, J=1.7 Hz), 8.65 (s, 1H). 13C
NMR (CDCl3) d 21.88, 37.35, 64.19, 76.08, 79.22, 85.26,
90.94, 119.07, 126.41, 128.59, 129.42, 129.49, 129.85,
130.41, 133.45, 138.60, 144.63, 166.28, 167.16, 169.45.
MS (IC) m/z=600 [M+H]+, 4 74 [ꢁMI+H]+, 119
[Tol]+. Anal. calcd for C28H26INO6: C, 56.11; H, 4.37;
N, 2.34. Found: C, 56.29; H, 4.44; N, 2.45.
116.77, 118.53, 120.60, 128.09, 134.44, 137.68, 148.41,
+
157.50, 168.09, 171.18. MS (FAB) m/z=384[M+Li]
.
Anal. calcd for C17H19N3O5S: C, 54.10; H, 5.07; N,
11.13. Found: C, 53.81; H, 5.11; N, 10.89.
C-(20-Deoxy-ꢀ-D-ribofuranosyl-10)-N-[3-(2-N-methyl-N-
tert - butoxycarbonyl - thiazole - 5 - yl)phenyl]carboxamide
(18). By the same procedure as described for the synth-
esis of 16, 15 (491 mg, 0.71 mmol) was converted into
compound 18 (white solid, 300ꢂmg, 93%). Rf 0.17 (95:5
C-(20-Deoxy-30,50-di-O-toluoyl-ꢀ-D-ribofuranosyl-10)-N-
[3-(2-N-methyl-N-tert-butoxycarbonyl-thiazole-5-yl) phe-
nyl]carboxamide (15). To a solution of compound 14
(530 mg, 0.88 mmol) in dry DMF (10 mL) were succes-
sively added, under Argon atmosphere, thiazolyltin
derivative 5 (1.2 equiv, 542 mg), Pd(PPh3)4 (60 mg) and
CuI (10 mg). The mixture was stirred at 65 ꢂC for 1 h
and then filtred through Celite. The filtrate was evapo-
rated to give a crude residue which was purified by flash
chromatography on silica gel, using 20% of ethyl ace-
tate in heptane, to afford 15 as a white foam (555 mg,
92%). Rf 0.69 (1:1 heptane/ethyl acetate). 1H NMR
(CDCl3) d 1.60 (s, 9H), 2.32 (s, 3H), 2.35 (m, 1H), 2.43
(s, 3H), 2.75 (m, 1H), 3.57 (s, 3H), 4.57 (m, 1H), 4.70
(dd, 1H, J=12.0 and 4.9 Hz), 4.82 (m, 2H), 5.54 (d, 1H,
J=5.7 Hz), 7.12 (d, 2H, J=7.9 Hz), 7.27 (m, 4H), 7.58
(m, 2H), 7.77 (s, 1H), 7.87 (d, 2H, J=8.2 Hz), 7.93 (d,
2H, J=8.2 Hz), 8.71 (s, 1H). 13C NMR (CDCl3) d
21.73, 21.83, 28.34, 34.04, 37.35, 64.26, 76.11, 79.29,
85.22, 117.30, 118.95, 122.22, 126.53, 129.41, 129.55,
129.85, 132.87, 133.08, 133.47, 138.04, 144.54, 166.0,
167.03, 169.51. MS (FAB) m/z=708 [M+Na]+, 119
[Tol]+. Anal. calcd for C37H39N3O8S: C, 64.80; H, 5.73;
N, 6.13. Found: C, 64.56; H, 5.74; N, 5.96.
1
CH2Cl2/MeOH). Mp 190–192 C. H NMR (CDCl3) d
1.59 (s, 9H), 2.33 (m, 2H), 3.55 (s, 3H), 3.73 (m, 1H),
3.85 (m, 1H), 3.94 (m, 1H), 4.34 (m, 1H), 4.63 (t, 1H,
J=7.2 Hz), 4.91 (d, 1H, J=4.6 Hz), 5.23 (t, 1H, J=5.1
Hz), 7.26 (m, 2H), 7.62 (d, 2H, J=9.7 Hz), 7.84(s, 1H),
9.81 (s, 1H). 13C NMR (CDCl3) d 27.50, 33.22, 39.09,
60.01, 69.27, 77.27, 82.50, 86.56, 116.44, 118.43, 120.76,
121.43, 123.83, 128.72, 131.82, 132.18, 132.60, 160.05,
171.59. MS (ES) m/z=472 [M+Na]+, 450 [M+H]+.
Anal. calcd for C21H27N3O6S: C, 56.11; H, 6.05; N,
9.35. Found: C, 56.01; H, 5.79; N, 9.22.
C-(20-Deoxy-50-O-dimethoxytrityl-ꢀ-D-ribofuranosyl-10)-
N-phenylcarboxamide (19). Compound 16 (120 mg, 0.5
mmol) was dissolved in dry DMF (8 mL). Triethyl-
amine (3 equiv, 0.2 mL), and DMTCl (1.5 equiv, 250
mg) were successively added. The mixture was stirred at
room temperature for 4h, then quenched by addition of
methanol and solvents were evaporated. The crude
product was dissolved in ethyl acetate and the organic
phase washed with water, dried (MgSO4), and con-
centrated in vacuo. Silica gel column chromatography
purification using 30% of ethyl acetate, 1% of triethyl-
amine in heptane afforded 19 as a white foam (215 mg,
80%). Rf 0.26 (1:1 heptane/ethyl acetate). 1H NMR
(CDCl3) d 2.38 (m, 2H), 3.21 (dd, 1H, J=4.4 and 10.3
Hz), 3.42 (dd, 1H, J=3.9 and 10.3 Hz), 3.74(s, 6H),
4.13 (m, 1H), 4.36 (m, 1H), 4.72 (t, 1H, J=7.8 Hz), 6.77
C-(20-Deoxy-ꢀ-D-ribofuranosyl-10)-N-phenylcarboxamide
(16). To a stirred solution of compound 12 (300 mg,
0.634mmol) in dry MeOH (10 mL) was added MeONa
(3 equiv, 102 mg). The reaction mixture was stirred at
room temperature for 16 h, then evaporated under
reduced pressure to give a crude oil. Silica gel column
chromatography purification using 5% of methanol in

2758
D. Guianvarc’h et al. / Bioorg. Med. Chem. 11 (2003) 2751–2759
(dd, 4H, J=1.4and 8.8 Hz), 7.09 (m, 1H), 7.15–7.41 (m,
16H), 8.60 (s, 1H). 13C NMR (CDCl3) d 38.93, 54.87,
62.89, 66.05, 72.24, 76.28, 86.06, 86.26, 112.91, 119.65,
124.07, 126.68, 127.63, 127.77, 128.54, 129.65, 135.30,
136.67, 144.01, 158.27, 170.49. MS (ES) m/z=562
[M+Na]+, 303 [DMT]+.
phosphoramidite 23 (white foam, 140 mg, 73%, mixture
of two diastereoisomers). Rf 0.30 and 0.38 (1:1 heptane/
ethyl acetate). 31P NMR (CDCl3, 243 MHz) d 147.08,
147.86. MS (ES) m/z=903 [M+Na]+, 881 [M+H]+,
303 [DMT]+.
Phosphoramidite (24). By the same procedure as descri-
bed above for the synthesis of phosphoramidite 22,
compound 21 (200 mg, 0.26 mmol) was converted into
phosphoramidite 24 (white foam, 200 mg, 80%, mixture
of two diastereoisomers). Rf 0.56 and 0.60 (1:1 heptane/
ethyl acetate). 31P NMR (CDCl3, 243 MHz) d 147.02,
147.83. MS (ES) m/z=952 [M+H]+, 303 [DMT]+.
C-(20-Deoxy-50-O-dimethoxytrityl-ꢀ-D-ribofuranosyl-10)-
N-[3-(2-acetylaminothiazole-4-yl)phenyl]
carboxamide
(20). By the same procedure as described for the syn-
thesis of 19, 17 (150 mg, 0.4mmol) was converted into
compound 20 (white foam, 220 mg, 81%). Rf 0.30 (1:4
heptane/ethyl acetate). 1H NMR (CDCl3) d 2.20 (s, 3H),
2.45 (m, 2H), 3.23 (dd, 1H, J=10.5 and 4.5 Hz), 3.43
(dd, 1H, J=3.5 and 10.5 Hz), 3.71 (s, 6H), 4.15 (m, 1H),
4.40 (m, 1H), 4.79 (t, 1H, J=7.7 Hz), 6.75 (d, 4H,
J=8.7 Hz), 6.86 (s, 1H), 7.18–7.49 (m, 13H), 7.90 (s,
1H), 8.77 (s, 1H), 10.10 (s, 1H). 13C NMR (CDCl3) d
23.18, 39.52, 55.30, 72.55, 78.04, 86.68, 108.18, 113.36,
118.04, 119.56, 122.05, 127.10, 128.06, 128.17, 129.28,
130.07, 135.13, 135.61, 137.47, 144.46, 149.50, 158.25,
158.64, 168.10, 171.48. MS (FAB) m/z=680 [M+H]+,
303 [DMT]+.
Automatedoligonucleotides synthesis
Oligonucleotides 18-mer were synthesized on a 1-mmol
scale using an Biosystem 392 synthesizer employing the
standard phosphoramidite chemistry. Following the
solid-phase synthesis of oligonucleotides, they were
cleaved from the solid support by treatment with con-
centrated ammonia (30%) for 24h at room tempera-
ture. The 50-DMT protected oligonucleotides were
purified by reversed-phase HPLC (Waters PrepPak
Cartridge Delta-Pak C18, 15 mm, 300 A, 25*100 mm)
using a 30 min linear gradient of solvent A (0.1 M tri-
ethylammonium acetate buffer (pH 7) containing 7% of
CH3CN) and solvent B (CH3CN) (100:0 to 60:40) at a
flow rate of 5 mL/min. The DMT and Boc protecting
groups were cleaved by treatment with 80% aq AcOH
solution for 60 min at room temperature and the mixture
was extracted twice with diethylether (20 mL). The oli-
gomers were precipitated using a AcONa buffer (pH
6.2)/EtOH solution and centrifugation at 0 ꢂC for 15
min. The pure oligonucleotides were subjected to analy-
tic HPLC (Waters Delta-Pak C18, 5 mm, 300 A, 3.9*150
mm), UV (260 nm) analysis and MALDI-TOF mass
spectrometry for identification. MALDI-TOF-MS
18TC-Sa1: calcd for C178H232N39O119P17 (MꢁH)ꢁ: 5348,
Found 5348. 18TC-Sa1: calcd for C182H236N41O119 P17S
(MꢁH)ꢁ: 5461, Found: 5461. 18TC-Sa: calcd for
C183H236N41O120P17S (MꢁH)ꢁ: 5491, Found: 5489.
18TC-S1: calcd for C183H238N41O119P17S (MꢁH)ꢁ:
C-(20-Deoxy-50-O-dimethoxytrityl-ꢀ-D-ribofuranosyl-10)-
N-[3-(2-N-methyl-N-tert-butoxycarbonyl-thiazole-5-yl)-
phenyl]carboxamide (21). By the same procedure as
described for the synthesis of 19, 18 (220 mg, 0.49
mmol) was converted into compound 21 (white foam,
1
300 mg, 81%). Rf 0.40 (1:1 heptane/ethyl acetate). H
NMR (CDCl3) d 1.59 (s, 9H), 2.42 (m, 2H), 3.25 (dd,
1H, J=10.4 and 4.3 Hz), 3.40 (dd, 1H, J=10.4and 3.5
Hz), 3.54 (s, 3H), 3.68 (s, 6H), 4.14 (m, 1H), 4.41 (dd,
1H, J=4.1 and 7.9 Hz), 4.73 (t, 1H, J=7.8 Hz), 6.72 (d,
4H, J=8.6 Hz), 7.15–7.28 (m, 11H), 7.40 (m, 2H), 7.59
(s, 1H), 8.71 (s, 1H). 13C NMR (CDCl3) d 27.68, 33.41,
38.81, 54.60, 62.69, 71.82, 77.41, 82.81, 86.05, 112.70,
116.86, 118.59, 121.63, 126.50, 127.42, 127.60, 128.92,
129.42, 132.21, 132.78, 135.02, 135.09, 137.12, 143.76,
+
158.04, 160.35, 170.69. MS (ES) m/z=774[M+Na]
752 [M+H]+, 303 [DMT]+.
,
Phosphoramidite (22). A solution of compound 19 (150
mg, 0.27 mmol) and N,N0-diisopropylethylamine (2
equiv, 100 mL) in 5 mL of anhydrous methylene chlor-
ide, under an inert atmosphere, was treated with 2-cya-
noethyl diisopropyl-chlorophosphoramidite (70 mL, 1.2
equiv). After stiring for 15 min at room temperature,
the reaction mixture was poured into aq NaHCO3
solution, extracted and whashed twice with water. The
organic layer was separated, dried over MgSO4, filtred
and the solvent removed in vacuo. The crude residue
was purified by silica gel chromatography using ethyl
acetate/heptane (1/9) as solvants with triethylamine
(1%) to provide 22 as a white foam (130 mg, 63%,
mixture of two diastereoisomers). Rf 0.57 and 0.62 (1:1
heptane/ethyl acetate). 31P NMR (CDCl3, 243 MHz) d
147.00, 147.84. MS (ES) m/z=762 [M+Na]+, 303
[DMT]+.
5477,
Found:
5475.
18TC-TSC:
calcd
for
C184H238N41O120P17S (M-H)ꢁ: 5503, Found: 5500.
18TC-CST: calcd for C183H237N42O119 P17S (MꢁH)ꢁ:
5488,
Found:
5488.
18TC-CSC
calcd
for
C182H236N43O118P17S, (MꢁH)ꢁ: 5473), Found: 5471.
Melting temperature experiments
All thermal denaturation studies were carried out on a
Uvikon 940 spectrophotometer and interfaced to an
IBM-AT computer for data collection and analysis.
Temperature control of the cell holder was achieved by
a Haake D8 circulating water bath. The temperature of
the water bath was increased from 0 to 90 ꢂC and
decreased back to 0 ꢂC at a rate of 0.15 ꢂC/min with
absorption readings at 260, 295, 500 nm taken every
1 ꢂC. The maxima of the first derivatives of the melting
curves gave a good approximation of the half-dissocia-
tion temperature (Tm) and allowed us to characterise the
stabilities of the complexes in a reproducible way. The
Phosphoramidite (23). By the same procedure as descri-
bed above for the synthesis of phosphoramidite 22,
compound 20 (150 mg, 0.22 mmol) was converted into

D. Guianvarc’h et al. / Bioorg. Med. Chem. 11 (2003) 2751–2759
2759
DNA melting experiments were carried out in a 10-mM
cacodylate buffer (pH 6.0) containing 100 mM NaCl, 10
mM MgCl2 and 0.5 mM spermine.
4264. (d) Kukreti, S.; Sun, J. S.; Loakes, D.; Brown, D. M.;
Nguyen, C. H.; Bisagni, E.; Garestier, T.; Helene, C. Nucleic
Acids Res. 1998, 26, 2179. (e) Gianolio, D. A.; McLaughlin,
L. W. J. Am. Chem. Soc. 1999, 121, 6334.
8. Communications: (a) Guianvarc’h, D.; Benhida, R. Four-
rey, J-L., Maurisse, R. and Sun, J. S., Chem. Commun 2001,
1814. (b) Guianvarc’h, D.; Fourrey, J.-L.; Maurisse, R.; Sun,
J. S.; Benhida, R. Org. Lett. 2002, 4, 4209.
9. (a) Basha, A.; Lipton, M.; Weinreb, S. M. Tetrahedron
Lett. 1977, 48, 4171. (b) Lipton, M. F.; Basha, A.; Weinreb,
M. Org. Synth. 1979, 59, 4 9.
10. Hovinen, J.; Salo, H. J. Chem. Soc., PerkinTrans. 1 1997,
3017.
11. Benhida, R.; Lecubin, F.; Fourrey, J.-L.; Quintero, L.
TetrahedronLett. 1999, 40, 5701.
Acknowledgements
The authors thank M. Thomas for oligonucleotide
synthesis, Dr. J.-C. Blais for MALDI-TOF/MS charac-
terisation of TFO. We also acknowledge the ANRS for
financial support and CNRS for a doctoral fellowship
(BDI) to DG. Thanks are also due to Dr. P. Vierling for
useful discussions.
12. (a) Stille, J. K. J. Am. Chem. Soc. 1979, 4992. (b) Stille,
J. K. Angew. Chem., Int. Ed. Engl. 1986, 25, 508.
13. Compound 5 was obtained in three steps from 2-ami-
nothiazole (see ref 8b).
References andNotes
1. The symbols ꢀ and x indicate Watson–Crick and Hoogsteen
hydrogen bonds, respectively.
14. Hoffer, M. Chem. Ber. 1960, 93, 2777.
15. We observed that the procedure using diethylaluminium
cyanide (Iyer, R.P.; Phillips, L.R.; Egan, W. Synth. Commun.
1991, 21, 2053) or sodium or potassium cyanide (Kolb, A.;
Huynh Dinh, T.; Igolen, J. Bull. Soc. Chim. France 1973, 3447)
gave low diastereoselectivity and low chemical yields because
the deprotection of toluoyl protecting groups was principally
observed. The introduction of nitrile group was done using
2. (a) Le Doan, T.; Perrouault, L.; Praseuth, D.; Habhoub, N.;
Decout, J. L.; Thuong, N. T.; Lhomme, J.; Helene, C. Nucleic Acids
Res. 1987, 15, 7749. (b) Moser, H. E.; Dervan, P. B. Science
1987, 238, 645. (c) For reviews see: Thuong, N. T.; Helene, C.
Angew. Chem., Int. Ed. Engl. 1993, 32, 666. (d) Frank-Kame-
netskii, M. T.; Mirkin, S. M. Annu. Rev. Biochem. 1995, 64,
65. (e) Neidle, S. Anti-Cancer Drug Des. 1997, 12, 433.
3. (a) Maher, J. L., III Cancer Invest. 1996, 14, 66. (b) Gio-
vannangeli, C.; Helene, C. Antisense Nucleic Acid Drug Dev.
1997, 7, 413. (c) Vasquez, K. M.; Wilson, J. H. Trends Bio-
chem. Sci. 1998, 23, 4. (d) Praseuth, D.; Guieysse-Peugeot,
A. L.; Helene, C. Biochim. Biophys. Acta 1999, 1489, 181.
4. (a) Mergny, J. L.; Sun, J. S.; Rougee, M.; Garestier, T.;
Barcelo, F.; Chomilier, J.; Helene, C. Biochemistry 1991, 30,
9791. (b) Greenberg, W. A.; Dervan, P. B. J. Am. Chem. Soc.
1995, 117, 5016. (c) Best, G. C.; Dervan, P. B. J. Am. Chem.
Soc. 1995, 117, 1187.
.
TMSCN and BF3 Et2O leading to a 1:3 mixture of a/b anom-
ers in 90% yield.
16. Mukayiama, T.; Usui, M.; Shimada, E.; Saigo, K. Chem.
Lett. 1975, 1045.
17. The DNA melting experiments were carried out in a 10
mM cacodylate buffer (pH 6.0) containing 0.1 M NaCl, 10
mM MgCl2 and 0.5 mM spermine.
18. Kiessling, L. L.; Griffin, L. C.; Dervan, P. B. Biochemistry
1992, 31, 2829.
19. (a) NMR structural studies of the Aꢀ TxG triplet have
explained this stabilization by the existence of a favourable
stacking interaction between guanine (of the triplet Aꢀ TxG)
and the thymine on the 50-side (Radhakrishnan, I.; Gao, X.;
De Los Santos, C.; Live, D.; Patel, D. J. Biochemistry 1991,
30, 9022). (b) Such a favourable stacking can also explain the
lower stability obtained in presence of C+ at positions Z1 and/
or Z2 when Z=S.
5. Reviews: (a) Sun, J. S.; Helene, C. Cur. Opin. Struct. Biol.
1993, 3, 345. (b) Doronina, S. O.; Behr, J. P. Chem. Soc. Rev.
1997, 63. (c) Gowers, D. M.; Fox, K. R. Nucleic Acids Res.
1999, 27, 1569, and references cited therein.
6. (a) Griffin, L. C.; Kiessling, L. L.; Beal, P. A.; Gillespie, P.;
Dervan, P. B. J. Am. Chem. Soc. 1992, 114, 7976. (b) Koshlap,
K. M.; Gillespie, P.; Dervan, P. B.; Feigon, J. J. Am. Chem.
Soc. 1993, 115, 7908. (c) Huang, C. Y.; Cushman, C. D.;
Miller, P. S. J. Org. Chem. 1993, 58, 5048. (d) Huang, C. Y.;
Miller, P. S. J. Am. Chem. Soc. 1993, 15, 10456. (e) Huang,
C. Y.; Bi, G.; Miller, P. S. Nucleic Acids Res. 1996, 24, 2606.
(f) Lehmann, T. E.; Greenberg, W. A.; Liberles, D. A.; Wada,
C. K.; Dervan, P. B. Helv. Chim. Acta 1997, 80, 2002.
7. (a) Durland, R. H.; Rao, T. S.; Bodepudi, V.; Seth, D. M.;
Jayaraman, K.; Revankar, G. R. Nucleic Acids Res. 1995, 23,
647. (b) Zhou, B.; Puga, E.; Sun, J. S.; Garestier, T.; Helene,
C. J. Am. Chem. Soc. 1995, 117, 10425. (c) Kukreti, S.; Sun,
J. S.; Garestier, T.; Helene, C. Nucleic Acids Res. 1997, 25,
20. In a similar work on compound D3, a 4-(5-benzimidazo-
phenyl)imidazole nucleobase that was found to equally stabi-
lise the triplex at both Aꢀ T and Gꢀ C sites, Dervan et al. have
shown that in a triplex containing an Aꢀ TxD3 triplet, only the
30-neighbouring triplet affected the triplex stability.18 Later,
NMR experiments showed D3 to be intercalated between the
Aꢀ T inverted base pair and the 30 neighbouring triplet thus
explaining the 30 neighbour effect (Koshlap, K.M.; Gillespie,
P.; Dervan, P.B.; Feigon, J. J. Am. Chem. Soc. 1993, 115,
7908). In the case of S, this absence of 30 or 50 effect seems to
be in favour of a recognition process which involves hydrogen
bond contacts.