Efficient synthesis of extended guanine analogues designed for recognition of an A·T inverted base pair in triple helix based-strategy

Article abstract of DOI:10.1016/j.tetlet.2004.06.095

We report herein an efficient synthesis of new nucleosides N1, N2, and N3 as extended guanine analogues, derived from aminobenzimidazole and thymine or 5-substituted uracil. These nucleosides were devised for the recognition of an A??T inverted base pair

Full text of DOI:10.1016/j.tetlet.2004.06.095

Tetrahedron
Letters
Tetrahedron Letters 45 (2004) 6243–6247
Efficient synthesis of extended guanine analogues designed
for recognition of an AÆT inverted base pair in triple helix
based-strategy
Nathalie Van Craynest, Dominique Guianvarcꢀh,ꢁ Corinne Peyron and
Rachid Benhida*
´
Laboratoire de Chimie Bioorganique UMR-CNRS 6001, Universite de Nice-Sophia Antipolis, Parc Valrose,
06108 Nice Cedex 2, France
Received 15 June 2004; accepted 22 June 2004
Abstract—We report herein an efficient synthesis of new nucleosides N1, N2, and N3 as extended guanine analogues, derived from
aminobenzimidazole and thymine or 5-substituted uracil. These nucleosides were devised for the recognition of an AÆT inverted base
pair by three hydrogen bonds, in triple helix-based technology.
ꢀ 2004 Elsevier Ltd. All rights reserved.
It is currently accepted that pyrimidine oligonucleotides
can bind in parallel orientation to a double stranded
DNA (ds-DNA) to form stable local triple-helical struc-
tures. This specific recognition process occurs in the
major groove of ds-DNA by formation of TÆA·T and
CÆG·C+ triplets (i.e., pyrimidine motif) through specific
Hoogsteen hydrogen bonds (Fig. 1).1 Accordingly, tri-
plex forming oligonucleotides (TFOs) have attracted a
considerable interest because of their potential use in
genes recognition-based technologies.2 Unfortunately,
the application of TFO approach in therapeutic and
biotechnology has a major intrinsic limitation since tri-
plex formation is restricted to oligopyrimidineÆoligo-
purine tracts. Thus, a single purineÆpyrimidine (AÆT or
GÆC) interruption in the oligopyrimidineÆoligopurine
ds-DNA is sufficient to induce high triplex destabiliza-
tion.3
Many efforts have been undertaken to extend the recog-
nition pattern of TFOs to polypurine sequences, which
are interrupted by a pyrimidine.4 However, the most re-
ported works in this area were addressed for the recog-
Figure 1. Hydrogen bonding within TÆA·T, CÆG·C+ and AÆT·G
triplets.
nition of GÆC rather than AÆT. Indeed, the recognition
of AÆT is highly hampered by the bulky methyl group
of thymine due to steric hindrance and hence, only a
few systems were devised for this purpose.
Keywords: Extended nucleosides; Triple helix; Molecular recognition.
*
Corresponding author. Tel.: +33-(0)4-9207-6153; fax: +33-(0)4-9207-
6151; e-mail: benhida@unice.fr
ꢁ Present address: Laboratoire de Chimie Organique Biologique,
We have previously reported on the recognition proper-
ties of several non-natural C-nucleosides, designed for
´
Universite Paris VI, UMR CNRS 7613, 4 Place Jussieu, 75252 Paris
cedex 05, France.
0040-4039/$ - see front matter ꢀ 2004 Elsevier Ltd. All rights reserved.
doi:10.1016/j.tetlet.2004.06.095

6244
N. Van Craynest et al. / Tetrahedron Letters 45 (2004) 6243–6247
Figure 2. Proposed recognition model within AÆT·N1 and AÆT·N2(N3) triplets.
the AÆT inverted base pair recognition on the basis of
molecular modeling studies, when incorporated into py-
rimidine-motif TFOs.5 Among these compounds, we
found that an analogue containing two unfused aro-
matic rings (3-aminophenyl-thiazole) linked to a 20-deo-
xyribose unit by an acetamide motif (named S) had
interesting properties. Indeed, it was able to recognize
an AÆT inversion in ds-DNA with a high affinity, very
close to those of canonical triplets (TÆA·T and
CÆG·C+), and a moderate but significant selectivity
for AÆT rather than GÆC inversion.5a,c In continuation
of our studies directed at overcoming the sequence lim-
itations in triple helix strategy,5 we investigated a novel
class of modified nucleobases designed for specific rec-
ognition of AÆT inversion. This study is based on the fact
that the guanine is known to recognize a T interruption
in oligopurine strand through formation of one hydro-
gen bond between the 2-amino group of guanine and
the oxygen O(4) of thymine (Fig. 1).6 Moreover, the
non-canonical AÆT·G triplet was found to be the most
stable triplet combination with natural nucleobases.
However, this stabilizing effect by guanine is insufficient
because it involves only one hydrogen bond. From this
fact, a promising way to increase triplex stability at
the AÆT inverted site should be the use of an extended
guanine analogue that can simultaneously and coopera-
tively bind to all free Hoogsteen sites of both thymine
and adenine Watson–Crick base pair with increase of
bases stacking.
We report herein the synthesis of novel extended nucle-
osides N1, N2, and N3 designed to contain the comple-
mentary donor–acceptor–donor (dad) binding sites for
the formation of base triplets as shown in Figure 2. In
these nucleobases, the thymine was employed to achieve
adenine recognition in Hoogsteen (N2 and N3) or reverse
Hoogsteen (N1) like configuration, and the 6-amino-
benzimidazole was chosen as a guanine analogue for
NH(6) benzimidazole-O(4) thymine interaction (Fig.
2). N2 and N3 only differ by the nature of their linker
(between aminobenzimidazole and thymine), which is
conformationally restricted in N2 and flexible in N3
(more degrees of freedom).
First, the common starting material 20-deoxynucleoside
1 was prepared by simple glycosylation of 5(6)-nitro-
benzimidazole and flash chromatography separation of
the obtained regioisomers.7 Catalytic hydrogenation of
1 provided quantitatively the amine derivative 2, which
was then coupled with 1-thyminyl-propionic acid8 using
Mukaiyamaꢀs reagent (2-chloro-1-methyl-pyridinium
iodide)9 to give the extended nucleoside 3 in 56% non-
optimized yield (Scheme 1). The use of standard
HOBT/EDC or pentafluorophenol activating agents
Scheme 1. Reagents and conditions: (a) H2–Pd/C, THF; (b) 1-thyminyl-propionic acid (1.2equiv), Mukaiyamaꢀs reagent (2equiv), Et3N (2equiv),
CH2Cl2, 50ꢁC; (c) MeONa (3equiv), MeOH; (d) DMTCl (3equiv), Et3N (3equiv), DMF; (e) (CN(CH2)2O)(iPr2N)PCl (2equiv), Hunigꢀs base
(2equiv).

N. Van Craynest et al. / Tetrahedron Letters 45 (2004) 6243–6247
6245
was found to be ineffective in such a reaction. The tolu-
oyl protecting groups were then cleaved (MeONa/
MeOH) leading to the free nucleoside 4 in 86% yield.
Successive 50-dimethoxytritylation and 30-phosphityla-
tion afforded the phosphoramidite derivative 5 in 52%
overall yield (Scheme 1).10
desired pyrimidine ended nucleoside 9 was obtained in
36% yield (path A). Alternatively 9 was best obtained
from 2 by simply inverting the reaction sequences (path
B). Thus, treatment of 2 with acryloyl chloride in THF
gave 8 which, after Pd coupling with 6 afforded 9 in
80% overall yield. Compound 9 was then quantitatively
converted to its free analogue 10 by using NH3/EtOH
solution instead of MeONa/MeOH, which induced, as
expected, the cleavage of SEM protecting group. The
50-dimethoxytritylation and 30-phosphitylation steps on
compound 10 were achieved in 72% overall yield, pro-
viding the fully protected phosphoramidite 11.
In the case of nucleosides N2 and N3, their incorpora-
tion into oligonucleotides requires the protection of
the reactive N-1 position of pyrimidine base. Hence,
we have selected for our synthesis the trimethylsilyleth-
oxymethyl (SEM) as a protecting group. This group
seems to be a good candidate because it could be easily
introduced and cleaved from oligonucleotides. The syn-
thesis of the targeted phosphoramidites 11 and 13, rela-
tive to N2 and N3, respectively, is outlined in Scheme 2.
In the same way, phosphoramidite building block 13
(i.e., N3), a flexible analogue of N2, was obtained in
good yield from intermediate 10 following (i) catalytic
hydrogenation (Pd/C, THF/AcOH) and (ii) successive
50- and 30-functionalization (Scheme 2).13
We first tried to obtain 7 following Heck coupling of 611
and methylacrylate then saponification (aqNaOH, diox-
ane) but, surprisingly, we observed the removal of SEM
protecting group during NaOH treatment.12 To avoid
the saponification step, transformation of 6 to 7 was
achieved in 62% yield by using acrylic acid in one oper-
ation. However, when 7 was coupled with amine 2, the
In order to point out the implication of the pyrimidine
ring of N1 and N2 in the proposed recognition process
(Fig. 2), we have efficiently synthesized the phosphoram-
idite guanine analogue 15 containing only one hydrogen
donor site (NH2), as a control. The 6-amino group was
Scheme 2. Reagents and conditions: (a) N,O-bis(trimethylsilyl)acetamide (BSA) (2.2equiv), SEMCl (1.2equiv), Bu4N+Iꢀ (10% molar), CH2Cl2,
50ꢁC; (b) acrylic acid (2.5equiv), Pd(OAc)2 (0.1equiv), PPh3 (0.2equiv), Et3N (2equiv), DMF, 100ꢁC; (c) Mukaiyamaꢀs reagent (2equiv), Et3N
(2equiv), CH2Cl2/DMF; (d) acryloyl chloride (1.5equiv), Et3N (2equiv), CH2Cl2; (e) Pd(OAc)2 (0.1equiv), PPh3 (0.2equiv), Et3N (2equiv), dioxane,
reflux; (f) EtOH, aq NH3; (g) DMTCl (2equiv), Et3N (3equiv), DMF then (CN(CH2)2O)(iPr2N)PCl (1.2equiv), Hunigꢀs base (2equiv), CH2Cl2; (h)
Pd/C (10% molar), AcOH/THF (1/1, v/v).

6246
N. Van Craynest et al. / Tetrahedron Letters 45 (2004) 6243–6247
Scheme 3. Reagents and conditions: (a) DMA–DMF, 80ꢁC; (b) MeONa (2.2equiv), MeOH; (c) DMTCl (2equiv), Et3N (3equiv), DMF; (d)
(CN(CH2)2O)(iPr2N)PCl (1.2equiv), Hunigꢀs base (2equiv), CH2Cl2.
1994, 2, 17–32; (f) Gowers, D. M.; Fox, K. R. Nucleic
Acids Res. 1997, 25, 3787–3794.
protected, for oligonucleotides synthesis, using dimeth-
ylformamide dimethylacetal (DMA) in DMF (Scheme
3).14
7. In our case, the glycosylation of 5(6)-nitrobenzimidazole
in the presence of a-2-deoxy-3,5-ditoluoylribosyl chloride
using NaH in CH3CN yielded the b-nucleoside 2 as a
mixture of 5- and 6-regioisomers in nearly 1/1 ratio. The
structure of both regioisomers was assigned by 2D COSY-
NOESY 1H NMR, in agreement with the previously
reported data Seela, F.; Bourgeois, W. Synthesis 1989, 12,
912–918.
8. 1-Thyminyl-propionic acid was synthesized as previously
reported Mokhir, A. A.; Richert, C. Nucleic Acids Res.
2000, 28, 4254–4265.
9. Mukaiyama, T.; Usui, M.; Shimada, E.; Saigo, E. Chem.
Lett. 1975, 1045.
In summary, we have designed and synthesized new ex-
tended nucleosides holding three specific dad sites for tri-
ple helix mediated recognition of both thymine and
adenine base pair in DNA duplexes having a polypurine
strand interrupted by one or more thymidine. The incor-
poration of the efficiently protected phosphoramidites 5,
11, 13, and 15 into TFO for triple helix hybridization
studies is in progress.
10. Analytical and spectral data for selected products.
3: 1H NMR (CDCl3, 200MHz) d 1.59 (3H, br s, Me), 2.35
(3H, s, Me), 2.43 (3H, s, Me), 2.60–3.30 (4H, m, CH2CO
and 20-H), 3.97 (2H, m, CH2N), 4.50–4.80 (m, 3H, H-40
and H-50), 5.74 (m, 1H, H-30), 6.42 (m, 1H, H-10), 7.10–
7.40 (6H, m, 4·H-Tol., H-5 and H-6-thymine), 7.58 (1H,
d, J=8.5Hz, H-4), 7.86 (2H, d, J=8.2Hz, H-Tol.), 7.98
(2H, d, J=8.2Hz, H-Tol.), 8.13 (1H, s, H-7), 8.29 (1H, s,
H-2), 9.62 (1H, br s, NH), 10.95 (1H, br s, NH); 13C NMR
(CDCl3) d 11.98, 21.60, 21.70, 35.72, 37.87, 45.66, 64.13,
74.84, 82.36, 85.23, 102.12, 110.02, 116.02, 120.22, 126.37,
126.64, 129.26, 129.58, 129.81, 132.90, 134.67, 140.45,
142.26, 144.09, 144.49, 151.64, 165.28, 165.97, 166.19,
168.95; MS (ES+) m/z 688 (MNa+).
Acknowledgements
´
We gratefully acknowledge the CNRS, Region PACA
and Universite de Nice-Sophia Antipolis for financial
´
support.
References and notes
1. (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, 7749–7760; (b) Moser, H. E.;
Dervan, P. B. Science 1987, 238, 645–650; (c) Praseuth, D.;
4: 1H NMR (CD3OD, 200MHz) d 1.81 (3H, br s, Me),
2.37–2.53 (m, 1H, H-20), 2.58–2.75 (1H, m, H-20), 2.81
(2H, t, J=6.4Hz, CH2CO), 3.67–2.87 (2H, m, H-50), 3.97–
2.15 (3H, m, H-40 and CH2N), 4.52 (1H, m, H-30), 6.35
(1H, t, J=6.7Hz, H-10), 7.20 (1H, dd, J=8.7 and 1.9Hz,
H-5), 7.44 (1H, br s, H-6-thymine), 7.57 (1H, d, J=8.7Hz,
H-4), 8.13 (1H, d, J=1.9Hz, H-7), 8.39 (1H, s, H-20); 13C
NMR (CD3OD) d 12.20, 36.58, 41.17, 46.33, 63.20, 72.38,
86.49, 88.95, 104.14, 110.83, 117.44, 120.21, 134.09,
135.66, 141.06, 142.96, 143.73, 152.83, 166.92, 171.07;
MS (ES+) m/z 430 (MH+), 314 (MH+ sugar).
´ `
Guieysse, A. L.; Helene, C. Biochim. Biophys. Acta 1999,
1489, 181–206.
2. (a) Casey, B. P.; Glazer, P. M. Prog. Nucleic Acid Res.
Mol. Biol. 2001, 67, 163–192; (b) Faria, M.; Giovannan-
geli, C. J. Gene Med. 2001, 3, 299–310.
´
3. (a) Mergny, J. L.; Sun, J. S.; Rougee, M.; Garestier, T.;
Barcelo, F.; Chomilier, J.; Helene, C. Biochemistry 1991,
´ `
30, 9791–9798; (b) Greenberg, W. A.; Dervan, P. B. J. Am.
Chem. Soc. 1995, 117, 5016–5022; (c) Best, G. C.; Dervan,
P. B. J. Am. Chem. Soc. 1995, 117, 1187–1193.
4. Recent reviews: (a) Gowers, D. M.; Fox, K. R. Nucleic
Acids Res. 1999, 27, 1569–1577; (b) Purwanto, M. G. M.;
Weisz, K. Curr. Org. Chem. 2003, 7, 427–446; (c) Buchini,
S.; Leumann, C. J. Curr. Opin. Chem. Biol. 2003, 7,
717–726, and references cited therein.
5: 31P NMR (CDCl3, 200MHz) d 148.88 and 148.96; MS
(ES+) m/z 932 (MH+), 303 (DMT+).
11. The SEM-protected 5-iodouracil 7 was obtained in quan-
titative yield from 5-iodouracil and used without purifica-
1
tion. H NMR (CDCl3, 250MHz) d 0.02 (9H, s, SiMe3),
5. (a) Guianvarcꢀh, D.; Benhida, R.; Fourrey, J.-L.; Maur-
isse, R.; Sun, J. S. Chem. Commun. 2001, 1814–1815; (b)
Guianvarcꢀh, D.; Fourrey, J.-L.; Maurisse, R.; Sun, J. S.;
Benhida, R. Org. Lett. 2003, 4, 4209–4212; (c) Guian-
varcꢀh, D.; Fourrey, J.-L.; Maurisse, R.; Sun, J. S.;
Benhida, R. Bioorg. Med. Chem. 2003, 11, 2751–2759.
6. (a) Griffin, L. C.; Dervan, P. B. Science 1989, 245,
967–971; (b) Kiessling, L. L.; Griffin, L. C.; Dervan, P. B.
Biochemistry 1992, 31, 2829–2834; (c) Wang, G.; Malek,
S.; Feigon, J. Biochemistry 1992, 31, 2514–2523; (d)
Chandller, S. P.; Fox, K. R. FEBS Lett. 1993, 322,
189–192; (e) Radhakrishnan, I.; Patel, D. J. Structure
0.90 (2H, t, J=8.0Hz, CH2), 3.58 (2H, t, J=8.0Hz, CH2),
5.12 (2H, s, CH2), 7.77 (1H, s, H-6), 10.09 (1H, br s, NH).
13C NMR (CDCl3) d ꢀ1.42, 17.83, 67.34, 76.18, 94.09,
147.82, 150.94, 160.71; MS (ES+) m/z 802 (MNa+), 780
(MH+), 242 (MH+ I).
12. Under these conditions only the deprotected 3-(5-thym-
inyl)-acrylic acid was obtained.
13. Analytical and spectral data for selected products. 8: H
1
NMR (CDCl3, 200MHz) d 2.37 (3H, s, Me), 2.44 (3H, s,
Me), 2.82–2.20 (2H, m, H-20), 4.64 (3H, m, H-40 and
H-50), 5.70 (2H, m, H-30 and H-acryl.), 6.37 (2H, m, H-
acryl.), 6.54 (1H, t, J=6.1Hz, H-10), 7.10–7.40 (2H, m, H-

N. Van Craynest et al. / Tetrahedron Letters 45 (2004) 6243–6247
6247
Ar), 7.21 (2H, d, J=8.2Hz, 2·H-Tol.), 7.33 (2H, d,
J=8.2Hz, 2·H-Tol.), 7.61 (1H, d, J=9.2Hz, H-Ar), 7.81
(2H, d, J=8.2Hz, 2·H-Tol.), 7.84 (1H, m, H-Ar), 8.00
(2H, d, J=8.2Hz, 2·H-Tol.); 13C NMR (CDCl3) d 21.68,
38.29, 64.14, 74.84, 82, 85.30, 102.28, 116.09, 120.37,
126.31, 126.59, 127.48, 129.35, 129.59, 131.26, 133.08,
134.51, 140.36, 144.29, 144.58, 163.97, 166.07, 166.25; MS
7.35 (3H, m, H-4, H-5 and H-acryl.), 7.57 (1H, d,
J=8.6Hz, H-acryl.), 7.94 (1H, s, H-7), 8.27 (1H, br s,
H-6 thymine), 8.34 (1H, s, H-2);13C NMR (CD3OD) d
ꢀ3.16, 16.83, 39.39, 61.22, 66.16, 70.35, 75.96, 83.60,
84.54, 86.88, 101.92, 109.50, 115.50, 118.32, 121.36,
132.82, 134.24, 139.20, 140.85, 143.15, 145.89; MS (ES+)
m/z 566 (MNa+), 544 (MH+). 12: 1H NMR (CD3OD/
CDCl3, 250MHz) d 0.02 (9H, s, SiMe3), 0.88 (2H, t,
J=7.9Hz, CH2Si), 2.51–2.96 (6H, m, 2·CH2 and H-20),
3.66 (2H, t, J=7.9Hz, OCH2), 3.90 (2H, m, H-50), 4.16
(1H, m, H-40), 4.65 (1H, m, H-30), 5.21 (2H, s, NCH2O),
6.51 (1H, t, J=6.0Hz, H-10), 7.03 (2H, m, H-4 and H-5),
7.63 (1H, s, H-7), 7.84 (1H, br s, H-6-thymine), 8.36 (1H,
s, H-2), 8.81 (1H, br s, NH); 13C NMR (CD3OD/CDCl3) d
ꢀ3.27, 16.60, 19.45, 22.98, 33.32, 35.95, 39.67, 46.00,
60.88, 65.59, 70.07, 75.34, 85.03, 87.10, 102.02, 112.22,
116.16, 117.64, 127.49, 134.60, 136.92, 140.95, 150.81,
163.99, 171.19; MS (ES+) m/z 568 (MNa+), 546 (MH+).
14. 14: two imine-isomers: 1H NMR (CDCl3, 250MHz) d
2.30–2.80 (2H, m, H-20), 2.92 (6H, s, 2·Me), 3.62 (2H, m,
H-50), 4.01 (m, 1H, H-40), 4.51 (1H, m, H-30), 6.19 (1H, t,
J=6.4Hz, H-10), 6.85 (1H, d, J=8.5Hz, H-5), 7.22 (1H, br
s, H-7), 7.31 (1H, m, H-4), 7.51 (1H, s, H-2), 8.13 and 8.31
(1H, s, NCH); 13C NMR (CDCl3) d 40.07, 40.30, 61.70,
62.09, 70.67, 71.03, 84.73, 85.10, 87.33, 87.91, 109.91,
110.41, 118.86, 128.99, 129.24, 140.96, 142.00, 144.14,
147.79, 153.85, 155.59, 162.41; MS (ES+) m/z 305 (MH+).
15: imine and phosphoramidite isomers: 31P NMR
(CDCl3, 200MHz) d 147.27, 147.50, 147.74, and 147.82.
1
(ES+) m/z 562 (MNa+), 540 (MH+). 9: H NMR (CDCl3,
250MHz) d 0.02 (9H, s, SiMe3), 0.96 (2H, t, J=8.4Hz,
CH2Si), 2.36 (3H, s, Me), 2.46 (3H, s, Me), 2.70–2.10 (2H,
m, H-20), 3.66 (2H, t, J=8.4Hz, OCH2), 4.60–2.80 (3H, m,
H-50 and H-40), 5.21 (2H, br s, NCH2O), 5.81 (1H, m, H-
30), 6.45 (1H, t, J=6.5Hz, H-10), 7.18 (2H, d, J=8.2Hz,
2·H-Tol.), 7.30 (2H, d, J=8.2Hz, 2·H-Tol.), 7.20–7.50
(2H, m, 5-H and H-acryl.), 7.58 (1H, d, J=13.1Hz, H-
acryl.), 7.60 (1H, s, H-7), 7.70 (1H, d, J=8.5Hz, H-4), 7.89
(2H, d, J=8.0Hz, 2·H-Tol.), 8.00 (2H, d, J=8.0Hz,
2·H-Tol.), 8.32 (1H, s, 6-H-thymine), 8.56 (1H, s, H-2),
9.52 (1H, br s, NH); 13C NMR (CDCl3) d ꢀ1.39, 8.63,
17.96, 21.79, 38.06, 46.27, 64.21, 67.45, 74.87, 82.41, 85.09,
102.06, 110.73, 116.21, 120.87, 123.41, 126.42, 126.66,
129.32, 129.70, 129.93, 132.99, 133.27, 135.09, 140.22,
140.70, 144.15, 144.49, 145.51, 150.01, 162.33, 165.07,
166.14, 166.42; MS (ES+) m/z 802 (MNa+), 780 (MH+).
1
10: H NMR (CD3OD, 250MHz) d 0.02 (9H, s, SiMe3),
0.91 (2H, t, J=8.1Hz, CH2Si), 2.45 (1H, m, H-20), 2.64
(1H, m, H-20), 3.65 (2H, t, J=8.1Hz, OCH2), 3.76 (2H, m,
H-50), 4.02 (1H, q, J=3.8Hz, H-40), 4.51 (1H, m, H-30),
5.17 (2H, s, NCH2O), 6.35 (1H, t, J=6.7Hz, H-10), 7.20–