Oligonucleotides containing a new type of acyclic, achiral nucleoside analogue: 1-[3-hydroxy-2-(hydroxymethyl)prop-1-enyl]thymine

Article abstract of DOI:10.1016/S0960-894X(03)00008-8

An achiral, acyclic nucleoside analogue has been incorporated once or twice in oligodeoxyribonucleotides by the phosphoramidite method, and conditions found which allow deprotection of the oligonucleotides containing a sensitive modified allylic unit. The

Full text of DOI:10.1016/S0960-894X(03)00008-8

Bioorganic & Medicinal Chemistry Letters 13 (2003) 847–850
Oligonucleotides Containing a New Type
of Acyclic, Achiral Nucleoside Analogue:
1-[3-Hydroxy-2-(hydroxymethyl)prop-1-enyl]thymine
Thomas Boesen,y Daniel Sejer Pedersen,{ Brian M. Nielsen, Asger B. Petersen,
Ulla Henriksen, Britta M. Dahl and Otto Dahl*
Department of Chemistry, University of Copenhagen, The H. C. Ørsted Institute, Universitetsparken 5,
DK-2100 Copenhagen, Denmark
Received 1 October 2002; revised 14 November 2002; accepted 20 December 2002
Abstract—An achiral, acyclic nucleoside analogue has been incorporated once or twice in oligodeoxyribonucleotides by the phos-
phoramidite method, and conditions found which allow deprotection of the oligonucleotides containing a sensitive modified allylic
unit. The binding affinity of the modified oligonucleotides towards complementary DNA and RNA was reduced compared to
unmodified DNA (ÁTm À2 to À6.5 ꢀC). An oligonucleotide with two modifications at the 30-end showed considerable resistance
towards cleavage with a 30-exonuclease.
# 2003 Elsevier Science Ltd. All rights reserved.
Oligonucleotides which are modified to be resistant to
degradation by nucleases are promising therapeutic
agents (antisense or antigene drugs).1,2 The attraction of
antisense drugs is their high selectivity for a com-
plementary mRNA target based on the fidelity of the
phosphoramidates);3 of the (deoxy)ribose unit (e.g., 20-
substitution with O-alkyl or F,2 40-substitution of O
with S or CH2,4 and introduction of bicyclic derivatives
of ribose, as in LNA5); and of the bases.6 More radical
modifications include oligonucleotides where the phos-
phate groups have been replaced (with e.g., amide,
hydroxylamine, or formacetal groups),3 the ribose has
been opened to acyclic analogues,7 or replaced with
pyranose, cyclohexene or anhydrohexitol groups.2 Even
the whole sugar-phosphate backbone has been replaced,
as in PNA and morpholino-DNA.3 Oligonucleotides
containing acyclic nucleoside analogues usually bind
poorly to DNA and RNA due to losses in entropy upon
duplex formation, caused by the flexibility of the acyclic
units.7 However, some acyclic analogues, notably PNA,
are restricted conformationally by rigid amide linkages
and hybridise well to DNA and RNA. The central
amide bond in PNA has been replaced with a C¼C
bond, however with a detrimental effect on the binding
properties.8
Watson–Crick
hydrogen
binding.
Modified
oligonucleotides have to fulfil several requirements in
order to be useful antisense drugs: they should bind
(hybridise) efficiently to RNA, be resistant to degrada-
tion, be soluble in water, be able to invade cells, and be
able to stop the in vivo processing or translation of
RNA. The last requirement may be fulfilled in several
ways, for example, by physical blocking of splicing of
RNA, physical blocking of initiation of translation of
RNA, or by activating RNase H cleavage of RNA
hybridised to the oligonucleotide.3
Many oligonucleotide modifications have been studied.
These include modifications of the phosphate groups
(e.g., phosphorothioates, methylphosphonates, and
The nucleoside analogue studied in this paper, 1-[3-
19
(Fig. 1) is an acyclic analogue conformationally restric-
ted by a C¼C bond. It was designed to mimic the nat-
ural nucleosides, and is radically different from PNA (or
its C¼C-analogue). According to models, 1 can exist in
*Corresponding author. Tel.: +45-353-20176; fax: +45-353-20212;
e-mail: dahlo@kiku.dk
hydroxy-2-(hydroxymethyl)prop-1-enyl]thymine
yCurrent address: Lica Pharmaceuticals A/S, Fruebjergvej 3, DK-
2100, Copenhagen, Denmark.
{Current address: University Chemical Laboratory, Lensfield Road,
Cambridge, CB2 1EW, UK.
0960-894X/03/$ - see front matter # 2003 Elsevier Science Ltd. All rights reserved.
doi:10.1016/S0960-894X(03)00008-8

848
T. Boesen et al. / Bioorg. Med. Chem. Lett. 13 (2003) 847–850
ammonia at the phosphortriester stage. Thus, a tetra-
decamer, dT7TbT6, gave, after treatment with concd aq
ammonia for 2 h at rt, mainly dT7p and pdT6 and no
full length product according to polyacrylamide capil-
lary electrophoresis (CE) and electrospray ionization
mass spectroscopy (ESI-MS). The Tb unit was unstable
to the milder deprotection/cleavage reagent K2CO3 in
MeOH13 as well. Allylic groups are known to be
removed from allylic phosphortriesters by concd aq
ammonia,14 but their removal in preference to removal
of the cyanoethyl group in our system was unexpected.
Since the allylic group is removed by an SN2 or SN20
process and the cyanoethyl group by a b-elimination
(Fig. 2), a hindered weakly nucleophilic base might
remove the cyanoethyl group preferentially. Indeed,
treatment with neat, dry Pri2NH for 14h at rt,15 followed
by concd aq ammonia for 2 h at rt, gave full length
dT7TbT6 tetradecamer, although the product was con-
taminated by shorter sequences according to CE. The
tetradecamer was purified by ion exchange HPLC, fol-
lowed by desalting on a NAP column, to give dT7TbT6
in ca. 20% yield, 92% pure according to CE.16 Two
other homothymine tetradecamers, containing one Ta in
the middle or two Tb unit at the 30-end, and one non-
amer with mixed sequence, dGCATbCTbCAC, were
prepared in a similar way. The purified sequences were
characterised by CE and ESI-MS.17
Figure 1.
conformations which positions the two hydroxy groups
and the thymine group close to the positions of the 30-
hydroxy group, the 50-hydroxy group, and thymine in
both natural dsDNA (B-type helixes) and dsRNA (A-type
helixes). Therefore 1 should be able to replace natural thy-
midines in both A- and B-type helixes without seriously
disturbing the helical structures. The present paper
describes our results on the preparation, by phosphor-
amidite synthesis, of oligodeoxyribonucleotides containing
1, their ability to hybridise to RNA and DNA, and their
resistance to cleavage by a 30-exonuclease.
Phosphoramidites were prepared from 19 in a standard
way. The mono-DMT derivatives 2a and 2b were
obtained each in 20–25% yield, together with the bis-
DMT derivative 3, from 1 and one equivalent of
DMTCl in pyridine (Fig. 1). The three compounds were
easily separated on a silica column, eluted with EtOAc–
MeOH–Et3N 98:1:1. The slowest moving compound
was identified as the Z-isomer 2b by NOE experiments.
Phosphitylation with 2-cyanoethyl N,N-diisopropyl-
phosphoramidochloridite gave the phosphoramidites 4a
or 4b in 60–70% yield after purification by column
chromatography and precipitation in hexane.10
Oligonucleotides containing the Ta and Tb units were
quite stable after the cyanoethyl groups were removed.
Thus dT7TbT6 was not visibly cleaved (CE) for at least 2
months in aq solutions at pH 7. In 32% aq ammonia,
7% was cleaved after 6 h at 55 ꢀC, but less than 1% was
cleaved after 24 h at rt. An oligonucleotide with two Tb
units at the 30-end was significantly more stable towards
30-exonucleolytic cleavage than the unmodified analogue.
Thus, dT11TbTb was 65% intact (CE) after cleavage of
dT11TbTbT with snake venom phosphordiesterase
(Phosphodiesterase I, Crotalus adamanteus Venom,
Pharmacia) for 2 h, under conditions where t1/2 for dT14
was 8 min.18
Modified oligonucleotides were synthesised on a T-suc-
cinate LCAA CPG support using standard reagents
(tetrazole activation, oxidation with ButOOH, capping
with NMI/Ac2O/Lu, detritylation with CCl3COOH).11
Unwanted phosphitylation at the bases occurred to
some degree with the very reactive 4a and 4b, but this
was remedied by an aqueous wash before oxidation.
Attempted oxidation with iodine/water cleaved the
allylic C–O bond. This is known to occur for other
phosphites with allylic or tertiary substituents.12 Sur-
prisingly, coupling of the next monomer upon the unit
Tb derived from 4b proceeded with a DMT efficiency of
only 80–85%. This low efficiency was independent of
the identity of the next phosphoramidite, and was not
improved by prolonged or repeated couplings. When
the unit was Ta derived from 4a, the next monomer
coupled with a better DMT efficiency (ca. 95%). The
reasons for these inefficient couplings are unknown at
present.
Deprotection and cleavage of the modified oligonucleo-
tides from the support had to be performed in two steps,
because the Ta and Tb units were unstable to aqueous
Figure 2.

T. Boesen et al. / Bioorg. Med. Chem. Lett. 13 (2003) 847–850
849
The hybridisation properties of the modified oligo-
Acknowledgements
nucleotides towards complementary DNA and RNA
single strands were evaluated by thermal denaturation
studies (Table 1). These data show that one Tb unit in
the central region of an oligomer decreased the Tm sig-
nificantly (ÁTm À5.0 to À6.5 ꢀC), and that the decrease,
as expected, was smaller (ÁTm À2.0 to À2.5 ꢀC) when
Tb was situated near the end. No large differences in
binding to DNA and RNA complements were observed.
A mismatched RNA complement resulted in a depression
of Tm which was as large (ÁTm À15.5 ꢀC) as that found
for unmodified dT14. A Ta unit resulted in a larger
decrease of Tm than a Tb unit, as expected since the Ta
unit is analogous to an a-DNA unit. The hybridisation
properties of the Tb unit compare favourably with those
of other acyclic nucleoside analogues, apart from PNA.
Thus the flexible analogue 5 (Fig. 3) built into central
regions of oligomers resulted in an average ÁTm of
À13.4 ꢀC, and for several other acyclic analogues ÁTm
values of À6.5 to À10.5 ꢀC have been found, 6 being the
best one.7 Although the saturated analogue of Tb, 7, has
not to our knowledge been evaluated for its hybridi-
sation properties, the presence of a double bond in the
Tb unit seems to have a positive effect on the binding.
We are indebted to Cureon A/S for the ESI-MS determi-
nations. Ms. Jette Poulsen and Ms. Anette W. Jørgensen are
thanked for expert technical assistance. The Danish Natural
Science Research Council and The Danish Technical
Research Council are acknowledged for financial support.
References and Notes
1. (a) Chadwick, D. J.; Cardew, G., Eds.; Oligonucleotides as
Therapeutic Agents. Ciba Foundation, John Wiley: 1997. (b)
Crooke, S. T., Ed.: Antisense Research and Application,
Springer: 1998.
2. Uhlmann, E. Curr. Opin. Drug Discov. Develop. 2000, 3,
203.
3. Micklefield, J. Current Med. Chem. 2001, 8, 1157.
4. Altmann, K.-H.; Cuenoud, B.; von Matt, P. Novel Chem-
istry. In Applied Antisense Oligonucleotide Technology; Stein,
C. A., Krieg, A. M., Eds.; Wiley-Liss: New York, 1998;
Chapter 4.
5. Meldgaard, M.; Wengel, J. J. Chem. Soc., Perkin Trans. 1
2000, 3539.
6. Herdewijn, P. Antisense Nucleic Acid Drug Dev. 2000, 10,
297.
This study shows that a very simple acyclic, achiral
monomer like 1 is able to replace a DNA monomer in
oligonucleotides without seriously compromising the
hybridisation properties. Work is in progress to optimise
the coupling and deprotection procedures in order to
prepare and study fully modified oligomers, and to
examine possible reasons (less than optimal preorganised
conformation, reduced solvation, or other factors) for
the somewhat unsatisfactory binding properties.
7. Nielsen, P.; Dreiøe, L. H.; Wengel, J. Bioorg. Med. Chem.
1995, 3, 19.
8. Schutz, R.; Cantin, M.; Roberts, C.; Greiner, B.; Uhlmann,
E.; Leumann, C. Angew. Chem., Int. Ed. 2000, 39, 1250.
9. Pedersen, D. S.; Boesen, T.; Eldrup, A. B.; Kiær, B.; Mad-
sen, C.; Henriksen, U.; Dahl, O. J. Chem. Soc., Perkin Trans.
1 2001, 1656.
10. Selected data for new compounds. (E)-1-[2-(Dimethoxy-
trityloxymethyl)-3-hydroxyprop-1-enyl]thymine (2a): Rf 0.40
(EtOAc–MeOH–Et3N 98:1:1). NMR (DMSO-d6): dH 11.39 (1H,
s, NH), 7.43–6.89 (14H, m, Ar+T-H6), 6.65 (1H, s, NCH¼C),
4.90 (1H, t, J=5.3 Hz, OH), 3.92 (2H, d, J=5.3 Hz, CH2OH),
3.74 (6H, s, OCH3), 3.70 (2H, s, CH2ODMT), 1.78 (3H, d, J=1.2
Hz, T-CH3). FAB+ MS: 515.2 (M+H+ calcd 515.2). (Z)-1-[2-
(Dimethoxytrityloxymethyl)-3-hydroxyprop-1-enyl]thymine (2b):
Rf 0.25 (EtOAc–MeOH–Et3N 98:1:1). NMR (DMSO-d6): dH
11.30 (1H, s, NH), 7.31–6.86 (14H, m, Ar+T-H6), 6.49 (1H, s,
NCH¼C), 5.15 (1H, t, J=5.0, OH), 4.16 (2H, d, J=5.0,
CH2OH), 3.73 (6H, s, OCH3), 3.47 (2H, s, CH2ODMT), 1.60 (3H,
d, J=1.2 Hz, T-CH3). dC 164.1, 158.9, 150.4, 144.5, 140.5, 136.1,
135.5, 130.1, 128.2, 127.3, 124.8, 113.5, 110.6, 87.3, 63.6, 58.9,
55.5, 12.6. FAB+ MS: 515.2 (M+H+ calcd 515.2). The Z con-
Table 1. Hybridisation data (Tm, ꢀC) for modified and unmodified
oligodeoxyribonucleotides with DNA and RNA complementsa
b
b
dA14
ÁTm
rA14
ÁTm
rA6CA7
ÁTm
dT14
36.0
31.0
26.0
31.0
33.5
28.0
27.0
29.0
18.0
12.5
À15.5
À15.5
dT7TbT6
dT7TaT6
dT11TbTbT
À5.0
À10.0
À2.5
À5.5
À6.5
À2.0
b
b
dGTGAGATGC ÁTm rGTGAGATGC ÁTm
1
figuration of 2b was determined by H NMR NOE effects from
dGCATCTCAC
dGCATbCTbCAC
39.0
28.0
41.0
27.5
NCH¼C to CH2OH. (Z)-1-[3-(Dimethoxytrityloxy)-2-[2-
cyanoethoxy(diisopropylamino)phosphinoxy - methyl]prop - 1-
enyl]thymine (4a): NMR (CDCl3): dH 7.39–7.14 (10H, m,
Ar+NH), 7.12 (1H, q, J 1.2, T-H6), 6.76 (4H, d, J=8.8 Hz,
Ar), 6.70 (1H, s, NCH¼C), 4.11–3.96 (2H, m, CH2OP), 3.80–
3.36 (6H, m, CH2ODMT, CH2CH2CN, CH(CH3)2), 3.72 (6H,
s, CH3O), 2.41 (2H, t, J 6.3, CH2CH2CN), 1.87 (3H, d, J=1.2
HZ, T-CH3), 1.08–0.95 (12H, m, CH(CH3)2). dP 150.7. (E)-1-[3-
(Dimethoxytrityloxy)-2-[2-cyanoethoxy(diisopropylamino)phos-
phinoxy-methyl]prop-1-enyl]thymine (4b): NMR (CDCl3): dH
7.40–7.18 (10H, m, Ar+NH), 7.14 (1H, s, T-H6), 6.81 (4H, d,
J=8.8 Hz, Ar), 6.70 (1H, s, NCH¼C), 4.38 (2H, AB of ABX
system, Á=27.5 Hz, JAB=13.4 Hz, JAX=8.2 Hz, CH2OP),
3.90–3.53 (6H, m, CH2ODMT, CH2CH2CN, CH(CH3)2), 3.78
(6H, s, CH3O), 2.61 (2H, t, J=6.3 Hz, CH2CH2CN), 1.71 (3H,
s, T-CH3), 1.19 (12H, pseudo-triplet, J=7.0 Hz, CH(CH3)2).
dC 164.1, 158.9, 150.2, 144.6, 140.7, 135.6, 133.3, 133.2, 130.2,
128.2, 128.1, 127.3, 125.3, 113.4, 110.4, 87.1, 64.0, 63.7, 58.9,
À5.5
À6.5
aTm was determined by measuring absorbance at 260 nm against
increasing temperature (0.5 ꢀC steps) on equimolar mixtures (3 mM in
each strand) of modified oligomer and its complementary DNA or
RNA strand in medium salt buffer (10 mM Na2HPO4, 100 mM NaCl,
0.1 mM EDTA, pH 7.0). Ta and Tb are explained in the text.
bChange in Tm per modification.
Figure 3.

850
T. Boesen et al. / Bioorg. Med. Chem. Lett. 13 (2003) 847–850
58.8, 55.5, 43.5, 43.4, 24.9, 24.8, 20.7, 20.6, 12.6. dP 149.8.
FAB+ MS: 715.3 (M+H+ calcd 715.3).
with a syringe, and washing with diethyl ether, followed by
shaking with concd aq ammonia (0.5 mL) for 2 h at rt. The
supernatant was removed with a syringe and the support
washed with water. Speed-Vac evaporation gave the crude
product which was purified by HPLC (Pharmacia FPLC)
using a Resource-Q 1 mL column and a 0–50% linear gradient
over 25 min of buffer A: 20 mM aq NaOH, buffer B: 20 mM aq
NaOH, 2.5 M aq NaCl, flow 1 mL/min, detection at 260 nm,
25 ꢀC. Fractions containing the full length product (latest
eluted peak) were neutralised with 0.5 M aq HCl, concentrated
to ca. 1 mL, and desalted on a NAP 10 column (Pharmacia) by
elution with water.
17. dT7TbT6, 92% pure, M 4167.0 (M calcd 4165.7), dT7TaT6,
81% pure, M 4166.8 (M calcd 4165.7), dT11TbTbT, 85% pure,
M 4136.5 (M calcd 4135.7), dGCATbCTbCAC, 97% pure, M
2598.6 (M calcd 2598.5).
18. To a solution (100 mL) of dT14 or dT11TbTbT (ca. 1.0 OD),
20 mM Tris, 10 mM MgCl2, pH 8.3 was added snake venom
phosphordiesterase (SVP, Phosphodiesterase I (Crotalus ada-
manteus Venom), ca. 0.005 u, Pharmacia) at 37 ꢀC. At 1, 2, 5,
10, 15, 25, 40, 60, and 120 min 10 mL samples were removed
and added to stop-solutions (50 mL, 1 M NH3, 2 mM EDTA).
The samples were analysed by CE (30 s electro-injection at
9 kv).
11. Oligonucleotide syntheses were performed on a Biosearch
8750 DNA Synthesizer using Pac-protected cyanoethyl phos-
phoramidites (Millipore) and 4a or 4b. The modified phos-
phoramidites (0.05 M in CH3CN) were manually coupled for
6 min using tetrazole (0.45 M in CH3CN) as activator, fol-
lowed by a washing step (10% H2O, 0.2% Ac2O and 0.2%
lutidine v/v/v in THF). The capping (standard Ac2O/lutidi-
ne+NMI) was followed by oxidation with tert-butyl hydro-
peroxide (80% in di-tert-butyl ether, ca. 0.5 M in CH2Cl2–
acetone 1:1 v/v) for 3 min, and detritylation (3% TCA in
CH2Cl2).
12. Scheuer-Larsen, C.; Dahl, B. M.; Wengel, J.; Dahl, O.
Tetrahedron Lett. 1998, 39, 8361.
13. Kuijpers, W. H. A.; Huskens, J.; Koole, L. H.; van
Boeckel, C. A. A. Nucleic Acids Res. 1990, 18, 5197.
14. Manoharan, M.; Lu, Y.; Casper, M. D.; Just, G. Org.
Lett. 2000, 2, 243.
15. Alul, R. H.; Singman, C. N.; Zhang, G.; Letsinger, R. L.
Nucleic Acids Res. 1991, 19, 1527.
16. Deprotection and cleavage were performed by shaking the
support-bound oligonucleotide with neat, dry diisopropyl-
amine (0.5 mL) for 14 h at rt, removal of the diisopropylamine