Synthesis and incorporation of an α-hexofuranosyl thymidine into oligodeoxynucleotides via its two exocyclic OH-groups

Article abstract of DOI:10.1016/j.bmcl.2003.11.075

1-(2,3-Dideoxy-3-amino-α-D-arabino-hexofuranosyl)thymine is considered as a conformationally restricted acyclic nucleoside using the furanose ring to link the diol backbone to the nucleobase. The appropriately substituted phosphoramidites were synthesised via 1-(5,6-di-O-acetyl-2,3- dideoxy-3-phthalimido-α-D-arabino-hexofuranosyl)thymine and used in oligodeoxynucleotide (ODN) synthesis. However, the binding affinity of the mixed ODNs towards complementary DNA and RNA was decreased compared to the wild-type oligos. The decrease was smaller when the monomer was inserted near the end of the sequence. The insertions into an α T sequence or in a β T sequence gave nearly the same dropping in melting temperature per modification which indicates that the new nucleotide modifications behave both as α and β nucleotides.

Full text of DOI:10.1016/j.bmcl.2003.11.075

Bioorganic & Medicinal Chemistry Letters 14 (2004) 581–584
Synthesis and incorporation of an ꢀ-hexofuranosyl thymidine into
oligodeoxynucleotides via its two exocyclic OH-groups
Vyacheslav V. Filichev and Erik B. Pedersen*
Nucleic Acid Center,y Department of Chemistry, University of Southern Denmark, DK-5230Odense M, Denmark
Received 3 October 2003; revised 14 November 2003; accepted 28 November 2003
Abstract—1-(2,3-Dideoxy-3-amino-a-d-arabino-hexofuranosyl)thymine is considered as a conformationally restricted acyclic
nucleoside using the furanose ring to link the diol backbone to the nucleobase. The appropriately substituted phosphoramidites
were synthesised via 1-(5,6-di-O-acetyl-2,3-dideoxy-3-phthalimido-a-d-arabino-hexofuranosyl)thymine and used in oligodeoxy-
nucleotide (ODN) synthesis. However, the binding affinity of the mixed ODNs towards complementary DNA and RNA was
decreased compared to the wild-type oligos. The decrease was smaller when the monomer was inserted near the end of the sequence.
The insertions into an a T sequence or in a b T sequence gave nearly the same dropping in melting temperature per modification
which indicates that the new nucleotide modifications behave both as a and b nucleotides.
# 2003 Elsevier Ltd. All rights reserved.
In recent years, much effort has been put into the design
and synthesis of new oligodeoxynucleotides (ODNs) in
order to improve the hybridisation affinity towards
complementary DNA or RNA sequences.1 Since the
discovery of PNA (peptide nucleic acid, Fig. 1), con-
structed from 2-aminoethyl glycine units to which
nucleobases are attached by methylenecarbonyl linkers,2
a considerable number of ODNs having acyclic DNA
backbone have been synthesised.3 Usually ODNs pos-
sessing an acyclic moiety bind poorly to DNA or RNA
due to loss in entropy upon duplex formation. However,
rigid carboxamide groups in PNA improve the hybridi-
sation properties. Hexofuranosyl nucleotides, by inser-
tions into ODNs via the two exocyclic OH-groups, are
here considered as conformationally restricted acyclo-
nucleotides constrained by the furanose ring between
the nucleobase and the ethyleneglycol backbone. Pre-
viously, this type of nucleosides have been synthesised
and investigated as potential anticancer and anti-viral
agents.4 In oligonucleotide chemistry hexofuranosyl
analogues were regarded as branched ODNs possessing
free 50-C-hydroxymethyl group that could be used for
conjugation reactions.5
In this investigation we decided to incorporate 1-(3-
amino-2,3-dideoxy-a-d-arabino-hexofuranosyl)thymine
into oligonucleotide sequences in order to evaluate the
thermal stability of ODN/DNA and ODN/RNA
duplexes. The positively charged amino group is expec-
ted to neutralise the electrostatic repulsion between
phosphodiester backbones and thus lead to better bind-
ing properties. Due to free rotation around C-40–C-50
bond, the hexofuranosyl nucleosides can exist in two
conformations 1 and 10 corresponding to a and b
anomers, respectively, of natural nucleosides (Fig. 2).
Therefore, oligos from this type of nucleosides may
have the advantage like PNA in forming duplexes
Keywords: Hexofuranosyl nucleoside; DNA; Backbone modification;
Duplex stability; Exocyclic DNA backbone.
* Corresponding author. Tel.: +45-6550-2555; fax: +45-6615-8780;
e-mail: ebp@chem.sdu.dk
y
A research center funded by The Danish National Research Foun-
dation for studies on nucleic acid chemical biology.
Figure 1.
0960-894X/$ - see front matter # 2003 Elsevier Ltd. All rights reserved.
doi:10.1016/j.bmcl.2003.11.075

582
V. V. Filichev, E. B. Pedersen / Bioorg. Med. Chem. Lett. 14 (2004) 581–584
procedure was used because the standard DMTCl/pyri-
dine procedure failed. Desilylation using tetra-n-butyl-
ammonium fluoride (TBAF) in THF gave compound 8
in 23% yield after 2 days of stirring at room temper-
ature and at 50 ꢀC overnight. The low yield observed
here is probably due to competitive partial deprotection
of N-phtaloyl group by the nucleophilic fluoride ion.
Phosphitylation of 8 afforded a diastereomeric mixture
(1:1) of the phosphoramidite 9 in 70% yield after silica
gel column purification.
Oligonucleotides were synthesised by the solid phase
phosphoramidite method on an automated Pharmacia
Gene AssemblerTM Special synthesiser in 0.2 mmol-scale
(7.5 mmol embedded per cycle, Pharmacia primer sup-
portTM) using an increased coupling time (10 min)
for compound 6 and 9 compared to 2 min couplings for
commercial phosphoramidites. The coupling efficiency
for both monomers using 3,5-dicyanoimidazole as acti-
vator was only 70–80% and was not improved by pro-
longed or repeated couplings. The 50-DMT-on derivatives
were cleaved off from the solid support (rt, 3 h) and
deprotected (60 ꢀC, overnight) using 32% aqueous
ammonia for standard support or the solution of 2%
LiCl in 32% aqueous ammonia (w/w) when using uni-
versal support for a nucleotides. The purification of
ODNs was performed on a preparative HPLC (Waters
Delta Prep 4000) using an Xterra2 column,7 followed by
removal of DMT (80% AcOH) with further precipita-
tion from ethanol. However for ODN12–ODN15 (Table
1) a second HPLC purification was needed due to fail-
ure of the precipitation procedure. In this case the
volatile salts were removed by several co-evaporations
with water.7 The removal of N-phthaloyl group during
standard oligonucleotide deprotection was confirmed by
MALDI-TOF analysis for the final ODNs (Table 1).
Figure 2.
simulating both b and a anomers. To put it another
way, this type of oligos may hybridise like PNA, both in
a parallel and an antiparallel fashion. This property
seems to be a prerequisite for strand invasion.2b
Another feature, the primary OH-group of the hexo-
furanosyl nucleoside can be considered as the 50-OH
group of natural occurring nucleosides as in 1 or the
nucleoside can be reverted so the same OH-group is
considered as the 30-OH as in 2 (Fig. 2 and Scheme 1).
The synthesis of target building blocks 6 and 9 started
from previously published 1-(5,6-di-O-acetyl-2,3-dideoxy-
3-phtalimido-a-d-arabino-hexofuranosyl)thymine (3)4a
that was selectively deprotected in the refluxing mixture
of Et3N/MeOH to give 4 in 95% yield.6 The primary
alcohol 4 was treated with an excess of DMTCl in pyridine
with further purification on a silica gel column affording
compound 5 in 64% yield. The final phosphoramidite 6
was obtained as a mixture of two diastereoisomers (3:4)
by treatment with 2-cyanoethyl N,N,N0,N0-tetraisopro-
pylphosphorodiamidite in the presence of N,N-diiso-
propylammonium tetrazolide in 75% yield. To
incorporate compound 4 into ODNin the opposite
direction, nucleoside 4 was reacted with tert-butyldi-
methylsilyl chloride (TBDMSiCl) in pyridine followed
by treatment with DMTCl in the presence of AgNO3
and pyridine in anhydrous THF for 2 days. The latter
The hybridisation properties of the modified ODNs
towards complementary DNA and RNA single strands
were evaluated by thermal denaturation studies. The
melting temperature (Tm) as a first derivative of
the melting curve and the differences between modified
and unmodified oligomers as the change in melting
Scheme 1. Reagents and conditions: (a) Et3N, MeOH, reflux, 4 h; (b) DMTCl, Py, rt, overnight; (c) NC(CH2)2OP(NPri2)2, N,N-diisopropy-
lammonium tetrazolide, CH2Cl2; (d) DNA synthesizer; (e) (i) TBDMSiCl, Py, 0 ꢀC, overnight; (ii) DMTCl, AgNO3, Py, THF, rt, 2 days; (f) Bu4NF,
THF, rt, 2 days, then 50 ꢀC, overnight.

V. V. Filichev, E. B. Pedersen / Bioorg. Med. Chem. Lett. 14 (2004) 581–584
583
Table 1. Oligodeoxynucleotides synthesised, melting experiments, calculated and experimental masses of ODNs from MALDI-TOF MSa
No.
Sequence
Tm (ꢀC)
(ODN/DNA)
ꢀTm
(C)
Tm (ꢀC)
(ODN/RNA)
ꢀTm
(ꢀC)
m/z, calc.
(Da)
m/z, found
(Da)
ODN1
ODN2
ODN3
ODN4
ODN5
ODN6
ODN7
ODN8
ODN9
50-GCTTCTCTAGAATG
50-GCTTC1CTAGAATG
50-GCTTCTC1AGAATG
50-GCTTC1C1AGAATG
50-GCT1CTCTAGAATG
50-GC11CTCTAGAATG
50-GCTTCTCTAGAA1G
50-GCTTC2CTAGAATG
50-GCTTCTC2AGAATG
42.6b
33.7
35.3
27.6
34.4
25.2
39.1
29.1
31.5
43.8c
36.9
37.7
25.3
37.6
29.5
42.0
28.6
29.6
À8.9
À7.3
À7.5
À8.2
À8.7
À3.5
À13.5
À11.1
À6.9
À6.1
À9.3
À6.2
À7.2
À1.8
À15.2
À14.2
4281.03
4281.03
4310.06
4281.03
4310.06
4281.03
4281.03
4281.03
4280.59
4281.28
4309.10
4279.91
4311.91
4280.07
4280.38
4282.38
ODN10
ODN11
ODN12
ODN13
ODN14
ODN15
dT14
adT14
31.8d
32.0
<10
<10
13.7
<10
28.3e
42.5
10.6
10.3
22.5
<10
50-TTTT1T1TTTTTTT
50-TTTT2T2TTTTTTT
a 50-(TTTT1T1TTTTTTT)
a 50-(TTTT2T2TTTTTTT)
À8.9
À9.0
4253.06
4253.06
4253.06
4253.06
4253.55
4252.11
4253.31
4253.65
À9.2
À10.0
Tm (ꢀC) (ODN/
ꢀTm (C)
30ÀCGAAGAGCTCTTAC)
ODN16
ODN17
ODN18
50-GCTTCTCTAGAATG
50-GCTTCTC1AGAATG
50-GCTTCTC2AGAATG
32.2
30.3
28.2
À10.4
À5.0
À3.3
ODN1 as references
ODN3 as references
ODN9 as references
a Tm was determined by measuring absorbance at 260 nm against increasing temperature (1.0 ꢀC) on equimolar mixtures (1.5 mM in each strand) of
modified oligomer and its complementary DNA/RNA and single mismatched DNA in medium salt buffer (10 mM NaH2PO4, 100 mM NaCl, 0.1
mM EDTA, pH 7.00); 1 and 2 are explained in the text; ꢀTm=change in Tm per modification.
b Target: 30-CGAAGAGATCTTAC.
c Target: 30-CGAAGAGAUCUUAC.
d Target: dA14
.
e Target: rA14
.
temperature per modification (ꢀTm) are listed in
Table 1.
into a b T sequence. Per modification, the drop in
melting temperature was nearly the same whatever
incorporated in an a T sequence or in a b T sequence.
This may indicate that the new nucleoside modification
behave both as a and b nucleoside.
The thermal melting data showed that insertion of 1 in
the central region of a mixed oligomer sequence
decreased the Tm significantly (ꢀTm À7.0 to À9.0 ꢀC),
and that the decrease was smaller (ꢀTm À2.0 to
À3.5 ꢀC) when 1 was situated near the end. This is in
agreement with previous studies for acyclonucleotide
systems.8À11 The insertion of 2 led to a more dramatic
decrease in Tm (ꢀTm À11.0 to À15.0 ꢀC) when com-
pared with the nucleotide 1. The fully modified oligomer
gave no melting neither towards DNA nor RNA. The
difference in ꢀTm between matched and a C/T mis-
matched duplexes was larger for insertion of 1 than for
2 (ODN16–ODN18). The destabilisation effect observed
in the case of 2 is in the same order of magnitude as
previously observed for the ethyleneglycol linker.12 The
consequence is that in case of incorporation 2 we have
very little base pairing and it could as well be considered
as a mismatch. However, insertions of the a-nucleoside
analogue 1 gave ODNs, where base pairing is more
likely as evaluated from mismatched meltings. Only
small differences in binding abilities to DNA and RNA
complements were observed. To evaluate further the
possible rotation around C-40–C-50 bond we synthesised
ODNs possessing incorporation as next-nearest neigh-
bours of 1 and 2 either in d(b T)14 (ODN12, ODN13,
Table 1) or in d(a T)14 (ODN14, ODN15). In both cases
decreases in Tm were observed. In the case of hybridi-
sation with RNA, melting data could be determined for
the nucleoside 1 which could allow comparison of
incorporation into an a T sequence with incorporation
The proposed system 1 having relatively short ethyl-
eneglycol backbone and thymine situated in the a-posi-
tion of the furanose ring was not diminished the
hybridisation properties larger than other acyclic
nucleotide analogues, apart from PNA.8 Thus an aver-
age ꢀTm of À6.0 to À10.5 ꢀC has been found for incor-
poration into central regions of oligomers summarised
by Nielsen et al.8 and acyclo DNA (T) (Fig. 1) being the
worst one with ꢀTm of À13.4 ꢀC.9 The conformational
restriction by C¼C bond decreased the binding towards
DNA and RNA from ꢀTm of À2.0 ꢀC in the end of the
duplex to ꢀTm of À5.5 ꢀC in the middle of the duplex.10
ODNs with (N-thymin-1-ylacetyl)-1-arylserinol back-
bone has been found quite efficient in their binding
capacity towards DNA duplex and triplex structures,
but the dropping ꢀTm of À9.0 ꢀC has also been
observed when the monomer has been inserted in the
middle of the sequence.11 The optimum flexibility and
rigidity of the sugar ring and phosphodiester backbone
in conjunction with the proper position of the base
moieties is a key factor in the modeling of new DNA/
RNA analogues. When this is combined by the above
findings of incorporation of 1 into a and b T sequences
we still think it is possible to develop new nucleotides
which can give interesting hybridisation properties for
incorporation in both a and b nucleotide sequences.
The perspective for the fully modified sequence is the

584
V. V. Filichev, E. B. Pedersen / Bioorg. Med. Chem. Lett. 14 (2004) 581–584
hybridisation both in a parallel and in an antiparallel
fashion, but the design of a proper phosphodiester
backbone linker is still needed to be found.
H-6, Phth), 8.60 (1H, br.s, NH). dP 150.6 (s) and 151.1 (s)
in the ratio 3:4. UV-MALDI MS: 926.348 ([M+Na]+ calcd
926.350). 1-[2,3-Dideoxy-6-O-(tert-butyltrimethylsilyl)-5-
O-(4,40-dimethoxytrityl)-3-phthalimido-a-d-arabino-hexo-
furanosyl]thymine (7): Rf 0.45 (5% MeOH/CHCl3). NMR
(CDCl3): dH 0.06 (6H, s, 2ꢁCH3 [TBDMSi]), 0.98 (9H, s,
t-C4H9), 2.09 (3H, s, CH3), 2.79 (2H, m, H-20), 3.50 (1H,
dd, J=2.6, 11.0 Hz, H-60), 3.75 (1H, dd, J=3.6, 11.0 Hz,
H-60), 3.90 (7H, m, OCH3, H-50), 4.98 (1H, t, J=5.7 Hz,
H-40), 5.40 (1H, m, H-30), 6.62 (1H, t, J=6.7 Hz, H-10),
6.70–6.80 and 7.00–7.50 (14H, m, DMT), 7.75–7.95 (5H,
m, H-6, Phth), 8.54 (1H, s, NH). dC À5.4, À5.6, 12.5, 18.2,
25.9, 39.1, 51.2, 55.1, 55.2, 61.2, 73.2, 82.2, 86.0, 87.9,
110.5, 112.6, 113.1, 123.3, 123.7, 126.6, 127.3, 127.8, 128.6,
129.1, 130.6, 131.9, 134.0, 135.9, 136.0, 136.8, 139.5,
144.3, 149.9, 150.0, 158.3, 163.6, 167.7. UV-MALDI
MS: 840.331 ([M+Na]+ calcd 840.329). 1-{2,3-Dideoxy-
6-O-[2-cyanoethoxy(diisopropylamino)phosphino]-5-O-
[4,40-dimethoxytrityl]-3-phthalimido-a-d-arabino-hexofur-
anosyl}thymine (9): Rf 0.35 (5% MeOH/CHCl3). NMR
(CDCl3): dH 0.95 (6H, m, CH3 [Pri]), 1.05 (6H, m, CH3
[Pri]), 1.90 (3H, s, CH3), 2.50–2.75 (6H, m, CH2CH2CN,
H-20), 3.25–3.60 (5H, m, H-60, H-50, CH [Pri]), 3.75 (6H, s,
OCH3), 4.70 (1H, m, H-30), 5.20 (1H, m, H-40), 6.40 (1H,
m, H-10), 6.50–6.70 and 7.00–7.50 (14H, m, DMT), 7.70–
7.90 (5H, m, H-6, Phth), 8.05 (1H, br.s, NH). dP 146.4 (s)
and 148.5 (s) in the ratio 1:1. UV-MALDI MS: 926.345
([M+Na]+ calcd 926.350).
References and notes
1. Leumann, C. J. Bioorg. Med. Chem. 2002, 10, 841.
2. (a) Nielsen, P. E.; Egholm, M.; Berg, R. H.; Buchardt, O.
Science 1991, 254, 1497. (b) Larsen, H. J.; Bentin, T.;
Nielsen, P. E. Biochem. Biophys. Acta 1999, 1489, 159.
3. Kumar, V. A. Eur. J. Org. Chem. 2002, 2021.
4. (a) Lau, J.; Pedersen, E. B.; Jensen, L. V. Arch. Pharm.
(Weinheim) 1991, 324, 83. (b) Lau, J.; Wengel, J.; Ped-
ersen, E. B.; Vestergaard, B. F. Synthesis 1991, 1183.
5. Fensholdt, J.; Wengel, J. Acta Chem. Scand. 1996, 50,
1157.
6. Selected data for new compounds. 1-(2,3-Dideoxy-6-O-
(4,40 -dimethoxytrityl)-3-phthalimido-a-d-arabino-hexo-
furanosyl)thymine (5): Rf 0.40 (5% MeOH/CHCl3).
NMR (CDCl3): dH 1.80 (3H, s, CH3), 2.60 (1H, br.s.,
OH), 2.75 (1H, m, H-20a), 2.95 (1H, m, H-20b), 3.32 (2H,
m, H-60), 3.50 (1H, m, H-50), 3.80 (6H, s, OCH3), 4.72
(1H, dd, J=5.8 and 3.6, H-30), 5.28 (1H, t, J=6.6, H-40),
6.34 (1H, t, J=6.2 Hz, H-10), 6.70–6.80 and 7.00–7.50
(14H, m, DMT), 7.70–7.90 (5H, m, H-6, Phth), 8.80 (1H,
s, NH). dC 12.3, 37.2, 51.3, 55.1, 55.2, 64.3, 69.5, 82.6,
86.2, 90.3, 110.3, 112.9, 113.1, 123.5, 123.7, 126.8, 127.7,
127.9, 128.1, 129.1, 129.9, 131.5, 134.2, 135.6, 135.9,
136.8, 139.5, 144.6, 149.7, 150.0, 158.5, 163.9, 168.3. UV-
MALDI MS: 726.245 ([M+Na]+ calcd 726.242). 1-{2,3-
Dideoxy - 6 - O - [4,40 - dimethoxytrityl] - 5 - O - [2 - cyano-
ethoxy(diisopropylamino)phosphino]-3-phthalimido-a-d-
arabino-hexofuranosyl}thymine (6): Rf 0.62 (5% MeOH/
CHCl3). NMR (CDCl3): dH 0.90 (6H, m, CH3 [Pri]), 1.05
(6H, m, CH3 [Pri]), 1.80 (3H, s, CH3), 2.19 (2H, m,
CH2CN), 2.60–2.90 (4H, m, CH2CH2CN, H-20), 3.20–
3.60 (5H, m, H-60, H-50, CH [Pri]), 3.78 (6H, s, OCH3),
5.00 (1H, m, H-30), 5.23 (1H, m, H-40), 6.40 (1H, m, H-10),
6.70–6.80 and 7.00–7.50 (14H, m, DMT), 7.70–7.90 (5H, m,
7. Filichev, V. V.; Pedersen, E. B. Org. Biomol. Chem. 2003,
1, 100.
8. Nielsen, P.; Dreiøe, L. H.; Wengel, J. Bioorg. Med. Chem.
1995, 3, 19.
9. Schneider, K. C.; Benner, S. A. J. Am. Chem. Soc. 1990,
112, 453.
10. Boensen, T.; Pedersen, D. S.; Nielsen, B. M.; Petersen,
A. B.; Henriksen, U.; Dahl, B. M.; Dahl, O. Bioorg. Med.
Chem. Lett. 2003, 13, 847.
11. Rana, V. S.; Kumar, V. A.; Ganesh, K. N. Tetrahedron
2001, 57, 1311.
12. Christensen, U. B.; Pedersen, E. B. Nucl. Acids Res. 2002,
30, 4918.