pH-independent triple-helix formation with 6-oxocytidine as cytidine analogue

Article abstract of DOI:10.1002/1521-3765(20000703)6:13<2409::AID-CHEM2409>3.0.CO;2-H

The syntheses of six different phosphoramidite building blocks of 6-oxocytosine and 5-allyl-6-oxocytosine as analogues of N(3)-protonated cytosine are described. These compounds have been incorporated into oligonucleotides by standard solid-phase synthesi

Full text of DOI:10.1002/1521-3765(20000703)6:13<2409::AID-CHEM2409>3.0.CO;2-H

FULL PAPER
pH-Independent Triple-Helix Formation with 6-Oxocytidine
as Cytidine Analogue
[a]
Uwe Parsch and Joachim W. Engels*
Abstract: The syntheses of six different
phosphoramidite building blocks of
6-oxocytosine and 5-allyl-6-oxocytosine
as analogues of N(3)-protonated cyto-
sine are described. These compounds
have been incorporated into oligonu-
cleotides by standard solid-phase syn-
thesis. Hybridization of 15-mer Hoogs-
teen strands with target 21-mer duplexes
was investigated. Comparison of the
triplex-forming abilities of the different
building blocks revealed that: i) 5-allyl
substitution has a negative influence on
triplex stability; ii) a uniform backbone
of the Hoogsteen strand stabilizes tri-
cytidine analogues investigated in this
study. CD experiments provided further
evidence for the presence or absence of
triplex structures. In the course of these
temperature-dependent CD measure-
ments we were able to detect duplex
and triplex melting independent from
each other at selected wavelengths. This
methodology is especially interesting in
cases where UV melting curves show
only one transition owing to spectral
overlap.
plexes relative to mixed backbones;
iii) RNA strands with 6-oxocytidine or
5-allyl-6-oxocytidine do not form a triple
helix with the DNA target duplex,
probably due to backbone torsional
constraints; and (iv) a 15-mer DNA
sequence with three isolated 2'-deoxy-
6-oxocytidines has the highest Tm of all
Keywords: bioorganic chemistry
´
circular dichroism ´ DNA recogni-
tion ´ oligonucleotides ´ triplexes
against G ´ C base pairs inside a cell. Although the cytosine
residue appears to be protonated at pH values much higher
than its intrinsic pKa, its contribution to triplex stability at
physiological pH is minimal.[11±15] Different approaches have
been attempted to circumvent this difficulty. The first one was
the use of pyrimidine nucleosides with substituents in the
5-position of the base.[11, 16±21] Another one was the use of
5-methylcytosine tethered with spermine at N(4).[22±26] A third
attempt was the utilization of 5-substituted 2-aminopyridine
C-nucleosides as protonated cytidine equivalents.[27±30] These
molecules are more basic than cytosine and are protonated to
a higher extent in neutral and slightly basic conditions. All of
these experiments led to stabilization of the triple helices, but
none overcame the intrinsic pH dependence. One of the most
promising efforts involved the use of artificial nucleosides
able to form two hydrogen bonds to guanine without
protonation. Therefore, several groups suggested neutral
heterocyclic mimics of protonated cytosine that permit the
synthesis of oligomers having enhanced specific affinity for
duplex DNA under physiological conditions. Ono et al.
showed that deoxypyrimidine oligomers which contain pseu-
doisocytosine in place of cytosine form triplexes with an
oligodeoxyribonucleotide duplex at pH 7.0.[31, 32] Dervan and
coworkers reported that a base analogue, 1-(2-deoxy-b-d-
ribofuranosyl)-3-methyl-5-amino-1H-pyrazolo[4,3-d]pyrimi-
din-7-one, can interact with G ´ C base pairs in duplex DNA
over an extended pH range.[33, 34] The groups of Matteucci and
Introduction
In 1957, only four years after the discovery of the DNA double
helix by Watson and Crick,[1] Felsenfeld, Davies, and Rich
reported the detection of a triple helix based on polynucleo-
tides.[2] The possibility that oligonucleotides could be used to
regulate gene expression at the DNA level led to renewed
interest in the formation of triple-stranded nucleic acids three
decades later.[3±6] Since then, the triplex methodology has
become the most promising technology for the sequence-
specific recognition of DNA.[7, 8] Triplexes are formed from a
duplex with purine-consecutive sequences and oligonucleo-
tides that bind as a third strand in the major groove. In the
parallel binding motif, a thymine in the third strand is used to
form a Hoogsteen pair with adenine at neutral pH (TÂ A ´ T
triad) (Scheme 1). There is, however, no natural base that can
bind in a Hoogsteen fashion to guanine at pH 7. Generally,
‡
protonated cytosine is used in this capacity (C Â G ´ C triad;
Scheme 1).[9, 10] However, as the pKa of protonated cytosine is
ꢀ4.3 and the intracellular pH is ꢀ7.4, protonated cytosine has
limited applicability as a component in a third strand targeted
[a] Prof. Dr. J. W. Engels, Dipl.-Chem. U. Parsch
Johann Wolfgang Goethe-Universität
Institut für Organische Chemie
Marie-Curie-Str. 11, D-60439 Frankfurt am Main (Germany)
Fax : (‡49)69-7982-9148
E-mail: engels@ews1.org.chemie.uni-frankfurt.de
Chem. Eur. J. 2000, 6, No. 13
ꢀ WILEY-VCH Verlag GmbH, D-69451 Weinheim, 2000
0947-6539/00/0613-2409 $ 17.50+.50/0
2409

U. Parsch, J. W. Engels
FULL PAPER
Miller used oligonucleotides with N6-methyl-8-oxo-2'-deoxy-
adenosine and 8-oxo-2'-deoxyadenosine, respectively. UV
melting experiments showed that these oligonucleotides had
slightly lower Tm values than did those containing 2'-
deoxycytidine.[35±37] A pyrazine base with a donor± donor±
acceptor motif was incorporated into an oligonucleotide by
von Krosigk and Benner. The Tm value increased by 58C
compared with that of a 2'-deoxycytidine-containing oligomer
at pH 7.3 with 1m NaCl in the buffer.[38] The use of 5-methyl-6-
oxocytosine was reported by two groups.[39±42] The 6-oxocyti-
dine molecule seems to be the most promising candidate so
far for stable and pH-independent triple-helix formation in
this field. The other candidates mentioned are either purine-
like structures which distort the oligonucleotide backbone or
are C-nucleosides, which possibly form weaker base pairs than
the corresponding N-nucleosides.[43]
Here, we report the design and synthesis of nonnatural
analogues that bind G ´ C base pairs without protonation in a
pyrimidine-motif triple-helical complex. Our design criteria
for a nucleoside analogue included a pyrimidine-like ring
system, to minimize anomalous conformational changes in the
backbone, and a bidentate hydrogen-bond donor, to interact
effectively with the O(6)-oxygen and N(7)-nitrogen of the
target guanine. One molecule that meets these criteria is the
nucleoside 6-oxocytidine (Scheme 1). In this work we com-
pare the effects of different substituents in the backbone (2'-
position) and the 5-position of the heterocycle on the stability
of triple helices. As physicochemical methods, UV and CD
spectroscopy were used.
‡
Scheme 1. T-A-T, C -G-C, and O-G-C base triads; O ˆ 6-oxocytosine,
R' ˆ H or substituent, R ˆ sugar.
Abstract in German: Die Synthesen von sechs verschiedenen
Phosphoramiditen mit 6-Oxocytosin bzw. 5-Allyl-6-oxocytosin
als Analoga von protoniertem Cytosin werden beschrieben.
Diese Bausteine konnten durch Festphasensynthese in Oligo-
nucleotide eingebaut werden. Die Hybridisierungseigenschaf-
ten von 15mer Hoogsteen-Strängen mit 21mer Doppelhelices
wurden untersucht. Der Vergleich der Triplex-bildenden Ei-
genschaften der verschiedenen Oligonucleotide erbrachte fol-
gende Ergebnisse: i) 5-Allyl Substitution hat einen negativen
Einfluss auf die Tripelhelix-Stabilität. ii) Ein einheitliches
Rückgrat des Hoogsteen-Stranges stabilisiert Tripelhelices
verglichen mit einem gemischten Rückgrat. iii) RNA-Stränge,
die 6-Oxocytidin oder 5-Allyl-6-oxocytidin enthalten, bilden
keine Tripelhelices mit der DNA Doppelhelix. Dies ist wahr-
scheinlich in einer ungewöhnlichen Zuckerkonformation die-
ser Derivate begründet, welche zu einer Verzerrung des
Oligonucleotid Rückgrats führt. iv) Eine 15mer DNA-Sequenz
mit drei isolierten 2'-Desoxy-6-oxocytidinen zeigt den höchsten
Tm-Wert aller untersuchten Cytidin-Analoga dieser Studie. CD-
Experimente gaben weitere Hinweise auf die Existenzbzw.
Nicht-Existenztripelhelicaler Assoziate. Während dieser Un-
tersuchungen gelang es, eine Methode zur selektiven Erken-
nung des Triplex- bzw. Duplex-Schmelzprozesses zu etablie-
ren. Diese Technik ist besonders in Fällen interessant, in denen
UV-Schmelzkurven wegen Überlappung nur einen Übergang
zeigen.
Results and Discussion
Chemical synthesis of the phosphoramidites: The synthesis of
6-oxocytidine (8) was performed using the glycosylation
procedure of Vorbrüggen (Scheme 2).[44, 45] Refluxing 6-oxo-
cytosine (4) with hexamethyldisilazane (HMDS) and subse-
quent reaction of the persilylated base with 1,2,3,5-tetra-O-
acetyl-b-d-ribofuranose (3) in the presence of the Lewis acid
trimethylsilyl trifluoromethanesulfonate in 1,2-dichloro-
ethane (DCE) afforded the desired 2',3',5'-tri-O-acetyl-6-
oxocytidine (6) in 83% yield.
5-Allyl-6-oxocytosine (5) was synthesized from ethyl cya-
noacetate (1) with allyl chloride in sodium ethanolate, and
reaction of the product with urea in sodium methanolate
under reflux conditions. Silylation of 5 with HMDS and
subsequent glycosylation according to the procedure descri-
bed above furnished 2',3',5'-tri-O-acetyl-5-allyl-6-oxocytidine
(7) in only 46% yield. Here we were able to isolate four
different products: 7, its N(3)-regioisomer, and two other
compounds of unknown structure.
Deprotection of the acetylated nucleosides 6 and 7 in
methanolic ammonia gave 6-oxocytidine (8) and 5-allyl-6-
oxocytidine (9) in 93% and 96% yield, respectively. The
assignment of 8 and 9 as the expected b-isomers relied on the
existence of a ROESY cross-peak between H1' and H4', while
2410
ꢀ WILEY-VCH Verlag GmbH, D-69451 Weinheim, 2000
0947-6539/00/0613-2410 $ 17.50+.50/0
Chem. Eur. J. 2000, 6, No. 13

Triple-Helix Formation
2409±2424
no cross-peak was observed
between H1' and H3'. The re-
giochemistry (N(1), N(3)) was
assigned by a comparison of the
chemical shifts of the amino
proton resonances in the
1H NMR spectra.[46]
To improve the low overall
yield of 9, we tested a palladi-
um-catalyzed coupling reaction
between 8 and allylbromide in
DMF, adding triethylamine and
tetrakistriphenylphosphinepal-
ladium(0) as catalyst. The palla-
dium-catalyzed variant gave an
overall yield of 29% (starting
from 6-oxocytosine (4), three
reaction steps) compared with
only 8% yield from the route
described above (starting from
ethyl cyanoacetate (1), four re-
action steps).
Scheme 2. Synthesis of 6-oxocytidine (8) and 5-allyl-6-oxocytidine (9): reagents and conditions: i) 1) NaOEt,
2) allyl chloride, reflux, 3 h; ii) 1) NaOMe, 2) urea, reflux, 6 h; iii) 1) HMDS, (NH4)2SO4, reflux, 24 h, 2) DCE,
TMSOSO2CF3, 1,2,3,5-tetra-O-acetyl-b-d-ribofuranose (3), RT, 24 h; iv) NH3, MeOH, RT, 24 h.
The exocyclic amino func-
tions of the unprotected nucleo-
sides 8 and 9 were protected
with the dimethylformamidine
protecting group (Scheme 3).
The products were converted
with 4,4'-dimethoxytriphenyl-
methyl chloride (DMTrCl) in
dry pyridine. In the next step we
had to find a practical synthesis
for the 2'-O-methylation of the
nucleosides 12 and 13. The
problems encountered in the
use of methyl iodide/silver ox-
ide or trimethylsilyl diazome-
thane as a 2m solution in n-
hexane have been described by
Berressem and Engels.[42] We
chose freshly prepared diazo-
methane/SnCl2 as our methyla-
tion system. The monomethyla-
tion of the cis-glycol systems of
12 and 13 was performed in
DMF according to a method
first described by Robins and
coworkers.[47] The regioisomer
assignment was easily achieved
by 1H,1H-COSY NMR spectro-
scopy. The last step in the
syntheses was the phosphityla-
tion of the 3'-OH groups. The
6-oxocytidine derivatives 14
and 15 were each dissolved in
dry acetonitrile, and N,N-diiso-
propylethylamine
and the chlorophosphoramidite
18 were added. This afforded
(DIPEA)
Scheme 3. Synthesis of the 2'-O-methylphosphoramidites 19 and 20: reagents and conditions:
i) (MeO)2CHNMe2, MeOH, reflux, 10 min; ii) DMTrCl, pyridine, 10: RT, 3.5 h, 11: RT, 6 h; iii) CH2N2, SnCl2,
DMF, 4 h, 08C ! RT; iv) iPrNPCl(OCH2CH2CN) (18), DIPEA, CH3N, 14: RT, 0.5 h, 15: RT, 2 h.
Chem. Eur. J. 2000, 6, No. 13
ꢀ WILEY-VCH Verlag GmbH, D-69451 Weinheim, 2000
0947-6539/00/0613-2411 $ 17.50+.50/0
2411

U. Parsch, J. W. Engels
FULL PAPER
the desired phosphoramidite
building blocks 19 and 20.
To obtain the 2'-deoxyphos-
phoramidites, we started from
4-N-(dimethylformamidine)-6-
oxocytidine 10 and 5-allyl-4-N-
(dimethylformamidine)-6-oxo-
cytidine 11, using the Barton
deoxygenation
reaction
(Scheme 4).[48] The use of 2'-
deoxy sugars in the Vorbrüggen
glycosylation reaction for the
direct synthesis of 2'-deoxy nu-
cleosides has certain disadvan-
tages, mainly the lack of stereo-
specificity of this reaction.
The 6-oxocytidine deriva-
tives 10 and 11 were protected
simultaneously at the 5'- and 3'-
positions by the Markiewicz
protecting group in 93% and
88% yield, respectively.[49] The
products 21 and 22 were dis-
solved in dry acetonitrile, and
N,N-dimethylaminopyridine
(DMAP) was added in each
case. The addition of phen-
oxythiocarbonyl chloride led
to the isolation of 23 (58%)
and 24 (52%). These two com-
pounds were deoxygenated in
freshly distilled toluene with
a,a'-azoisobutyronitrile
(AIBN) and tributyltin hydride
at 758C. The products 25 and 26
were obtained in 65% and
58% yield, respectively. After
cleavage of the Markiewicz
protecting group with tetrabu-
tylammonium fluoride (TBAF)
in THF, the 5'-OH groups of 25
and 26 were protected with
Scheme 4. Synthesis of the DNA phosphoramidites 31 and 32: reagents and conditions: i) TiPDSCl2, pyridine,
10: 0.5 h at 08C, 1.5 h at RT, 11: 0.5 h at 08C, 2.5 h at RT; ii) phenoxythiocarbonyl chloride, DMAP, CH3CN, 21:
2 h at RT, 22: 4 h at RT; iii) nBuSnH, AIBN, toluene, 758C, 23: 2 h, 24: 1.5 h; iv) TBAF, THF, RT, 1 h; v) DMTrCl,
pyridine, 27: 3 h at À408C, 5 h at RT, 28: 1 h at 08C, 8 h at RT; vi) iPrNPCl(OCH2CH2CN) (18), DIPEA, CH2Cl2,
29: 0.5 h at RT, 30: 4 h at RT.
DMTrCl in dry pyridine in high yields. The last step again
was the phosphitylation to obtain the phosphoramidites 31
and 32.
technique under standard conditions.[55±59] The different
6-oxocytidine derivatives could be incorporated into the
oligonucleotide strands with coupling efficiencies comparable
to those of common nucleoside phosphoramidites. The
oligonucleotides were purified by RP-HPLC (trityl on) for
all DNA strands and by anion-exchange HPLC (trityl off) for
all RNA strands. Finally, all sequences were characterized by
MALDI mass spectrometry; for details, see experimental
section.
Starting materials for the syntheses of the two RNA
phosphoramidites were 12 and 13 (Scheme 5). Each of these
compounds was dissolved in THF/pyridine 1:1, and silver
nitrate and a 1m solution of tert-butyldimethylsilyl chloride
(TBDMSCl) in THF were added.[50±53] Good yields of the
desired products were obtained with a high preference for the
2'-protected species. The phosphitylation of the ribonucleo-
sides was accomplished according to the method of Scarringe
et al.[54] The phosphoramidites 37 and 38 were obtained in
72% and 71% yield, respectively.
UV thermal denaturation studies and CD measurements:
Thermal denaturation experiments were performed in
50 mm sodium cacodylate buffer at pH 5.0, 6.0, 7.0, and 7.4
containing 20 mm magnesium chloride and 100 mm sodium
chloride at oligonucleotide concentrations of 2 mm for each
strand. The Tm values were determined by a versus T plots,
where a is the fraction of molecules base-paired. Details of
Oligonucleotide synthesis, purification, and characterization:
The 15-mer Hoogsteen strands as well as the 21-mer duplex
strands were prepared by the phosphoramidite solid-phase
2412
ꢀ WILEY-VCH Verlag GmbH, D-69451 Weinheim, 2000
0947-6539/00/0613-2412 $ 17.50+.50/0
Chem. Eur. J. 2000, 6, No. 13

Triple-Helix Formation
2409±2424
independent triple-helix forma-
tion. A comparison of the two
triplexes shows that a uniform
backbone (DNA in S04) is
more stable than a mixed back-
bone (one 2'-O-methyl in an
otherwise DNA sequence in
S03). This finding allows the
conclusion that 2'-O-methyl nu-
cleosides are not able to stabi-
lize triplexes when incorporat-
ed in mixed-backbone Hoogs-
teen strands. Interestingly, T04
is slightly more stable at phys-
iological pH than under acidic
conditions. To our surprise,
RNA strand S05 was not able
to form a triple helix with its
target duplex D1. We suggest
that this unusual behavior is
caused by an extraordinary sug-
ar conformation of 6-oxocyti-
dine that distorts the backbone
in such a way as to prevent the
hybridization of the third
strand. This suggestion is sup-
ported by a comparison of the
CD spectra of the single strands
S02, S05, and S08 (see Fig-
ure 8). The sugar conformation
of the nucleoside was deter-
mined by X-ray crystallograph-
ic analysis.[63, 64]
The triplexes T06 and T07,
which contain one 5-allyl-6-ox-
ocytosine in the middle of their
sequences, exhibited very low
Tm values (Table 2). Obviously
Scheme 5. Synthesis of the RNA phosphoramidites 37 and 38: reagents and conditions: i) 1m solution of
TBDMSCl in THF, AgNO3, THF/pyridine 1:1, RT, 18 h; ii) iPrNPCl(OCH2CH2CN) (18), 1-methylimidazole,
sym-collidine, CH3CN, RT, 2 h.
the transformation of an absorbance versus temperature
curve to an a versus T curve have been published by various
authors.[60±62] The CD spectra were measured with the same
samples. For details, again, see experimental section.
the allyl substituent caused destabilization (compare T03 and
T04 with T06 and T07 in Table 2). Advantageous base
stacking or hydrophobic forces, which are responsible for the
effect of the methyl group in stabilizing 5-methylcytosine
compared with cytosine, cannot play a role in this case. A
more detailed investigation utilizing NMR techniques is
required to provide details of the structural perturbation
caused by the allyl function.
Hoogsteen strand S11, which contains three cytosine bases,
hybridizes with its target duplex D3 (S11 ‡ D3 ! T11). We
found a very strong pH dependence for the triplex stability
(Table 2). The RNA strand S12 formed a more stable triple
helix over the pH range investigated than its DNA counter-
part S11 (see Figure 1 and Table 2). At pH 5, triplex and
duplex melt together, showing only one transition in the UV
melting curve. The formation of triplex T13 was pH
independent, but the Tm was degraded compared to T03.
This confirms the hypothesis that a mixed backbone leads to
destabilization. Triplex T14 (S14 ‡ D3), with a uniform DNA
backbone and three 6-oxocytosine bases in the sequence, gave
the best results in this study. We found pH-independent triple-
Triplex studies: After the synthesis and characterization of all
oligonucleotides (Table 1), the triplex formation between the
cytosine- or 6-oxocytosine-containing Hoogsteen strands
S01 ± S08 or S11 ± S18 and their target duplexes D1 or D3
were studied. The melting temperatures (Tm) are summarized
in Table 2. Samples that have formed a triple helix undergo
two transitions, while samples without a triplex exhibit only
the second transition, characterizing denaturation of the
duplex.
We started with the hybridization of strand S01 with its
target duplex D1 (S01 ‡ D1 ! T01). As expected, the UV
melting curves revealed a pH-dependent triplex stability
caused by the protonation of the central cytosine base of the
Hoogsteen strand S01. Use of the RNA strand S02 gave
almost the same results (Table 2). The hybridization of strands
S03 and S04 with duplex D1 resulted in a virtually pH-
Chem. Eur. J. 2000, 6, No. 13
ꢀ WILEY-VCH Verlag GmbH, D-69451 Weinheim, 2000
0947-6539/00/0613-2413 $ 17.50+.50/0
2413

U. Parsch, J. W. Engels
FULL PAPER
Table 1. Oligonucleotide sequences, their calculated and detected masses, and calculated extinction coefficients.
Abbr.
Oligonucleotide sequence[b]
Calcd mass [Da]
Detected mass [Da] (MALDI)[c]
Extinction coefficient at 260 nm
D1
d(CCCAAAAAAAGAAAAAAACCC)
d(GGGTTTTTTTCTTTTTTTGGG)
d(TTTTTTTCTTTTTTT)
6387.30
6461.23
4485.97
4529.59
4532.00
4501.97
4545.59
4572.06
4542.04
4585.66
6386.33
6460.73
4485.05
4528.28
4530.98
4522.52
4568.73
4571.05
4542.31
4581.89
224.54
192.50
127.26
147.06
127.26
127.26
147.06
127.26
127.26
147.06
S01
S02
S03
S04
S05
S06
S07
S08
r(UUUUUUUCUUUUUUU)
d(TTTTTTTHOOMeTTTTTTT)
d(TTTTTTTHOHTTTTTTT)
r(UUUUUUUHOOHUUUUUUU)
d(TTTTTTTAOOMeTTTTTTT)
d(TTTTTTTAOHTTTTTTT)
r(UUUUUUUAOOHUUUUUUU)
D3
d(CCCAAAAGAAGAAGAAAACCC)
d(GGGTTTTCTTCTTCTTTTGGG)
d(TTTTCTTCTTCTTTT)
6419.30
6431.21
4455.95
4527.62
4594.04
4503.95
4575.62
4714.22
4624.16
4695.83
6418.68
6431.45
4454.75
4526.80
4595.09
4526.76
4595.51
4713.36
4623.39
4695.64
224.54
191.06
125.82
140.48
125.82
125.82
140.48
125.82
125.82
140.48
S11
S12
S13
S14
S15
S16
S17
S18
r(UUUUCUUCUUCUUUU)
d(TTTTHOOMeTTHOOMeTTHOOMeTTTT)
d(TTTTHOHTTHOHTTHOHTTTT)
r(UUUUHOOHUUHOOHUUHOOHUUUU)
d(TTTTAOOMeTTAOOMeTTAOOMeTTTT)
d(TTTTAOHTTAOHTTAOHTTTT)
r(UUUUAOOHUUAOOHUUAOOHUUUU)
K[a]
d(TTTTTTTTTTTTTTT)
4500.96
4500.72
127.98
H
A
[a] K ˆ control. [b] HOOMe ˆ 2'-O-methyl-6-oxocytidine, HOH ˆ 2'-deoxy-6-oxocytidine, OOH ˆ 6-oxocytidine, OOMe ˆ 5-allyl-2'-O-methyl-6-oxocytidine,
AOH ˆ 5-allyl-2'-deoxy-6-oxocytidine, AOOH ˆ 5-allyl-6-oxocytidine. [c] All molecular peaks found were in good agreement with the calculated molweights. In
some cases sodium adducts were found.
Table 2. Tm values of triplexes and duplexes.
Abbr.
Tm values [8C][a]
pH 5.0
pH 6.0
pH 7.0
pH 7.4
T01
T02
T03
T04
T05
T06
T07
T08
33.5
35.0
13.0
16.0
±
3.5
4.5
±
26.5
27.5
12.5
16.0
±
3.0
4.0
±
18.5
18.0
13.0
17.0
±
2.5
4.5
±
16.0
14.0
12.5
18.0
±
3.0
5.0
±
T11
T12
T13
T14
T15
T16
T17
T18
62.0
66.5
9.0
27.5
±
±
6.0
±
42.5
53.0
10.0
27.0
±
±
5.5
±
25.5
36.0
10.0
27.0
±
±
5.0
±
19.5
30.0
9.0
27.0
±
±
4.5
±
Figure 1. UV absorbance melting curves of triplex T12 at different pH
values in 50 mm sodium cacodylate buffer, 100 mm NaCl, and 20 mm
MgCl2.
K
±
±
±
±
D1
D3
64.5
64.5
65.5
67.0
65.0
66.5
65.0
66.5
[a] Errors: Æ 0.58C; ± indicates that no triplex was formed. The Tm values
are means of at least four measurements.
helix formation with a Tm of 278C (Figure 2 and Table 2). This
means a stabilization of 7.58C at pH 7.4 compared with its
natural analogue T11. Figure 3 shows the Tm values of
triplexes T11 ± T14 at different pH values. The almost linear
pH dependence over the whole pH range under investigation
gives rise to the assumption that the cytosines are still (in part)
protonated under neutral and slightly basic conditions. This is
in agreement with the results of other groups.[11±15] The
differences in stability are large and cannot be explained
simply by the presence and absence of one hydrogen bond.
We assume that factors like base stacking and electrostatic
Figure 2. UV absorbance melting curves of triplex T14 at different pH
values in 50 mm sodium cacodylate buffer, 100 mm NaCl, and 20 mm
MgCl2.
2414
ꢀ WILEY-VCH Verlag GmbH, D-69451 Weinheim, 2000
0947-6539/00/0613-2414 $ 17.50+.50/0
Chem. Eur. J. 2000, 6, No. 13

Triple-Helix Formation
2409±2424
between 6-oxocytosine and guanine really exist. Figure 4
shows the result of the control experiment in direct compar-
ison with the melting curve of triplex T14. It can be seen that
triplex T14 is much more stable than KD3. In the control
experiment only the upper half of the triplex ± duplex
transition is visible. An exact determination of the Tm is not
possible, but the experiment gives strong evidence that the
6-oxocytosine ± guanine ± cytosine base triad exists as pro-
posed in Scheme 1.
To confirm the triplex formation, different CD measure-
ments were performed.[66±69] First we measured wavelength-
dependent spectra of triplexes at different temperatures. All
the triple helices investigated had a B-DNA structure, which
was expected because of the underlying DNA duplexes. Only
if the third strand was RNA (T02 and T12) did some
structural features of A-DNA appear (data not shown).
Triplex CD spectra exhibit a negative band between 210 and
220 nm, which proves the triplex formation. This band
disappears at temperatures above the Tm (Figure 5). Differ-
ence spectra (in which the sum of the duplex and single-strand
spectra is subtracted from the triplex spectrum) are able to
show the parts of the spectrum specific to the triplex (data not
shown). Careful analysis of the CD spectra reveals that not
only can triplex melting be seen at 210 ± 220 nm, but also
duplex melting at wavelengths between 240 and 250 nm
(Figure 5). In this way it is possible to detect triplex and
duplex melting independently of each other, by performing
temperature-dependent experiments at selected wave-
lengths.[70] To detect triplex melting we chose 217 nm as the
optimal wavelength. For duplex melting, 246 nm was ideal.
This method worked well with DNA and RNA as third strand,
and also when artificial nucleobases like 6-oxocytosine were
incorporated into the Hoogsteen strand. An example is given
in Figure 6. These selective CD melting curves are particularly
useful in cases where duplex and triplex melting overlap. The
UV melting curves of T11 and T12 showed only one
transition at pH 5.0. Use of CD melting experiments at
selected wavelengths allowed the individual observation of
each melting process (Figure 7). Finally we investigated the
CD spectra of the Hoogsteen single strands. This makes it
possible to gain insight into conformational changes between
strands which contain only natural nucleotides and those with
6-oxocytidines. In fact the CD spectra of the RNA single
Figure 3. The Tm values of triplexes T11 ± T14 at different pH values.
contributions of the positive charge at N(3) with the negative
phosphate backbone play an important role in triple-helix
stabilization.[15, 65]
The 6-oxocytidine-containing RNA strands S15 and S18
did not form triplexes for the reasons discussed above.
Studying the hybridization of strands S16 and S17 confirmed
the strong destabilization caused by the allyl group.
The specificity of the triple-helix formation from strands
including 6-oxocytosine has already been investigated by
Berressem and Engels.[42] Their studies showed the specificity
of 6-oxocytosine to guanine in a Watson ± Crick base pair.
Triplex formation with reduced hypochromicity also occurred
when 6-oxocytosine paired with adenine. This affinity to
adenine was explained by the existence of an additional
tautomer of 6-oxocytosine.[42] In the present work, we
designed another control experiment, which should corrobo-
rate the existence of the two hydrogen bonds between
6-oxocytosine and guanine as postulated in Scheme 1. We
synthesized a 15-mer polydeoxythymidine (strand K in
Table 1). Thymine is able to form one hydrogen bond to
guanine. If 6-oxocytosine forms only one hydrogen bond to
guanine, the triplex between K and D3 should have a similar
Tm to that of T14. If the Tm of KD3 is clearly lower than the
one of triplex T14, one can assume that two hydrogen bonds
Figure 4. Comparison of UV melting curves of T14 (left) and KD3 (right) at pH 7.0.
Chem. Eur. J. 2000, 6, No. 13 ꢀ WILEY-VCH Verlag GmbH, D-69451 Weinheim, 2000
0947-6539/00/0613-2415 $ 17.50+.50/0
2415

U. Parsch, J. W. Engels
FULL PAPER
Figure 7. Selective CD melting experiments: detection of triplex (l ˆ
217 nm) and duplex (l ˆ 246 nm) melting for triplex T11 at pH 5.0. The
absolute value for the variation of ellipticity is 15.5 mdeg (triplex) and
7.0 mdeg (duplex). The UV melting experiment shows only one transition
in this case.
Figure 5. Wavelength-dependent CD spectra of triplex T11 (pH 7.0).
Figure 8. Wavelength-dependent CD spectra of the RNA single strands
S02, S05, and S08 (pH 7.0).
Determination of thermodynamic data: The measured UV
melting curves were used to calculate the thermodynamic
parameters DH8, DS8, and DG8 for triplexes and duplexes.
The vanꢁt Hoff plot (RlnK ˆ ÀDH8 ´ 1/T‡ DS0) yields DH8
and DS8. The standard equation DG8 ˆ DH8 À TDS8 gives the
free energy. Alternatively, DG8 can be calculated from the
equation DG8 ˆ ÀRTlnK. All methods utilized and their
underlying formulas were derived and explained in ref.[62]
.
The calculated values are given in Tables 3 and 4. In the cases
of T11 and T12 it was impossible to calculate DH8 and TDS8,
because duplex and triplex transitions were not separated
from each other. This means that the requirement of the two-
state model was not fulfilled.[62] DG8 could be calculated in
these cases from the equation DG8 ˆ ÀRTlnK.
Figure 6. CD melting experiment for triplex T14 (pH 7.0). Measurements
were made at 217 nm and show selectively the triplex-to-duplex transition.
The absolute value for the variation of ellipticity is 10.0 mdeg.
strands give further evidence that 6-oxocytidine leads to
backbone distortion. The base stacking of strands S05 and
S08 was much weaker than that in strand S02, as can be seen
from the changes in the spectra between 260 and 290 nm
(Figure 8). On the other hand, the single-strand spectra did
not indicate why the 5-allyl function reduces the melting
temperature. The spectra of S01, S06, and S07 showed only
minimal differences (data not shown).
Altogether, the free energy values show a tendency similar
to that of the Tmꢁs; the enthalpies and entropies are relative
constant. Only triplex T02 (third strand RNA) has enthalpy
and entropy values a little lower. ÀDH8 varies between À2.9
and À3.9 kcalmolÀ1 per Hoogsteen base pair, depending on
sequence and pH, which correlates with other published
results.[8, 71] We could not find any distinct effect of 6-oxocy-
tosine on DH8 and DS8. This lack may be due to the small
2416
ꢀ WILEY-VCH Verlag GmbH, D-69451 Weinheim, 2000
0947-6539/00/0613-2416 $ 17.50+.50/0
Chem. Eur. J. 2000, 6, No. 13

Triple-Helix Formation
2409±2424
Table 3. Thermodynamic data of triplexes.
substitution has a destabilizing influence on triplex stability.
An important finding was that a uniform backbone is more
stable than a mixed one. While triple helices with RNA third
strands that contain only standard nucleosides are more stable
than triplexes with DNA third strands (T12 is more stable
than T11), this study revealed that RNA strands with
6-oxocytidine or 5-allyl-6-oxocytidine did not form a triple
helix with a DNA target duplex. We suppose this is because of
an unusual conformation of the sugar of 6-oxocytidine, which
leads to backbone torsional constraints. CD spectroscopic
investigations gave further evidence for this interpretation. A
15-mer DNA sequence with three isolated 2'-deoxy-6-oxocy-
tidines exhibited the highest Tm of all cytidine analogues
investigated in this study. Temperature-dependent CD meas-
urements permitted us to recognize duplex and triplex
melting individually at selected wavelengths. This is especially
interesting in cases where UV melting curves show only one
transition.
Triplex
pH
À DH8
À TDS8
À DG8
r2[a]
[kcalmolÀ1
]
[kcalmolÀ1
]
[kcalmolÀ1
]
(Tˆ 298 K)
(Tˆ 298 K)
T01
5.0
6.0
7.0
7.4
5.0
6.0
7.0
7.4
5.0
6.0
7.0
7.4
5.0
6.0
7.0
7.4
5.0
6.0
7.0
7.4
5.0
6.0
7.0
7.4
5.0
6.0
7.0
7.4
54 Æ 5
52 Æ 6
51 Æ 2
50 Æ 2
48 Æ 2
47 Æ 2
44 Æ 2
45 Æ 4
54 Æ 3
54 Æ 4
52 Æ 5
53 Æ 4
52 Æ 2
53 Æ 1
55 Æ 4
54 Æ 2
44 Æ 5
44 Æ 6
44 Æ 2
44 Æ 2
38 Æ 2
38 Æ 2
37 Æ 2
39 Æ 4
48 Æ 3
48 Æ 4
46 Æ 5
47 Æ 4
45 Æ 2
46 Æ 1
48 Æ 4
47 Æ 2
9.5 Æ 0.1
8.6 Æ 0.2
7.1 Æ 0.1
6.8 Æ 0.2
9.7 Æ 0.1
8.8 Æ 0.2
7.3 Æ 0.3
6.5 Æ 0.3
6.4 Æ 0.2
6.0 Æ 0.1
6.4 Æ 0.3
6.2 Æ 0.3
7.2 Æ 0.4
7.3 Æ 0.2
7.1 Æ 0.2
7.1 Æ 0.3
13.5 Æ 0.4
10.8 Æ 0.3
8.3 Æ 0.1
6.6 Æ 0.2
12.6 Æ 0.3
11.2 Æ 0.1
9.9 Æ 0.2
9.2 Æ 0.2
8.3 Æ 0.2
8.4 Æ 0.1
8.3 Æ 0.2
8.4 Æ 0.1
0.999
0.998
0.998
0.997
1.000
0.998
0.994
0.992
0.993
0.987
0.992
0.989
0.986
0.990
0.991
0.995
T02
T03
T04
T11
T12
T14
Experimental Section
Melting points were not corrected. Microanalyses were carried out by the
analytical laboratory of the institute. Yields refer to analytically pure
compounds. TLC was performed with Merck silica gel plates 60F254, and
column chromatography by standard techniques on silica gel. All reactions
were carried out under an inert, dry atmosphere (Argon). All solvents were
purchased in dry p.a. quality and stored over molecular sieves. The 1H, 13C,
31P, 1H,1H-COSY, and HSQC NMR spectra were recorded on the following
instruments: Bruker AM-250, WH-270, and AM-X400. The NMR spectra
were referenced to the central [D6]DMSO line at d ˆ 2.49 (1H) and 39.5
(13C) or CDCl3 at d ˆ 7.27 (1H). The mass spectra were determined by the
electrospray technique on a VG Analytical Platform II instrument.
53 Æ 7
57 Æ 4
56 Æ 3
57 Æ 3
45 Æ 7
49 Æ 3
48 Æ 3
49 Æ 3
0.992
0.998
0.996
0.999
[a] r2 ˆ regression factor.
Table 4. Thermodynamic data of duplexes.
Ethyl-2-cyanopent-4-enoate (2): Ethyl cyanoacetate (53.3 mL, 500 mmol)
was slowly added to a solution of sodium (11.5 g, 500 mmol) in ethanol
(250 mL). A white precipitate of the sodium salt appeared. Allyl chloride
(42.7 mL, 525 mmol) was added and the suspension was refluxed until
neutrality was reached (3 h). The major part of the ethanol (180 mL) was
removed by distillation. After the remaining mixture had been cooled to
RT, ice water was added until a clear solution appeared (150 mL). Aqueous
and organic phases were separated and the aqueous phase was extracted
twice with ether. The collected organic phases were dried over MgSO4. A
distillation (30 cm Vigreux) did not lead to pure product. Finally, column
chromatography was performed (n-hexane/ethyl acetate 4:1). The product
was obtained as a colorless liquid (24.5 g, 32%). B.p. 2238C; Rf ˆ 0.31 (n-
Duplex
pH
À DH8
À TDS8
À DG8
r2[a]
[kcalmolÀ1
]
[kcalmolÀ1
]
[kcalmolÀ1
]
(Tˆ 298 K)
(Tˆ 298 K)
D1
5.0
6.0
7.0
7.4
5.0
6.0
7.0
7.4
104 Æ 7
105 Æ 8
101 Æ 2
103 Æ 3
122 Æ 1
115 Æ 6
115 Æ 5
119 Æ 4
86 Æ 6
85 Æ 7
81 Æ 2
84 Æ 3
103 Æ 1
95 Æ 6
95 Æ 5
100 Æ 4
18.7 Æ 0.6
19.8 Æ 0.9
19.4 Æ 0.2
19.1 Æ 0.5
19.4 Æ 0.1
19.2 Æ 0.2
19.4 Æ 0.1
19.5 Æ 0.6
0.969
0.981
0.981
0.971
0.988
0.976
0.978
0.981
D3
1
hexane/ethyl acetate 4:1); H NMR (250 MHz, CDCl3, 258C): d ˆ 1.33 (t,
[a] r2 ˆ regression factor.
J ˆ 7.1 Hz, 3H, CH3), 2.68 (m, 2H, allyl CH2), 3.57 (dd, J ˆ 6.4 Hz, 0.9 Hz,
1H, CH), 4.27 (q, J ˆ 7.1 Hz, 2H, CH2), 5.26 (m, 2H, allyl CH2), 5.83 (m,
1H, allyl CH); 13C NMR (63 MHz, CDCl3, 258C): d ˆ 13.88 (CH3), 33.71
(allyl CH2), 37.38 (CH), 62.73 (CH2), 115.97 (CN), 119.83 (allyl CH2), 131.35
sequence differences and the relatively large uncertainty of
the calculated data (see Tables 3 and 4).
Watson ± Crick base pairs are enthalpically more stable
(À4.8 to À5.8 kcalmolÀ1 per base pair in this study) than
Hoogsteen base pairs.
ˆ
(allyl CH), 165.41 (C O).
5-Allyl-6-oxocytosine (5): Urea (2.40 g, 40 mmol) and ethyl-2-cyanopent-4-
enoate (2; 6.14 mL, 40 mmol) were added to a solution of sodium (1.84 g,
80 mmol) in methanol (100 mL), and the mixture was refluxed for 6 h.
Some methanol (90 mL) was removed by distillation. Warm water (60 mL)
was added and the solution was acidified with acetic acid (4.6 mL,
80 mmol). After filtration of the precipitate, a white solid was obtained
(3.67 g, 55%). M.p. > 2508C; Rf ˆ 0.17 (dichloromethane/methanol 9:1);
UV: lmax (H2O, pH 7.0) ˆ 271 nm; 1H NMR (250 MHz, [D6]DMSO, 258C):
d ˆ 2.87 (d, J ˆ 5.9 Hz, 2H, allyl CH2), 4.90 (ddd, J ˆ 17.1, 10.0, 2.1 Hz, 2H,
allyl CH2), 5.68 (m, 1H, allyl CH), 5.86 (brs, 2H, NH2), 9.89 (brs, 1H, NH),
10.21 (brs, 1H, NH); 13C NMR (63 MHz, [D6]DMSO, 258C): d ˆ 25.78
(allyl CH2), 82.42 (C-5), 113.73 (allyl CH2), 136.20 (allyl CH), 150.18 (C-2),
151.49 (C-4), 163.88 (C-6); elemental analysis calcd (%) for C7H9N3O2
(167.17): C 50.29, H 5.43, N 25.14; found C 50.57, H 5.51, N 24.88; MS
ESI(À): m/z (%): 165.9 (100) [M À H]À, ESI(‡): m/z (%): 167.8 (100)
Conclusion
We have shown in this study that 6-oxocytosine and 5-allyl-6-
oxocytosine can be transformed into protected phosphorami-
dites and incorporated into oligonucleotides by standard
solid-phase synthesis. Comparison of the triplex-forming
abilities of the different building blocks showed that a 5-allyl
‡
[M‡H] .
Chem. Eur. J. 2000, 6, No. 13
ꢀ WILEY-VCH Verlag GmbH, D-69451 Weinheim, 2000
0947-6539/00/0613-2417 $ 17.50+.50/0
2417

U. Parsch, J. W. Engels
FULL PAPER
General procedure for the synthesis of protected nucleosides 6 and 7: The
nucleobase was refluxed in 10 equiv hexamethyldisilazane (HMDS) for
24 h, with a small amount of ammonium sulfate as starting reagent, and under
exclusion of moisture. After cooling to RT, the HMDS was completely
removed under reduced pressure, and 1,2-dichloroethane (100 mL) was
added to the residue. Subsequently 1,2,3,5-tetra-O-acetyl-b-d-ribofuranose
and trimethylsilyl trifluoromethanesulfonate were added, and the clear
solution was stirred at RT for 24 h. The reaction mixture was extracted with
dichloromethane and saturated NaHCO3 solution, and washed with brine.
The collected organic phases were dried with MgSO4, filtered and eva-
porated. The product was crystallized from ethanol/ether (6) or obtained
after column chromatography (dichloromethane/methanol 95:5) (7).
heated to 65 ± 708C for 21 h. Then it was cooled to RT and the solvent was
removed under reduced pressure. The residue was extracted in dichloro-
methane/water and the collected organic phases were dried over MgSO4,
filtered, and evaporated to dryness. The syrup was purified by column
chromatography (dichloromethane/methanol 9:1). After evaporation the
product (110 mg, 37%) was obtained as a white foam. M.p. 237 ± 2388C;
Rf ˆ 0.26 (ethyl acetate/methanol 4:1); UV: lmax (H2O, pH 7.0) ˆ 275 nm;
1H NMR (250 MHz, [D6]DMSO, 258C): d ˆ 2.91 (d, J ˆ 5.9 Hz, 2H, allyl
CH2), 3.40 (m, 1H, 5'a-H), 3.60 (m, 2H, 4'-H, 5'b-H), 4.06 (q, J ˆ 5.6 Hz,
1H, 3'-H), 4.45 (q, J ˆ 5.5 Hz, 1H, 2'-H), 4.57 (t, J ˆ 4.8 Hz, 1H, 5'-OH),
4.75 (d, J ˆ 6.5 Hz, 1H, 3'-OH), 4.93 (ddd, J ˆ 17.2, 10.0, 2.2 Hz, 2H, allyl
CH2), 4.94 (d, J ˆ 5.3 Hz, 1H, 2'-OH), 5.69 (m, 1H, allyl CH), 6.04 (d, J ˆ
3.9 Hz, 1H, 1'-H), 6.06 (brs, 2H, NH2), 10.24 (brs, 1H, NH); 13C NMR
(63 MHz, [D6]DMSO, 258C): d ˆ 26.39 (allyl CH2), 62.44 (C-5'), 70.17 (C-
3'), 70.98 (C-2'), 82.43 (C-5), 84.07 (C-4'), 87.26 (C-1'), 113.91 (allyl CH2),
135.95 (allyl CH), 149.93 (C-2), 150.60 (C-4), 162.45 (C-6); elemental
analysis calcd (%) for C12H17N3O6 (299.28): C 48.16, H 5.73, N 14.04; found
C 47.93, H 5.93, N 13.93; MS ESI(À): m/z (%): 298.2 (100) [M À H]À.
General procedure for protection of the exocyclic amino function of 8 and
9: The unprotected nucleosides 8 and 9 were suspended in dry methanol.
The nucleoside was dissolved by heating. N,N-Dimethylformamide di-
methyl acetal was added to the boiling solution. After 10 min the solution
was cooled to RTand the solvent was removed under reduced pressure. The
residue was purified by column chromatography with dichloromethane/
methanol 9:1 ! 4:1.
4-N-Dimethylformamidine-6-oxocytidine (10): 6-Oxocytidine (8; 3.80 g,
14.66 mmol) in methanol (100 mL) with N,N-dimethylformamide dimethyl
acetal (10.6 mL, 73.3 mmol, 5 equiv) afforded 10 (4.14 g, 90%) as a white
foam. Rf ˆ 0.35 (dichloromethane/methanol 4:1), Rf ˆ 0.11 (ethyl acetate/
methanol 4:1); 1H NMR (250 MHz, [D6]DMSO, 258C): d ˆ 2.96 (s, 3H,
N(CH3)), 3.09 (s, 3H, N(CH3)), 3.41 (m, 1H, 5'a-H), 3.56 (m, 1H, 5'b-H),
3.65 (m, 1H, 4'-H), 4.08 (q, J ˆ 6.2 Hz, 1H, 3'-H), 4.48 (q, J ˆ 4.8 Hz, 1H, 2'-
H), 4.59 (d, J ˆ 6.0 Hz, 1H, 5'-OH), 4.78 (d, J ˆ 6.3 Hz, 1H, 3'-OH), 4.96 (d,
J ˆ 5.2 Hz, 1H, 2'-OH), 5.05 (s, 1H, 5-H), 6.04 (d, J ˆ 3.8 Hz, 1H, 1'-H),
8.14 (s, 1H, NCHN), 10.75 (brs, 1H, NH); 13C NMR (63 MHz, [D6]DMSO,
258C): d ˆ 34.34 (N(CH3)), 40.34 (N(CH3)), 62.45 (C-5'), 70.22 (C-3'), 70.95
(C-2'), 81.04 (C-5), 84.17 (C-4'), 87.00 (C-1'), 151.19 (C-2), 157.00 (NCHN),
159.22 (C-4), 163.54 (C-6); elemental analysis calcd (%) for C12H18N4O8
(314.30): C 45.86, H 5.77, N 17.83; found C 45.95, H 6.04, N 16.61; MS
ESI(À): m/z (%): 313.1 (100) [M À H]À.
5-Allyl-4-N-dimethylformamidine-6-oxocytidine (11): 5-Allyl-6-oxocyti-
dine 9 (4.26 g, 14.23 mmol) in methanol (120 mL) with N,N-dimethylfor-
mamide dimethyl acetal (10.26 mL, 71.1 mmol, 5 equiv) afforded 11 (4.33 g,
88%) as a white foam. Rf ˆ 0.51 (dichloromethane/methanol 4:1); 1H NMR
(250 MHz, [D6]DMSO, 258C): d ˆ 2.94 (s, 3H, N(CH3)), 2.99 (d, J ˆ 6.3 Hz,
2H, allyl CH2), 3.03 (s, 3H, N(CH3)), 3.41 (m, 1H, 5'a-H), 3.63 (m, 2H, 4'-
H, 5'b-H), 4.10 (q, J ˆ 6.1 Hz, 1H, 3'-H), 4.48 (q, J ˆ 5.8 Hz, 1H, 2'-H), 4.59
(t, J ˆ 5.8 Hz, 1H, 5'-OH), 4.80 (d, J ˆ 6.5 Hz, 1H, 3'-OH), 4.87 (ddd, J ˆ
17.2, 10.0, 2.2 Hz, 2H, allyl CH2), 4.99 (d, J ˆ 5.3 Hz, 1H, 2'-OH), 5.74 (m,
1H, allyl CH), 6.08 (d, J ˆ 3.7 Hz, 1H, 1'-H), 7.97 (s, 1H, NCHN), 10.58
(brs, 1H, NH); 13C NMR (63 MHz, [D6]DMSO, 258C): d ˆ 28.11 (allyl
CH2), 33.78 (N(CH3)), 40.05 (N(CH3)), 62.50 (C-5'), 70.26 (C-3'), 71.02 (C-
2'), 84.24 (C-4'), 87.55 (C-1'), 95.36 (C-5), 113.98 (allyl CH2), 137.17 (allyl
CH), 150.68 (C-2), 153.43 (C-4), 155.52 (NCHN), 163.45 (C-6); elemental
analysis calcd (%) for C15H22N4O6 (354.36): C 50.84, H 6.26, N 15.81; found
C 50.56, H 6.32, N 15.92; MS ESI(À): m/z (%): 353.2 (100) [M À H]À.
2',3',5'-Tri-O-acetyl-6-oxocytidine (6): 6-Oxocytosine (4, Aldrich; 5.0 g,
39.3 mmol), HMDS (83 mL, 393 mmol), 1,2,3,5-tetra-O-acetyl-b-d-ribofur-
anose (11.9 g, 37.3 mmol), and trimethylsilyl trifluoromethanesulfonate
(7.1 mL, 39.3 mmol) afforded 6 (11.93 g, 83%) as a white solid. M.p. 228 ±
2298C; Rf ˆ 0.27 (dichloromethane/methanol 9:1); 1H NMR (270 MHz,
[D6]DMSO, 258C): d ˆ 1.99 (s, 3H, CH3), 2.02 (s, 3H, CH3), 2.05 (s, 3H,
CH3), 4.05 (m, 2H, 5'-H), 4.30 (m, 1H, 4'-H), 4.55 (s, 1H, 5-H), 5.46 (t, J ˆ
7.1 Hz, 1H, 3'-H), 5.63 (dd, J ˆ 2.6, 6.6 Hz, 1H, 2'-H), 6.14 (brs, 1H, 1'-H),
6.43 (brs, 2H, NH2), 10.58 (brs, 1H, NH); 13C NMR (63 MHz, [D6]DMSO,
258C): d ˆ 20.19 (CH3), 20.31 (CH3), 20.42 (CH3), 63.07 (C-5'), 69.87 (C-3'),
72.58 (C-2'), 73.58 (C-5), 77.71 (C-4'), 84.50 (C-1'), 150.30 (C-2), 154.33 (C-
ˆ
ˆ
ˆ
4), 161.86 (C-6), 169.28 (C O), 169.52 (C O), 170.04 (C O); elemental
analysis calcd (%) for C15H19N3O9 (385.34): C 46.75, H 4.97, N 10.90; found
C 46.90, H 5.02, N 10.71; MS ESI(À): m/z (%): 384.1 (100) [M À H]À.
2',3',5'-Tri-O-acetyl-5-allyl-6-oxocytidine (7): 5-Allyl-6-oxocytosine (5.95 g,
35.6 mmol), HMDS (75 mL, 356 mmol), 1,2,3,5-tetra-O-acetyl-b-d-ribofur-
anose (10.82 g, 34.0 mmol), and trimethylsilyl trifluoromethanesulfonate
(6.46 mL, 35.6 mmol) afforded 7 (6.65 g, 46%) as a white solid. M.p. 215 ±
2168C; Rf ˆ 0.48 (dichloromethane/methanol 9:1); 1H NMR (270 MHz,
[D6]DMSO, 258C): d ˆ 1.99 (s, 3H, CH3), 2.02 (s, 3H, CH3), 2.06 (s, 3H,
CH3), 2.92 (d, J ˆ 5.9 Hz, 2H, allyl CH2), 4.05 (m, 2H, 5'a-H, 4'H), 4.31 (m,
1H, 5'b-H), 4.94 (ddd, J ˆ 17.2, 10.0, 2.1 Hz, 2H, allyl CH2), 5.49 (t, J ˆ
7.1 Hz, 1H, 3'-H), 5.63 (dd, J ˆ 7.2, 2.1 Hz, 1H, 2'-H), 5.71 (m, 1H, allyl CH),
6.14 (brs, 3H, 1'-H and NH2), 10.41 (brs, 1H, NH); 13C NMR (63 MHz,
[D6]DMSO, 258C): d ˆ 20.24 (CH3), 20.36 (CH3), 20.48 (CH3), 26.30 (allyl
CH2), 62.98 (C-5'), 69.76 (C-3'), 72.67 (C-2'), 77.69 (C-4'), 82.15 (C-5), 85.10
(C-1'), 114.06 (allyl CH2), 135.79 (allyl CH), 149.66 (C-2), 150.93 (C-4),
ˆ
ˆ
ˆ
161.63 (C-6), 169.32 (C O), 169.57 (C O), 170.07 (C O); elemental
analysis calcd (%) for C18H23N3O9 (425.39): C 50.82, H 5.45, N 9.88; found
C 50.64, H 5.48, N 9.59; MS ESI(À): m/z (%): 426.1 (100) [M À H]À.
General procedure for the deprotection of 6 and 7: Blocked nucleosides
were dissolved in a solution of ammonia gas in methanol (saturated at
À788C). After stirring for 24 h at RT, the solvent was removed under re-
duced pressure and the residue was crystallized from water (8) or obtained
after column chromatography (dichloromethane/methanol 4:1) (9).
6-Oxocytidine (8): Compound
6 (5.0 g, 13.0 mmol) in a solution of
ammonia gas in methanol (200 mL) gave 8 (3.12 g, 93%) as a white solid.
M.p. 204 ± 2068C; Rf ˆ 0.08 (ethyl acetate/methanol 4:1); UV: lmax (H2O,
pH 7.0) ˆ 268 nm; 1H NMR (270 MHz, [D6]DMSO, 258C): d ˆ 3.38 (m, 1H,
5'a-H), 3.55 (m, 1H, 5'b-H), 3.63 (m, 1H, 4'-H), 4.05 (q, J ˆ 6.2 Hz, 1H, 3'-
H), 4.44 (q, J ˆ 5.3 Hz, 1H, 2'-H), 4.53 (s, 1H, 5-H), 4.58 (t, J ˆ 4.9 Hz, 1H,
5'-OH), 4.77 (d, J ˆ 6.4 Hz, 1H, 3'-OH), 4.94 (d, J ˆ 5.3 Hz, 1H, 2'-OH),
6.01 (d, J ˆ 3.9 Hz, 1H, 1'-H), 6.34 (brs, 2H, NH2), 10.42 (brs, 1H, NH);
13C NMR (63 MHz, [D6]DMSO, 258C): d ˆ 62.48 (C-5'), 70.23 (C-3'), 71.04
(C-2'), 74.03 (C-5), 84.09 (C-4'), 86.80 (C-1'), 150.73 (C-2), 154.06 (C-4),
162.85 (C-6); elemental analysis calcd (%) for C9H13N3O6 (259.22): C 41.70,
H 5.06, N 16.21; found C 41.64, H 5.02, N 16.17; MS ESI(À): m/z (%): 258.0
(100) [M À H]À.
General procedure for dimethoxytritylation of 10 and 11: The nucleosides
10 and 11 were each dissolved in dry pyridine and, after formation of the
clear solution, 4,4'-dimethoxytriphenylmethyl chloride (DMTrCl) was
added (1.2 equiv). After stirring the mixture at RT for 3.5 h (10) or 6 h
(11), methanol was added. The pyridine was removed at 258C under
reduced pressure and coevaporated with toluene three times. The residue
was extracted with chloroform/water, and the collected organic phases
were dried with MgSO4, filtered and evaporated to dryness. The residue
was purified by column chromatography (dichloromethane/methanol
95:5).
5-Allyl-6-oxocytidine (9): Compound 7 (6.37 g, 14.97 mmol) in a solution of
gaseous ammonia in methanol (200 mL) gave 5-allyl-6-oxocytidine (4.29 g,
96%) after chromatographic purification as a white foam.
Palladium-catalyzed method: 6-oxocytidine (8; 260 mg, 1.0 mmol) was
dissolved in DMF (4 mL). Triethylamine (167 mL, 1.2 mmol) was added.
The solution was stirred for 15 min. In a second flask tetrakistriphenyl-
phosphinepalladium(0) (58 mg, 0.05 mmol) was suspended in DMF (1 mL).
After addition of allyl bromide (87 mL, 1.0 mmol) a clear yellow solution
was formed. The solution from the first flask was poured into the second
flask and stirred at RT for 30 min. After that the reaction mixture was
5'-O-(4,4'-Dimethoxytriphenylmethyl)-4-N-dimethylformamidine-6-oxo-
cytidine (12): Nucleoside 10 (3.96 g, 12.6 mmol) in pyridine (90 mL) with
DMTrCl (5.12 g, 15.1 mmol) afforded 12 (6.37 g, 82%) as a white solid.
Rf ˆ 0.41 (dichloromethane/methanol 9:1), Rf ˆ 0.33 (ethyl acetate/meth-
2418
ꢀ WILEY-VCH Verlag GmbH, D-69451 Weinheim, 2000
0947-6539/00/0613-2418 $ 17.50+.50/0
Chem. Eur. J. 2000, 6, No. 13

Triple-Helix Formation
2409±2424
anol 4:1); 1H NMR (270 MHz, [D6]DMSO, 258C): d ˆ 2.97 (s, 3H,
N(CH3)), 3.08 (s, 3H, N(CH3)), 3.08 (m, 2H, 5'-H), 3.71 (s, 3H, OCH3),
3.72 (s, 3H, OCH3), 3.82 (m, 1H, 4'-H), 4.15 (q, J ˆ 7.5 Hz, 1H, 3'-H), 4.31
(m, 1H, 2'-H), 4.75 (d, J ˆ 7.5 Hz, 1H, 3'-OH), 4.99 (d, J ˆ 4.7 Hz, 1H, 2'-
OH), 5.06 (s, 1H, 5-H), 6.11 (brs, 1H, 1'-H), 6.81 ± 7.41 (m, 13H, Har), 8.14
(s, 1H, NCHN), 10.70 (brs, 1H, NH); 13C NMR (63 MHz, [D6]DMSO,
258C): d ˆ 34.33 (N(CH3)), 40.51 (N(CH3)), 54.96 (OCH3), 65.04 (C-5'),
70.40 (C-3'), 71.85 (C-2'), 80.95 (C-5), 81.87 (C-4'), 85.25 (DMTr), 87.90 (C-
1'), 113.04, 126.54, 127.72, 129.67, 129.83, 135.76, 145.14 (DMTr), 150.98 (C-
2), 157.02 (NCHN), 157.96 (DMTr), 159.29 (C-4), 163.30 (C-6); elemental
analysis calcd (%) for C33H36N4O8 (616.68): C 64.27, H 5.88, N 9.09; found C
64.06, H 5.96, N 8.80; MS ESI(À): m/z (%): 615.7 (100) [M À H]À.
5-Allyl-5'-O-(4,4'-dimethoxytriphenylmethyl)-4-N-dimethylformamidine-
6-oxocytidine (13): Nucleoside 11 (1.19 g, 3.36 mmol) in pyridine (25 mL)
with DMTrCl (1.37 g, 4.03 mmol) afforded 13 (2.00 g, 91%) as a white solid.
Rf ˆ 0.32 (dichloromethane/methanol 95:5), Rf ˆ 0.47 (dichloromethane/
methanol 9:1); 1H NMR (250 MHz, [D6]DMSO, 258C): d ˆ 2.94 (s, 3H,
N(CH3)), 2.99 (s, 2H, allyl CH2), 3.03 (s, 3H, N(CH3)), 3.08 (m, 2H, 5'-H),
3.70 (s, 3H, OCH3), 3.72 (s, 3H, OCH3), 3.84 (q, J ˆ 5.6 Hz, 1H, 4'-H), 4.18
(q, J ˆ 7.2 Hz, 1H, 3'-H), 4.32 (q, J ˆ 4.3 Hz, 1H, 2'-H), 4.78 (d, J ˆ 7.6 Hz,
1H, 3'-OH), 4.90 (ddd, J ˆ 17.3, 9.8, 1.8 Hz, 2H, allyl CH2), 5.02 (d, J ˆ
4.7 Hz, 1H, 2'-OH), 5.72 (m, 1H, allyl CH), 6.15 (s, 1H, 1'-H), 6.80 ± 7.40
(m, 13H, Har), 7.96 (s, 1H, NCHN), 10.59 (brs, 1H, NH); 13C NMR
(63 MHz, [D6]DMSO, 258C): d ˆ 28.16 (allyl CH2), 33.76 (N(CH3)), 40.04
(N(CH3)), 54.98 (OCH3), 65.05 (C-5'), 70.44 (C-3'), 71.91 (C-2'), 84.09 (C-
4'), 85.26 (DMTr), 88.20 (C-1'), 95.30 (C-5), 113.08 (DMTr), 113.91 (allyl
CH2), 126.54, 127.79, 129.69, 129.84, 135.70 (DMTr), 135.86 (allyl CH),
137.22, 145.17 (DMTr), 153.50 (C-2), 155.46 (NCHN), 157.94 (C-4), 158.01
(DMTr), 163.20 (C-6); elemental analysis calcd (%) for C36H40N4O8
(656.74): C 65.84, H 6.14, N 8.53; found C 65.55, H 6.22, N 8.62; MS
ESI(À): m/z (%): 655.4 (100) [M À H]À.
5-Allyl-5'-O-(4,4'-dimethoxytriphenylmethyl)-4-N-(dimethylformamidine)-
2'-O-methyl 6-oxocytidine (15) and 5-allyl-5'-O-(4,4'-dimethoxytriphenyl-
methyl)-4-N-(dimethylformamidine)-3'-O-methyl 6-oxocytidine (17):
A
solution of tin(ii) chloride (143 mg, 0.756 mmol, 0.33 equiv) in DMF
(5 mL) was added to a solution of 13 (1.50 g, 2.28 mmol) in DMF (15 mL).
The solution was cooled with an ice bath. A solution of diazomethane in
ether (5 mL, approx. 0.3 M) was slowly added. (Diazomethane was freshly
prepared with the DiazaldTM kit diazomethane generator from Aldrich.)
Further portions of diazomethane (5 mL) were added every 60 min. After
4 h the reaction was stopped by addition of a few drops of conc. ammonia
solution. The solvent was removed under reduced pressure, and the syrup
was extracted with chloroform/water. The collected organic phases were
dried over MgSO4, filtered, and evaporated to dryness, then purified by
column chromatography (chloroform/acetonitrile/methanol 60:24:1.
15: Rf ˆ 0.50 (chloroform/acetonitrile/methanol 20:8:1); 1H NMR (270
MHz, [D6]DMSO, 258C): d ˆ 2.94 (s, 3H, N(CH3)), 3.00 (d, J ˆ 6.1 Hz, 2H,
allyl CH2), 3.03 (s, 3H, N(CH3)), 3.11 (m, 2H, 5'-H), 3.32 (s, 3H, 2'-OCH3),
3.71 (s, 3H, OCH3), 3.72 (s, 3H, OCH3), 3.83 (m, 1H, 4'-H), 4.07 (dd, J ˆ 3.7,
2.5 Hz, 1H, 2'-H), 4.28 (m, 1H, 3'-H), 4.76 (d, J ˆ 8.2 Hz, 1H, 3'-OH), 4.88
(ddd, J ˆ 17.1, 10.0, 2.1 Hz, 2H, allyl CH2), 5.73 (m, 1H, allyl CH), 6.18 (brs,
1H, 1'-H), 6.81 ± 7.40 (m, 13H, Har), 7.97 (s, 1H, NCHN), 10.61 (brs, 1H,
NH); the assignment was confirmed by a 1H,1H-COSY experiment;
13C NMR (68 MHz, [D6]DMSO, 258C): d ˆ 28.09 (allyl CH2), 33.71
(N(CH3)), 39.98 (N(CH3)), 54.93 (OCH3), 57.72 (2'-OCH3), 64.64 (C-5'),
70.43 (C-3'), 81.44 (C-2'), 82.02 (C-4'), 85.24 (DMTr), 85.98 (C-1'), 95.24 (C-
5), 113.03 (DMTr), 113.88 (allyl CH2), 126.47, 127.74, 129.64, 129.76, 135.65
(DMTr), 135.77 (allyl CH), 137.13, 145.04 (DMTr), 153.42 (C-2), 155.43
(NCHN), 157.90 (C-4), 157.95 (DMTr), 162.96 (C-6); the assignment was
confirmed by an HSQC experiment; MS ESI(À): m/z (%): 669.4 (100)
[M À H]À.
17: Rf ˆ 0.41 (chloroform/acetonitrile/methanol 20:8:1); 1H NMR (250
MHz, [D6]DMSO, 258C): d ˆ 2.94 (s, 3H, N(CH3)), 3.00 (d, J ˆ 4.9 Hz,
2H, allyl CH2), 3.03 (s, 3H, N(CH3)), 3.11 (m, 2H, 5'-H), 3.25 (s, 3H, 3'-
OCH3), 3.71 (s, 3H, OCH3), 3.72 (s, 3H, OCH3), 3.93 (m, 1H, 4'-H), 3.99
(dd, J ˆ 7.2, 6.0 Hz, 1H, 3'-H), 4.56 (dt, J ˆ 6.2 Hz, 2.5 Hz, 1H, 2'-H), 4.88
(ddd, J ˆ 17.1, 9.9, 1.5 Hz, 2H, allyl CH2), 5.00 (d, J ˆ 6.6 Hz, 1H, 2'-OH),
5.73 (m, 1H, allyl CH), 6.15 (brs, 1H, 1'-H), 6.82 ± 7.39 (m, 13H, Har), 7.96
(s, 1H, NCHN), 10.56 (brs, 1H, NH); the assignment was confirmed by a
1H,1H-COSY experiment; 13C NMR (63 MHz, [D6]DMSO, 258C): d ˆ
28.09 (allyl CH2), 33.72 (N(CH3)), 39.98 (N(CH3)), 54.93 (OCH3), 54.95
(OCH3), 57.34 (3'-OCH3), 64.50 (C-5'), 70.45 (C-2'), 79.94 (C-3'), 80.06 (C-
4'), 85.22 (DMTr), 88.20 (C-1'), 95.30 (C-5), 113.06 (DMTr), 113.86 (allyl
CH2), 126.53, 127.73, 129.61, 129.73, 135.62 (DMTr), 135.78 (allyl CH),
137.18, 145.02 (DMTr), 153.45 (C-2), 155.41 (NCHN), 157.93 (C-4), 157.98
(DMTr), 163.13 (C-6); the assignment was confirmed by an HSQC
experiment; MS ESI(À): m/z (%): 669.5 (100) [M À H]À.
3'-O-(2-Cyanoethoxy-N,N-diisopropylphosphino)-5'-O-(4,4'-dimethoxytri-
phenylmethyl)-4-N-(dimethylformamidine)-2'-O-methyl-6-oxocytidine
(19): N,N-diisopropylethylamine (DIPEA) (0.64 mL, 3.71 mmol, 4.5 equiv)
was added to a solution of 14 (520 mg, 0.824 mmol) in acetonitrile (10 mL).
The reaction mixture was cooled with an ice bath, and 18 (280 mL,
1.24 mmol, 1.5 equiv) was added slowly. After 10 min the ice bath was
removed and the solution was stirred for another 20 min. The mixture was
diluted with dichloromethane and extracted with 2% aqueous sodium
carbonate solution. The collected organic phases were dried over MgSO4,
filtered, and evaporated to dryness. The crude product was chromato-
graphed with chloroform/acetonitrile/methanol 60:24:1 to afford 19 as a
white foam (397 mg, 58%). Compound 19 was obtained as a mixture of two
diastereomers with identical Rf values in a ratio of 1.7:1. The underlined
NMR spectroscopic data refer to the main diastereomer. Rf ˆ 0.32 (ethyl
acetate/acetonitrile 9:1 ‡ 5% triethylamine), Rf ˆ 0.48 (chloroform/aceto-
nitrile/methanol 20:8:1), 31P NMR (163 MHz, CDCl3, 258C): d ˆ 149.57,
150.54; 1H NMR (400 MHz, CDCl3, 258C): d ˆ 0.94 ± 1.15 (m, 12H,
isopropyl CH3), 2.35, 2.65 (2m, 2H, CH2CN), 3.02, 3.02 (2s, 3H,
N(CH3)), 3.10, 3.10 (2s, 3H, N(CH3)), 3.13 ± 3.60 (m, 4H, 5'-H and
POCH2), 3.39, 3.41 (2s, 3H, 2'-OCH3), 3.76, 3.76, 3.77, 3.77 (4s, 3H, OCH3),
3.80 ± 3.93 (m, 2H, isopropyl CH), 4.16 (m, 1H, 4'-H), 4.25, 4.32 (2dd, J ˆ
6.2, 2.5 Hz, J ˆ 6.1, 2.8 Hz, 1H, 2'-H), 4.52, 4.74 (2ddd, J ˆ 10.5, 6.2, 7.5 Hz,
J ˆ 12.1, 6.3, 8.0 Hz, 1H, 3'-H), 5.02, 5.03 (2s, 1H, 5-H), 6.39, 6.42 (2d, J ˆ
2.4 Hz, J ˆ 2.5 Hz, 1H, 1'-H), 6.74 ± 7.48 (m, 13H, Har), 7.78 (s, 1H, NCHN);
MS ESI(À): m/z (%): 829.1 (100) [M À H]À.
5'-O-(4,4'-Dimethoxytriphenylmethyl)-4-N-(dimethylformamidine)-2'-O-
methyl-6-oxocytidine (14) and 5'-O-(4,4'-dimethoxytriphenylmethyl)-4-N-
(dimethylformamidine)-3'-O-methyl-6-oxocytidine (16):
A solution of
tin(ii) chloride (0.27 g, 1.42 mmol, 0.25 equiv) in DMF (10 mL) was added
to a solution of nucleoside 12 (3.50 g, 5.68 mmol) in DMF (60 mL). The
reaction flask was cooled in an ice bath. A solution of diazomethane in
ether (14 mL, approx. 0.3m) was added slowly. (Diazomethane was freshly
prepared with the DiazaldTM kit diazomethane generator from Aldrich.)
Further portions of diazomethane (14 mL) were added every 45 min. After
4 h the reaction was stopped by addition of a few drops of conc. ammonia
solution. The solvent was removed under reduced pressure, and the syrup
was extracted with chloroform/water. The collected organic phases were
dried over MgSO4, filtered, and evaporated to dryness, then purified by
column chromatography (chloroform/acetonitrile/methanol 60:24:1.
14: Rf ˆ 0.36 (chloroform/acetonitrile/methanol 20:8:1); 1H NMR (270
MHz, [D6]DMSO, 258C): d ˆ 2.96 (s, 3H, N(CH3)), 3.08 (s, 3H, N(CH3)),
3.09 (m, 2H, 5'-H), 3.31 (s, 3H, 2'-OCH3), 3.71 (s, 3H, OCH3), 3.72 (s, 3H,
OCH3), 3.82 (m, 1H, 4'-H), 4.03 (dd, J ˆ 6.2, 2.5 Hz,1H, 2'-H), 4.24 (m, 1H,
3'-H), 4.78 (d, J ˆ 8.2 Hz, 1H, 3'-OH), 5.07 (s, 1H, 5-H), 6.13 (brs, 1H, 1'-
H), 6.81 ± 7.40 (m, 13H, Har), 8.14 (s, 1H, NCHN), 10.77 (brs, 1H, NH);
13C NMR (63 MHz, [D6]DMSO, 258C): d ˆ 34.31 (N(CH3)), 40.42
(N(CH3)), 54.94 (OCH3), 57.69 (2'-OCH3), 64.55 (C-5'), 70.39 (C-3'),
80.90 (C-5), 81.40 (C-2'), 81.85 (C-4'), 85.25 (DMTr), 85.30 (C-1'), 113.05,
126.50, 127.70, 129.65, 129.77, 135.72, 145.04 (DMTr), 150.80 (C-2), 157.02
(NCHN), 157.95 (DMTr), 159.26 (C-4), 163.07 (C-6); the assignment was
confirmed by an HSQC experiment; MS ESI(À): m/z (%): 629.2 (100)
[M À H]À.
16: Rf ˆ 0.28 (chloroform/acetonitrile/methanol 20:8:1); 1H NMR (270
MHz, [D6]DMSO, 258C): d ˆ 2.96 (s, 3H, N(CH3)), 3.09 (s, 3H, N(CH3)),
3.10 (m, 2H, 5'-H), 3.25 (s, 3H, 3'-OCH3), 3.72 (s, 3H, OCH3), 3.73 (s, 3H,
OCH3), 3.91 ± 3.99 (m, 2H, 3'-H, 4'-H), 4.54 (dt, J ˆ 2.6, 5.9 Hz, 1H, 2'-H),
5.02 (d, J ˆ 6.6 Hz, 1H, 2'-OH), 5.06 (s, 1H, 5-H), 6.10 (d, J ˆ 1.5 Hz, 1H, 1'-
H), 6.82 ± 7.40 (m, 13H, Har), 8.15 (s, 1H, NCHN), 10.74 (brs, 1H, NH);
13C NMR (63 MHz, [D6]DMSO, 258C): d ˆ 34.34 (N(CH3)), 40.49
(N(CH3)), 54.98 (OCH3), 57.35 (3'-OCH3), 64.51 (C-5'), 70.45 (C-2'),
79.87 (C-3'), 79.97 (C-4'), 80.90 (C-5), 85.27 (DMTr), 88.10 (C-1'), 113.11,
126.59, 127.76, 129.66, 129.79, 135.80, 145.05 (DMTr), 151.00 (C-2), 157.02
(NCHN), 158.02 (DMTr), 159.33 (C-4), 163.31 (C-6); the assignment was
confirmed by an HSQC experiment; MS ESI(À): m/z (%): 629.2 (100)
[M À H]À.
Chem. Eur. J. 2000, 6, No. 13
ꢀ WILEY-VCH Verlag GmbH, D-69451 Weinheim, 2000
0947-6539/00/0613-2419 $ 17.50+.50/0
2419

U. Parsch, J. W. Engels
FULL PAPER
5-Allyl-3'-O-(2-cyanoethoxy-N,N-diisopropylphosphino)-5'-O-(4,4'-dimeth-
oxytriphenylmethyl)-4-N-(dimethylformamidine)-2'-O-methyl-6-oxocyti-
dine (20): N,N-Diisopropylethylamine (DIPEA) (0.37 mL, 2.16 mmol,
4.5 equiv) was added to a solution of 15 (322 mg, 0.48 mmol) in acetonitrile
(8 mL). The reaction mixture was cooled with an ice bath and 18 (160 mL,
0.72 mmol, 1.5 equiv) was slowly added. After 10 min the ice bath was
removed and the solution was stirred for another 2 h. The mixture was
diluted with dichloromethane and extracted with 2% aqueous sodium
carbonate solution. The collected organic phases were dried over MgSO4,
filtered, and evaporated to dryness. The crude product was chromato-
graphed with chloroform/acetonitrile/methanol 60:24:1 to afford 20 as a
white foam (297 mg, 71%). Compound 20 was obtained as a mixture of two
diastereomers with identical Rf values in a ratio of 2:1. The underlined
NMR spectroscopic data refer to the main diastereomer. Rf ˆ 0.67 (chloro-
form/acetonitrile/methanol 20:8:1); 31P NMR (163 MHz, CDCl3, 258C):
d ˆ 150.10, 150.58; 1H NMR (400 MHz, CDCl3, 258C): d ˆ 0.95 ± 1.17 (m,
12H, isopropyl CH3), 2.34, 2.65 (2m, 2H, CH2CN), 3.02 (s, 3H, N(CH3)),
3.07 (d, J ˆ 6.0 Hz, 2H, allyl CH2), 3.09 (s, 3H, N(CH3)), 3.11 (m, 2H, 5'-H),
3.25 ± 3.61 (m, 2H, POCH2), 3.39, 3.42 (2s, 3H, 2'-OCH3), 3.76, 3.76, 3.77,
3.77 (4s, 3H, OCH3), 3.79 ± 3.93 (m, 2H, isopropyl CH), 4.13 (m, 1H, 4'-H),
4.28, 4.34 (2dd, J ˆ 6.3, 2.5 Hz, J ˆ 6.2, 2.7 Hz, 1H, 2'-H), 4.59, 4.78 (2ddd,
J ˆ 10.6, 6.4, 7.6 Hz, J ˆ 12.0, 6.4, 8.0 Hz, 1H, 3'-H), 4,96 (m, 2H, allyl CH2),
5.92 (m, 1H, allyl CH), 6.37, 6.39 (2s, 1H, 1'-H), 6.74 ± 7.47 (m, 13H, Har),
(N(CH3)), 62.27 (C-5'), 71.28 (C-3'), 72.82 (C-2'), 80.37 (C-4'), 87.77 (C-
1'), 95.10 (C-5), 113.83 (allyl CH2), 137.06 (allyl CH), 150.32 (C-2), 153.19
(C-4), 155.48 (NCHN), 162.96 (C-6); MS ESI(‡): m/z (%): 597.4 (100)
‡
[M‡H] .
4-N-(Dimethylformamidine)-2'-O-phenoxythiocarbonyl-3',5'-O-(1,1,3,3-
tetraisopropyldisiloxane-1,3-diyl)-6-oxocytidine (23): Nucleoside 21
(6.79 g, 12.19 mmol) was dissolved in acetonitrile (100 mL), and N,N-
dimethylaminopyridine (DMAP) (3.13 g, 25.61 mmol, 2.1 equiv) and
phenoxythiocarbonyl chloride (1.86 mL, 13.41 mmol, 1.1 equiv) were
added. After 2 h the solvent was removed under reduced pressure. The
residue was dissolved in ethyl acetate and extracted with 0.1m HCl, water,
sat. aqueous NaHCO3 solution, and brine. The collected organic phases
were dried over MgSO4, filtered and evaporated to dryness. The residue
was purified by column chromatography with dichloromethane/methanol
98:2. The product 23 (4.90 g, 58%) was obtained as a yellow foam. Rf ˆ 0.53
(dichloromethane/methanol 95:5); 1H NMR (400 MHz, [D6]DMSO, 258C):
d ˆ 0.93 ± 1.07 (m, 28H, isopropyl), 2.96 (s, 3H, N(CH3)), 3.09 (s, 3H,
N(CH3)), 3.72 (m, 1H, 4'-H), 3.95 (m, 2H, 5'-H), 5.11 (s, 1H, 5-H), 5.16 (m,
1H, 3'-H), 6.21 (brs, 1H, 1'-H), 6.29 (dd, J ˆ 5.9, 1.5 Hz, 1H, 2'-H), 7.11 ±
7.50 (m, 5H, Har), 8.15 (s, 1H, NCHN), 10.93 (brs, 1H, NH); the assignment
was confirmed by
a
1H,1H-COSY experiment; 13C NMR (100 MHz,
[D6]DMSO, 258C): d ˆ 12.25, 12.36, 12.58, 12.86 (4  CH), 16.67 ± 17.24
(8 Â CH3), 34.33 (N(CH3)), 40.35 (N(CH3)), 61.32 (C-5'), 70.24 (C-3'), 80.33
(C-4'), 80.99 (C-5), 84.26 (C-1'), 85.00 (C-2'), 121.51, 126.68, 129.28, 129.73
‡
7.81 (s, 1H, NCHN); MS ESI(‡): m/z (%): 871.5 (78.2) [M‡H] .
ˆ
(Car), 152.78 (C-2), 157.21 (NCHN), 157.88 (C S), 159.61 (C-4), 163.18 (C-
4-N-(Dimethylformamidine)-3',5'-O-(1,1,3,3-tetraisopropyldisiloxane-1,3-
diyl)-6-oxocytidine (21): A solution of 1,3-dichloro-1,1,3,3-tetraisopropyl-
disiloxane (4.56 mL, 14.55 mmol, 1.11 equiv) in 1,2-dichloroethane
(3.30 mL) was added to a solution of 10 (4.12 g, 13.11 mmol) in pyridine
(80 mL). After stirring at 08C for 30 min, pyridinium hydrochloride
precipitated. The reaction mixture was stirred for another 1.5 h at RT,
methanol was added, and the solution was evaporated to dryness. The
residue was dissolved in dichloromethane and extracted with saturated
aqueous NaHCO3 solution. The organic phase was dried over MgSO4,
filtered, and evaporated, followed by coevaporation with toluene. The
product was purified by column chromatography with dichloromethane/
methanol 95:5. Compound 21 (6.79 g, 93%) was obtained as a white foam.
Rf ˆ 0.58 (dichloromethane/methanol 9:1); 1H NMR (400 MHz,
[D6]DMSO, 258C): d ˆ 0.92 ± 1.07 (m, 28H, isopropyl), 2.95 (s, 3H,
N(CH3)), 3.09 (s, 3H, N(CH3)), 3.67 (m, 1H, 4'-H), 3.88 (m, 2H, 5'-H),
4.38 (t, J ˆ 4.4 Hz, 1H, 2'-H), 4.68 (d, J ˆ 4.2 Hz, 1H, 2'-OH), 4.86 (m, 1H,
3'-H), 5.05 (s, 1H, 5-H), 6.01 (s, 1H, 1'-H), 8.13 (s, 1H, NCHN), 10.73 (brs,
1H, NH); the assignment was confirmed by a 1H,1H-COSY experiment;
13C NMR (100 MHz, [D6]DMSO, 258C): d ˆ 12.06, 12.12, 12.50, 12.66 (4 Â
CH), 16.86 ± 17.35 (8 Â CH3), 34.28 (N(CH3)), 40.29 (N(CH3)), 62.42 (C-5'),
71.44 (C-3'), 72.81 (C-2'), 80.34 (C-4'), 80.95 (C-5), 87.55 (C-1'), 149.56 (C-
2), 156.97 (NCHN), 159.20 (C-4), 163.09 (C-6); the assignment was
confirmed by an HSQC experiment; elemental analysis calcd (%) for
C24H44N4O7Si2 (556.81): C 51.77, H 7.97, N 10.06; found C 51.52, H 7.90, N
9.81; MS ESI(À): m/z (%): 555.5 (100) [M À H]À, ESI(‡): m/z (%): 557.6
6); the assignment was confirmed by an HSQC experiment; MS ESI(‡): m/
‡
‡
z (%): 693.2 (32.2) [M‡H] , 539.3 (100) [M À OCSOC6H5] .
5-Allyl-4-N-(dimethylformamidine)-2'-O-phenoxythiocarbonyl-3',5'-O-
(1,1,3,3-tetraisopropyldisiloxane-1,3-diyl)-6-oxocytidine (24): Nucleoside
22 (2.57 g, 4.31 mmol) was dissolved in acetonitrile (40 mL), and N,N-
dimethylaminopyridine (DMAP) (1.10 g, 9.04 mmol, 2.1 equiv) and phe-
noxythiocarbonyl chloride (0.66 mL, 4.74 mmol, 1.1 equiv) were added.
After 4 h the solvent was removed under reduced pressure. The residue was
dissolved in ethyl acetate and extracted with 0.1m HCl, water, sat. aqueous
NaHCO3 solution, and brine. The collected organic phases were dried over
MgSO4, filtered, and evaporated to dryness. The residue was purified by
column chromatography with dichloromethane/methanol 99:1. The prod-
uct 24 (1.63 g, 52%) was obtained as a yellow foam. Rf ˆ 0.33 (dichloro-
methane/methanol 98:2); 1H NMR (400 MHz, [D6]DMSO, 258C): d ˆ
0.96 ± 1.09 (m, 28H, isopropyl), 2.94 (s, 3H, N(CH3)), 3.00 (d, 2H, allyl
CH2), 3.03 (s, 3H, N(CH3)), 3.73 (m, 1H, 4'-H), 3.94 (m, 2H, 5'-H), 4.87
(ddd, J ˆ 17.1, 10.1, 1.5 Hz, 2H, allyl CH2), 5.21 (d, J ˆ 6.0 Hz, 1H, 3'-H),
5.76 (m, 1H, allyl CH), 6.24 (s, 1H, 1'-H), 6.30 (d, J ˆ 5.9 Hz, 1H, 2'-H),
7.12 ± 7.48 (m, 5H, Har), 8.02 (s, 1H, NCHN), 10.82 (brs, 1H, NH); the
assignment was confirmed by
a
1H,1H-COSY experiment; 13C NMR
(100 MHz, [D6]DMSO, 258C): d ˆ 12.25, 12.36, 12.58, 12.62 (4  CH),
16.74 ± 17.25 (8 Â CH3), 28.05 (allyl CH2), 33.74 (N(CH3)), 40.13 (N(CH3)),
61.13 (C-5'), 70.07 (C-3'), 80.12 (C-4'), 83.21 (C-1'), 84.32 (C-2'), 95.22 (C-5),
113.98 (allyl CH2), 121.52, 126.68, 129.28, 129.73 (Car), 136.87 (allyl CH),
‡
(80.0) [M‡H] .
ˆ
152.77 (C-2), 153.61 (C-4), 155.69 (NCHN), 157.27 (C S), 162.89 (C-6); the
assignment was confirmed by an HSQC experiment; MS ESI(‡): m/z (%):
5-Allyl-4-N-(dimethylformamidine)-3',5'-O-(1,1,3,3-tetraisopropyldisilox-
ane-1,3-diyl)-6-oxocytidine (22): A solution of 1,3-dichloro-1,1,3,3-tetra-
isopropyldisiloxane (1.75 mL, 5.59 mmol, 1.1 equiv) in 1,2-dichloroethane
(1.20 mL) was added to a solution of 11 (1.80 g, 5.08 mmol) in pyridine
(25 mL). After the reaction mixture had been stirred at 08C for 30 min,
pyridinium hydrochloride precipitated. The mixture was stirred for another
2.5 h at RT, methanol was added, and the solution was evaporated to
dryness. The residue was dissolved in dichloromethane and extracted with
saturated aqueous NaHCO3 solution. The organic phase was dried over
MgSO4, filtered and evaporated, followed by coevaporation with toluene.
The product was purified by column chromatography with dichloro-
methane/methanol 98:2. Compound 22 (2.66 g, 88%) was obtained as a
white foam. Rf ˆ 0.66 (dichloromethane/methanol 95:5), Rf ˆ 0.15 (di-
chloromethane/methanol 98:2); 1H NMR (250 MHz, [D6]DMSO, 258C):
d ˆ 0.92 ± 1.07 (m, 28H, isopropyl), 2.93 (s, 3H, N(CH3)), 2.99 (d, 2 H allyl
CH2), 3.02 (s, 3H, N(CH3)), 3.68 (m, 1H, 4'-H), 3.88 (m, 2H, 5'-H), 4.39 (t,
J ˆ 5.1 Hz, 1H, 2'-H), 4.70 (d, J ˆ 4.2 Hz, 1H, 2'-OH), 4.87 (ddd, J ˆ 17.1,
10.1, 1.7 Hz, 2H, allyl CH2), 4.95 (d, J ˆ 6.3 Hz, 1H, 3'-H), 5.73 (m, 1H, allyl
CH), 6.06 (s, 1H, 1'-H), 7.99 (s, 1H, NCHN), 10.59 (brs, 1H, NH); 13C NMR
(63 MHz, [D6]DMSO, 258C): d ˆ 12.02, 12.09, 12.47, 12.63 (4  CH),
16.82 ± 17.34 (8 Â CH3), 28.09 (allyl CH2), 33.69 (N(CH3)), 39.95
‡
‡
733.4 (38.1) [M‡H] , 579.4 (100) [M À OCSOC6H5] .
2'-Deoxy-4-N-(dimethylformamidine)-3',5'-O-(1,1,3,3-tetraisopropyldisil-
oxane-1,3-diyl)-6-oxocytidine (25): Nucleoside 23 (4.90 g, 7.07 mmol) was
dissolved in freshly distilled toluene (160 mL). The solution was degassed
five times. a,a'-Azoisobutyronitrile (0.23 g, 1.41 mmol, 0.2 equiv) and
tributyltin hydride (2.81 mL, 10.61 mmol, 1.5 equiv) were added and the
solution was heated to 758C. After 2 h the solution was cooled to RT and
the solvent was removed by evaporation. The residue was dissolved in ethyl
acetate and extracted with 5% aqueous NaHCO3 solution. The organic
phase was dried over MgSO4, filtered, and evaporated to dryness. The
residue was purified by column chromatography with dichloromethane/
methanol 98:2. The product 25 (2.49 g, 65%) was obtained as a yellow
foam. Rf ˆ 0.42 (dichloromethane/methanol 95:5); 1H NMR (400 MHz,
[D6]DMSO, 258C): d ˆ 0.95 ± 1.10 (m, 28H, isopropyl), 2.25 (m, 1H, 2'a-H),
2.63 (m, 1H, 2'b-H), 2.94 (s, 3H, N(CH3)), 3.07 (s, 3H, N(CH3)), 3.63 (m,
1H, 4'-H), 3.88 (m, 2H, 5'-H), 5.01 (s, 1H, 5-H), 5.04 (m, 1H, 3'-H), 6.49
(dd, J ˆ 9.8, 2.6 Hz, 1H, 1'-H), 8.10 (s, 1H, NCHN), 10.66 (brs, 1H, NH);
the assignment was confirmed by a 1H,1H-COSY experiment; 13C NMR
(100 MHz, [D6]DMSO, 258C): d ˆ 11.98, 12.54, 12.66, 12.86 (4  CH),
16.79 ± 17.42 (8 Â CH3), 34.24 (N(CH3)), 38.87 (C-2'), 40.25 (N(CH3)), 64.27
2420
ꢀ WILEY-VCH Verlag GmbH, D-69451 Weinheim, 2000
0947-6539/00/0613-2420 $ 17.50+.50/0
Chem. Eur. J. 2000, 6, No. 13

Triple-Helix Formation
2409±2424
(C-5'), 74.19 (C-3'), 78.68 (C-1'), 81.19 (C-5), 84.61 (C-4'), 150.83 (C-2),
2'-Deoxy-5'-O-(4,4'-dimethoxytriphenylmethyl)-4-N-(dimethylformamidine)-
6-oxocytidine (29): Nucleoside 27 (600 mg, 2.01 mmol) was dissolved in dry
pyridine (30 mL). The solution was cooled to À408C with acetone/liquid
nitrogen. DMTrCl (750 mg, 2.21 mmol, 1.1 equiv) was added. After 3 h the
reaction mixture has reached RT and further DMTrCl (140 mg, 0.2 equiv)
was added. After 8 h dichloromethane (20 mL) was added and the solution
was extracted with 5% aqueous NaHCO3 solution. The collected organic
phases were dried over MgSO4, filtered, and evaporated to dryness. The
residue was coevaporated with toluene. Purification was achieved by
column chromatography with dichloromethane/methanol 95:5. The prod-
uct 29 (955 mg, 79%) was obtained as a white foam. Rf ˆ 0.43 (dichloro-
methane/methanol 9:1); 1H NMR (250 MHz, [D6]DMSO, 258C): d ˆ 1.97
(m, 1H, 2'a-H), 2.58 (m, 1H, 2'b-H), 2.95 (s, 3H, N(CH3)), 3.06 (m, 1H, 5'a-
H), 3.08 (s, 3H, N(CH3)), 3.22 (m, 1H, 5'b-H), 3.71 (s, 3H, OCH3), 3.72 (s,
3H, OCH3), 3.79 (m, 1H, 4'-H), 4.26 (m, 1H, 3'-H), 4.99 (d, J ˆ 5.5 Hz, 1H,
3'-OH), 5.03 (s, 1H, 5-H), 6.58 (dd, J ˆ 8.9, 4.2 Hz, 1H, 1'-H), 6.80 ± 7.40 (m,
13H, Har), 8.12 (s, 1H, NCHN), 10.60 (brs, 1H, NH); 13C NMR (63 MHz,
[D6]DMSO, 258C): d ˆ 34.24 (N(CH3)), 37.39 (C-2'), 40.50 (N(CH3)), 54.89
(OCH3), 54.93 (OCH3), 64.92 (C-5'), 71.50 (C-3'), 79.83 (C-1'), 81.14 (C-5),
85.16 (DMTr), 85.29 (C-4'), 112.99, 126.46, 127.67, 129.58, 129.80, 135.82,
145.24 (DMTr), 150.73 (C-2), 156.83 (NCHN), 157.30 (DMTr), 159.01 (C-4),
163.45 (C-6); MS ESI(À): m/z (%): 599.4 (100) [M À H]À.
156.89 (NCHN), 159.00 (C-4), 163.20 (C-6); the assignment was confirmed
‡
by an HSQC experiment; MS ESI(‡): m/z (%): 541.4 (100) [M‡H] .
5-Allyl-2'-deoxy-4-N-(dimethylformamidine)-3',5'-O-(1,1,3,3-tetraisopro-
pyldisiloxane-1,3-diyl)-6-oxocytidine (26): Nucleoside 24 (1.49 g,
2.03 mmol) was dissolved in freshly distilled toluene (50 mL). The solution
was degassed five times. a,a'-Azoisobutyronitrile (67 mg, 0.407 mmol,
0.2 equiv) and tributyltin hydride (0.81 mL, 3.05 mmol, 1.5 equiv) were
added and the solution was heated to 758C. After 1.5 h the solution was
cooled to RTand the solvent was removed by evaporation. The residue was
dissolved in ethyl acetate and extracted with 5% aqueous NaHCO3
solution. The organic phase was dried over MgSO4, filtered, and
evaporated to dryness. The residue was purified by column chromatog-
raphy with dichloromethane/methanol 98:2. The product 26 (0.68 g, 58%)
was obtained as a yellow foam. Rf ˆ 0.46 (dichloromethane/methanol 95:5);
1H NMR (400 MHz, [D6]DMSO, 258C): d ˆ 0.94 ± 1.11 (m, 28H, isopropyl),
2.27 (m, 1H, 2'a-H), 2.64 (m, 1H, 2'b-H), 2.93 (s, 3H, N(CH3)), 3.00 (d, J ˆ
6.3 Hz, 2H, allyl CH2), 3.02 (s, 3H, N(CH3)), 3.64 (m, 1H, 4'-H), 3.85 (dd,
J ˆ 11.1, 3.6 Hz, 1H, 5'a-H), 3.93 (dd, J ˆ 11.0, 8.5 Hz, 1H, 5'b-H), 4.85
(ddd, J ˆ 17.1, 10.0, 2.3 Hz, 2H, allyl CH2), 5.09 (yq, J ˆ 7.5 Hz, 1H, 3'-H),
5.72 (m, 1H, allyl CH), 6.56 (dd, J ˆ 9.8, 2.5 Hz, 1H, 1'-H), 7.98 (s, 1H,
NCHN), 10.54 (s, 1H, NH); the assignment was confirmed by a 1H,1H-
COSY experiment; 13C NMR (100 MHz, [D6]DMSO, 258C): d ˆ 11.98,
12.55, 12.66, 12.86 (4 Â CH), 16.76 ± 17.41 (8 Â CH3), 28.08 (allyl CH2), 33.66
(N(CH3)), 38.88 (C-2'), 40.13 (N(CH3)), 64.24 (C-5'), 74.15 (C-3'), 79.08 (C-
1'), 84.71 (C-4'), 95.25 (C-5), 113.76 (allyl CH2), 137.11 (allyl CH), 150.15
(C-2), 153.02 (C-4), 155.45 (NCHN), 163.11 (C-6); the assignment was
confirmed by an HSQC experiment; MS ESI(‡): m/z (%): 581.5 (100)
5-Allyl-2'-deoxy-5'-O-(4,4'-dimethoxytriphenylmethyl)-4-N-(dimethylfor-
mamidine)-6-oxocytidine (30): Nucleoside 28 (210 mg, 0.62 mmol) was
dissolved in dry pyridine (8 mL). The solution was cooled to 08C in an ice
bath. DMTrCl (252 mg, 0.74 mmol, 1.2 equiv) was added. The ice bath was
removed after 1 h. After further 4 h, more DMTrCl (21 mg, 0.1 equiv) was
added. Another 4 h later dichloromethane (10 mL) was added and the
solution was extracted with 5% aqueous NaHCO3 solution. The collected
organic phases were dried over MgSO4, filtered, and evaporated to dryness.
The residue was coevaporated with toluene, and purified by column
chromatography with dichloromethane/methanol 95:5. The product 30
(340 mg, 86%) was obtained as a white foam. Rf ˆ 0.57 (dichloromethane/
methanol 9:1); 1H NMR (400 MHz, [D6]DMSO, 258C): d ˆ 2.00 (m, 1H,
2'a-H), 2.60 (m, 1H, 2'b-H), 2.94 (s, 3H, N(CH3)), 3.00 (d, J ˆ 6.3 Hz, 2H,
allyl CH2), 3.02 (s, 3H, N(CH3)), 3.05 (m, 1H, 5'a-H), 3.25 (m, 1H, 5'b-H),
3.70 (s, 3H, OCH3), 3.72 (s, 3H, OCH3), 3.82 (m, 1H, 4'-H), 4.29 (m, 1H, 3'-
H), 4.86 (ddd, J ˆ 17.1, 10.0, 2.2 Hz, 2H, allyl CH2), 5.00 (d, J ˆ 5.5 Hz, 1H,
3'-OH), 5.73 (m, 2H, allyl CH), 6.64 (dd, J ˆ 8.9, 4.3 Hz, 1H, 1'-H), 6.78 ±
7.39 (m, 13H, Har), 7.94 (s, 1H, NCHN), 10.48 (s, 1H, NH); the assignment
‡
[M‡H] .
General procedure for the deprotection of 25 and 26: The protected
nucleosides 25 and 26 were each dissolved in THF (30 mL). A 1m solution
of tetrabutylammonium fluoride (TBAF) in THF was added. After 1 h at
RT, full deprotection was achieved. The solvent was removed and the oily
residue was dissolved in water. This was extracted with dichloromethane
(three times) and ether (two times). The aqueous phase was evaporated to
dryness and coevaporated with methanol. The product was purified by
column chromatography with dichloromethane/methanol 9:1, to yield 27 or
28 as a white foam.
2'-Deoxy-4-N-(dimethylformamidine)-6-oxocytidine (27): Protected nu-
cleoside 25 (2.49 g, 4.60 mmol) was dissolved in THF (30 mL), and TBAF
(10.13 mL, 10.13 mmol) was added. The product 27 was obtained in 1.35 g
(99%) yield. Rf ˆ 0.19 (dichloromethane/methanol 9:1); 1H NMR
(400 MHz, [D6]DMSO, 258C): d ˆ 1.84 (m, 1H, 2'a-H), 2.70 (m, 1H, 2'b-
H), 2.95 (s, 3H, N(CH3)), 3.08 (s, 3H, N(CH3)), 3.43 (m, 1H, 5'a-H), 3.57
(m, 1H, 5'b-H), 3.65 (m, 1H, 4'-H), 4.29 (m, 1H, 3'-H), 4.57 (t, J ˆ 4.2 Hz,
1H, 5'-OH), 5.01 (d, J ˆ 4.7 Hz, 1H, 3'-OH), 5.03 (s, 1H, 5-H), 6.52 (yt, J ˆ
7.3 Hz, 1H, 1'-H), 8.11 (s, 1H, NCHN), 10.67 (brs, 1H, NH); the
was confirmed by
a
1H,1H-COSY experiment; 13C NMR (100 MHz,
[D6]DMSO, 258C): d ˆ 28.11 (allyl CH2), 33.67 (N(CH3)), 37.55 (C-2'),
40.12 (N(CH3)), 54.89 (OCH3), 54.93 (OCH3), 64.99 (C-5'), 71.60 (C-3'),
80.42 (C-1'), 85.15 (DMTr), 85.56 (C-4'), 95.33 (C-5), 112.97 (DMTr), 113.76
(allyl CH2), 126.44, 127.73, 129.56, 129.81, 135.72 (DMTr), 137.23 (allyl CH),
145.27 (DMTr), 150.17 (C-2), 153.10 (C-4), 155.31 (NCHN), 157.95 (DMTr),
163.32 (C-6); the assignment was confirmed by an HSQC experiment; MS
ESI(À): m/z (%): 639.4 (100) [M À H]À.
assignment was confirmed by
a
1H,1H-COSY experiment; 13C NMR
3'-O-(2-Cyanoethoxy-N,N-diisopropylphosphino)-2'-deoxy-5'-O-(4,4'-di-
methoxytriphenylmethyl)-4-N-(dimethylformamidine)-6-oxocytidine (31):
N,N-Diisopropylethylamine (DIPEA) (449 mL, 2.62 mmol, 5 equiv) and
2-cyanoethyl-N,N-diisopropyl chlorophosphoramidite (18; 234 mL,
1.05 mmol, 2 equiv) were added to a solution of 29 (315 mg, 0.524 mmol)
in dichloromethane (12 mL). After 30 min the reaction mixture was diluted
with dichloromethane and extracted with 2% aqueous sodium carbonate
solution. The collected organic phases were dried over MgSO4, filtered,
and evaporated to dryness. The crude product was chromatographed with
dichloromethane/methanol 95:5 to afford 31 as a white foam (315 mg,
75%). Compound 31 was obtained as a mixture of two diastereomers with
different Rf values in a ratio of 1:1.2. The underlined NMR spectroscopic
data refer to the main diastereomer. Rf ˆ 0.33/0.38 (dichloromethane/
methanol 95:5); 31P NMR (163 MHz, CDCl3, 258C): d ˆ 148.80, 148.98;
1H NMR (400 MHz, CDCl3, 258C): d ˆ 0.98 ± 1.15 (m, 12H, isopropyl
CH3), 2.23 (m, 1H, 2'a-H), 2.37, 2.57 (2m, 2H, CH2CN), 2.84 (m, 1H, 2'b-
H), 3.01 (s, 3H, N(CH3)), 3.10, (s, 3H, N(CH3)), 3.26 (m, 1H, 5'a-H), 3.35
(m, 1H, 5'b-H), 3.40 ± 3.72 (m, 4H, POCH2 and isopropyl CH), 3.75, 3.76,
3.76, 3.77 (4s, 3H, OCH3), 4.08 (m, 1H, 4'-H), 4.61, 4.70 (2m, 1H, 3'-H),
4.99 (s, 1H, 5-H), 6.73 ± 7.48 (m, 14H, 1'-H and Har), 7.75, 7.76 (2s, 1H,
NCHN); MS ESI(À): m/z (%): 799.5 (100) [M À H]À.
(100 MHz, [D6]DMSO, 258C): d ˆ 34.32 (N(CH3)), 36.61 (C-2'), 40.32
(N(CH3)), 62.36 (C-5'), 71.32 (C-3'), 80.56 (C-1'), 81.23 (C-5), 87.17 (C-4'),
151.08 (C-2), 156.97 (NCHN), 159.10 (C-4), 163.66 (C-6); the assignment
was confirmed by an HSQC experiment; MS ESI(À): m/z (%): 297.2 (100)
[M À H]À.
5-Allyl-2'-deoxy-4-N-(dimethylformamidine)-6-oxocytidine 28: Protected
nucleoside 26 (0.58 g, 0.998 mmol) was dissolved in THF (7 mL), and
TBAF (2.20 mL, 2.20 mmol) was added. The product 28 was obtained in
330 mg (98%) yield. Rf ˆ 0.28 (dichloromethane/methanol 9:1); 1H NMR
(400 MHz, [D6]DMSO, 258C): d ˆ 1.87 (m, 1H, 2'a-H), 2.72 (m, 1H, 2'b-H),
2.93 (s, 3H, N(CH3)), 2.99 (d, J ˆ 6.3 Hz, 2H, allyl CH2), 3.02 (s, 3H,
N(CH3)), 3.46 (m, 1H, 4'-H), 3.56 (m, 1H, 5'a-H), 3.65 (m, 1H, 5'b-H), 4.31
(m, 1H, 3'-H), 4.56 (dd, J ˆ 7.1, 4.4 Hz, 1H, 5'-OH), 4.87 (ddd, J ˆ 17.1, 10.0,
2.2 Hz, 2H, allyl CH2), 5.00 (d, J ˆ 4.7 Hz, 1H, 3'-OH), 5.74 (m, 1H, allyl
CH), 6.57 (yt, J ˆ 7.3 Hz, 1H, 1'-H), 7.95 (s, 1H, NCHN), 10.55 (brs, 1H,
NH); the assignment was confirmed by a 1H,1H-COSY experiment;
13C NMR (100 MHz, [D6]DMSO, 258C): d ˆ 28.02 (allyl CH2), 33.72
(N(CH3)), 36.66 (C-2'), 40.17 (N(CH3)), 62.35 (C-5'), 71.30 (C-3'), 81.01
(C-1'), 87.22 (C-4'), 95.43 (C-5), 113.87 (allyl CH2), 137.16 (allyl CH), 150.51
(C-2), 153.27 (C-4), 155.45 (NCHN), 163.48 (C-6); the assignment was
confirmed by an HSQC experiment; MS ESI(À): m/z (%): 337.2 (100)
[M À H]À.
5-Allyl-3'-O-(2-cyanoethoxy-N,N-diisopropylphosphino)-2'-deoxy-5'-O-
(4,4'-dimethoxytriphenylmethyl)-4-N-(dimethylformamidine)-6-oxocyti-
Chem. Eur. J. 2000, 6, No. 13
ꢀ WILEY-VCH Verlag GmbH, D-69451 Weinheim, 2000
0947-6539/00/0613-2421 $ 17.50+.50/0
2421

U. Parsch, J. W. Engels
FULL PAPER
dine (32): N,N-Diisopropylethylamine (DIPEA) (376 mL, 2.20 mmol,
5 equiv) and 18 (147 mL, 0.66 mmol, 1.5 equiv) were added to a solution
of 30 (282 mg, 0.440 mmol) in dichloromethane (6 mL). After 2 h another
portion of 18 (49 mL, 0.5 equiv) was added. After 4 h the reaction mixture
was diluted with dichloromethane and extracted with 2% aqueous sodium
carbonate solution. The collected organic phases were dried over MgSO4,
filtered, and evaporated to dryness. The crude product was chromato-
graphed with dichloromethane/methanol 98:2 to afford 32 as a white foam
64.09, H 6.90, N 7.67; found C 63.93, H 6.95, N 7.71; MS ESI(À): m/z (%):
729.4 (100) [M À H]À.
5-Allyl-5'-O-(4,4'-dimethoxytriphenylmethyl)-4-N-dimethylformamidin-2'-
O-tert-butyldimethylsilyl-6-oxocytidine (34) and 5-allyl-5'-O-(4,4'-dimeth-
oxytriphenylmethyl)-4-N-dimethylformamidin-3'-O-tert-butyldimethylsil-
yl-6-oxocytidine (36): Nucleoside 13 (825 mg, 1.256 mmol) in THF/pyridine
1:1 (8 mL), silver nitrate (256 mg, 1.51 mmol, 1.2 equiv), and TBDMSCl
(1.51 mL, 1.51 mmol, 1.2 equiv ‡ 0.25 mL, 0.25 mmol, 0.2 equiv) afforded
34 (560 mg, 58%) and 36 (260 mg, 27%).
(262 mg, 71%). Compound 32 was obtained as
a mixture of two
diastereomers with different Rf values in a ratio of 1.2:1. The underlined
NMR spectroscopic data refer to the main diastereomer. Rf ˆ 0.55/0.61
(dichloromethane/methanol 95:5); 31P NMR (163 MHz, CDCl3, 258C): d ˆ
149.06, 149.14; 1H NMR (400 MHz, CDCl3, 258C): d ˆ 0.98 ± 1.15 (m, 12H,
isopropyl CH3), 2.23 (m, 1H, 2'a-H), 2.35, 2.57 (2t, J ˆ 6.6 Hz, J ˆ 6.3 Hz,
2H, CH2CN), 2.88 (m, 1H, 2'b-H), 3.01 (s, 3H, N(CH3)), 3.03 (d, J ˆ 5.9 Hz,
2H, allyl CH2), 3.07 (s, 3H, N(CH3)), 3.24 (dd, J ˆ 10.0, 3.4 Hz, 1H, 5'a-H),
3.30 (dd, J ˆ 10.0, 3.5 Hz, 1H, 5'b-H), 3.42 ± 3.57 (m, 4H, POCH2 and
isopropyl CH), 3.74, 3.75, 3.75, 3.76 (4s, 3H, OCH3), 4.07 (m, 1H, 4'-H),
4.66, 4.73 (2m, 1H, 3'-H), 4.95 (m, 2H, allyl CH2), 5.89 (m, 1H, allyl CH),
6.73 ± 7.48 (m, 14H, 1'-H and Har), 7.75, 7.76 (2s, 1H, NCHN), 8.38 (brs, 1H,
NH); MS ESI(À): m/z (%): 839.6 (100) [M À H]À.
General procedure for the conversion of 12 and 13 with TBDMSCl: Silver
nitrate and a 1m solution of tert-butyldimethylsilyl chloride (TBDMSCl)
were added to a solution of 12 or 13 in THF/pyridine 1:1. Silver chloride
precipitated. After the mixture had been stirred for 16 h at RT, another
portion of TBDMSCl was added. The reaction was stopped by addition of a
5% aqueous NaHCO3 solution, and the suspension was filtered through a
plug of Celite and washed. After extraction with dichloromethane, the
collected organic phases were dried over MgSO4, filtered, and evaporated
to dryness. The residue was coevaporated with toluene and purified by
column chromatography with dichloromethane/ethyl acetate 8:2. This
afforded the pure products as white foams.
34: Rf ˆ 0.34 (dichloromethane/ethyl acetate 8:2), Rf ˆ 0.60 (dichlorome-
thane/methanol 95:5); 1H NMR (400 MHz, [D6]DMSO, 258C): d ˆ À0.01
(s, 3H, SiCH3), 0.00 (s, 3H, SiCH3), 0.83 (s, 9H, SiC(CH3)3), 2.94 (s, 3H,
N(CH3)), 3.01 (d, J ˆ 7.0 Hz, 2H, allyl CH2), 3.02 (s, 3H, N(CH3)), 3.12 (m,
2H, 5'-H), 3.71 (s, 3H, OCH3), 3.72 (s, 3H, OCH3), 3.84 (m, 1H, 4'-H), 4.16
(m, 1H, 3'-H), 4.38 (d, J ˆ 8.3 Hz, 1H, 3'-OH), 4.59 (dd, J ˆ 6.1, 2.9 Hz, 1H,
2'-H), 4.86 (ddd, J ˆ 17.2, 10.0, 1.8 Hz, 2H, allyl CH2), 5.73 (m, 1H, allyl
CH), 6.14 (brs, 1H, 1'-H), 6.81 ± 7.41 (m, 13H, Har), 7.96 (s, 1H, NCHN),
10.58 (brs, 1H, NH); the assignment was confirmed by a 1H,1H-COSY
experiment; 13C NMR (100 MHz, [D6]DMSO, 258C): d ˆ À5.13 (SiCH3),
À4.80 (SiCH3), 18.06 (quart. C, tBu), 25.68 (CH3, tBu), 27.98 (allyl CH2),
33.72 (N(CH3)), 39.96 (N(CH3)), 54.92 (OCH3), 54.94 (OCH3), 64.42 (C-5'),
70.22 (C-3'), 73.31 (C-2'), 81.77 (C-4'), 85.18 (DMTr), 85.81 (C-1'), 95.45 (C-
5), 113.02 (DMTr), 113.66 (allyl CH2), 126.48, 127.65, 127.74, 129.67, 129.76,
135.67, 135.74 (DMTr), 137.08 (allyl CH), 145.04 (DMTr), 153.26 (C-2),
155.40 (NCHN), 157.91 (C-4), 157.96 (DMTr), 163.10 (C-6); the assignment
was confirmed by an HSQC experiment; elemental analysis calcd (%) for
C42H54N4O8Si (771.00): C 65.43, H 7.06, N 7.27; found C 65.31, H 7.11, N 7.07;
MS ESI(À): m/z (%): 769.6 (100) [M À H]À.
36: Rf ˆ 0.11 (dichloromethane/ethyl acetate 8:2), Rf ˆ 0.51 (dichlorome-
thane/methanol 95:5); 1H NMR (400 MHz, [D6]DMSO, 258C): d ˆ À0.10
(s, 3H, SiCH3), À0.05 (s, 3H, SiCH3), 0.74 (s, 9H, SiC(CH3)3), 2.94 (s, 3H,
N(CH3)), 3.00 (m, 1H, 5'a-H), 3.00 (d, J ˆ 7.7 Hz, 2H, allyl CH2), 3.02 (s,
3H, N(CH3)), 3.17 (m, 1H, 5'b-H), 3.71 (s, 3H, OCH3), 3.71 (s, 3H, OCH3),
3.87 (m, 1H, 4'-H), 4.38 (m, 2H, 2'-H, 3'-H), 4.72 (d, J ˆ 6.1 Hz, 1H, 2'-OH),
4.88 (ddd, J ˆ 17.2, 10.0, 1.7 Hz, 2H, allyl CH2), 5.74 (m, 1H, allyl CH), 6.14
(brs, 1H, 1'-H), 6.81 ± 7.40 (m, 13H, Har), 7.97 (s, 1H, NCHN), 10.57 (brs,
1H, NH); the assignment was confirmed by a 1H,1H-COSY experiment;
13C NMR (100 MHz, [D6]DMSO, 258C): d ˆ À5.25 (SiCH3), À4.67
(SiCH3), 17.86 (quart. C, tBu), 25.66 (CH3, tBu), 28.07 (allyl CH2), 33.71
(N(CH3)), 39.97 (N(CH3)), 54.93 (OCH3), 64.50 (C-5'), 71.39 (C-2'), 71.80
(C-3'), 81.37 (C-4'), 85.39 (DMTr), 86.26 (C-1'), 95.20 (C-5), 113.01 (DMTr),
113.83 (allyl CH2), 126.49, 127.63, 127.71, 129.59, 129.69, 135.61, 135.70
(DMTr), 137.16 (allyl CH), 144.91 (DMTr), 153.30 (C-2), 155.37 (NCHN),
157.95 (C-4), 157.99 (DMTr), 163.23 (C-6); the assignment was confirmed
by an HSQC experiment; elemental analysis calcd (%) for C42H54N4O8Si
(771.00): C 65.43, H 7.06, N 7.27; found C 65.18, H 7.25, N 7.01; MS ESI(À):
m/z (%): 769.5 (100) [M À H]À.
5'-O-(4,4'-Dimethoxytriphenylmethyl)-4-N-dimethylformamidin-2'-O-
tert-butyldimethylsilyl-6-oxocytidine (33) and 5'-O-(4,4'-dimethoxytriphe-
nylmethyl)-4-N-dimethylformamidin-3'-O-tert-butyldimethylsilyl-6-oxocy-
tidine (35): Nucleoside 12 (1.233 g, 2.0 mmol) in THF/pyridine 1:1 (12 mL),
silver nitrate (408 mg, 2.4 mmol, 1.2 equiv), and TBDMSCl (2.4 mL,
2.4 mmol, 1.2 equiv ‡ 0.4 mL, 0.4 mmol, 0.2 equiv) afforded 33 (833 mg,
57%) and 35 (395 mg, 27%).
33: Rf ˆ 0.15 (dichloromethane/ethyl acetate 8:2), Rf ˆ 0.54 (dichlorome-
thane/methanol 95:5); 1H NMR (250 MHz, [D6]DMSO, 258C): d ˆ 0.00 (s,
3H, SiCH3), 0.02 (s, 3H, SiCH3), 0.84 (s, 9H, SiC(CH3)3), 2.97 (s, 3H,
N(CH3)), 3.10 (s, 3H, N(CH3)), 3.15 (m, 2H, 5'-H), 3.73 (s, 6H, 2 Â OCH3),
3.82 (m, 1H, 4'-H), 4.15 (m, 1H, 3'-H), 4.37 (d, J ˆ 8.2 Hz, 1H, 3'-OH), 4.59
(dd, J ˆ 6.0, 2.7 Hz, 1H, 2'-H), 5.07 (s, 1H, 5-H), 6.10 (d, J ˆ 2.0 Hz, 1H, 1'-
H), 6.83 ± 7.42 (m, 13H, Har), 8.16 (s, 1H, NCHN), 10.73 (brs, 1H, NH); the
assignment was confirmed by
a
1H,1H-COSY experiment; 13C NMR
(63 MHz, [D6]DMSO, 258C): d ˆ À5.13 (SiCH3), À4.81 (SiCH3), 18.06
(quart. C, tBu), 25.67 (CH3, tBu), 34.28 (N(CH3)), 40.29 (N(CH3)), 54.91
(OCH3), 54.93 (OCH3), 64.43 (C-5'), 70.19 (C-3'), 73.30 (C-2'), 80.95 (C-4'),
81.58 (C-5), 85.17 (DMTr), 87.80 (C-1'), 113.01, 126.47, 127.64, 127.73,
129.66, 129.74, 135.66, 135.72, 145.02 (DMTr), 157.01 (C-2), 157.91
(NCHN), 157.95 (DMTr), 159.21 (C-4), 163.15 (C-6); the assignment was
confirmed by an HSQC experiment; elemental analysis calcd (%) for
C39H50N4O8Si (730.94): C 64.09, H 6.90, N 7.67; found C 64.07, H 6.95, N
7.60; MS ESI(À): m/z (%): 729.4 (100) [M À H]À.
General procedure for the phosphitylation of 33 and 34: The 2'-O-TBDMS-
protected compounds 33 and 34 were each dissolved in acetonitrile. Sym-
collidine, 1-methyl imidazole, and (after cooling with an ice bath)
2-cyanoethyl-N,N-diisopropylchloro phosphoramidite (18) were added.
The ice bath was removed after 15 min. After 2 h the reaction mixture was
diluted with dichloromethane and 5% aqueous NaHCO3 solution. The
extraction of the aqueous phase with dichloromethane was repeated three
times. The collected organic phases were dried over MgSO4, filtered and
evaporated to dryness. The residue was coevaporated with toluene five
times and purified by column chromatography with ethyl acetate/acetoni-
trile 9:1. The products 37 and 38 were obtained as mixtures of two
diastereomers with identical Rf values. The underlined NMR spectroscopic
data refer to the main diastereomer.
35: Rf ˆ 0.03 (dichloromethane/ethyl acetate 8:2), Rf ˆ 0.43 (dichlorome-
thane/methanol 95:5); 1H NMR (250 MHz, [D6]DMSO, 258C): d ˆ À0.10
(s, 3H, SiCH3), À0.04 (s, 3H, SiCH3), 0.74 (s, 9H, SiC(CH3)3), 2.97 (s, 3H,
N(CH3)), 3.02 (m, 1H, 5'a-H), 3.10 (s, 3H, N(CH3)), 3.18 (m, 1H, 5'b-H),
3.73 (s, 6H, 2 Â OCH3), 3.87 (m, 1H, 4'-H), 4.36 (m, 2H, 2'-H, 3'-H), 4.70 (d,
J ˆ 5.3 Hz, 1H, 2'-OH), 5.07 (s, 1H, 5-H), 6.10 (brs, 1H, 1'-H), 6.83 ± 7.41
(m, 13H, Har), 8.15 (s, 1H, NCHN), 10.70 (brs, 1H, NH); the assignment
3'-O-(2-Cyanoethoxy-N,N-diisopropylphosphino)-5'-O-(4,4'-dimethoxytri-
phenylmethyl)-4-N-(dimethylformamidine)-2'-O-tert-butyldimethylsilyl-6-
oxocytidine (37): Compound 33 (260 mg, 0.356 mmol) in acetonitrile
(5 mL), sym-collidine (236 mL, 1.78 mmol, 5 equiv), 1-methyl imidazole
(14 mL, 0.178 mmol, 0.5 equiv), and 18 (119 mL, 0.534 mmol, 1.5 equiv)
afforded 37 (240 mg, 72%) as a white foam. Rf ˆ 0.55 (ethyl acetate/
acetonitrile 9:1); 31P NMR (163 MHz, CDCl3, 258C): d ˆ 148.80, 149.48
(ratio 1.2:1); 1H NMR (400 MHz, CDCl3, 258C): d ˆ 0.02, 0.02, 0.04, 0.05
(4s, 3H, SiCH3), 0.88, 0.88 (2s, 9H, SiC(CH3)3), 1.02 ± 1.14 (m, 12H,
isopropyl CH3), 2.36, 2.63 (2m, 2H, CH2CN), 3.02 (s, 3H, N(CH3)), 3.11 (s,
was confirmed by
a
1H,1H-COSY experiment; 13C NMR (63 MHz,
[D6]DMSO, 258C): d ˆ À5.28 (SiCH3), À4.69 (SiCH3), 17.83 (quart. C,
tBu), 25.64 (CH3, tBu), 34.30 (N(CH3)), 40.30 (N(CH3)), 54.94 (OCH3),
64.55 (C-5'), 71.41 (C-2'), 71.74 (C-3'), 81.05 (C-4'), 81.16 (C-5), 85.39
(DMTr), 87.80 (C-1'), 113.01, 126.49, 127.63, 127.69, 129.58, 129.69, 135.59,
135.68, 144.88 (DMTr), 156.94 (C-2), 157.94 (NCHN), 157.99 (DMTr),
159.21 (C-4), 163.28 (C-6); the assignment was confirmed by an HSQC
experiment; elemental analysis calcd (%) for C39H50N4O8Si (730.94): C
2422
ꢀ WILEY-VCH Verlag GmbH, D-69451 Weinheim, 2000
0947-6539/00/0613-2422 $ 17.50+.50/0
Chem. Eur. J. 2000, 6, No. 13

Triple-Helix Formation
2409±2424
3H, N(CH3)), 3.17 ± 3.53 (m, 4H, 5'-H and POCH2), 3.77, 3.77, 3.77, 3.78 (4s,
3H, OCH3), 3.91 ± 3.96 (m, 2H, isopropyl CH), 4.19 (m, 1H, 4'-H), 4.40,
4.47 (2ddd, J ˆ 11.4, 5.9, 6.1 Hz, J ˆ 12.3, 5.9, 6.4 Hz, 1H, 3'-H), 4.88, 4.95
(2dd, 1H, 2'-H), 5.01, 5.02 (2s, 1H, 5-H), 6.29 (d, J ˆ 3.7 Hz, 1H, 1'-H),
6.76 ± 7.49 (m, 13H, Har), 7.77 (s, 1H, NCHN); MS ESI(‡): m/z (%): 931.9
858C was reached. Since a correct melting curve is an equilibrium
measurement, this assumption was tested by measuring the melting curve
at a lower heating rate (here 0.28CminÀ1). The curves were identical. Every
melting curve was determined twice and the Tm values were calculated
from a versus T plots. At low temperatures (<208C) the cuvette chamber
was flushed with nitrogen to prevent condensation.
‡
‡
(100) [M‡H] , 830.5 (79.3) [M À N(iPr)2] .
CD measurements: The samples of the UV spectroscopic investigations
were used directly. All CD measurements were performed on a JASCO
J-710 spectropolarimeter equipped with a Neslab RTE-100 temperature
control unit. The wavelength-dependent spectra were recorded between
330 and 210 nm with a resolution of 0.2 nm, a scan speed of 50 nmminÀ1, a
5-Allyl-3'-O-(2-cyanoethoxy-N,N-diisopropylphosphino)-5'-O-(4,4'-dimeth-
oxytriphenylmethyl)-4-N-(dimethylformamidine)-2'-O-tert-butyldimethyl-
silyl-6-oxocytidine (38): Compound 34 (285 mg, 0.370 mmol) in acetonitrile
(5 mL), sym-collidine (245 mL, 1.848 mmol, 5 equiv), 1-methylimidazole
(15 mL, 0.185 mmol, 0.5 equiv), and 18 (124 mL, 0.554 mmol, 1.5 equiv)
afforded 38 (255 mg, 71%) as a white foam. Rf ˆ 0.68 (ethyl acetate/
acetonitrile 9:1); 31P NMR (163 MHz, CDCl3, 258C): d ˆ 149.11, 149.54
(ratio 1.3:1); 1H NMR (400 MHz, CDCl3, 258C): d ˆ À0.01, 0.01, 0.04, 0.05
(4s, 3H, SiCH3), 0.86, 0.87 (2s, 9H, SiC(CH3)3), 0.89 ± 1.14 (m, 12H,
isopropyl CH3), 2.35, 2.62 (2m, 2H, CH2CN), 3.01 (s, 3H, N(CH3)), 3.08 (d,
J ˆ 7.2 Hz, 2H, allyl CH2), 3.09 (s, 3H, N(CH3)), 3.23 ± 3.58 (m, 4H, 5'-H
and POCH2), 3.76, 3.76, 3.77, 3.77 (4s, 3H, OCH3), 3.82 ± 3.98 (m, 2H,
isopropyl CH), 4.15 (m, 1H, 4'-H), 4.45 (m, 1H, 3'-H), 4.90 ± 4.98 (m, 3H,
2'-H and allyl CH2), 5.94 (m, 1H, allyl CH), 6.27, 6.28 (2s, 1H, 1'-H), 6.75 ±
7.48 (m, 13H, Har), 7.72, 7.72 (s, 1H, NCHN); MS ESI(‡): m/z (%): 971.9
response of 1 s, and
a bandwidth of 1.0 nm. Every spectrum was
accumulated five times and smoothed (noise reduction). At each temper-
ature the spectrum of the buffer was subtracted.
The temperature-dependent spectra were recorded between 58C or 108C
and 808C with a heating rate of 0.58CminÀ1 and a resolution of 0.1 nm, a
response of 2 s, and a bandwidth of 2.0 nm. A data point was registered
every 0.18C and the curves were smoothed. For a better comparison of the
curves in Figures 7 and 8 the ellipticity [q] has been normalized. This was
done by means of the equation qnorm ˆ (q(T) À ql)/(qh À ql), where (ql ˆ
ellipticity at the lowest temperature (108C), qh ˆ ellipticity at the highest
temperature (808C)).
‡
‡
(100) [M‡H] , 870.6 (60.1) [M À N(iPr)2] .
Oligonucleotide synthesis, purification, and characterization: Synthesis of
the oligonucleotides was performed by phosporamidite solid-phase chem-
istry by standard techniques with automated synthesizers (PerSeptive
Biosystem Expedite 8905 for DNA, Eppendorf Biotronik D300 ‡ for
RNA) on a 1 mmol scale. Therefore we used a 500 Š CPG support. 1H-
Tetrazole was used as coupling activator in all cases. The coupling times
were 30 s for the standard DNA building blocks (PerSeptive Biosystems)
and 300 s for the 2'-deoxy- and 2'-methoxy-6-oxocytidine compounds. For
all RNA building blocks a coupling time of 10 min was used. The
6-oxocytidine phosphoramidites were freeze-dried and employed as 0.1m
solutions in dry acetonitrile. All coupling efficiencies were comparable with
those of the common nucleoside phosphoramidites, judged by the color of
the liberated DMTr cation. DNA strands were synthesized in the trityl-on
mode and deprotected with conc. NH3 overnight at 558C. The purification
of the crude oligomers was achieved by RP-HPLC (Merck ± Hitachi) by
means of a Lichrosphor RP18 column filled with POROS R3 silica gel
(PerSeptive Biosystems). The mobile phase was a linear gradient of 0 ±
25% acetonitrile over a period of 10 min in 0.1m aqueous triethylammo-
nium acetate buffer (pH 7.0). The collected oligonucleotides were reduced
in volume, detritylated with 80% aqueous acetic acid (30 min, RT), and
precipitated with ethanol. RNA strands were synthesized in the trityl-off
mode and deprotected with conc. NH3/methanol 3:1 overnight at 558C. The
crude oligomers were deprotected at the 2'-position with 1 mL triethyl-
amine trihydrofluoride for 24 h at RT. Subsequently the crude RNA was
precipitated by butanol (À208C, 2 h). The oligonucleotides were dissolved
in DEPC/water and purified by anion exchange HPLC (JASCO) using a
Dionex NucleoPacTM PA100 column. The mobile phase was a linear
gradient of buffer B in A (0 to 70% in 40 min). Buffer A was DEPC/water
(pH 8.0):acetonitrile 9:1, buffer B 1m LiCl (pH 8.0):acetonitrile 9:1. The
collected oligonucleotides were reduced in volume and desalted on
Sephadex G25 columns. Finally the yields of all oligonucleotides were
measured at 260 nm. The extinction coefficients were calculated according
to the data given by Puglisi and Tinoco and by Gray et al.[61, 67] The
characterization of the oligonucleotides was performed on a VG Biotec
Tofspec MALDI mass spectrometer.
Acknowledgements
We wish to thank Dr. Zimmermann and his coworkers for all NMR
spectroscopic measurements. I. Priess and H. Brill for the measurements of
mass spectra, M. Christof for performing the microanalyses, Dr. J. Bats for
X-ray analysis of 6-oxocytidine and K. Jahn-Hofmann, B. Conrady, and T.
Strube for supporting synthesis and purification of the oligonucleotides. We
gratefully acknowledge the financial support of this work by a grant from
the Deutsche Forschungsgemeinschaft. U.P. wishes to thank the Fonds der
Chemischen Industrie for a Ph.D. fellowship.
[1] J. D. Watson, F. H. C. Crick, Nature 1953, 171, 737 ± 738.
[2] G. Felsenfeld, D. R. Davies, A. Rich, J. Am. Chem. Soc. 1957, 79,
2023 ± 2024.
[3] H. E. Moser, P. B. Dervan, Science 1987, 238, 645 ± 650.
[4] T. Le Doan, L. Perrouault, D. Praseuth, N. Habhoub, J. L. Decoult,
 Á
N. T. Thuong, J. Lhomme, C. Helene, Nucleic Acids Res. 1987, 15,
7749 ± 7760.
[5] M. Cooney, G. Czernuszewicz, E. H. Postel, S. J. Flint, M. E. Hogan,
Science 1988, 241, 456 ± 459.
[6] V. I. Lyamichev, S. M. Mirkin, M. D. Frank-Kamenetskii, C. R.
Cantor, Nucleic Acids Res. 1988, 16, 2165 ± 2178.
 Á
[7] N. T. Thuong, C. Helene, Angew. Chem. 1993, 105, 697 ± 723; Angew.
Chem. Int. Ed. Engl. 1993, 32, 666 ± 690.
[8] V. N. Soyfer, V. N. Potaman, Triple-Helical Nucleic Acids, Springer,
New York, Berlin, Heidelberg, 1996.
[9] C. De los Santos, M. Rosen, D. Patel, Biochemistry 1989, 28, 7282 ±
7289.
[10] P. Rajagopal, J. Feigon, Biochemistry 1989, 28, 7859 ± 7870.
[11] L. E. Xodo, G. Manzini, F. Quadrifoglio, G. A. van der Marel, J. H.
van Boom, Nucleic Acids Res. 1991, 19, 5625 ± 5631.
[12] S. F. Singleton, P. B. Dervan, J. Am. Chem. Soc. 1992, 114, 6957 ± 6965.
[13] L. Lavelle, J. R. Fresco, Nucleic Acids Res. 1995, 23, 2692 ± 2705.
[14] S. Wang, A. E. Friedman, E. T. Kool, Biochemistry 1995, 34, 9774 ±
9784.
[15] J. L. Asensio, A. N. Lane, J. Dhesi, S. Bergqvist, T. Brown, J. Mol. Biol.
1998, 275, 811 ± 822.
[16] J. S. Lee, M. L. Woodsworth, L. J. P. Latimer, A. R. Morgan, Nucleic
Acids Res. 1984, 12, 6603 ± 6614.
UV thermal denaturation studies: Thermal denaturation experiments were
performed on a Varian Cary 1 UV/Vis spectrophotometer equipped with a
Peltier temperature control unit in 1 mL of sodium cacodylate buffer
(50 mm) at pH 5.0, 6.0, 7.0, and 7.4 containing magnesium chloride (20 mm)
and sodium chloride (100 mm). The corresponding target duplex was
prepared by dissolving equimolar amounts of the 21-mer strands in the
buffer according to their molar extinction coefficients at 260 nm in a
cuvette with 10 mm length. The target duplex was denatured at 808C for
10 min and renatured with at 58CminÀ1. After the mixture had reached
208C an equimolar amount (2 mm) of the third strand was added and the
mixture was cooled to 08C for 20 min to reach equilibrium. Longer cooling
times (e.g. overnight) did not change the results. Then measurement was
started by increasing the temperature at a rate of 0.58CminÀ1, and the UV
absorbance was determined at 260 and 274 nm for every 0.58C step until
[17] T. J. Povsic, P. B. Dervan, J. Am. Chem. Soc. 1989, 111, 3059 ± 3061.
[18] B. C. Froehler, S. Wadwani, T. J. Terhorst, S. R. Gerrard, Tetrahedron
Lett. 1992, 33, 5307 ± 5310.
[19] N. Colocci, P. B. Dervan, J. Am. Chem. Soc. 1994, 116, 785 ± 786.
[20] A. K. Phipps, M. Tarköy, P. Schultze, J. Feigon, Biochemistry 1998, 37,
5820 ± 5830.
Chem. Eur. J. 2000, 6, No. 13
ꢀ WILEY-VCH Verlag GmbH, D-69451 Weinheim, 2000
0947-6539/00/0613-2423 $ 17.50+.50/0
2423

U. Parsch, J. W. Engels
FULL PAPER
[21] L. Yang, M. Liu, W. Deng, C. Wang, C. Bai, L.-S. Kan, Biophys. Chem.
1999, 76, 25 ± 34.
[52] G. H. Hakimelahi, Z. A. Proba, K. K. Ogilvie, Tetrahedron Lett. 1981,
22, 4775 ± 4778.
[22] D. A. Barawkar, K. G. Rajeev, V. A. Kumar, K. N. Ganesh, Nucleic
Acids Res. 1996, 24, 1229 ± 1237.
[53] G. H. Hakimelahi, Z. A. Proba, K. K. Ogilvie, Can. J. Chem. 1982, 60,
1106 ± 1113.
[23] K. G. Rajeev, V. R. Jadhav, K. N. Ganesh, Nucleic Acids Res. 1997, 25,
4187 ± 4193.
[54] S. A. Scarringe, C. Francklyn, N. Usman, Nucleic Acids Res. 1990, 18,
5433 ± 5441.
[24] K. N. Ganesh, K. G. Rajeev, P. S. Pallan, V. S. Rana, D. A. Barawkar,
V. A. Kumar, Nucleosides Nucleotides 1997, 16, 1271 ± 1278.
[25] K. N. Ganesh, V. A. Kumar, D. A. Barawkar, K. G. Rajeev, V. S. Rana,
Pure Appl. Chem. 1998, 70, 289 ± 292.
[26] J. Bijapur, M. D. Keppler, S. Bergqvist, T. Brown, K. R. Fox, Nucleic
Acids Res. 1999, 27, 1802 ± 1809.
[27] P. S. Bates, C. A. Laughton, T. C. Jenkins, D. C. Capaldi, P. D. Roselt,
C. B. Reese, S. Neidle, Nucleic Acids Res. 1996, 24, 4176 ± 4184.
[28] S. Hildbrand, C. Leumann, Angew. Chem. 1996, 108, 2100 ± 2102;
Angew. Chem. Int. Ed. Engl. 1996, 35, 2100 ± 2102.
[29] S. Hildbrand, A. Blaser, S. P. Parel, C. J. Leumann, J. Am. Chem. Soc.
1997, 119, 5499 ± 5511.
[30] S. A. Cassidy, P. Slickers, J. O. Trent, D. C. Capaldi, P. D. Roselt, C. B.
Reese, S. Neidle, K. R. Fox, Nucleic Acids Res. 1997, 25, 4891 ± 4898.
[31] A. Ono, P. O. P. Tsꢁo, L.-S. Kan, J. Am. Chem. Soc. 1991, 113, 4032 ±
4033.
[32] A. Ono, P. O. P. Tsꢁo, L.-S. Kan, J. Org. Chem. 1992, 57, 3225 ± 3230.
[33] J. S. Koh, P. B. Dervan, J. Am. Chem. Soc. 1992, 114, 1470 ± 1478.
[34] E. S. Priestley, P. B. Dervan, J. Am. Chem. Soc. 1995, 117, 4761 ± 4765.
[35] S. H. Krawczyk, J. F. Milligan, S. Wadwani, C. Moulds, B. C. Froehler,
M. D. Matteucci, Proc. Natl. Acad. Sci. USA 1992, 89, 3761 ± 3764.
[36] P. S. Miller, P. Bhan, C. D. Cushman, T. L. Trapane, Biochemistry 1992,
31, 6788 ± 6793.
[37] P. S. Miller, G. Bi, S. A. Kipp, V. Fok, R. K. DeLong, Nucleic Acids
Res. 1996, 24, 730 ± 736.
[38] U. von Krosigk, S. A. Benner, J. Am. Chem. Soc. 1995, 117, 5361 ±
5362.
[39] G. Xiang, W. Soussou, L. W. McLaughlin, J. Am. Chem. Soc. 1994, 116,
11155 ± 11156.
[40] G. Xiang, R. Bogacki, L. W. McLaughlin, Nucleic Acids Res. 1996, 24,
1963 ± 1970.
[41] G. Xiang, L. W. McLaughlin, Tetrahedron 1998, 54, 375 ± 392.
[42] R. Berressem, J. W. Engels, Nucleic Acids Res. 1995, 23, 3465 ± 3472.
[43] J. J. Voegel, S. A. Benner, J. Am. Chem. Soc. 1994, 116, 6929 ± 6930.
[44] H. Vorbrüggen, K. Krolikiewicz, B. Bennua, Chem. Ber. 1981, 114,
1234 ± 1255.
[45] H. Vorbrüggen, G. Höfle, Chem. Ber. 1981, 114, 1256 ± 1268.
[46] C. E. Müller, Tetrahedron Lett. 1991, 32, 6539 ± 6540.
[47] M. J. Robins, S. R. Naik, A. S. K. Lee, J. Org. Chem. 1974, 39, 1891 ±
1899.
[55] S. L. Beaucage, M. H. Caruthers, Tetrahedron Lett. 1981, 22, 1859 ±
1862.
[56] S. L. Beaucage, R. P. Iyer, Tetrahedron 1992, 48, 2223 ± 2311.
[57] N. Usman, K. K. Ogilvie, M.-Y. Jiang, R. J. Cedergren, J. Am. Chem.
Soc. 1987, 109, 7845 ± 7854.
[58] T. Wu, K. K. Ogilvie, R. T. Pon, Nucleic Acids Res. 1989, 17, 3501 ±
3517.
[59] M. H. Lyttle, P. B. Wright, N. D. Sinha, J. D. Bain, A. R. Chamberlin, J.
Org. Chem. 1991, 56, 4608 ± 4615.
[60] L. A. Marky, K. J. Breslauer, Biopolymers 1987, 26, 1601 ± 1620.
[61] J. D. Puglisi, I. Tinoco, Methods Enzymol. 1989, 180, 304 ± 325.
[62] U. Parsch, J. W. Engels, Collection Symposium Series, Vol. 2 (Eds.: A.
Holy, M. Hocek), Institute of Organic Chemistry and Biochemistry,
Academy of Sciences of the Czech Republic, Prague, 1999, pp. 15 ± 21.
[63] J. Gorski, P. Tollin, Cryst. Struct. Commun. 1982, 11, 543 ± 546.
[64] J. W. Bats, U. Parsch, J. W. Engels, unpublished data. The main
features of the elucidated structure are: The five-membered ring has a
C2'-endo envelope conformation with atoms C1', C3', C4' and O4'
almost in a plane and ring atom C2' 0.59 Š above this plane. N1 and
O2' occupy pseudoequatorial positions, O3' a pseudoaxial position
À
with respect to the five-membered ring. The C5' O5' group is in a
gauche ± gauche orientation. The pyrimidine ring is approximately
planar. The glycosyl bond has an anti orientation with a torsion angle
O4'-C1'-N1-C2 of À118.1(1)8. The molecule is stabilized by an
À
intramolecular hydrogen bond between the O5' H5'O hydroxyl
group and the keto oxygen atom O6. The crystal packing involves a
three-dimensional network of hydrogen bonds. Each possible hydro-
gen-bond donor is included in the hydrogen-bond system. There are
three intermolecular hydrogen bonds between pyrimidione and
ribofuranosyl groups of neighboring molecules. The water molecule
donates two hydrogen bonds and accepts two hydrogen bonds and is
hydrogen-bonded to four neighboring nucleoside molecules.
[65] M. D. Keppler, K. R. Fox, Nucleic Acids Res. 1997, 25, 4644 ± 4649.
[66] D. M. Gray, R. L. Ratliff, M. R. Vaughan, Methods Enzymol. 1992,
211, 389 ± 406.
[67] D. M. Gray, S.-H. Hung, K. H. Johnson, Methods Enzymol. 1995, 246,
19 ± 34.
[68] W. C. Johnson, Circular DichroismÐPrinciples and Applications,
VCH, Weinheim, New York, Basel, Cambridge, 1994, pp. 523 ± 540.
[69] R. W. Woody, Methods Enzymol. 1995, 246, 34 ± 71.
[70] U. Parsch, K. Wörner, J. W. Engels, Nucleosides Nucleotides 1999, 18,
1641 ± 1643.
[48] M. J. Robins, J. S. Wilson, F. Hansske, J. Am. Chem. Soc. 1983, 105,
4059 ± 4065.
[49] a) W. T. Markiewicz, J. Chem. Research (S) 1979, 24 ± 25; b) W. T.
Markiewicz, J. Chem. Research (M) 1979, 181 ± 187.
[50] K. K. Ogilvie, K. L. Sadana, E. A. Thompson, Tetrahedron Lett. 1974,
2861 ± 63.
[71] Y.-K. Cheng, B. M. Pettitt, Prog. Biophys. Molec. Biol. 1992, 58, 225 ±
257.
[51] K. K. Ogilvie, S. L. Beaucage, A. L. Schifman, N. Y. Theriault, K. L.
Sadana, Can. J. Chem. 1978, 56, 2768 ± 2780.
Received: November 25, 1999 [F2158]
2424
ꢀ WILEY-VCH Verlag GmbH, D-69451 Weinheim, 2000
0947-6539/00/0613-2424 $ 17.50+.50/0
Chem. Eur. J. 2000, 6, No. 13