Synthesis and triplex forming properties of pyrrolidino pseudoisocytidine containing oligodeoxynucleotides

Article abstract of DOI:10.1039/b502799c

Pyrrolidino pseudo-C-nucleosides are isosteres of natural deoxynucleosides which are protonated at the pyrrolidino ring nitrogen under physiological conditions. As constituents of a triplex forming oligodeoxynucleotide (TFO), the positive charge is expect

Full text of DOI:10.1039/b502799c

View Article Online / Journal Homepage / Table of Contents for this issue
A R T I C L E
Synthesis and triplex forming properties of pyrrolidino
pseudoisocytidine containing oligodeoxynucleotides
Alain Mayer, Adrian Ha¨berli and Christian J. Leumann*
Department of Chemistry & Biochemistry, University of Berne, Freiestrasse 3, CH-3012, Berne,
Switzerland. E-mail: leumann@ioc.unibe.ch; Fax: +41(0)31 631 3422; Tel: +41(0)31 631 4355
Received 24th February 2005, Accepted 21st March 2005
First published as an Advance Article on the web 7th April 2005
Pyrrolidino pseudo-C-nucleosides are isosteres of natural deoxynucleosides which are protonated at the pyrrolidino
ring nitrogen under physiological conditions. As constituents of a triplex forming oligodeoxynucleotide (TFO), the
positive charge is expected to stabilise DNA triple helices via electrostatic interactions with the phosphodiester
backbone of the target DNA. We describe the synthesis of the pyrrolidino isocytidine pseudonucleoside and the
corresponding phosphoramidite building block and its incorporation into TFOs. Such TFOs show substantially
increased DNA affinity compared to unmodified oligodeoxynucleotides. The increase in affinity is shown to be due to
the positive charge at the pyrrolidino subunit.
Introduction
Sequence specific binding of triplex forming oligonucleotides
(TFOs) in the major groove of double stranded DNA can
control and modulate gene expression at the level of tran-
scription and is the basis of antigene technology.1 TFOs can
bind to homopurine/homopyrimidine tracts of DNA in either
parallel or antiparallel alignment relative to the purine target
strand, forming Hoogsteen or reversed-Hoogsteen base-triples.
Unfortunately, the use of triplex forming oligonucleotides
(TFOs) as antigene agents suffers from several limitations.
Low thermal stability of the triplexes, low biostability of
the TFOs in vivo, low bioavailability, sequence restrictions to
homopurine/homopyrimidine DNA tracts and a strong pH
dependence in the parallel binding motif due to C+·G–C triplet
formation are among the major disadvantages that reduce the
efficacy of antigene agents and that slow down the development
and applications of antigene technology. Many strategies have
been exploited to avoid these drawbacks,2 including the use of
conformationally restricted DNA analogues to increase thermal
stability3 and the design of new selective nucleobase analogues
Fig. 1 Top: ‘dual recognition’ principle for pyrrolidino pseudonucleo-
to alleviate target sequence restriction.4
sides (strand 2 corresponds to the purine strand and strand 3 to the TFO
in a modeled triplex); bottom: pyrrolidino pseudonucleoside analogues
of deoxyuridine (dpwU), thymidine (dpwT) and deoxycytidine (dpwiC)
as well as their deoxyribo-equivalents used for comparison.
As part of our current research we are interested in increasing
TFO affinity to a DNA target by applying the ‘dual recognition’
approach, originally developed by Cuenoud and coworkers.5 In
this approach, a target dsDNA is not only recognised by selective
forming properties of TFOs containing dpwU(T) and dpwiC
base–base interactions but also via an additional, non-specific
were evaluated using UV melting curve analysis, CD spec-
salt bridge between an appropriately placed positive charge in
troscopy and gel mobility shift experiments. For comparison,
the TFO and a negatively charged phosphate oxygen in the target
2ꢀ-deoxypseudoisocytidine (drwiC) and 2ꢀ-deoxypseudouridine
DNA backbone.
(drwU) containing TFOs in the same sequence context were
Based on a recent X-ray data on a parallel DNA triplex,6
studied as well, allowing for an analysis of the contribution of
molecular modelling showed that replacement of the 4ꢀ-oxygen
the pyrrolidino modification. For this we recently also developed
in a deoxynucleoside residue of a TFO by a basic nitrogen would
a convenient synthesis of the drwiC phosphoramidite building
potentially place a positive charge next to a non-bridging pro-
block.10
R-phosphate oxygen of the purine strand of a target dsDNA
(Fig. 1). In recent studies we explored pyrrolidino pseudonu-
Results and discussion
cleosides containing the base uracil (dpwU) and N-1-methyl
uracil (dpwT, Fig. 1) for A–T base-pair recognition.7,8 Analysis
of corresponding TFOs showed a destabilisation relative to
thymidine in a sequence dependent manner. Interestingly, duplex
formation was only marginally destabilised and the sequence
effects were less pronounced.
In order to complement this study we focused now on the
pyrrolidino analogue of pseudoisocytidine in order to recognise
target G–C base-pairs.9 Here we report on the synthesis of pyrro-
lidino pseudoisocytidine (dpwiC, Fig. 1) via Heck chemistry
and its incorporation into oligodeoxynucleotides. The triplex
Synthesis of the monomer
We based our synthesis of the dpwiC monomer 8 on a palladium
catalysed Heck reaction as already applied successfully for the
synthesis of the dpwU and dpwT monomers.7 As precursors, we
envisioned the CBz protected enamine 4 and the corresponding
protected iodoisocytosine 3. Enamine 4 could be obtained in
seven steps and 52% overall yield starting from trans-3-hydroxy-
L-proline.7 The necessary 2-N-benzoyl-5-iodoisocytosine (3) was
prepared in two steps from commercially available isocytosine 1
T h i s j o u r n a l i s
T h e R o y a l S o c i e t y o f C h e m i s t r y 2 0 0 5
O r g . B i o m o l . C h e m . , 2 0 0 5 , 3 , 1 6 5 3 – 1 6 5 8
1 6 5 3
©

View Article Online
via iodination with NIS in AcOH, followed by benzoylation of
the exocyclic amino function under standard conditions in good
yield (Scheme 1).
synthesis. The hardly soluble product 9 was subsequently
subjected to dimethoxy-tritylation (→ 10), followed by standard
phosphitylation to yield the phosphoramidite building block 11
in 17% overall yield from 4.
Synthesis of oligodeoxynucleotides
Oligodeoxynucleotides were synthesised on a 1.0 or 1.3 lmol
scale on a DNA synthesiser using standard solid-phase phos-
phoramidite chemistry. Minor modifications to the synthesis
cycle were introduced for the incorporation of the non-natural
building blocks. More precisely, the coupling time was extended
from 1.5 to 6 min and the standard activator tetrazole was
replaced by the more powerful (S-benzylthio)-1H-tetrazole.
Coupling efficiencies for the modified units were typically >97%,
according to trityl assay. Standard ammonia deprotection led to
cleavage of the crude oligonucleotides from the solid support
and removal of all protecting groups (including Fmoc). Crude
oligomers were purified by ion exchange HPLC and charac-
terised by ESI-mass spectrometry. A summary of the modified
oligomers synthesised and their mass spectrometric analysis is
provided in Table 1. The triplex forming properties of the modi-
fied oligodeoxynucleotides were subsequently examined by UV
melting curve analysis, CD spectroscopy and gel electrophoresis.
Scheme 1 Reagents and conditions: (a) NIS, AcOH, 70 ◦C → 100 ◦C,
1 h, 87%; (b) benzoic anhydride, DMF, 100 ◦C, 1.5 h, 76%.
The palladium catalyzed Heck reaction of enamine 4 with
base 3 using Pd(OAc)2 as catalyst, AsPh3 as a soft ligand and
diisopropylethylamine as base gave selectively the b-pyrrolidino
derivative 5 in 63% yield (Scheme 2). In situ silylation of 3 with
BSA was necessary due to its poor solubility in DMF. The b-
configuration at the pseudo anomeric centre was spectroscopi-
cally confirmed at a later stage. After removal of the silyl protect-
ing groups with HF–pyridine in acetonitrile, the resulting ketone
was diastereoselectively reduced with NaBH(OAc)3 to yield diol
6. The cleavage of the CBz group (6 → 7) was then accomplished
by catalytic hydrogenation. A sample of compound 7 was
converted (conc. NH3) to the fully deprotected nucleoside 8 for
spectroscopic analysis. The relative configurations at C(1ꢀ) and
C(3ꢀ) (nucleoside numbering scheme) in 8 could unambiguously
be assigned by 1H-NMR-NOE experiments (see Experimental).
Pairing properties with DNA duplexes in the parallel motif
The Tm of TFO dissociation from a 21-mer dsDNA target are
summarized in Table 2. Since protonation of the pyrrolidino ring
nitrogen is pH-dependent, the melting experiments were carried
out at different pH values. NMR titration experiments with the
monomer dpwU revealed a pKa of 7.9.8 Therefore we chose to
investigate triplex formation in the pH range from 6 to 9.
At pH 6, where the pyrrolidino ring nitrogen is fully proto-
nated, the triplexes with TFOs dp1–3, bearing dpwiC residues
(Table 2), were more stable than the one formed with the
reference TFO Ref1, in which the modified units were replaced
by deoxycytidine. Each modification contributed to triplex
stability by 1.9 to 2.4 ◦C. The most stable of all investigated
triplexes was found to be dp3 with five dpwiC units and a Tm that
was ca. 10 ◦C higher than that of the corresponding unmodified
triplex. The TFOs dr1–3, containing pseudoisocytidine units
(Table 2), were used to identify affinity differences arising
exclusively from the change of the nucleobase. We found a
decrease in Tm per mod. by ca. 2.5–5.0 ◦C for these TFOs, relative
to the DNA control Ref1. This clearly shows that the pyrrolidino
modification is contributing significantly to stability, as it is able
to overcompensate for the generally destabilising nature of the
pseudobases under the given conditions.
Scheme 2 Reagents and conditions: (a) i). 3, BSA, DMF, rt, 1 h; ii). 4,
iPr2NEt, Pd(OAc)2, AsPh3, DMF, 80 ◦C, 22 h, 63%; (b) HF–pyridine,
MeCN, rt, 6 h; (c) NaBH(OAc)3, AcOH, MeCN, −15 ◦C → rt, 5 h, 79%;
(d) H2, Pd/C, MeOH, rt, 7 h, 99%; (e) conc. NH3, 55 ◦C, 18 h, 83%;
(f) Fmoc–OSu, THF, dioxane, 5% aq. NaHCO3, rt, 3 h; (g) DMT–Cl,
pyridine, rt, 2 h, 47%; (h) (iPr2N)(NCCH2CH2O)PCl, iPr2NEt, THF, rt,
2 h, 88%.
Of special interest was TFO dp3 in which all deoxycytidines
were replaced by dpwiC units. In contrast to deoxycytidine,
the isocytidine pseudonucleosides need no protonation at the
bases upon binding to their target G–C base-pairs. Thus, any
dependence of target binding of dp3 from pH in the interval of
pH 6–9 must be due to a change of the protonation state of the
pyrrolidine ring nitrogen. Indeed, raising the pH from 6.0 to 9.0
leads to a reduction in triplex Tm by 25 ◦C, corresponding to
The pyrrolidino ring nitrogen in 7 was then Fmoc protected
(→ 9), as this base labile protecting group is compatible with
the regular protecting group scheme in oligodeoxynucleotide
3
◦C per mod. In contrast, and as expected, dr3, containing
deoxypseudoisocytidine residues instead of the pyrrolidino
Table 1 Sequence and ESI-MS data of the modified oligonucleotides investigated in this study
TFO
Sequence
Mod.
[M–H]− (calcd)
[M–H]− (found)
dp1
dr1
dp2
dr2
dp3
dr3
dp4
dp5
dr5
5ꢀ-d(TTTTTCTXTCTCTCT)
5ꢀ-d(TTTTTCTXTCTCTCT)
5ꢀ-d(TTTTTXTCTXTCTCT)
5ꢀ-d(TTTTTXTCTXTCTCT)
5ꢀ-d(TTTTTXTXTXTXTXT)
5ꢀ-d(TTTTTXTXTXTXTXT)
5ꢀ-d(YYYYYXYXYXYXYXT)
5ꢀ-d(TTTXXXXTTTTXTTT)
5ꢀ-d(TTTXXXXTTTTXTTT)
X = dpwiC
4423.9
4424.9
4422.9
4424.9
4419.8
4424.9
4284.9
4420.0
4424.9
4424.4
4424.7
4423.6
4424.8
4420.3
4424.8
4285.0
4420.3
4425.0
X = drwiC
X = dpwiC
X = drwiC
X = dpwiC
X = drwiC
X = dpwiC, Y = dpwU
X = dpwiC
X = drwiC
1 6 5 4
O r g . B i o m o l . C h e m . , 2 0 0 5 , 3 , 1 6 5 3 – 1 6 5 8

View Article Online
Table 2 Tm data (◦C) of third strand dissociation from UV melting curves (260 nm)
Target
TFO
5ꢀ-d(GCTAAAAAGAGAGAGAGATCG) 3ꢀ-d(CGATTTTTCTCTCTCTCTAGC)
a
pH
Tm
DTm/mod.
Ref1b
dp1
dr1
dp2
dr2
dp3
6.0
6.0
6.0
6.0
6.0
6.0
7.0
8.0
9.0
6.0
7.0
8.0
9.0
6.0
43.1
45.5
39.9
46.9
37.9
53.0
44.4
37.1
28.0
16.4
11.2
13.5
10.0
n.d.c
0
+2.4
−3.2
+1.9
−2.6
+2.0
–
–
–
dr3
dp4
−5.3
–
–
–
n.d.c
a Single strand concentration = 1.2 lM in 140 mM KCl, 7 mM NaH2PO4, 0.5 mM MgCl2. Tm of target duplex = 57.0
d(TTTTTCTCTCTCTCT). c No Tm detectable.
1.0◦C. b Ref1 = 5ꢀ-
pseudoisocytidines, essentially shows no pH dependence in
target binding, indicating that no reversible protonation events
are involved in target binding.
These results unambiguously highlight the importance of the
protonated state of the pyrrolidino units in target binding and
are in agreement with the ‘dual recognition’ principle by specific
base–base interactions and unspecific salt bridge formation
between the protonated pyrrolidino nitrogens of the TFO and
phosphate units of the target DNA.
The fully modified dpwU and dpwiC containing TFO dp4
(Table 2) did not show any triplex formation with its target. The
observed monophasic UV-melting transition could be attributed
to the melting of the Watson–Crick duplex. This result is
in line with earlier observations with TFOs containing only
dpwU units that were found to strongly destabilise triplexes.8
Obviously the destabilising effect of the nine dpwU units cannot
be compensated by the five dpwiC units in dp4.
30 ◦C. This Tm is even higher than that of Ref2 at pH 6. The TFO
dr5, containing five drwiC units, was again used for dissecting
the role of the sugar modification. As expected, the Tms between
pH 6.0 and 8.0 are similar ruling out protonation events upon
target binding. In addition, their values between 11–16 C are
considerably lower than that of dp5, again demonstrating the
superiority of the dpwiC units.
To confirm the negative binding result of the fully modified
dp4 (where only one transition was observed in the UV melting
curve) and to corroborate the positive results obtained for
the dpwiC containing TFOs dp3 and dp5, a gel retardation
experiment was performed (Fig. 2).
◦
In order to test for sequence effects, a different sequence con-
text was explored. The corresponding modified 15-mer TFOs,
their DNA target and the Tm values for third strand melting at
pH values between 6 and 8 are summarized in Table 3. These
data were again compared with those of the reference TFO Ref2
bearing 5-methyldeoxycytidine (MeC) instead of deoxycytidine at
the X-positions.
At pH 6 the oligonucleotide dp5, containing five dpwiC units,
binds to the target duplex with considerably higher affinity
Fig. 2 UV shadowing of a non-denaturing 20% polyacrylamide gel at
pH 7.1. Lanes 1 and 2: purine rich and pyrimidine rich single strands
of the DNA target for dp3 and dp4; lane 3–5: dp3, dp5 and dp4 (single
strands); lane 6 and 7: target duplexes; lane 8: ternary mixture of dp5 with
its target duplex strands; lane 9: ternary mixture of dp3 with its target
duplex strands; lane 10: ternary mixture of dp4 with its target duplex
strands. 0.7 nmol of each strand was used with equal stoichiometry in
mixtures.
◦
than the reference TFO Ref2. The Tm is ca. 13 C higher and
the DTm per mod. of 2.6 ◦C is of the same order as in the
previous sequence context. Again, the Tm drop upon increasing
the pH from 6–8 due to partial deprotonation of the pyrrolidino
nitrogen. Remarkably, even at pH 8 where Ref2 has no detectable
affinity to its target, dp5 still binds dsDNA with a Tm of ca.
Table 3 Tm data (◦C) of third strand dissociation from UV melting curves (260 nm)
Target
TFO
5ꢀ-d(GCTAAAGGGGAAAAGAAATCG) 3ꢀ-d(CGATTTCCCCTTTTCTTTAGC)
a
pH
Tm
DTm/mod.
Ref2b
6
7
8
6
7
8
6
7
8
27.2
n.d.c
n.d.c
40.0
36.2
29.3
16.5
11.8
11.0
0
0c
0c
dp5
+2.6
–
–
dr5
−2.1
–
–
a Single strand concentration = 1.2 lM in 140 mM KCl, 7 mM NaH2PO4, 0.5 m MgCl2. Tm of target duplex = 60.0
d(TTTCCCCTTTTCTTT), C = 5MeC. c No Tm detectable.
1.0 ◦C. b Ref2 = 5ꢀ-
O r g . B i o m o l . C h e m . , 2 0 0 5 , 3 , 1 6 5 3 – 1 6 5 8
1 6 5 5

View Article Online
From the ternary mixtures of the TFOs dp3, dp5 and dp4 with
their duplex targets, it clearly emerges that no triplex is formed
in the latter case (lane 10) while stable triplexes occur in the two
former cases (lane 8 and 9).
contexts. By comparison with TFOs containing ribo-pseudo-
iC units (drwiC-series), it clearly emerges that the increase in
stability is associated with the protonation of the pyrrolidino
ring nitrogen which confirms the presence of attractive electro-
static interactions between the additional positive charge in the
TFOs and the phosphodiester backbone of the target duplex.
A puzzling situation remains in that this additional elec-
trostatic stabilisation seems to be restricted to dpwiC units
and does not occur with dpwU units, as reported earlier (see
also binding properties of dp4).8 In the case of the dpwU
and its N-1-methyl derivative, the pyrrolidino modification was
shown to destabilise a triplex, while in the case of dpwiC the
pyrrolidino unit significantly stabilises the triplex. On the basis of
a X-ray triplex structure we found that O-(4ꢀ)-pro-R-phosphate
In order to get additional information on target binding by
dpwiC containing TFOs, the triplex involving dp3 was analyzed
by CD spectroscopy. The CD spectra were measured at different
temperatures below and above the Tm of the third strand (Fig. 3,
top). The resulting curves were almost identical in shape, except
in the region below 235 nm. At low temperature, the curves
show a minimum around 220 nm, characteristic of a triple
helical structure, which evolves into a maximum upon raising
the temperature to above 50 ◦C. This change coincides with
the melting of the triplex to the duplex. Consequently a CD
melting curve at a fixed wavelength of 217 nm was recorded
(Fig. 3, bottom). As in the UV melting experiments, a heating–
cooling–heating cycle was applied and sigmoidal curves were
obtained. The two heating ramps were superimposable while
the cooling ramp showed slight hysteresis. A Tm value of 48 ◦C
was determined which is in agreement with that from the UV-
melting results.
˚
oxygen distances vary between 3.4–4.0 A from base-triple to
base-triple.6 Given the fact that electrostatic interactions are
strongly distance dependent, it is possible that local structural
variations in the N–O distances (sequence effects) might account
for the differential stability. Another reason might be a generally
unfavourable TFO conformation of oligodeoxyribonucleotides
for dual recognition (B- vs. A-conformation). Indeed, dual
recognition of sugar modified TFOs was so far only found to
be effective for those preferring A helix conformations (e.g. the
2ꢀ-aminoethyl nucleosides). It might therefore be of interest to
study RNA-TFOs containing single dpw-modifications or the
corresponding pyrrolidino pseudonucleosides with a 2ꢀ-hydroxyl
function in the ribo-configuration.
Experimental
General
Solvents for chromatography and extraction were distilled prior
to use. Chemicals for synthesis were generally reagent grade.
All moisture sensitive reactions were performed in oven dried
glassware in anhydrous solvents under an Ar atmosphere.
External bath temperatures were used to record all reaction
temperatures. Spectral data (NMR, MS) were measured with
standard equipment and are indicated in standard format. J
values are given in Hz and chemical shifts were referenced to the
residual, non-deuterated solvent used. Multiplicity assignments
for 13C-signals were from DEPT spectra. Flash chromatography
(FC) was performed using silica gel 60 (230–400 mesh). Thin-
layer chromatography (TLC) was performed on pre-coated silica
gel plates SIL-G-25 UV254 (Macherey-Nagel).
2-Amino-5-iodo-4-oxo-3,4-dihydropyrimidine 2. To a solu-
tion of 2-amino-4-oxo-3,4-dihydropyrimidine 1 (5.20 g, 46.8
mmol) in AcOH (84 cm3) at 70 ◦C was added N-iodosuccinimide
(11.6 g, 51.6 mmol) and the suspension was heated to 100 ◦C. Af-
ter 1 h, the mixture was cooled to rt and H2O (84 cm3) was added.
The solid was filtered over a filter paper (Whatman 1PS), washed
with H2O and dried under HV to give 2 (9.68 g, 87%) as a white
solid. Mp 245–250 ◦C (decomp.). dH (300 MHz; (CD3)2SO) 6.70
(2 H, br s, NH2), 7.93 (1 H, s, C(6)H), 11.26 (1 H, br s, NH); dC
(75 MHz; (CD3)2SO) 71.3 (s, C(5)), 156.8 (d, C(6)), 160.0, 161.4
(2 × s, C(2, 4)); m/z (FAB+) 237.9497 (M+ + H. C4H5N3OI
requires 237.9477).
Fig. 3 CD spectra of the triplex containing TFO dp3; top: CD spectra
measured at various temperatures; bottom: temperature scan CD spectra
from 10 to 80 ◦C at 217 nm; single strand concentration = 1.2 lM in
140 mM KCl, 7 mM NaH2PO4, 0.5 mM MgCl2, pH 6.0.
Conclusions
Introduction of positive charges at strategic positions within a
TFO has proven in the past to be a valuable tool to improve
triplex stability by favourable additional electrostatic interac-
tions with the negatively charged phosphodiester backbone
of the target duplex. Most of the strategies developed so
far were based on the attachment of corresponding chemical
functions at various positions of the sugar,5a–c the base11 or
the phosphate backbone.12 The pyrrolidino pseudonucleoside
dpwiC, as reported here, is a valuable extension of the ‘dual
recognition’ concept and provides a novel molecular platform.
TFOs containing pyrrolidino pseudoisocytidine units
(dpwiC-series) exhibit a significant increase in triplex stability
compared to unmodified TFOs in at least two different sequence
2-(N-Benzoylamino)-5-iodo-4-oxo-3,4-dihydropyrimidine 3.
A suspension of 2 (5.00 g, 21.1 mmol) and benzoic anhydride
(12.6 g, 55.7 mmol) in dry DMF (80 cm3) was heated to
100 ◦C. After 1.5 h, a clear, slightly yellow solution was
obtained. Most of the DMF was evaporated and the residue
suspended in EtOH (150 cm3). The solid was filtered off
(Whatman 1PS) and washed with cold EtOH (150 ml). Drying
under HV afforded dihydropyrimidine 3 (5.46 g, 76%) as a white
◦
powder. Mp 296–299 C (decomp.). dH (300 MHz; (CD3)2SO)
7.54 (2 H, t, J 7.5, Bz), 7.67 (1 H, t, J 7.4, Bz), 8.03 (2 H, d,
J 7.4, Bz), 8.31 (1 H, s, C(6)H), 12.25 (2 H, br s, 2 × NH); dC
(75 MHz; (CD3)2SO) 81.2 (s, C(5)), 128.7 (d, Bz), 132.6 (s, Bz),
1 6 5 6
O r g . B i o m o l . C h e m . , 2 0 0 5 , 3 , 1 6 5 3 – 1 6 5 8

View Article Online
133.36 (d, Bz), 152.5 (d, C(6)), 159.0, 160.7 (2 × s, C(2, 4)),
pyrrolidine 7 (211 mg, 99%) as a white solid. dH (300 MHz;
CD3OD) 2.06–2.25 (2 H, m, C(3)H2), 3.29–3.34 (1 H, m,
C(5)H), 3.70–3.75 (2 H, m, CH2OH), 4.24–4.30 (1 H, m,
C(4)H), 4.44–4.50 (1 H, m, C(2)H), 7.53 (2 H, t, J 7.0, Bz),
7.61 (1 H, d, J 7.0, Bz), 7.91 (1 H, s, C(6ꢀ)H), 8.05 (2 H, d, J
7.4, Bz); dC (75 MHz, CD3OD) 41.2 (t, C(3)), 57.9 (d, C(2)),
63.6 (t, CH2OH), 70.1 (d, C(5)), 74.4 (d, C(4)), 129.6, 129.9,
130.8, 133.8, 134.5 (5 × d; Bz), 150.5 (d, C(6ꢀ)); m/z (FAB+)
331.1406 (M+ + H. C16H19N4O4 requires 331.1406), 242 (10%),
227 (23%).
+
+
=
169.8 (s, C O); m/z (FAB ) 341.9752 (M + H. C11H9N3O2I
requires 341.9739).
(2R,5R)-N-(Benzyloxy)carbonyl-2-[2-(N-benzoylamino)-4-oxo-
3,4-dihydropyrimidin-5-yl]-4-[(tert-butyl)dimethylsilyl]oxy-5-[(tert-
butyl)dimethylsilyl]oxymethyl-aza-cyclopent-3-ene 5. To a sus-
pension of 3 (2.73 g, 8.00 mmol) in dry DMF (16 cm3) was added
dropwise N,O-bis(trimethylsilyl)acetamide (BSA, 2.45 cm3,
10.0 mmol). After 1 h, N,N-diisopropylethylamine (1.87 cm3,
10.9 mmol) and pyrroline 4 (1.60 g, 3.34 mmol) were added
to the clear solution. To a separate solution of triphenylarsine
(408 mg, 1.33 mmol) in dry DMF (48 cm3) was added Pd(OAc)2
(150 mg, 0.67 mmol). After 30 min, this solution was added
dropwise to the first solution and the mixture was heated at
(2R,4S,5R)-2-(2-Amino-4-oxo-3,4-dihydropyrimidin-5-yl)-4-
hydroxy-5-hydroxymethyl pyrrolidine 8.
A solution of 7
(120 mg, 0.36 mmol) in 25% conc. NH3 (5 cm3) was heated to
60 ◦C for 18 h. The solution was evaporated to dryness and the
residue dissolved in MeOH (3 cm3). The product precipitated
after the addition of EtOAc. Filtration and drying under HV
gave nucleoside 8 (68 mg, 83%) as a slightly brown powder. dH
(300 MHz, D2O) 1.88–1.96 (1 H, m, C(3) Ha), 2.00–2.11 (1 H,
m, C(3)Hb), 3.18–3.23 (1 H, m, C(5)H), 3.54–3.67 (2 H, m,
CH2OH), 4.15–4.19 (1 H, m, C(4)H), 4.26 (1 H, dd, J 6.6 and
11.0, C(2)H), 7.46 (1 H, s, C(6ꢀ)H); dC (75 MHz, D2O) 38.4 (t,
C(3)), 56.2 (d, C(2)), 61.5 (t, CH2OH), 67.8 (d, C(5)), 72.6 (d,
C(4)), 112.1 (s, C(5ꢀ)), 148.5 (d, C(6ꢀ)), 158.9 (s, C(2ꢀ)), 171.3 (s,
C(4ꢀ)); difference-NOE (500 MHz, D2O) 1.88–1.96 (C(3)Ha) →
2.00–2.11 (13%, C(3)Hb), 4.26 (1.9%, C(2)H); 2.00–2.11
(C(3)Hb) → 1.88–1.96 (13%, C(3)Ha), 4.15–4.19 (4.7%, C(4)H),
7.46 (1.7%, C(6ꢀ)H); 3.18–3.23 (C(5)H) → 3.54–3.67 (3.0%,
CH2OH), 4.15–4.19 (1.5%, C(4)H), 4.26 (1.3%, C(2)H);
3.54–3.67 (CH2OH) → 3.18–3.23 (1.6%, C(5)H), 4.15–4.19
(2.0%, C(4)H); 4.15–4.19 (C(4)H) → 2.00–2.11 (1.6%, C(3)Hb),
3.54–3.67 (1.4%, CH2OH); 4.26 (C(2)H) → 1.88–1.96 (2.2%,
C(3) Ha), 3.18–3.23 (1.5%, C(5)H), 7.46 (2.0%, H–C(6ꢀ));
7.46 (H–C(6ꢀ)) → 2.00–2.11 (1.9%, C(3)Hb), 4.26 (2.0%,
C(2)H); m/z (ESI−) 225.0988 (M− − H. C9H13N4O3 requires
225.0987).
◦
80 C for 22 h. The reaction was quenched by the addition of
H2O (20 cm3) and most of the solvents were evaporated. The
residue was diluted with EtOAc, washed with H2O, the organic
phase dried (MgSO4) and evaporated. FC (toluene–THF 7 : 1 →
4 : 1 + 0.1% NEt3) afforded compound 5 (1.45 g, 63%) as an
orange foam. Rf (CH2Cl2–EtOH 20 : 1) 0.55. dH (300 MHz;
CDCl3) −0.03, 0.02, 0.06, 0.08, 0.16, 0.20 (12 H, 6 × s, 4 ×
SiMe), 0.84, 0.86, 0.94 (18 H, 3 × s, 2 × SiC(CH3)3), 3.84–3.88
(1 H, m, CH2-OTBDMS), 4.06 (0.40 H, d, J 10.3, C(5)H),
4.23–4.34 (1.60 H, m, C(5)H, CH2-OTBDMS), 4.84 (0.60 H, s,
C(2)H), 4.95 (0.40H, s, C(2)H), 5.03–5.24 (2 H, m, CH2–C6H5),
5.59–5.62 (1 H, m, C(3)H), 7.18–7.37 (5 H, m, C6H5CH2),
7.47–7.53 (2 H, m, Bz), 7.58–7.64 (1 H, m, Bz), 7.90–8.05 (3 H,
m, C(6ꢀ)H, Bz); dC (75 MHz; CDCl3) −5.4, −5.1, −4.9 (3 × q,
4 × SiMe), 17.9, 18.5 (2 × s, 2 × (CH3)3C-Si), 25.5, 25.9 (2 ×
q, 2 × (CH3)3C–Si), 57.9, 59.0 (2 × d, C(5)), 60.5, 61.3 (2 ×
t; CH2-OTBDMS), 65.2, 65.7 (2 × d, C(2)), 66.7, 67.0 (2 ×
t, CH2–C6H5), 100.5, 100.7 (2 × d, C(3)), 127.6, 127.8, 127.9,
128.1, 128.3, 128.5, 128.6, 128.9 (8 × d, Bz, C6H5CH2), 132.3 (d,
C(6ꢀ)), 148.4, 148.8 (2 × s, C(4)); m/z (FAB+) 691.3344 (M+ +
H. C36H51N4O6Si2 requires 691.3347), 515 (53%), 423 (45%).
(2R,4S,5R)-N-(Benzyloxy)carbonyl-2-[2-(N-benzoylamino)-
4-oxo-3,4-dihydropyrimidin-5-yl]-4-hydroxy-5-hydroxymethyl
pyrrolidine 6. A solution of HF–pyridine (70% HF, 0.42 cm3,
ca. 16 mmol HF) in MeCN (7 cm3) was added dropwise to a
solution of 5 (600 mg, 0.87 mmol) in MeCN (18 cm3). After
6 h, the obtained suspension was diluted with AcOH (5 cm3)
and evaporated. The residue was suspended in a mixture of
AcOH (16 cm3) and MeCN (16 cm3), cooled to −15 ◦C and
NaBH(OAc)3 (720 mg, 3.40 mmol) was added in portions.
After 2 h additional NaBH(OAc)3 (360 mg, 1.70 mmol) was
(2R,4S,5R)-N-[(9-Fluorenylmethoxy)carbonyl]-2-[2-(N-benzoyl-
amino)-4-oxo-3,4-dihydropyrimidin-5-yl]-4-hydroxy-5-hydroxy-
methyl pyrrolidine 9. A solution of N-(9-fluorenylmethoxycar-
bonyloxy)succinimide (440 mg, 1.30 mmol) in THF (6 cm3) was
added to a suspension of 7 (218 mg, 0.66 mmol) in dioxane
(6 cm3) and 5% aq. NaHCO3 (8 cm3). After 4 h the solution
was evaporated to dryness. Most of the hardly soluble crude
product was used in the next step without further purification.
A small amount was purified for NMR characterization. It was
distributed between EtOAc and H2O. The organic phase was
separated, filtered over celite and evaporated. Product 9 was
isolated by FC (toluene–acetone–MeOH 10 : 10 : 1 → 5 : 5 :
1). Rf (CH2Cl2–EtOH 20 : 3) 0.45. dH (300 MHz; (CD3)2SO)
1.88–2.21 (2 H, m, C(3)H2), 3.44–3.76 (3 H, m, C(5)H, CH2OH),
4.15–4.28 (4 H, m, C(4)H, Fmoc-CH, Fmoc-CH2), 4.75 (1 H, s,
C(2)H), 4.96–5.10 (2 H, m, 2 OH), 7.23–8.08 (14 H, m, C(6ꢀ)H,
Fmoc-H, Bz), 11.87, 12.31 (2 × H, 2 × br s, 2 × NH); m/z
(FAB+) 553.2086 (M+ + H. C31H29N4O6 requires 553.2087), 179
(79%), 149 (100).
◦
added in portions. The suspension was slowly warmed to 0 C
and stirred for another 3 h. The mixture was then distributed
between EtOAc and H2O. The organic phase was filtered over
celite and evaporated. FC (CH2Cl2–EtOH 20 : 3 → 4 : 1)
gave compound 6 (322 mg, 79%) as a slightly yellow solid.
Rf(CH2Cl2–EtOH 20 : 3) 0.48. dH (300 MHz; CD3OD) 2.28–2.35
(2 H, m, C(3)H2), 3.79–4.01 (3 H, m, C(5)H, CH2OH), 4.39
(1 H, s, C(4)H), 4.88–5.25 (3 H, m, C(2)H, CH2C6H5), 7.16–7.40
(5 H, m, C6H5CH2), 7.58–7.62 (2 H, m, Bz), 7.68–7.83 (2 H,
m, C(6ꢀ)H, Bz), 8.04–8.07 (2 H, m, Bz); dC (75 MHz, CD3OD)
40.1, 40.9 (2 × t, C(3)), 57.3, 58.0 (2 × d, C(2)), 63.1, 63.3 (2 ×
t, CH2–OH), 68.5 (t, CH2C6H5), 70.8, 71.3 (2 × d, C(5)), 73.2,
73.5 (2 × d; C(4)), 123.5, 124.2 (2 × s, C(5ꢀ)), 129.6, 129.1,
129.3, 129.5, 129.7, 130.1, 130.2 (7 × d, Bz, C6H5CH2), 134.1,
134.7 (2 × s, Bz, C6H5CH2), 137.9, 138.1 (2 × d, C(6ꢀ)); m/z
(FAB+) 465.1773 (M+ + H. C24H25N4O6 requires 465.1774), 331
(11%), 242 (21%).
(2R,4S,5R)-N-[(9-Fluorenylmethoxy)carbonyl]-2-[2-(N-benz-
oylamino)-4-oxo-3,4-dihydropyrimidin-5-yl]-5-[(4,4ꢀ-dimethoxy)-
triphenylmethyl]-oxymethyl-4-hydroxy pyrrolidine 10. To
a
suspension of crude 9 (365 mg, 0.66 mmol) in dry pyridine
(7 cm3) was added in portions 4,4ꢀ-dimethoxytrityl chloride
(268 mg, 0.79 mmol). Further additions of 4,4ꢀ-dimethoxytrityl
chloride every 30 min were added until no starting material was
detected by TLC. The solution was diluted with EtOAc, washed
with water, dried (MgSO4) and evaporated. The crude material
was purified by FC (CH2Cl2–EtOH–NEt3 25 : 1 : 0.03 → 20 : 1 :
0.03) to give trityl compound 10 (265 mg, 47% over 2 steps) as
a white solid. Rf (CH2Cl2–EtOH–NEt3 20 : 1: 0.03) 0.47; m/z
(2R,4S,5R)-2-[2ꢀ-(N-Benzoylamino)-4-oxo-3,4-dihydropyrimidin-
5-yl]-4-hydroxy-5-hydroxymethyl pyrrolidine 7. A mixture of
6 (300 mg, 0.65 mmol) and 10% Pd/C (60 mg) in MeOH
(24 cm3) was vigorously stirred under an H2 atmosphere. After
7 h, the solution was filtered over celite and evaporated to give
O r g . B i o m o l . C h e m . , 2 0 0 5 , 3 , 1 6 5 3 – 1 6 5 8
1 6 5 7

View Article Online
(ESI−) 853.3215 (M− − H. C52H45N4O8 requires 853.3237) 631
(7%, M − Fmoc), 283 (44%).
Circular dichroism spectra
CD-spectra were recorded on a Jasco J-715 spectropolarimeter
equipped with a Jasco PFO-350S temperature controller. The
temperature was measured directly in the sample. The strand
concentration was 1.2 lM in 140 mM KCl, 7 mM NaH2PO4,
0.5 mM MgCl2 at pH 6. The samples were scanned at a speed of
50 nm min−1, band width of 1 nm, response of 1 sec and in a 210–
320 nm range at constant temperature. Each spectrum was taken
as an average of three scans using a 10 mm cell. Subsequently, the
graphs were smoothed with a noise filter. For the temperature
scan, the spectra were measured at 217 nm between 10 and 80 ◦C
with a resolution step of 0.1 ◦C and a slope of 50 ◦C h−1.
(2S,4R,5S)-N-[(9-Fluorenylmethoxy)carbonyl]-2-[2ꢀ-(N-benz-
oylamino)-4ꢀ-oxo-3ꢀ,4ꢀ-dihydropyrimidin-5ꢀ-yl)-4-[(2ꢀ-cyanoethoxy)-
(N,Nꢀ-diisopropylamino)phosphin]oxy-5-[(4,4ꢀ-dimethoxy)triphen-
ylmethyl]-oxymethyl pyrrolidine 11. To a solution of 10
(158 mg, 0.18 mmol) in dry THF (5 cm3) was added EtNiPr2
(158 mm3, 0.92 mmol) followed by 2-cyanoethyl diisopropyl-
chlorophosphoramidite (103 mm3, 0.46 mmol). After 1 h, H2O
was added and the product extracted twice with EtOAc. The
combined organic layers were dried (MgSO4) and evaporated.
FC (CH2Cl2–EtOH–NEt3 50: 1 : 0.03 → 40 : 1 : 0.03) gave
phosphoramidite 11 (172 mg, 88%) as a white foam. Rf(CH2Cl2–
EtOH–NEt3 30 : 1 : 0.03) 0.30, 0.36; dP (161.9 MHz; CDCl3)
151.04, 151.49; m/z (ESI−) 1053.4290 (M− − H, C61H62N6O9P
requires 1053.4316).
Gel retardation experiments
Non-denaturing 20% polyacrylamide gel were prepared with
a 19 : 1 acrylamide to bisacrylamide ratio. The gel and the
buffer contained TBM (90 mM Tris, 90 mM boric acid, 10 mM
MgCl2, pH 7.1). After conditioning of the gel (1 h at 120 V) and
annealing of the oligonucleotides, the samples were loaded onto
the gel (loading buffer) and electrophoresis was performed at
120 V for 18 h at 4 ◦C. Bands were detected by UV shadowing
on a TLC plate at 254 nm.
Synthesis of oligonucleotides
All oligomers were synthesized on a 1.0 or 1.3 lm scale on
an Applied Biosystems Expedite 8900 or a Pharmacia LKB
Gene Assembler Special DNA synthesizer, using standard phos-
phoramidite chemistry. Natural phosphoramidites were pur-
chased from Glen Research. The pyrrolidino phosphoramidites
were prepared as described. The 2ꢀ-deoxypseudoisocytidine
phosphoramidite was prepared as described previously.10 As
starting units, commercially available natural nucleosides bound
to CPG-solid support were chosen. The standard synthetic
procedure was used with minor modifications. The coupling
time was extended from 1.5 to 6 min for the modified monomers.
Coupling efficiencies were >97% as determined by trityl assay
with (S-benzylthio)-1H-tetrazole (0.25 M in CH3CN) instead
of tetrazole as activator. After synthesis, the oligonucleotides
were cleaved from the solid support and deprotected in conc.
NH3 at rt or 55 ◦C. The crude oligonucleotides were purified by
ion exchange-HPLC and desalted over Sep-Pak C-18 cartridges
(Waters) and were obtained in yields of 17–42% (OD 260 nm).
The purity of the oligomers was confirmed to be >95% by
RP-HPLC and their structural integrity, especially the cleavage
of the Fmoc- and benzoyl-protecting groups, was proven by
electrospray mass spectrometry. The difference between the
calculated and the measured mass was always <0.05%.
Acknowledgements
The authors would like to thank the Swiss National Science
Foundation for financial support of this project. Thanks also
goes to Dr Sabrina Buchini for precious help with gel elec-
trophoresis.
References
1 (a) D. Praseuth, A. L. Guieysse and C. He´le`ne, Biochim. Biophys.
Acta, 1999, 1489, 181–206; (b) R. V. Guntaka, B. R. Varma and K. T.
Weber, Int. J. Biochem. Cell Biol., 2003, 35, 22–31; (c) M. M. Seidman
and P. M. Glazer, J. Clin. Invest., 2003, 112, 487–494.
2 (a) S. Buchini and C. J. Leumann, Curr. Opin. Chem. Biol., 2003, 7,
717–726; (b) K. R. Fox, Curr. Med. Chem., 2000, 7, 17–37.
3 (a) S. Obika, T. Uneda, T. Sugimoto, D. Nanbu, T. Minami, T. Doi
and T. Imanishi, Bioorg. Med. Chem., 2001, 9, 1001–1011; (b) T.
Imanishi and S. Obika, Chem. Commun., 2002, 1653–1659; (c) M.
Koizumi, K. Morita, M. Daigo, S. Tsutsumi, K. Abe, S. Obika and T.
Imanishi, Nucleic Acids Res., 2003, 31, 3267–3273; (d) D. Renneberg
and C. J. Leumann, ChemBioChem, 2004, 5, 1114–1118.
4 (a) I. Pre´vot-Halter and C. J. Leumann, Bioorg. Med. Chem. Lett.,
1999, 9, 2657–2660; (b) S. Obika, Y. Hari, M. Sekiguchi and T.
Imanishi, Chem. Eur. J., 2002, 8, 4796–4802; (c) S. Buchini and C. J.
Leumann, Angew. Chem., Int. Ed., 2004, 43, 3925–3928.
5 (a) B. Cuenoud, F. Casset, D. Hu¨sken, F. Natt, R. M. Wolf, K.-H.
Altmann, P. Martin and H. E. Moser, Angew. Chem., Int. Ed., 1998,
37, 1288–1291; (b) M. J. J. Blommers, F. Natt, W. Jahnke and B.
Cuenoud, Biochemistry, 1998, 37, 17714–17725; (c) M. Sollogoub,
R. A. J. Darby, B. Cuenoud, T. Brown and K. R. Fox, Biochemistry,
2002, 41, 7224–7231; (d) V. Roig and U. Asseline, J. Am. Chem. Soc.,
2003, 125, 4416–4417.
6 S. Rhee, Z.-J. Han, K. Liu, H. T. Miles and D. R. Davies,
Biochemistry, 1999, 38, 16810–16815.
7 A. Ha¨berli and C. J. Leumann, Org. Lett., 2001, 3, 489–492.
8 A. Ha¨berli and C. J. Leumann, Org. Lett., 2002, 4, 3275–3278.
9 A. Ha¨berli, A. Mayer and C. J. Leumann, Nucleosides, Nucleotides
Nucleic Acids, 2003, 22, 1187–1189.
UV-melting curves
UV-melting curves were measured on a Varian Cary 3E UV/Vis
spectrophotometer and monitored at 260 nm. A heating–
cooling–heating cycle in the temperature range 20–90 ◦C or
0–90 ◦C was applied, with a linear gradient of 0.5 C min−1.
◦
While the heating cycles were superimposable, the cooling cycles
usually showed slight hysteresis for the association of the third
strands. Tm values were defined as the maximum of the first
derivative of the melting curves (heating ramp) using the Varian
WinUV or Microcal Origin softwares. For temperatures <20 ◦C,
nitrogen was passed through the spectrophotometer to avoid
condensation on the cuvettes. To avoid evaporation of the
solutions, 6–8 drops of dimethylpolysiloxane were added on
top of the samples in the cell. Oligonucleotides were mixed in
stoichiometric amounts by using the UV extinction coefficients
of natural oligonucleotides. The buffer (140 mM KCl, 7 mM
NaH2PO4, 0.5 mM MgCl2) was chosen to approximate the
intracellular cationic environment as closely as possible.
10 A. Mayer and C. J. Leumann, Nucleosides, Nucleotides Nucleic Acids,
2003, 22, 1919–1925.
11 (a) J. Bijapur, M. D. Keppler, S. Bergqvist, T. Brown and K. R. Fox,
Nucleic Acids Res., 1999, 27, 1802–1809; (b) K. G. Rajeev, V. R.
Jadhav and K. N. Ganesh, Nucleic Acids Res., 1997, 25, 4187–4193.
12 J. M. Dagle and D. L. Weeks, Nucleic Acids Res., 1996, 24, 2143–2149.
1 6 5 8
O r g . B i o m o l . C h e m . , 2 0 0 5 , 3 , 1 6 5 3 – 1 6 5 8