Synthesis and pairing properties of α-tricyclo-DNA

Article abstract of DOI:10.1002/chem.200600597

We describe the synthesis of the phosphoramidite building blocks of α-tricyclo-DNA (α-tc-DNA) covering all four natural bases, starting from the already known corresponding α-tc-nucleosides. These building blocks were used for the preparation of three α-t

Full text of DOI:10.1002/chem.200600597

DOI: 10.1002/chem.200600597
Synthesis and Pairing Properties of a-Tricyclo-DNA
Simon P. Scheidegger and Christian J. Leumann*[a]
Abstract: We describe the synthesis of
the phosphoramidite building blocks of
a-tricyclo-DNA (a-tc-DNA) covering
all four natural bases, starting from the
already known corresponding a-tc-nu-
cleosides.These building blocks were
used for the preparation of three a-tc-
oligonucleotide 10-mers representing a
homopurine, a homopyrimidine, and a
mixed purine/pyrimidine base se-
quence.The base-pairing properties
with complementary parallel and anti-
parallel oriented DNA and RNA were
studied by UV-melting analysis and
CD spectroscopy.We found that a-tc-
DNA binds preferentially to parallel
nucleic acid complements through
Watson–Crick duplex formation, with a
preference for RNA over DNA.In
comparison with natural DNA, a-tc-
DNA shows equal to enhanced affinity
to RNA and also pairs to antiparallel
DNA or RNA complements, although
with much lower affinity.In the mixed-
base sequence these anatiparallel du-
plexes are of the reversed Watson–
Crick type, while in the homopurine/
homopyrimidine sequences Hoogsteen
and/or reversed Hoogsteen pairing is
observed.Antiparallel duplex forma-
tion of two a-tc-oligonucleotides was
also observed, although the thermal
stability of this duplex was surprisingly
low.The base-pairing properties of a-
tc-DNA are discussed in the context of
a-DNA, a-RNA, and a-LNA.
Keywords: DNA
DNA oligonucleotides
tricyclo-DNA
recognition
·
·
·
Introduction
Oligonucleotide-based gene silencing by antisense[1] and an-
tigene[2] agents or by small interfering RNAs and micro-
RNAs[3] is of prime interest[4] in human therapy and in
DNA-based diagnostics.Chemically modified oligonucleoti-
des have proven very useful in this context, providing solu-
tions to the relatively low biostabilities and target affinities
of natural oligonucleotides,[5] which are in many cases inca-
pable of eliciting biological responses.One of the first oligo-
nucleotide analogues investigated in detail was a-DNA,
made up of a-anomeric 2’-deoxyribonucleosides (Figure 1).
In early modeling[6] and structural[7,8] investigations it was
found that a-DNA forms parallel oriented Watson–Crick
duplexes with natural DNA, exhibiting thermal and thermo-
dynamic stabilities similar to those of the corresponding nat-
ural duplexes.[9] In its own series, a-DNA forms antiparallel,
Watson–Crick paired duplexes.[10] a-Oligodeoxynucleotides
are substantially more stable than natural oligonucleotides
Figure 1.Chemical structures of a-DNA and analogues.
towards degradation by nucleases[11] and, furthermore,
hybrid a-DNA/b-RNA duplexes are not substrates for the
RNA-cleaving enzyme RNase H.[12] In a similar vein, a-
anomeric oligoribonucleotides have also been synthesized
and shown to bind to the corresponding natural DNA in
parallel orientation, although with reduced thermal stabili-
ty.[13]
The search for novel carbohydrate-modified oligonucleoti-
des has given rise to a class of conformationally constrained
analogues that display high RNA and DNA affinity due to
reductions in entropic cost upon duplex formation.Members
of this class include, among others, the locked nucleic acids
(LNAs),[14,15] the hexitol nucleic acids (HNAs),[16] and tricy-
clo-DNA (tc-DNA).[17,18] These third-generation analogues
have in the past proven to be very useful antisense agents
with generally superior efficacy in biological assays in rela-
[a] Dipl-.Chem.S.P.Scheidegger, Prof.C.J.Leumann
Department of Chemistry & Biochemistry
University of Bern
Freiestrasse 3, 3012 Bern (Switzerland)
Fax : (+41)31-631-3422
E-mail: leumann@ioc.unibe.ch
8014
ꢀ 2006 Wiley-VCH Verlag GmbH & Co.KGaA, Weinheim
Chem. Eur. J. 2006, 12, 8014 – 8023

FULL PAPER
tion to their first- (phosphorothioates) or second-generation
(2’-alkyl-RNA) analogues.
a-Anomeric forms of many of these carbohydrate-modi-
fied oligonucleotide analogues also exist, but data on the
pairing properties of these are limited.Typically, though,
binding properties inferior to those of the corresponding b-
series are observed.In the case of a-d-LNA (Figure 1), how-
ever, it had previously been shown that pyrimidine sequen-
ces bind to complementary RNA in a parallel fashion and
with high affinity, whilst, interestingly, no binding to DNA
was observed.[19–21] In this article we report on the synthesis
and base-pairing properties of a-tc-DNA with complementa-
ry RNA, DNA, and with itself.
Results
Scheme 2.a) HF·pyridine, pyridine, RT, 2–5 h, 60–100%. b) DMT-OTf,
pyridine, RT, 1–6 h, 91–99%.c) ClP[(OCH 2CH2CN)N
CH3CN, RT, 2 h, 58–83%.
N
G
Synthesis of the phosphoramidite building blocks: For a-tc-
guanosine (Scheme 1), a more practical synthesis of the
tions finally yielded the phosphoramidites 16–19 in reasona-
ble yields.While desilylation and tritylation posed no prob-
lems, phosphitylation required a slight excess of the chloro-
phosphite reagent, which resulted after quenching with H2O
in an H-phosphonate by-product that was difficult to
remove by extraction or column chromatography.A viable
solution to this problem was found in quenching of the reac-
tion with glycerol, to a large extent allowing extraction of
excess reagent into the aqueous phase.
The synthesis of a-tc-oligonucleotides by the established
phosphoramidite synthesis method on a DNA synthesizer
(1 mmol scale) was then examined, but it soon became clear
that the synthesis was more difficult than expected and re-
quired considerable optimization.In initial attempts using
standard a-tc-amidite solutions (0.1 mm in CH3CN) we
found coupling efficiencies in the 50% range (trityl assay)
even with extended coupling times (12 min).Increasing the
amidite concentration twofold (0.2 mm) and using a 0.4m
solution of 5-(ethylthio)-1H-tetrazole (ETT) as activator fi-
nally provided coupling yields of 95% per step.A second
problem encountered was the oxidation step, which proved
to be incomplete with the use of standard iodine oxidation
procedures.This issue was solved by employment of a con-
centration of 20 mm and an oxidation time extended by 20 s.
The third problem arose from the universal solid support
used, which rapidly released a linker–oligonucleotide conju-
gate into solution during deprotection, whilst use of an ex-
tended deprotection time (60 h, 558C) was necessary to
remove the linker unit from the oligomer.As in the case of
b-tc-DNA synthesis, the 5’-end had to be phosphorylated to
prevent loss of the 5’-terminal nucleotide unit initiated by
thermal cyclopropanol ring-opening, followed by b-elimina-
tion of the 5’-keto nucleoside formed in situ.[25] This was
done by addition of one further nucleotide unit (n+1)
during synthesis, resulting after deprotection in the 5’-phos-
phorylated n-mer oligonucleotide.Using this procedure we
prepared the three a-tc-DNA decamers 20–22 listed in
Scheme 1.a) 2-Amino-6-chloropurine, BSA, TMSOTf, 55 8C, 4 h, 48%.
b) 3-Hydroxypropionitrile, NaH, THF, 08C, 3 h, then CH3OH/NH3 3:2,
558C, 16 h, 91%.c) HC
N
N
N,N-dimethylformamidino-protected nucleoside tc-Gdmf (4)
was developed, starting from the sugar building block 1 and
the base 2-amino-6-chloropurine in analogy to a procedure
in the LNA series.[22] This synthesis proved to be more effi-
cient than that for the previously reported isobutyryl-pro-
tected tc-Gibu [17]
The synthesis of the other a-tc-nucleoside
.
phosphoramidites with the bases ABz, T, and CBz started
from the known a-anomeric, base-protected tc-nucleosides
5–7 (Scheme 2).[17,23]
Nucleosidation of building block 1 with 2-amino-6-chloro-
purine in dichloroethane with TMSOTf as Lewis acid yield-
ed nucleoside 2 in 48% yield, together with its b-isomer
(40%).Subsequent conversion into tc-guanosine 3 with 2-
cyanoethanol/NaH, followed by installation of the dimethyl-
formamidine group by treatment with DMF dimethylacetal
yielded the base-protected nucleoside 4 in 85% yield over
two steps.
The 5’-protected a-tc-nucleosides 4–7 were then desilylat-
ed in HF/pyridine to give the dihydroxynucleosides 8–11.
Subsequent tritylation with DMT-OTf in pyridine[24] afford-
ed the 5’-protected nucleosides 12–15 in yields in the 90%
range, and phosphitylation under standard reaction condi-
Chem. Eur. J. 2006, 12, 8014 – 8023
ꢀ 2006 Wiley-VCH Verlag GmbH & Co.KGaA, Weinheim
www.chemeurj.org
8015

C.J.Leumann and S.P.Scheidegger
À
Table 1.Analysis by ESI mass spectrometry confirmed the
ture (Figure 2), which shows significant deviation with a sig-
moidal melting curve in the case of the mixture.
expected masses and in addition revealed the presence of
minor amounts (10–25%) of the 3’-phosphorylated oligonu-
À
cleotides, arising from partial P O bond hydrolysis between
the 3’-terminal phosphate and the universal support linker.
In our hands these mixtures were inseparable by standard
RP-HPLC.
Table 1.Sequences of a-tc-oligonucleotides 20–22 and their ESIÀ-MS
data.
Sequence
m/z
m/z
calcd
found
20
21
22
a-tc(pGAGAAGGAAA)
a-tc(pTTTCCTTCTC)
a-tc(pGCACTGTCAA)
3595.6
3380.4
3472.5
3594.9
3379.6
3472.1
Figure 2.UV-melting curves (260 nm) of duplexes indicated. For experi-
&
*
*
:
mental conditions see Table 2. : 20/pDNA; &: 20/pRNA; : 20/21;
20+21.
UV-melting experiments: With the three a-tc-oligonucleoti-
des 20–22—representing a pyrimidine, a purine, and a mixed
purine/pyrimidine sequence—to hand we investigated their
pairing properties with parallel and antiparallel complemen-
tary DNA and RNA by UV-melting.Table 2 shows a sum-
mary of the corresponding Tm data.Relevant melting curves
are reproduced in Figures 2–5.None of the a-tc-DNA single
strands 20–22 displayed any cooperative transition, ruling
out self-complex formation (data not shown).
In analogy to parallel a-DNA/DNA
conclude that parallel a-tc-DNA/DNA(RNA) duplexes are
also of the Watson–Crick type (Figure 5).This is supported
by the Tm of 20 with its parallel DNA complement at pH 6.0
(Table 2), which is identical (within the error limits) to that
at pH 7, thus excluding duplex-
(RNA) duplexes we
AHCTREUNG
es of the Hoogsteen or re-
Table 2.Summary of Tm data (260 nm). ct = 5.1 mm in 0.15m NaCl, 10 mm NaH2PO4, at pH 7.0.
Complementary DNA[a] Complementary RNA[a]
parallel antiparallel
versed-Hoogsteen type, which
would require cytosine proto-
nation.
parallel
antiparallel
20 a-tc(pGAGAAGGAAA)
21 a-tc(pTTTCCTTCTC)
22 a-tc(pGCACTGTCAA)
23 d(GAGAAGGAAA)
24 r(GAGAAGGAAA)
25 d(AACTGTCACG)
26 r(AACUGUCACG)
33.0 (32.0[b]
16.5
34.7
)
14.0 (31.5)
<5
17.8 (18.1)
25.9
41.5
44.4
44.8
32.8
45.1
18.3 (28.0)
14.7
27.0 (28.0)
12
47.8
42.5
The homopurine sequence
20 also undergoes duplex for-
mation with apDNA and
apRNA, although with dis-
tinctly lower Tm values than
seen with the parallel comple-
ments at neutral pH.In these
cases duplex formation is
strongly pH-dependent, with
<5
<10
23.2
19.1
n.m.[c]
n.m.[c]
n.m.[c]
n.m.[c]
40.1
52.2
[a] Tm values in parenthesis were determined at pH 6.0. [b] ct = 2.6 mm. [c] n.m. = not measured.
The homopurine oligonucleotide 20 forms stable duplexes
with parallel oriented DNA (pDNA) and RNA (pRNA),
with a distinctly higher Tm for the RNA complement
(Figure 2, Table 2).The hyperchromicities found are in the
same range, indicating similar extents of base stacking in the
corresponding duplexes.In both cases the Tm values are con-
siderably higher than those for the corresponding antiparal-
lel (ap) DNA or RNA duplexes with the sequence-identical
deoxyoligonucleotide 23 and slightly reduced in relation to
those with the sequence-identical oligoribonucleotide 24.In-
terestingly, the pure ap a-tc-DNA duplex 20/21 is signifi-
cantly less stable than any of the mixed backbone duplexes
and showed much less hyperchromicity, which is in accord
with reduced base stacking in the duplex.Evidence of
duplex formation in the latter case was provided by a com-
parison of a plot of the sum of the melting curves of the in-
dividual single strands 20 + 21 with that of the 20/21 mix-
higher Tm values at lower pH.Figure 3 illustrates this with
the example of the duplex of 20 with apDNA.The pH de-
pendence is a safe indication that protonated cytosines are
involved, pointing towards Hoogsteen and/or reversed-
Hoogsteen base pairing (Figure 5).Support for this interpre-
tation is also provided by the known fact that a-DNA can
bind as a third strand in parallel and antiparallel orienta-
tions to the purine strand of the underlying duplex through
reversed-Hoogsteen and Hoogsteen base-pair forma-
tion.[26,27]
The homo-pyrimidine a-tc-oligonucleotide 21 also strong-
ly favors binding to pDNA or pRNA, although with reduced
Tm values relative to the purine sequence.Again, the duplex
with pRNA is more stable than that with pDNA.For com-
parison, the oligodeoxyribonucleotide of identical sequence
to 21 binds to apRNA and apDNA with slightly enhanced
thermal stability, whilst the sequence-identical oligoribonu-
8016
www.chemeurj.org
ꢀ 2006 Wiley-VCH Verlag GmbH & Co.KGaA, Weinheim
Chem. Eur. J. 2006, 12, 8014 – 8023

a-Tricyclo-DNA
FULL PAPER
DNA/a-DNA duplexes exhibiting noncanonical G–A base
pairs.[30] In this mixed sequence context, a-tc-DNA/RNA
duplex stability is approximately the same as DNA/DNA
duplex stability, lying between DNA/RNA and RNA/RNA
duplex stability (Table 2).
Figure 3.UV-melting curves (260 nm) of duplexes indicated. For experi-
~
&
mental conditions see Table 2. : 20/apDNA, pH 7.0; : 20/apDNA,
pH 6.0.
cleotide shows weaker binding to DNA but stronger binding
to RNA.
To address a more general sequence context we prepared
the a-tc-oligonucleotide 22, containing all four bases, and in-
vestigated its pairing properties with parallel and antiparal-
lel DNA and RNA (Table 2, Figure 4).Pairing to comple-
mentary pRNA and pDNA is also preferred here, but some
promiscuity exists in that duplexes with the antiparallel ori-
ented complements were also observed.All duplexes show
similar hyperchromicities, indicating similar extents of base
pairing and base stacking.While Watson–Crick-paired struc-
tures are expected with parallel complements, the matter of
the pairing mode in the anti-
Figure 4.UV-melting curves (260 nm) of duplexes indicated. For experi-
&
*
:
*
:
mental conditions see Table 2.
22/apDNA;
:
22/pDNA;
22/
apRNA; 22/pRNA.
We also determined the thermodynamic data for duplex
formation by curve fitting to the experimentally derived
melting curves in those cases in which the lower baseline
was well defined.[31] These data are summarized in Table 3.
The order of the free enthalpies of duplex formation
(DG) in all cases parallels the order of the corresponding
parallel duplex was open; pH
Table 3.Thermodynamic data for duplex formation between a-tc-oligonucleotides and parallel complementa-
variation did not result in Tm
differences, ruling out Hoogs-
teen and reversed-Hoogsteen
structures, so the most likely
interaction of the bases is the
reversed-Watson–Crick mode
(Figure 5), as is also observed
in parallel b-DNA duplex-
es[28,29] and in antiparallel b-
ry DNA and RNA determined by curve fitting.
DH [kcalmolÀ1
]
DS [calKÀ1 molÀ1
]
DG258C [kcalmolÀ1
]
a-tc(pGAGAAGGAAA) b-d(CTCTTCCTTT)
a-tc(pGAGAAGGAAA) b-r(CUCUUCCUUU)
a-tc(pTTTCCTTCTC) b-r(AAAGGAAGAG)
a-tc(pGCACTGTCAA) b-d(GCATGTCAA)
a-tc(pGCACTGTCAA) b-r(GCAUGUCAA)
b-r(GAGAAGGAAA) b-d(TTTCCTTCTC)
b-r(GAGAAGGAAA) b-r(UUUCCUUCUC)
À60.5
À64.8
À63.0
À64.1
À64.9
À81.0
À89.7
À170.0
À177.3
À177.8
À180.2
À177.0
À229.0
À250.9
À9.8
À12.0
À10.0
À10.4
À12.1
À12.6
À14.8
Figure 5.Chemical formulae of the four possible bi- or tridentate A–T(U) and G–C base pairs.
Chem. Eur. J. 2006, 12, 8014 – 8023
ꢀ 2006 Wiley-VCH Verlag GmbH & Co.KGaA, Weinheim
www.chemeurj.org
8017

C.J.Leumann and S.P.Scheidegger
Tm.The enthalpies ( DH) are also very similar for all a-tc-du-
plexes but distinctly smaller (less negative) than in the anti-
parallel DNA/RNA and RNA/RNA control duplexes.This
is an indication that base stacking in the unnatural duplexes
is less efficient than in the natural series, an obstacle that is
compensated for by a smaller entropic penalty upon duplex
formation (DS), most probably due to the higher preorgani-
zation of the a-tc-DNA backbone.
detection of changes in the base pair geometry within the
stack in a comparative sense.
The CD traces of the mixed sequence 22 with its parallel
and antiparallel RNA complements are very similar, indicat-
ing close base-pair geometries in both cases.This is not un-
usual and is also observed for parallel reversed-Watson–
Crick and antiparallel Watson–Crick paired DNA.[28] In the
cases of pDNA or apDNA as the complement, the two posi-
tive ellipticity maxima at 220 and 275 nm are inverted in
their relative amplitude.As can be seen from the curves at
high temperatures, at which the CD reflects the sum of the
CD traces of the single strands, this inversion of maxima is
to a large extent due to the spectral properties of the single
strands and so is not a specific property of the ordered struc-
ture.From these observations we assume that Watson–Crick
and reversed-Watson–Crick base pairs are also formed in
the parallel and antiparallel duplexes in these cases.
The situation is somewhat different in the antiparallel du-
plexes of 20 (a-tc-homopurine sequence) with antiparallel
DNA and RNA.In the DNA hybrid a relatively small but
significant positive CD band is
CD spectroscopy: We recorded CD spectra of 20 as an ex-
ample of a homopurine sequence and of 22 as an example
of a mixed-base sequence with their respective parallel and
antiparallel DNA and RNA complements at different tem-
peratures.Relevant spectra are reproduced in Figure 6.The
temperature dependence of the CD spectra reflect the coop-
erative structural transitions from duplex to single strands
observed in the UV-melting curves in all cases.Unfortunate-
ly, CD spectra do not provide detailed structural informa-
tion on a duplex, but since CD signals arise from transition
dipoles of adjacent base pairs they are very well suited to
found around 290 nm in the duplex,
together with a stronger positive
band at 260 nm.Both fade away
upon duplex melting and result in
one large band around 270 nm in
the single strands.This behavior is
not observed in the a-tc-hybrid with
RNA as a complement.As found
before, the Tm values of both du-
plexes are pH-sensitive, indicating
Hoogsteen or reversed-Hoogsteen
base pairing.Given the differences
in the CD spectra it is likely that
one hybrid prefers the Hoogsteen
and the other the reversed-Hoogs-
teen pairing mode.Promiscuity in
Hoogsteen versus reversed-Hoogs-
teen base pairing with a-DNA has
been observed in DNA triplexes,
where it was shown that a-DNA
homopyrimidine third strands can
bind, in
a
sequence-dependent
manner, to the major groove of a
DNA duplex in either the parallel
or the antiparallel modes.[26,27]
Discussion
The synthesis of a-tc-DNA by phos-
phoramidite chemistry was more
difficult than had been anticipated
from previous experience with a-
DNA and b-tc-DNA synthesis.In
particular, the coupling step re-
quired higher a-tc-phosphoramidite
Figure 6.CD spectra of duplexes indicated at different pre- and post-melting temperatures.Experimental
conditions were as indicated in Table 2.
8018
www.chemeurj.org
ꢀ 2006 Wiley-VCH Verlag GmbH & Co.KGaA, Weinheim
Chem. Eur. J. 2006, 12, 8014 – 8023

a-Tricyclo-DNA
FULL PAPER
concentrations and a larger excess of monomeric unit per
coupling step than is usual in natural DNA or b-tc-DNA
synthesis.We ascribe this lack of reactivity during coupling
to increased steric constraints imposed by the nucleobases
at the activated PIII center.A further problem identified was
the reduced release rate (in relation to oligodeoxynucleoti-
des) from the universal solid support, which required pro-
longed treatment with conc.ammonia (60 h, 60 8C instead of
16 h, 608C) and produced not a uniformly free 3’-OH termi-
nus but also 10–25% 3’-phosphorylated termini arising from
the differences are much more pronounced in the case of a-
tc-DNA.Furthermore, the relative affinities to DNA and
RNA, especially in purine/pyrimidine mixed sequences, are
much higher in the case of a-tc-DNA, so the coherent pic-
ture of the superiority of natural nucleic acid recognition by
the tc-DNA backbone relative to its bc-DNA counterpart
also pertains in the a-anomeric series.
From NMR structures of a-DNA/DNA duplexes it is
known that the deoxyribose units in the a-strand are pre-
dominantly in the 3’-exo conformation, giving rise to a right-
handed duplex of the B conformational family.[35] X-ray
analysis of the a-tc-C nucleoside showed the furanose unit
to adopt a 3’-endo conformation.[17] An extrapolation of this
structural feature to the a-tc backbone would result in a
conformation similar to that of A-DNA.It is known that
the ribose units of a-ribonucleosides also exist predominant-
ly in N-type conformations (2’-exo).[36] a-RNA is known to
form complementary parallel duplexes with both DNA and
RNA complements, albeit with relatively low Tm values.[13]
From these considerations it becomes clear that differences
in ribose conformations in the three a-backbone systems
alone do not account for the observed differences in affinity.
It may well be that furanose conformation-dependent
changes in the glycosidic torsion angles play an important
role, which could also explain the rather peculiar situation
encountered with a-d-LNA, in which the ribose units are
locked in a 3’-endo conformation.In contrast with a-RNA,
a-d-LNA recognizes complementary RNA with high affinity,
but not DNA.[19] It may well be that the glycosidic torsion
angle is also modulated here both by the sugar pucker and
À À
alternative P O linker bond scission.It has thus become
clear that further improvement in a-tc-DNA synthesis
would be necessary if larger quantities and a wider variety
of a-tc-oligonucleotides were needed for, for example, struc-
tural or biological tests.
We compared affinities and preferred strand orientations
of a-tc-DNA with those of a-DNA.[32] In mixed-base se-
quence contexts, both a-anomeric oligonucleotide types
prefer parallel alignment over antiparallel in duplexes with
DNA or RNA.Thermal stabilities with RNA complements
are generally higher than those observed for DNA comple-
ments in the case of a-tc-DNA while no significant differen-
ces were found in the case of a-DNA.In a homopurine/ho-
mopyrimidine sequence context it has become clear that a-
tc-DNA and a-DNA show inverted preferences.Tc-purine
sequences form more stable duplexes with DNA and RNA
than tc-pyrimidine sequences, whilst the opposite is ob-
served for a-DNA.The relative thermal stabilities of a-
DNA and a-tc-DNA with b-DNA are similar, implying that
there is no major difference in thermal stabilities between
the two backbone systems.
A direct comparison with b-DNA (Table 2) reveals that
the affinity of a-tc-DNA to a parallel RNA complement is
roughly of the same order as that of a b-DNA to an antipar-
allel RNA complement.Pure RNA duplexes of the same
base sequence are generally slightly more stable than the
mixed backbone duplexes.In summary, this analysis shows
that, in mixed sequence contexts, a-tc-DNA performs simi-
larly to DNA and a-DNA in terms of target affinity, whilst
in homopurine sequence contexts a-tc-DNA outperforms
the other two backbone systems.
An interesting difference between a-DNA and a-tc-DNA
resides in the complementary base-pairing properties in
their own series.While a-DNA forms relatively stable anti-
parallel duplexes with itself,[10,32] this is not the case for a-tc-
DNA, as can be seen from the Tm values of 20/21.This is a
rather unusual feature, as for most backbone-modified oli-
gonucleotide analogues—such as LNA[33] and b-tc-
DNA[17]—the homo-backbone duplexes are more stable
than hetero-backbone duplexes.
We also compared the pairing behavior of a-tc-DNA with
that of a-bicyclo-(bc)-DNA, the structurally simpler version
of tricyclo-DNA lacking the cyclopropane ring.[34] Because
only a-bc-oligonucleotides with adenine and thymine as
bases were investigated, only a very general comparison is
possible.We find the same trend in that homo-backbone du-
plexes are less stable than hetero-backbone duplexes, but
À
by the C-4’ O-2’-bridge.This is further supported by the
UV-melting curves, which show weak hyperchromicities, and
by the CD spectra, which are totally different from those of
a-DNA and a-tc-DNA, pointing to duplexes with RNA of
an as yet unknown structure.
Conclusion
a-tc-DNA has been prepared and its pairing properties with
RNA and DNA have been analyzed.Like a-DNA, a-tc-
DNA prefers Watson–Crick duplex formation with parallel-
stranded RNA and DNA complements.The thermal stabili-
ties of such duplexes are equal to or slightly enhanced in re-
lation to corresponding DNA duplexes.Interestingly, anti-
parallel duplex formation within the a-tc-backbone system
is weak.This special feature, together with the expected
high biostability of a-tc-DNA, lends hope for successful ap-
plications as antisense agents or novel tools for biotechnolo-
gy in general, and as DNA double-strand invaders in partic-
ular.Work in this direction is currently in progress.
Experimental Section
General: Reactions were performed under argon in distilled, anhydrous
solvents.All chemicals were reagent grade from Fluka or Aldrich.
Chem. Eur. J. 2006, 12, 8014 – 8023
ꢀ 2006 Wiley-VCH Verlag GmbH & Co.KGaA, Weinheim
www.chemeurj.org
8019

C.J.Leumann and S.P.Scheidegger
1H NMR (300 MHz, 500 MHz) spectra were recorded on
a
Bruker
al of the solvent under reduced pressure the residue was purified by CC
AC 300 or a Bruker DRX 500 spectrometer; the chemical shifts d in ppm
were referenced to residual undeuterated solvent (CHCl3: 7.27,
CHD2OD: 3.35), J in Hz. 13C NMR (75 MHz) were recorded on a Bruker
AC 300 instrument; the chemical shifts d were referenced to residual un-
deuterated solvent (CHCl3: 77.00, CHD2OD: 49.3). 31P NMR spectra
(silica gel, CH2Cl2/CH3OH 8:1) to yield title compound 4 (882 mg,
1.86 mmol, 94%) as a yellow foam. Rf = 0.31 (CH2Cl2/CH3OH 7:1);
1H NMR (300 MHz, [D6]DMSO): d = 0.01, 0.05 (2”s, 6H; Me2Si), 0.76
(m, 10H; Me3C, H-C(8’)), 0.99 (m, 1H; H-C(8’)), 1.48 (m, 1H; H-C(6’)),
1.65 (d, J = 13.6 Hz, 1H; H-C(7’)), 2.24 (dd, J = 4.4 Hz, 1H; 13.6 Hz,
H-C(7’)), 2.64 (d, J = 7.1 Hz, 2H; H2C(2’)), 3.03, 3.13 (2”s, 6H; Me2N),
4.25 (s, 1H; H-C(4’)), 5.37 (s, 1H; OH), 6.19 (t, J = 7.1 Hz, 1H; H-
C(1’)), 8.00 (s, 1H; H-C(8)), 8.66 (s, 1H; N=CHNMe2), 11.32 ppm (br,
1H; NH); 13C NMR (75.5 MHz, [D6]DMSO): d = À4.0, À3.6, 16.7, 17.7,
23.8, 25.8, 34.8, 39.7, 40.8, 45.9, 65.7, 82.6, 84.8, 88.4, 119.9, 136.7, 149.7,
157.4, 157.8, 158.4 ppm; MS (ESI-MS+): m/z: calcd for: C22H34N6O4Si:
474.24; found: 475.27 [M+H]+.
(162 MHz) were recorded on
a Bruker DRX 400 spectrometer; the
chemical shift d was referenced to 85% H3PO4 as external standard.ESI-
MS mass spectra were recorded on a Fisons Instrument VG Platform.
For TLC, pre-coated plates (SIL-G UV254, Macherey–Nagel) were used,
these being viewed by UV and/or by dipping into a solution of Ce(SO4)2
G
(10.5 g), phosphomolybdic acid (21 g), H2SO4 (60 mL), and H2O
(900 mL).Flash chromatography (FC) was performed with silica gel 60
(230–400 mesh).
(3’S,5’R,6’R)-N2-(N,N-Dimethylformamidino)-9-(2’-deoxy-3’,5’-ethano-
5’,6’-methano-a-d-ribofuranosyl)guanine (8): HF·pyridine (70% HF,
3.1 mL) was added to a solution of 4 (1.11 g, 2.34 mmol) in pyridine
(25 mL).After the system had been kept for 2 h at RT, silica gel (2 g)
was added and the solvent was evaporated.Purification by CC (silica gel,
CH2Cl2/CH3OH 7:1) yielded fully deprotected 8 (506 mg, 1.40 mmol,
(3’S,5’R,6’R)-2-Amino-6-chloro-9-{5’-O-[(tert-butyl)-dimethylsilyl]-3’O-
trimethylsilyl-2’-deoxy-3’,5’-ethano-5’,6’-methano-a-d-ribofuranosyl}pur-
ine (2): Tricyclo-sugar 1 (1.5 g, 5 mmol) was dissolved in 1,2-dichloro-
ethane (19.5 mL), and 2-amino-6-chloropurine (1.3 g, 7.7 mmol) was
added, followed by N,O-bis(trimethylsilyl)acetamide (3 mL, 12.3 mmol).
The reaction mixture was heated to 408C until a clear solution appeared
(30 min) and was then cooled to 08C.Trimethylsilyl triflate (13. mL,
7.2 mmol) was added over 15 min and the reaction mixture was heated to
60%) as
a white foam. Rf =
0.15 (CH2Cl2/CH3OH 7:1); 1H NMR
(300 MHz, [D6]DMSO): d = 0.73 (m, 1H; H-C(8’)), 0.90 (t, J = 4.5 Hz,
1H; H-C(8’)), 1.36 (m, 1H; H-C(6’)), 1.65 (d, J = 13.6 Hz, 1H; H-C(7’)),
2.20 (dd, J = 4.3 Hz, 13.6 Hz, 1H; H-C(7’)), 2.65 (d, J = 7.0 Hz, 2H;
H2C(2’)), 3.03, 3.14 (2”s, 6H; Me2N), 4.21 (s, 1H; H-C(4’)), 5.30, 5.51
(2”s, 2H; 2”OH), 6.21 (t, J = 7.0 Hz, 1H; H-C(1’)), 8.03 (s, 1H; H-
C(8)), 8.68 (s, 1H; N=CHNMe2), 11.33 ppm (br, 1H; NH); 13C NMR
(75.5 MHz, [D6]DMSO): d = 16.45, 23.4, 40.8, 46.3, 48.8, 63.6, 82.6, 84.8,
88.5, 119.9, 124.1, 136.3, 137.0, 149.7, 149.8, 157.3, 157.8, 158.5 ppm;
HRMS (ESI-MS+): m/z: calcd for C16H21N6O4: 361.1546; found: 361.1624
[M+H]+.
558C for 4 h, allowed to cool to RT, and poured into sat.NaHCO
3
(40 mL).CHCl (24 mL) was added and the mixture was stirred for
3
30 min.The phases were separated and the aqueous phase was extracted
with CHCl3 (3”20 mL).The combined organic phases were washed with
sat.NaHCO /brine (1:1, 40 mL), dried (MgSO4), filtered, and concentrat-
3
ed to give a yellow oil.Purification by CC (hexane/EtOAc 2:1) gave title
compound 2 (1.06 g, 48%), together with the corresponding b-anomer
(890 mg, 40%).Data for the a-anomer: Rf = 0.29 (hexane/EtOAc 2:1);
1H NMR (300 MHz, CDCl3): d = 0.10, 0.13 (2”s, 15H; Me3Si, Me2Si),
0.85 (s, 9H; Me3C), 0.95, 1.01 (2”m, 2H; H2-C(8’)), 1.57 (m, 1H; H-
C(6’)), 1.82 (d, J = 13.8 Hz, 1H; H-C(7’)), 2.31 (dd, J = 4.9, 13.8 Hz,
1H; H-C(7’)), 2.62 (dd, J = 6.2, 13.2 Hz, 1H; H-C(2’)), 2.76 (dd, J =
6.7, 13.1 Hz, 1H; H-C(2’)), 4.36 (s, 1H; H-C(4’)), 5.20 (s, 2H; NH2), 6.25
(3’S,5’R,6’R)-1-(2’-Deoxy-3’,5’-ethano-5’,6’-methano-a-d-ribofuranosyl)th-
ymine (9): Compound 5 (320 mg, 0.81 mmol) was dissolved in pyridine
(10.2 mL) and HF·pyridine (70% HF, 0.8 mL, 2.4 mmol) was added drop-
wise. After 3 h a second aliquot of HF·pyridine (0.3 mL, 0.9 mmol) was
added, and after 2 h the reaction was quenched by the addition of silica
gel (0.5 g), the solvent was evaporated, and the residue was purified by
CC (silica, EtOAc/EtOH 9:1).Compound 9 (225 mg, 0.80 mmol, 99%)
(t, J
=
6.4 Hz, 1H; H-C(1’)), 8.04 ppm (s, 1H; H-C(8)); 13C NMR
(75.5 MHz, CDCl3): d = À3.80, À3.76, 1.9, 17.9, 18.3, 23.9, 25.7, 40.2,
46.9, 64.9, 84.4, 88.0, 90.8, 140.6, 151.3, 153.1, 159.0 ppm; HRMS (ESI-
MS+): m/z: calcd for C22H36ClN5O3Si2Na: 532.1943; found: 532.1927
[M+Na]+.
was readily isolated as a white foam. Rf
= 20 (hexane/iPrOH 3:1);
1H NMR (300 MHz, CH3OD): d = 0.93 (m, 1H; H-C(8’)), 1.00 (m, 1H;
H-C(8’)), 1.57 (m, 1H; H-C(6’)), 1.75 (d, J = 14.0 Hz, 1H; H-C(7’)), 1.96
(d, J = 1.1 Hz, 3H; Me-C(5)), 2.25 (m, 2H; H-C(2’), H-C(7’)), 2.57 (dd,
J = 6.1, 12.9 Hz, 1H; H-C(2’)), 4.32 (s, 1H; H-C(4’)), 6.22 (dd, J = 6.1,
7.8 Hz, 1H; H-C(1’)), 7.64 ppm (s, 1H; H-C(6)); 13C NMR (75.5 MHz,
CH3OD): d = 12.77, 17.72, 25.55, 25.65, 41.63, 65.40, 86.81, 86.87, 90.79,
112.15, 138.07, 152.60, 166.65 ppm; HRMS (ESI-MS+): m/z: calcd for
C13H17N2O5: 281.1138; found: 281.1137 [M+H]+.
(3’S,5’R,6’R)-9-{5’-O-[(tert-Butyl)-dimethylsilyl]-2’-deoxy-3’,5’-ethano-
5’,6’-methano-a-d-ribofuranosyl}guanine (3): A solution of 3-hydroxypro-
pionitrile (0.95 mL, 13.91 mmol) in anhydrous THF (20 mL) was cooled
to 08C.After careful addition of NaH (60%, 673 mg, 169.7 mmol) the
ice-bath was removed, the mixture was stirred for 30 min at RT, and com-
pound 2 (1.57 g, 3.08 mmol) dissolved in THF (20 mL) was then added at
08C.The reaction was complete after 3 h at RT.A 1:1 mixture of brine
(3’S,5’R,6’R)-N6-Benzoyl-9-(2’-deoxy-3’,5’-ethano-5’,6’-methano-a-d-ribo-
furanosyl)adenine (10): Compound 6 (817 mg, 1.41 mmol) was dissolved
in pyridine (15 mL) and HF·pyridine (70% HF, 1.11 mL, 3.5 mmol) was
added at RT.After 4 h the reaction was complete and was quenched by
the addition of silica gel (2 g).Evaporation of the solvent and purifica-
tion by CC (silica gel, CH2Cl2/CH3OH 9:1) yielded the deprotected nu-
and sat.NaHCO (60 mL) was then slowly added, and the organic phase
3
was separated and dried over MgSO4.Evaporation of the solvent yielded
a mixture of 3 and 3’-O-TMS-3 in a ratio of 1:5.The crude solid was dis-
solved in CH3OH (60 mL) and conc.NH (40 mL) and the mixture was
3
heated at 558C overnight.After evaporation of the solvent and CC
(silica gel, CH2Cl2/CH3OH 7:1) title compound 3 (1.17 g, 2.78 mmol,
91%) was isolated as a white foam. Rf = 0.21 (CH2Cl2/CH3OH 7:1);
1H NMR (300 MHz, [D6]DMSO): d = 0.00, 0.04 (2”s, 6H; Me2Si), 0.76
(m, 10H; Me3C, H-C(8’)), 0.96 (m, 1H; H-C(8’)), 1.48 (m, 1H; H-C(6’)),
1.67 (d, J = 13.6 Hz, 1H; H-C(7’)), 2.18 (dd, J = 6.0, 13.0 Hz, 1H; H-
C(7’)), 2.48–2.59 (m, 2H; H2C(2’)), 4.20 (s, 1H; H-C(4’)), 5.28 (s, 1H;
OH), 6.06 (t, J = 7.1 Hz, 1H; H-C(1’)), 6.44 (s, 2H; NH2), 7.89 (s, 1H;
H-C(8)), 10.63 ppm (br, 1H; NH); 13C NMR (75.5 MHz, [D6]DMSO): d
= À3.9, À3.6, 16.6, 17.7, 24.0, 25.8, 40.1, 46.4, 65.6, 82.4, 84.8, 88.3, 116.9,
135.4, 151.0, 153.8, 156.9 ppm; HRMS (ESI-MS+): m/z: calcd for
C19H29N5O4Si: 420.1989; found: 420.2089 [M+H]+.
cleoside 10 (450 mg, 1.15 mmol, 81%) as a white foam. Rf
= 0.30
(CH2Cl2/CH3OH 9:1); 1H NMR (300 MHz, [D6]DMSO): d = 0.81 (m,
1H; H-C(8’)), 0.90 (m, 1H; H-C(8’)), 1.44 (m, 1H; H-C(6’)), 1.66 (d, J =
13.7 Hz, 1H; H-C(7’)), 2.26 (dd, J = 4.5, 13.4 Hz, 1H; H-C(7’)), 2.71 (dd,
J = 6.7, 13.1 Hz, 1H; H-C(2’)), 2.82 (dd, J = 7.0, 13.2 Hz, 1H; H-C(2’)),
4.27 (s, 1H; H-C(4’)), 5.39, 5.58 (2”s, 2H; 2”OH), 6.46 (t, J = 6.8 Hz,
1H; H-C(1’)), 7.54 (m, 2H; arom. H), 7.64 (m, 1H; arom. H), 8.04 (d, J
=
7.1 Hz, 2H; arom. H), 8.69, 8.77 (2”s, 2H; H-C(2), H-C(8)),
11.21 ppm (brs, 1H; HN-C(6)); 13C NMR (75.5 MHz, [D6]DMSO): d =
16.7, 23.7, 45.9, 63.5, 83.5, 85.0, 86.9, 89.0, 126.0, 128.6, 132.6, 133.6, 143.3,
150.6, 151.8 ppm.
(3’S,5’R,6’R)-N2-{N,N-Dimethylformamidino)-9-(5’-O-[(tert-butyl)dime-
thylsilyl]-2’-deoxy-3’,5’-ethano-5’,6’-methano-a-d-ribofuranosyl}guanine
(4): Compound 3 (830 mg, 1.98 mmol) was dissolved in DMF (13 mL),
and N,N-dimethylformamide dimethylacetal (0.5 mL, 3.70 mmol) was
added.The mixture was heated at 55 8C and stirred for 3 h.After remov-
(3’S,5’R,6’R)-N4-Benzoyl-1-(2’-deoxy-3’,5’-ethano-5’,6’-methano-a-d-ribo-
furanosyl)cytosine (11): HF·Et3N (70% HF, 1.2 mL, 4.65 mmol, 3 equiv)
was added to a solution of nucleoside 7 (750 mg, 1.55 mmol) in pyridine
(19.5 mL) and the mixture was stirred for 3 h, after which another aliquot
of HF·Pyr (0.3 mL, 1.67 mmol) was added. After an additional 2 h the re-
8020
www.chemeurj.org
ꢀ 2006 Wiley-VCH Verlag GmbH & Co.KGaA, Weinheim
Chem. Eur. J. 2006, 12, 8014 – 8023

a-Tricyclo-DNA
FULL PAPER
action was quenched with silica gel (1 g), the solvent was evaporated, and
the residue was purified by CC (silica gel, EtOAc/EtOH 9:1) to yield
compound 11 as a white foam (100%). Rf = 0.62 (EtOAc/EtOH 9:1);
1H NMR (300 MHz, CD3OD): d = 0.80 (m, 1H; H-C(8’)), 0.95 (m, 1H;
H-C(8’)), 1.48 (m, 1H; H-C(6’)), 1.59 (d, J = 13.9 Hz, 1H; H-C(7’)), 2.10
(dd, J = 6.2, 13.4 Hz, 1H; H-C(7’)), 2.24 (dd, J = 5.1, 13.9 Hz, 1H; H-
C(2’)), 2.71 (dd, J = 6.2, 13.4 Hz, 1H; H-C(2’)), 4.32 (s, 1H; H-C(4’)),
6.09 (t, J = 6.2 Hz, 1H; H-C(1’)), 7.45 (m, 2H; arom. H), 7.55 (m, 2H;
arom. H), 7.99 (m, 2H; arom. H), 8.21 ppm (d, J = 7.5 Hz, 1H; H-C(6));
13C NMR (75 MHz, CD3OD): d = 18.4, 18.7, 41.5, 58.61, 87.3, 89.8, 91.7,
98.8, 129.5, 130.1, 134.4, 135.0, 146.3, 158.0, 165.1, 169.4 ppm; HRMS
(ESI-MS+): m/z: calcd for C19H20N3O5: 370.1400; found: 370.1402
[M+H]+.
(3’S,5’R,6’R)-N2-(N,N-Dimethylformamidino)-9-{2’-deoxy-5’-O-(4,4’-di-
methoxytrityl)-3’,5’-ethano-5’,6’-methano-a-d-ribofuranosyl}guanine (12):
DMT-OTf (1.2 g, 2.36 mmol) was added to a solution of 8 (460 mg,
1.27 mmol) in anhydrous pyridine (4.4 mL) and ethyldiisopropylamine
(2.2 mL, 13.2 mmol). After 1 h at RT the reaction mixture was diluted
133.7, 144.1, 146.2, 150.4, 158.4, 158.5 ppm; HRMS (ESI-MS+): m/z:
calcd for C41H38N5O6: 696.2787; found: 696.2822 [M+H+].
(3’S,5’R,6’R)-N4-Benzoyl-1-{2’-deoxy-5’-O-(4,4’-dimethoxytrityl)-3’,5’-
ethano-5’,6’-methano-a-d-ribofuranosyl}cytosine (15): DMT-OTf (1.01 g,
2.20 mmol, 2.1 equiv) was added to a solution of compound 11 (392 mg,
1.06 mmol) in pyridine (4.2 mL) and the mixture was stirred for 4 h at
RT.The reaction mixture was then diluted with EtOAc (10 mL) and
washed with sat.NaHCO (10 mL), the organic phase was dried over
3
MgSO4, the solvent was evaporated, and purification by CC (silica gel,
EtOAc/EtOH 9:1
+
1% Et3N) yielded compound 15 (706 mg,
1.05 mmol, 99%) as a yellow foam. Rf = 0.70 (EtOAc/EtOH 9:1 + 1%
Et3N); 1H NMR (300 MHz, CD3OD): d = 0.70 (t, J = 5.7 Hz, 1H; 1H-
C(8’)), 1.28 (t, J = 7.0 Hz, 1H; 1H-C(8’)), 1.63 (d, J = 13.7 Hz, 1H; H-
C(7’)), 1.82 (m, 1H; H-C(6’)), 1.98 (dd, J = 5.0, 14.1 Hz, 1H; H-C(7’)),
2.33 (dd, J = 5.1, 13.7 Hz, 1H; H-C(2’)), 2.75 (s, 1H; H-C4’)), 2.88 (m,
1H; H-C(2’)), 3.84, 3.85 (2”s, 6H; 2”OMe), 6.11 (t, J = 5.8 Hz, 1H; H-
C(1’)), 6.91 (m, 4H; arom. H), 7.15 (m, 1H; arom. H), 7.16 (m, 4H;
arom. H), 7.41 (m, 4H; arom. H), 7.50 (m, 2H; arom. H), 7.62 (m, 4H;
arom.H), 8.02 ppm (d, J = 7.2 Hz, 2H; arom. H); 13C NMR (75.5 MHz,
with EtOAc (5 mL) and washed with sat.NaHCO (5 mL).The organic
3
phase was then dried over MgSO4 and the solvent was evaporated under
reduced pressure.CC (silica gel, 5% CH 3OH in CH2Cl2) yielded 12
(816 mg, 1.23 mmol, 97%) as a yellow foam. Rf = 0.68 (CH2Cl2/CH3OH
9:1); 1H NMR (300 MHz, CDCl3): d = 0.49 (m, 1H; H-C(8’)), 1.23 (m,
1H; H-C(8’)), 1.83 (d, J = 43.7 Hz, 1H; H-C(7’)), 1.87 (m, 1H; H-C(6’)),
2.29 (dd, J = 4.9, 13.7 Hz, 1H; H-C(7’)), 2.50 (s, 1H; H-C(4’)), 2.59 (dd,
J = 2.3, 14.7 Hz, 1H; H-C(2’)), 2.78 (dd, J = 8.2, 14.6 Hz, 1H; H-C(2’)),
3.09, 3.13 (2”s, 6H; NMe2), 3.49, 3.71 (2”s, 6H; 2”MeO), 5.13 (s, 1H;
OH), 6.10 (dd, J = 2.3, 8.0 Hz, 1H; H-C(1’)), 6.63 (m, 4H; arom. H),
7.14 (m, 3H; arom. H), 7.31 (m, 4H; arom. H), 7.42 (m, 2H; arom. H),
7.71 (s, 1H; H-C(8)), 8.11 ppm (s, 1H; N=CHNMe2); HRMS (ESI-MS+):
m/z: calcd for C37H39N6O6: 663.2853; found: 663.2918 [M+H]+.
CD3OD): d = 7.02, 25.76, 56.06, 68.26, 87.70, 90.14, 114.03, 114.09,
128.81, 129.45, 130.14, 130.25, 130.75, 132.63, 132.80, 138.57, 148.29,
152.28, 160.71, 160.81, 217.98, 234.12 ppm; HRMS (ESI-MS+): m/z: calcd
for C40H38N3O7: 672.2720; found: 672.2709 [M+H]+.
(3’S,5’R,6’R)-N2-(N,N-Dimethylformamidino-9-{3’-O-[2-cyanoethoxy)-
A
ethano-5’,6’-methano-a-d-ribofuranosyl}guanine (16): Tritylated nucleo-
side 12 (389 mg, 0.59 mmol) was dissolved in anhydrous CH3CN (5.4 mL)
and ethyldiisopropylamine (0.4 mL, 2.32 mmol). Chloro-(2-cyanoethyl)-
diisopropylaminophosphine (0.27 mL, 1.29 mmol) was then added and
the solution was stirred for 2 h at RT.The reaction was quenched by the
addition of anhydrous glycerol (0.1 mL, 1.33 mmol), and stirring was con-
tinued for 15 min.The mixture was diluted with EtOAc (10 mL) and
(3’S,5’R,6’R)-1-{2’-Deoxy-5’-O-(4,4’-dimethoxytrityl)-3’,5’-ethano-5’,6’-
methano-a-d-ribofuranosyl}thymine
(13):
DMT-OTf[24]
(760 mg,
washed with a H2O/sat.NaHCO (10:1) solution (10 mL).After drying of
3
1.64 mmol, 2 equiv) was added to a solution of compound 9 (230 mg,
0.82 mmol) in pyridine (3.1 mL). The mixture was protected from light,
stirred for 6 h at RT, and then diluted with CH2Cl2 (20 mL) and washed
with sat.NaHCO .The organic phase was dried over MgSO and the sol-
vent was evaporated.The resulting oil was purified by CC (silica,
EtOAc/hexane 9:1 + 1% Et3N) to yield the tritylated compound 13
the organic phase over MgSO4 and evaporation of the solvent, the resi-
due was subjected to CC (silica gel, 3% CH3OH in CH2Cl2) and phos-
phoramidite 16 (290 mg, 0.34 mmol, 58%) was obtained as a white foam.
Rf = 0.59 (6% CH3OH in CH2Cl2); 1H NMR (300 MHz, CDCl3): d =
0.65, 0.80 (2”t, J = 5.6 Hz, 1H; H-C(8’)), 1.10 (m, 12H; (Me3-CH)2N),
1.19 (m, 1H; H-C(8’)), 1.26 (m, 1H; H-C(7’)), 1.87 (m, 1H; H-C(6’)),
2.25 (m, 1H; H-C(7’)), 2.51 (m, 3H; H-C(4’), NC-CH2), 2.61 (m, 1H; H-
C(2’)), 2.73 (m, 1H; H-C(2’)), 3.12, 3.13 (2”s, 3H; MeN), 3.21 (s, 3H;
MeN), 3.49 (m, 3H; CH2O, CHMe2), 3.77, 3.79 (2”s, 6H; 2”MeO), 6.35
(m, 1H; H-C(1’)), 6.71 (m, 4H; arom. H), 7.14 (m, 3H; arom. H), 7.28
(m, 4H; arom. H), 7.42 (m, 2H; arom. H), 8.62, 8.65 (2”s, 1H; H-C(8)),
3
4
(436 mg, 0.75 mmol, 91%) as a yellowish foam. Rf
= 0.18 (EtOAc/
1
hexane 9:1 + 1% Et3N); H NMR (300 MHz, CD3OD): d = 0.64 (t, J =
5.6 Hz, 1H; H-C(8’)), 1.21 (m, 1H; H-C(8’)), 1.62 (d, J = 14.0 Hz, 1H;
H-C(7’)), 1.80 (m, 1H; H-C(6’)), 1.94 (d, J = 1.1 Hz, 3H; Me-C(5)), 2.28
(m, 2H; H-C(2’), H-C(7’)), 2.61 (m, 2H; H-C(2’), H-C(4’)), 3.82, 3.83 (2”
s, 6H; 2”MeO), 6.19 (t, J
=
6.6 Hz, 1H; H-C(1’)), 6.86 (m, 4H;
9.06 ppm (br, 1H; N=CHNMe2); 31P NMR (161.9 MHz, CDCl3): d
=
arom. H), 7.23–7.33, 7.35–7.42 ppm (2”m, 7H; arom. H); 13C NMR
(75.5 MHz, CD3OD): d = 56.0, 68.5, 87.6, 90.4, 114.0, 128.7, 130.3, 132.6,
132.7, 137.7, 138.7, 150.4, 178.7 ppm; HRMS (ESI-MS+): m/z: calcd for
C34H34N2O7Na: 605.2261; found: 605.2263 [M+Na]+.
142.13, 144.33 ppm; HRMS (ESI-MS+): m/z: calcd for C46H56N8O7P:
863.4009; found: 863.4005 [M+H]+.
(3’S,5’R,6’S)-1-{3’-O-[(2-Cyanoethoxy)(diisopropylamino)phosphino]-2’-
deoxy-5’-O-(4,4’-dimethoxytrityl)-3’,5’-ethano-5’,6’-methano-a-d-ribofura-
nosyl}thymine (17): Compound 13 (400 mg, 0.69 mmol) and ethyldiiso-
propylamine (0.69 mL, 3.97 mmol) were dissolved in dry CH3CN
(3.5 mL), chloro-(2-cyanoethyl)diisopropylaminophosphine (0.30 mL,
1.84 mmol) was added dropwise, and the clear solution was stirred for 2 h
at RT.The reaction was quenched by the addition of glycerol (05. mL)
and stirring was continued for an additional hour.After dilution with
(3’S,5’R,6’R)-N6-Benzoyl-9-{2’-deoxy-5’-O-(4,4’-dimethoxytrityl)-3’,5’-
ethano-5’,6’-methano-a-d-ribofuranosyl}adenine (14): DMT-OTf (1.0 g,
2.28 mmol, 2 equiv) was carefully added to a solution of 10 (450 mg,
1.14 mmol) in pyridine (4.6 mL) and the mixture was stirred at RT for
3 h.After addition of EtOAc (20 mL) the yellow solution was washed
twice with sat.NaHCO
(30 mL), the organic phase was dried over
3
MgSO4 and evaporated, and the residue was then purified by CC (silica
gel, EtOAc + 5% Et3N).Tritylated compound 14 (730 mg, 1.05 mmol,
92%) was readily obtained as a yellow foam. Rf = 0.19 (EtOAc + 5%
Et3N); 1H NMR (300 MHz, CDCl3): d = 0.57 (t, J = 5.6 Hz, 1H; H-
C(8’)), 1.15 (m, 1H; H-C(8’)), 1.89 (m, 1H; H-C(6’)), 1.97 (d, J = 4.9 Hz,
1H; H-C(7’)), 2.06 (s, 1H; H-C(4’)), 2.31 (dd, J = 5.0, 13.8 Hz, 1H; H-
C(7’)), 2.86 (m, 2H; H2C(2’)), 3.66, 3.67 (2”s, 6H; 2”MeO), 5.79 (s, 1H;
OH), 6.17 (m, 1H; H-C(1’)), 6.52 (m, 4H; arom. H), 7.01 (m, 1H;
arom. H), 7.09 (m, 2H; arom. H), 7.27 (m, 4H; arom. H), 7.39 (d, J =
7.5 Hz, 2H; arom. H), 7.56 (m, 3H; arom. H), 8.03, 8.68 (2”s, 2H; H-
C(2), H-C(8)), 8.09 (d, J = 7.0 Hz, 2H; arom. H), 9.22 ppm (brs, 1H;
C(6)-NH); 13C NMR (75.5 MHz, CDCl3): d = 14.8, 16.4, 24.7, 34.4, 41.6,
55.2, 66.3, 86.8, 87.4, 89.6, 112.3, 112.3, 127.2, 127.9, 128.6, 130.7, 132.9,
EtOAc (30 mL) and washing with sat.NaHCO (30 mL), the organic
phase was dried over MgSO4 and the solvent was evaporated.CC (3 g
silica conditioned with EtOAc + 1% Et3N, elution with EtOAc) afford-
3
ed phosphoramidite 17 (432 mg, 0.55 mmol, 80%) as a white foam. Rf
=
0.39, 0.34 (2 diastereomers, hexane/EtOAc/Et3N 10:20:1); 1H NMR
(300 MHz, CDCl3): d = 0.64, (t, J = 5.6 Hz, 1H; H-C(8’)), 1.11 (m,
12H; 2”Me3CH), 1.18 (m, 2H; H-(7’), H-(8’)), 1.81 (m, 1H; H-C(6’)),
1.95 (d, J = 1.0 Hz, 3H; MeC(5)), 2.20 (m, 2H; H-C(6’), H-C(7’)), 2.5–
2.7 (2”m, 4H; H2C(2’), H-C(4’), NC-CH2), 3.41 (m, 2H; P-O-CH2-C),
3.5–3.7 (2”m, 2H; 2”CH(Me)2), 3.79, 3.80 (2”s, 6H; 2”MeO), 6.18 (m,
1H; H-C(1’)), 6.79 (m, 4H; arom. H), 6.91, 6.94 (2”s, 1H; H-C(6)), 7.23
(m, 3H; arom. H), 7.38 (m, 4H; arom. H), 7.49 (m, 2H; arom. H), 8.13
(s, 1H; H-N(3)) ppm; 31P NMR (161.9 MHz, CDCl3):
d
=
142.51,
Chem. Eur. J. 2006, 12, 8014 – 8023
ꢀ 2006 Wiley-VCH Verlag GmbH & Co.KGaA, Weinheim
www.chemeurj.org
8021

C.J.Leumann and S.P.Scheidegger
144.28 ppm; HRMS (ESI-MS+): m/z: calcd for C43H52N4O8P: 783.3489;
found: 783.3522 [M+H+].
(3’S,5’R,6’S)-N6-Benzoyl-9-{3’-O-[(2-cyanoethoxy)(diisopropylamino)-
phosphino]-2’-deoxy-5’-O-(4,4’-dimethoxytrityl)-3’,5’-ethano-5’,6’-metha-
no-a-d-ribofuranosyl}adenine (18): Ethyldiisopropylamine (0.93 mL,
MS data are given in Table 1.Natural oligonucleotides were synthesized
and purified by standard methodology, mass spectrometry also being
used here as quality control.
UV melting curves: UV melting curves were determined at 260 nm on a
Varian Cary 3E spectrophotometer fitted with a Peltier block with use of
Varian WinUV software.Complementary oligodeoxynucleotides were
mixed to 1:1 stoichiometry with a final total single-strand concentration
of 5 mm.A heating-cooling-heating cycle (0–90 8C) was applied with a
temperature gradient of 0.5–1.08CminÀ1.Each Tm values was defined as
the maximum of the first derivative of the melting curve with the aid of
the OriginTM v 5.0 software package. Thermodynamic data for duplex for-
mation were obtained as indicated in the text.
5.35 mmol)
and
chloro-(2-cyanoethyl)diisopropylaminophosphine
(0.49 mL, 3.01 mmol) were added to a stirred solution of compound 14
(540 mg, 0.78 mmol) in dry CH3CN (4.5 mL). After 2 h the reaction was
quenched with glycerol (0.5 mL) and after another hour the mixture was
diluted with EtOAc (20 mL) and washed with NaHCO3/sat.H 2O 1:10.
The organic phase was then dried over MgSO4 and evaporated.Purifica-
tion of the residue by CC (10 g silica, conditioned with EtOAc + 1%
Et3N, eluent: EtOAc) afforded amidite 18 (528 mg, 0.59 mmol, 83%) as a
CD spectra: CD spectra were recorded on a Jasco J-715 spectropolarime-
ter with a Jasco PFO-350S temperature controller.The temperature was
measured directly in the cell (path length 10 mm).
yellow foam. Rf
= 0.51 (EtOAc +
1% Et3N); 1H NMR (300 MHz,
CDCl3): d = 0.68, 0.82 (2”m, 1H; H-C(8’)), 1.00–1.15 (m, 13H; 2”
CMe2, H-C(8’)), 1.87 (m, 1H; H-C(7’)), 2.12 (m, 1H; H-C(6’)), 2.31 (m,
1H; H-C(7’)), 2.42, 2.47 (2”t, J = 6.2 Hz, 2H; NC-CH2-), 2.55 (s, 1H;
H-C(4’)), 2.77, 2.84 (2”m, 2H; H2(C2’)), 3.50 (m, 4H; 2”CHCH3,
G
-OCH2-), 3.78, 3.78, 3.79, 3.79 (4”s, 6H; 2”MeO), 6.45, 6.52 (2”m, 1H;
H-C(1’)), 6.72 (m, 4H; arom. H), 7.20 (m, 3H; arom. H), 7.33 (m, 4H;
arom. H), 7.44 (m, 2H; arom. H), 7.58 (m, 3H; arom. H), 7.99, 8.04, 8.77,
8.81 (4”s, 2H; H-C(2), H-C(8)), 8.05 (m, 2H; arom. H), 9.05 ppm (brs,
Acknowledgements
1H; NH); 31P NMR (161.9 MHz, CDCl3):
d = 143.22, 144.35 ppm;
HRMS (ESI-MS+): m/z: calcd for C50H55N7O7P: 896.3908; found:
896.3908 [M+H]+.
We thank the Swiss National Science Foundation (grant-No.: 200020-
100178) for financial support of this project.
(3’S,5’R,6’S)-N4-Benzoyl-1-{3’-O-[(2-cyanoethoxy)(diisopropylamino)-
phosphino]-2’-deoxy-5’-O-(4,4’-dimethoxytrityl)-3’,5’-ethano-5’,6’-metha-
no-a-d-ribofuranosyl}cytosine (19): Ethyldiisopropylamine (0.26 mL,
[1] T.Aboul-Fadl, Curr. Med. Chem. 2005, 12, 2193–2214.
[2] S.Buchini, C.J.Leumann,
726.
Curr. Opin. Chem. Biol. 2003, 7, 717–
1.52 mmol)
and
chloro-(2-cyanoethyl)diisopropylaminophosphine
(0.19 mL, 0.92 mmol, 3 equiv) were added drop by drop to a solution of
the tritylated nucleoside 15 (204 mg, 0.30 mmol) in dry CH3CN (1.9 mL).
After 2 h at RT the reaction mixture was quenched by the addition of
glycerol (0.3 mL) and allowed to stir for an additional hour. The mixture
[3] M.A.Valencia-Sanchez, J.Liu, G.J.Hannon, R.Parker,
2006, 20, 515–524.
[4] L.M.Jarvis, Chem. Eng.News 2006, April 17, 13–18.
[5] J.Kurreck, Eur. J. Biochem. 2003, 270, 1628–1644.
[6] U.SØquin, Experientia 1973, 29, 1059–1062.
[7] F. Morvan, B. Rayner, J.-L. Imbach, D.-K. Chang, J. W. Lown, Nu-
cleic Acids Res. 1986, 14, 5019–5035.
[8] F.Morvan, B.Rayner, J-.L.Imbach, M.Lee, J.A.Hartley, D-.K.
Chang, J.W.Lown, Nucleic Acids Res. 1987, 15, 7027–7044.
Genes Dev.
was then diluted with EtOAc (4 mL) and washed with sat.NaHCO
3
(6 mL), and the organic phase was dried over MgSO4 and evaporated.
Purification by CC (5 g silica, conditioned with EtOAc + 1% Et3N, elu-
tion with EtOAc) afforded amidite 19 (186 mg, 0.21 mmol, 71%) as a
white foam. Rf
= 0.53, 0.45 (2 diastereomers, EtOAc + 1% Et3N);
1H NMR (300 MHz, CDCl3): d = 0.60, 0.76 (2”m, 1H; H-C(8’)), 1.06
(m, 12H; 2”CMe2), 1.13 (m, 1H; H-C(8’)), 1.84 (m, 1H; H-C(6’)), 2.05–
2.35 (2”m, 2H; H2C(7’)), 2.38 (m, 1H; H-C(2’)), 2.56 (m, 2H; NC-CH2),
2.57, 2.77 (2”s, 1H; H-C(4’)), 2.86, 3.03 (2”m, 1H; H-C(2’)), 3.38 (m,
2H; OCH2), 3.55, 3.66 (2”m, 2H; 2”CHMe2), 3.81, 3.82 (2”s, 6H; 2”
MeO), 6.12 (m, 1H; H-C(1’)), 6.83 (m, 4H; arom. H), 7.29 (m, 2H;
arom. H), 7.37 (m, 4H; arom. H), 7.50 (m, 6H; arom. H), 7.61 (m, 2H;
arom. H), 7.90 (m, 2H; arom. H), 8.66 ppm (brs, 1H; NH-C(4));
31P NMR (161.9 MHz, CDCl3): d = 143.22, 144.35 ppm; HRMS (ESI-
MS+): m/z: calcd for C49H55N5O8P: 872.3822; found: 872.3788 [M+H]+.
[9] J.Paoletti, D.Bazile, F.Morvan, J-.L.Imbach, C.Paoletti,
Nucleic
Acids Res. 1989, 17, 2693–2704.
[10] F. Morvan, B. Rayner, J.-L. Imbach, D.-K. Chang, J. W. Lown, Nu-
cleic Acids Res. 1987, 15, 4241–4255.
[11] T.A.Bacon, F.Morvan, B.Rayner, J-.L.Imbach, E.Wickstrom,
Biochem. Biophys. Methods 1988, 10, 311–318.
[12] E.Bloch, M.Lavignon, J-.R.Bertrand, F.Pognan, F.Morvan, C.
J.
Malvy, B.Rayner, J-.L.Imbach, C.Paoletti,
[13] F.Debart, B.Rayner, G.Degols, J-.L.Imbach,
1992, 20, 1193–1200.
Gene 1988, 72, 349–360.
Nucleic Acids Res.
a-tc-Oligonucleotide synthesis: The syntheses of oligonucleotides 20–22
were performed by a modified phosphoramidite method on a 1.0 mmol
scale in the trityl-off mode on an Applied Biosystems Expedite 8909
DNA synthesizer.The following modifications of the standard DNA syn-
thesis cycle were made: i) the commonly used activator tetrazole was re-
placed by the more reactive 5-(ethylthio)-1H-tetrazole (0.4m in MeCN),
ii) 90 mL of 0. 2 mm amidite solution in CH3CN were used per coupling,
iii) the coupling time was increased to a total of 15 min, and iv) a 20 mm
I2 oxidizing solution was used and the oxidation time was extended to
40 s.All a-tc-oligonucleotides were synthesized on a universal solid sup-
port (universal support UL1, CTGen, Milpitas, CA).5 ’-End phosphoryla-
tion was effected by terminal coupling with an additional a-tc-C building
block, which was lost upon deprotection, leaving behind a 5’-phosphate
terminus.[25] After synthesis, the solid support was suspended in concen-
trated NH3 solution (ca.1 mL) and left for 60 h at 60–65 8C.The crude
oligonucleotides were purified by RP-HPLC on a Source 15RPC ST 4.6/
100 column (Amersham Pharmacia Biotech) with a CH3CN gradient in
Et3N/AcOH buffer (0.1m, pH 7.0, 608C).Purified oligonucleotides were
desalted over SEP-PAK C-18 cartridges (Waters).All modified oligonu-
cleotides were analyzed by ESIÀ mass spectrometry; the corresponding
[14] M.Petersen, J.Wengel, Trends Biotechnol. 2003, 21, 74–81.
[15] T.Imanishi, S.Obika, Chem. Commun. 2002, 1653–1659.
[16] C.Hendrix, H.Rosemeyer, I.Verheggen, F.Seela, A.Van Aerschot,
P.Herdewijn, Chem. Eur. J. 1997, 3, 110–120.
[17] D.Renneberg, C.J.Leumann,
6002.
J. Am. Chem. Soc. 2002, 124, 5993–
[18] C.J.Leumann, Bioorg. Med. Chem. 2002, 10, 841–854.
[19] P. Nielsen, N.K. Christensen, J.K. Dalskov,
Chem. Eur. J. 2002, 8,
712–722.
[20] N.K.Christensen, T.Bryld, M.D.Sorensen, K.Arar, J.Wengel, P.
Nielsen, Chem. Commun. 2004, 282–283.
[21] P.Nielsen, J.K.Dalskov, Chem. Commun. 2000, 1179–1180.
[22] C.Rosenbohm, D.S.Pedersen, M.Frieden, F.R.Jensen, S.Arent, S.
Larsen, T.Koch, Bioorg. Med. Chem. 2004, 12, 2385–2396.
[23] R.Steffens, C.Leumann, Helv. Chim. Acta 1997, 80, 2426–2439.
[24] M.Tarkçy, M.Bolli, C.Leumann,
744.
Helv. Chim. Acta 1994, 77, 716–
[25] R.Steffens, C.J.Leumann, J. Am. Chem. Soc. 1999, 121, 3249–3255.
[26] D.Praseuth, L.Perrouault, T.Le Doan, M.Chassignol, N.Thuong,
C.HØl›ne, Proc. Natl. Acad. Sci. USA 1988, 85, 1349–1353.
8022
www.chemeurj.org
ꢀ 2006 Wiley-VCH Verlag GmbH & Co.KGaA, Weinheim
Chem. Eur. J. 2006, 12, 8014 – 8023

a-Tricyclo-DNA
FULL PAPER
[27] J.S.Sun, C.Giovannangeli, J.C.FranÅois, R.Kurfurst, T.Montenay-
Garestier, U.Asseline, T.Saison-Behmoaras, NT..Thuong, C.
HØl›ne, Proc. Natl. Acad. Sci. USA 1991, 88, 6023–6027.
[32] F.Morvan, B.Rayner, J-.L.Imbach,
521–529.
[33] AA. .Koshkin, P.Nielsen, M.Meldgaard, VK. .Rajwanshi, SK. .
Anti-Cancer Drug Des. 1991, 6,
[28] A.K.Shchyolkina, O.F.Borisova, M.A.Livshits, G.E.Pozmogova,
Singh, J.Wengel, J. Am. Chem. Soc. 1998, 120, 13252–13253.
BK. .Chernov, R.Klement, TM. .Jovin,
10034–10044.
Biochemistry 2000, 39,
[34] M.Bolli, P.Lubini, C.Leumann,
Helv. Chim. Acta 1995, 78, 2077–
2096.
[29] J.H. van de Sande, N.B. Ramsing, M.W. Germann, W. Elhorst,
B.W.Kalisch, E.v. Kitzing, R.T.Pon, R.C.Clegg, T.M.Jovin,
ence 1988, 241, 551–557.
[35] G.Lancelot, J-.L.Guesnet, F.Vovelle,
7878.
[36] M.Sundaralingam, J. Am. Chem. Soc. 1971, 93, 6644–6647.
Biochemistry 1989, 28, 7871–
Sci-
[30] B.W.Kalisch, M.W.Germann, J.H.van de Sande,
FEBS Lett. 1998,
427, 301–304.
Received: April 28, 2006
Published online: August 17, 2006
[31] C.Epple, C.Leumann, Chem. Biol. 1998, 5, 209–216.
Chem. Eur. J. 2006, 12, 8014 – 8023
ꢀ 2006 Wiley-VCH Verlag GmbH & Co.KGaA, Weinheim
www.chemeurj.org
8023