α-LNA (locked nucleic acid with α-D-configuration): Synthesis and selective parallel recognition of RNA

Article abstract of DOI:10.1002/1521-3765(20020201)8:3<712::AID-CHEM712>3.0.CO;2-0

α-LNA is presented as a stereoisomer of LNA (locked nucleic acid) with α-D-configuration. Three different approaches towards the thymine α-LNA monomer as well as the 5-methylcytosine α-LNA monomer are presented. Different α-LNA sequences have been synthesised and their hybridisation with complementary DNA and RNA has been evaluated by means of thermal stability experiments and circular dichroism spectroscopy. In a mixed pyrimidine sequence, α-LNA displays unprecedented parallel-stranded and selective RNA binding. Furthermore, a remarkable selectivity for hybridisation with RNA over DNA is indicated.

Full text of DOI:10.1002/1521-3765(20020201)8:3<712::AID-CHEM712>3.0.CO;2-0

FULL PAPER
a-LNA (Locked Nucleic Acid with a-d-Configuration):
Synthesis and Selective Parallel Recognition of RNA
[a]
Poul Nielsen,*Nanna K. Christensen, and Jakob K. Dalskov
Abstract: a-LNA is presented as a stereoisomer of LNA (locked nucleic acid) with
a-d-configuration. Three different approaches towards the thymine a-LNA mono-
mer as well as the 5-methylcytosine a-LNA monomer are presented. Different a-
LNA sequences have been synthesised and their hybridisation with complementary
DNA and RNA has been evaluated by means of thermal stability experiments and
circular dichroism spectroscopy. In a mixed pyrimidine sequence, a-LNA displays
unprecedented parallel-stranded and selective RNA binding. Furthermore, a
remarkable selectivity for hybridisation with RNA over DNA is indicated.
Keywords: locked nucleic acid
nucleosides oligonucleotides
RNA recognition
¥
¥
¥
imidines into a-ODNs was found to increase the thermal
affinities towards both complementary DNA and RNA
sequences, as similarly reported for the corresponding b-
ODNs.[9] The most significant improvements of a-DNA as
regards binding to RNA and DNA have been obtained with
modified phosphodiester linkages.[10] Thus, with nonionic
phosphoramidate or methylphosphonate linkages, the ther-
mal affinities of a-ODNs towards complementary RNA and,
especially, DNA sequences have been significantly improved
with increases in thermal stability of 0 18C and 1 38C per
modification, respectively, when compared to the correspond-
ing unmodified a-ODNs.[10] In contrast, similar modifications
of the b-ODNs were found to destabilise the duplexes.[10]
In general, the most successful approach towards ODNs
with high-affinity recognition of complementary nucleic acid
sequences has probably been the introduction of conforma-
tionally restricted analogues.[1, 13, 14] Thus, ONs in which the
nucleoside monomers contain bi- or tricyclic carbohydrate
moieties have been found to display high-affinity binding of,
in particular, complementary RNA.[14] The most enhanced
properties to date have been obtained with LNA (locked
nucleic acid) (Scheme 1).[15] LNA sequences are defined as
ONs that contain one or more LNA monomers, which are
nucleosides locked in an N-type conformation due to a
bicyclo[2.2.1]heptane carbohydrate skeleton (see the thymine
monomer 1; Scheme 1).[15] By the introduction of LNA
monomers into ONs, the formation of remarkably thermody-
namically favoured duplexes with both complementary DNA
and RNA has been demonstrated. Thus, both fully modified
LNA sequences as well as sequences that contain only a few
LNA monomers in mixmers with unmodified ribo- or 2'-
deoxynucleosides display very strong thermal affinities to-
wards both DNA and RNA, with increases in thermal stability
Introduction
The increasing demand for potential antisense agents and
diagnostic probes has motivated an intensive search for
nucleic acid analogues showing selective and high-affinity
recognition of complementary nucleic acid sequences.[1, 2]
Therefore, a plethora of chemically modified oligonucleotide
(ON) sequences have been introduced and investigated with
regard to their specific hybridisation with complementary
DNA and RNA.[2] A preliminary but extensively investigated
modification has been the anomeric inverted isomer of DNA,
namely a-DNA.[3 a-DNA sequences, or a-oligodeoxynu-
5]
cleotides (a-ODNs), have demonstrated efficient hybridisa-
tion with complementary nucleic acids, adopting an unusual
parallel strand orientation.[3, 4] Furthermore, a-ODNs are
highly resistant towards degradation by nucleases.[3] Like the
nucleosides of DNA and RNA, the a-2'-deoxynucleosides in
a-DNA exist in an equilibrium between the two low-energy
N- and S-type conformational ranges (Scheme 1).[6] In natural
b-configured nucleic acids, the former is predominant in
A-type duplexes and is preferred by the ribonucleosides in
RNA, while the latter is predominant in B-type duplexes and
is usually preferred by the 2'-deoxynucleosides in DNA.[7] The
a-2'-deoxynucleosides also display a preference for S-type
conformations.[6]
Chemically modified analogues of a-DNA have also been
12]
investigated.[8
Thus, the introduction of 5-C-propynylpyr-
[a] P. Nielsen, N. K. Christensen, J. K. Dalskov
Department of Chemistry
University of Southern Denmark
Odense University, 5230 Odense M (Denmark)
Fax : (‡45)6615-8780
E-mail: pon@chem.sdu.dk
712
¹ WILEY-VCH Verlag GmbH, 69451 Weinheim, Germany, 2002
0947-6539/02/0803-0712 $ 17.50+.50/0
Chem. Eur. J. 2002, 8, No. 3

712 722
Scheme 1. a) Conformational equilibrium for nucleosides between N- and S-type conformations; top: the b-(2'-deoxy)nucleosides in DNA and RNA;
bottom: the a-2'-deoxynucleosides in a-DNA; b) the structure of LNA and a-LNA; c) the a- and b-LNA thymine monomers.
of 3 88C per modification relative to unmodified ONs and
ODNs.[15] Furthermore, LNA displays a strong resistance
towards degradation by nucleases,[15] and preliminary but very
promising in vivo antisense activity.[16]
modification relative to the corresponding a-DNA sequence
(aT10) and 2.58C per modification relative to T10 (Table 1).[20]
Thus, a-LNA was found to display an affinity towards RNA
comparable to that of other stereoisomers of LNA,[19] and
Among the chemically modified analogues of a-DNA,
conformationally restricted analogues that include two differ-
ent a-configured bicyclic nucleoside monomers[11] have also
been introduced, but no improved DNA or RNA binding was
observed.[8, 11] However, these nucleoside analogues were
restricted to S-type conformations.[8, 11] No analogue of a-
DNA in which the a-nucleoside monomers are restricted in
N-type conformations has hitherto been described. However,
the N3'-P5'-phosphoramidite linkage, which has been a very
successful modification of b-ODNs due to their induced
preference to adopt N-type conformations,[17] has also been
introduced in a-ODNs.[12] Nonetheless, the latter seem to
prefer the S-type conformation, at least for pyrimidine
nucleotides, and large decreases in thermal affinity of mixed
pyrimidine sequences towards both DNA and RNA were
observed.[12] On the other hand, purine nucleotides seem to
prefer N-type conformations, and a fully N3'-P5'-modified
aA10 displays almost the same affinity towards complemen-
tary DNA and RNA as the native aA10.[12]
As a consequence of these results, a chemically modified
analogue of a-DNA with a strong conformational restriction
in the N-type conformational range might form the basis for
high-affinity parallel recognition of DNA and RNA sequen-
ces. Evidently, this could be addressed by the preparation and
oligomerisation of the anomeric inverted isomer of LNA, that
is a-LNA (or a-d-LNA; see the general structure and the
thymine monomer 2, Scheme 1), in which the bicyclo[2.2.1]-
heptane carbohydrate skeleton ensures a locked N-type
conformation. Thus, the stereoisomers of LNA, including
the enantiomer of a-LNA, that is a-l-LNA,[18] have recently
been demonstrated to be a class of very efficient RNA-
binding nucleic acid analogues.[19] Recently, in a preliminary
form, we introduced the synthesis and hybridisation proper-
ties of a-LNA containing only the thymine monomer 2.[20, 21]
Thus, a fully modified a-LNA sequence (aTL10† was found to
form a very stable duplex with complementary RNA (rA14).
The increase in thermal stability found was 1.28C per
Table 1. Hybridisation data of a-LNA sequences.
ODN
DNA (dA14)
RNA (rA14)
mm RNA (rA6CA7)
sequences Tm (DTm) [8C][a] Tm (DTm) [8C][a]
Tm (DTm) [8C][a]
21 5'-T14
33.0
32.0
30.0
43.0
22 5'-aT14
23 5'-aT7TLT6
24 5'-aT5TL4 T5
25 5'-T10
26 5'-aT10
27 5'-aTL10
25.5 (À6.5)[b] 35.0 (À8.0)[b]
26.0 (À1.5)[b] 24.5 (À4.6)[b]
22.0
18.0
< 10
20.0
33.5
22.0 (À11.5)[e]
37.0 (À8.0)[f]
45.0 (‡1.2[c]; ‡2.5[d]
)
[a] Melting temperatures obtained from the maxima of the first derivatives of the
melting curve (A260 vs. temperature) recorded in a buffer containing Na2HPO4
(10 mm), NaCl (100 mm), EDTA (0.1 mm), pH 7.0 with 1.5mm concentrations of
each strand. Values in brackets show the changes in Tm values per modification
compared with the reference strands. [b] Relative to 22. [c] Relative to 26.
[d] Relative to 25. [e] Relative to 26: rA14. [f] Relative to 27: rA14
.
although not quite as high as that of the original LNA,[15, 19]
this is unprecedented among a-d-configured nucleic acid
analogues. However, these preliminary results could neither
answer the question as to whether a-LNA prefers parallel or
antiparallel strand orientation, nor explain the apparently
strong selectivity for hybridisation with complementary RNA
over DNA. In this paper, the synthesis of a-LNA is described
in full detail, including different synthetic approaches towards
two different pyrimidine monomers. Furthermore, the hybrid-
isation properties of both oligothymidylate as well as mixed
pyrimidine a-LNA sequences with DNA and RNA comple-
ments are assessed by means of thermal stability experiments
and circular dichroism (CD) spectroscopy.
Results
Chemical synthesis: For the synthesis of a-LNA monomers,
several starting materials were considered. Thus, in a prelimi-
nary study, the coupling of thymine to precyclised bicy-
Chem. Eur. J. 2002, 8, No. 3
¹ WILEY-VCH Verlag GmbH, 69451 Weinheim, Germany, 2002
0947-6539/02/0803-0713 $ 17.50+.50/0
713

P. Nielsen et al.
FULL PAPER
clo[2.2.1]heptane carbohydrate precursors was investigat-
ed.[22] However, this approach has not been optimised to give
the target compound 2 in reasonable yields,[22] and a simpler
strategy has been applied.[20, 21] Thus, the same initial steps
were used and diacetone-d-allose 3 was converted to the diol
4[23] in four standard steps. In three further known steps,
including a selective benzylation,[24] 4 was converted to an
anomeric mixture of methyl furanosides 5 (Scheme 2).[22, 25]
Scheme 3. Alternative synthesis of a-LNA thymine monomer 2: a) MsCl,
pyridine, 92%; b) HCl, MeOH, H2O, 95%; c) i) TMSCl, N,O-bis(trimeth-
ylsilyl)acetamide, thymine, MeCN, then TMS triflate; ii) TBAF, THF, 24%
from 8; d) NaOH, EtOH, H2O, 46%; e) H2, Pd(OH)2/C, EtOH, dioxane,
91%.
nucleobase coupling reaction occurred without significant
stereoselectivity. Therefore, another approach was investigat-
ed in which a substrate was preconstructed for a stereo-
selective synthesis of an appropriate a-d-configured nucleo-
side (Scheme 4). Thus, the enantiomer of the target nucleo-
side 2 (the a-l-LNA thymine monomer)[18] has recently been
Scheme 2. Synthesis of a- and b-LNA thymine monomers 1 and 2:
a) ref. [23], 4 steps, approximately 80%; b) refs. [22, 24, 25],
3 steps,
approximately 65%; c) i) TMSCl, N,O-bis(trimethylsilyl)acetamide, thy-
mine, MeCN, then TMS triflate; ii) TBAF, THF; iii) NaH, DMF, 57% from
5; d) H2, Pd(OH)2/C, EtOH, 97%.
Subsequently, and as an alternative to cyclisation,[22] 5 was
directly used as the substrate in a modified Vorbr¸ggen-type
coupling reaction with thymine involving in situ silylation of
the 2-hydroxyl group and the nucleobase and with TMS
triflate as a Lewis acid catalyst. After a long reaction time
(4 days), a mixture of nucleosides was obtained which, after
desilylation, was treated with sodium hydride to afford an
anomeric mixture (a:b ꢀ 1.3:1) of bicyclic nucleosides 6 in a
reasonable yield (57%) over the three steps. Other conditions
were investigated for the coupling reaction, but neither the
yield nor the anomeric ratio were improved. The benzyl
ethers of 6 were cleaved by hydrogenation to give, after
chromatographic separation, the two anomeric LNA mono-
mers 1[15] and 2 in high yields.[20]
Scheme 4. Synthesis of a potential precursor for a-LNA monomers:
a) ref. [31], 3 steps, approximately 50%; b) ref. [32], 2 steps, 62%;
c) i) (COCl)2, DMSO, Et3N, CH2Cl2; ii) CH2O, NaOH, H2O, THF, 63%
from 12; d) ref. [29], 3 steps on enantiomers, 82%; e) refs. [27, 29], 5 steps
on enantiomers, 40%.
In order to avoid the selective benzylation in the prepara-
tion of 5, another approach was investigated using a bis(meth-
ylsulfonic) ester (Scheme 3). Thus, compound 4 was esterified
to give 7,[26] which was further converted to a mixture of
methyl furanosides 8. Various conditions for the coupling of
thymine to this substrate were investigated, without any
significant success, and after chromatographic purification the
a-nucleoside 9 could only be obtained in a relatively low yield.
Nevertheless, 9 was converted to the target bicyclic nucleoside
2 by treatment with strong aqueous base to simultaneously
bring about ring-closure and removal of the remaining
sulfonic ester group to give 10, followed by hydrogenation.
No explanation was found for the low yields in the coupling of
thymine with 8 and it was concluded that the original route via
5 (Scheme 2) was in fact superior.
synthesised[27] by using diacetone-d-glucose as the starting
material, which was converted to the 3'-epimer of 4 (the
enantiomer of 13; Scheme 4)[28] as an intermediate prod-
uct.[27, 29] Therefore, 2 should be accessible from 13, suggesting
the use of another carbohydrate precursor that may be
converted to 13. As the only reasonably cheap starting
material with this potential, d-arabinose was chosen. In
two[30] or three[31] steps, this was converted to the furanose
derivative 11, which was further converted to 12
(Scheme 4).[32] Subsequently, the primary alcohol function of
12 was oxidised using the Swern protocol to give an aldehyde
and, after the usual aldol condensation and Cannizzarro
reaction, 13 was obtained in good yield. We have not
employed this material in the synthesis of 2, but as 13, by
comparison of its NMR data, was shown to be the enantiomer
Even though the first route (Scheme 2) afforded the target
a-LNA monomer 2 in a reasonable overall yield, the
714
¹ WILEY-VCH Verlag GmbH, 69451 Weinheim, Germany, 2002
0947-6539/02/0803-0714 $ 17.50+.50/0
Chem. Eur. J. 2002, 8, No. 3

Locked Nucleic Acids
712 722
of a known compound,[28] and as that compound has been used
in the stereoselective eight-step synthesis of the a-l-LNA
thymine monomer,[27, 29] that is, through stereoselective cou-
pling of thymine to the enantiomers of 13a and without the
application of any chiral reagents, we concluded that this
strategy could afford 2 in the same overall yield. However,
and despite the lack of stereoselectivity, the original method
(Scheme 2) is still considered as superior due to the fact that it
requires fewer reaction steps and the key intermediate 4[23] is
more readily available compared to its stereoisomer 13.
To incorporate the a-LNA monomer into oligonucleotides,
2 was converted to the appropriately protected phosphor-
amidite derivative (Scheme 5). Thus, the primary alcohol
function of 2 was easily protected with the DMT (4,4'-
dimethoxytrityl) group to afford 14, which was further
with the corresponding unmodified a-2'-deoxynucleoside
thymine and 5-methyl cytosine phosphoramidites, which were
obtained according to literature methods.[36, 41] Hence, a-DNA
and a-LNA sequences were obtained with >98% stepwise
coupling yields (Tables 1 and 2) by using tetrazole activation
and coupling times of 10 15 min for 15 and 20. All modified
oligonucleotides were obtained on universal CPG support
(Biogenex) using the DMT-ON mode. This allowed the
synthesis of fully modified sequences after cleavage from the
solid support using LiCl in aqueous ammonia. The oligomers
were purified by using disposable reversed-phase chromatog-
raphy cartridges (Cruachem) yielding products with >90%
purity as judged on the basis of capillary gel electrophoresis.
The compositions of all the a-DNA and a-LNA sequences
were verified from their MALDI mass spectra.
Thermodynamic stability: The hybridisation between a-LNA
sequences and unmodified complementary DNA and RNA
sequences was explored by means of thermodynamic stability
examinations (Tables 1 and 2). Thus, the oligothymidylate a-
LNA sequences 23, 24 and 27, as well as their a-DNA
counterparts 22 and 26, were mixed with the corresponding
DNA and RNA complements, and the melting temperatures
of the complexes were determined (Table 1).[20] As described
in our preliminary communication,[20] the affinities of the a-
ODNs 22 and 26 towards the corresponding 14-mer DNA
sequence were, as expected,[4, 5] slightly decreased relative to
those of the unmodified ODNs 21 and 25. On the other hand,
and as reported in the literature,[4, 5] the same a-DNA
sequences displayed a significant increase in affinity towards
complementary RNA. One a-LNA monomer incorporated in
an otherwise unmodified a-DNA sequence, 23, resulted in
large decreases in affinity when compared to the unmodified
sequence 22. However, a block of a-LNA monomers, as in
sequence 24, diminished the relative decreases in affinity
introduced by each modification towards both DNA and
RNA.[20] As mentioned above, the fully modified decameric a-
LNA sequence 27 displayed large increases in binding affinity
to complementary RNA.[20, 42] On the other hand, no stable a-
LNA:DNA complex was observed. The melting curves
leading to these results are shown in Figure 1. The very clear
sigmoidal melting transitions indicate the formation of
duplexes between both a-DNA and a-LNA with comple-
mentary RNA. The observed hyperchromicity is even more
pronounced with a-LNA than with a-DNA. On the other
hand, the rather unstable duplex between a-DNA and RNA is
only slightly indicated, whereas no transition can be observed
for the mixture of a-LNA and complementary DNA. A very
weak transition at around 458C might be suggested from the
curve, but as mentioned in our former communication,[20] the
presence of a duplex can, in our opinion, be excluded by the
fact that this transition is not seen at a higher temperature
when measured at a higher ionic strength, in contrast to the
other transitions (data not shown).
Scheme 5. Synthesis a-LNA phosphoramidite building blocks 15 and 20.
a) DMTCl, AgNO3, pyridine, THF, DMF, 71%; b) NC(CH2)2OP(Cl)-
N(iPr)2, EtN(iPr)2, CH2Cl2, 96%; c) TBDMSCl, imidazole, DMF, 79%;
d) i) 1,2,4-triazole, POCl3, pyridine; ii) NH3, H2O, dioxane, 44% from 16;
e) N-benzoyltetrazole, CH3CN, 100%; f) TBAF, THF, 83%; g) see b),
60%.
converted to the phosphoramidite 15.[20] In order to permit the
preparation of mixed sequences, the well-known conversion
of a thymine moiety to a 5-methyl cytosine moiety was
evaluated.[33] However, the reported direct conversion of
thymidine phosphoramidite derivatives to give triazole pre-
cursors of 5-methyl cytidine phosphoramidite derivatives in
one step[34, 35] was found not to be successful in this case;
instead a conventional multistep conversion was employed
(Scheme 5).[33, 35, 36] Thus, the secondary alcohol function of 14
was converted to a silyl ether 16 and further converted to the
5-methyl cytidine derivative 17 via a triazole intermediate.
Protection of the exocyclic amino group of 17 as a benzoyl
amide proved surprisingly troublesome and treatment with
benzoyl chloride or benzoic anhydride in combination with
pyridine or DMAP did not afford 18 in more than 20% yield.
However, the use of the recently reported N-benzoyltetrazole
for this purpose[37] afforded 18 quantitatively. Subsequent
desilylation to give 19 was followed by phosphitylation to give
20 as a building block for oligonucleotide synthesis.[38, 39]
For the automated solid-phase synthesis of oligonucleotides
by the phosphoramidite approach,[40] the two a-LNA mono-
meric building blocks 15 and 20 were used in combination
Hybridisation data for the mixed decameric pyrimidine a-
LNA sequences (Table 2) demonstrate the preference for
parallel-stranded duplex formation for both a-DNA and a-
LNA. Thus, this decameric sequence was chosen as a non-self-
complementary and nonpalindromic sequence requiring only
Chem. Eur. J. 2002, 8, No. 3
¹ WILEY-VCH Verlag GmbH, 69451 Weinheim, Germany, 2002
0947-6539/02/0803-0715 $ 17.50+.50/0
715

P. Nielsen et al.
FULL PAPER
Table 2. Hybridisation data of mixed a-LNA sequences.
DNA (p)[a]
DNA (ap)[b]
RNA (p)[a]
RNA (ap)[b]
mm RNA (p)[a]
dGAGGAAGAAA dAAAGAAGGAG rGAGGAAGAAA rAAAGAAGGAG rGAGGCAGAAA
ODN sequences
Tm (DTm) [8C][c]
Tm (DTm) [8C][c]
Tm (DTm) [8C][c]
Tm (DTm) [8C][c]
Tm (DTm) [8C][c]
28 5'-mCTmCmCTTmCTTT
< 10
43.0
36.0
< 10
< 10
< 10
< 10
16.0
32.0
26.0 (À6.0)[d]
22.0 (À1.7)[d]
61.5 (‡2.9)[d]
44.5
< 10
< 10
< 10
< 10
29 5'-a-mCTmCmCTTmCTTT
30 5'-a-mCTmCmCTTLmCTTT
31 5'-a-mCTLmCmCTLTLmCTLTLTL
32 5'-a-mCLTLmCLmCLTLTLmCLTLTLTL
30.5 (À12.5)[d]
< 10
< 10
47.0 (À14.5)[e]
[a] p ˆ parallel. [b] ap ˆ antiparallel. [c] See Table 1. Values in brackets show the changes in Tm values per modification compared with the reference strands.
[d] Relative to 29. [e] Relative to 32: RNA(p).
played in Figure 2. Again, a very clear transition verifies
duplex formation between the mixed a-DNA sequence and
complementary parallel RNA, but in this case the corre-
sponding a-LNA:RNA duplex, despite having a much higher
Figure 1. Absorption versus temperature curves displaying the melting
profiles of homothymidine homoadenine duplexes. a-DNA corresponds
to 26, a-LNA corresponds to 27, DNA and RNA correspond to dA14 and
rA14, respectively.
two different building blocks, and the 5-methylcytosine
monomer was chosen instead of cytosine due to its ready
availability from the corresponding thymine monomer (vide
Figure 2. Absorption versus temperature curves displaying the melting
profiles of duplexes formed by the mixed sequences. a-DNA corresponds
to 29, a-LNA corresponds to 32, DNA and RNA correspond to their
supra) and as the 5-methyl groups can be expected to further
stabilise the duplex structures.[1, 2] As expected,[4] the unmodi-
fied a-ODN 29 only formed stable duplexes with comple-
mentary parallel DNA and RNA sequences, whereas the
corresponding unmodified ODN 28 formed the most stable
duplexes with antiparallel complements. When one a-LNA
thymine monomer was incorporated, the sequence 30 showed
an even larger decrease in affinity towards the parallel DNA
complement than that observed for the oligothymidylate
sequence 23 (À12.58C per modification compared to
À6.58C). For 30, however, and in contrast to 23, the decrease
in thermal stability was smaller with complementary RNA. A
mixed sequence 31 containing six a-LNA T-monomers and
four conventional a-DNA 5-methylcytosine monomers dis-
played no recognition of either parallel or antiparallel
complementary DNA, whereas a duplex was obtained with
parallel RNA, for which the relative decrease in thermal
stability introduced by each bicyclic monomer was smaller
than that observed for 30 (À1.7 8C per modification compared
to À6.08C).
parallel complements (see Table 2).
thermal stability, displays a smaller hyperchromicity. A very
clear transition between a-DNA and complementary parallel
DNA can be observed, whereas no indication whatsoever of
any complex between a-LNA and DNA can be observed. No
sign of a melting transition was observed with 32 alone (not
shown). As with the homothymidine sequence 27 (Table 1),[20]
the presence of a duplex with RNA rather than an intra-
molecular complex was further verified by the observation
that a stable duplex was also formed with a single mismatch
RNA sequence with an expected decrease in thermal stability
(À14.58C, Table 2). Furthermore, no significant changes in
thermal stability were observed for the duplexes of 29 and 32
with RNA when measured at pH 5.6 (data not shown).
Circular dichroism: CD spectra were obtained for all the
uniformly modified a-DNA, a-LNA, DNA and RNA se-
quences as single strands and in the different mixtures using
the same buffers and strand concentrations as in the melting
experiments. Selected spectra are shown in Figures 3 and 4.
As expected, the CD spectrum of the homothymidine a-
LNA:RNA duplex (Figure 3) is an almost perfect mirror
image of the spectrum reported earlier for the similar
sequence of a-l-LNA mixed with l-RNA.[42] The correspond-
The fully modified mixed a-LNA sequence 32 displayed,
like 27, no binding to complementary DNA, whereas a very
stable parallel-stranded duplex was formed with complemen-
tary RNA. Thus, an increase in thermal stability of almost 38C
for each bicyclic monomer was observed in comparison with
the a-DNA sequence 29. The melting transitions are dis-
716
¹ WILEY-VCH Verlag GmbH, 69451 Weinheim, Germany, 2002
0947-6539/02/0803-0716 $ 17.50+.50/0
Chem. Eur. J. 2002, 8, No. 3

Locked Nucleic Acids
712 722
Figure 4. CD spectrocopy of the mixed sequences: a) CD spectra recorded
at 258C for the duplexes and mixtures. b) CD spectra recorded at 258C for
the single strands. a-DNA corresponds to 29, a-LNA corresponds to 32,
DNA and RNA correspond to their parallel complements (see Table 2).
Figure 3. CD spectrocopy of homothymidine homoadenine sequences:
a) CD spectra recorded at 258C for the duplexes and mixtures. b) CD
spectra recorded at 258C for the single strands. a-DNA corresponds to 26,
a-LNA corresponds to 27, DNA and RNA correspond to dA14 and rA14
,
respectively.
ing spectrum of the a-DNA:RNA duplex displays some
similarities, but with a much larger negative ellipticity value at
250 nm and a much larger positive ellipticity value at 265 nm.
As expected, this spectrum is very similar to that reported for
a similar complex.[5] For the mixtures of homothymidine a-
DNA and a-LNA with DNA, the CD spectra are relatively
similar, although the latter displays a negative ellipticity value
at 270 nm. As the spectra were obtained at 258C, no duplexes
should be present in the latter two cases. For the single
strands, no similarities between a-DNA and a-LNA were
observed, the latter displaying a remarkably low general CD
activity giving only a small band at around 270 nm (Figure 3).
For the mixed sequences (Table 2), a different picture
emerges (Figure 4). Thus, for the a-LNA:RNA duplex, an
unusual CD spectrum is obtained, dominated by a large
negative ellipticity value at 270 nm. The spectrum of the
corresponding a-DNA:RNA duplex is completely different,
with a large positive ellipticity value at 220 nm. The a-
LNA:DNA mixture gives rise to a CD spectrum with some
similarities to that obtained with a-LNA:RNA, but it also
shows a positive ellipticity value at 245 nm. The a-DNA:DNA
duplex also gives rise to a somewhat similar CD spectrum.
The single-stranded a-LNA has a CD spectrum dominated by
the same large negative ellipticity value, but slightly shifted
towards 280 nm. However, this spectrum displays no similar-
ities with the spectra of the single-stranded a-DNA or of the
complementary DNA and RNA sequences.
Discussion
Four synthetic approaches towards the a-LNA thymine
monomer 2 have been investigated, three of which are
described in this paper. In the first (Scheme 2), 2 was obtained
in eleven steps and 15% overall yield. In the alternative
approach avoiding the selective benzylation (Scheme 3), 2
was obtained in ten steps but only 7% overall yield. Following
the third approach (Scheme 4), the synthesis of 2 was not
accomplished but a precursor has been obtained, and, in
theory, 2 could be obtained in 14 steps and 6.5% overall yield.
Therefore, despite the two inconvenient isomer separations,
the first approach (Scheme 2) has in general been applied for
the bulk synthesis of 2. However, the latter method
(Scheme 4) might be superior in the forthcoming synthesis
of the remaining pyrimidine and purine monomers of a-LNA,
as the diacetylate 13a might be much more efficient as a
glycoside donor in nucleobase coupling reactions compared to
the methyl furanoside 5, when the other nucleobases are
Chem. Eur. J. 2002, 8, No. 3
¹ WILEY-VCH Verlag GmbH, 69451 Weinheim, Germany, 2002
0947-6539/02/0803-0717 $ 17.50+.50/0
717

P. Nielsen et al.
FULL PAPER
applied. As expected, the a-LNA phosphoramidites 15 and
20, like the parent LNA phosphoramidites,[15] can be very
conveniently used in automated solid-phase synthesis of
oligonucleotides.[43]
with the postulate that, due to stereoelectronic effects, a-2'-
deoxynucleosides are more restricted towards S-type con-
formations than the b-2'-deoxynucleosides.[6] Thus, the
anomeric effect and the O3' O4' gauche effect both push
the a-2'-deoxynucleosides towards S-type conformations,[6]
whereas they work in opposite directions for b-2'-deoxynu-
cleosides.[6, 7] Therefore, the N-type conformations cannot
favourably be obtained for the a-2'-deoxynucleotides, and
unfavourably preorganised oligonucleotides with a mixture of
N- and S-type conformations are obtained. Thus, when
comparing the parallel duplexes obtained with RNA, 32 with
31, an extraordinary decrease in thermal stability of À108C
for each inserted unmodified a-2'-deoxynucleotide can be
observed. This further confirms the assumption that the pure
a-LNA:RNA duplex has a completely different structure
compared to the a-DNA:RNA duplex. This property of a-
LNA diminishes the potential of tuning the parallel recog-
nition of RNA by the design of mixmer sequences in the same
way as is possible for LNA/DNA mixmers.[15, 16]
In general, a-LNA in the present sequences displays the
highest affinity towards RNA obtained with any a-d-config-
ured oligonucleotide analogue. Furthermore, the a-LNA:
RNA duplex probably displays the hitherto most stable
parallel-stranded duplex. Importantly, the selectivity for
parallel over antiparallel complementary sequences seems
to be remarkably high on the basis of our data (Table 2).
Furthermore, the discrimination of single mismatches in the
recognition of complementary parallel RNA, though estab-
lished with only two C T mismatches (Tables 1 and 2), seems
to be comparable to the selectivity obtainable for natural
ONs. For the present a-LNA sequence 32, some duplex
stabilisation is probably obtained through the complete
5-methylation of the pyrimidines. However, and even though
this has to be verified experimentally, we certainly expect
other mixed sequences containing the remaining natural
pyrimidine and purine nucleobases to reveal similar stabilisa-
tion through the locked N-type conformations. In this way, a-
LNA reveals a potential for a new concept in the design of
selective nucleic acid recognising ONs for use as antisense
agents or diagnostic probes. Parallel-stranded recognition is as
yet a relatively unexplored and perhaps underestimated
phenomenon in this context.
A strong selectivity of a-LNA for RNA over DNA has also
been strongly indicated. Thus, the thermal transition curves
(Figures 2 and 3) indicate no melting of a duplex in the case of
the fully modified a-LNA sequences 27 and 32 with comple-
mentary DNA. The CD spectra are not very informative.
Nevertheless, the CD spectrum of the a-LNA:DNA mixture
(27:rA14, Figure 4) may be interpreted as the expected sum of
the spectra of the two single strands, thereby confirming that
no complex between the strands is present. However, for the
mixed a-LNA:DNA mixture (32:pDNA, Figure 5), the same
conclusion cannot immediately be drawn from the CD
spectra, and the alternative that complexes not detectable
by UV at 260 nm might in theory be present has to be
considered. Thus, a duplex structure in which the overall base-
stacking is not very different to the combined base-stacking of
the two single strands would give a very small and undetect-
able hyperchromicity in the melting transition. Nevertheless,
The hybridisation data (Tables 1 and 2, Figures 1 and 2)
clearly demonstrate that a-LNA forms very stable complexes
with complementary RNA. However, and in contrast to the
behaviour of LNA in a context of 2'-deoxynucleotides,[15]
ODNs with single a-LNA monomers in combination with a-
2'-deoxynucleotides, that is a-LNA/a-DNA mixmers, form
destabilised duplexes. Compared to the nucleosides in a-
DNA, the nucleosides in a-LNA have been locked in an
N-type conformation, which is unprecedented for a-DNA
analogues. This conformation dictates a completely different
duplex structure of the fully modified a-LNA:RNA duplexes,
as indicated by the CD spectra. In general, CD spectroscopy
gives a convenient but only very superficial indication of the
structural behaviour of nucleic acids. However, the CD
spectra of the a-LNA:RNA duplexes (Figures 3 and 4),
although not very conclusive, strongly indicate a structure not
resembling either the corresponding a-DNA:RNA duplexes
or typical A- or B-type duplexes. Thus, the base-pairing mode
followed is not necessarily the Watson Crick type found in,
for example, a parallel a-DNA:DNA duplex.[44] Nevertheless,
Hoogsteen or reversed-Hoogsteen base-pairing can be ex-
cluded by the fact that in a preliminary experiment neither the
mixed pyrimidine a-LNA:RNA duplex nor the a-DNA:RNA
duplex displayed any increase in thermal stability when
measured at a lower pH. As exemplified by a bicyclo-DNA
system,[45] Hoogsteen or reversed-Hoogsteen base-pairing
systems display a strong pH dependence as protonated
cytidines have to be involved in the base-pairing. Moreover,
the CD spectra of the two single-stranded a-LNA oligomers
(Figures 3 and 4) are very peculiar, especially with regard to
the very low overall ellipticities for the oligothymidylate
sequence 27. Furthermore, remarkable differences are gen-
erally observed between oligothymidylate and mixed sequen-
ces. This indicates only that the perturbation of the nucleo-
bases by the chiral centres, especially the anomeric centre,
which dictates the shape of the CD spectrum, is very different
for the locked compared to the unlocked nucleotides and is
highly dependent on the nucleotide sequence. Thus, the
structure of the a-LNA:RNA hybrid as well as of the a-LNA
single strands remains unknown and further studies on the
exact duplex structure will be performed in due course.
The locked N-type conformation of the a-LNA monomer
might also provide a logical explanation for the unfavourable
hybridisation behaviour of a-LNA/a-DNA mixmer ODNs.
Thus, for the parent LNA in LNA/DNA mixmers, NMR
studies have demonstrated the ability of the LNA monomers
to alter the conformational equilibrium of neighbouring 2'-
deoxynucleotides in favour of N-type conformations.[46] This
might be the reason for the very thermodynamically stable
duplexes formed by LNA/DNA mixmers with complemen-
tary DNA and RNA (vide supra).[15] Apparently, the a-LNA
monomers, which are also locked in N-type conformations,
cannot induce a similar conformational shift on the neigh-
bouring unmodified a-2'-deoxynucleotides, at least not a shift
that is favourable for duplex formation. This is in agreement
718
¹ WILEY-VCH Verlag GmbH, 69451 Weinheim, Germany, 2002
0947-6539/02/0803-0718 $ 17.50+.50/0
Chem. Eur. J. 2002, 8, No. 3

Locked Nucleic Acids
712 722
aqueous NaHCO3 solution (2 Â 200 mL) and water (2 Â 200 mL), dried
(MgSO4) and concentrated in vacuo. The residue was purified by column
chromatography (CH2Cl2/MeOH, 96:4) to afford the product as a white
solid in an anomeric mixture (3.90 g; 57%, a:b ꢀ 1.3:1). 1H NMR
(300 MHz, CDCl3, 258C, TMS): d ˆ 9.05 (brs), 8.98 (brs), 7.49 (d,
we suggest on the basis of the present data that an unpre-
cedented selectivity for RNA has been introduced. This prop-
erty certainly deserves further investigation as an RNA-selec-
tivity so pronounced is unprecedented and potentially very
useful in the design of therapeutic ONs and diagnostic probes.
3
3J(H,H) ˆ 1.0 Hz), 7.46 (d, J(H,H) ˆ 1.0 Hz), 7.36 7.26 (m), 5.95 (s), 5.65
(s), 4.67 4.50 (m), 4.23 (s), 4.16 (d, 3J(H,H) ˆ 6.3 Hz), 4.03 3.96 (m),
3.85 3.80 (m), 1.95 (s), 1.63 (s); 13C NMR (75 MHz, CDCl3, 258C, TMS):
d ˆ 163.83, 150.49, 149,83, 137.53, 136.99, 135.50, 134.88, 128.65, 128.59,
128.56, 128.23, 128.15, 128.03, 127.87, 127.76, 127.73, 127.71, 110.21, 109.72,
89.38, 87.40, 87.30, 87.20, 79.53, 75.61, 73.80, 73.40, 72.18, 71.99, 65.55, 64.46,
Conclusion
‡
12.60, 12.15; MS FAB: m/z (%): 451(70) [M‡H] ; elemental analysis calcd
The synthesis of two a-LNA monomers and oligomerisation
of these into a-LNA oligonucleotide sequences have been
efficiently accomplished. In this way, the successful concept of
locking the nucleoside monomers of oligonucleotides in
N-type conformations by the use of the LNA-type bicyclic
nucleoside structure has proven its strength in the a-DNA
series as well. Thus, a-LNA displays an unprecedented
parallel recognition of RNA and forms the strongest parallel
orientated duplex known. A new concept for high-affinity
nucleic acid recognition has thus been introduced. This
extraordinary nucleic acid recognising behaviour of a-LNA,
including the apparent selectivity for complementary RNA
over DNA, is conducive to further investigations, as well as to
the synthesis of mixed a-LNA sequences containing all the
natural nucleobases.
(%) for C25H26N2O6 ¥ 0.25H2O (454.99): C 66.00, H 5.87, N 6.16; found: C
66.06, H 5.70, N 6.53.
(3R)- and (3S)-(1S,4R,7S)-7-Hydroxy-1-hydroxymethyl-3-(thymin-1-yl)-
2,5-dioxabicyclo[2.2.1]heptane (1 and 2):
A solution of 6 (520 mg,
1.15 mmol) in ethanol (3.3 mL) was stirred at room temperature and
20% palladium hydroxide on carbon (250 mg) was added. The mixture was
degassed several times with argon and then placed under a hydrogen
atmosphere. After stirring for 5 days, the mixture was filtered through a
layer of Celite, the filtrate was concentrated and the residue was purified by
column chromatography (CH2Cl2/MeOH, 9:1) to give three fractions as
white solids: 1 (57.5 mg; 18%), a mixture of 1 and 2 (87.5 mg; 28%, ratio ꢀ
7:1), and 2 (160 mg, 51%); 1: All NMR data were in accordance with
literature data.[15] Data for 2: 1H NMR (300 MHz, [D6]DMSO, 258C, TMS):
d ˆ 11.34 (s, 1H; NH), 7.62 (d, 3J(H,H) ˆ 0.9 Hz, 1H; 6-H), 5.85 (s, 1H; 1'-
3
3
H), 5.84 (d, J(H,H) ˆ 4.0 Hz, 1H; 3'-OH), 4.92 (t, J(H,H) ˆ 5.5 Hz, 1H;
5'-OH), 4.25 (d, 3J(H,H) ˆ 4.0 Hz, 1H; 3'-H), 4.19 (s, 1H; 2'-H), 3.92 (m,
2H; 5''-H), 3.72 (d, 3J(H,H) ˆ 5.5 Hz, 2H; 5'-H), 1.82 (d, 3J(H,H) ˆ 0.9 Hz,
3H; CH3); 13C NMR (75 MHz, [D6]DMSO, 258C, TMS): d ˆ 164.00,
150.43, 135.96, 107.81, 90.93, 86.36, 78.78, 72.40, 71.92, 57.33, 12.11; NMR
data were in accordance with literature data on the enantiomer;[27] MS EI:
‡
m/z (%): 270 (100) [M] ; elemental analysis calcd (%) for C11H14N2O6 ¥
0.33H2O (276.25): C 47.78, H 5.35, N 10.14; found: C 47.96, H 4.99, N 9.73.
Experimental Section
3-O-Benzyl-1,2-O-isopropylidene-5-O-methylsulfonyl-4-C-methylsulfonyl-
oxymethyl-a-d-ribofuranose (7):
A
solution of the diol 4[23] (9.24 g,
All commercial reagents were used as supplied. All reactions were
performed under an atmosphere of nitrogen. Column chromatography
was carried out by using glass columns packed with silica gel 60 (0.040
0.063 mm). NMR spectra were obtained on a Bruker AC250, a Varian
Gemini2000, or a Varian Unity500 spectrometer. Chemical shifts are
quoted relative to TMS as an internal standard. 1H,1H-COSY and 1H NOE
difference spectra were recorded for compound 2. Assignments of NMR
signals follow standard carbohydrate and nucleoside style. However,
bicyclic compounds are named according to the von Baeyer nomenclature.
FAB mass spectra were recorded in positive-ion mode on a Kratos
MS50TC spectrometer; EI mass spectra were recorded on an SSQ710
Finnigan MAT spectrometer. High-resolution MALDI mass determina-
tions were performed on an Ionspec Ultima Fourier transform mass
spectrometer. Microanalyses were performed at The Microanalytical
Laboratory, Department of Chemistry, University of Copenhagen.
29.8 mmol) in anhydrous pyridine (94 mL) was cooled to 08C, whereupon
methanesulfonyl chloride (11.5 mL, 149 mmol) was added. The mixture
was stirred at room temperature for 2 h, and then the reaction was
quenched by adding iced water (450 mL). The resulting mixture was
extracted with dichloromethane (3 Â 250 mL). The combined organic
phases were washed with saturated aqueous NaHCO3 solution (250 mL),
dried (MgSO4) and concentrated in vacuo. The residue was purified by
column chromatography (CH2Cl2/MeOH, 95:5) to give the product (12.8 g,
92%) as an oil. 1H NMR (300 MHz, CDCl3, 258C, TMS): d ˆ 7.38 7.33 (m,
3
3
5H), 5.79 (d, J(H,H) ˆ 3.9 Hz, 1H), 4.88 (d, J(H,H) ˆ 11.9 Hz, 1H), 4.77
(d, 3J(H,H) ˆ 11.5 Hz, 1H), 4.65 (dd, 3J(H,H) ˆ 3.9, 4.7 Hz, 1H), 4.57 (d,
3J(H,H) ˆ 11.5 Hz, 1H), 4.42 (d, 3J(H,H) ˆ 11.9 Hz, 1H), 4.32 (d,
3J(H,H) ˆ 10.9 Hz, 1H), 4.20 (d, 3J(H,H) ˆ 4.7 Hz, 1H), 4.15 (d,
3J(H,H) ˆ 10.9 Hz, 1H), 3.08 (s, 3H), 2.98 (s, 3H), 1.69 (s, 3H), 1.34 (s,
3H); 13C NMR (75 MHz, CDCl3, 258C, TMS): d ˆ 136.62, 128.61, 128.37,
128.10, 113.97, 104.47, 83.20, 78.32, 77.77, 72.84, 69.44, 68.62, 37.99, 37.43,
(3R/S)-(1S,4R,7S)-7-Benzyloxy-1-benzyloxymethyl-3-(thymin-1-yl)-2,5-di-
oxabicyclo[2.2.1]heptane (6): A mixture of methyl furanosides 5[22, 25]
(6.79 g, 15.0 mmol) and thymine (3.78 g, 30.0 mmol) was dried and
dissolved in anhydrous CH3CN (55 mL). TMSCl (1.90 mL, 15.0 mmol)
and N,O-bis(trimethylsilyl)acetamide (29.3 mL, 120 mmol) were added
and the mixture was stirred at 608C for 1 h. The solution was then cooled to
08C and TMS triflate (13.6 mL, 75 mmol) was added dropwise. The mixture
was stirred at 708C for 4 days and then quenched by pouring it into an ice-
cold saturated aqueous solution of NaHCO3 (300 mL). The resulting
mixture was extracted with dichloromethane (400 mL) and the organic
fraction was washed with saturated aqueous NaHCO3 solution (2 Â
300 mL), dried (MgSO4) and concentrated in vacuo. The residue was
redissolved in THF (40 mL) and to this solution was added a 1.0m solution
of TBAF in THF (15.0 mL). The resulting mixture was stirred for 1 h and
then diluted with dichloromethane (250 mL). The mixture obtained was
washed with saturated aqueous NaHCO3 solution (2 Â 250 mL) and water
(2 Â 250 mL), dried (MgSO4) and concentrated in vacuo. The residue was
redissolved in anhydrous DMF and the solution was stirred at 08C. A 60%
dispersion of NaH in mineral oil (1.20 g, 30 mmol) was added and the
mixture was stirred at room temperature for 16 h. Dichloromethane
(250 mL) was then added and the mixture was washed with saturated
‡
‡
26.17, 25.61; MS FAB: m/z (%): 489 (3) [M‡Na] , 451 (55) [M À CH3] ;
elemental analysis calcd (%) for C18H26O10S2 (466.52): C 46.34, H 5.61;
found: C 46.74, H 5.47.
Methyl (3-O-benzyl-5-O-methylsulfonyl-4-C-methylsulfonyloxymethyl)-d-
ribofuranoside (8):
A solution of the bis(sulfonic ester) 7 (1.46 g,
3.13 mmol) in water (6.0 mL) and methanol (11.5 mL) was cooled to 08C
and a 28% solution of HCl in methanol (31 mL) was added. The mixture
was stirred at room temperature for 19 h, then neutralised with NaHCO3
(s), and water (60 mL) was added. The resulting mixture was extracted with
dichloromethane (2 Â 70 mL). The combined organic phases were washed
with water (60 mL), dried (MgSO4) and concentrated in vacuo. The residue
was purified by column chromatography (CH2Cl2/MeOH, 99:1) to give the
product as an anomeric mixture (1.31 g; 95%, a:b ꢀ 1:10) as an oil.
1H NMR (300 MHz, CDCl3, 258C, TMS): d ˆ 7.42 7.33 (m), 4.90 4.00
(m), 3.48 (s), 3.33 (s), 3.06 (s), 3.03 (s), 3.00 (s), 2.99 (s); 13C NMR (75 MHz,
CDCl3, 258C, TMS): (major isomer) d ˆ 136.36, 128.84, 128.66, 128.25,
107.72, 81.58, 81.51, 73.88, 73.74, 69.56, 69.34, 55.42, 37.47, 37.38; MS FAB:
‡
m/z (%): 441 (32) [M‡H] ; elemental analysis calcd (%) for C16H24O10S2
(440.48): C 43.63, H 5.49; found: C 43.55, H 5.14.
Chem. Eur. J. 2002, 8, No. 3
¹ WILEY-VCH Verlag GmbH, 69451 Weinheim, Germany, 2002
0947-6539/02/0803-0719 $ 17.50+.50/0
719

P. Nielsen et al.
FULL PAPER
1-(3-O-Benzyl-5-O-methylsulfonyl-4-C-methylsulfonyloxymethyl-a-d-ribo-
50 mL), and the combined organic phases were washed with saturated
aqueous NaHCO3 solution (50 mL), dried (MgSO4) and concentrated in
vacuo. The residue was purified by column chromatography (CH2Cl2/
MeOH, 97:3) to give the product (703 mg, 63%) as a white solid. 1H NMR
(in accordance with literature data on the enantiomer[28]) (300 MHz,
CDCl3, 258C, TMS): d ˆ 7.39 7.26 (m, 5H), 6.02 (d, 3J(H,H) ˆ 4.4 Hz, 1H),
4.79 4.75 (m, 2H), 4.55 (d, 3J(H,H) ˆ 11.6 Hz, 1H), 4.12 (d, 3J(H,H) ˆ
1.7 Hz, 1H), 3.77 3.58 (m, 4H), 2.42 2.31 (m, 2H), 1.54 (s, 3H), 1.35 (s,
3H); 13C NMR (75 MHz, CDCl3, 258C, TMS): d ˆ 136.79, 128.65, 128.22,
127.70, 113.13, 104.89, 89.91, 85.88, 84.88, 72.63, 63.80, 63.38, 27.31, 26.77.
furanosyl)thymine (9):
A mixture of methyl furanosides 8 (325 mg,
0.74 mmol) and thymine (186 mg, 1.48 mmol) were dried and dissolved in
anhydrous CH3CN (2.7 mL). TMSCl (0.094 mL, 0.74 mmol) and N,O-
bis(trimethylsilyl)acetamide (1.45 mL, 5.9 mmol) were added and the
mixture was stirred at 608C for 1 h. The solution was cooled to 08C,
whereupon TMS triflate (0.67 mL, 3.7 mmol) was added dropwise. The
mixture was stirred at 708C for 4 days, after which a second portion of TMS
triflate (0.27 mL, 1.5 mmol) was added. After stirring for a further 2 days, a
third portion of TMS triflate (0.27 mL, 1.5 mmol) was added; the mixture
was stirred for a further 2 days and then finally quenched by pouring it into
an ice-cold saturated aqueous NaHCO3 solution (20 mL). The resulting
mixture was extracted with dichloromethane (30 mL) and the organic
fraction was washed with saturated aqueous NaHCO3 solution (2 Â 20 mL),
dried (MgSO4) and concentrated in vacuo. The residue was redissolved in
THF (2.1 mL) and treated with a 1.0m solution of TBAF in THF (0.74 mL).
The solution was stirred for 1 h and then diluted with dichloromethane
(20 mL). The resulting solution was washed with saturated aqueous
NaHCO3 solution (2 Â 20 mL) and water (2 Â 20 mL), dried (MgSO4) and
concentrated in vacuo. The residue was purified by column chromatog-
raphy (CH2Cl2/MeOH, 96:4) to afford the product as a white solid (93 mg;
24%) together with other impure fractions. 1H NMR (300 MHz, CDCl3,
258C, TMS): d ˆ 10.95 (brs, 1H), 7.85 (s, 1H), 7.42 7.32 (m, 5H), 6.11 (d,
3J(H,H) ˆ 2.3 Hz, 1H), 5.19 (d, 3J(H,H) ˆ 11.8 Hz, 1H), 5.08 (brs, 1H), 4.85
(d, 3J(H,H) ˆ 11.8 Hz, 1H), 4.60 (d, 3J(H,H) ˆ 11.8 Hz, 1H), 4.44 4.35 (m,
5H), 4.26 (d, 3J(H,H) ˆ 10.6 Hz, 1H), 3.07 (s, 3H), 2.90 (s, 3H), 1.83 (s,
3H); 13C NMR (75 MHz, CDCl3, 258C, TMS): d ˆ 166.37, 150.69, 139.54,
136.86, 128.74, 128.47, 128.26, 108.71, 87.14, 82.33, 77.94, 72.16, 69.41, 68.55,
67.90, 37.76, 37.35, 11.81.
(1R,3S,4R,7S)-1-(4,4'-Dimethoxytrityl)oxymethyl-7-hydroxy-3-(thymin-1-
yl)-2,5-dioxabicyclo[2.2.1]heptane (14):
A
solution of
2
(152 mg,
0.56 mmol), AgNO3 (191 mg, 1.1 mmol), 4,4'-dimethoxytrityl chloride
(209 mg, 0.62 mmol) and anhydrous pyridine (0.45 mL, 5.6 mmol) in
anhydrous DMF (0.5 mL) and THF (5.0 mL) was stirred at room temper-
ature for 22 h. The reaction was then quenched by the addition of methanol
(1.0 mL) and the mixture was concentrated in vacuo. The residue was
purified by column chromatography (CH2Cl2/MeOH/pyridine, 95:4.5:0.5)
to give the product (227 mg, 71%) as a white solid, which was used in the
1
next step without further purification; H NMR (300 MHz, CDCl3, 258C,
TMS): d ˆ 9.39 (brs, 1H), 7.56 (d, 3J(H,H) ˆ 1.2 Hz, 1H), 7.46 7.20 (m,
9H), 6.84 (d, 3J(H,H) ˆ 9.1 Hz, 4H), 5.99 (s, 1H), 4.52 (s, 1H), 4.44 (s, 1H),
4.15 (d, 3J(H,H) ˆ 8.9 Hz, 1H), 4.06 (d, 3J(H,H) ˆ 8.9 Hz, 1H), 3.79 (s, 6H),
3.56 3.37 (m, 2H), 1.99 (s, 3H); 13C NMR (75 MHz, CDCl3, 258C, TMS):
d ˆ 164.19, 158.75, 150.60, 144.50, 135.57, 135.47, 135.40, 130.09, 129.20,
129.08, 128.27, 128.06, 127.88, 127.83, 127.11, 127.07, 113.31, 113,18, 109.66,
89.63, 87.34, 86.38, 79.12, 74.19, 72.83, 59.93, 55.16, 12.71; MS FAB: m/z
‡
(%): 573 (20) [M‡H] .
(1R,3S,4R,7S)-1-(4,4'-Dimethoxytrityl)oxymethyl-7-(O-2-cyanoethyl-N,N-
diisopropylphosphityl)oxy-3-(thymin-1-yl)-2,5-dioxabicyclo[2.2.1]heptane
(15): Compound 14 (121 mg, 0.211 mmol) was freed of solvents by co-
evaporation with anhydrous CH3CN and redissolved in anhydrous
dichloromethane (1.2 mL). The solution was stirred at room temperature
and diisopropyl ethylamine (0.171 mL) and NC(CH2)2OP(Cl)N(iPr)2
(0.081 mL, 0.36 mmol) were added. The mixture was stirred for 2 h, then
diluted with EtOAc (15 mL), washed with saturated aqueous NaHCO3
solution (2 Â 10 mL) and brine (2 Â 10 mL), dried (Na2SO4) and concen-
trated in vacuo. The residue was purified by column chromatography
(CH2Cl2/Et3N, 99:1), dissolved in toluene (1.0 mL), and precipitated from
petroleum ether (65 mL) at À308C to afford the product as a white solid
(156 mg, 96%). 31P NMR (121.5 MHz, CDCl3, 258C, 85% H3PO4 external):
d ˆ 150.9, 151.1.
(1S,3S,4R,7S)-7-Benzyloxy-1-hydroxymethyl-3-(thymin-1-yl)-2,5-dioxabi-
cyclo[2.2.1]heptane (10):
A solution of the a-nucleoside 9 (344 mg,
0.64 mmol) in ethanol (7 mL) and water (7 mL) was treated with 2m
aqueous NaOH solution (3.2 mL) and the mixture was stirred at 758C
for 3 days. Concentrated aqueous NaOH solution (1.0 mL) was then added
and the mixture was stirred for a further 2 days. It was then neutralised with
HCl and extracted with dichloromethane (3 Â 15 mL). The combined
organic fractions were washed with saturated aqueous NaHCO3 solution
(3 Â 20 mL), dried (Na2SO4) and concentrated in vacuo. The residue was
purified by column chromatography (CH2Cl2/MeOH, 96:4) to afford the
product as a white solid (107 mg; 46%) together with its intermediate 5'-O-
methylsulfonic ester (29 mg; 10%); 1H NMR (300 MHz, CDCl3, 258C,
TMS): d ˆ 7.46 (d, 3J(H,H) ˆ 1.0 Hz, 1H), 7.38 7.30 (m, 5H), 5.95 (s, 1H),
4.72 (d, 3J(H,H) ˆ 12.0 Hz, 1H), 4.64 (d, 3J(H,H) ˆ 12.0 Hz, 1H), 4.57 (s,
1H), 4.25 (s, 1H), 4.19 (d, 3J(H,H) ˆ 8.7 Hz, 1H), 3.95 3.90 (m, 3H), 1.95
(s, 3H).
(1R,3S,4R,7S)-7-(tert-Butyldimethylsilyl)oxy-1-(4,4'-dimethoxytrityl)oxy-
methyl-3-(thymin-1-yl)-2,5-dioxabicyclo[2.2.1]heptane (16): A solution of
14 (117 mg, 0.205 mmol) and imidazole (38 mg, 0.56 mmol) in anhydrous
pyridine (2.5 mL) was treated with TBDMSCl (45 mg, 0.30 mmol) and the
mixture was stirred at room temperature for 22 h. Another portion of
TBDMSCl (90 mg, 0.60 mmol) was then added and the mixture was stirred
at room temperature for 5 days. It was subsequently concentrated in vacuo,
the residue was taken up in dichloromethane (25 mL), and the resulting
solution was washed with saturated aqueous NaHCO3 solution (3 Â
25 mL). The organic phase was dried (MgSO4) and concentrated in vacuo.
The residue was purified by column chromatography (CH2Cl2/MeOH/
pyridine, 98:1.5:0.5) to give the product (110 mg, 79%) as a white solid,
which was used in the next step without further purification. 1H NMR
(300 MHz, CDCl3, 258C, TMS): d ˆ 8.62 (brs, 1H), 7.59 (d, 3J(H,H) ˆ
1.1 Hz, 1H), 7.47 7.23 (m, 9H), 6.84 (d, 3J(H,H) ˆ 8.3 Hz, 4H), 6.04 (s,
1H), 4.34 (s, 1H), 4.33 (s, 1H), 4.01 (m, 2H), 3.79 (s, 6H), 3.41 (d,
3J(H,H) ˆ 10.6 Hz, 1H), 3.33 (d, 3J(H,H) ˆ 10.6 Hz, 1H), 2.01 (d,
3J(H,H) ˆ 1.1 Hz, 3H), 0.75 (s, 9H), 0.05 (s, 3H), À0.04 (s, 3H); 13C NMR
(75 MHz, CDCl3, 258C, TMS): d ˆ 163.78, 158.56, 150.44, 144.46, 135.65,
135.53, 135.42, 129.95, 129.81, 127.96, 127.86, 126.88, 113.16, 109.46, 90.09,
87.20, 86.11, 78.76, 74.29, 73.42, 59.83, 55.21, 25.42, 17.73, 12.90, À4.72,
Alternative preparation of 2: A solution of 10 (105 mg, 0.291 mmol) in
ethanol (1.0 mL) and dioxane (0.5 mL) was stirred at room temperature
and 20% palladium hydroxide on carbon (65 mg) was added. The mixture
was degassed several times with argon and then placed under a hydrogen
atmosphere. After stirring for 3 h, the mixture was filtered through a layer
of Celite. The filter was rinsed with a mixture of dichloromethane and
MeOH (1:1; 10 mL) and the combined filtrates were concentrated in vacuo
to give the product (72 mg, 91%) as a white solid.
3-O-Benzyl-4-C-hydroxymethyl-1,2-O-isopropylidene-a-d-arabinofura-
nose (13): A solution of oxalyl chloride (0.39 mL, 4.51 mmol) in anhydrous
dichloromethane (11 mL) was stirred at À788C and a solution of DMSO
(0.64 mL, 9.0 mmol) in dichloromethane (9 mL) was slowly added. A
solution of the primary alcohol 12[32] (1.012 g, 3.61 mmol) in anhydrous
dichloromethane (9 mL) was then slowly added, followed by Et3N
(2.5 mL). The resulting mixture was stirred at À788C for 15 min. and at
room temperature for 2 h. The reaction was quenched by the addition of
water (20 mL) and after separation of the layers the aqueous phase was
extracted with dichloromethane (3 Â 20 mL). The combined organic phases
were washed with 1m aqueous HCl (20 mL), saturated aqueous NaHCO3
solution (20 mL) and water (20 mL), dried (MgSO4) and concentrated in
vacuo. The residue was redissolved in water (5 mL) and THF (5 mL) and
the solution was stirred at 08C. A 36% aqueous solution of formaldehyde
(1.3 mL, 15.5 mmol) was added, and then 1.0m aqueous NaOH solution
(5.2 mL) was slowly added. The mixture was stirred at room temperature
for 17 h and then treated with a saturated aqueous NaHCO3 solution
(25 mL). The resulting mixture was extracted with dichloromethane (3 Â
‡
À5.14; MS FAB: m/z (%): 687(4) [M‡H] .
(1R,3S,4R,7S)-7-(tert-Butyldimethylsilyl)oxy-1-(4,4'-dimethoxytrityl)oxy-
methyl-3-(5-methylcytosin-1-yl)-2,5-dioxabicyclo[2.2.1]heptane (17): 1,2,4-
Triazole (68 mg, 0.99 mmol) was added to a solution of 16 (97 mg,
0.141 mmol) in anhydrous pyridine (1.2 mL) and the mixture was stirred
at 08C. POCl3 (0.031 mL, 0.33 mmol) was added and the mixture was
stirred at room temperature for 15 h. It was subsequently diluted with
dichloromethane (8 mL), washed with saturated aqueous NaHCO3 solu-
720
¹ WILEY-VCH Verlag GmbH, 69451 Weinheim, Germany, 2002
0947-6539/02/0803-0720 $ 17.50+.50/0
Chem. Eur. J. 2002, 8, No. 3

Locked Nucleic Acids
712 722
tion (2 Â 5 mL), dried (MgSO4), and concentrated in vacuo. The residue
was purified by column chromatography (CH2Cl2/MeOH/pyridine,
99:0.5:0.5) to afford the triazole intermediate (72.3 mg, 69%). This
compound (113 mg, 0.153 mmol) was dissolved in dioxane (1.3 mL) and
treated with 20% aqueous ammonia (0.38 mL). The mixture was stirred at
room temperature for 22 h, then diluted with dichloromethane (5 mL) and
washed with water (5 mL). The organic phase was dried (MgSO4) and
concentrated in vacuo. The residue was purified by column chromatog-
raphy (CH2Cl2/MeOH/pyridine, 98.5:1:0.5) to give the product (68 mg;
64%) as a white solid, which was used in the next step without further
purification; 1H NMR (300 MHz, CDCl3, 258C, TMS): d ˆ 7.60 (s, 1H),
7.47 7.22 (m, 9H), 6.83 (d, 3J(H,H) ˆ 9.0 Hz, 4H), 6.14 (s, 1H), 4.44 (s,
1H), 4.35 (s, 1H), 4.02 (d, 3J(H,H) ˆ 8.4 Hz, 1H), 3.96 (d, 3J(H,H) ˆ 8.4 Hz,
1H), 3.79 (s, 6H), 3.42 (d, 3J(H,H) ˆ 10.7 Hz, 1H), 3.32 (d, 3J(H,H) ˆ
10.7 Hz, 1H), 2.02 (s, 3H), 0.74 (s, 9H), 0.04 (s, 3H), À0.05 (s, 3H);
13C NMR (75 MHz, CDCl3, 258C, TMS): d ˆ 165.65, 158.52, 156.02, 144.55,
138.60, 135.63, 135.51, 129.97, 127.99, 127.83, 126.82, 113.13, 100.26, 89.77,
87.87, 86.03, 78.77, 74.43, 73.35, 60.14, 55.16, 25.44, 17.72, 13.60, À4.72,
22 24, 26, 27, and 29 32 was performed on a 0.2mmol scale by using a-
thymidine 2-cyanoethyl phosphoramidite,[41] N-benzoyl-protected a-5-
methylcytosine 2-cyanoethyl phosphoramidite,[36] as well as compounds
15 and 20. The synthesis followed the regular protocol for the DNA
synthesizer and a universal CPG support (Biogenex) was employed.
However, for compounds 15 and 20, a prolonged coupling time of 10 min
was used. Coupling yields for all 2-cyanoethyl phosphoramidites were
>98%. The 5'-O-DMT-ON oligonucleotides were removed from the
universal solid support by treatment with 2% LiCl in concentrated
ammonia at 558C for 20 h, which also removed the protecting groups.
Subsequent purification using disposable reversed-phase cartridges, in-
cluding 5'-O detritylation, afforded the pure oligonucleotides. MALDI-MS
‡
[M‡H] gave the following results (found/calcd): 22 (4196.9/4196.8); 23
(4227.1/4224.8); 24 (4309.2/4308.8); 26 (2976.4/2980.0); 27 (3261.6/3260.1);
29 (2976.6/2976.0); 30 (3001.9/3004.0); 31 (3145.0/3144.1); 32 (3257.2/
3256.1). The purities of all the oligonucleotides were verified by capillary
gel electrophoretic experiments.
Melting experiments: UV melting experiments were carried out on a
Perkin Elmer Lambda 2 spectrometer. Samples were dissolved in a
medium salt buffer containing Na2HPO4 (10mm), NaCl (100mm) and
EDTA (0.1mm), pH 7.0 with 1.5mm concentrations of the two comple-
mentary sequences. The extinction coefficients were calculated assuming
the extinction coefficients for all thymine nucleotides to be identical, as
well as those for all 5-methyl cytosine nucleotides to be identical. The
increase in absorbance at 260 nm as a function of time was recorded while
the temperature was increased linearly from 8 to 808C at a rate of
0.58CminÀ1 by means of a Pertier temperature programmer. The melting
temperature was determined as the local maximum of the first derivatives
of the absorbance versus temperature curve. All melting curves were found
to be reversible.
‡
À5.18; MS FAB: m/z (%): 686 (3) [M‡H] .
(1R,3S,4R,7S)-3-(4-N-Benzoyl-5-methylcytosin-1-yl)-7-(tert-butyldimeth-
ylsilyl)oxy-1-(4,4'-dimethoxytrityl)oxymethyl-2,5-dioxabicyclo[2.2.1]hep-
tane (18): Compound 17 (32 mg, 0.047 mmol) was freed of solvents by co-
evaporation with anhydrous CH3CN and redissolved in anhydrous CH3CN
(0.2 mL). A solution of DMAP (5.9 mg, 0.048 mmol) in anhydrous CH3CN
(0.2 mL) and N-benzoyltetrazole[37] (16.5 mg, 0.095 mmol) were added to
the reaction mixture. The resulting mixture was stirred at 658C for 15 min
and then cooled to room temperature. Saturated aqueous NaHCO3
solution (10 mL) was added and the mixture was extracted with dichloro-
methane (2 Â 10 mL). The combined organic phases were dried (MgSO4)
and concentrated in vacuo. The residue was purified by column chroma-
tography (CH2Cl2/MeOH/pyridine, 98.5:1:0.5) to afford the product as a
CD experiments: CD experiments were carried out on a Jasco J710
spectropolarimeter at 258C using 1 cm cuvettes. The samples contained the
same buffer and the same concentrations as used in the melting experi-
ments.
1
white solid (37 mg, 100%); H NMR (300 MHz, CDCl3, 258C, TMS): d ˆ
8.34 (d, 3J(H,H) ˆ 6.8 Hz, 2H), 7.74 (d, 3J(H,H) ˆ 0.8 Hz, 1H), 7.47 7.26
(m, 12H), 6.85 (d, 3J(H,H) ˆ 8.6 Hz, 4H), 6.09 (s, 1H), 4.40 (s, 1H), 4.35 (s,
1H), 4.06 (d, 3J(H,H) ˆ 8.7 Hz, 1H), 4.00 (d, 3J(H,H) ˆ 8.7 Hz, 1H), 3.80 (s,
6H), 3.45 (d, 3J(H,H) ˆ 10.7 Hz, 1H), 3.35 (d, 3J(H,H) ˆ 10.7 Hz, 1H), 2.21
(d, 3J(H,H) ˆ 0.8 Hz, 3H), 0.76 (s, 9H), 0.06 (s, 3H), À0.03 (s, 3H); MS
‡
FAB : m/z (%): 790 (13) [M‡H] , HR MALDI: calcd for C45H51N3O8-
Si‡Na: 812.3338; found: 812.3338.
Acknowledgements
(1R,3S,4R,7S)-3-(4-N-Benzoyl-5-methylcytosin-1-yl)-1-(4,4'-dimethoxytrit-
yl)oxymethyl-7-hydroxy-2,5-dioxabicyclo[2.2.1]heptane (19): A solution of
compound 18 (58 mg, 0.073 mmol) in THF (0.7 mL) was treated with a
1.0 m solution of TBAF in THF (0.080 mL) and the mixture was stirred at
room temperature for 30 min. It was then concentrated in vacuo and the
residue was redissolved in dichloromethane (10 mL). This solution was
washed with saturated aqueous NaHCO3 solution (2 Â 10 mL), dried
(MgSO4) and concentrated in vacuo. The residue was purified by column
chromatography (CH2Cl2/MeOH/pyridine, 98:1.5:0.5) to afford the product
as a white solid (41 mg, 83%). 1H NMR (300 MHz, CDCl3, 258C, TMS):[38]
d ˆ 8.32 (d, 3J(H,H) ˆ 7.0 Hz, 2H), 7.70 (s, 1H), 7.47 7.26 (m, 12H), 6.86 (d,
3J(H,H) ˆ 8.9 Hz, 4H), 6.00 (s, 1H), 4.58 (s, 1H), 4.42 (s, 1H), 4.10 (d,
3J(H,H) ˆ 8.9 Hz, 1H), 4.03 (d, 3J(H,H) ˆ 8.9 Hz, 1H), 3.80 (s, 6H), 3.54
The Danish Natural Science Research Council and The Danish National
Research Foundation are thanked for financial support. Ms. Britta M.
Dahl, Department of Chemistry, University of Copenhagen, is thanked for
synthesising oligonucleotide sequences. Ms. Karen J˘rgensen, Department
of Chemistry, University of Copenhagen, is thanked for recording CD-
spectra. Dr. Henrik M. Pfundheller, Exiqon A/S, is thanked for recording
MALDI-MS spectra.
[1] a) E. Uhlmann, Curr. Opin. Drug Discovery Dev. 2000, 3, 203 213;
b) E. T. Kool, Chem. Rev. 1997, 97, 1473 1487.
[2] a) S. M. Freier, K.-H. Altmann, Nucleic Acids Res. 1997, 25, 4429
4443; b) O. Seitz, Angew. Chem. 1999, 38, 3674 3677; Angew. Chem.
Int. Ed. 1999, 38, 3466 3469.
‡
(m, 2H), 2.18 (s, 3H); MS FAB: m/z (%): 676 (11) [M‡H] , HR MALDI:
calcd for C39H37N3O8‡Na: 698.2473; found: 698.2476.
(1R,3S,4R,7S)-3-(4-N-Benzoyl-5-methylcytosin-1-yl)-1-(4,4'-dimethoxytrit-
yl)oxymethyl-7-(O-2-cyanoethyl-N,N-diisopropylphosphityl)oxy-2,5-dioxa-
bicyclo[2.2.1]heptane (20): Compound 19 (40.5 mg, 0.060 mmol) was
freed of solvents by co-evaporation with anhydrous CH3CN and redis-
solved in anhydrous dichloromethane (0.33 mL). The solution was stirred
at room temperature and diisopropyl ethylamine (0.048 mL) and
NC(CH2)2OP(Cl)N(iPr)2 (0.023 mL, 0.069 mmol) were added. The mixture
was stirred for 2 h, then diluted with EtOAc (5 mL), washed with saturated
aqueous NaHCO3 solution (2 Â 5 mL) and brine (2 Â 5 mL), dried
(MgSO4) and concentrated in vacuo. The residue was purified by column
chromatography (CH2Cl2/Et3N, 199:1), dissolved in toluene (0.5 mL) and
precipitated from petroleum ether (20 mL) at À308C to afford the product
as a white solid (31.5 mg; 60%). 31P NMR (121.5 MHz, CDCl3, 258C, 85%
H3PO4 external): d ˆ 150.9, 150.7.
[3] F. Morvan, B. Rayner, J.-L. Imbach, S. Thenet, J.-R. Bertrand, J.
Paoletti, C. Malvy, C. Paoletti, Nucleic Acids Res. 1987, 15, 3421 3437.
¬ ¡
[4] a) N. T. Thuong, U. Asseline, V. Roig, M. Takasugi, C. Helene, Proc.
Natl. Acad. Sci. USA 1987, 84, 5129 5133; b) C. Gagnor, J.-R.
Bertrand, S. Thenet, M. LemaÓtre, F. Morvan, B. Rayner, C. Malvy, B.
Lebleu, J.-L. Imbach, C. Paoletti, Nucleic Acids Res. 1987, 15, 10419
10436; c) C. Gagnor, B. Rayner, J.-P. Leonetti, J.-L. Imbach, B. Lebleu,
Nucleic Acids Res. 1989, 17, 5107 5114.
¬ ¡
[5] M. Durand, J. C. Maurizot, N. T. Thuong, C. Helene, Nucleic Acids
Res. 1988, 16, 5039 5053.
[6] C. Thibaudeau, A. Fˆldesi, J. Chattopadhyaya, Tetrahedron 1998, 54,
1867 1900.
[7] W. Saenger, in Principles of Nucleic Acid Structure, Springer, New
York, 1984.
[8] C. H. Gotfredsen, J. P. Jacobsen, J. Wengel, Bioorg. Med. Chem. 1996,
4, 1217 1225.
Synthesis of oligodeoxynucleotides: Oligonucleotide synthesis was carried
out by using
a Biosearch model 8750 automated DNA synthesizer
following the phosphoramidite approach. Synthesis of a-oligonucleotides
[9] F. Morvan, J. Zeidler, B. Rayner, Tetrahedron 1998, 54, 71 82.
Chem. Eur. J. 2002, 8, No. 3
¹ WILEY-VCH Verlag GmbH, 69451 Weinheim, Germany, 2002
0947-6539/02/0803-0721 $ 17.50+.50/0
721

P. Nielsen et al.
FULL PAPER
[10] a) S. Peyrottes, J.-J. Vasseur, J.-L. Imbach, B. Rayner, Tetrahedron
Lett. 1996, 37, 5869 5872; b) F. Debart, A. Meyer, J.-J. Vasseur, B.
Rayner, Nucleic Acids Res. 1998, 26, 4551 4556; c) A. Laurant, M.
Naval, F. Debart, J.-J. Vasseur, B. Rayner, Nucleic Acids Res. 1999, 27,
4151 4159.
[27] A. E. Hakansson, A. A. Koshkin, M. D. S˘rensen, J. Wengel, J. Org.
Chem. 2000, 65, 5161 5166.
[28] T. F. Tam, B. Fraser-Reid, Can. J. Chem. 1979, 57, 2818 2822.
[29] V. K. Rajwanshi, R. Kumar, M. Kofod-Hansen, J. Wengel, J. Chem.
Soc. Perkin Trans. 1 1999, 1407 1414.
[11] a) M. Bolli, P. Lubini, C. Leumann, Helv. Chim. Acta 1995, 78, 2077
2096; b) R. Zou, M. D. Matteucci, Bioorg. Med. Chem. Lett. 1998, 8,
3049 3052.
[30] O. Dahlman, P. J. Garegg, H. Mayer, S. Schramek, Acta Chem. Scand.
Ser. B 1986, 40, 15 20.
¬
[31] M. P. Vazquez-Tato, J. A. Seijas, G. W. J. Fleet, C. J. Mathews, P. R.
[12] K. Pongracz, S. M. Gryaznov, Nucleic Acids Res. 1998, 26, 1099 1106.
[13] a) P. Herdewijn, Liebigs Ann. 1996, 1337 1348; b) P. Herdewijn,
Biochim. Biophys. Acta 1999, 1489, 167 179.
Hemmings, D. Brown, Tetrahedron 1995, 51, 959 974.
[32] Z. Pakulski, A. Zamojski, Tetrahedron 1995, 51, 871 908.
[33] C.-H. Kim, V. E. Marquez, S. Broder, H. Mitsuya, J. S. Driscoll, J. Med.
Chem. 1987, 30, 862 866.
[34] a) T. R. Webb, M. D. Matteucci, J. Am. Chem. Soc. 1986, 108, 2764
2765; b) Y.-Z. Xu, Q. Zheng, P. F. Swann, J. Org. Chem. 1992, 57,
3839 3845; c) see ref. [15c].
[35] K.-H. Altmann, R. Kesselring, U. Pieles, Tetrahedron 1996, 52, 12699
12722.
[36] R. Kurf¸rst, V. Roig, M. Chassignol, U. Asseline, N. T. Thuong,
Tetrahedron 1993, 49, 6975 6990.
[37] B. Bhat, Y. S. Sanghvi, Tetrahedron Lett. 1997, 38, 8811 8814.
[38] The corresponding a-l-LNA 5-methylcytosine monomer has been
prepared very recently, and the NMR spectra of 19 and its enantiomer
proved to be identical; M. D. S˘rensen, Dept. of Chemistry, Univer-
sity of Copenhagen, and J. Wengel, Dept. of Chemistry, University of
Southern Denmark, personal pommunication.
[39] The preference of 4-N-benzoylated 5-methyl cytosine derivatives for
the alternative 4-imino tautomer has been reported: R. Busson, L.
Kerremans, A. Van Aerschot, M. Peeters, N. Blaton, P. Herdewijn,
Nucleosides Nucleotides 1999, 18, 1079 1082; however, a similar
behaviour of 18 20 can neither be confirmed nor excluded on the
basis of our NMR data.
[40] M. H. Caruthers, Acc. Chem. Res. 1991, 24, 278 284.
[41] M. Chassignol, N. T. Thuong, C.R. Acad. Sci. Paris Ser. II 1987, 305,
1527 1530.
[42] The complex between the a-l-LNA sequence (5'-aLTL)9T and com-
plementary l-RNA has also been examined, see ref. [19].
[43] We have recently employed the a-LNA thymine phosphoramidite 15
for the convenient incorporation into ODNs with polarity reversals:
N. K. Christensen, B. M. Dahl, P. Nielsen, Bioorg. Med. Chem. Lett.
2001, 11, 1765 1768.
[14] M. Meldgaard, J. Wengel, Perkin 1 2000, 3539 3554.
[15] a) S. K. Singh, P. Nielsen, A. A. Koshkin, J. Wengel, Chem. Commun.
1998, 455 456; b) A. A. Koshkin, S. K. Singh, P. Nielsen, V. K.
Rajwanshi, R. Kumar, M. Meldgaard, C. E. Olsen, J. Wengel,
Tetrahedron 1998, 54, 3607 3630; c) S. Obika, D. Nanbu, Y. Hari, J.
Andoh, K. Morio, T. Doi, T. Imanishi, Tetrahedron Lett. 1998, 39,
5401 5404; d) A. A. Koshkin, P. Nielsen, M. Meldgaard, V. K.
Rajwanshi, S. K. Singh, J. Wengel, J. Am. Chem. Soc. 1998, 120,
13252 13253; e) J. Wengel, Acc. Chem. Res. 1999, 32, 301 310.
[16] C. Wahlestedt, P. Salmi, L. Good, J. Kela, T. Johnsson, T. Hˆkfelt, C.
Broberger, F. Porreca, J. Lai, K. Ren, M. Ossipov, A. Koshkin, N.
Jakobsen, J. Skouv, H. Oerum, M. H. Jacobsen, J. Wengel, Proc. Natl.
Acad. Sci. USA 2000, 97, 5633 5638.
[17] For a review, see S. Gryaznov, Biochim. Biophys. Acta 1999, 1489,
131 140.
[18] a) V. K. Rajwanshi, A. E. Hakansson, B. M. Dahl, J. Wengel, Chem.
Commun. 1999, 1395 1396; b) V. K. Rajwanshi, A. E. Hakansson, R.
Kumar, J. Wengel, Chem. Commun. 1999, 2073 2074; c) A. E.
Hakansson, J. Wengel, Bioorg. Med. Chem. Lett. 2001, 11, 935 938.
[19] V. K. Rajwanshi, A. E. Hakansson, M. D. S˘rensen, S. Pitsch, S. K.
Singh, R. Kumar, P. Nielsen, J. Wengel, Angew. Chem. 2000, 39, 1722
1725; Angew. Chem. Int. Ed. 2000, 39, 1656 1659.
[20] Parts of the present work have been described in a short communi-
cation: P. Nielsen, J. K. Dalskov, Chem. Commun. 2000, 1179 1180.
[21] Parts of the present work have been presented at the XIV Interna-
tional Roundtable, Nucleosides, Nucleotides and Their Biological
Applications (San Francisco, USA) 2000, Abstract 108 and in a
corresponding proceeding paper; N. K. Christensen, J. K. Dalskov, P.
Nielsen, Nucleosides Nucleotides Nucleic Acids 2001, 20, 825 828.
[22] P. Nielsen, J. Wengel, Chem. Commun. 1998, 2645 2646; the two
anomeric bicyclic methyl furanosides as well as the corresponding
phenyl thioglycosides were obtained directly from 5 (Scheme 2).
[23] R. D. Youssefyeh, J. P. H. Verheyden, J. G. Moffatt, J. Org. Chem.
1979, 44, 1301 1309.
[44] G. Lancelot, J.-L. Guesnet, F. Vovelle, Biochemistry 1989, 28, 7871
7878.
[45] M. Bolli, J. C. Litten, R. Sch¸tz, C. J. Leumann, Chem. Biol. 1996, 3,
197 206.
[46] M. Petersen, C. B. Nielsen, K. E. Nielsen, G. A. Jensen, K. Bondens-
gaard, S. K. Singh, V. K. Rajwanshi, A. A. Koshkin, B. M. Dahl, J.
Wengel, J. P. Jacobsen, J. Mol. Recognit. 2000, 13, 44 53.
[24] T. Waga, T. Nishizaki, I. Miyakawa, H. Ohrui, H. Meguro, Biosci.
Biotech. Biochem. 1993, 57, 1433 1438.
[25] L. KvÒrn˘, R. Kumar, B. M. Dahl, C. E. Olsen, J. Wengel, J. Org.
Chem. 2000, 65, 5167 5176.
[26] A. A. Koshkin, J. Fensholdt, H. M. Pfundheller, C. Lomholt, J. Org.
Chem. 2001, 66, 8504 8512.
Received: August 13, 2001 [F3484]
722
¹ WILEY-VCH Verlag GmbH, 69451 Weinheim, Germany, 2002
0947-6539/02/0803-0722 $ 17.50+.50/0
Chem. Eur. J. 2002, 8, No. 3